Full download ebook at ebookname.

com

Hidden Structure Music Analysis Using Computers

David Cope

https://ebookname.com/product/hidden-structure-music-
analysis-using-computers-david-cope/

OR CLICK BUTTON

DOWLOAD NOW

Download more ebook from https://ebookname.com

More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Using OpenCL Programming Massively Parallel Computers

J. Kowalik

https://ebookname.com/product/using-opencl-programming-massively-
parallel-computers-j-kowalik/

Fraud Analysis Techniques Using ACL 1st Edition David

Coderre

https://ebookname.com/product/fraud-analysis-techniques-using-
acl-1st-edition-david-coderre/

Composing music with computers 1st ed Edition Eduardo

Miranda

https://ebookname.com/product/composing-music-with-computers-1st-
ed-edition-eduardo-miranda/

Using Computers in Linguistics A Practical Guide 1st

Edition John M. Lawler

https://ebookname.com/product/using-computers-in-linguistics-a-
practical-guide-1st-edition-john-m-lawler/
Using Information Technology A Practical Introduction
to Computers Communications 9th Edition Brian Williams

https://ebookname.com/product/using-information-technology-a-
practical-introduction-to-computers-communications-9th-edition-
brian-williams/

The Hidden Wealth of Nations 1st Edition David Halpern

https://ebookname.com/product/the-hidden-wealth-of-nations-1st-
edition-david-halpern/

Mathematics and Music David Wright

https://ebookname.com/product/mathematics-and-music-david-wright/

Music and Probability David Temperley

https://ebookname.com/product/music-and-probability-david-
temperley-2/

Music and Probability David Temperley

https://ebookname.com/product/music-and-probability-david-
temperley/
DAS23CoverCope 2/28/14 1:22 PM Page 1

VOLUME 23 THE COMPUTER MUSIC AND DIGITAL AUDIO SERIES

VOLUME 23 THE COMPUTER MUSIC AND DIGITAL AUDIO SERIES

Hidden Structure
Hidden Structure

Hidden Structure
Music Analysis
Using Computers Music Analysis
David Cope
Today’s computers provide music theorists with unprecedented opportunities to analyze
Using Computers
music more quickly and accurately than ever before. Where analysis once required several
weeks or even months to complete—often replete with human errors, computers now pro-
vide the means to accomplish these same analyses in a fraction of the time and with far

Music Analysis Using Computers

more accuracy. However, while such computer music analyses represent significant improve-
ments in the field, computational analyses using traditional approaches by themselves do
not constitute the true innovations in music theory that computers offer. In Hidden Structure:
Music Analysis Using Computers David Cope introduces a series of analytical processes
that—by virtue of their concept and design—can be better, and in some cases, only accom-
plished by computer programs, thereby presenting unique opportunities for music theorists to
understand more thoroughly the various kinds of music they study.
Following the introductory chapter that covers several important premises, Hidden Struc-
ture focuses on several unique approaches to music analysis offered by computer programs.
While these unique approaches do not represent an all-encompassing and integrated
global theory of music analysis, they do represent significantly more than a compilation of
loosely related computer program descriptions. For example, Chapter 5 on function in post-
tonal music, firmly depends on the scalar foundations presented in chapter 4. Likewise, chap-
ter 7 presents a multi-tiered approach to musical analysis that builds on the material found
in all of the preceding chapters. In short, Hidden Structure uniquely offers an integrated view
of computer music analysis for today’s musicians.

Of Related Interest
The Algorithmic Composer
By David Cope
DAS 16 ISBN 978-0-89579-454-3 (2000) xiii + 130 pp. (with CD-ROM) $49.95

Hyperimprovisation: Computer-Interactive Sound Improvisation

By Roger Dean
DAS 19 ISBN 978-0-89579-508-3 (2003) xxvi + 206 pp. (with CD-ROM) $49.95

New Digital Musical Instruments: Control and Interaction Beyond the Keyboard
By Eduardo Reck Miranda and Marcelo M. Wanderley
DAS 21 ISBN 978-0-89579-585-4 (2006) xxii + 286 pp. (with CD-ROM) $49.95

David Cope
I S B N 978-0-89579-640-0
A-R Editions, Inc. 90000
8551 Research Way, Suite 180
Middleton, WI 53562
800-736-0070
608-836-9000
http://www.areditions.com 9 780895 796400

Í David Cope
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page i

HIDDEN STRUCTURE
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page ii

THE COMPUTER MUSIC AND DIGITAL AUDIO SERIES

John Strawn, Founding Editor
James Zychowicz, Series Editor

Digital Audio Signal Processing Knowledge-Based Programming for

Edited by John Strawn Music Research
John W. Schaffer and Deron McGee
Composers and the Computer
Edited by Curtis Roads Fundamentals of Digital Audio
Alan P. Kefauver
Digital Audio Engineering
Edited by John Strawn The Digital Audio Music List: A Critical
Guide to Listening
Computer Applications in Music: Howard W. Ferstler
A Bibliography
Deta S. Davis The Algorithmic Composer
David Cope
The Compact Disc Handbook
Ken C. Pohlman The Audio Recording Handbook
Alan P. Kefauver
Computers and Musical Style
David Cope Cooking with Csound
Part I: Woodwind and Brass Recipes
MIDI: A Comprehensive Introduction Andrew Horner and Lydia Ayers
Joseph Rothstein
William Eldridge, Volume Editor Hyperimprovisation: Computer-
Interactive Sound Improvisation
Synthesizer Performance and Roger T. Dean
Real-Time Techniques
Jeff Pressing Introduction to Audio
Chris Meyer, Volume Editor Peter Utz

Music Processing New Digital Musical Instruments:

Edited by Goffredo Haus Control and Interaction Beyond
the Keyboard
Computer Applications in Music: Eduardo R. Miranda and
A Bibliography, Supplement I Marcelo M. Wanderley, with a
Deta S. Davis Foreword by Ross Kirk
Garrett Bowles, Volume Editor
Fundamentals of Digital Audio
General MIDI New Edition
Stanley Jungleib Alan P. Kefauver and David Patschke

Experiments in Musical Intelligence Hidden Structure: Music Analysis

David Cope Using Computers
David Cope
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page iii

Volume 23 • THE COMPUTER MUSIC AND DIGITAL AUDIO SERIES

HIDDEN STRUCTURE

Music Analysis Using Computers

David Cope

Í A-R Editions, Inc.

Middleton, Wisconsin
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page iv

Library of Congress Cataloging-in-Publication Data

Cope, David, 1941–

Hidden structure : music analysis using computers / David Cope.
p. cm. — (Computer music and digital audio series : v. 23)
Includes bibliographical references and index.
ISBN 978-0-89579-640-0
1. Music—Data processing. 2. Musical analysis—Data processing.
I. Title.

ML74.C69 2008
781.0285—dc22
2008019352
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page v

Pulchritudo est Splendor Ordinis.

Beauty is the splendor of order.
Saint Augustine (345–430)
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page vi
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page vii

Contents

List of Figures xi

Preface xxi
Description of CD-ROM xxvii
Chapter One Background 1
Principles and Definitions 1
A Brief History of Algorithmic Analysis 7
A Brief Survey of Computational Music Analysis 22
Musical Examples 39
Program Description 41
Conclusions 43

Chapter Two Lisp, Algorithmic Information Theory, 45

and Music
Lisp 45
Definitions 50
Musical Algorithmic Information Theory 62
Musical Examples 74
Program Description 90
Conclusions 97
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page viii

Chapter Three Register and Range in Set Analysis 99

Basics of Set Theory 99
Register 112
Ranges and Vectors 118
Comparisons 120
Musical Examples 127
Program Description 139
Conclusions 144

Chapter Four Computer Analysis of Scales in 145

Post-Tonal Music
Mathematical Sequences 145
Scales 148
Vector Classes and Metaclasses 153
Varèse’s Density 21.5 160
Schoenberg’s Six Little Piano Pieces, op. 19, no. 1 (1911) 169
Other Musical Examples 172
Program Description 182
Conclusions 188

Chapter Five Function and Structure in Post-Tonal 189

Music
Object-Oriented Programming 189
Definitions 197
The Acoustic Theory of Chord Roots 201
Musical Tension 205
Context and SPEAC 210
Function 215
Form and Structure 217
Musical Examples 225
Program Description 225
Conclusions 228
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page ix

Chapter Six Generative Models of Music 231

Modeling 231
Recombinancy 234
Probabilities 249
Rules and Markov Chains 252
Musical Examples 269
Program Description 269
Conclusions 274

Chapter Seven A Look to the Future 275

Principles 275
Mathematics 277
Artificial Intelligence 290
Muse 295
Musical Examples 305
The Future 310
Conclusions 316

Bibliography 321
Glossary 331
Index 337
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page x
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xi

List of Figures

Chapter 1
Figure 1.1 From Aristoxenus, The Harmonics of Aristoxenus, trans.. Henry S.
Macran (New York: Oxford University Press, 1902), 249.
Figure 1.2 A woodcut of Boethius with an instrument designed for methodologi-
cally deriving tunings (an algorithm).
Figure 1.3 An example of Gregorian chant, Parce Domine, notated in neumes; the
second staff is a continuation of the first.
Figure 1.4 An example of medieval organum.
Figure 1.5 The Guidonian hand, in which a specific note is assigned to each part
of the hand. By pointing to a part of the hand, the conductor can indi-
cate to a group of singers which note to sing.
Figure 1.6 An example of three-voice counterpoint around the time of Tinctoris.
Figure 1.7 The division of the octave according to Zarlino.
Figure 1.8 An exercise by Beethoven for his student Archduke Rudolph based on
Fux’s Gradus ad Parnassum.
Figure 1.9 Two diagrams from Rameau’s Treatise on Harmony describing (a) the
major (perfect) triad and its inversions, and (b) the dominant (funda-
mental) seventh chord and its inversions.
Figure 1.10 First published after Mozart’s death in 1793 by J. J. Hummel in Berlin,
this musikalisches Würfelspiel instruction page shows the first two
phrases of number selections (chosen by throws of the dice) that were
keyed to measures of music. Also shown, an example page of music by
Haydn for this own musikalisches Würfelspiel.
Figure 1.11 A sample Schenker analysis of Bach’s organ prelude Wenn wir in höch-
sten Noten sein.
Figure 1.12 Charles Babbage’s first Difference Machine (1833), a forerunner of the
modern-day computer.
Figure 1.13 The author, Horizons for Orchestra, in graphic format. Time moves left
to right.

xi
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xii

xii HIDDEN STRUCTURE

Figure 1.14 From Experiment 2 from Hiller and Isaacson, Illiac Suite, using the
Illiac computer at the University of Illinois, ca. 1956.
Figure 1.15 Flowcharts for MUSANA and its analysis module.
Figure 1.16 Ten commands found in Humdrum.
Figure 1.17 Bach Chorale 002 from the Music Theory Workbench, version 0.1
(http://pinhead.music.uiuc.edu/~hkt/mtw/pdf/), by Heinrich Taube
(accessed 24 July 2007).
Figure 1.18 An example of the OpenMusic visual programming environment of mu-
sic for composition and analysis.
Figure 1.19 The first few measures of Stravinsky, Three Pieces for String Quartet
(1914), 3rd movement.
Figure 1.20 Two screens from the Sets and Vectors program on the CD-ROM that
accompanies this book.
Chapter 2
Figure 2.1 The first 2000 digits of π.
Figure 2.2 Morse’s code for English letters and numbers.
Figure 2.3 An example of the kind of three-dimensional modeling of musical style
by Böker-Heil.
Figure 2.4 Cellular automata output generated by the code in the text.
Figure 2.5 Examples of retrograde (b), inversion (c), and retrograde inversion (d)
of a motive (a).
Figure 2.6 Two matrixes of the four-note motive of Figure 2.5. The first matrix
provides the motive and its eleven transpositions (down from right
to left) along with its retrograde and its eleven transpositions (read
right to left). The second matrix provides the inversion and its eleven
transpositions along with its retrograde inversion and its eleven
transpositions.
Figure 2.7 A simple downward scale beginning on different pitches, thus produc-
ing different arrangements of whole and half steps.
Figure 2.8 Brief monophonic samplings of music by (a) J. S. Bach, Suite no. 1 in G
Major for Violoncello Solo (1720), Minuet, mm. 1–8; (b) W. A. Mozart,
Symphony no. 40 in G Minor, K. 550 (1788), 1st movement, mm. 1–5;
(c) Ludwig van Beethoven, Symphony no. 5, op. 67 (1808), 1st move-
ment, mm. 1–5; (d), Johannes Brahms, Symphony no. 1, op. 68 (1876),
4th movement, mm. 30–38; (e) Anton Webern, Variations for Piano, op.
27 (1936), 2nd movement, mm. 1– 4; (f) Ernst Krenek, Suite for Violon-
cello, op. 84 (1939), 1st movement, mm. 1–4; (g) The author, Three
Pieces for Solo Clarinet(1965), 1st movement, mm. 1–3.
Figure 2.9 Bartók, Mikrokosmos, no. 81.
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xiii

LIST OF FIGURES xiii

Figure 2.10 AIT structural analysis of the music presented in Figure 2.9.
Figure 2.11 A simple graph showing four parameters of a melody by Brahms (pre-
sented at the top of the figure).
Figure 2.12 Pitches reduced to intervals for pattern matching.
Figure 2.13 Bartók, Mikrokosmos, no. 81, from Figure 2.9, in a Dynamic Musical AIT
graph.
Figure 2.14 Multigraph output for the beginning of Stravinsky, Three Pieces for
String Quartet (1914), shown in Figure 1.19.
Figure 2.15 Two mazurkas in the style of Chopin for analytical comparison.
Figure 2.16 Two DMAIT graphs representing the Chopin mazurkas shown in Figure
2.15.
Figure 2.17 A prelude by Chopin comparable in length to the pieces in Figure 2.15.
Figure 2.18 A Dynamic Musical Algorithmic Information Theory graph of the Cho-
pin prelude shown in Figure 2.17.
Figure 2.19 Sample Multigraph output.

Chapter 3
Figure 3.1 Symbols representing several ways in which sets can be more formally
compared.
Figure 3.2 A Venn graphic of set theory logical deductions.
Figure 3.3 Root positions of triads have a smaller range between their outer notes.
Figure 3.4 A clock face arranged to allow clearer visualization of pitch-class
relationships.
Figure 3.5 The [1,6,t] pitch-class set does not resolve to the pitch-class set
[0,5,9] (9 distance when rotated to 0) because of the smaller range of
[6,t,1] (7 distance and equating to [0,4,7] when rotated to 0).
Figure 3.6 A straightforward example of the use of set theory to find similarities
in post-tonal music.
Figure 3.7 (a) [e,0,2,5,7,8] is [0,1,3,6,8,9] when rotated one number clockwise;
(b) [6,7,9,0,2,3] is [0,1,3,6,8,9] when rotated six numbers clockwise;
(c) [7,6,4,1,e,t] is [0,1,3,6,8,9] when rotated seven numbers counter-
clockwise; (d) [0,e,9,6,4,3] is [0,1,3,6,8,9] when rotated twelve num-
bers counterclockwise.
Figure 3.8 The opening 4 measures of Schoenberg, Six Little Piano Pieces, Op. 19,
no. 6.
Figure 3.9 The opening measures of Claude Debussy, La cathédrale engloutie
(The Sunken Cathedral), 1909.
Figure 3.10 A conversion chart relating decimal to duodecimal notation.
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xiv

xiv HIDDEN STRUCTURE

Figure 3.11 A more understandable and musical way to think in terms of duodeci-
mal notation.
Figure 3.12 Conversion methods used for decimal to duodecimal numbers and
vice versa.
Figure 3.13 Subtraction, addition, multiplication, and division using arbitrary
numbers in duodecimal notation.
Figure 3.14 Two two-set progressions showing pitch classes (first), t-normal form
(second), and prime (third).
Figure 3.15 Same reduction as Figure 3.14 but with registers intact in the form of
duodecimal notation.
Figure 3.16 Range notation shown to the lower right.
Figure 3.17 Arnold Schoenberg, Suite for Piano, op. 25 (1923), Gavotte, mm. 1–6,
with groupings circled and numbered.
Figure 3.18 Computer analysis of Schoenberg‘s Gavotte by tetrachords and tri-
chords as shown circled and numbered in figure 3.17.
Figure 3.19 Webern, Variations (1936), mm. 1–7.
Figure 3.20 Webern analysis by dyads (a) and trichords by hand (b) as shown
circled—dyads by solid line and trichords by dotted line—in Figure
3.19.
Figure 3.21 Nine measures from Boulez, Structures (1952).
Figure 3.22 Boulez analysis by tetrachords by hand as shown circled in Figure
3.21.
Figure 3.23 Webern, Concerto for 9 Instruments, op. 24, 1st movement, mm. 1–7.
Figure 3.24 (a) Calculations of registral information in Webern, Concerto for 9 In-
struments, op. 24, 1st movement, mm. 1–3.
Figure 3.25 Webern, Concerto for 9 Instruments, op. 24, 1st movement, last 7
measures.
Figure 3.26 Calculations of registral information from Figure 3.25.
Figure 3.27 Boulez, Second Sonata for Piano (1948), 2nd movement, beginning.
Figure 3.28 Calculations of registral information in mm. 1–6 in Boulez, Second
Sonata for Piano (1948), 2nd movement.
Figure 3.29 The author, Triplum for flute and piano (1975), beginning.
Figure 3.30 Calculations of registral information in the author’s Triplum for flute
and piano (1975), mm. 1–7.
Figure 3.31 The author, Triplum for flute and piano (1975), mm. 98–104.
Figure 3.32 Another passage from the author’s Triplum for flute and piano (1975),
m. 43.
Figure 3.33 The 97 possible origin sets for [0,2,6].
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xv

LIST OF FIGURES xv

Chapter 4
Figure 4.1 Beethoven, Piano Sonata in C Major, op. 53, beginning.
Figure 4.2 The numbers of iterations per pitch of the 298 pitches present in
Beethoven, Piano Sonata in C Major, op. 53, beginning (C, 48; C-sharp,
1; D, 26; E-flat, 11; E, 17; F, 57; F-sharp, 2; G, 52; A-flat, 16; A 13; B-flat,
31; B, 24).
Figure 4.3 A list of all 77 linear interval vector classes.
Figure 4.4 Four sets of scales related by symmetry (a, b, and c) and repeating
pattern (d).
Figure 4.5 Beethoven, Piano Sonata op. 7, mm. 97–105.
Figure 4.6 A series of possible scales, with each successive scale possibility lack-
ing one of the previous scale’s notes, the one with the shortest overall
duration. Each line here represents the scale’s transposed pitch-class
set beginning on 0, its linear interval vector, and the original pitch-
class set in ascending order.
Figure 4.7 Varèse, Density 21.5 for solo flute.
Figure 4.8 Computer scale analysis of the phrases of the three sections of Varèse,
Density 21.5 (see Figure 4.7), with author’s determination of best scale
possibility in boldface.
Figure 4.9 Computer analysis of Varèse, Density 21.5, by complete section with
author’s determination of best scale possibility in boldface and sec-
ondary possibilities marked by “*” and “†.”
Figure 4.10 Unordered but nontransposed scale pitch classes for each section in
Figure 4.7.
Figure 4.11 The first section of Varèse, Density 21.5, with non-scale tones circled.
Figure 4.12 Schoenberg, Six Little Piano Pieces, op. 19, no. 1 (1911).
Figure 4.13 Computer readouts from a three-phrase analysis of Schoenberg, Six Little
Piano Pieces op. 19, no. 1.
Figure 4.14 Computer analysis of the entirety of Schoenberg, op. 19, no. 1.
Figure 4.15 From the author’s Concerto for Cello and Orchestra (1994).
Figure 4.16 The author, Concerto for Cello and Orchestra (1994), computer analy-
sis by phrase.
Figure 4.17 Computer analysis of the excerpt from the author’s concerto in Figure
4.15, by section.
Figure 4.18 Computer analysis of the entirety of the excerpt from the author’s
concerto shown in Figure 4.15.
Figure 4.19 Selected measures of the solo cello part of the author’s concerto. The
scale used in the third movement has a related complement used
much as the dominant key might be used in a tonal concerto.
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xvi

xvi HIDDEN STRUCTURE

Figure 4.20 A scale analysis of Stravinsky, Three Pieces for String Quartet (1914),
3rd movement (opening), as shown in Figure 1.19.
Figure 4.21 A pitch field that covers a span of two-plus octaves and contains two
exclusionary hexachord pitch-class sets.
Figure 4.22 A scale with a range of two-plus octaves for the example shown in Fig-
ure 4.21.
Figure 4.23 An example where the notes C-sharp, G, and B appear only in the up-
per register and the notes E-flat, F, and A appear only in the lower reg-
ister (Arnold Schoenberg, Das Buch der hängenden Gärten, II).
Figure 4.24 The scale differences between a one-octave representation and a two-
octave representation of the music in Figure 4.23.

Chapter 5
Figure 5.1 One possible overview of the structure of a work of music that is at
one and the same time a visual/musical representation and an OOP
representation.
Figure 5.2 Examples where (a) roots remain the same but tensions change and
(b) tensions remain fairly equal but roots change, demonstrating how
neither process alone can determine function in post-tonal music.
Figure 5.3 The overtone series from the fundamental C.
Figure 5.4 (a) Interval derivations from the overtone series along with their root
designations, and (b) interval root strengths shown in order of strength.
Figure 5.5 Six groupings to serve as examples of the root-identification process.
Figure 5.6 A list of intervals weighted according to their tension levels.
Figure 5.7 Figuring interval weights from the bass note by using the chart in Fig-
ure 5.6 produces 0.3 (M3 at 0.2 + P5 at 0.1), .5 (m3 at 0.225 + m6 at
0.275), and 0.8 (P4 at 0.55 + M6 at 0.25), respectively. The minor triad
and its inversions produce 0.325 (m3 at 0.225 + P5 at 0.1), .45 (M3 at
0.2 + M6 at 0.25), and 0.825 (P4 at 0.55 + m6 at 0.275).
Figure 5.8 The augmented triad at 0.475 (M3 at 0.2 + m6 at 0.275), the diminished
triad at 0..775 (m3 at 0.225 + A4 at 0.55), the dominant seventh chord
at 1.0 (M3 at 0.2 + P5 at 0.1 + m7 at 0.7), and the diminished seventh
chord at 1.125 (m3 at 0.225 + A4 at 0.65 + M6 at 0.25).
Figure 5.9 Four chords producing tensions of 1.2 (M3 at 0.2 + P5 at 0.1 + M7 at
0.9), 1.25 (m3 at 0.225 + P5 at 0.1 + m7 at 0.7), 0.55 (M3 at 0.2 + P5 at
0.1 + M6 at 0.25), and 2.0 (M3 at 0.2 + P5 at 0.1 + M7 at 0.9 + M2 at 0.8).
Figure 5.10 Chord tensions adding to 1.0 (M2 at 0.8 + M3 at 0..2), 1.225 (m2 at 1.0 +
m3 at 0.225), 1.8 (m2 at 1.0 + M2 at 0.8), and 1.475 (P4 at 0.55 + m7 at
0.7 + m3 at 0.225).
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xvii

LIST OF FIGURES xvii

Figure 5.11 A simple lookup table of metric tensions for principal beats in twelve
different meters.
Figure 5.12 (a) J. S. Bach, Chorale no. 42, demonstrating the same chord in differ-
ing contexts; (b) Schoenberg, Six Pieces for Piano, op. 19, no. 6, open-
ing measures, demonstrating the same types of contextual differences.
Figure 5.13 SPEAC analysis of the music in Figure 5.12a and b.
Figure 5.14 An example of a unification as derived from Schoenberg, Three Piano
Pieces, op. 11, no. 1: (a) the unification, and (b) the first 8 measures of
the work from which the unification was drawn. Note the other unifi-
cations present here as well.
Figure 5.15 Two patterns, a target pattern and a potential matching pattern. If the
*intervals-off* controller were set to 2, the two patterns would
not match. However, if the *intervals-off* controller were set to
3, the two patterns would match.
Figure 5.16 A phrase of music by Mozart (a); a machine-composed replication (b);
the meta-pattern that binds them together (c).
Figure 5.17 Various levels (gradient map) of music revealed by applying filters.
Figure 5.18 The opening few beats of a graphical analysis of Stravinsky, Three
Pieces for String Quartet (1914), 3rd movement.
Figure 5.19 A root analysis of Stravinsky, Three Pieces for String Quartet (1914),
3rd movement.
Figure 5.20 SPEAC analysis (by chord change) of Stravinsky, Three Pieces for
String Quartet (1914), 3rd movement.
Figure 5.21 A structural graph of a section from the author’s Triplum for flute and
piano.

Chapter 6
Figure 6.1 (a) Mozart, Piano Sonata K. 279, 1st movement, mm. 1–3; (b) results of
the Melodic Predictor; (c) Mozart, Piano Sonata K. 533, 1st movement,
mm. 1–4; (d) results of the Melodic Predictor.
Figure 6.2 Recombination process leading to the completion of a new Bach-like
chorale phrase.
Figure 6.3 An example of differing types of internal patterns causing problems in
recombination: (a) and (b) originals, (c) a recombination without re-
gard to texture sensitivity.
Figure 6.4 An example of Experiments in Musical Intelligence output: the begin-
ning of the Rondo Capriccio for cello and orchestra arguably in the
style of Mozart.
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xviii

xviii HIDDEN STRUCTURE

Figure 6.5 An example of allusion in (a) an Experiments in Musical Intelligence

replication; and (b) a Beethoven bagatelle Op. 119, no. 1 (1820), upon
which it is partially based.
Figure 6.6 A Bach chorale as example of a jigsaw puzzle.
Figure 6.7 The Bayes rule.
Figure 6.8 A state-transition matrix for a first-order Markov chain involving all
twelve notes of the chromatic scale.
Figure 6.9 The upper left-hand corner of a second-order Markov chain state tran-
sition matrix.
Figure 6.10 (a) Original, Bartók, Mikrokosmos, no. 71, transposed to begin on G;
(b–f) five computer extensions beginning five notes from the end of
the passage.
Figure 6.11 (a) Original, Bartók, Mikrokosmos, no. 77, transposed to begin on D;
(b–d) three computer extensions beginning five notes from the end of
the passage.
Figure 6.12 A simple two-voice counterpoint in which each voice moves stepwise
in various directions.
Figure 6.13 A post-tonal example, where the various two-note groupings do not
have clear functionalism.
Figure 6.14 Two possible nonidentical computer extensions (b and c) to the origi-
nal by Bartók (Mikrokosmos, no. 80) presented in (a).
Figure 6.15 Three different forms of analysis provided by Alice.
Figure 6.16 The rule (((3 3) 2 1) ((5 1) -2 2) ((0 0) 2 3) ((8 5) -1 4)) in music
notation.
Figure 6.17 Extending voice-leading rules.
Figure 6.18 Several student-created extensions (b–d) based on Bartók (a),
Mikrokosmos, no. 81.
Figure 6.19 The rule (((19 16) -3)) in music notation.

Chapter 7
Figure 7.1 The top ten levels of Pascal’s triangle.
Figure 7.2 Pascal’s triangle shown stacked to the left and modulo 2 with the ze-
ros identifying the Sierpinski gasket (a), and zeros removed (b) to
make the graphic more readable.
Figure 7.3 A simple 5 × 5 magic square in which all horizontal ranks and vertical
columns sum to 65.
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xix

LIST OF FIGURES xix

Figure 7.4 A magic square containing intervals that equate (when added to-
gether) to 5 in top-to-bottom and left-to-right directions, along with
musical examples.
Figure 7.5 The author, Concerto for Cello and Orchestra, 2nd movement
(excerpt).
Figure 7.6 A magic cube with both incremental numbers and music intervals
(with directions).
Figure 7.7 A formal analysis of the first 89 measures of Bartók, Music for Strings,
Percussion and Celeste (Lendvai 1983, p. 74), based on the golden
mean and the Fibonacci series.
Figure 7.8 The first 14 measures of Varèse, Density 21.5, with much the same
pitch classes as the silver variation.
Figure 7.9 A simple model of a neural network.
Figure 7. 10 A simple example of an association network.
Figure 7.11 An example of Muse’s grouping process.
Figure 7.12 All of the possible permutations of a six-note list of nonrepeating num-
bers not including the original arrangement ([0,2,7] [1,5,8]) and its ret-
rograde ([1,5,8] [0,2,7]).
Figure 7.13 A view of the listener window after the completion of one analysis by
Muse.
Figure 7.14 Muse’s program includes a function called sleep.
Figure 7.15 The Eine Kleine Stück arguably in the style of Arnold Schoenberg by
Experiments in Musical Intelligence.
Figure 7.16 Schoenberg, Three Piano Pieces, op. 11, no. 1 (opening).
Figure 7.17 The first page of the score to Beethoven, Symphony no. 5, revision.
Figure 7.18 Dies irae, a medieval Latin sequence and the first words of the Re-
quiem Mass.
Figure 7.19 (a) A passage from the Mystic Circle of the Young Girls in Stravinsky,
The Rite of Spring ; (b) a passage from Stravinsky, Three Pieces for
String Quartet.
Figure 7.20 (a) A passage from the 3rd movement of Stravinsky, Three Pieces for
String Quartet; (b) a similar passage in the Symphonies of Wind Instru-
ments.
Figure 7.21 A post-tonal work by the author resembling Stravinsky’s style.
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xx
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxi

Preface

Today’s computers provide music theorists with unprecedented opportunities to

analyze music more quickly and accurately than ever before. Whereas analysis once
required several weeks or even months to complete and was often replete with hu-
man errors, computers now provide the means to accomplish these same analyses
in a fraction of the time and with far more accuracy. However, although such com-
puter music analyses represent significant improvements in the field, computational
analyses using traditional approaches do not in themselves constitute the true inno-
vations in music theory that computers offer. In this book, I introduce a series of an-
alytical processes that—by virtue of their concept and design—can better, and in
some cases only, be accomplished by computer programs, presenting unique oppor-
tunities for music theorists to better understand the music they study.
As an example, when comparing computational and human analysis, we need to
distinguish between time differences of minutes or hours and of months or years.
Analyses that may take a human minutes, hours, or days may still be undertaken
by hand. Analyses that require a human months or even years may not even be
considered feasible. Without these latter analyses, however, our range of choices of
analytical processes becomes arbitrarily limited. With computers able to reduce
our months or years to fractions of seconds and accuracy to near perfect, music
analysts must rethink many theories that they may have previously dismissed as
impossible due to the magnitude of the undertaking. Computers can not only re-
duce our fatigue and increase our accuracy, but they can open whole new worlds
previously considered unachievable—if considered at all.
Many readers of a book like this might value an annotated bibliography of cur-
rently available software for tonal and post-tonal music analysis, but most would
find this kind of listing woefully out of date the very day it became available; the
field simply changes too rapidly to make this kind of offering practical. Instead, this
book—with the exception of the broad introductory first chapter—focuses on sev-
eral unique approaches to music analysis offered by computer programs. Although
these approaches do not represent an all-encompassing and integrated global
theory of music analysis, they do represent significantly more than a compilation of
loosely related computer program descriptions. Chapter 5, on function in post-
tonal music, for example, firmly depends on the scalar foundations presented in

xxi
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxii

xxii HIDDEN STRUCTURE

Chapter 4. Likewise, the logic of Chapter 7’s presentation of a multi-tiered approach

to musical analysis depends on the material of all of the preceding chapters. In
short, this book represents an integrated view of computer music analysis.
Hidden Structure also centers on post-tonal rather than tonal music for several
reasons. First, as we shall see in this chapter, tonal music analysis has a long and
distinguished history that details approaches to melody, harmony, counterpoint,
form, and structure. Although the advent of computer technology can certainly add
tools and allow the discovery of new concepts, the range of potential for truly new
approaches is, I feel, somewhat limited compared to the potential for analytical
possibilities in post-tonal music. Second, many diverse and useful programs for
tonal music analysis already exist. In fact, most early computer computational ex-
periments with music analysis involved tonal music. On the other hand, although
many computer pitch-class set programs exist, little beyond this easily created soft-
ware is currently available. Third, and possibly most important, post-tonal music
represents the lingua franca of today’s concert music, the music in which I am most
interested and to which I wish to devote my research time.
The ideas described in this book are ordered from simple to complex, and from
more traditional to possibly more innovative. Chapter 2’s focus on information
theory mirrors the work done by early music theorists using computers, while
Chapter 6’s concentration on computer modeling parallels more recent work in
computer composition using rule acquisition. Most of these ideas also share com-
mon perspectives. For example, they have the same post-tonal focus, use similar
artificial-intelligence approaches, and share the same computational foci. Thus, this
book represents an integrated view of music analysis rather than a potpourri of
interesting but unrelated analytical perspectives.
Although this book is meant for a general audience, with the only prerequisite
being the ability to read music, it is particularly intended for both musicians inter-
ested in the application of computers to the understanding of music and computer
scientists—both professionals and those amateurs who enjoy the sport of benign
hacking—wishing to use their expertise to better understand music. Given my sus-
picion that many of these individuals do not share a great deal of common ground,
I have attempted to educate each group in the territory of the other. I therefore
define the terms and concepts used very carefully. For example, music theorists
will require an understanding of information theory, programming, and so on. Like-
wise, non–music theorists will require an understanding of musical set theory,
chord roots, and so on. By doing this, however, I fear that both will consider many
of the chapters here uneven, at least in the sense that computer scientists, for ex-
ample, understand set theory, just not musical set theory, and vice versa for musi-
cians. Thus, readers may find themselves in shallow water for one section of a
chapter and then suddenly in deep water. I apologize for the lack of uniformity that
may result. Hopefully, the chapter subheadings will allow those already familiar
with one or more of the concepts described to skip to the next section relevant to
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxiii

PREFACE xxiii

their interests if they so desire. I believe that those who stick with this book, despite
its bumpy ride, will gain a new appreciation for the views of music offered here.
Chapter 1 begins with principles and definitions, followed by a survey of histori-
cal algorithmic analysis processes and a broad overview of the computer music
programs in use today. This overview does not just list the many different imple-
mentations of particular analytical processes, but describes representative exam-
ples of them. For example, pitch-class set analysis programs in Java can be found at
dozens of Internet sites at present, the variations between them limited primarily
to differences in graphical user interfaces. Therefore, the coverage here is relatively
limited. Other programs, such as the Humdrum Toolkit, that provide more univer-
sal algorithms are discussed in more detail. Chapter 2, on music and algorithmic
information theory, describes how a form of compression (replacing redundant ma-
terial with signifiers to conserve space) can reveal important statistics about post-
tonal music. It also proposes an analytical technique that incorporates a dynamic
integrative approach to aspects of pitch, rhythm, texture, and dynamics in an at-
tempt to understand the constantly changing foci of musical works.
Chapter 3, on set analysis of register and range, introduces a duodecimal nota-
tion for representing pitch-class sets, a notation that, by virtue of its having a base
(radix) of twelve, allows both register and pitch class to share one relatively simple
notation. Integrating these two interrelated aspects of musical pitch in a pitch-class
abstraction enables the comparison of equivalent prime-form sets with significant
registral differences, revealing contrasts as significant as those between prime-form
sets that share very similar registral arrangements. Chapter 4 presents a computer
analysis of musical scales and describes a method for generating all possible equal-
tempered scales and an approach to efficiently discovering and cataloging these
scales, particularly in post-tonal music. Because scales have scale degrees that in
tonal music signify musical functions, the concept of post-tonal scale analysis
has far-reaching implications. This is the subject of Chapter 5, which more fully
explores the implications of discovering function in post-tonal music, defining and
analyzing potential musical progressions, cadences, chromatic harmonies, hierar-
chies, and so on.
Chapter 6 focuses on generating rules from music itself, rather than imposing
user-prescribed rules. I end in Chapter 7 with a possibly brazen (but hopefully use-
ful) look to the future at some of the areas of musical analysis I feel will develop
over the next few decades thanks to the availability of new computational tools.
I have chosen commonly analyzed musical examples for this book so that read-
ers will be able to better compare the results of the analytical processes I describe
here with the results of other, more standard analytical techniques. I have also in-
terpolated examples of my own music here and there in order that readers may
more directly ascertain how the techniques described further prove valuable in the
creative process. I explicitly used the analytical processes described in these com-
positions, and these works represent the best cases I can prove of such usage.
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxiv

xxiv HIDDEN STRUCTURE

I have limited the scope of study for this book to classical post-tonal music.
Other genres of music might serve equally well. However, classical post-tonal music
provides a comprehensive range of music over a significant historical period. As
well, my own background consists almost exclusively of classical tonal and post-
tonal music, and hence I lack the expertise to intelligently discuss other genres of
music. Readers may offset this shortcoming by applying the techniques defined
and described here to whatever style of music they know best.
The various computer programs discussed in this book, along with MP3 versions
of all of the book’s musical examples, are available on the CD-ROM that accompa-
nies this book. The programs are written in Common Lisp, of which two versions
are available: (1) Macintosh platform, or (2) any platform that supports Common
Lisp. To ensure that these latter programs will perform in different environments
requires the omission of platform-dependent code such as all Musical Instrument
Digital Interface (MIDI) and graphical user interface (GUI) functions. Full documen-
tation and operating instructions are included with each program.
As time permits, I will provide code updates for new versions of the Macintosh
operating system on my Web site (arts.ucsc.edu/faculty/cope). However, if history
proves accurate, as soon as I write new platform-dependent programs, system
hardware or software changes will once again render the code almost immediately
obsolete. Furthermore, the software for the programs in this book, although very
helpful in demonstrating the principles of each chapter and in clarifying the ana-
lytical principles proposed, is not critical to the understanding of the material
presented.
Whenever possible in this book, I have included the thoughts of music theorists,
musicologists, mathematicians, computer scientists, and cognitive scientists whose
work complements or poses the greatest challenges to the ideas presented here.
My apologies to those whose work may seem relevant, but to which I have not re-
ferred here due to space limitations.
Many individuals have advised me in this study of computer music analysis, par-
ticularly graduate students and colleagues at the University of California at Santa
Cruz (notably Paul Nauert, Ben Carson, and Daniel Brown). I also owe immense
gratitude to Nico Schüler for forwarding a copy of his dissertation (2000), without
which I would have floundered more than I already have in the first chapter’s his-
tory of computer music analysis. Keith Muscutt, Eric Nichols, and Irene Natow pro-
vided much-needed advice (editorial and otherwise). Many of the ideas in this book
originated from my teaching over the years, and I thank the many classes of stu-
dents who acted as guinea pigs for my theoretical explorations. Without support
from colleagues and students such as these, this book could not have been
completed.
Hidden Structure describes a few of the ways in which I believe computer pro-
grams can contribute significantly to the analysis of music, particularly to the often
difficult-to-understand post-tonal music of our time. Although computer programs—
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxv

PREFACE xxv

through their incredible speed and accuracy—certainly aid our ability to analyze
this music in more or less traditional ways, creating programs that analyze music
using more indigenous computational approaches offers greater challenges and po-
tential rewards. I hope that my descriptions of several such programs here will en-
courage others to continue to seek newer and more revealing analytical processes
in the future.

Acknowledgements
The author and publisher are grateful for the use of the following material:

Figure 1.14 Theodore Presser Company

Figure 1.15 Permission granted by Nico Schuller.
Figure 1.17 Bach Chorale 002 from the Music Theory Workbench version 0.1
(http://pinhead.music.uiuc.edu/~hkt/mtw/pdf/), Heinrich Taube, author.
Figure 1.18 Permission granted by IRCAM.
Figure 1.19 © Copyright 1923 by Hawkes & Son (London) Ltd
Reproduced by kind permission of Boosey & Hawkes Music Publisher Ltd
Figure 2.9 © Copyright 1940 by Hawkes & Son (London) Ltd
Figure 2.15 Reproduced by permission of Spectrum Press (spectrumpress.com)
Figure 3.8 © 1913, 1940 by Universal Edition A.G. Wien/UE 5069
Figure 3.17 Used by permission of Belmont Music Publishers, Los Angeles
Figure 3.19 © 1937 by Universal Edition A.G. Wien/UE 10881
Figure 3.21 © 1955 by Universal Edition (London) Ltd. London/UE 12267
Figure 3.23 © 1948 by Universal Edition A.G. Wien/UE 11830
Figure 3.25 © 1948 by Universal Edition A.G. Wien/UE 11830
Figure 3.27 Heugel and Co. 1950
Figure 3.29 1975 Carl Fischer Music Publishers
Figure 3.31 1975 Carl Fischer Music Publishers
Figure 3.32 1975 Carl Fischer Music Publishers
Figure 4.7 G. Ricordi and Co.
Figure 4.11 G. Ricordi and Co.
Figure 4.12 © 1913, 1940 by Universal Edition A.G. Wien/UE 5069
Figure 4.23 Dover Publications 1995
Figure 5.12b © 1913, 1940 by Universal Edition A.G. Wien/UE 5069
Figure 5.13b © 1913, 1940 by Universal Edition A.G. Wien/UE 5069
Figure 5.14b Belmont Music Publishers, Los Angeles 1910
Figure 5.19 © Copyright 1923 by Hawkes & Son (London) Ltd
Reproduced by kind permission of Boosey & Hawkes Music Publisher Ltd
Figure 6.5 Reproduced by permission of Spectrum Press (spectrumpress.com)
Figure 6.10a © Copyright 1940 by Hawkes & Son (London) Ltd
Figure 6.11a © Copyright 1940 by Hawkes & Son (London) Ltd
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxvi

xxvi HIDDEN STRUCTURE

Figure 6.14a © Copyright 1940 by Hawkes & Son (London) Ltd

Figure 6.18a © Copyright 1940 by Hawkes & Son (London) Ltd
Figure 7.7 Lendvai, Ernö. 1983. The Workshop of Bartók and Kodály. Budapest: Editio
Musica.
Figure 7.8 G. Ricordi and Co.
Figure 7.15 Reproduced by permission of Spectrum Press (spectrumpress.com)
Figure 7.16 Belmont Music Publishers, Los Angeles 1910
Figure 7.19 © Copyright 1923 by Hawkes & Son (London) Ltd
Reproduced by kind permission of Boosey & Hawkes Music Publisher Ltd
Figure 7.20 © Copyright 1923 by Hawkes & Son (London) Ltd
Reproduced by kind permission of Boosey & Hawkes Music Publisher Ltd
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxvii

Description of the
CD-ROM

The CD-ROM that accompanies this book contains materials that augment the text
of Hidden Structure.

MP3S ON THE CD-ROM

The CD-ROM contains MP3s of all of the musical examples presented in this book,
labeled according to figure number. These MP3s are machine performed using MIDI
(Musical Instrument Digital Interface) files and Macintosh Quicktime Musical Instru-
ments. The results are thus mechanistic in performance and relatively shallow in
timbral quality and are not intended as masterful performances. They simply pro-
vide simple aural replications of the figures in the book for those unable to play or
otherwise hear them.

SOFTWARE ON THE CD-ROM

The software found on the CD-ROM includes all of the programs described in this
book. Details on how to operate this software are found below, in the chapters of
this book, and in the files containing the software. Triangulation of these three
sources should provide readers who lack a background in computer programming
with the ability to run the code.
The software found on the CD-ROM is written in Common Lisp in source code
form (readable as opposed to object code; unreadable compiled binary code). The
source code requires Common Lisp in order to function. For programs labeled (CL),
any form of Common Lisp (of which there are many free available forms for use on
any computer platform) may be used. For such programs, simply download any
Common Lisp, load the program into the CL program by any of several methods
(see the CL program itself for a description of these processes), and then follow the
instructions provided at the top of each source code file, which is readable in any
text program. For programs labeled MCL (Macintosh Common Lisp), you must
download a form of MCL that works with your particular computer (see the digi-
tool.com Web site for more information) and follow the instructions provided.
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxviii

xxviii HIDDEN STRUCTURE

The source code consists of a series of variable declarations (preceded by “def-

var”) and function definitions (preceded by “defun”) enclosed in various parenthe-
ses that help connect variables and processes. Objects (denoted by “defclass”)
define windows, menus, and so on. No knowledge other than this general prescrip-
tion is required to run the code provided here.
I created the software here primarily for demonstrating the principles expressed
in the associated chapter of this book. All of the program runs in the figures, for ex-
ample, result from using this software. However, the code is not bulletproof. In
other words, loading some music (e.g., performed music) may create problems for
some of the functions and thus cause programs themselves to fail. Ensuring that
code can endure any input not only would take hundreds of hours of testing and
even then be incomplete, it would also require several times the amount of code
provided here, making the result unreadable except by experts. Problems encoun-
tered with these programs should be reported to the author at howell@ucsc.edu.

SOFTWARE ON THE CD-ROM BY CHAPTER

Chapter 1
Sets (CL/MCL)
Visualize (MCL)
Chapter 2
Comparison (CL)
Multigraph (MCL)
Chapter 3
SetMath (CL)
Set Multiples (CL)
Set Database Analysis (CL)
Register (CL)
Chapter 4
Scale Analysis (CL)
Chapter 5
Root (CL)
Structure Map (CL)
Structure Graph (MCL)
Chapter 6
Markov (CL)
Extend (CL/MCL)
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxix

DESCRIPTION OF THE CD-ROM xxix

Chapter 7
Neural Net (CL)
SPEAC (MCL)
Muse (CL)
01_DAS_FM_ppi-xxx 1/29/09 1:06 PM Page xxx
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 1

ONE
Background

Throughout the centuries, the arts have undergone transformations that paral-
leled two essential creations of human thought: the hierarchical principle and
the principle of numbers. In fact, these principles have dominated music, partic-
ularly since the Renaissance, down to present-day procedures of composition.
Iannis Xenakis, Formalized Music (1971), 204

This chapter begins with definitions of the principles and terms that act as founda-
tions for the ideas, processes, and programs to follow. It continues with a survey of
historical algorithmic music analyses (often called paper algorithms) and a brief
look at the computer music programs in use today and some representative exam-
ples of particular analytical processes. It concludes with a description of two sim-
ple programs, one for visualizing music of all types and the other for delineating
pitch-class sets, particularly in post-tonal music.

PRINCIPLES AND DEFINITIONS

As I mentioned in the preface, Hidden Structure does not build a single theory of
post-tonal music, but rather describes several diverse methodologies for computer
analysis. However, this book does focus on four central principles:
1. all music consists of patterns;
2. all pitch patterns can be reduced to scales;
3. all elements of scales have different functions; and
4. all patterns, scales, and functions in music are best understood by modeling
their processes.

The first three principles listed here seem plain enough. The last, however, de-
serves elaboration. The term “modeling” as used here refers to the building of
accurate replicas of objects or phenomena in order to better understand them.
The concept of modeling as an alternative to reverse engineering, though not new,

1
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 2

2 HIDDEN STRUCTURE

has—with the advent of computational technology—gained a significant following

in recent years. For example, Stephen Wolfram (2002) has argued that traditional
scientific methodology is like a salmon losing ground against a strong current, wag-
ing a futile battle in its attempts to answer the important scientific questions of our
day. He further argues that the only true revelations of the universe will come from
computationally modeling it. Without either agreeing or disagreeing with Wolfram’s
point of view on science, I argue vigorously for the usefulness of modeling as an ap-
proach to musical analysis, as later chapters in this book will testify.
Each of the chapters in this book relates to all of the above principles to some
extent. However, Chapters 1, 2, 3, 6, and 7 concentrate particularly on Principle 1,
Chapters 4 and 7 on Principle 2, Chapters 5 and 7 on Principle 3, and Chapters 6
and 7 on Principle 4. These principles then provide a framework on which a more
general theory might be developed. As the processes described in this book unfold,
I will remind readers of how the material under consideration exemplifies these
four principles.
Because the word atonal is somewhat vague and its invocation often has deroga-
tory inferences (see Straus 1990, v), I use the term post-tonal here instead. There
are several current definitions of this word, ranging from “not explicitly tonal” to
“twelve-tone” or “serial.” In this book, I will use a more liberal definition. This
means that terms such as polytonal, octatonal, and serial all fall under the umbrella
term post-tonal. This broad definition will make it unnecessary to differentiate be-
tween terms for music by, for example, Igor Stravinsky, Béla Bartók, Edgard Varèse,
Arnold Schoenberg, and Anton von Webern.
The terms music theory and music analysis seem interchangeable to some. In this
book, however, I define music theory as a more or less descriptive account of the
principles of musical structure, and music analysis as the more or less factual ac-
count of what literally occurs in music itself. Anthony Pople has it right: “There is a
broad historical distinction between music theory—which studies musical works in
order to deduce ‘more general principles of musical structure’—and music analy-
sis, in which the interest is focused on individual pieces of music” (Pople 2004,
127). Therefore, whenever I refer to music theory, I am speaking of basic tenets that
govern all music. When I refer to music analysis, I am speaking of specific instances
of those principles in a more or less restricted body of music. For the most part, if
not stated otherwise, this book deals with music analysis rather than music theory.
To some extent, then, music analysis as referred to here resembles cryptogra-
phy, or code breaking. As Nicholas Cook puts it:

Music is a code in which the deepest secrets of humanity are written: this heady
thought assured musical studies their central place in ancient, medieval and re-
naissance thought. And though the study [analysis] of music no longer occupies
quite so elevated a role in intellectual circles, some of today’s most important
trends in the human sciences still owe it a debt. (Cook 1987, 1)
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 3

BACKGROUND 3

We have already partially broken many such musical codes; after all, the discovery
of function in tonal music, along with its ancillary, harmonic syntax, represents
hundreds of years of cryptographic investigations. Yet, in the view of some ana-
lysts, more code breaking remains. For example, no convincing corollary has yet
been found for function in post-tonal music, nor has a clear process for defining
musical structure been revealed in post-tonal music, nor has a defined syntax been
presented for most of the post-tonal music of the twentieth and twenty-first cen-
turies. Even tonal music holds many mysteries yet undiscovered. Hopefully, the
code breaking that remains can be enhanced with the aid of computers.
The term algorithm will be used often in this book and requires a clear defini-
tion. An algorithm (see Cope 2000, 1ff., for a more complete definition) is a recipe
for achieving a goal in a finite number of clearly defined steps. For example, the in-
structions “extract the dysfunctional bulb by turning it counterclockwise, insert
new bulb, turn new bulb clockwise several times until tight, and turn on the light”
constitutes a simple algorithm for replacing a burned-out light bulb. Deoxyribonu-
cleic acid (DNA) replication, blinking, breathing, and so on also represent algo-
rithms. Note that algorithmic analysis does not depend on computer hardware or
software.
The most interesting example of the notion of an algorithm is Alan Turing’s 1952
experiment with computer chess. Turing—the oft-credited father of artificial
intelligence—had so tired of waiting to challenge the first computer chess-playing
program that he invited a friend to play a game of computer chess, with Turing as-
suming the role of computer. Turing spent days creating a paper algorithm, a list of
rules required to respond to whatever imaginable move his opponent might make.
Turing and his opponent’s subsequent game lasted three or so hours, during which
Turing rigorously followed the rules he had compiled. Turing eventually lost the
game. He argued, however, that he lost because his program, called Turochamp,
had not been complete or accurate enough. Thus, his point was made—by follow-
ing his algorithm exactly he had fulfilled his role as computer. If a human computer
could play chess algorithmically, then certainly a hardware computer following
human-programmed rules could do so as well (see Standage 2002, 226–229, for
more information).
Obviously, however, a computer can play such games only as well as its human-
provided rules allow it to play. My point here is simple: computers simply obey the
dictates of their programmers; they are limited only by the programmer’s skill in
developing the right algorithm. Computers are tools, nothing more and nothing
less. Thus, although computers process data faster and more accurately than hu-
mans, algorithms for analyzing music have existed for centuries.
In tonal music, using dominant-tonic cadences, avoiding parallel fifths, resolving
dissonances, creating formalistic canons and fugues, and other creative processes
result from algorithms that composers invoke—whether they realize it or not—in
order to plan their compositions and participate in the musical style of their day.
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 4

4 HIDDEN STRUCTURE

One could also argue the algorithmic foundation of much post-tonal music on the
basis of the revelations of set theory alone. Serial music requires less substantia-
tion. Even Cage’s indeterminate music often relies on rigorously applied algorithms.
In short, algorithms play significant roles in all music.
Interestingly, the traditional processes of tonal analysis that pervades music
academia today represent an algorithm. The rules by which students reduce Bach
chorales to registerless triads, reveal their inversions, and ultimately use roman
and arabic numerals for describing functions and their inversions within keys
clearly denote a set of instructions to achieve a particular goal—the standard defi-
nition of an algorithm.
Because computers offer the fastest, most accurate, and most efficient manner
of applying algorithms to data, using computers for music analysis is a natural con-
sequence of these ideas. Indeed, computers can process information exponentially
faster than humans. For the most part, computers offer users the ability to extend
themselves in ways that no other means can match. Given the right algorithm,
then, computers present the best opportunity for music analysts to understand the
music of the future and to better understand the music of the present and the past.
Before beginning a brief history of related music research and analysis, and be-
cause Hidden Structure conjures up very different meanings to different people, I
would like to preface my historical survey with what I hope is a clear definition of
what the title means for this book. Before stating this definition, however, I feel
obliged to articulate what the title does not mean, along with my reasons for ex-
cluding such definitions.
Many readers, composers in particular, may take the title Hidden Structure: Music
Analysis Using Computers to mean analysis of musical timbre. After all, most jour-
nals dealing with computer music composition (Computer Music Journal, Perspec-
tives of New Music, Journal of New Music Research, etc.) routinely publish articles
devoted to constructing and deconstructing musical sound, called synthesis. I sus-
pect that were one to take a survey, the great majority of books and articles cover-
ing computer music analysis do so from the perspective of the analysis of timbre,
rather than from the perspective of the analysis of compositional technique. Analy-
sis of musical timbre, however, will not appear in this book, because the analysis to
which I refer in its title covers the relationships between sounds and not the
sounds themselves.
The title of this book may indicate to some the computer analysis of musical per-
formance. Creating software to perform in the style of human performers has long
been a focus of computer music research. Indeed, the performer’s nuances of
rhythm, meter, dynamics, articulation, and other parameters are a critical compo-
nent of human appreciation of music, and their analysis would seem vitally impor-
tant. Because this book can only cover so much territory, however, I have opted to
not include computer music analysis of performance. In fact, including perfor-
mance analysis would most likely obscure many of the points I wish to make.
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 5

BACKGROUND 5

Computers have also proven valuable in the translation of notated music (either
in autograph or printed form) into digital data for computer performance, engraved-
quality scores and part printing, storage, and other forms of digital coding. Such
digitizing, however, though extremely important, does not in itself provide any true
analysis of music and therefore does not represent a logical subject for this book.
Music perception is an absolutely critical component to understanding music,
and, as such, one might expect to find music cognition an actual cornerstone of a
book titled Hidden Structure: Music Analysis Using Computers. My reasons for exclud-
ing it here are many, but two are salient. First, studies of music cognition have only
recently begun, and many of the initial results conflict with one another to the
point of seriously confusing the current state of research. Second, this book is
based, at least in part, on the assumption that the analytical process discussed can
be programmed. Research in music cognition often results from studies relating to
statistics of how human subjects react to musical input, rather than objective—and
thus programmable—principles and data.
Readers interested in the analysis of electronic music (see Simoni 2006), impro-
vised music, music of oral traditions, and other such topics will also be disap-
pointed, because this book will not shed much light on these important topics. At
the same time, most of the principles described here can be applied to these con-
nected areas of research. I leave such application, however, to other, more qualified
individuals.
What, then, does the title Hidden Structure: Music Analysis Using Computers actu-
ally mean? In brief, this book focuses on the analytical study of Western classical-
music notated scores that can be represented digitally. Although a number of
individuals have attempted to quantify important aspects of music such as perfor-
mance practice (Lawson and Stowell 1999; Rink 1995) and other nonnotatable infor-
mation that composers may have intended beyond the actual score, printed music
remains the most reliable resource available for pitch, rhythm, dynamics, and other
of its parameters. Thus, I have limited the investigations presented here to notated
scores only.
Given my exclusion of several very important aspects of music, my actual defini-
tion of computer music analysis for this book may seem to many readers quite lim-
ited. This is intentional. The programs that accompany the chapters on this book’s
associated CD-ROM require such limitations. My work here is intended not to con-
jecture, but to prove. In order to provide these proofs, I must divide and conquer—
a mantra I repeat ad nauseam to my students. Hidden Structure thus will, I hope,
provide useful grist for others interested in understanding music, particularly mu-
sic of the twentieth and twenty-first centuries.
This book’s original title, “Algorithmic Music Theory,” better describes the his-
tory that follows this definition of the subject. At least half of this history covers
theories of music prior to the advent of computers, theories that nonetheless share
a single important process—the algorithm. Thus, despite the revised title, I ask all
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 6

6 HIDDEN STRUCTURE

readers to consider much of this chapter’s contents in particular and the historical
sections’ contents in general as guided by my original title, focusing on algorithms
rather than on the computer’s ability to process data more quickly and more accu-
rately than humans.
Before progressing to subsequent sections of this chapter, I should mention a
few other caveats. Throughout the book, the processes described focus almost ex-
clusively on pitch. I include rhythm (Chapters 2 and 5, for example), timbre, dy-
namics, articulations, and so on, but only occasionally. I do not intend this pitch
bias to suggest that pitch represents the nadir of musical parameters, but rather to
allow development of one aspect of music more thoroughly. I would rather not dis-
cuss, say, rhythm at all than discuss it shallowly just to ensure balance.
I also do not consider composer intent a factor in the music analysis discussed
here. Although music analysts should not ignore intent, especially when the com-
poser has made this intent public, composers are often unaware of many of the
processes they use while composing. To ignore these other processes while focus-
ing on the processes the composer has claimed as central to a composition would
render analysis practically impotent. In short, at least for this book, composer in-
tent will not take precedence over any other logical process that may be equally or
more revealing.
Finally, although listening should inform analysis, the reverse should also be
true. In fact, because I cannot predict what readers will hear in any of the examples
in this book, I take the viewpoint that the analyses presented here inform listening
and not vice versa. Listeners are biased by the music that they have heard, by their
own unique biological and chemical constitutions, by cultural and other learned
musical habits, and, of course, by personal musical sensibilities and aesthetics.
For millennia, astronomers and other celestial observers assumed that the sun
and stars revolved around the earth. However, in the Renaissance, three noted as-
tronomers, Nicolaus Copernicus (1473–1543), Johannes Kepler (1571–1630), and
Galileo Galilei (1564–1642), independently formulated, developed, and defended a
very different concept of the heavens: heliocentrism, the idea that the earth and
other planets revolved around the sun. So objectionable was this new idea to most
of the Western world that Galileo suffered severe scientific, social, and, in particu-
lar, religious persecution. Today, of course, we know that Copernicus, Kepler, and
Galileo were correct, and their theories have subsequently been applied to star
clusters, galaxies, and even clusters of galaxies called superclusters. In this in-
stance, our original analytical approach—observing the sun and stars move across
our sky—was actually an impediment to understanding the real physics involved.
Although the original theories seem almost silly today, they can provide a very im-
portant lesson for music analysts: analytical methods for better understanding mu-
sic, no matter their apparent worth in the short run, may in fact actually prevent
our ability to really understand the music we study. Although I do not pretend to
know the truth of this possibility, or, if true, a better way to proceed, I profoundly
believe that alternate ways to analyze music should not be ignored, no matter how
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 7

BACKGROUND 7

improbable they may seem. I hope that readers will bear these thoughts in mind as
the book unfolds.

A BRIEF HISTORY OF ALGORITHMIC ANALYSIS

Almost all music analysis is algorithmic in the sense that it compares musical
processes in a work under study to a corpus of known rules. I use two strategies to
focus the following survey on particular algorithms: choosing only those algorith-
mic processes that have made significant and quantifiable changes in the manner
in which music analysts analyze music, and wherever possible pointing readers in-
terested in pursuing more information on those processes I do mention—or even
processes that I do not explicitly mention but to which I implicitly refer—toward
important general and specific sources. My use of the word algorithm in precompu-
tational analysis also relates to analyses that are clearly programmable in some
meaningful way. It is in this spirit, then, that I begin the history of computer music
analysis long before the advent of actual computational hardware.
Furthermore, the following narrative in no way fully covers the subject of algo-
rithmic music theory, which itself could fill a very large book. Indeed, if my premise
is true—that most music analysis is algorithmic—then most volumes on the gen-
eral subject of the history of music analysis written in the last millennium them-
selves cover this subject. Therefore, particularly because history is not my primary
subject here, I include this brief outline in order to provide a broad context for the
discussion that follows and, hopefully, direct readers toward many sources that
they themselves can and should research further.
Note how, as the following historical outline proceeds, the reference to three of
the four principles described at the outset of this chapter follow in incremental or-
der, with early analytical approaches involving issues of pattern (shape, repetition,
cadence, etc.), medieval studies interested in scale (modes, etc.), and eighteenth-
and nineteenth-century analysis concerned with function (chord roots, etc.). The
section that follows that will then deal with a similar incremental ordering, but,
with the advent of computer analysis, include Principle 4—modeling in addition to
reverse engineering.
Although Pythagoras (582–500 BCE; see Cazden 1958), Plato (427–347 BCE, espe-
cially The Republic), and Aristotle (384–322 BCE, especially The Politics) postulated
many new ideas about the theory and analysis of music, most of these ideas were
philosophical or mathematical in regard to tuning (the ratios of Pythagoras princi-
pally) and thus suggest algorithmic frequency determination not particularly re-
lated to the purpose of this history (Winnington-Ingram 1929). Aristoxenus (b. 354
BCE), in his two major books Harmonics and Rhythmics, attempted to define inter-
vals (Principle 1), scales (Principle 2), keys, melody, and consonance, as well as
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 8

8 HIDDEN STRUCTURE

various related ideas (Winnington-Ingram 1932). Although Aristoxenus avoided

whenever possible defining his subjects in terms of tuning ratios—a very popular
approach during this period—he was also an astronomer, and he found it advanta-
geous to use numerical measurements when describing musical phenomena such
as intervals, scales, and consonance. Figure 1.1 presents Aristoxenus’s division of
the tetrachord into thirty-two equal parts, each part being a twelfth of a tone. An
early twentieth-century translation by Macran (1902) presents a series of enhar-
monic graphs representing the types and subtypes of the intervals and scales de-
scribed more fully in the book itself.
Without the groundwork established by Aristoxenus, Claudius Ptolemy (second
century CE, especially his Harmonics) could not have described the full comple-
ment of the Greek modes (Book III) and detailed them in ways that today could eas-
ily be programmed for analysis and composition (Shirlaw 1955). Ptolemy’s limiting
of the overall range of modes to two octaves (systema teleion or “perfect system”)
forced the modes in transposition to lose members at the upper limits of their
range and regain them at the lower limit of their range, a particularly computable
process. He also accepted the notion of seven modes (Principle 2), because that
number produced the central octave of the male vocal range.
Ptolemy’s studies, among others, prompted Aristides Quintilianus in the third
century CE to divide the study of music into three principal categories: theoretical,
practical, and numerical. Quintilianus’s Perí mousikês (On Music) further separates
the purely theoretical study of music into four distinct groupings (scientific, techni-
cal, critical, and historical). Although these categories do not strictly apply to mu-
sic theory, analytical processes are attached to aspects of each. Quintilianus’s
treatment of music as a numerical art connected it directly to mathematics and
involved patterns (Principle 1) such as triads, which he considered an important
concept allied with beauty.
Anicius Boethius (480–524), notably in his De institutione musica, inherited much
of Ptolemy’s modal concepts while continuing to advocate traditional Greek analyti-
cal processes: “Greatly influenced by Greek writers such as Nicomachus, Ptolemy,
Euclid, Plato, and Aristotle, the young Boethius set out to write works treating arith-
metic, music, geometry, and astronomy as disciplines that lead the soul to its first
encounter with incorporeal knowledge” (Bower 2002, 141). Boethius’s De institutione
musica (c. 500) follows the ideal of the Greek quadrivium, and its conservatism
deeply influenced the music theory of Western Europe for nearly the next millen-
nium (Patch 1935). The five books of De institutione musica cover harmonics, pro-
portions (Principle 1), semitones, and scales (Principle 2), and ends with a review of
harmonics as seen particularly through Ptolemy’s eyes. Boethius’s seminal work
provides a foundation for the compositional rules of his time (Boethius 1967). His
prescriptive formulae allow for many computational translations (see Figure 1.2).
Centuries passed before Hucbald (840–930), in his De harmonica institutione,
described the nomenclature necessary to realize many of Boethius’s theories.
Whereas Boethius’s treatise is primarily theoretical, Hucbald’s concentrates on the
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 9

BACKGROUND 9

FIGURE 1.1 From Aristoxenus, The Harmonics of Aristoxenus, trans. Henry S. Macran
(New York: Oxford University Press, 1902), 249.
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 10

10 HIDDEN STRUCTURE

FIGURE 1.2 A woodcut of Boethius with an instrument designed for methodologi-

cally deriving tunings (an algorithm).

practical. De harmonica institutione assumes its readers’ knowledge of Gregorian

chant—see the example in Figure 1.3—and thus is directed to musicians and not to
laymen (Weakland 1956).
The Musica enchiriadis, originally attributed to Hucbald but now considered
anonymous, appeared concomitantly with De harmonica institutione and provided
the first known instruction in organum (two-voice parallel-motion music, an exam-
ple of which appears in Figure 1.4; Principle 1 and Principle 2) and thus demarcated
a major turning point from monophony into an early type of polyphony. Like
Hucbald’s De harmonica institutione, the Musica enchiriadis had a completely practi-
cal intent as well as an algorithmic perspective in the sense of providing clear rules
for composition and analysis:
For example, he [Hucbald] (like Musica enchiriadis) uses no numerical interval
ratios or monochord division. Instead, the scale, and the intervals that structure
it, are taught empirically by means of concrete examples drawn from the plain-
chant melodies and intonation formulas of the cantus tradition, demonstrating by
direct experience the connection between the two. This characteristically Car-
olingian pragmatic approach is evident throughout in Hucbald’s continual cita-
tion of specific chant melodies to exemplify theoretical concepts. The concepts
themselves, however, are adapted from late Roman writings on the ancient ars
musica, especially Boethius’s De institutione musica. (Cohen 2002, 318)
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 11

BACKGROUND 11

FIGURE 1.3 An example of Gregorian chant, Parce Domine, notated in neumes; the
second staff is a continuation of the first.

& ww ww ww ww
ww ww ww ww ww ww ww
Sit glor - ri - a Do - mi - ni in sae - cu - la

From the booklet edited by Gerald Abraham

accompanying The History of Music in Sound
(booklet published by Oxford University press);
examples reprinted with permission of the publisher

FIGURE 1.4 An example of medieval organum.

About this same time, Guido of Arezzo (995–1050) completed his Micrologus
(1026), a work influenced by both Boethius and Hucbald as well as many of the
theorists that came before them. The Micrologus was one of the most significant
treatises on medieval music. Written in order to train the choir of the Arezzo Cathe-
dral, Micrologus covers intervals (Principle 1), scales (Principle 2), species conso-
nance, and the proper division of the monochord, all in great detail. The book is
rich in insight, but limited by its focus—training vocalists. Guido developed clear
descriptions of phrase structure and rhythmic meanings of neumes, among other
things:

Three brilliant pedagogical ideas have traditionally been attributed to Guido,

earning him his honored place in the history of music pedagogy: staff notation,
the system of hexachords, and his “classroom visual aid” for sight-singing
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 12

12 HIDDEN STRUCTURE

performance, the “Guidonian Hand” . . . these three innovations are so towering,

that it is less often noted that the “Micrologus,” besides being in effect an early
sight-singing manual, is also one of the very first in another long line of music-
pedagogical genres: the treatise on composition. (Wason 2002, 48–49)

The Guidonian hand, a visual aid for locating the semitones in the central part of
the gamut (see Figure 1.5), serves as a kind of algorithm in itself, a simple organiza-
tion of rules for memorization. Each portion of the Guidonian hand represents a
specific pitch (in solfège, of which this is one of the first known instances) within
the hexachord. The hand here spans nearly three octaves. Conductors or instruc-
tors would point to parts of their hands to indicate a sequence of pitches that the
choir or students would then sing in proper order. More than any of the treatises
discussed so far, Guido’s Micrologus represents a perfect historical model for how
theories and algorithms from any period can transfer to computation (Crocker
1958).
Guido’s important contributions influenced the work of Johannes de Garlandia
(1195–1272). Garlandia’s De mensurabili musica proposed a new theory of conso-
nances (Principle 1), dividing them into perfect (unisons and octaves) and imperfect
(major and minor thirds) types, with fourths and fifths relegated to intermediate sta-
tus. Garlandia classified dissonances into similar categories: perfect (minor seconds,
tritones, and major sevenths), imperfect (major sixths and minor sevenths), and
intermediate (major seconds and minor sixths). Garlandia also defined classes of
organum, pitches, ligatures, and many other aspects of thirteenth-century music,
making De mensurabili musica one of the most comprehensive and enumerative
texts of its time (Crocker 1962). De mensurabili musica, though often overly exhaus-
tive and confining in its limitations, provides clear algorithms for the practical use
of rules in analyzing the music of its day.
Jacques de Liège’s (1270–1340) Speculum musice (1340), an encyclopedic de-
scription of the music theory of the early fourteenth century, followed and built
upon the work of Guido and Garlandia. Liège classified music theory into five
categories: heavenly (celestis), cosmic (mundana), human (humana), instrumental
(sonorus), and analysis (practica). However, most of his treatise concerns sonorus
and practica, which involved modes—Principle 1—and measured (metered) music.
Philippe de Vitry (1291–1361), especially in his Ars nova (1320), developed a series
of signs that represented the division of notes into various short durations:

Several of the texts which describe these new mensurations refer to the com-
poser Philippe de Vitry as the inventor of the new system. Whether this is true
or not, de Vitry was certainly a well-known advocate of the “new art” of musical
composition in the fourteenth century, of which mensural innovations play such
a prominent role. Indeed it is from the title of one of his treatises that the term
ars nova was taken to describe this new style. . . . (Berger 2002, 635)
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 13

BACKGROUND 13

FIGURE 1.5 The Guidonian hand, in which a specific note is assigned to each part
of the hand. By pointing to a part of the hand, the conductor can indi-
cate to a group of singers which note to sing.
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 14

14 HIDDEN STRUCTURE

Both Liège’s and de Vitry’s writings have clear mathematical (programmable) as-
pects, even though, like many medieval theorists, these authors often navigate
through religious and philosophical realms as well (Werner 1956).
Johannes Tinctoris (1436–1511) honed the work of Liège and de Vitry (especially
in his Liber de arte contrapuncti of 1477). Tinctoris convincingly demonstrated that
his theories of the uses and construction of dissonant suspensions (Principle 1)
could help reveal the intricacies of the counterpoint of his time. His Proportionale
musices (1473–1474) comprehensively describes temporal relationships based on
proportions. Although Tinctoris was not a particularly original thinker, his prolific
output and careful description of the rules that apply to the various forms of
fifteenth-century music make his work vital to the history of theory and provide a
clear model for rule-based algorithmic analysis. Figure 1.6 presents an example of
florid counterpoint—free rhythmic values in all voices—from Tinctoris’s era.
As sixteenth-century polyphony developed and flourished, the work of Gioseffo
Zarlino (1517–1590), particularly his Le istitutioni harmoniche (1558) and Dimonstra-
tioni harmoniche (1571); Henricus Glarean (1488–1563), notably his Dodecachordon
(1547); and Nicola Vicentino (1511–1572), especially his L’antica musica ridotta alla
moderna prattica (1555), helped to decipher the developing complex counterpoint
and introduced chromatics into the diatonic modes (Principle 2). Zarlino was par-
ticularly influential in describing just intonation (see Figure 1.7) and the proper use
of chromaticism. Many of the principles these music theorists espoused translate
well to algorithms, particularly the rules of modal counterpoint (see Cope 2004).
Building on the work of Zarlino, Glarean, and Vicentino, several late sixteenth-
and early seventeenth-century theorists (e.g., Thomas Morley in A Plaine and Easie
Introduction to Praticall Musicke [1597]; Thomas Campion in A New Way of Making
Fowre Parts in Counter-point [1618]; and Michael Praetorius in Syntagma musicum
[1618]) developed algorithms for both composition and performance. Christoph
Bernhard in his Musica autoschediastika (1601) and Musica poetica (1606) actually
developed a system of classifying musical style. His descriptions and naming of
several embellishments (Principle 1) and expressions in music were the first of
their kind (Boorman 1980):

Bernhard, working in North Germany in the decades just after the publication of
Kircher’s Musurgia universalis in 1650, made style the very foundation of his clas-
sificatory system. He retained, perhaps unconsciously, an underlying link to the
Burmeister tradition, in that he saw musical figures as ornamenting a plain, dia-
tonic musical style. But now, that plain style is explicitly identified as an actual
style in the real musical world. (McCreless 2002, 862)

In the early eighteenth century, Joseph Johann Fux (1660–1741) completed his
seminal Gradus ad Parnassum (1725), a book that survives to this day as a guide for
writing species counterpoint. Fux had studied the writings of Zarlino and others,
and his dialogue between master and pupil in Gradus lays out both the didactic and
the aesthetic rules for sixteenth-century counterpoint (Principles 1 and 2) in ways
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 15

BACKGROUND 15

À& C w ˙ ˙ ˙ ˙ œ œ ˙ ˙ ˙
˙ ˙ ˙ ˙
Et resurrexit
˙ ww w ẇ ˙ ww w
?C w ˙
à w

FIGURE 1.6 An example of three-voice counterpoint around the time of Tinctoris.

FIGURE 1.7 The division of the octave according to Zarlino.

02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 16

16 HIDDEN STRUCTURE

that make computational implementation not only possible but also practical (see
my description of the Gradus program in Cope 2004):

Whereas the rules of Cochlaeus, Zarlino, and Aaron deal with surface situations,
Fux’s four rules operate abstractly, at a high level of generality. True, these rules
do not cover all of the situations addressed by earlier theorists. For example, be-
ginning and end, proximate location, and so on, are left to later discussion in
specific contexts. In effect, the rules cover an infinitely wider range of situations.
Taking two classes of consonance and mapping them on to three types of move-
ment result in a regulation of great power and memorability. (Bent 2002, 560–561)

The four rules mentioned here refer to the four ways that perfect and imperfect
consonances can precede and follow one another. Documented evidence exists
that many composers of note (e.g., Beethoven, Brahms, and others) used Fux’s
Gradus for studying counterpoint. Figure 1.8 presents a manuscript page based
on this text, attributed by some to Beethoven, who, these scholars believe, tran-
scribed and embellished it for his student Archduke Rudolph.
Around this same time, Jean-Philippe Rameau (1683–1764), building on the theo-
retical studies of many of the theorists discussed thus far, described chords as
emanating from a single source pitch (root).

Rameau acknowledged that the inspiration for this breakthrough came from
Descartes’ method, which was to build a system of natural law on a self-evident
principle. In his Traité de l’harmonie réduite à ses principes naturels (1722)
Rameau identified this first principle as the first six divisions of the string; these
could be shown to generate all of the consonant and dissonant intervals and
chords as well as the rules for their interconnection. But it was first necessary to
recognise as an a priori fact that a note and its octave-replicates were identical.
From this ensued the principle of inversion. (Boorman 1980, 756)

Thus, Rameau laid the foundation for the now traditional functional analysis (Prin-
ciple 3) of tonal music, the fundamental bass—a foundation that has proven emi-
nently computational, as demonstrated in the next section.
Figure 1.9 presents two diagrams from Rameau’s treatise on harmony (see
Rameau 1971) describing (a) the major (perfect) triad and its inversions, and (b) the
dominant (fundamental) seventh chord and its inversions (Principle 3). Rameau
comments that “the largest triangle will contain the perfect chord, the source and
the root of the other chords; these others will be contained in the two smaller tri-
angles” (Rameau 1971, 41). The pitch names here refer to solfège scale degrees, and
the numbers refer to partials of the overtone series projected from the bass pitch
(root) of the chord. Thus, Do in Figure 1.9a is partial 4 of the C overtone series, two
octaves above the fundamental C, and La in Figure 1.9b is partial 20 of the F over-
tone series. Although Figures 1.5 and 1.9 are separated by more than six hundred
years, the algorithmic notation for both suggest at least one common perspective:
methods of automatic generation and analysis of music.
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 17

BACKGROUND 17

FIGURE 1.8 An exercise by Beethoven for his student Archduke Rudolph based on
Fux’s Gradus ad Parnassum.

Johann Philip Kirnberger (1721–1783) continued to develop many of Rameau’s

ideas, especially the triad and its inversions, and extended these ideas into a no-
tion of melodic function. He also divided dissonance into two categories: essential
(such as the dominant seventh, Principle 3) and incidental (such as the suspen-
sion). His many books on theory (especially Die Kunst des reinen Satzes in der Mu-
sick of 1774–1779) could almost be used in classrooms to this day, so closely do
they approximate current approaches to tonal analysis.
Kirnberger also invented the musikalisches Würfelspiele, or musical dice games,
algorithmic combinatoria that also interested C. P. E. Bach, Haydn, and Mozart
(Cope 1996). Figure 1.10 presents the first two pages of a published version of a
musikalisches Würfelspiel attributed to Mozart, with the two matrices representing
phrases of music and each column giving numerical choices of measures of music
(not shown).
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 18

18 HIDDEN STRUCTURE

a.

b.

FIGURE 1.9 Two diagrams from Rameau’s Treatise on Harmony describing (a) the
major (perfect) triad and its inversions, and (b) the dominant (funda-
mental) seventh chord and its inversions.

Like Kirnberger, Joseph Riepel developed musikalisches Würfelspiele, often in-

cluding such algorithmic exercises in his treatises (see Grundregeln zur Tonordnung
insgemein [1755] and Gründkliche Erklärung der Tonordnung insbesondere [1757]):
Especially noteworthy is the wide variety of ways in which he [Riepel] character-
izes the organization and content of phrases. He thus distinguishes them on the
02_DAS_Chap1_pp1-44 1/29/09 1:31 PM Page 19

BACKGROUND 19

a.

a.

FIGURE 1.10 (a) First published after Mozart’s death in 1793 by J. J. Hummel in
Berlin, this musikalisches Würfelspiel instruction page shows the first
two phrases of number selections (chosen by throws of the dice)
that were keyed to measures of music. (b) Also shown, an example
page of music by Haydn for this own musikalisches Würfelspiel.

basis of their rhythmic activity (a concern rarely addressed by eighteenth-

century theorists), their overall melodic contour, their underlying harmonic sup-
port, their degree of melodic closure, and their length in terms of measure num-
bers. (Caplin 2002, 671)

Riepel further proposed the creation of major works using combinatorial proce-
dures. Ratner comments that “Riepel proceeds along these lines as he works out
melodic combinations in the construction of minuets, concertos, and symphonies.
Within a given model he seeks to achieve optimum effects by substituting figures,
phrases, and cadences” (Ratner 1970, 351).
Influenced primarily by Rameau, Moritz Hauptmann (1792–1868) in his Die Natur
der Harmonik und Metrik of 1853 outlines a system of harmony based on logic.
Although adopting Rameau’s system of roots and inversion, Hauptmann developed
harmonic successions based on common tones rather than exclusively on root pro-
gression (Principle 3). He avoided relating his theories to the overtone series,
primarily due to its inclusion of both consonant and dissonant intervals, and also
Another random document with
no related content on Scribd:
 THE ANTI-BURGLARS
I
The letter was addressed to Miss Mary Stavely. It ran:
“My dear Mary,
“I have just received five pounds that I had given up for
lost, and, remembering what you told me at Easter of the
importance of distributing a little money in the village, I
think you had better have it and become my almoner. An
almoner is one who gives away money for another. I shall
be interested in hearing how you get on.
“Your affectionate
“Uncle Herbert.”
Inside the letter was a five-pound note.
Mary read the letter for the twentieth time, and for the twentieth
time unfolded the crackling five-pound note—more money than she
had ever seen before. She was thirteen.
“But what shall I do with it?” she asked. “So many people want
things.”
“Oh, you mustn’t ask me,” said her mother. “Uncle Herbert wants
you to decide entirely for yourself. You must make a list of every one
in the village who wants help, and then look into each case very
carefully.”
“Yes,” said Harry, Mary’s brother, as he finished breakfast, “and don’t
forget me. My bicycle ought to be put right, for one thing, and, for
another, I haven’t any more films for my camera. If that isn’t a
deserving case I’d like to know what is.”
 II
In a few days’ time the list was ready. It ran like this:
£ s. d.
Mrs. Meadows’ false teeth want mending.
It can be done for 0 12 6
Tommy Pringle ought to go to a Nursing
Home by the sea for three weeks. This
costs 7s. a week and 5s. 4d. return fare 1 6 4
Old Mrs. Wigram really must have a new
bonnet 0 4 6
Mrs. Ryan has been saving up for months
to buy a sewing machine. She had it all
ready, but Sarah’s illness has taken away
10s. I should like to make that up 0 10 0
The little Barretts ought to have a real ball.
It isn’t any fun playing with a bit of
wood 0 1 0
Mr. Eyles has broken his spectacles again 0 2 6
Old Mr. and Mrs. Snelling have never been
in London, and they’re both nearly eighty.
I’m sure they ought to go. There is an
excursion on the first of the month at 3s.
return each, and their grandson’s wife
would look after them there. Fraser’s
cart to the station and back would be 4s. 0 10 0
Mrs. Callow will lose all her peas and currants
again if she doesn’t have a net 0 3 0
The schoolmaster says that the one thing
that would get the boys to the village
room is a gramaphone like the one at the
public-house. This is 15s., and twelve
tunes for 9s. 1 4 0
 Mrs. Carter’s mangle will cost 8s. to be
mended, but it must be done 0 8 0
Thomas Barnes’ truck is no good any more,
and his illness took away all the money
he had; but he will never take it if he
knows it comes from us 1 10 0
Mary read through her list and once more added up the figures.
They came to £6 11s. 10d.
“Dear me!” she said, “I hadn’t any idea it was so difficult to be an
almoner.”
She went through the list again, and brought it down to £5 0s. 10d.
by knocking off one week of Tommy Pringle’s sea-side holiday and
depriving the village room of its gramaphone.
“I suppose I must make up the tenpence myself,” she said.

III
That afternoon Mary went to call on Mr. Verney. Mr. Verney was an
artist who lived at the forge cottage. He and Mary were great
friends. She used to sit by him while he painted, and he played
cricket with her and Harry and was very useful with a pocket-knife.
“No one,” she said to herself, “can help me so well as Mr. Verney,
and if I decide myself on how the money is to be spent, it will be all
right to get some help in spending it.”
Mr. Verney liked the scheme immensely. “But I don’t see that you
want any help,” he said. “You have done it so far as well as possible.”
“Well,” said Mary, “there’s one great difficulty: Thomas Barnes would
never take anything from our house. You see, we once had his son
for a gardener, and father had to send him away because of
something he did; but though it was altogether his son’s fault,
Thomas Barnes has never spoken to father since, or even looked at
him. But he’s very old and poorly, and very lonely, and it’s most
important he should have a new hand-truck, because all his living
depends on it; but it’s frightfully important that he shouldn’t know
who gave it to him.”
“Wouldn’t he guess?” Mr. Verney said.
“Not if nobody knew.”
“Oh, I see: no one is to know. That makes it much more fun.”
“But how are we to do it?” Mary asked. “That’s why I want you to
help. Of course, we can post most of the money, but we can’t post a
truck. If Thomas Barnes knew, he’d send it back directly.”
“Well,” said Mr. Verney, after thinking for some time, “there’s only
one way: we shall have to be anti-burglars.”
“Anti-burglars!” cried Mary. “What’s that?”
“Well, a burglar is some one who breaks into a house and takes
things away; an anti-burglar is some one who breaks into a house
and leaves things there. Just the opposite, you see.”
“But suppose we are caught?”
“That would be funny. I don’t know what the punishment for anti-
burgling is. I think perhaps the owner of the house ought to be
punished for being so foolish as to interrupt. But tell me more about
Thomas Barnes.”
“Thomas Barnes,” said Mary, “lives in a cottage by the cross-roads all
alone.”
“What does he do?”
“He fetches things from the station for people; he carries the
washing home from Mrs. Carter’s; he runs errands—at least, he
doesn’t run them: people wish he would; he sometimes does a day’s
work in a garden. But he really must have a new barrow, and his
illness took all his money away, because he wouldn’t belong to a
club. He’s quite the most obstinate man in this part of the country.
But he’s so lonely, you know.”
“Then,” said Mr. Verney, “we must wait till he goes away on an
errand.”
“But he locks his shed.”
“Then we must break in.”
“But if people saw us taking the barrow there?”
“Then we must go in the night. I’ll send him to Westerfield suddenly
for something quite late—some medicine, and then he’ll think I’m ill
—on a Thursday, when there’s the midnight train, and we’ll pop
down to his place at about eleven with a screw-driver and things.”
After arranging to go to Westerfield as soon as possible to spend
their money, Mary ran home.
Being an almoner was becoming much more interesting.

IV
Mr. Verney and Mary went to Westerfield the next day, leaving a very
sulky Harry behind.
“I can’t think why Uncle Herbert didn’t send that money to me,” he
grumbled. “Why should a girl like Mary have all this almoning fun? I
could almon as well as she can.”
As a matter of fact, Uncle Herbert had made a very wise choice.
Harry had none of Mary’s interest in the village, nor had he any of
her patience. But in his own way he was a very clever boy. He
bowled straight, and knew a linnet’s egg from a greenfinch’s.
Mr. Verney and Mary’s first visit was to the bank, where Mary handed
her five-pound note through the bars, and the clerk scooped up four
sovereigns and two half-sovereigns in his little copper shovel and
poured them into her hand.
Then they bought a penny account-book and went on to Mr. Flower,
the ironmonger, to see about Thomas Barnes’ truck. Mr. Flower had
a secondhand one for twenty-five shillings, and he promised to
touch it up for two shillings more; and he promised, also, that
neither he nor his man should ever say anything about it. It was
arranged that the barrow should be wrapped up in sacking and
taken to Mr. Verney’s, inside the waggon, and be delivered after
dark.
“Why do you want it?” Mary asked him.
“That’s a secret,” he said; “you’ll know later.”
Mr. Flower also undertook to send three shillings’ worth of netting to
Mrs. Callow, asking her to do him the favour of trying it to see if it
were a good strong kind.
Mary and Mr. Verney then walked on to Mr. Costall, the dentist, who
was in Westerfield only on Thursdays between ten and four. It was
the first time that Mary had ever stood on his doorstep without
feeling her heart sink. Mr. Costall, although a dentist, was a smiling,
happy man, and he entered into the scheme directly. He said he
would write to Mrs. Meadows and ask her to call, saying that some
one whom he would not mention had arranged the matter with him.
And when Mary asked him how much she should pay him, he said
that ten shillings would do. This meant a saving of half a crown.
“How nice it would be always to visit Mr. Costall,” Mary said, with a
sigh, “if he did not pull out teeth.”
Mary and Mr. Verney then chose Mrs. Wigram’s new bonnet, which
they posted to her at once. Mr. Verney liked one with red roses, but
Mary told him that nothing would ever induce Mrs. Wigram to wear
anything but black. The girl in the shop recommended another kind,
trimmed with a very blue bird; but Mary had her own way.
Afterwards they bought a ball for the Barretts; and then they bought
a postal order for eight shillings for Mrs. Carter, and half a crown for
Mr. Eyles, and ten shillings for Mrs. Ryan, and fourteen shillings for
Mrs. Pringle. It was most melancholy to see the beautiful sovereigns
dropping into other people’s tills. Mary put all these amounts down
in her penny account-book. She also put down the cost of her return
ticket.
When they got back to the village they saw Mr. Ward, the station-
master. After telling him how important it was to keep the secret,
Mary bought a return ticket to the sea for Tommy Pringle, without
any date on it, and two excursion tickets for old Mr. and Mrs.
Snelling for the 1st of next month. Mr. Ward did not have many
secrets in his life, and he was delighted to keep these.
While they were talking to him a curious and exciting thing
happened. A message began to tick off on the telegraph machine.
Mr. Verney was just turning to go away when Mr. Ward called out,
“Stop a minute, please! This message is for Miss Stavely.”
Mary ran over to the machine and stood by Mr. Ward while he wrote
down the message which the little needle ticked out. She had never
had a telegram before, and to have one like this—“warm from the
cow,” as Mr. Ward said—was splendid. Mr. Ward handed it to her at
last.
“Mary Stavely, Mercombe.
“How is the almoning? I want to pay all extra expenses.—
Uncle Herbert.”
The reply was paid; but Mary had to write it out several times before
it satisfied her and came within the sixpence. This was what she
said:
“Stavely, Reform Club, London.
“All right. Will send accounts. Expenses small.—Mary.”
On the way home they spoke to Fraser, who let out carriages and
carts. Fraser liked the plan as much as every one else did. He
promised to call in on the Snellings in a casual way, on the morning
on which they would receive the tickets, and suggest to them that
they should let him drive them to the station and bring them home
again. When Mary offered to pay him, Mr. Fraser said no, certainly
not; he would like to help her. He hadn’t done anything for anybody
for so long that he should be interested in seeing what it felt like.
This meant a saving of four shillings.
Mary went to tea at Mr. Verney’s. After tea he printed addresses on a
number of envelopes, and put the postal orders inside, with a little
card in each, on which he printed the words, “From a friend, for
Tommy to go to the sea-side home for a fortnight”; “From a friend,
for Mr. and Mrs. Snelling to go to London”; “From a friend, for Mr.
Eyles’ spectacles,” and so forth, and then he stamped them and
stuck them down, and put them all into a big envelope, which he
posted to his sister in Ireland, so that when they came back they all
had the Dublin postmark, and no one ever saw such puzzled and
happy people as the recipients were.
“Has your mother any friends in Dublin, Miss Mary?” Mrs. Snelling
asked a day or so later, in the midst of a conversation about sweet
peas.
“No,” said Mary. It was not until afterwards that she saw what Mrs.
Snelling meant.

V
Next Thursday came at last, the day on which Thomas Barnes’ shed
was to be anti-burgled. At ten o’clock, having had leave to stay up
late on this great occasion, Mary put on her things, and Mr. Verney,
who had come to dinner, took her to his rooms. There, in the
outhouse which he used for a studio, he showed her the truck.
“And here,” he said, “is my secret,” pointing out the words—
THOMAS BARNES,
PORTER, MERCOMBE.
which he had painted in white letters on the side.
“He’s bound to keep it now, whatever happens,” Mr. Verney said. “In
order to make as little noise as possible to-night,” he added, “I have
wrapped felt round the tyres.”
He then took a bag from the shelf, placed it on the barrow, and they
stole out. Mr. Verney’s landlady had gone to bed, and there was no
sound of anyone in the village. The truck made no noise.
After half a mile they came to the cross-roads where Thomas
Barnes’ cottage stood, and Mr. Verney walked to the house and
knocked loudly.
There was no answer. Indeed, he had not expected one, but he
wished to make sure that Thomas had not returned from Westerfield
sooner than he should.
“It’s all right,” he whispered. “Now for the anti-burgling.”
He wheeled the truck to the side of the gate leading to the shed,
and, taking the bag, they passed through. Mr. Verney opened the
bag and took out a lantern, a hammer, and a screw-driver.
“We must get this padlock off,” he said, and while Mary held the
lantern he worked away at the fastenings. It was more difficult than
he expected, especially as he did not want to break anything, but to
put it back exactly as it had been. Several minutes passed.
“There,” he cried; “that’s it.”
At the same moment a sound of heavy footsteps was heard, and
Mary gave a little scream and dropped the lantern.
A strong hand gripped her arm.
“Hullo! Hullo!” said a gruff voice. “What’s this? Housebreaking,
indeed!”
Mr. Verney had stooped for the lantern, and as he rose the
policeman—for he it was—seized him also.
“You’d better come along with me,” the policeman said, “and make
no trouble about it. The less trouble you make, the easier it’ll be for
you before the magistrates.”
 WHILE MARY HELD THE LANTERN, HE WORKED AWAY
AT THE FASTENINGS.
“But look here,” Mr. Verney said, “you’re making a mistake. We’re not
housebreaking.”
The policeman laughed. “Now, that’s a good’un,” he said. “Dark
lantern, screw-driver, hammer, eleven o’clock at night, Thomas
Barnes’ shed—and you’re not housebreaking! Perhaps you’ll tell me
what you are doing, you and your audacious female accomplice
here. Playing hide-and-seek, I suppose?”
“Well,” said Mr. Verney, suddenly striking a match with his free hand,
and holding it up so that the light fell full on his own and on Mary’s
face, “we’ll tell you the whole story.”
“Miss Stavely!” cried the policeman, “and Mr. Verney. Well, this is a
start. But what does it all mean?”
Then Mr. Verney told the story, first making Dobbs promise not to
tell it again.
The policeman grew more and more interested as it went on. Finally
he exclaimed: “You get the door open, sir, and I’ll fetch the truck
through. Time’s getting along.”
He hurried out of the yard and returned carrying the truck on his
shoulders. Then he stripped off the felt with his knife and ran it into
the shed, beside the old broken-down barrow that had done service
for so many years.
Mr. Verney soon had the padlock back in its place as if nothing had
happened, and after carefully gathering up the felt they hurried off,
in order to get home before Thomas Barnes should call with the
medicine that he had been sent to buy.
“Let me carry the bag, sir,” the policeman said.
“What, full of burgling tools!” said Mr. Verney.
“Mum’s the word,” the policeman replied, “mum’s the word.”
At the forge cottage he wished them good-night.
“Then you don’t want us in court to-morrow?” Mr. Verney asked.
“Mum’s the word,” was all that Dobbs replied, with a chuckle.
Thomas Barnes’ train being late, Mary did not get to bed until after
twelve that night. She laid her head on the pillow with particular
satisfaction, for the last and most difficult part of the distribution of
Uncle Herbert’s money was over.

VI
The next day Mary sent Uncle Herbert a long description of her
duties as his almoner, and enclosed the account. What with postages
and her railway fare, she had spent altogether £4. 18s. 11d.
Two days later this letter came back from Uncle Herbert:
“Dear Mary,
“You are as good an almoner as I could wish, and I hope
that another chance of setting you to work will come. Put
the thirteen pence that are over in a box labelled ‘The
Almoner’s Fund.’ Then take the enclosed postal order for a
pound and get it cashed, and the next time you are in
Westerfield buy Mr. Verney a box of cigarettes, but be sure
to find out first what kind he likes. Also give Harry six
shillings. I dare say he has broken his bicycle or wants
some more films: at any rate, he will not say no. The rest
is for yourself to buy something purely for yourself with.
Please tell your mother that I am coming on Saturday by
the train reaching you at 5.8. I shall walk from the station,
but I want Thomas Barnes to fetch my bag.
“Your affectionate
“Uncle Herbert.”
Whether or no Thomas Barnes knew where the truck came from we
never found out; but at Christmas-time he was discovered among
the waits who sang carols on the Stavelys’ lawn.
SIR FRANKLIN AND THE LITTLE
MOTHERS
 SIR FRANKLIN AND THE LITTLE
MOTHERS
I
Once upon a time there was a very rich gentleman named Sir
Franklin Ingleside, who lived all alone in a beautiful house in
Berkeley Square. He was so rich that he could not possibly spend
more than a little of his money, although he gave great sums away,
and had horses and carriages, and bought old pictures and new
books.
He lived very quietly, rode a little, drove a little, called on old friends
(chiefly old ladies), usually dined alone, and afterwards read by the
fire.
Although the house was large and full of servants, all Sir Franklin’s
wants were supplied by his own particular man, Pembroke.
Pembroke was clean-shaven, very neat, spoke quietly, and never
grew any older or seemed ever to have been any younger. It was
impossible to think of Pembroke as a baby, or a boy, or a person
with a Christian name. One could think of him only as a grave man
named Pembroke. No one ever saw him smile in Berkeley Square,
but a page boy once came home with the news that he had passed
Mr. Pembroke talking to a man in the street at Islington, and heard
him laugh out loud. But page boys like inventing impossible stories,
and making your flesh creep.
Pembroke lived in a little room communicating by bells with all the
rooms which Sir Franklin used; so that whenever the bell rang
Pembroke knew exactly where his master was. Pembroke did not
seem to have any life but his master’s; and the one thing about
which he was always thinking was how to know beforehand exactly
what his master wanted. Pembroke became so clever at this that he
would often, after being rung for, enter the room carrying the very
thing that Sir Franklin was going to request him to get.
Sir Franklin once asked him how he did it, and Pembroke said that
he did not know; but part of the secret was explained that very year
quite by chance. It was like this. In the autumn Sir Franklin and
Pembroke always went to Scotland, and that year when they were in
Scotland the Berkeley Square house was done up and all the old
pull-bells were taken away and new electric ones were put in
instead. When Sir Franklin came back again he noticed that
Pembroke was not nearly so clever in anticipating needs as he had
been before; and when he asked him about it, Pembroke said: “My
opinion, sir, is that it’s all along of the bells. The new bells, which
you press, ring the same, however you press them, and startle a
body too, whereas the old bells, which you used to pull, sir, told me
what you wanted by the way you pulled them, and never startled
one at all.”
So Sir Franklin and Pembroke went to Paris for a week while the new
press bells were taken away and the old pull-bells put back again,
and then Pembroke became again just as clever as before. (But that
was, of course, only part of the secret.)

II
It was at a quarter to nine on the evening of December 18, 1907,
that Sir Franklin, who was sitting by the fire reading and thinking,
suddenly got up and rang the bell.
Pembroke came in at once and said, “I’m sorry you’re troubled in
your mind, sir. Perhaps I can be of some assistance.”
“I’m afraid not,” said Sir Franklin. “But do you know what day this
is?”
“We are nearing the end of December 18,” said Pembroke.
“Yes,” said Sir Franklin, “and what is a week to-day?”
“A week to-day, sir,” said Pembroke, “is Christmas Day.”
“And what about children who won’t get any presents this
Christmas?” Sir Franklin asked.
“Ah, indeed, sir,” said Pembroke.
“And what about people in trouble, Pembroke?” Sir Franklin asked.
“Ah, indeed, sir,” said Pembroke.
“And that reminds me,” Pembroke added after a pause, “that I was
going to speak to you about the cook’s brother-in-law, sir: a worthy
man, sir, but in difficulties.”
Sir Franklin asked for particulars.
“He keeps a toy-shop, sir, in London, and he can’t make it pay. He’s
tried and tried, but there’s no money in toys in his neighbourhood—
except penny toys, on which the margin of profit is, I am told, sir,
very small.”
Sir Franklin poked the fire and looked into it for a little while. Then,
“It seems to me, Pembroke,” he said, “that the cook’s brother-in-
law’s difficulties and the little matter of the children can be solved in
the same action. Why shouldn’t we take over the toy-shop and let
the children into it on Christmas Eve to choose what they will?”
Pembroke stroked his chin for a moment and then said, “The very
thing, sir.”
“Where does the cook’s brother-in-law live?” Sir Franklin asked.
Pembroke gave the address.
“Then if you’ll call a hansom, Pembroke, we’ll drive there at once.”

III
It does not matter at all about the visit which Sir Franklin Ingleside
and Pembroke paid to the cook’s brother-in-law. All that need be
said is that the cook’s brother-in-law was quite willing to sell Sir
Franklin his stock-in-trade and to make the shop over to him, and Sir
Franklin Ingleside rode back to Berkeley Square not only a
gentleman who had horses and carriages and who bought old
pictures and new books, but perhaps the first gentleman in Berkeley
Square to have a toy-shop too.
On the way back he talked to Pembroke about his plans.
“There’s a kind of child, Pembroke,” said Sir Franklin, “that I
particularly want to encourage and reward. It is clear that we can’t
give presents to all; and I don’t want the greedy ones and the
strongest ones to be as fortunate as the modest ones and the weak
ones. So my plan is, first of all to make sure that the kind of child
that I have in mind is properly looked after, and then to give the
others what remains. And the particular children I mean are the little
girls who take care of their younger brothers and sisters while their
parents are busy, and who go to the shops and stalls and do the
marketing. Whenever I see one I always say to myself, ‘There goes a
Little Mother!’ and it is the Little Mothers whom this Christmas we
must particularly help.
“Now what you must do, Pembroke, during the next few days, is to
make a list of the streets in every direction within a quarter of a mile
of the toy-shop, and then find out, from the schoolmistresses, and
the butchers, and the publicans’ wives, and the grocers, and the oil-
shops, and the greengrocers, and the more talkative women on the
doorsteps, which are the best Little Mothers in the district and what
is the size of their families, and get their names and addresses. And
then we shall know what to do.”
By this time the cab had reached Berkeley Square again, and Sir
Franklin returned to his books.

IV
The next few days were the busiest and most perplexing that
Pembroke ever spent. He was in Clerkenwell, where the toy-shop
was situated, from morning till night. He bought all kinds of things
that he did not want—cheese and celery, mutton-chops and beer,
butter and paraffin—just to get on terms with the people who know
about the Little Mothers.
Although naturally rather silent and reserved, he talked to butchers
and bakers and women on doorsteps, and schoolmistresses, and
even hot-potato men, as if they were the best company in the world,
and bit by bit he made a list of twenty-two Little Mothers of first-
class merit, and fifty-one of second-class merit, and all their
children.
Having got these down in his book, Pembroke was going home on
the evening of the 21st very well pleased with himself on the whole,
but still feeling that Sir Franklin would be disappointed not to have
the name of the best Little Mother of all, when an odd thing
happened. He had stopped in a doorway not very far from the toy-
shop, to light his pipe, when he heard a shrill voice saying very
decidedly, “Very well, then, William Kitchener Beacon, if that’s your
determination you shall stay here all night, and by and by the rats
will come out and bite you.”
Pembroke stood still and listened.
 A LITTLE PROCESSION PASSED THE DOORWAY.
“I don’t want to go home,” a childish voice whimpered. “I want to
look in the shops.”
“Come home you must and shall,” said the other. “Here’s Lucy tired
out, and Amy crying, and John cold to his very marrer, and Tommy
with a sawreel, and father’ll want his dinner, and mother’ll think
we’re all run over by a motor-car; and come home you must and
shall.”
Sounds of a scuffle followed, and then a little procession passed the
doorway. First came a sturdy little girl of about ten, carrying a huge
string-basket filled with heavy things, and pulling behind her by the
other hand a small and sulky boy, whom Pembroke took to be
William Kitchener Beacon. Then came the others, and lastly Tommy,
limping with the sore heel.
Pembroke stopped the girl with the bag, and asked her if she lived
far away, and finding that it was close to the toy-shop, he said he
should like to carry the bag, and help the family home. He was not
allowed to carry the bag, but no objection was raised to his lifting
Tommy on his back, and they all went home together.
On the way he discovered that the Little Mother was named Matilda
Beacon, and that she lived at 28, Pulvercake Buildings, Clerkenwell.
She was nine years old, an age when most of you are still running to
your nurses to have this and that done for you. But Matilda, in
addition to doing everything for herself very quietly and well, had
also to do most things for her mother, who went out charing every
day, except Sunday, and for her brothers and sisters, of whom she
had five—three brothers aged seven, six, and three, and two sisters,
who were twins and both five. Matilda got them up and put them to
bed; picked them up when they fell, and dried their tears; separated
them when they quarrelled, which was very often; bought their food
and cooked it, and gave it to them, and saw that they did not eat
too fast; and was, in short, the absolute mistress of the very tiny flat
where the Beacons lived.
Mr. Beacon worked on the line at St. Pancras, and if he was late
home, as he very often was, Mrs. Beacon was always sure that he
had been run over by a passing train and cut into several pieces; so
that in addition to all her other work Matilda had also to comfort her
mother.
The next day, when he came again to the toy-shop district,
Pembroke was delighted to find that by general consent Matilda
Beacon was considered to be the best Little Mother in Clerkenwell;
but who do you think came next in public opinion? Not Carrie
Tompsett, although she had several strong backers; and not Lou
Miller, although she had her supporters too, and was really a very
good little thing, with an enormous family on her hands. No, it was
neither of these. Indeed, it was not a Little Mother at all, so I don’t
see how you could have guessed. It was a “Little Father.” It was
generally agreed by the butchers and bakers and oilmen and hot-
potato men and publicans and the women on the doorsteps, that the
best Little Mother next to Matilda Beacon was Artie Gillam, who,
since his mother had died last year and his father had not yet
married again, had the charge of four sisters and two brothers.
All these things Pembroke reported to his master; and Sir Franklin
was so much interested in hearing about Matilda Beacon that he told
Pembroke to arrange so that Mrs. Beacon might stay at home one
day and let Matilda come to see him. So Matilda put on her best hat
and came down from Clerkenwell to Berkeley Square on the blue bus
that runs between Highbury and Walham Green.

V
When the splendid great door was opened by a tall and handsome
footman Matilda clung to Pembroke as if he were her only friend in
the world, as, indeed, he really seemed to be at that moment in that
house. She had never seen anything so grand before; and after all, it
is rather striking for a little girl of nine who has all her life been
managing a large family in two small rooms in Clerkenwell, to be
brought suddenly into a mansion in Berkeley Square to speak to a
gentleman with a title. Not that a gentleman with a title is
necessarily any more dreadful than a policeman; but Matilda knew
several policemen quite intimately, and was, therefore, no longer
afraid of them, although she still found their terribleness useful when
her little brothers and sisters were naughty. “I’ll fetch a policeman to
you!” she used to say, and sometimes actually would go downstairs
a little way to do so and come back stamping her feet; and this
always had the effect of making them good again.
Sir Franklin was sitting in the library with a tea-table by his side set
for two, and directly Matilda had dared to shake his hand he told
Pembroke to bring the tea.
Matilda could not take her eyes from the shelves of books which ran
all round the room. She did not quite know whether it might not be
a book-shop and Sir Franklin a grand kind of bookseller; and then
she looked at the walls and wondered if it was a picture-shop; and
she made a note in her mind to ask Mr. Pembroke.
Her thoughts were brought back by Pembroke bringing in a silver
tea-pot and silver kettle, which he placed over a spirit lamp; and
then Sir Franklin asked her if she took sugar.
(If she took sugar? What a question!)
She said, “Yes, please, sir,” very nicely, and Sir Franklin handed her
the basin.
Would she have bread and butter or cake? he asked next.
(Or cake? What a question again!)
She said she would like cake, and she watched very carefully to see
how Sir Franklin ate his, and at first did the same; but when after
two very small bites he laid it down and did not pick it up again,
Matilda very sensibly ceased to copy him.
When they had finished tea and had talked about various things that
did not matter, Sir Franklin asked her suddenly, “How would you like
to keep shop, Matilda?”
Matilda gasped. “What sort of a shop?” she asked at last.
“A toy-shop,” said Sir Franklin.
“Oh, but I couldn’t,” she said.
“Only for one day,” Sir Franklin added.
“One day!” Her eager eyes glistened. “But what about Tommy and
Willy and the twins?”
“Your mother would stay at home that day and look after them. That
could easily be arranged.
“You see,” Sir Franklin went on, “I want to give all the children in
your street and in several other streets near it a Christmas present,
and it is thought that the best way is to open a toy-shop for the
purpose. But it is necessary that the toy-shop keeper should know
most of the children and should be a capable woman of business,
and that is why I ask you. The salary will be a sovereign; the hours
will be from two to eight, with an interval for tea; and you shall have
Mr. Pembroke to help you.”
Matilda did not know how to keep still, and yet there was the least
shade of disappointment, or at least perplexity, on her face.
“Is it all right?” Sir Franklin asked.
“Ye-e-s,” said Matilda.
“Nothing you want to say?”
“No-o-o,” said Matilda; “I don’t think so.”
And yet it was very clear that something troubled her a little.
Sir Franklin was so puzzled by it that he went out to consult
Pembroke. Pembroke explained the matter in a moment.
“I ought to have said,” Sir Franklin remarked at once on returning,
“that the shopkeeper, although a capable business woman, may play
at being a little girl, too, if she likes, and will find a doll and a work-
basket for herself, and even sweets too, just like the others.”
Matilda’s face at once became nothing but smiles.
“You will want a foreman,” Sir Franklin then said.
“Yes,” said Matilda, who would have said yes to anything by this
time.
“Well, who will you have?”
“I don’t think Tommy would do,” said Matilda. “He’s that thoughtless.
And Willy’s too small.”
“How about Frederick?” said Sir Franklin, ringing the bell twice.
Matilda sat still and waited, wondering who Frederick was.
After a moment or two the door opened, and a very smart boy, all
over buttons, came in. “You can take away the tea-things,” said Sir
Franklin.
“That was Frederick,” said Sir Franklin, when the boy had gone.
“Oh!” said Matilda.
“Would he do for foreman?” Sir Franklin asked.
Matilda hesitated. She would have preferred some one she knew,
but she did not like to say so.
“Too buttony?” suggested Sir Franklin.
Matilda agreed.
“Then,” said Sir Franklin, “is there anyone you know?”
“I think Artie Gillam——” said Matilda.
“Very well, then,” said Sir Franklin, “it shall be Artie Gillam. His
wages will be ten shillings.”
And thus everything was settled, and Matilda was sent home with
Frederick the page boy, the happiest and most responsible Little
Mother in London, with an armful of good things for the family.

VI
Meanwhile Pembroke had been to Houndsditch buying quantities of
new toys: for every Little Mother a large doll and a work-basket, and
smaller dolls and other toys for the others, together with sweets and
oranges and all kinds of other things, and everything was ready by
the day before Christmas Eve, and all the tickets were distributed.
The tickets were Pembroke’s idea, because one difficulty about
opening a free toy-shop in a poor district of London for one day only
is that even the invited children, not having had your opportunities
of being brought up nicely and learning good manners, are apt to
push and struggle to get in out of their turn, and perhaps even to try
to get in twice, while there would be trouble, too, from the children
who did not belong to the district. Pembroke knew this, and thought
a good deal about the way to manage it so that there should be no
crowding or difficulty. In the end Sir Franklin engaged a large hall, to
which all the children were to come with their tickets, and from this
hall they were to visit the shop in little companies of ten, make their
choice of toys, and then go straight home. Of course, a certain
number of other children would gather round the shop, but that
could not be helped, and perhaps at the close of the afternoon,
when all the others had been looked after, they might be let in to
choose what was left. And in this state were the things the night
before Christmas Eve.

VII
Pembroke managed everything so well that the great day went off
without a hitch. At half-past two the Little Mothers with their families
began to arrive, and they were sent off to the shop in companies of
ten or thereabouts, two or three families at once. A couple of
friendly policemen kept the crowd away from the shop, so that the
children had plenty of time and quiet to choose what they wanted.
All the Little Mothers, as I have said, had each a doll and a work-
basket; but the younger children might make their choice of two
things each, and take two things for any little brother or sister who
could not come—Clerkenwell being full of little boys and girls who
are not very well.
When they were chosen, Artie Gillam wrapped them up, and off the
children went to make room for others.
Matilda was a splendid shopkeeper. She helped the smaller children
to choose things in a way that might be a real lesson to real keepers
of toy-shops, who always seem tired.
“Now then, Lizzie Hatchett,” she said, “you don’t want that jack-in-
the-box. What’s the good of a jack-in-the-box to you if your brother’s
got one? One in a family’s plenty. Better have this parasol: it lasts
longer and is much more useful.
“Here’s a nice woolly lamb for Jenny Rogers’s baby brother,” she
cried, taking away a monkey on a stick. “He’ll only suck the paint off
that and be deathly ill.
“Now, Tommy Williams, don’t bother about those ninepins. Here’s a
clockwork mouse I’ve been keeping for you.” And so on. Matilda’s
bright, quick eyes were everywhere.
Only one or two uninvited children squeezed in with the others. One
of these was a very determined little rascal, who actually got in
twice. The first time he went away not only with toys of his own, but
with something for a quite imaginary brother with whooping-cough.
This made him so bold that he hurried away and fought another little
boy in the next street and took away his coat and cap. The coat was
red and the cap had flaps for the ears, so that they made him look
quite different. Wearing these, he managed to mix with the next
little party coming from the hall. But he had forgotten one thing, and
that was that the little boy whom he had fought was Artie Gillam’s
cousin. Artie at once recognized both the cap and the coat, and told
Mr. Pembroke, and Mr. Pembroke told one of the policemen, who
marched into the shop, looking exceedingly fierce, and seizing the
interloper by the arm, asked him whose coat he had on. At this the
boy began to cry, and said he would never do it again. But it was too
late. The policeman took hold of his wrist and marched him out of
the shop and through all the other children in the street, who
followed them in a procession, to the home of Artie’s cousin, and
there he had to give back the coat. Then he was allowed to go,
because Artie’s cousin’s father was out, and Artie’s cousin’s mother
(who was Artie’s aunt) was not at all the kind of woman to thrash
little boys.
So the time went on until all the children in Pembroke’s list had got
their toys and the hall was empty, and then the many others who
had been waiting outside were let in, one by one, until all the toys
were gone, and the policemen sent the rest away.
“Now,” said Pembroke, “we must shut the shop.” So Artie Gillam
went outside and put up the shutters, and Matilda put on her jacket
and hat.
Then Pembroke took some money out of his pocket to pay the
manager and her foreman their salaries.
“How will you have it?” he said to Matilda.
“Please I don’t know what you mean,” Matilda replied.
“Gold or silver?” Pembroke explained.
Matilda had never seen gold yet, except in jewellers’ windows. Her
mother’s wedding-ring was silver. “Oh, gold, please,” she gasped.
“One sovereign or two halves?” Pembroke asked.
“Two halves,” Matilda said.
Pembroke gave them to her.
Artie Gillam, on the other hand, wanted his ten shillings in as many
coins as he could have, and his pocket was quite heavy with it.
“And now,” said Pembroke, “I suppose you’re going home. Be careful
of your money on the way.”
“Oh no,” said Matilda, “I’m not going home yet. I’ve got some
shopping to do.”
“To-morrow’s dinner?” Pembroke suggested.
“No,” said Matilda mysteriously. “That’s all bought. Father won a
goose in the Goose Club.”
“Then what are you going to buy?” Pembroke asked, for he wished
to take as long and full a story home to Sir Franklin as might be.
“I’m going shopping for myself,” said Matilda. “I’m going to buy
some Christmas presents.”
“May I come with you?” Pembroke asked.
“Oh yes, please, I want you to. I’m only going to spend one of these
half-sovereigns. The other I shall put away. But I must buy
something for mother, and something for father, and I want to buy
something else, too, for somebody else.”
So Pembroke and Matilda and Artie, having turned out the gas and
locked up the shop, which, however, now contained nothing
whatever but paper and string and straw, walked off to the shops.
They first went into a draper’s, where Matilda looked at some shawls
and bought a nice thick woollen one for her mother, and also a pair
of grey wool mittens for her father. These came to five-and-six.
Then they went to an ironmonger’s and bought a cover for a plate to
keep things warm, which Matilda said was for her father’s dinner,
because he was often late while her mother thought he was being
cut in pieces. This cost ninepence.
Then they went to a tobacconist’s and bought a pipe with a silver
band on it, and two ounces of tobacco. These came to one-and-
fourpence and were also for her father.
Then they went to a china-shop and bought a hot-water bottle for a
shilling. “That,” said Matilda, “is for the old woman next door to us,
who nursed mother when she was ill. She can’t sleep at night
because her feet are so cold.”
“And now,” said Pembroke, “it’s my turn,” and he took the children
into a greengrocer’s shop and bought a shilling’s worth of holly and
mistletoe for each of them. “If you like,” he said, “I will carry this
home for you.”
Matilda thanked him very heartily, but said that she still had one
more present she must buy, and led the way to a little fancy shop,
kept by an old maid.
“Please,” said Matilda, “I want a kettle-holder.”
The old lady took out a drawer and laid it on the counter. It was full
of kettle-holders, some made in wool-work, others in patch-work.
Matilda looked at them very carefully one by one, and at last chose
one in scarlet and bright yellow wool-work. When it was done up in
a neat little packet and she had paid for it—sixpence—she handed it
to Pembroke.
“That,” she said, “is a present for the gentleman. When I had tea
with him I noticed that he hadn’t got one, and of course every family
ought to have a kettle-holder. I should have liked to make one for
him myself, but there hasn’t been time.”
 VIII
Sir Franklin Ingleside did not use the kettle-holder. He hung it on a
nail by the fire-place, and whenever he is asked about it, or people
smile at its very striking colours, he says, “I value that very highly;
that is the profit that I made out of a toy-shop which I once kept.”