Audacious Euphony

OXFORD STUDIES IN MUSIC THEORY

Series Editor Richard Cohn

Studies in Music with Text, David Lewin

Music as Discourse: Semiotic Adventures in Romantic Music, Kofi Agawu

Playing with Meter: Metric Manipulations in Haydn and Mozart’s Chamber Music
for Strings, Danuta Mirka

Songs in Motion: Rhythm and Meter in the German Lied, Yonatan Malin

A Geometry of Music: Harmony and Counterpoint in the Extended Common

Practice, Dmitri Tymoczko

In the Process of Becoming: Analytic and Philosophical Perspectives on Form in Early

Nineteenth-Century Music, Janet Schmalfeldt

Tonality and Transformation, Steven Rings

Audacious Euphony: Chromaticism and the Triad’s Second Nature, Richard Cohn
 Audacious Euphony

Chromaticism and the Triad’s Second Nature

Richard C ohn

1
1
Oxford University Press, Inc., publishes works that further
Oxford University’s objective of excellence
in research, scholarship, and education.

Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto

With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright © 2012 by Oxford University Press

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016
www.oup.com

Oxford is a registered trademark of Oxford University Press

All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.

Library of Congress Cataloging-in-Publication Data

Cohn, Richard Lawrence, 1955-
Audacious euphony : chromaticism and the consonant triad’s second nature / Richard Cohn.
p. cm. — (Oxford studies in music theory)
ISBN 978-0-19-977269-8 (hardback : alk. paper) — ISBN 978-0-19-983282-8 (companion website)
1. Harmony. 2. Triads (Music) I. Title. II. Series.
MT50.C736 2011
781.2΄5—dc22 2011008754

Publication of this book was supported by the Otto Kinkeldey Endowment of the
American Musicological Society.
135798642
Printed in the United States of America
on acid-free paper
 in memoriam
John Clough David Lewin
This page intentionally left blank
CONTENTS

Introduction ix
About the Companion Web Site xviii

1 Mapping the Triadic Universe 1

Three Ways to Calculate Triadic Distance 1
Triads in Chromatic Space 8
Remarks on Syntax and Maps 13
2 Hexatonic Cycles 17
A Minimal-Work Model of the Triadic Universe 17
The Hexatonic Trance 20
Contrary Motion and Balance 24
Hexatonic Progressions, Tonnetz Representations,
and Triadic Transformations 25
Near Evenness, Minimal Voice Leading, and the
Central Role of Augmented Triads 33
Remarks on Dualism 37
Triadic Structure Generates Pan-Triadic Syntax 39
Triads Are Homophonous Diamorphs 40
3 Reciprocity 43
The Historical Emergence of Augmented Triads 43
Consonance/Dissonance Reciprocity 46
Two Early-Century Examples: Beethoven and Schubert 48
Three Late-Century Examples: Liszt, Rimsky-Korsakov, Fauré 49
Reciprocity in Weitzmann’s Der Ubermässige Dreiklang 56
4 Weitzmann Regions 59
The Structure of a Weitzmann Region 59
Weitzmann Transformations and N/R Cycles 61
Remarks on the Tonnetz 65
Historical Origins of Weitzmann Regions 67
The Double-Agent Complex 72
Expanded N/R Chains 76
Weitzmann Regions without Sequences: Wagner and Strauss 78
viii  Contents

5 A Unified Model of Triadic Voice-Leading Space 83

How Hexatonic and Weitzmann Regions Interact 83
Chromatic Sequences 89
Transformational Substitutions 95
Voice-Leading Zones 102
Remarks on Disjunction and Entropy 106
6 Navigating the Triadic Universe: Three Compositional Scripts 111
Neighborhoods and Pitch Retention Loops 113
Departure → Return Scripts 121
Continuous Upshifts 131
7 Dissonance 139
Four Eighteenth-Century Approaches to Dissonance 139
Reduction to a Triadic Subset 142
Hexatonic Poles in Parsifal 145
The Tristan Genus as Nearly Even Tetrachord 148
Circumnavigating the Tristan-Genus Universe 159
Scriabin’s Mystic Species and Generalized Weitzmann Regions 166
8 Syntactic Interaction and the Convertible Tonnetz 169
Some Previous Proposals 169
The Diatonic Tonnetz 175
Horizontal Extensions 179
Vertical Extensions 184
The Convertible Tonnetz 186
Two Analytical Vignettes: Wagner and Brahms 189
9 Double Syntax and the Soft Revolution 195
A Summary Example from Schubert 195
Double Syntax and Its Skeptics 199
Code Switching and Double Determination 201
Cognitive Opacity 203
The Soft Revolution 205
On Musical Overdetermination 208

Glossary 211
Bibliography 215
Index 229
I N T R O D U C T IO N

The admittedly audacious but effective and euphonious progression shown [above]
defies definition in terms of an older doctrine of key. But . . . it consists only of
closely related chords contrasted with the tonic triad.
—Hugo Riemann, s.v. “Tonalität,” Musik-Lexicon, 1909

Two questions arise. First, what notion of harmony underlies Hugo Riemann’s
judgment that these chords are closely related? Current textbooks have inherited
the eighteenth-century formulation that triads are closely related if their epony-
mous scales are identical to within one degree of difference: they share at least six
out of seven tones. These harmonies don’t come close to qualifying: their associ-
ated scales share three tones out of seven. Riemann’s conception of harmonic dis-
tance is evidently rather different from our own. How can we construct and
represent that conception? Second, if the triads are closely related, why does
Riemann call the progression “audacious”? Close relations are unmarked, well
formed, normal—not the stuff of audacity. These questions lead to different kinds
of responses. The first is susceptible in principle to a systematic inquiry. If we can
establish that Riemann and his contemporaries calculated harmonic distance in a
consistent way, even if distinct from the way we do so, then we might have a chance
to understand the basis for his judgment. The second, because it identifies a para-
dox, is an invitation to an interpretation.
Audacious Euphony reconstructs conceptions of triadic distance that were
proper to nineteenth-century harmonic thought but have since been stripped
from music theory’s inheritance. What effect do these alternative conceptions have
on our understanding of how the nineteenth-century ear understood harmonic
relations and how nineteenth-century composers crafted strategies and made
choices that both appealed to and molded that ear? How did these alternative
conceptions, and the strategies and choices they motivated, contribute to the
charismatic, entraining, and sublime qualities that we still hear in many compositions
of that era?
Harmonic theorists of the nineteenth century provided a partial response to
these questions. This book expands that sketch into a fully realized proposal using
x  Introduction

conceptual, technical, and representational resources unavailable a century ago. Its

principal thesis is that major and minor triads are not only the familiar acoustic
consonances of eighteenth-century classical theory. They also have a second prop-
erty, equally rare and equally generative of syntax.1 That property underlies an
alternative method for measuring harmonic proximity, leading naturally to an
account of the triadic universe and to its representation through an interrelated set
of maps.
This book also explores the paradox identified in Riemann’s formulation, a
paradox that is writ large across the music of the long nineteenth century.
Chromatic progressions of triads excited the Romantic imagination not because
they conformed to expectations about triadic behavior and succession but because
they confounded them. In song, opera, and programmatic music of the period,
such progressions are often explicitly affiliated with altered or heightened realities:
Schubert’s magic harp, Wagner’s magic sleep, Rimsky-Korsakov’s magic and exotic
kingdoms, Liszt’s mountain-top meditations (Kurth 1923; Taruskin 1985). Even in
nonprogrammatic instrumental music, they are capable of evoking the strange,
magical, and inscrutable (Hoffmann 1989 [1813–14], 318; Cohn 2004). To address
how chromatic progressions summon those responses, we will need to explore
their position on a field animated by opposing forces: of classical and romantic
syntax, of normativity and aberration. How do the old and the new syntaxes coex-
ist and interact in nineteenth-century music, and what are the implications of this
coexistence for our understanding of musical cognition and of the transformation
of musical syntax across the long nineteenth century?
In order to cultivate an unfamiliar conception of the harmonic universe, it will
be useful to suspend some overlearned habits. Because some species of classical
diatonic theory is the first, and in many cases the only, music theory that we are
exposed to, its status as theory, a way of looking, easily hardens into nature: the
way that music is. In response, I will occasionally invite readers to “forget” that
major and minor triads are acoustic consonances generated from roots that occupy
a position in a diatonic scale with respect to some tonic. This is a cultivated and
strategic denial, motivated by the desire to prevent our default habits from flood-
ing back in and drowning those properties and relations that I wish to foreground.
The move is perhaps akin to one that physicians make when they bracket off the
aesthetic qualities of the bodies that they are examining—qualities that might
interest them in other contexts. Those qualities are not relevant to the diagnosis or
treatment of some ailment or condition, and to focus on them would be a distrac-
tion from the task before them. As we explore the relations of the two syntaxes in
the final chapters, readers will increasingly find themselves on familiar ground,
although their perspective on that ground will have been altered in some mean-
ingful way. That, at least, is my hope and intention.
In establishing the viability and utility of the alternative syntax whose proposal
is at the core of this book, Audacious Euphony draws on four different kinds of
evidence. The syntactic model will be shown to be theoretically powerful, in the

1. Syntax is a problematic term whose chief virtue is that it is better than its alternatives. It nonetheless
requires clarification, which I provide in the final section of chapter 1.
 Introduction  xi

sense that it generalizes in surprising and productive ways, beyond the consonant
triads for which it was designed. It will be shown to be analytically productive, in
the sense that it leads to new ways of hearing, conceiving, and representing music
composed by a broad cross section of nineteenth-century composers. It will be
shown to be historically appropriate, in the sense already indicated: its approach
emerges from indigenously nineteenth-century ways of thinking and writing
about musical materials and their relations, although it is often couched in lan-
guage that those historical figures would find alien. And it will be shown to be
historically productive, affording a new way to address the transformation of
musical style across the long nineteenth century. My hope is that each of these four
types of evidence will bear its own intrinsic and independent value for readers, so
that, for example, readers who do not connect with my analyses, who find my
historical readings tendentious, or for whom abstract modes of thinking about
music are distasteful, might nonetheless find value elsewhere in the book. My aim
has been to write a book that is as robust, multidimensional, and overdetermined
as its topic.

Genesis and Relation to Prior Work

In 1990, a passage from Schubert’s Eᅈ Piano Trio called my attention to a voice-

leading property of consonant triads. (The passage is partly reproduced as figure
2.8 in chapter 2.) The property was special; it was shared by no other three-note
chords, and the few chords of other sizes that shared it nontrivially included the
diatonic and pentatonic collections, which are equally privileged within the
European musical tradition, and within others as well. After seeking it in chords
and scales in universes with more or fewer than twelve tones, I began to formulate
some conjectures about this property and its impact on historical repertories.
Informal correspondence led John Clough to convene a small conference in
Buffalo during the summer of 1993, where Jack Douthett and David Lewin refined
and generalized my conjectures and suggested a scope for them beyond the limited
domain within which I had initially envisioned them (Douthett 1993; Lewin 1996;
Douthett and Steinbach 1998). These unanticipated applications to living reperto-
ries (rather than hypothetical ones that might be constructed ad hoc) helped build
my confidence in the significance of my observations and the conjectures that they
had prompted.
A second wave of reinforcements arrived in 1995–96, during a residency at the
University of Chicago’s Franke Institute of the Humanities. Reading in and around
nineteenth-century harmonic theory, aided by the recent dissertations of David
Kopp (1995) and Michael Kevin Mooney (1996), I discovered affinities with my
developing conception. An article of Larry Todd’s (1988) led me to Carl Friedrich
Weitzmann’s short monograph on The Augmented Triad (1853), which provided
the final building block for a model of triadic space that I presented at a second
Buffalo conference in 1997, the proceedings of which became a special issue of the
Journal of Music Theory (vol. 42, no. 2, 1998).
xii  Introduction

In the fourteen years since, the ideas emerging from the two Buffalo
conferences have been further generalized, merged with other theoretical con-
cerns, and adapted to various analytical and historical purposes, in dissertations
(and subsequent publications) of David Clampitt (1997), Edward Gollin (2000),
Robert Cook (2001), Nora Engebretsen (2002), Michael Siciliano (2002), Julian
Hook (2002), So-Yung Ahn (2003), and Steven Rings (2006). Those ideas are also
absorbed into, and significantly transformed by, the writings of Dmitri Tymoczko,
which have appeared in a steady stream since 2005 and culminated in the publica-
tion of A Geometry of Music (2011). The theoretical power of that emerging work,
and the pedagogical gifts of its author, prompted me to divert the flow of my own
developing book away from some of its early theoretical ambitions, the better to
focus on analytical and historical questions pertinent to its music-historical
moment.
With so many horses out of the barn and reproducing, I have two distinct sets
of motivations for bringing this material together into a book now. First, I have
aimed to make my ideas about nineteenth-century music accessible to readers
uncomfortable with or uninterested in the technical literature in scholarly journals.
The core of the exposition assumes no knowledge beyond a semester of under-
graduate harmony. I intend that this book be read not only by music theorists
and their students but also by music historians and psychologists, performers of
nineteenth-century repertories, composers who find creative vitality in triads
and seventh chords, and anyone else who has the appropriate background and
temperament to work through a close argument about musical properties and
relational systems and close readings of particular compositions.
I have avoided, to the extent possible, the mathematical modes of discourse
that have dominated much of the technical literature on this topic. Most of the
book will ask readers to be comfortable with counting small quantities, represent-
ing them as numbers, and adding or subtracting their values; for example, one
apple over here and two apples over there make three apples in all. Some parts of
the book use modulo twelve arithmetic, which you have been performing all your
life, although perhaps beneath your threshold of awareness. If you worked from
nine this morning until six this afternoon, how long was your shift? If you board
the train at eleven, and the journey lasts five hours, what time do you arrive? If you
know the answers spontaneously, or can calculate them quickly, then you are per-
forming arithmetic modulo twelve. My hope is that this level of mathematical
rigor will not drive away readers who are inclined to head for the hills when a
scientistic odor wafts about. For those who are comfortable with such discourses,
references to the technical literature are furnished. At the ends of chapters 2, 5,
and 7, the technical bar rises.
Second, I intend this book as a fresh contribution to music theory and analysis.
Although readers familiar with my papers on chromatic harmony will recognize
traces of many of them, almost every sentence and analysis presented here is
new. My earlier publications have offered a kaleidoscope of concepts and analyti-
cal perspectives, all radiating from my initial observation about the consonant
triad’s special voice-leading properties and their relationship to nineteenth-
century triadic chromaticism. A principal concern of Audacious Euphony is to
 Introduction  xiii

show that these perspectives constitute paths through a unified field, reflecting an
integrated vision. A second concern is to work out a model of the relationship
between classical and romantic harmony and to pursue its historical implications,
building on a program that I first sketched at the end of my initial publication on
the topic (Cohn 1996). Several readers, evidently having overlooked that sketch,
have interpreted me as claiming no such relationship (Fisk 2000; Lerdahl 2001;
Kopp 2002; Horton 2004; Damschroder 2010), and I am intent to put those
responses to rest. A related aim is to stake out positions with respect to some tech-
nical topics of particular interest to music theorists. These discussions are mostly
confined to subsections titled “Remarks.” Readers who find these concerns
parochial may skip these passages without loss of continuity.
A final goal is to convince readers that these assumptions, concepts, and repre-
sentational modes have the capacity to illuminate aspects of musical artworks that
they may cherish. Prior work, of mine and others, has focused mostly on refining
and generalizing the model, rather than pursuing its implications for understanding
particular compositions and the responses they provoke in musicians and listeners
(Hook 2002; Rings 2007). The core chapters include analyses of passages of music,
and sometimes entire compositions, by Schubert, Chopin, Liszt, Wagner, Brahms,
Rimsky-Korsakov, Dvořák, Bruckner, Fauré, and Richard Strauss. I hope that these
analyses will be satisfying on their own terms and that they might provide models
for readers to explore other compositions from this glorious repertory that continues
to occupy a prominent position in concert halls, opera houses, and mp3 players.
My prior articles on chromatic harmonic flew under the banner of neo-
Riemannianism, and I owe it to readers to explain why I avoid that term in the
present book. The first of my two distinct motivations is that it gives too much
credit to Hugo Riemann. It was David Lewin’s reading of Riemann’s harmonic
writings that constituted the originary moment for this branch of theorizing, and
the insights that he fashioned from that reading fuel my approach. But I have come
to believe that, with respect to the nineteenth-century ideas most at the heart
of this book, Riemann was functioning more as transmitter than generator
(Cohn 1998a).
My second motivation is that the domain of neo-Riemannian theory has never
been very stable (Hook 2002), and I am not comfortable with all of the views that
have been attributed to it or with all of the practices that have been performed
under its name. For example, Tymoczko (2009a) claims that a commitment to
harmonic dualism is at the heart of neo-Riemannian theory. On this basis, I am no
neo-Riemannian. At the same time, a strain of neo-Riemannianism has arisen
whose focus is on the application of transformational labels to harmonic progres-
sions. Although these labels are descriptively useful, they do not in themselves
lead to an understanding of triadic syntax any more than Roman numeral or
set-class designations alone constitute explanatorily adequate accounts of compo-
sitions from the tonal and atonal repertories. I view these labels as a bridge to a
first approximation; what lies on the other side of the bridge, if one is lucky, is an
understanding of how the moves designated by these labels behave as part of a
compositional system. It is the exploration, representation, and application of that
system that is the primary aim of this book.
xiv  Introduction

Although neo-Riemannian properly labels an approach toward hearing

and representing musical relations, the term has also been applied to the musical
relations themselves; thus, a tonally indeterminate progression of triads, or a com-
position filled with the same, is said to be neo-Riemannian. The term I favor for
the compositions or passages that invite such modes of audition, conceptualiza-
tion, and representation is pan-triadic, first used by Evan Copley (1991). The term
is inspired by its obverse, pan-diatonic (Slonimsky 1937): both terms designate
music that uses fundamental materials of tonality in tonally indeterminate ways,
one by using diatonic scales without triads, and the other by using triads without
diatonic scales.

Organization of Chapters

Figure 0.1 models the organization of the book. The nine chapters form three
groups of three: premises and theses, model construction and analysis, extensions
and implications. This symmetry is perturbed, however, by the beginning of chap-
ter 2, where significant aspects of the model are put into place. The form of the
book in some respects mirrors an aspect of its thesis: that it is the consonant triad’s
perturbed symmetry that motivates those triadic progressions most characteristic
of the nineteenth century.
Chapter 1 probes the limits of the classical view of harmonic distance, presents
some alternative approaches cultivated by nineteenth-century theorists, and sug-
gests that those approaches point the direction toward a model of nineteenth-
century triadic syntax that operates independently of classical principles. The first
half of chapter 2 introduces a preliminary model of the triadic universe, based on
the idea that triadic distance is conditioned by voice-leading proximity among
consonant triads, and shows how that model captures a number of characteristic
passages that extend chronologically from Mozart and Haydn of the 1780s to
Wagner and Brahms of the 1880s. The central role of major thirds and augmented
triads in the model leads to a proposal for why chromatic major third relations are
affiliated throughout the nineteenth century with altered and destabilized psycho-
logical states, the supernatural and the uncanny, as Richard Taruskin and others
have observed.

Figure 0.1. The organization of this book.

 Introduction  xv

In the middle of chapter 2, the arc of model construction is suspended in

order to advance the major theoretical claim of the book, which follows from ele-
ments already in play. That claim is that pan-triadic syntax follows from an inter-
nal property that consonant triads hold independently of the acoustic properties
that are claimed to generate classical syntax. One aspect of that claim, that aug-
mented triads are central to pan-triadic syntax even when they are absent from the
musical surface, may be counterintuitive, especially to readers whose understand-
ing of harmony is unilaterally shaped by textbook accounts of harmonic tonality.
Chapter 3 aims to defuse any skepticism that is aroused, by clarifying the nature of
the assertion and providing historical and analytical support for it. Following ideas
introduced by François-Joseph Fétis in the 1840s and refined by Ernst Kurth
around 1920, I recognize that repetition, conceived broadly, has the power to offset
the monolithy of diatonic tonality.
Some ideas from Carl Friedrich Weitzmann’s 1853 treatise on augmented
triads, introduced at the end of chapter 3, serve as the basis for a second prelimi-
nary model of the triadic universe, based on the voice-leading proximity of each
consonant triad to one of the four augmented triads, and partitioning the triads
into four Weitzmann regions. Chapter 4 traces the historical growth of Weitzmann
regions from basic diatonic routines; shows how Schubert, Chopin, and others use
four-chord subsets of these regions, forming what I call a double-agent complex, to
tease the enharmonic seam; and analyzes some passages of Schubert, Liszt, Wagner,
and Richard Strauss that plunge through that seam in order to explore the full
chromatic extent of a Weitzmann region.
The opening of chapter 5 demonstrates that the two preliminary models stand
in figure–ground relationship and that a comprehensive model of the triadic uni-
verse results from their synthesis. Chapters 5 and 6 explore various modes of geo-
metric representation and take the model for an analytical ride across some
familiar and lovely musical terrain. Chapter 5 focuses on sequential passages and
their transformations, with examples from Chopin, Liszt, Brahms, and Bruckner.
Chapter 6 documents three strategies for exploring triadic space using less strictly
patterned surfaces and analyzes complete short compositions and large sections of
longer ones, including Schubert’s “Der Doppelgänger,” “Auf dem Flusse,” and Bᅈ
major Piano Sonata, Brahms’s Second Symphony, and Dvořák’s “New World”
Symphony. A major focus of these chapters, following an implication of Daniel
Harrison’s work (1994), is to position Riemann’s harmonic dualism as a particular
instance of a more fundamental phenomenon, the relationship between upward
and downward melodic motion.
Chapters 7 and 8 take up two issues that have been bracketed off until the
model could be constructed on its own terms. Chapter 7 proposes several ways to
integrate dissonant harmonies into the model, with a particular analytic focus on
Chopin’s piano music and Wagner’s late operas. The second half of the chapter
constructs a model of seventh chords by analogy to the pan-triadic model, leading
to a general version of the model whose breadth is suggested by its applicability
to Alexander Scriabin’s treatment of mystic chords. Chapter 8 responds to a ques-
tion implicitly teased but explicitly skirted throughout the preceding chapters:
how does pan-triadic syntax interact, in alternation or overlay, with the familiar
xvi  Introduction

diatonic syntax of classical tonality? After reviewing the approaches of prior theo-
rists, including David Lewin, Brian Hyer, Steven Rings, and Candace Brower, as
well as those offered in my own publications, I present a new proposal and analyze
passages from Liszt, Wagner, and Brahms. My proposal asks readers to accept
the evidently controversial proposition that good music can be fashioned from
multiple syntaxes, and that listeners are cognitively capable of integrating them.
Chapter 9 supports that proposition’s epistemological value and uses it to approach
the question of how the harmonic practice of Mozart mutated into that of Webern.
It is this material that provides the basis for the claim that the model is historically
productive. The analysis of a passage from Schubert’s C major Symphony pre-
sented at the beginning of that chapter summarizes the approach developed in this
book and might serve as an effective initial exposure for readers who want to get a
sense of its central ideas.

Acknowledgments

First thanks go to Heather, whose love calmed and focused me, and whose
generosity sustained and nourished me, through many years of work on this
project.
The encouragement of David Lewin and John Clough, to whose memories this
book is dedicated, endured in me long after they passed away, within months of
each other, in 2003. David’s ideas were foundational, and the power of his ideas is
reflected in the number of citations to them in my text. John’s interest in my ideas
led him, on three occasions, to gather music theorists to Buffalo to grapple
with their implications. I will be forever grateful for the honor, as well as for the
ideas and perspectives that I took away from those sessions. Jack Douthett’s
early extensions of my ideas are amply represented in this book, and his skill at
communicating complex insights about musical systems has been a continuing
inspiration. I have also benefited from vigorous discussions with Robert Cook
and Michael Siciliano, who worked closely with me at a crucial point in the
development of the ideas presented here.
I am grateful to William Benjamin and an anonymous reader for energetic
readings of an initial partial draft; their impact was immense. Prof. Benjamin also
contributed an insightful reading of a later complete draft. Dmitri Tymoczko
provided a generous end-stage reading of the entire manuscript, corrected a
number of errors, and helped clarify some questions concerning the relationship
of my work to his. Steven Rings was an invaluable consultant during the
final stages of the project. Specific passages in the book benefited from friendly
communications from Robert Bowers, Candace Brower, Scott Burnham, Deborah
Burton, Adrian Childs, Thomas Christensen, David Clampitt, Suzannah Clark,
Cliff Eisen, Daniel Goldberg, Floyd Grave, Berthold Hoeckner, Christoph Hust,
Stefano Mengozzi, Tahirih Motazedian, Ian Quinn, Ramon Satyendra, and Judith
Schwartz. I regret that I am unable to honor by name other students and colleagues
whose responses to my ideas left traces that detached from their sources.
 Introduction  xvii

I am pleased to acknowledge the dedication and skill of Andrew Maillet, who

produced the graphics for the book and for the accompanying Web site, and of
Christopher Brody, a meticulous bibliographer, indexer, reference checker, and
proofreader. Their work was partly funded by a grant from the Otto Kinkeldey
Fund of the American Musicological Society, for whose assistance I am grateful. At
an early stage, I benefited from institutional support of the Franke Institute of the
Humanities at the University of Chicago, and from Philip Gossett and Janel
Mueller, both former deans of the Humanities Division there. Toshiyuki Shimada
and Brian Robinson kindly provided access to the recordings archive of the Yale
Symphony Orchestra and granted permission to use selections to accompany the
animated graphics on the companion Web site. Patrick McCreless performed the
piano excerpts that appear there and accompanied singer Paul Berry in the perfor-
mance of “Der Doppelgänger.” I am grateful to both of them, and to Mateusz
Zechowski, who recorded and edited the sound files.
Final thanks go to Heather’s mother, Meryl, whose loving gift of time made it
possible for me to finish on schedule; to my mother, Dorrit, a model writer and
teacher and an intrepid theorist; and to my daughter, Sylvia, who has tickled my
soul every day for the last seven years.
A B O U T T H E C OM PA N IO N W E B SI T E

www.oup.com/us/audaciouseuphony
The Web site contains scores for those passages that are subjects of sustained anal-
ysis, and whose length exceeds a single page. These scores are in pdf format and
may be downloaded to be viewed on screen or printed. Every effort has been made
to render the notes and rhythms as accurately as necessary for study purposes, but
the scores lack dynamics, articulations, and other performance indications. Scores
are not provided for passages discussed only briefly. Most of these scores are widely
available and also are available online, for instance, via the International Music
Score Library Project.
The Web site also contains recorded performances of some passages, coordi-
nated with graphic animations that correspond to static graphics in the book.
Recordings are limited by their availability under fair-use statutes of copyright law.
The orchestral excerpts are performed by the Yale Symphony Orchestra, under the
direction of Toshiyuko Shimada. Patrick McCreless performed the excerpts for
piano, and Paul Berry sang “Der Doppelgänger.”
References in the form “Web figure x.y” direct readers to these Web resources,
accompanied by the symbol .
User name: Music5 Password: Book1745
C HA P T E R
One
Mapping the Triadic Universe

Three Ways to Calculate Triadic Distance

It is self-evident that those keys whose scales have most notes in common are most
closely related.
—Johann Phillip Kirnberger, The Art of Strict Musical Composition, 1771

It is as though a hidden, sympathetic bond often connected the most remotely

separated keys, and as though under certain circumstances an insuperable
idiosyncrasy separated even the most closely related keys.
—E. T. A. Hoffmann, Kreisleriana, 1814

In the age of Mozart, distance between keys is linear and easily calculated. In the
age of Beethoven, the matter is more complicated, although Hoffmann (writing as
Kreisler) is unprepared to say why.
As key proximity became more complicated in the age of Beethoven, so too
did the calculation of distances between the triadic harmonies of which keys are
composed.1 This is because for theorists of the time, triadic relations tracked those
of their eponymous keys. Jean-Philippe Rameau proclaimed in 1722 that “every
note that supports a perfect chord should be considered a tonic” (Dahlhaus 1990
[1967], 28). Adolph Bernhard Marx (1841–47) “understands every consonant
triad to be ‘borrowed’ from the key in which it is the tonic, and he claims that these
triads stand in the same relation to one another as the keys they represent”
(Engebretsen 2002, 70). Hugo Riemann (1897, 86) wrote pithily that “key relation
is nothing other than the relation of their two tonic triads.” And Heinrich Schenker
collapsed the distinction altogether, regarding “keys” as triads under prolongation
(Schachter 1987).

1. Throughout this book, “triad,” in its unmodified form, refers restrictively to the twenty-four conso-
nant triads. Particular triads are identified in the standard manner, by root and mode. Roots
for minor triads are sometimes in lower case. In the figures and tables, major and minor are often
abbreviated as plus and minus signs, respectively; thus, C+ stands for C major, and c– for c minor.

1
2  Audacious Euphony

Figure 1.1. Schubert, Sonata in Bᅈ major, D. 960, 1st mvt., mm. 217–56.

To introduce the porous boundary between chord and key, as well as the com-
plicated proximity judgments at both levels of relation, consider figure 1.1, which
opens the first-movement recapitulation of Schubert’s Bᅈ major Piano Sonata
(D. 960, 1828). The first theme, which ends at m. 233, and its counterstatement,
which begins at m. 254, are separated by three spans that respectively prolong Gᅈ
major, fᅊ minor, and A major. Each span is locally diatonic. That is, within each
local span’s own context, the role of each note and chord is specified, consistently
and without ambiguity, by any of the protocols (e.g., Roman numerals, Schenkerian
graphs, Riemannian functions) that represent the detailed inner workings of dia-
tonic tonality. From a global perspective, too, the passage is normatively tonal: it
begins and ends in the tonic.
Everything that we have observed so far points to the conclusion that the music
of figure 1.1 adheres to the syntactic principles of classical diatonic tonality. It
would be premature, however, to conclude that the passage is determinately tonal
in all of its aspects. We have yet to consider how the local keys (or, from a different
perspective, the triads prolonged by the local spans) relate to one another and
how they work together to express the global tonic of Bᅈ major. If we are unable
to do so, we just have a bunch of tubs floating around on their own bottoms.
Each vessel is internally coherent and occupies a space bounded by the Bᅈ shores.
But in relation to one another, their relation is random, for all we know. And
that is no way to express a tonality. We can’t just goᇳBᅈ major, Cough, Wheeze,
Honk, Bᅈ majorᇴand pretend that we have made coherent music in Bᅈ major
(Straus 1987). If a tonal theory is to meet its claims of explanatory adequacy, it
needs to be able to specify the role, with respect to tonic, of the harmonies that
 CHAPTER 1 Mapping the Triadic Universe  3

separate the bounding tonics. The fact that each harmony may also be a tonic of its
own local context in no way relieves it of that responsibility, any more than my role
in my own home relieves me of my role in the community.
Let us locate each triad, in turn, with respect to Bᅈ major:

[1] The opening triad, Bᅈ major, is rooted on the tonic axiomatically.

[2] Gᅈ major is rooted on the flatted 6th scale degree of Bᅈ major, as notated.
We know this because we identify the cantus Bᅈ of m. 235 with the cantus
at the previous cadence, which we identified as tonic in [1] above, and we
hear the bass as a consonant third beneath it.
[3] “fᅊ minor” is a notational surrogate for gᅈ minor, rooted on the flatted 6th
degree of Bᅈ major. We know this because we identify the bass pitch at m.
239 with the bass pitch in the measures just preceding, which we described
as Gᅈ in [2] above.
[4] “A major” is a notational surrogate for Bᅈᅈ major, rooted on the flatted 1st
degree. We know this because we identify the bass pitch at m. 241 as
octave-related to the cantus pitch in the preceding measure, which was a
consonant third above the bass pitch that we identified as Gᅈ in [3] above,
and because the bass proceeds from “Fᅊ” to “A” through three steps of a
scale in m. 240.
[5] “Bᅈ major” is a surrogate for Cᅈᅈ major, rooted on the doubly flatted 2nd
degree. We know this because we identify the cantus “D” at m. 255 as a
notational surrogate for Eᅈᅈ, the proper tonic of the previous dominant
seventh, which was rooted on Bᅈᅈ, as identified in [4] above; and we hear
the bass of m. 255 as a consonant major third below that Eᅈᅈ.

But our syllogisms have led us astray! No amount of logical sophistry can dislodge
us from the conviction that the final chord of the progression represents the
tonic degree, not the doubly flatted second. There must be an error to repair.
Perhaps we can find it by retroengineering the analysis: [5] The final chord is Bᅈ
major, axiomatically; and so [4] its immediate predecessor is rooted on its leading
tone, qua dominant of its mediant, just as notated; and so [3] the chord just prior,
a minor third below the leading tone, represents the fifth degree, Fᅊ, again just
as notated; and so [2] despite its notation, “Gᅈ major” represents Fᅊ major; and so
[1] the cadence at m. 233 is on Aᅊ major.
The problem remains unrepaired. We have backed ourselves into another
corner, on the opposite side of the room. Fortunately, there are still some options
to explore. Taking them in reverse order: [4] Perhaps the roots of the last two
chords are separated by a chromatic rather than a diatonic semitone? [3] Perhaps
the root of the third chord lies an augmented second beneath that of its successor?
(But then the bass of the latter represents a different scale degree than the soprano
of the former; moreover, the stepwise approach in the bass signifies that the con-
secutive roots are not related by step.) [2] Perhaps the Fᅊ and the Gᅈ really do
represent different degrees, just as Schubert notated it? (This is implausible prima
facie: if you sing the bass while you perform the passage, nothing will persuade
4  Audacious Euphony

you to fracture the sustained pitch Gᅈ2 into two noncommunicating entities.)
[1] Perhaps the root of the second chord represents Fᅊ despite its notation? (But
then, the soprano Bᅈ4 represents a different scale degree than it did a moment ago.)
Are any of these perceptions plausible? What might motivate one to make a
case for any one of them, other than the desire to preserve the initial premise,
which is that the passage is determinately tonal in all of its aspects? If none of these
questions can be answered in the affirmative, then we can only conclude that the
passage is not entirely determined by the logic of classical diatonic tonality. This
conclusion is independent of how we choose to regard the status of the entities that
are progressing, that is, whether they are placed under the auspices of harmonic or
modulatory or linear-prolongational syntax.
We can corroborate this conclusion by means of a simple measure of diatonic
coherence: how many pairs of triads (not limiting to those presented in immediate
succession) share membership in at least one diatonic collection?2 In a typical dia-
tonic passage in major mode, a randomly selected group of four distinct triads share
membership in a single diatonic collection; ipso facto, so do the six pairs that they
form. In a passage with a single applied dominant chord, four or five of the six pairs
coexist in some diatonic collection (although not a single unified one). In figure 1.1,
only a single pair, A major/fᅊ minor, shares membership in some diatonic collec-
tion. This is, of course, very low on the spectrum of possibilities: of the 33,649
(= 23 choose 5) sextets of distinct triads that include Bᅈ major, only eight contain
fewer (= zero) common diatonic memberships.3 From a diatonic standpoint, this
progression is among the most entropic. To the extent that Schubert is employing
the logic of diatonicism here, it is in a negative sense: it is present in its absence.
We might then conclude that Schubert is being disjunctive, irrational, or arbi-
trary. To do so would place us in good company (Clark 2011a). Some critics of
Schubert’s time “described harmonic indirection as a kind of aimless wandering
towards extraneous goals, which injected a quality of randomness and lack of
plan into the music” (Shamgar 1989, 530–31). The more progressive of them
placed high aesthetic value on tonal ruptures and disjunctions, connecting their
inexplicability to the mysterious and sublime qualities so valued in the Romantic
imagination (e.g., Hoffmann 1989 [1813–14], 131–36). A related view became the
inheritance of historical musicology in its poststructural phase, for which ruptures
constitute traces of ideological, sociocultural, and psychological formations that
are otherwise occluded by the passage of historical time.4

2. I regard this measure as more suggestive than definitive. A more useful metric might additionally
track the number of pairs that share membership in some harmonic minor scale, although that
introduces other problems; for example, is {Aᅈ, B, Eᅈ} a diatonic chord in c minor?
3. These eight include Bᅈ major together with one chord from each of the following pairs: {G major,
e minor}; {E major, cᅊ minor}; {Dᅈ major, bᅈ minor}.
4. Examples include Kramer 1986, 233; Subotnik 1987; Abbate 1991; and McClary 1994, 223. Carolyn
Abbate’s analysis of a scene from Die Walküre in her Unsung Voices (1991) is a particularly fertile
garden for such tropic varietals; it refers to “harmonic irrationality and incongruence” (189),
“unstructured harmonic improvisation” (192), “cannot be heard as a logical harmonic progression”
(194), “disjunctive gap” (194), “no progress, no development . . . a repeated succession of discontinuous
chords” (199).
 CHAPTER 1 Mapping the Triadic Universe  5

But there is another available interpretation: perhaps diatonic distance is not

the best metric for the situation at hand. In a treatise published in 1796, Francesco
Galeazzi estimated the relationship between C major and d minor triads as “very
irregular and poor” (irregolarissimo e pessimo), even though each has diatonic
status when the other is tonic (Galeazzi 1796, 264).5 Yet he values the relation
between C major and E major as “regular and good” (regolare e buono), although
there is no diatonic collection that includes them both. Why does he judge the
diatonic progression less normative than the chromatic one? Because the latter has
a common tone that the former lacks. He considers the relation between C major
and e minor to be “even better” (migliore) than the previous two. Because the rela-
tion is diatonic? No—because the chords share two common tones (see Galeazzi
1796, 263–64).
Diatonic collections play no role in the model of triadic proximity that under-
lies Galeazzi’s judgments. Nor, for that matter, do harmonic roots have any role
to play (although they are implicitly present to the extent that they furnish
labels). Galeazzi’s judgments are based on properties and relations that are inde-
pendent of those identified by classical theory, such as acoustic consonance and
diatonic inclusion. If we are willing to suffer the anachronism and the scientistic
odor, we can express Galeazzi’s implicit conception in the language of modern
mathematical set theory: triadic proximity correlates with cardinality of pitch-class
intersection.
Galeazzi’s association of harmonic proximity with common-tone preservation
recurs consistently in music theory treatises throughout the nineteenth century.
K. C. F. Krause asserted in 1827 that “the most closely related consonant triads are
those that have two notes in common with the given triad, then follow those with
one note in common with the given chord” (qtd. in Engebretsen 2002, 69 n. 1).
Nora Engebretsen notes that Krause “presents his view without any fanfare, in a
manner suggesting that this is the standard approach” (69). Ten years later, Marx
offered the opinion that co-occurrence of triads in a diatonic collection counted as
a “superficial unity” but that “a more distinct tie exists in the connecting notes
which each of our chords has in common with its neighbors” (Marx 1841–47, qtd.
in Engebretsen 2002, 69). Two influential treatises from midcentury, by Moritz
Hauptmann and Hermann Helmholtz, were equally dedicated to the common-
tone basis of harmonic proximity, even though their epistemological bases (respec-
tively, in idealist philosophy and scientific empiricism) were diametrically opposed
(Hauptmann 1888 [1853], 45; Helmholtz 1885 [1877], 292). In the final decades of
the century, the common-tone view of the harmonic Verwandschaft began to lose
ground to a renewed interest in acoustic generation and consonant root relations
(Engebretsen 2008). But the two methods are nonetheless frequently seated side
by side. For example, Tchaikovsky’s 1872 Guide to the Practical Study of Harmony
distinguishes between “inner” relations, based on root distance on the circle of
fifths, and “external” connections based on common tones (Tchaikovsky 1976
[1872], 11–13; compare Riemann 1897, 85ff.).

5. For an annotated translation of Galeazzi’s treatise, see Burton and Harwood (forthcoming).
6  Audacious Euphony

Table 1.1(a). Number of common tones between each triadic pair in figure 1.1
Gᅈ major fᅊ minor A major
Bᅈ major 1 0 0
Gᅈ major 2 1
fᅊ minor 2

Applying this criterion to the Schubert passage gives a rather different picture
of its coherence. For example, the fᅊ minor triad, which is the most difficult of the
four to integrate into a Bᅈ major tonal framework, shares two common tones, the
maximum possible, with both its predecessor, Gᅈ major, and its successor, A major.
Table 1.1(a) counts the number of common tones between each pair of triads in
the progression, disregarding their order of presentation. The total of six common
tones positions the progression toward the upper end of the range for quartets of
triads, which extends from zero to nine, and well above the average, which is just
below four.6
In counting common-tone connections in a particular passage, we have implic-
itly assumed that voice leading is idealized.7 In most compositions, tones freely
transfer registers, and multioctave tone doublings liberally appear and disappear.
We say that two triads have a common tone even when, in a particular setting,
those tones appear one or more octaves apart. Identity of tones, then, is indepen-
dent of the particular register in which those tones appear. When we speak of
common tones, then, we are adopting a conception of tone that is allied with pitch
class rather than pitch. There is nothing special about idealized voice leading;
music theory teachers and scholars assume it every day of their working lives. It is
so familiar, indeed, that it takes a special effort to acknowledge it.
Idealized voice leading is also assumed by a related method for calculating the
distance between triads, which attends not only to the number of moving voices
but also to the absolute distance of motion.8 We define a unit of voice-leading work
as the motion of one voice by one semitone. The initial Schubert progression, Bᅈ
major → Gᅈ major, requires two units of work: the voices containing F and D both
move by semitone (up and down, respectively), while the voice containing Bᅈ stays
put. The progression Gᅈ major → fᅊ minor involves only a single unit of work, Bᅈ
to A (assuming no surcharge for enharmonic exchanges). And the progression
from fᅊ minor to A major involves two units of work, Fᅊ → E. Table 1.1(b) calcu-
lates the work for the six pairs of triads in the Schubert progression. Summing
the values in the table, the progression as a whole involves fourteen units of work.

6. An example of the maximum is {C major, a minor, e minor, c minor}. An example of the minimum
is {G major, eᅈ minor, Dᅈ major, a minor}; see figure 5.25(b) in chapter 5.
7. Proctor 1978 attributes the term to Godfrey Winham.
8. Not all theorists agree that voice leading should be idealized when voice-leading measurements are
assessed. Tymoczko (2005, 2009c, 2011b) presents an argument in favor of measuring voice leading
along paths in circular pitch class space, distinguishing between upward and downward motions. My
own views are flexible on this matter, in accordance with the position taken in Rings 2011, 51–54.
 CHAPTER 1 Mapping the Triadic Universe  7

Table 1.1(b). Number of semitonal displacements (“voice-leading work”)

between each triadic pair in figure 1.1
Gᅈ major fᅊ minor A major
Bᅈ major 2 3 3
Gᅈ major 1 3
fᅊ minor 2

This is on the lower end: for a set of four triads, the minimal total work is ten, the
maximum twenty-eight.9
The assumptions underlying this method of calculating triadic proximity are
even more venerable than Galeazzi’s. Already in the early fourteenth century,
Marchettus of Padua was articulating a “closest approach” preference for semitonal
voice leading (Schubert 2002, 506).10 Gioseffo Zarlino wrote that “when from the
third we wish to arrive at the unison . . . the third should always be minor—this
being closer” (1968 [1558], 79). Early-nineteenth-century theorists cultivated
melodic fluency as an alternative to fundamental bass progression (Engebretsen
2002), and at the turn of the twentieth century, Georg Capellen proposed that
triadic connections are based on a combination of common tones and semitonal
motions (Bernstein 1986, 142).
Maximum common tone retention and minimal voice-leading work are so
closely related to each other that one might be tempted to think of them as equiva-
lent. They are conflated, for example, in the “law of least motion,” which decrees
that voices should move by minimal intervals, holding common tones in the
same voice.11 This principle has the status of a robust prescription if one takes the
classical view that voice leading is secondary to harmony. If one first selects a pair
of chords and then considers how most economically to join them, maximum
common-tone preservation entails minimal voice-leading work. If, however, these
metrics serve as a primary determinant for selecting harmonies, rather than as a
criterion invoked only after the harmonies have been selected, then they do not
yield identical judgments about triadic proximity (Cook 2005, Tymoczko 2009b).
In some cases, voice-leading work makes a finer set of distinctions than does
common-tone retention, since the former spreads its results across six distinct
values whereas the latter returns only three (see figure 4.7 in chapter 4). For
example, in figure 1.1, fᅊ minor shares two common tones with both the
preceding Gᅈ major and the subsequent A major. Yet the moving voice travels by
semitone in the first case, whole tone in the second, a distinction that disappears
when one is merely counting common tones. In other cases the two metrics make

9. An example of the minimum is {Bᅈ major, bᅈ minor, Gᅈ major, fᅊ minor} (see chapter 2). The
maximum is fulfilled by {G major, eᅈ minor, Dᅈ major, a minor}, which is the minimum of note 6.
10. Dahlhaus (1990 [1967], 335 n. 7) speculates on even earlier origin.
11. The law of least motion was erroneously deposited in Arnold Schoenberg’s theoretical account, but,
like so many of the other treasures banked there (the chart of regions from Weber, the emancipa-
tion of dissonance from Weitzmann), it was siphoned from the accounts of predecessors. The “law”
was a staple of thoroughbass theory and debuted no later than Charles Masson’s 1694 treatise.
8  Audacious Euphony

contradictory claims about relative proximity. Is C major closer to g minor or to

gᅊ minor? The former preserves a common tone where the latter has none, and it
is thus closer on one criterion. But the latter involves three units of voice-leading
work ({C, B} = 1; {E, Dᅊ} = 1; {G, Gᅊ} = 1), whereas the former involves four
({C, Bᅈ = 2; {E, D} = 2; {G, G} = 0), producing a proximity judgment that contradicts
the previous case.
In summary, we have reviewed three distinct metrics, each of which formalizes
a different set of intuitions about triadic proximity. The classical metric evaluates
triadic proximity in terms of mutual membership in diatonic collections and inter-
prets figure 1.1 as very disjunct. The same passage is interpreted by Galeazzi’s
common-tone metric to be fairly conjunct, and by the voice-leading metric to be
very conjunct. The diatonic collection, which plays a central role in the first metric,
has no privilege whatsoever in the remaining two. In the voice-leading metric,
which most successfully captures the intuition that the triads in Schubert’s
progression inhabit a similar neighborhood, it is the chromatic collection that
explicitly comes forward as the template against which distance is assessed.

Triads in Chromatic Space

To view consonant triads against the background of chromatic space is to decline

to interpret them in terms of the number of diatonic degrees that separate their
root from some tonic. This choice cuts against the multiple denominations of
classical tonal theory and their pedagogical offshoots, which all teach that chro-
matic harmonies are primarily to be understood as transformations of some
underlying diatonic one. The idea that the diatonic collection conceptually
precedes and regulates the interpretation of the chromatic one, already implicit in
the names of notes, their position on the staff, and the system of key signatures,
became canonized with respect to classical tonality in the early nineteenth century,
at roughly the same historical moment that musical education became institution-
alized in conservatories, analysis evolved into its own discipline, a theory of
tonality began to congeal under that name, and Roman numerals became the
default first-level descriptors for triads (Wason 1985, 53). That idea has proven
hardy indeed, as can be confirmed with reference to any English-language
harmony textbook.
The diatonic view of chromaticism has prevailed for good reason. Triads and
diatonic scales together constitute the foundational organizing materials of classi-
cal tonality. Although the diverse traditions of classical theory assign consonant
triads and diatonic scales different values in relation to each other (Dahlhaus 1990
[1967]), they all agree that it is through their coordination that major and minor
keys are established. For an acculturated listener, a major or minor triad, sounded
in isolation and without prior context, signals the tonic status of its root by default.
In a process first described by Gottfried Weber (1846 [1817–21]), a listener
spontaneously imagines an isolated triad housed within a diatonic collection,
signifying a tonic that bears its name.
 CHAPTER 1 Mapping the Triadic Universe  9

Yet there is a difference between a default interpretation and a necessary

one. The tonic status of a triad requires confirmation, weakly through the remain-
ing tones of its associated diatonic collection; more strongly by arranging those
tones into a local cadence; more strongly yet by repeating that cadence, perhaps
with supplementary rhetorical packaging, at the end of the movement or composi-
tion. Such a confirmation is by no means inevitable. Sometimes an initial triad
comes to be understood as representing a nontonic degree (Kirnberger 1982
[1771–76], 45; Aldwell and Schachter 1989, 135). And sometimes it initiates a
succession of triads, any of which could claim tonic status under appropriate
conditions (Kirnberger 1982, 114; Schenker 1954 [1906], 254). The more tones put
in play, the less likely their alignment with respect to a diatonic collection that
organizes their position and role with respect to some tonic. Until such a collec-
tion emerges and is cadentially crowned, the triadic progressions are diatonically
indeterminate.
Once essential enharmonic relations arise, indeterminacy evolves into contra-
diction. By essential enharmonicism, I am excluding those notated enharmonic
conversions that arise as artifacts of notational pragmatics, as might happen if a
composer presumes that a performer is more comfortable reading a signature of
four sharps than eight flats. What distinguishes essential enharmonicism is that
the composer has no choice but to convert between sharps and flats in order to
retain global diatonic logic (Schenker 1954 [1906], 333–34). In such cases, the
exact point where the composer notates the conversion is a pragmatic matter
without significance; in a phenomenological sense, such a conversion happens
everywhere and nowhere, which is tantamount to saying that it is distributed
evenly across all of the possible moments when it could occur (Proctor 1978, 177;
Telesco 1998; Harrison 2002a). When enharmonically paired pitch classes are jux-
taposed directly, the ear cannot avoid identifying them as the product of a single
tone. As we saw with reference to figure 1.1, the splitting of a tone’s scale-degree
constituency can have a ripple effect, destabilizing the diatonic collection and the
tonic that is claimed to anchor it.
The recognition that triadic music is not always fully determined by the
principles of diatonic tonality is by no means a new one. As already noted, early-
nineteenth-century critics intuited that contemporary music defied familiar logic
(although they disagreed as to whether this was a good thing). The impulse to
systematize these intuitions was first acted on in the writings of François-Joseph
Fétis, a Belgian music critic working in Paris, initially in a series of articles from
1832, eventually in his 1844 harmony treatise. Fétis divided tonality into four sty-
listic species, each representing a stage of historical development, and each defined
by its own syntactic principles and affective properties. The most progressive of the
four species, omnitonality, is distinguished by a proliferation of enharmonic rela-
tions that indicate a “multiplicity, or even the universality of the keys” (Fétis 2008
[1844], 190), a process that Fétis predicted would lead to “the total destruction of
the scale in certain cases, and the beginnings of an acoustic division of the musical
scale into twelve equal semitones” (Berry 2004, 257, quoting Fétis 1832). For Fétis,
the objects of omnitonality are chromatically intensified dissonant harmonies,
rather than the consonant triads that concern us in the present study. It is rather in
10  Audacious Euphony

the historically earliest of Fétis’s four species, unitonality, that one finds tonally
indeterminate chromatic successions of triads, as in some music of Marenzio
and Gesualdo from the turn of the seventeenth century. Such successions fail to
define a key because their constituent triads do not communicate with each other:
“No attraction is evident, because every perfect chord is a harmony of repose”
(Fétis 2008 [1844], 163). If every chord is a potential tonic, then no chord can
fulfill that potential by functioning as one. Each tub is on its own bottom, bobbing
around the sea independently of the others.
Although each of Fétis’s four tonal species arises at a particular historical
moment, the later species do not supplant the earlier ones. According to his his-
torical model, they are cumulative; once available, the best composers know how
to combine them in a single work of art (Berry 2004, 255). It is evident, then, that
Fétis conceives of classical tonality (“transitonality”) as a category whose constitu-
ent elements are not integral “pieces”—compositions or complete movements—
but rather musical moments. On Fétis’s view, the faculty of (transi)tonal listening
is capable of spontaneous suspension and reengagement without notice or fuss,
like a carpenter exchanging a screwdriver for a hammer. He recognizes a similar
dynamic in a purely diatonic environment, as when a sequence arises midphrase.
At the moment that the sequence is recognized, the “law of tonality” is placed in
abeyance, as our cognition is submitted to a “law of uniformity.” “The mind,
absorbed in the contemplation of the progressive series, momentarily loses the
feeling of tonality, and regains it only at the final cadence, where the normal order
is reestablished” (Fétis 2008 [1844], 27).
The idea of simultaneously accessible tonal schemata was developed specifi-
cally with relation to pan-triadic progressions seventy-five years later by Ernst
Kurth, who was raised in an era of rampant, fully ramified omnitonal chromati-
cism that Fétis could only divine. Kurth’s Romantische Harmonik und ihre Krise in
Wagners “Tristan,” initially published in 1920, proposed that many chromatic
progressions, particularly those that involved root relations by third, introduced
rifts, wedges, and fissures into the fabric of tonality. The identity and function of
these chords are found in their internal structure and in their local connections to
their immediate antecedents and successors. When concatenated with sufficient
intensity and persistence, such absolute progressions bring about “the total disrup-
tion of the original embracing tonal unity” (Kurth 1991, 120). Kurth discovered an
agent of tonal disruption in chromatic sequences, which, like Fétis’s diatonic ones,
are governed by the logic of repetition. Such progressions are “extratonal” in the
sense that their relation to the tonal pillars that bound them on either side is not
tonally determined.
After Kurth’s 1920 treatise, it became a commonplace of German musicology
that neither the appearance of consonant triads nor their framing by occasional
cadential progressions was sufficient to justify the judgment that their syntax was
governed by the principles of classical tonality; other factors were necessary in
addition (Adorno 1964; Kunze 1970; Dahlhaus 1980a [1974]; Motte 1976). Among
the adherents of this view were Theodor Adorno and Carl Dahlhaus, both of
whom eventually acquired a significant readership in North America, one result of
which was that Kurth’s views immigrated into the arena of American musicology
 CHAPTER 1 Mapping the Triadic Universe  11

(e.g., Newcomb 1981; Meyer 1989, 302; Agawu 1989, 27; Abbate 1991, 192). Similar
views can also be found, perhaps surprisingly, in the writings of American
Schenkerians, who otherwise are committed to the vision of the masterwork as
organically unified by Ursatz emanations that function uniformly at all composi-
tional levels. These included Adele Katz, for whom the Magic Sleep music from
Die Walküre “lack[s] . . . tonal implication” (1945, 213); William J. Mitchell, who
noted that a triadic circle of fifths “can be arrested at any point or it can just as
easily go on in perpetuity” (1962, 9); and Felix Salzer and Carl Schachter, who
wrote that “we register the equal intervallic progressions without referring them to
a supposed diatonic original. This temporary lack of a diatonic frame of reference
creates, as it were, a suspension of tonal gravity” (1969, 215).
The dissemination of this view has not, however, dislodged a broadly shared
commitment to the notion that the chromatic triadic progressions characteristic
of the nineteenth century are determined by their position with respect to some
tonal center. This commitment is evident not just in the profusion of inflected
Roman numerals or function symbols that dominate the textbook teaching of
nineteenth-century harmony on both sides of the Atlantic. It also dominates vari-
ous branches of research, whether based in Roman-numeral/fundamental bass
traditions (Lerdahl 2001), Schenkerian/linear approaches (Darcy 1993; Brown
2005), Riemannian functions (Harrison 1994), or Lewinian transformations
(Kopp 2002). Although these denominations interpret triadic harmony according
to quite different sets of assumptions, and express those interpretations using dis-
tinct modes of representation, they all share a base in the late-eighteenth-century
classical harmonium, from which they reach out to lay claim to the chromatic
triadic music of the nineteenth century.
I can think of three reasons that analysts of nineteenth-century triadic music
have continued to dance to a modified eighteenth-century beat, despite the many
stumbles induced by the terrain. First is the promiscuity of triadic descriptive
categories, combined with the illusion that to describe is to explain. Roman numer-
als are flexible enough to furnish a first-level description of almost any triad in
almost any key (Dahlhaus 1980a [1974], 68; Hyer 1989, 229–30). Many Roman
numeral practices are satisfied, moreover, with finding a local tonic for each har-
mony, without any demand that local tonics be reconciled to each other and to a
global tonic. Riemannian functions likewise are catchall categories, such that
“a student of Riemann’s system can analyze virtually any chord into any one
of the three functions should the occasion demand” (Harrison 1994, 284).
Schenkerian approaches allow chromatic triads to degrade into coordinated linear
spans (Benjamin 1976; Smith 1986), which serve as carpets under which to sweep
enharmonic paradoxes.
A second reason for the continued resistance to alternative views of triadic
chromaticism is that it requires an embrace of some form of double syntax. Most
nineteenth-century passages that can be seen to juxtapose triads according to
nonclassical principles exist in close proximity to other behaviors that are normal
under classical diatonic tonality. The Schubert excerpt with which we began
(figure 1.1) is not atypical: while the local spans are classically tonal, the middle-
ground tonics adhere to a different logic. To analyze such a composition requires
12  Audacious Euphony

not only that we navigate, sometimes in rapid alternation, between two or more
syntaxes, as Fétis imagined listeners moving between his four kinds of tonality. It
also requires the capacity to simultaneously process two distinct sets of syntactic
principles that unscroll at different speeds. Can music of high aesthetic value really
partake of two systemic modes of organization, shuttle between them quasi instan-
taneously, and even overlay them? Are our musical brains wired in such a way that
we have the capacity to shift between these syntaxes as if at the click of a switch, or
to multitask between them? If the responses of several prominent music scholars
are representative, it seems that there is a strong motivation to reject any such idea
on a priori grounds, which is to say, independently of the details of the proposal
under which a double syntax program might be carried out (Dahlhaus 1990 [1967],
111; Smith 1986, 109; Lerdahl 2001, 85). Chapter 9 considers and responds to
this line of objection; readers who share this prima facie skepticism may wish to
teleport there before proceeding with the linear exposition.
The final reason pertains to the absence of a fully ramified alternative. We are
inclined to come out from under familiar technologies only when we are prepared
to substitute for them an alternative that is plausible, coherent, and productive.
To acknowledge that chromatic progressions of triads might be based in some
syntactic principles other than those of diatonic tonality is to clear a space, but
that is not the same thing as building a house. One needs to be able to say some-
thing about what that syntax is, not just what it is not. To say that “Beethoven’s
third period seemed destined to shake the absolutist regime of the main tonality
for the first time” (Draeseke 1987 [1861], 315) or that some Wagnerian progres-
sions “stand . . . in certain opposition to tonal unity” (Kurth 1923, 249; my
translation) constitutes a necessary first step. To allow for the existence of a
“countersyntax” that stands in “dialectical” relation to classical tonality (Kramer
1986) constitutes a significant second one. But to posit the terms of that counter-
syntax, it is necessary to do more than substitute a Latin adjective for its Greek
equivalent (as occurs whenever a writer feels that they have scratched an explana-
tory itch when they have attributed a chromatic harmony to a “coloristic” effect),12
or refer to linear processes without being prepared to specify anything beyond
pointing to lots of semitones (e.g., Dahlhaus 1980a [1974], Agawu 1989). In
addition, one wants to know what principles underlie the syntax, how it
operates, how its analyses are represented. Are its claims consistent, well formed,
and free of internal contradiction? How is the syntax motivated by the lexicon;
that is, what properties do triads possess that qualify them for the job that (the
syntax claims) nineteenth-century composers put them up to perform? What
sorts of problems does this syntax help solve? Does it generate analyses that
reflect some aspect, however obliquely and abstractly, of a musician’s or listener’s
experience? Does it lead us to notice interesting things about a score, or about its
relationship to other scores, that would have otherwise escaped attention? Does
it help us think differently about historical problems of genre, style, evolution,
12. See Tischler 1964, 233; Rosen 1980, 245; Kramer 1986, 203; Todd 1988, 94; Meyer 1989, 299; Ratner
1992, 113; Somer 1995, 219; and Taruskin 2005, 69. “This chromaticism has a coloristic effect” has
roughly the propositional status and explanatory value of “this box is so heavy because it weighs a
lot.” David Kopp (1995, 345) makes a similar point.
 CHAPTER 1 Mapping the Triadic Universe  13

and the like, or about the relationship between music and the historical conditions
of the individual, society, or culture that produced it?
This book responds to these questions by adapting a conceptual framework
erected between 1955 and 1980 by the field of atonal pitch-class theory, whose
great achievement was to develop a systematic approach for exploring the proper-
ties, potentials, and interrelations of chords (“sets”) within the chromatic universe.
Atonal theorists of that era were not much interested in consonant triads, as their
analytic interests were focused on a repertory whose principal phonological con-
straint was, on some accounts, their absence (Boulez 1971 [1963]; Forte 1972; but
see Straus 1990). Reciprocally, music scholars of that era who were open to the
cultivation of alternative approaches to nineteenth-century triadic music were
alienated from American atonal theory because of geography, the serendipities of
disciplinary configuration, or the low priority that cold-war theorists placed on
disciplinary outreach. In exploring the properties and potentials of consonant
triads using a method adapted from atonal theory, I hope to defuse the suspicion
that “applying analytical techniques derived from contemporary music” to late-
Romantic repertory is “menial and easily accomplished” (Dahlhaus 1989 [1980],
381–82), or an act of desperation (Harrison 1994, 2).

Remarks on Syntax and Maps

Syntax is a central term in the study of natural language, and not all of the mean-
ings that it accumulates there can be transferred into music. Syntax is that branch
of linguistics that studies how words and their constituent particles combine to
form coherent sentences, independently (in principle) of how those sentences rep-
resent concepts and states, or motivate actions, in the world. I use syntax in this
book in three different ways. First, syntax contrasts with phonology and lexicon,
which respectively treat the internal structure of atomistic units and their first-
level bundling into units of signification or reference. Because music, under ordi-
nary conditions, lacks the referential dimension of language, phonology and
lexicon come close to fusing: a lexicon is a list of available sounds (chord, scales,
sets), and phonology provides a principled account of what properties make those
sounds available for use. Second, syntax is the study of the ordering of events as
they sequentially unfold in time: how triads “progress” in a moment-by-moment
sense, and perhaps also in a middleground sense where such interpretations are
appropriate. Third, and most important for present purposes, I use syntax in the
same sense that the Greeks used harmonia, the “means of codifying the relation-
ship between those notes that constituted the framework of the tonal system”
(Dahlhaus 1980b, 175). This broader domain is roughly equivalent to what Roger
Sessions (1950, 33) designated as “the relationships between tones, and . . . the
organization which the ear deduces . . . from those relationships” and what David
Lewin (1969, 61) characterized as the way that “sound [is] conceptually struc-
tured, categorically prior to any one specific piece.” It is at this most abstract level
that we can also refer to Fétis’s four types of tonality as evincing distinct syntaxes,
14  Audacious Euphony

or think of his laws of tonality and uniformity as manifesting distinct syntactic

principles. They are distinct in the sense that they generate different orderings of
the harmonies, doublings of chords, and expectations about dissonance treatment.
But they are also distinct in the sense that they evoke different modes of musical
cognition.
In this third sense, musical syntax has long benefited from geometric and
graphical representation. Geometric models of pitch space have been in use for
some 1,500 years (Popovic 1992; Westergaard 1996). During the eighteenth and
nineteenth centuries, they were frequently applied to relations among keys, and
later among chords. American music theory of the postwar era generally favored
algebraic models, for their compactness (a significant consideration in print
media) and their strong generalizing capacity. A surge in geometric models began
in the 1980s, and has intensified in the last decade, in part due to the increasing
accessibility of graphics software and the economies afforded by electronic
space. Geometric models can thrive as effective modes of exploration and com-
munication only if the phenomenon being modeled meets certain structural and
psychological conditions. Structural problems arise if there are more conceptual
dimensions than are available in the physical medium. In linear space, one dimen-
sion is great, two’s fine, three’s the limit, and four blows the mind. In cyclic space,
even a second dimension introduces falsifications and distortions, like the Bering
Strait problem familiar from Eurocentric world maps. Moreover, Euclid’s logic
often collides with that of the psyche. The symmetry of spatial distance may lack
psychological salience for someone walking uphill, and the triangle inequality
prohibition is violated whenever two miles walked in intense conversation feels
shorter than one mile alone on a sore ankle. Fortunately for my project, in the case
of triadic distance measurements these problems are kept to a minimum. The
cyclic structure of chromatic space will create some Bering Strait problems, but we
will find that these are easily negotiable with the help of a supplementary “legend”
that guides interpretation of the map.
The supreme advantage afforded by musical maps is their capacity to reflect
judgments about the psychological proximity of musical objects or states (Popovic
1992). Elementally, such judgments come in binary form (“these two notes sound
close, those two sound distant”) that lead naturally to comparison (‘“these notes
sound closer together than those”). When structural and psychological conditions
align, a map has the capacity to draw together a family of pairwise distance assess-
ments. Such a map then acquires the capacity to capture syntactic judgments,
which might take the form of conjunct versus disjunct, normal versus unusual,
or acceptable versus unacceptable. Moreover, it earns the potential to aid in the
exploration of semantic predicates, such as “betweenness” (there is a gap that
one expects will be filled), “orientation” (we can chart our distance from and
direction with respect to some “home”), or “momentum” (there is a pattern whose
continuation we anticipate).
Like a geographical map, a good representation of musical space does not
merely sit there, as a static structure. It acts as a stage upon which imaginative per-
formances are mounted, thus serving the same function as a geographical map for
a child with a toy car, or for a medieval monk tracking a crusade (Connolly 1999).
 CHAPTER 1 Mapping the Triadic Universe  15

A musical map can illuminate compositional decisions as selections from a finite

menu. It can move composers to ask, “How many ways are there to connect these
two chords?” “What chord stands halfway between these two chords?” “How can
I form a cycle, with some desired number of elements, that begins and ends at the
same element?” or “If I’m at A, what state B should ensue, if I want to mimic the
gesture that carried Q to R?” It invites analysts to ask, “Is this a step or a leap?” “Is
this connection the most direct one?” “Are these two paths parallel?” or “Is this
path an embellishment of that one?” Questions of this type are concerned with
corpuswide ideals, norms, and limitations, as well as with “motivic” elements that
shape and individuate a particular composition in dialogue with those norms
and ideals. These are the musical equivalents of langue and parole, language and
utterance, the topics most central to the syntactic study of natural language.
Steven Rings (2006) suspects me of using coherence, in related contexts, as a
fourth-order stalking horse for the universalization of nineteenth-century German
aesthetic ideology, by way of the intermediate terms “unity,” “autonomy of the
artwork,” and so forth, and he and others may suspect that syntax just heaps
another shell or two of derivatives on top. I am not committed to either italicized
term and would by happy for readers simply to substitute some other, or perhaps
some neutral, term (e.g., X-factor) in their place. I do think that there is some
profit in acknowledging that, among communities, some musical phenomena
“go down easy” and some “go down hard.” Asyntactic and incoherent signify the
neighborhood of aesthetic responses that might alternatively take the form “doesn’t
make sense,” “sounds weird,” “sounds erratic,” “doesn’t fit,” “sounds random,”
“sounds awful,” “I don’t get it,” “I wasn’t expecting that,” “that’s not normal,” “not
immediately intelligible.” Someone with a historical sense might posit those same
responses through comparison, similar to the way that Forkel responded to that
odd passage from C. P. E. Bach’s f minor Piano Sonata (Kramer 2008, 11), or that
some Viennese critics heard Schubert’s modulations (Shamgar 1989), or that the
first European heard South Asian music or the first Indian heard European music.
Although I don’t care what term is used to make the distinction, I am pretty sure
that there is a distinction to be made.
I would even go so far as to suggest that that distinction is universal, on the
hypothesis that, for individuals or communities or cultures, there are things that
make sense and things that don’t, things that go down easy and things that go
down hard, things that are familiar and things that are foreign, and so forth. An
anthropologist might make this distinction with the term emic (fits the world-view
of the folks who live there), an intellectual historian with episteme (fits the world-
view of the folks who lived then), and a linguist with syntactic (has potential mean-
ing within that linguistic community). What I mean by “syntactic,” then, is the
musical equivalent of all of those.
This page intentionally left blank
C HA P T E R
Two
Hexatonic Cycles

Chapter 1 proposed that triads could be related by voice leading, independently of

roots, diatonic collections, and other central premises of classical theory. This
chapter pursues that proposal, considering two triads to be closely related if they
share two common tones and their remaining tones are separated by semitone.
Motion between them thus involves a single unit of work. Positioning each triad
beside its closest relations produces a preliminary map of the triadic universe. The
map serves some analytical purposes, which are explored in this chapter. Because
it is not fully connected, it will be supplemented with other relations developed in
chapters 4 and 5.
The simplicity of the model is a pedagogical advantage, as it presents a circum-
scribed environment in which to develop some central concepts, terms, and modes
of representation that are used throughout the book. The model highlights the
central role of what is traditionally called the chromatic major-third relation,
although that relation is theorized here without reference to harmonic roots. It
draws attention to the contrary-motion property that is inherent in and exclusive
to triadic pairs in that relation. That property, I argue, underlies the association of
chromatic major-third relations with supernatural phenomena and altered states
of consciousness in the early nineteenth century. Finally, the model is sufficient to
provide preliminary support for the central theoretical claim of this study: that the
capacity for minimal voice leading between chords of a single type is a special
property of consonant triads, resulting from their status as minimal perturbations
of perfectly even augmented triads. The consequences of that claim are the focus
of the final sections of this chapter.

A Minimal-Work Model of the Triadic Universe

We will say that two triads are in the minimal-work relation if motion between
them involves the displacement of a single voice by semitone. According to this

17
18  Audacious Euphony

definition, each triad is in the minimal-work relation to two triads of the opposite
mode. Each major triad is in the minimal-work relation with its parallel minor and
with the minor triad whose root lies four semitones above it. For example, C major
is in the specified relation with c minor and with e minor. Reciprocally, each minor
triad is in the specified relation with its parallel major and with the major triad
whose root lies four semitones below it. For example, c minor can reach both C
major and Aᅈ major by a single semitonal displacement.
Each consonant triad is thus situated in a chain of alternating major and minor
triads. C major is flanked by c minor and e minor, producing the three-element
chain {c minor, C major, e minor}. That trio is nested within a five-element chain,
{Aᅈ major, {c minor, C major, e minor}, E major}. That quintet is, in turn, nested
within a seven-element chain, {aᅈ minor, {Aᅈ major, {c minor, C major, e minor},
E major}, gᅊ minor}. The external elements of that septet are identified under
enharmonic equivalence, and so the outer elements are “glued together,” convert-
ing the seven-element chain into a six-element cycle.
Figure 2.1 presents that cycle in its upper quadrant, along with three other
cycles, each germinated in an analogous way from other triadic seeds. Together,
the four cycles partition the twenty-four consonant triads.1 Connections among
the cycles will emerge as a central project of chapters 5 and 6. This chapter focuses
on the internal workings of the individual cycles. Because they relate to each other
by transposition, internal features proper to one are proper to all.
Two triads are situated in the same cycle if they have the same roots, or if their
roots are four semitones apart. Classical theory presents various languages for
identifying the collections of co-cyclic triadic roots. We might say that they form

Figure 2.1. A graph of the twenty-four triads under single semitonal displacement,
producing the four hexatonic cycles.

1. Derek Waller called attention to these cycles in a 1978 article in the Mathematical Gazette, which
anticipated some aspects of my presentation in Cohn 1996.
 CHAPTER 2 Hexatonic Cycles  19

an augmented triad. If we are concerned with projecting neutrality about enhar-

monic choices, we might say that the roots form a symmetric division of the octave
into three equal parts. In circumnavigating the cycle in either direction, the
progression consists of paired triads transposed by major third, or (again more
neutrally) by an interval of four semitones. We might then be inclined to refer to it
as a T4 or T8 cycle, where “T” refers to transposition and the subscripted number
indicates the size of the transposition, measured in semitones. We might be
inclined to view this progression in terms of a chromatic sequence whose generat-
ing interval is the major third or minor sixth, or whose generating transpositions
are T4 and T8. All of these characterizations are more or less interchangeable, and
I shall treat them as equivalent.
It has become standard to view the cyclic progressions portrayed in figure 2.1
in terms of “third relations” and thereby to throw them into the same pot with
progressions that transpose by a series of minor thirds. This conflation, which is
encouraged both by the labeling conventions for intervals and by the appearance
of both types of cycles at roughly the same historical moment, obscures a funda-
mental distinction: cycles generated by major thirds exhibit balanced voice leading,
alternating between up and down, whereas those generated by minor thirds lead
their voices in a uniform direction. For example, in transposing by major third
from C major to E major, under idealized voice leading one voice moves up
(G → Gᅊ) and one moves down (C → B), while the third voice, E, holds its place. In
transposing by minor third from C major to Eᅈ major, again under idealized voice
leading, both moving voices move down (C → Bᅈ and E → Eᅈ) while G holds its
place. The latter case is the norm. Under least-motion voice leading, recursive
transposition by any interval other than major third generates uniformly directed
voice leading. It is transposition by four or eight semitones that is special: these
alone generate transposition cycles whose voice leading is balanced.2
A significant entailment of balanced voice leading is contrary motion: for every
voice that rises, another falls by the same magnitude. Contrary motion among
triads of the same mode thus inheres exclusively to triads that are transpositionally
related by major third (Cohn 1996). This entailment, which is at the heart of the
approach developed in this book, has gone unrecognized by the many scholars
who have studied third relations during the last thirty years.3 I will argue, in this
chapter and beyond, that the contrary motion of major-third relations underlies
both their central role in the syntax of pan-triadic progressions and their associa-
tion with the semiotics of the supernatural. The reason that the major third has
this special status is that it divides the octave into as many equal parts as the triad
has notes. One job of this chapter is to show why.

2. Balanced voice leading corresponds to balls atop an arch. Some released balls will fall east, some
west, and the behavior of one ball does not predict the behavior of its successor. Uniform voice
leading corresponds to the situation of balls on a tilted plane. If one ball falls east, they all fall east;
one can predict the behavior of all from the behavior of any one.
3. These include Proctor 1978; Krebs 1980; Taruskin 1985, 1996, 2005; Bailey 1985; Cinnamon 1986;
Aldwell and Schachter 1989; Agawu 1989; Kraus 1990; Rosen 1995; Somer 1995; Kopp 2002; and
Bribitzer-Stull 2006. Tymoczko 2011b is a recent and notable exception.
20  Audacious Euphony

Figure 2.2. Voice leading through a hexatonic cycle. Arrowheads indicate the
direction of semitonal displacement between adjacent triads.

The Hexatonic Trance

Figure 2.2 progresses clockwise about one of the cycles from figure 2.1, in a format
that emphasizes the behavior of the individual voices. The repeat sign suggests a
continuously cyclic process. (The features that we identify in this progression will
also be present in its retrograde, as voice-leading features are independent of cyclic
direction.) Arrowheads indicate the location and direction of semitonal displace-
ments. Each individual voice holds its pitch for three “beats” and then displaces to
a new pitch that is likewise sustained for three beats before returning to the origi-
nal pitch. The entire cycle engages only six pitches, two for each of its three voices.
Accordingly, the progression is referred to as a hexatonic cycle. Ordered linearly
within an octave, the six tones form a hexatonic scale, alternating semitone and
minor third.4
The triple periodicity within each individual voice interlocks with the duple
periodicity resulting from their combination. This duple periodicity has both a
directly perceptible aspect and a more abstract structural one. The direct aspect
emerges in the periodic alternation of upward and downward motion, from one
“beat” to the next. The more abstract aspect emerges from the relations between
the three voices, each of which executes its displacements two beats later and four
semitones lower than its predecessor (assuming octave equivalence). The voices
thus combine to form a hocket canon, a structure familiar to music historians
from the caccia of late-medieval polyphony (Bukofzer 1940). The interlocking
of duple and triple periodicities, induced respectively within and between the
individual voices, forms a 3:2 phasing (also referred to variously as polyrhythm,
grouping dissonance, hemiola, and cross pattern). (Readers may find it useful
at this point to kinetically engrave and aurally entrain these periodicities by play-
ing through the repeating hexatonic cycle with one hand at a keyboard, or by
arpeggiating the successive triads on a single-line instrument.)
Figure 2.3 juxtaposes nonadjacent triads of the same hexatonic cycle.
Next-adjacent triads are of the same mode; either major (a) or minor (b).

4. Jazz players know the same pattern as the augmented scale, because the collection combines two
adjacent augmented triads. It has also been referred to variously as the Ode to Napoleon collection,
Miracle hexachord, Liszt model, source-set E, 1:3 collection, and set class 6–20.
 CHAPTER 2 Hexatonic Cycles  21

Figure 2.3. Voice leading between nonadjacent triads within a hexatonic cycle.

Diametric triads, three positions apart, are of opposite mode (c). The number of
semitonal displacements in each progression is equivalent to the cyclic distance of
its constituents. In figure 2.3, (a) and (b) show the double displacements that occur
in major-to-major and minor-to-minor juxtapositions, and (c) shows the triple
displacement between nonadjacent triads of opposite mode. All three progres-
sions involve balanced voice leading, in the sense that at least one voice moves in
each direction. These progressions can be seen to compress the serial alternation
of up and down in figure 2.2, transforming the directional alternation into a
simultaneous contrary motion.
In music of the nineteenth century, and throughout the history of music for
film, the progressions illustrated in figure 2.3 frequently depict sublime, super-
natural, or exotic phenomena. In “Nacht und Traüme,” Schubert slips directly
from B major to G major to depict a nocturnal fixation on evanescent dreams
(Schachter 1983b). In the Ring, Wagner uses the progression at figure 2.3(b) to
portray the Tarnhelm, which makes its wearer invisible. Rimsky-Korsakov used
the same progression, transposed through the minor triads of a hexatonic cycle, to
depict the reclusive Antar adrift on the sands of the Sahara desert (see figure 3.6
in chapter 3). Wagner subsequently used the same cycle in Parsifal to depict the
sorcerer Klingsor, another recluse in the Arabian wasteland. The diametric pro-
gression at figure 2.3(c) depicts a range of uncanny phenomena. The many exam-
ples cited in Cohn 2004 include Kundry’s de-souling (Parsifal), the dead Siegfried
shaking his fist (Götterdämmerung), Scarpia’s murder (Tosca), Aase’s arrival at
St. Peter’s Gates (Peer Gynt), Strauss’s Salomé singing to the severed head of
Jochanaan, and Schoenberg’s self-portrait in death (String Trio, Op. 45).
In a chapter titled “Music and Trance,” Richard Taruskin (2005) recognizes the
semantic charge of chromatic progressions by major third but implies that their
link to altered and uncanny states is conventional, relying on what Swiss linguist
Ferdinand de Saussure referred to as the arbitrary bond between signifier and
signified.5 Although convention is certainly an element of this semiotic system,
there is something else at work: the affective power of the progressions in figure 2.3
derives from a paradoxical characteristic that is inherent to them, when they are

5. Taruskin’s association of chromatic progressions by major third with trance, the uncanny, etc., seems
to be limited to the major mode, in which all of his examples occur. In minor, the submediant is
“naturally” flat, and so ᅈVI and “flat submediant” are not meaningful designations.
22  Audacious Euphony

heard against the expectations of classical diatonic tonality. By default, the classi-
cally conditioned ear interprets a relation between two tones as diatonic rather
than chromatic (Agmon 1986, 185; Temperley 2001, 128). The empty ear filling
with music interprets a two-semitone interval as a major second rather than a
diminished third, a three-semitone interval as a minor third rather than an aug-
mented second, and so forth. (A canny composer cultivates various resources for
reversing these defaults, just as an engineer can raise objects in physical space; but
they nonetheless hold in an “everything else equal” context.) This same principle
dictates that a single semitone be heard, again ceteris paribus, as a change of
diatonic degree, rather than as a chromatic inflection of an invariant degree.
Applied to a perfect fifth whose voices move outward by semitone, as in the
outer voices of the three progressions at figure 2.3, the default principle dictates
that the interval between them increase by two diatonic degrees, producing a
diminished seventh. Yet when heard in their own insular context, the default prin-
ciple dictates that two tones separated by an interval of nine semitones express a
major sixth.6 The principle thus produces contradictory information: on the first
application, the nine-semitone interval is dissonant; on the second, it is consonant.
In the attempt to reconcile these interpretations, the ear is caught in a liminal
space, where the binary distinction between consonance and dissonance is eroded.
Such breakdowns in the division between otherwise securely demarcated catego-
ries, prototypically the boundary between reality and illusion, or life and death, are
a mark of the psychological uncanny.7 The capacity of chromatic root progressions
by major third to signify altered and unstable mental states is thus based not on
mere convention but on a homology between the signifying progression and the
signified affect. In Peircean terms, the progressions in figure 2.3 are icons rather
than symbols of altered or destabilized mental states.
Consider, for example, the Tarnhelm progression (figure 2.4), whose first two
chords match those of figure 2.3(b). Warren Darcy (1993, 170) writes of the “aural
sense of . . . eerie power” and of “radical disjunction. . . . the motif seems almost to
have fallen in from another world.” For Carolyn Abbate (2006, qtd. in Parly 2009,
166), that world is subterranean, “as if excavated from primeval time.” Nila Parly
(2009, 167) attributes this effect to the open B/Fᅊ fifth, an archaic symbol that
evokes an “air of something ‘uncanny,’” by virtue of incongruity, when embedded
into Wagner’s progressive harmonic language.

6. Riemann (1890, 38) wrote that, in these cases, one “will more or less always feel the inclination . . .
and indeed with good reason,” to hear one of the triads as a dissonance, spelled as a consonance only
for convenience. Kopp 1995, 141ff., contains a translation and exegesis. Prout (1903, 256) calls them
“false triads.” See also Louis and Thuille 1982 [1913], 409–10, and Hull 1915, 42. Louis and Thuille
note that Liszt frequently spells mediant triads as dissonances. Lendvai (1988, I: 60) observes a
specific instance in Verdi’s Otello.
7. Although the analogy may initially seem far-fetched, it taps into a significant history of theorizing
about consonance and dissonance. Heinrich Schenker considered consonances to uniquely possess
life-generative capacities, in the form of the capacity for prolongation; in one passage, he refers to the
tonic triad as the maternal womb. He also considered dissonant harmonies to be false and illusory.
Thus, to confound consonance with dissonance was literally to confound musical reality with
musical illusion. Similar passages can be found in writings of his contemporaries, Ernst Kurth and
Alfred Lorenz. See Cohn 2004. I return to this theme in chapter 7, in connection with Parsifal.
 CHAPTER 2 Hexatonic Cycles  23

Figure 2.4. The Tarnhelm progression from Wagner’s Ring.

In his initial notation of the progression, Wagner struggled with enharmonic

decisions (Darcy 1993, 168–69). He initially sketched the first chord as aᅈ minor.
He led cantus Eᅈ to Fᅈ, as the diatonic logic suggests, and then retracted it in
favor of E, as part of an e minor triad that supplies the G leading tone. The third
presentation of that chord, though, is rewritten as an fᅈ minor triad. The approach
pursued here suggests that the indeterminate enharmonics and the uncanny
and disjunctive semantics stem from the same source. The melodic Eᅈ ought to
move diatonically to Fᅈ; the bass Aᅈ ought to move diatonically to (tenor) G;
and the interval between bass and melody ought to be a diatonic and consonant
major sixth. One of these imperatives must be discarded so that the other two can
survive.
But the effect is overdetermined. Hexatonic progressions are also able to depict
the world of noumena by virtue of their tonal multistability. Each Tarnhelm triad
contains the other’s leading tone and hence signifies the other as tonic, like the
Escher hands drawing each other’s sleeves. A hexatonic cycle abjures the cadential
resources of classical tonality, such as fifth-related roots, dissonances, and diatonic
coordination. A composer can suggest one of its constituent triads as a tonic
“factitiously by virtue of its recurrence” (Taruskin 1996, I: 259) but can secure it
only by recruiting external syntactic routines. The six triads are equally likely
recipients of rhetorical or cadential benefaction; the progression itself is neutral
with respect to its potential tonics. In this way, hexatonic progressions resemble
the so-called “standard” bell pattern (variously called bembé and gagokoe) of
West Africa or the Caribbean, whose temporal “tonic” (i.e., downbeat) can occur
at several points along the cycle. The cycle itself is neutral with respect to these
bestowals and is in this sense multistable (Pressing 1983).
The analogy that I have just ventured takes us into a speculative terrain that
merits light treading, because it is counterintuitive from a historical standpoint.
Yet a small cluster of features line up suggestively in support of pursuing it at least
around one or two bends in the road, before beating a hasty retreat back to the
main path. We observed, in connection with figure 2.2, that a hexatonic cycle
embeds a hocket canon that projects a 3:2 phasing. This particular combination of
attributes is likewise identified with African musical traditions associated with
spirit possession and trance. Rouget 1985 suggests that such concerns are not
entirely remote from Europe. In the post-Enlightenment world, these concerns
bore scientific trappings. Dr. Franz Mesmer’s theories of animal magnetism and
universal fluids, as well as his therapeutic practice of hypnosis, held considerable
interest among artists and intellectuals in Biedermeier Vienna; Schubert’s sus-
tained encounter with them is documented in Feurzeig 1997. One can imagine,
then, that a composer of the early nineteenth century might be interested in
exploring ways to depict states or sensations associated with hypnosis, as an
24  Audacious Euphony

uncanny state liminally perched at the juncture between reality and illusion, or life
and death, and might find in hexatonic progressions a number of homologous
attributes.

Contrary Motion and Balance

How is it that contrary motion is a special case, when it is universally acknowl-

edged as a prescriptive norm in voice-leading theory and pedagogy? The resolu-
tion to this peculiarity points to an overlooked circumstance that turns out to be
fundamental to our understanding of pan-triadic harmony. When three-voice
triads (i.e., lacking tone doublings) are connected by idealized voice leading, con-
trary motion is indeed a special case: it arises only between triads whose roots are
related by major third and hence share membership in a hexatonic cycle. Any
given triad (e.g., C major) can be juxtaposed with eighteen other triads that lie
outside of its cycle. The progression to one of these triads (in the given case, A
minor) involves motion in a single voice, to which the categories of paired motion
do not apply. The progression to each of the remaining seventeen triads involves
similar or parallel motion.8
The bidirectional voice leading in a hexatonic cycle, whether successive, as in
the case of incremental motion through the cycle, or simultaneous, as in the case
of motion between nonadjacent triads, invites an interpretation in terms of
internal points of symmetry. In figure 2.2, each triad has a point of balance whose
location is roughly constant throughout the progression, despite local fluctuations.
This even balance results from the quasi-uniform size of the intervals of which
the triad is composed, which fall within a narrow compass of from three to five
semitones. The particular point of balance depends, of course, on the registral
ordering of the voices, which is arbitrarily selected in figure 2.2. But the general
feature of balance would remain under any other ordering of the pitches, provided
that the ambitus remained within an octave. (Readers comfortable with represen-
tations of pitch-class space will recognize that the center of symmetry is more
properly shown as a vector cutting through a cycle, but the basic point holds in
pitch space as well.) As the triads move incrementally through the hexatonic cycle,
the semitonal alternation of up and down causes the center of balance to toggle
back and forth between two pitches separated by one-third of a semitone.
Circumnavigation of a hexatonic cycle thus produces neither upward nor down-
ward motion through the pitch spectrum. Like a walker or a waterfall, the incessant

8. This observation has a significant entailment for one aspect of the history of harmonic tonality. Late-
medieval contrapuntal theory privileged small melodic intervals within voices and contrary motion
between them. Under triadic tonality, these two principles come into conflict. Any three-voice pro-
gression between diatonic triads must violate one or the other of them, if it involves two moving
voices. This conflict may have been a stimulant to collective creativity, a problem whose solution
incidentally introduced a number of varietals that ultimately enriched the compositional soil of
harmonic tonality: among them, the addition of a fourth voice that either doubled a triadic pitch
class or introduced a dissonance that resolved in contrary motion to the other voices.
 CHAPTER 2 Hexatonic Cycles  25

local fluctuations are underlain by a global stasis. This stasis, however, coalesces
around the prolongation not of a tonic, in any standard construal of the term, but
rather of a zone of voice-leading space (or voice-leading zone). What is meant by the
italicized term is taken up in chapter 5, after several more elements that support it
are put in place.

Hexatonic Progressions, Tonnetz Representations,

and Triadic Transformations

Having observed hexatonic cycles in the laboratory, we are now in a position to

study them in the field. Figures 2.5 through 2.8 present four of the earliest hexa-
tonic cycles, from four decades of Viennese music. It is characteristic of these
chronologically early examples that tonicizing dissonances buffer the stations of
the cycle, which are presented at a thinly veiled layer of the middleground. Like the
Schubert excerpt examined in chapter 1, each of these excerpts instantiates what
Ramon Satyendra (1992) calls layered tonality: classically tonal at both the most
global and the most local level but chromatic at the level of the local modulations
(also see McCreless 1996, 102). The diatonic indeterminacy resides in the succes-
sion of tonics, as symptomized by the essential enharmonic transformation that a
hexatonic cycle requires.
Figure 2.5 presents a passage from the final movement of Mozart’s Symphony
in Eᅈ major, K. 543 (1788). The excerpt initiates the developmental core of a mono-
thematic sonata-form movement. Beginning in Aᅈ major with a partial quotation
of the principal theme, the music proceeds, via sostenuto wind chords, through
a gᅊ minor triad to E major at m. 115. A series of violin/cello stretti, based on the

Figure 2.5. Mozart, Symphony in Eᅈ major, K. 543, finale, mm. 109–26.

26  Audacious Euphony

Figure 2.6. Haydn, Symphony no. 98, finale, mm. 148–98.

one-measure head, unfolds in the subsequent measures, connecting through

e minor and C major to c minor at m. 123, and then intensifies through a diatonic
sequence that ultimately prolongs c minor.
Figure 2.6 presents a passage from a Haydn symphony composed four years
later; the passage likewise constitutes the developmental core of a final movement.
The development begins with a series of four-measure thematic segments that
tonicize Aᅈ major, its submediant f minor, and its subdominant Dᅈ major, the
latter tonicization precipitating a sudden caesura. For four measures, the first
violin paws at the rising-third incipit of the principal theme, finally grabbing hold
of a Sturm und Drang arpeggiation of Cᅊ minor, the enharmonic parallel of the
immediately previous tonic. Gᅊ → A leads to a thematic statement in A major, from
which vantage point we can reinterpret f minor → Dᅈ major, initially heard in Aᅈ
major diatonic space, as the first step in a journey through a hexatonic cycle that is
completed by an a minor arpeggiation (m. 178) that ultimately leads to F major
at m. 190.
The Mozart and Haydn passages are harmonically open and texturally diverse,
as befits a developmental core. In the following two passages, similar progressions
are harmonically closed, unfold at a leisurely pace, and involve block transposi-
tions of lyrical thematic material. Figure 2.7 presents the sixteen-measure period
in the center of the ternary-form Adagio of Beethoven’s “Spring” Sonata, Op. 24
(1802).9 After a cadence in Bᅈ major, the antecedent phrase arises from a modal
mutation to bᅈ minor and leads after eight measures to a cadence in its submedi-
ant, Gᅈ major. The consequent phrase likewise begins with a modal mutation to

9. Schenker models this passage in Free Composition, figure 100.6(b), commenting only that “here we
have a descending register transfer by means of three major thirds” (1979 [1935], 82). Proctor 1978,
175–76, provides an astute commentary.
 CHAPTER 2 Hexatonic Cycles  27

Figure 2.7. Beethoven, Sonata for Violin and Piano, Op. 24, 2nd mvt., mm. 38–54.

fᅊ minor and leads to a tonicization of its submediant D major, but this tonal
motion is transacted in half the time. This acceleration leaves a balance of
four measures, which are filled by yet a third transposition of the same progres-
sion, leading from d minor to Bᅈ major for the start of the final section of the
movement.
Figure 2.8 presents the first of three phrases from the conclusion of the
immense initial movement of Schubert’s Piano Trio in Eᅈ major (1828). The
unmelodied accompaniment (or “vamp”) mutates a major chord to its parallel
minor, leading to an eight-measure period that modulates down a major third to
the latter’s submediant. The phrase given in figure 2.8 is followed twice by its
exact transposition, beginning first in B major and then in G major, the latter
modulating back to the closing Eᅈ major tonic.
Figure 2.9 models these four hexatonic passages in a graphic format that brings
out some features obscured by the cyclic graphs of figure 2.1. Points represent
individual tones, rather than the triads formed by their combination, and edges
connect tones that form consonant intervals. This graphic format is a fragment of
the Tonnetz (“tonal network” in German), a planar figure that coordinates axes
representing the consonant interval classes. In the version that will be used

Figure 2.8. Schubert, Piano Trio in Eᅈ major, Op. 100, 1st mvt., mm. 584–595.
28  Audacious Euphony

Figure 2.9. Tonnetz models of figures 2.5–2.8.

throughout this book, perfect fifths rise from left to right along the horizontal axis,
minor thirds rise from northwest to southeast, and major thirds from southwest
to northeast.10 What appears as clockwise motion in figure 2.1 is converted in
figure 2.9 to downward motion through a strip whose external boundaries form
augmented triads. Each strip’s interior is tiled into triangles, representing conso-
nant triads. Major triads extend upward, and minor triads subtend downward,
from their shared perfect-fifth edge. Internal edges represent shared dyads. The
motion of each passage through the strip is modeled by an arrow. Numbers inside
the triangles reference bar numbers of the score.
These planar graphs suffer from a Bering Strait flaw. Each tone and edge at the
top of a strip reappears at its bottom, masking identities and distorting distances.
Were these identities honored by “gluing together,” the strip would convert to a
cylinder. Our failure to so honor them is a concession to the dimensional limits of
the printed page. These limitations are worth tolerating because of the many
advantages that the triangularly tiled planar representations afford. A few of those
advantages will become apparent in this chapter; many more will accumulate as
we penetrate more deeply into the heart of the model.

10. The Tonnetz was first presented by Leonhard Euler in 1739. It was revived by German harmonic
theorists in the second half of the nineteenth century and was independently reinvented numerous
times, and for numerous reasons, by late-twentieth-century music theorists, psychologists, and dil-
ettantes interloping from other academic fields. The angled format was introduced by Ottakar
Hostinský in 1879 and adopted by Hugo Riemann in his later publications. For more on the history
of the Tonnetz, see Vogel 1993 [1975], Mooney 1996, Gollin 2006, and Cohn 2011. One innovation
adopted here, following an idea of Daniel Harrison (2002b), is the double labeling of nodes that
correspond to enharmonic exchanges in the score being modeled (e.g., the Aᅈ/Gᅊ along the left
border of the Mozart Tonnetz in figure 2.9). This makes it easier to identify triads and track their
progressions.
 CHAPTER 2 Hexatonic Cycles  29

Figure 2.10. A progression on a Tonnetz strip, as incremental moves through a

hexatonic cycle.

Figure 2.10 extracts one of these graphs for closer study. It breaks the single
arrow into a series of local ones that indicate individual pitch-class displacements.
Downward arrows, indicating chromatic-semitone descents, alternate with diago-
nal ones, indicating diatonic-semitone ascents. To distinguish these two species of
local progression, we draw on a tradition initiated by Arthur von Oettingen (1866),
who identified several exchange operations (Wechsel) that connect opposite-mode
triads. These include a leading-tone exchange (Leittonwechsel) that connects
opposite-mode triads that share a minor-third dyad, as exemplified by the diago-
nal arrows in figure 2.10. Oettingen’s exchange operations were developed by
Riemann (1880), and some of them were revived and formalized a century later by
David Lewin (1982, 1987).11 Following Brian Hyer’s 1989 adaptation from Lewin,
I will use the letter P (parallel major/minor) to indicate the motion between triads
that share two common tones and a common root, and L (Leittonwechsel) to
indicate triads that share two common tones and whose roots are a major third
apart. Both operations are involutions, which is to say that they “undo themselves”:
two consecutive applications produce an identity. Table 2.1 summarizes the
information about these two transformations.
Figure 2.11 presents Tonnetz models of four additional passages, which
traverse hexatonic cycles in a manner more characteristic of the nineteenth
century. Figure 2.11(a) models the passage presented at figure 2.12, from the first
movement of Brahms’s Concerto for Violin and Cello, Op. 102 (1887). The passage
resembles those studied above in connection with Figures 2.5 through 2.9, but the
diatonic buffers have been removed. Each station along the path is approached

11. The history of transformations, on their own and as they relate to the Tonnetz, is documented in
Klumpenhouwer 1994, Mooney 1996, Gollin 2000, Kopp 2002, and Engebretsen 2002. The exchange
operations were theorized as contextual inversions in Clough 1998.
30  Audacious Euphony

Table 2.1. Incremental hexatonic transformations

Name Symbol Root Common Semitonal Planar angle
motion dyad species on Tonnetz
Parallel P No change Perfect fifth Chromatic 0°
Leittonwechsel L Major third Minor third Diatonic 120°

directly from its predecessor, and as a result the passage sounds only the six tones
of the cycle’s associated hexatonic scale.
Figure 2.11(b) represents the second sentence of Liszt’s Consolation no. 3
(1840s), for which a score is available at Web score 8.3 . The passage presents
upward motion through the strip, equivalent to counterclockwise cyclic motion,
for the first time.12 A second new feature is the omission of one member of the
cycle, cᅊ minor, resulting in the direct juxtaposition of two triads of the same mode
(see figure 2.3(a)). Although the name for this transformation, LP, suggests a
compound transformation, with intermediate cᅊ minor deleted, I prefer to think
of the transformation (and its inverse, PL) as a unitary Gestalt whose name
happens to have two syllables.13

Figure 2.11. Four hexatonic passages from the nineteenth century.

12. For analyses, see Lewin 1967, Gollin 2000, Kopp 2002, Santa 2003, and chapter 8 of this book.
13. Kopp (2002, 159) suggests that LP is a compound operation by virtue of its compound name.
Gollin 2000 (6–12) argues that this need not be so; any compound transformation, such as “retro-
grade inversion,” can be furnished with a unary name (he suggests “George”). Although there is
heuristic value in the compound name, there is no necessary significance to it. The same is true
in natural language, where words like breakfast and handicap autonomously accrue and shed
meanings apart from their compound origins.
 CHAPTER 2 Hexatonic Cycles  31

Figure 2.12. Brahms, Concerto for Violin and Cello, Op. 102, 1st mvt., mm. 270–79.

The final two passages modeled in figure 2.11 are presented at figure 2.13: the
opening of the Sanctus movement from Schubert’s Eᅈ major Mass (1828) and a
chromatic Grail distortion from Wagner’s Parsifal (1882).14 Both passages begin
with Eᅈ major → b minor, diametrically across a hexatonic cycle. Two intermediate
triads on the hexatonic cycle are omitted, and all three voices move simultaneously
by semitone, as in figure 2.3(c). I will describe this relation as a hexatonic pole,
and the corresponding transformation with the label H (after Cook 1994). Like
all mode-switching operations, H is an involution that “undoes itself ” when
performed twice consecutively.
Both passages also continue by indirectly connecting b minor → eᅈ minor, but
they do so in different ways. Wagner interpolates G major, which shares two

Figure 2.13. Two similar hexatonic progressions.

14. For a consideration of the Schubert passage, see Salzer and Schachter 1969 (215–18). Concerning
the Grail distortion, see Lewin 1984 and 1992, Clampitt 1998, Lerdahl 2001, Cohn 2006, Lerdahl
and Krumhansl 2007, Brower 2008, and Rings 2011.
32  Audacious Euphony

common tones with its predecessor and none with its successor. As a result, the
closing G major → eᅈ minor transposes the opening H progression, Eᅈ major → b
minor. Schubert interpolates g minor, which shares one tone with both of its
neighbors, dividing the four units of voice-leading work evenly between the two
progressions. The progression into and out of g minor transposes a minor triad
downward by major third, meriting the label LP.
That label is familiar from the Liszt Consolation (figure 2.11(b)), where it
identified the transposition of a major triad upward by a major third. Here we
arrive at a circumstance that has come to be viewed by some as the Achilles heel
of triadic transformational theory. In furnishing g minor → eᅈ minor, which termi-
nates the Schubert progression, with the same label as A major → Dᅈ major, which
terminates the Liszt one, we are implicitly claiming that the two progressions
are in some sense equivalent. Three sorts of objections have been raised against
this claim. First, the claim is seen to violate the intuition that progressions, in order
to be considered equivalent, ought to move roots not only by the same magnitude
but also in the same direction (Kopp 2002; Tymoczko 2009b). Second, the claim is
seen to result solely from the way that the transformational logic plays out. The LP
operations, and their PL inverses, become the boorish in-laws that need to be
tolerated if we want to marry into the otherwise attractive and well-behaved
kinship system. Third, the claim seems to entail a commitment to the harmonic
dualism under whose banner it first arose in the writing of Oettingen and Riemann:
the notion that major and minor triads are generated by equal but opposite meta-
physical or physical forces. Such a commitment would be embarrassing on a priori
grounds, since the metaphysics is obsolete and the physics apocryphal (Harrison
1994; Rehding 2003).
All three objections are neutralized by an appeal to voice leading, a dimension
of experience that is, in principle, independent of root motion. What C major → E
major shares with c minor → aᅈ minor is the behavior of each individual voice:
the G voice moves up by semitone, the C voice moves down by semitone, and the
remaining voice holds constant. More generally, any LP operation sends its
perfect-fifth dyad to a major sixth (or perfect fourth to minor third), and any
PL operation does the reverse.
Consider figure 2.14, which juxtaposes two passages from Richard Strauss’s
“Frühling” of 1949. A score is available at Web score 4.19 . After an orchestral
alternation between c minor and aᅈ minor, the singer enters on the wave of that
same progression at m. 5. The second stanza opens on a C major triad that moves
directly to E major in 64 position. The two passages feature major-third transposi-
tions but in opposite directions. What they share is their voice leading: both pro-
gressions lead C down by semitone and G up by semitone while keeping their third
voice invariant. It is this voice-leading equivalence that the label LP captures.15
Although there is no appeal to harmonic dualism here, there is, nonetheless, a
more benign melodic dualism lurking about in the wings, whose implications are
treated below.

15. See also Lewin 1992, figure 3; for a similar case involving PL, see Cohn 1999.
 CHAPTER 2 Hexatonic Cycles  33

Figure 2.14. Two LP transformations in Strauss’s “Frühling” (Four Last Songs).

Near Evenness, Minimal Voice Leading, and the

Central Role of Augmented Triads

Our work with the Tonnetz strips (figures 2.9 and 2.11) has focused on the rela-
tions between the triangles at their interior. In a move that will have significant
repercussions for the remainder of this book, we now attend to the augmented
triads that bound the strips. The role that they play, in the hexatonic passages that
those strips represent, is not immediately apparent. In none of the passages does
an augmented triad appear as a surface harmony. One of the augmented triads, on
the left boundary of the strip, does have a certain salience in the bass register,
where its tones slowly unfold one by one. We might be inclined, by habit, to say
that this augmented triad is arpeggiated, or even prolonged. But we may not be
prepared to take on board some implications of such a claim. Are the tones of
the augmented triad fused into a corps sonore, which generates the passage by
distributing its components across time and sprouting triads from each one? The
invocation of arpeggiation and prolongation in this context has a metaphorical
component whose heuristic value has been a site of heated controversy among
Schenkerian theorists since the 1950s.16
The notion that augmented triads are “prolonged” is particularly problematic
in the current context, since it suggests that the smooth voice leading of these pas-
sages is a by-product. If we consider semitonal voice leading as primary, then such
a conception places the tail where the head should be. It is the semitonal voice
leading that stands at the core. The disjunct bass is merely running about town
making calls where its services are needed (see Schachter 1983b, 75). Moreover,
even if we are comfortable with assigning fundamental status to the augmented
triad at the left boundary of each strip, what is the role of its partner at its right
boundary?
In proposing an equal role for the two augmented triads that bound a Tonnetz
strip, we arrive at a major theoretical claim of this book: when triadic progressions
are pursuing the logic of smooth voice leading rather than that of acoustic

16. For a nuanced discussion of this issue, see Proctor 1978, 157ff.
34  Audacious Euphony

consonance, augmented triads play a central role in their syntax, even when
occluded from the music’s surface and hence not directly accessible to perception.
By virtue of their status as perfectly even trisections of the octave, augmented triads
are the invisible axes about which pan-triadic progressions spin. Consonant triads
acquire their distinctive voice-leading features in chromatic space by virtue of
their status as minimal perturbations of the perfectly even augmented triads. The
crucial property that consonant triads bear is one that Dmitri Tymoczko has
named near evenness.17 Major triads are nearly even because they can be formed
from an augmented triad by a single semitonal displacement downward; minor
triads, conversely, are formed by a semitonal displacement upward. A consonant
triad is thus like a wheel that is dented just enough to affect a wobble but not so
much that it is knocked off its rotation. It is this property that makes possible the
minimal voice leading that hexatonic progressions feature.
Why does a nearly even chord bear the capacity to connect to an equivalently
structured chord by minimal voice leading? Why does it uniquely bear that capac-
ity? More generally, what is the connection between degree of evenness (proper to
a chord’s internal structure) and voice-leading magnitude (proper to its external
relations)? We can begin to explore these questions by inspecting figure 2.15(a), a
circle that intersects the vertices of an equilateral triangle. The vertices are labeled
with the tones of an augmented triad. Each vertex is flanked by two black circles,
labeled in two ways: with the tone that semitonally displaces a member of the
augmented triad, and with the major or minor triad that results when that tone is
combined with those of the remaining fixed vertices.18
Figure 2.15(b) shows how a sample triad, C major, is generated from the aug-
mented triad by the displacement Gᅊ → G. Our interest is in exploring the behavior
of this triad as it is subjected to those shape-preserving transformations that
involve minimal change. When the triangle is transformed in such a way that two
vertices are preserved, how far is the third vertex displaced?
The shape-preserving transformations are of two types: rotation, equivalent to
pitch-class transposition; and reflection, equivalent to pitch-class inversion. As the
triangle in figure 2.15(b) is scalene, it cannot be rotated in such a way that two of
its vertices are invariant. (This corresponds to the impossibility of transposing a
consonant triad such that two tones are preserved.) Reflection, though, is another
matter: to preserve two vertices, exchange their position by reflecting them about
an axis halfway between them.
As there are three pairs of vertices, there are three axes around which such a
reflection is possible. They are positioned as broken lines in the three components
of figure 2.16. In each component, the C major triangle is presented in half-tones,
and its inversion about the axis is presented in full tones. A double-headed arrow
indicates the exchange of the two common tones. A single-headed arrow indicates

17. The connection between smooth voice leading and near evenness was initially suggested to me by
John Clough in 1993, mentioned in Cohn 1996, 39n40, and elaborated in Cohn 1997, although
using different terms. Tymoczko 2011b (14, 61, 85–93) positions the connection between near
evenness and voice leading within a very broad framework with many concrete applications.
18. Figure 2.15(a) resembles some graphs that appear in the teaching materials of jazz guitarist
Pat Martino (Capuzzo 2006). See also Siciliano 2005a.
 CHAPTER 2 Hexatonic Cycles  35

Figure 2.15. Derivation of consonant from augmented triads via single semitonal
displacement.

the path from the remaining tone of the C major triad to the tone that replaces it
as a result of inversion about that axis. That path’s magnitude is double the distance
that separates the tone from the axis. This is, of course, how inversion/reflection
works: the farther the distance of some object from the axis, the farther that object
is projected by inversion about it.

• In figure 2.16(a), the axis (represented as a broken line intersecting the

center of the circle) is positioned halfway between E and G, so that those
two tones reflect into each other (as indicated by the double-headed arrow).
Under reflection about that same axis, the remaining tone, C, is replaced
by B (as indicated by the single-headed arrow). Interpreted in pitch-class
space, this reflection models the application of the L transformation on
C major to produce e minor.

Figure 2.16. Double common tone transformation of C major, depicted as

inversion about an axis that holds two of its tones invariant.
36  Audacious Euphony

• In figure 2.16(b), the axis mediates C and G, which invert into each other.
The remaining tone, E, reflects into Eᅈ. This reflection models P as it
transforms C major to c minor.
• In figure 2.16(c), the axis mediates C and E, which invert into each other.
The remaining tone, G, reflects into A. As this transformation involves two
units of work—small, but not minimal—it has not arisen in this chapter;
we shall study it in chapter 4.

For comparison, figure 2.17 shows the behavior under reflection of two
other chord types, which are, respectively, more and less even than the C major
triad of figure 2.16. Both chords share the CE dyad with C major; what defines
their degree of evenness is the position of the third voice. In figure 2.17(a) that
third voice is subject to a “null” perturbation, remaining on Gᅊ and creating a
perfectly even augmented triad. At (b), the third voice is perturbed by three
units, moving to B and creating a relatively uneven [015]-type trichord. In the
perfectly even case, the axis that reflects C and E into each other also reflects
Gᅊ into itself. Consequently, the perfectly even augmented triad cannot be
distinguished from its reflection, and the transformation is a phantom. In the
uneven case, reflection about the same axis slings F a tritone away, to B. This exer-
cise demonstrates that, in order to create a small but recognizable displacement of
a single voice under inversion, the trichord must be as even as possible, but not
perfectly even.
What is true of nearly even trichords in a chromatic space of twelve tones is
equally true of nearly even chords of any cardinality in a chromatic space of any
size.19 I mention here briefly several other cases of near evenness that are familiar
to music theorists, or that we will have occasion to refer to in chapter 7.

Figure 2.17. Double common tone transformations of two dissonant chords.

19. The general principles are established, in different ways, in Douthett 2008 and in Tymoczko
2011b.
 CHAPTER 2 Hexatonic Cycles  37

• Dominant and half-diminished seventh chords, as minimal displacements

of perfectly symmetric fully diminished seventh chords, voice lead to each
other more smoothly than any other tetrachord type.
• Mystic chords, as minimal displacements of whole tone collections, have a
similar capacity among hexachords.
• Diatonic collections, which are maximally but not perfectly even (Clough
and Douthett 1991), communicate with each other via minimal
displacement, a circumstance that underlies the system of key signatures at
the heart of musical notation and theory of modulation. The same is true
of their pentatonic complements (Kopp 1997).
• Diatonic triads, which are maximally but not perfectly even in seven-tone
diatonic space, communicate with each other by single stepwise motion, a
circumstance that underlies the system of diatonic third-relations central
to some theories of harmonic function (e.g., Agmon 1991, 1995).

For the more circumscribed case that concerns us in this chapter, the central
point is this: Via single semitonal displacement, each major triad communicates
with two minor triads, and each minor triad with two major triads, precisely
because major and minor triads are nearly even. One can thus draw a direct
connection between near evenness and the unique ability of triads to participate
in hexatonic chains and cycles. The role of near evenness, with respect to the
participation of triads in hexatonic progressions, is thus analogous to that of
consonance, with respect to their participation in diatonic ones. This suggests
that the role of the augmented triad in the former case is analogous to that of the
harmonic series in the latter: it is the concealed noumenon that gives rise to
the revealed phenomenon. We might then find value in the claim that augmented
triads generate pan-triadic syntax by way of nearly even trichords, just as many
theorists have found value in the claim that the harmonic series generates
diatonic syntax by way of consonant triads. This conjecture is given close scrutiny
in chapter 3, which focuses more closely on the relationship between consonant
and augmented triads in the music and the music theory of the nineteenth
century.

Remarks on Dualism

The above discussion places us in a position to refine our understanding of the way
that the system of triadic transformations rest on a dualistic basis. David Kopp
(2002, 155) frames an extensive discussion of my early work on this topic under
the subtitle “A Dualist Transformation System” and suggests without comment
that the dualism is a limitation. Dmitri Tymoczko (2009a) makes a more sweeping
claim: that the system of triadic transformations is “fundamentally dualist” and
hence ripe for wholesale rejection. It is hardly necessary to explain why attribu-
tions of dualism are a priori problematic, since the term is identified with the
38  Audacious Euphony

discredited harmonic dualism of Oettingen and Riemann under which the triadic
transformations were initially conceived.
There are two points to make in response. The first is that concepts may be
detached from the framework in which they were initially conceived, in principle.
This point will be familiar to music theorists, who identify Schenkerian prolonga-
tions and progressions without fear that they will be suspected of subscribing to
the ideas of German racial superiority in support of which those ideas were evi-
dently conceived. The second point is that relations that may appear to be funda-
mentally dualist may arise as epiphenomena of other relations. This is the central
argument of Tymoczko 2011a: “Nineteenth-century composers were not explicitly
concerned with inversional relationships as such; instead, these relationships
appear as necessary by-products of a deeper and more fundamental concern with
efficient voice leading (253).” The same argument applies to twenty-first-century
theorists. Major and minor triads constitute a fundamental and coherent class of
objects not because they are related to each other by inversion about an axis. They
are related because they share the property of near evenness, and degree of even-
ness is invariant under inversion. Their inversional relation is a consequence of the
capacity of the tones of an augmented triad to be semitonally perturbed in two
directions, up and down.
What is true of the objects is true of their transformations: any hexatonic
transformation introduced in this chapter can be defined with respect to its
structure as a minimally displaced augmented triad. For example, L can be defined
as the transformation that acts upon the tone that is further from the displaced
tone, moving the former a semitone closer to the latter; and P as the transforma-
tion that acts upon the tone that is nearer to the displaced tone, moving the former
a semitone further from the latter. Similar formulations can be used to character-
ize any of the remaining transformations introduced in this chapter and, indeed,
any of the transformations and transformation classes introduced in chapters 4, 5,
and 7.20
In adopting this position, we are not washing our hands of harmonic dualism,
which Henry Klumpenhouwer characterizes as “music-theoretical work that
accept[s] the absolute structural equality of major and minor triads as objects
derived from a single, unitary process that structurally contains the potential for
twofold, or binary, articulation” (2002, 459). Instead, we are viewing harmonic
dualism as the product of a more fundamental melodic dualism, which posits that
melodic motion proceeds in two opposite directions, which we figure in our cul-
ture as “up” and “down,”21 and that there exists a sense in which it is productive
to grant equivalent status to directed motions of equivalent magnitude but not
equivalent direction. This dualism is so familiar as to be transparent; we invoke

20. To be sure, these formulations are cumbersome, so I often use inversion with reference to the
triadic transformations. But, as I stressed in note 13 above, heuristics should not be confused
with ontology. One can refer to inversional relations without believing that they are essential,
just as one can refer to a sunset while still believing that the sun’s position is fixed and the earth’s is
variable.
21. In other cultures it is figured as old/young, sharp/flat, skinny/plump, etc. See Zbikowski 2002,
67–68.
 CHAPTER 2 Hexatonic Cycles  39

it implicitly whenever we refer to arpeggiation, passing or neighboring motion,

or refer to an interval without specifying its direction. On this position, the uni-
tary process to which Klumpenhouwer refers is not a harmonic one, as it was
for nineteenth-century German theorists, but rather a melodic one: the single
semitonal displacement of the augmented triad. The binary articulation involves
neither overtones/undertones nor having/being, but rather an acknowledgment
that such displacements may proceed either up or down.

Triadic Structure Generates Pan-Triadic Syntax

Material may suggest what process it should be run through (content suggests
form), and processes may suggest what sort of material should be run through
them (form suggests content).
—Steve Reich, Writings about Music

One of the desirable qualities of a theory is the ability to demonstrate a relationship

between the internal properties of an object and its function within a system.
A successful model of triadic music ought to give a coherent account not only
of triadic behavior but also of why composers have selected triads to do the behav-
ing. Accordingly, one of the most enduring features of classical tonal theory is its
capacity to generate syntax from the phonological properties of its constituent
objects. Remarkably, the conviction that phonological consonance generates syn-
tactic proximity is held by consensus across the many denominations of classical
theory, which conceive and represent their subject in distinct and often competing
ways. Riemann’s functions, Piston’s Roman numerals, Schoenberg’s structural func-
tions, Schenker’s Ursätze, and Lerdahl’s pitch-space grids say different things about
tonal syntax, but the acoustic properties of major and minor triads are foundational
to each.
If triads are nothing but quintessentially consonant objects, why should they
be asked to generate a syntax that is not predicated on their consonance? The good
composer listens to the musical object, identifies its properties and tendencies, and
recognizes the transformations that will extract the dynamic life from the object’s
interior, just as the good sculptor recognizes and extracts the form latent in the
stone or the fallen log. It is a poor composer who runs any old object through any
old machine and calls the result “art.” A passage from Daniel Harrison’s “Three
Essays on Neo-Riemannian Theory,” written in 2001 but only recently published
(2011), gives some sense of what is at stake:

Transformational theory in general requires a separation of object and activity, of

what something is and what is done with it. . . . Objects are inert and without
tendency, and all activity and meaning are supplied by transformations applied to
them. From this far vantage point, transformational theory appears to model the
metaphor of musical motion by constructing a ventriloquist’s dummy; it only
appears to be alive, but is in fact a construction of lifeless parts that are made to
move by some external force. (552)
40  Audacious Euphony

Although transformational theory, in its broadest outlines, may suffer from

Harrison’s gruesome vision, that branch of transformational theory that takes
consonant triads for its objects, and subjects those objects to transformations
that minimize voice leading, is immune from it, by virtue of the work we have
done in this chapter. Establishing near evenness as a unique feature of consonant
triads places us in a position to see that the structure of triads, as objects, is inti-
mately related to their function, as participants in hexatonic (and, more broadly,
pan-triadic) syntax.

Triads Are Homophonous Diamorphs

Identifying the triad as an optimal voice-leading structure by virtue of its near

evenness does not detract from its status as an optimal acoustic structure by virtue
of its consonance. What it suggests is that the triad is a homophonous diamorph:
one sound, two forms.22 There are two distinct, independent reasons for selecting
major and minor triads as primary structures on which to build a musical syntax.23
Even in some alternative universe where major and minor triads were acoustically
dissonant, there would still be a musical motivation for inventing them and basing
a musical system upon their properties. Dmitri Tymoczko (2011b, 64) makes a
similar point with a Deist parable:

Suppose God asked you, at the dawn of time, to choose the chords that humanity
would use in its music. There are two different choices you might make. You might
say, “. . . I’d like some nearly even chords that allow us to combine efficient voice
leading and harmonic consistency. . . .” And God would hand you a suitcase con-
taining nearly even chords, including the perfect fifth, the major triad, and domi-
nant-seventh chord. On the other hand, you might say “. . . I’d really like to hear
chords that sound good—chords whose intrinsic consonance will put a smile on my
face.” In this case, God would hand you a suitcase containing . . . the perfect fifth,

22. The term is appropriated from linguistic theories of code switching (Muysken 2000, 123), whose
connections with music are explored in chapter 9.
23. The assertion that near evenness is independent of consonance is complicated by the strong correla-
tion between the two properties. Tymoczko points out that nearly even chords in the twelve-note
universe are among the most consonant of their cardinality (2011b, 14), and that “for small chords,
maximal consonance implies near evenness” (61). The implication goes only one way—it is not the
case that near evenness implies maximal consonance. This is true whether one assesses consonance
by an overtone method or by an interval vector method. With an overtone method, the maximally
consonant chord of cardinality n is the one that most closely approximates the first n odd partials
of some generating fundamental. On this standard, the [02469] major ninth chord is the maximally
consonant pentachord; but it is less even than the [02479] “usual” pentatonic. With an interval
vector method, the maximally consonant chord of a given cardinality is the one whose interval
vector entries for classes 3, 4, and 5 sum to the highest value. On this standard, the [0347] split third
is the maximally consonant tetrachord, but it is less even than the nearly even [0258] dominant/
half-diminished chord. The divergence between the two properties becomes more acute in larger
“microtonal” universes, where the nearly even dyad moves away from the perfect fourth and closer
to the tritone.
 CHAPTER 2 Hexatonic Cycles  41

the major triad, the dominant-seventh chord. In other words, he would hand you
the very same chords, no matter which choice you made.

The general phenomenon described here turns up frequently in the natural

and social worlds, where it is referred to as overdetermination, robustness,
or Babylonianism. These terms characterize “the use of multiple means of
determination to ‘triangulate’ on the existence and character of a common phe-
nomenon, object, or result” (Wimsatt 1981, 125).24 These might include “using
different assumptions, models, or axiomatizations to derive the same result or
theorem” (127).
Human physiology presents many easily accessible examples of overdetermi-
nation. Mouths are for eating, talking, and breathing; ears serve auditory and
vestibular functions; male urethrae channel both excretory and reproductive
fluids. Organs transform seamlessly between the different functions that they ful-
fill. In humans, these organs are housed within a body that also includes an organ
responsible for achieving and articulating awareness of the world within, without,
and beyond. And yet these transformations evade recognition by that organ, under
ordinary circumstances. Organs fulfill their overdetermined potentials well
beneath the threshold of awareness.
But music is different, as its active production and passive experience neces-
sarily involve, in some measure, the participation of a conscious, aware mental
faculty. How, then, does triadic music execute the transformation between the
multiple potentials of its constituent objects? This question will emerge as explic-
itly central in the closing chapters of this book. In the meantime, our concern will
be directed toward refining the model of triadic syntax introduced in this chapter,
which is predicated exclusively on fulfilling the triad’s syntactic potential as a
nearly even object.
We will nonetheless be unable to forget that the triad is also something else
that we have long known it to be. We will preserve that memory in our terms of
reference: I will continue to refer to them collectively as consonant triads, even
though we are more interested in their extensionally identical but intensionally
distinct status as nearly even trichords. I will continue to name them individually
by their roots, even though roots have no theoretical status in the theory of pan-
triadic syntax. (Heuristics ≠ ontology; see notes 13 and 20.) And, in discussions of
particular passages, I shall continue to casually invoke all manner of overlearned
ascriptions and categories, in ways that contain unspoken implications about how
composers move between or overlay diatonic and pan-triadic syntax, the nearly
even trichord with the consonant triad. What I defer, until chapters 8 and 9, is the
explicit theorization of this process. Our docket is otherwise full, as we seek to
refine the notion of how the nearly even trichord generates pan-triadic syntax on
its own terms, independently of diatonic tonality.

24. Wimsatt 1981 traces the origin of the tradition to Aristotle, “who valued having multiple explana-
tions of a phenomenon” (125). Freud’s Interpretation of Dreams (1900) is the moment when over-
determination emerges to prominence in modern thought. Feynman 1965, 46, referred to this
approach to knowledge as “Babylonian.”
This page intentionally left blank
C HA P T E R
Three
Reciprocity

The Historical Emergence of Augmented Triads

Chapter 2 proposed that pan-triadic progressions, exemplified by hexatonic cycles,

arise from the status of consonant triads as minimal perturbations of the perfectly
even augmented triad. Some readers might worry that too much weight is being
placed on a relatively slender shoot. When an augmented triad appears in music
before 1830, its behavior is normally well regulated and unobtrusive, tucked into the
middle of a phrase rather than exposed at its boundaries, passed through quickly
and lacking metric accent. In an 1853 monograph titled The Augmented Triad,
Carl Friedrich Weitzmann portrayed his protagonist as a serf, scurrying in and out
the rear entrance, occasionally showing his face but never intruding on the con-
versation in the salon. After agitating on behalf of “granting [the augmented triad]
an abiding place in the kingdom of tones,” Weitzmann “gave its further fate over to
our enlightened composers” (Weitzmann 2004 [1853], 144, 224; my translations).
As if in response, some late-nineteenth-century composers featured augmented
triads as motivic emblems in individual compositions: Brahms in his Alto
Rhapsody (Forte 1983) and Wagner in Siegfried Idyll (Anson-Cartwright 1996)
and his musical portraits of the Valkyries and of Amfortas. Liszt and Wolf acted
more boldly, incorporating augmented triads into their normative sonic core.1
It is nonetheless difficult to make a case that augmented triads ever achieved a
normative, unmarked status.
Charles Moomaw’s 1985 dissertation is the most comprehensive English-
language source concerning the augmented triad’s origins and early history.
Moomaw locates the chord in France as early as 1636, typically when the fifth of a
dominant triad is displaced up a diatonic semitone (Moomaw 1985, 251). He also
reports that figured bass treatises consistently instruct that the +5 figure be

1. On augmented triads in Liszt, see Forte 1987, Todd 1988, and Satyendra 1992. Hantz 1982 analyzes
the augmented triads in Liszt’s “Blume und Duft” in a way that particularly relates to the approach
developed here. On augmented triads in Wolf, see McKinney 1993.

43
44  Audacious Euphony

rendered with a seventh or ninth above the bass even when the latter is not
explicitly ciphered (128).
Georg Andreas Sorge was evidently the first to recognize the augmented triad
as a primary harmony, although initially in 1745 he did so with great reluctance:
“The best thing about this harsh harmony, if one may speak of it as one, is that it
seldom appears” (Sorge 1980 [1745], 440). In 1760, Sorge upgraded its status
incrementally, observing that the augmented triad is tolerable when it results
from a chromatic passing tone that connects fifth-related major triads (Moomaw
1985, 323). It is in such passing contexts, bisecting a whole step, that the aug-
mented triad most characteristically and frequently occurs in music of the late
eighteenth and early nineteenth centuries.
During the 1770s, French theorists began to accept the augmented triad as a
fundamental sonority, bearing a distinctive character, and even a capacity to sup-
port accretions (Gessele 1994, 84–86). This acceptance becomes evident in a
remarkable D major Minuet that has been attributed to Mozart.2 The short compo-
sition contains seven augmented triads, of which only the first behaves in the
manner sanctioned by contemporaneous treatises. The remaining six dissonances
are anomalously accented in three independent ways: each initiates a phrase,
occurs on a metric downbeat, and is marked sforzando. Howard Boatwright
astutely observed that “each augmented chord has a different melodic origin and a
different harmonic function” (1966, 30) and concluded that the sonority has
motivic value, in and of itself, rather than as a diminutional accretion to some
other formation (also see Sobaskie 1987).
Abbé Georg Vogler’s 1802 Handbuch zur Harmonielehre was the first treatise
to explore the augmented triad’s potential for enharmonic reinterpretation.
Writing that the augmented triad “appears to consist of three similar major
thirds,” Vogler claimed that its proper roost was the third scale degree of harmonic
minor and that “each III chord in minor . . . can be multiply interpreted as a
III chord in three different keys” (1802, 103, 109; my translations). Vogler illus-
trated this potential for Mehrdeutigkeit (multiple meaning) through the progres-
sion given here as figure 3.1, whose anacrusis/downbeat combinations form a
hexatonic cycle. The third beat of each measure hosts some spelling of the
CEGᅊ augmented triad, acting successively as dominant of each triad on the
following beat.

Figure 3.1. From Georg Vogler’s Handbuch zur Harmonie (1802).

2. The Köchel number is K. 355/576b. The attribution, from an 1801 publication, is suspicious on inter-
nal grounds (Oster 1966) and has never been corroborated. In any case, no evidence exists as to date
of composition (Cliff Eisen, e-mail correspondence with the author, 2007).
 CHAPTER 3 Reciprocity  45

Figure 3.2. Schubert, Symphony no. 2, 4th mvt., mm. 300–12.

Vogler’s progression serves as a template for the passage presented in linear

reduction at figure 3.2, from the finale of Schubert’s Second Symphony in Bᅈ major
(1812).3 The core of the development consists of three transpositionally identical
phrases that divide the octave by major thirds, beginning and ending with the
F major that terminates the exposition. Immediately preceding each tonic is a
locally appropriate spelling of the CEGᅊ augmented triad. Prolongation does not
capture the relation of these chords to one another, because the chord functions as
a local dominant to a series of well-articulated tonics. Nor is motivic association
quite adequate; it is too static to capture the phenomenology of the passage. More
than waving their hands and crying, “Remember me, here I am again,” these
augmented triads are also saying, “You thought I was this; well think again, ’cause
I can be that too.”
If prolongation is at work in this passage in any form, then its more plausible
object is the FACᅊ triad that unites the three tonics. In graphs of similar passages
by Beethoven and Wolf, Heinrich Schenker implied that he understood the arpeg-
giated augmented triad as the prolonged displacement, by chromatic neighbor, of
a major triad.4 In figure 3.2, that triad would be F major, which appears as a tonic
at the end of the exposition, returns at m. 352, and ultimately acquires a retransi-
tional seventh (m. 392). When Dᅈ major is tonicized at m. 312, C is displaced by a
Dᅈ neighbor, which continues to be locally supported (qua Cᅊ) when A major is
tonicized at m. 332, and only returns to C at m. 352, when F major is retonicized.
If, as Schenker implies in analogous passages, FADᅈ is the prolonged harmony
from m. 312 to m. 347, then it follows that both Dᅈ major and A major are subor-
dinated to a controlling dissonance. Yet the score contains no vertical slice or con-
tiguous patch, even an egregiously gerrymandered one, to be circled and labeled as
a “controlling harmony.” In passages such as these, then, the augmented triad is

3. Seidel (1963) draws attention to this passage. Wason (1985, 19) speculates on a possible lineage from
Vogler to Schubert. Vogler was a peripatetic, ambitious, and charismatic personality who lived in
Vienna from 1802 to 1805 and later taught composition to such prominent figures as Carl Maria von
Weber, Gottfried Weber, and Meyerbeer (Grave and Grave 1987).
4. See Schenker’s analyses of passages from Beethoven’s “Appassionata” and “Spring” Sonatas and Wolf ’s
“Ständchen” (2005 [1924], 41–64; 1979 [1935]: fig. 100.6). Many scholars (e.g., Slatin 1967; Morgan
1976; Proctor 1978; Stein 1985) have observed that his treatment of middleground equal divisions
cannot be reconciled to his pronouncements elsewhere that only consonant harmonies are suscep-
tible to composing out. What is of primary interest here is that Schenker found dissonant prolonga-
tions aurally and conceptually plausible, even if they “prolong” an idea that dissonates with the
fundaments of tonality.
46  Audacious Euphony

not directly available to perception. Its status, as a collocation of bass pitches or

triadic roots or local tonics, is virtual and liminal.
This analysis suggests that the relationship between consonant and dissonant
harmonies is not diodic. Consonant harmonies provide the context in which dis-
sonant harmonies can operate, as a rule. But, as Robert P. Morgan showed in
“Dissonant Prolongation” (1976), there are situations where these priorities are
reversed, and consonant triads subordinate to dissonant ones, not only locally but
across spans of significant duration. The relation between consonance and disso-
nance, then, is fluid in principle. The potential for this fluidity opens up a compo-
sitional dynamic, where a terrain of fixed relations is transformed into a site for
negotiation. Consonant and augmented triads gain the potential, in principle, for
reciprocity.

Consonance/Dissonance Reciprocity

The nineteenth century was familiar with reciprocity as a general cultural condi-
tion. Kant developed it in his influential Critique of Pure Reason (1982 [1787])
as his third analogy of experience. The term was imported into music theory by
Simon Sechter (1853–54), who noted (following Kirnberger 75 years earlier) that,
lacking further context, two fifth-related triads are tonally indeterminate.5 C serves
as dominant to F, which serves as subdominant to C, triggering a recursive circle
whose resolution requires external intervention (see Lewin 2006, 64). A similar
situation arises in the case of diatonic third relations, whether relative major/
minor (C major/a minor) or Leittonwechsel (C major/e minor). Both of these spe-
cies focus their tonal indeterminacy at a single melodic fulcrum, a whole step in
the first case and a semitone in the second. The potential indeterminacy of the
former case is well documented, particularly with respect to Schumann (e.g.,
Rosen 1995, 674). That of the latter case is encoded into its German name. The
modern conception of leading tone is restricted to the relationship between a
tonic and its semitonal lower neighbor. For German theorists of the middle of the
nineteenth century, this relationship captured only one half of a duality: Leitton
applies equally to the relationship of dominant and its semitonal upper neighbor. 6
Accordingly, when C major and e minor are juxtaposed, the attraction of C to B
^ ^ ^ ^
(as 6–5 in e minor) is as strong as that of B to C (as 7–8 in C major). The semi-
tonal relation thus projects an unstable force field that pulls simultaneously in
both directions.
Carl Friedrich Weitzmann identified a third type of reciprocity that shared
aspects of those identified above: the relationship between a minor triad and its
major dominant, which he regarded as equivalent to that of a major triad and

5. Kirnberger (1982 [1771–76], 44–45). See also Hauptmann 1888 [1853]. Such bilateralism is also
characteristic of the sixteenth-century view (Dahlhaus 1990 [1967], 241).
6. The conception originates in dualist thinking but was sensible enough that it was taken up by theo-
rists with no commitment to dualism, such as Louis and Thuille 1982 [1913], Kurth 1923, and Lorenz
1933. Harrison 1994 provides an excellent elaboration on these matters.
 CHAPTER 3 Reciprocity  47

its minor subdominant. The reciprocal leading tone energies are divided between
the thirds, which Daniel Harrison (1994) calls the agents: the upward-pressing E,
borrowed from f minor’s parallel major, and the downward-pressing Aᅈ, borrowed
from C major’s parallel minor.7 Although this triadic relation plays a central role
in the writings of many theorists after 1850, it never achieved a stable name. I shall
refer it to using Weitzmann’s term, nebenverwandt, which Janna Saslaw translates
as “adjacency relation” (Weitzmann 2004 [1853]). Of particular relevance for
present purposes is Arthur von Oettingen’s name for the same triadic pairing:
reciprocal, a German/English cognate (Mooney 1996, 56).
The situations examined so far involve a relation between two consonant
triads. Such relations are bilateral in principle, since no consonant triad is more
stable than any other absent a particular context. The reciprocity that we identified
with regard to figure 3.2, however, is of a different kind, as it involves the relation-
ship between a consonant and dissonant triad. The scale is inherently out of
balance and can only be leveled through the application of external forces. In the
crudest cases, such as Schubert’s “Die Stadt”, a dissonant harmony achieves a quasi
stability by squatting like a brute and appropriating the rhetorical garments
normally reserved for consonances (first, last, loudest, longest; see Harrison
1994, 75ff.).
In contrast to such ad hoc solutions, François-Joseph Fétis recognized a way to
override the forces of tonality by cultivating more systematic resources, which he
referred to under the terms uniformity and symmetry. In a passage quoted in
chapter 1, Fétis described the experience of a diatonic sequence in phenomeno-
logical terms: “the succession and . . . movement fix the attention of the mind,
which holds on to the form so strongly that any irregularity of tonality is not
noticed. . . . The mind, absorbed in the contemplation of the progressive series,
momentarily loses the feeling of tonality. . . . The attention of the musical sense is
diverted from the feeling of tonality by symmetry of movement and succession”
(2008 [1844], 27, 30). Fétis writes that a sequence levels the distinction between
consonance and dissonance. A diminished fifth no longer requires resolution; in
this context, its behavior is indistinguishable from that of the perfect fifth.
In a diatonic sequence, the law of uniformity is kept in check by the prior com-
mitment to the diatonic scale. Although each pattern iteration replicates the
generic intervals of its predecessor, its specific intervals are channeled within the
banks of the diatonic scale. The forces that Fétis identifies become more fully
unleashed in chromatic sequences, such as the hexatonic cycles explored in chap-
ter 2 or the Schubertian third-divisions represented by figure 3.2. In these cases,
the law of uniformity has a monolithic force, and the rapid turnover of chromatic
pitch classes ruptures any ability of the diatonic collection to hold a focus on a
particular global tonic.
The binary distinction between diatonic and chromatic sequences is a particu-
lar manifestation of a more general dynamic that arises in many passages that
we would not consider to be sequential per se. Whenever a motivic fragment
migrates across a series of transpositional levels, or a fugal point of imitation is

7. Smith 2006 identifies several Brahms compositions that thematize this reciprocity as an ambiguity.
48  Audacious Euphony

replicated on a different degree of the scale, the absolute sizes of the intervals may
conform to the locally governing scale, or they may be preserved at the expense of
subverting or even rupturing that government. The pressures toward uniformity
may be confined within the diatonic channel or may jump those banks and lay
down their own channels. Later theorists formulate this same duality in terms of
diatony versus repetition (Schenker 1954 [1906]) and magnetism versus inertia
(Larson 1994).
To see how this duality manifests in the relation between consonant and aug-
mented triads, consider the following classroom situation. Two students are pre-
sented with a melodic gesture from C up to E and asked to replicate that gesture
beginning on E. One responds with E up to Gᅊ, projecting an augmented triad; the
second with E up to Gᅉ, projecting a consonant one. Both responses are correct,
but one interprets replication as raw uniformity; the other, as tempered to the
diatonic collection. Gᅉ and Gᅊ displace each other across the melodic fulcrum
upon which the diatonality/uniformity tension is balanced, in the same way that
the same two tones constitute the modal fulcrum in the case of an E tonic, or the
melodic fulcrum in a Leittonwechsel relation between c minor and Aᅈ major, or
one of two such fulcrums in the nebenverwandt relation between C major and
f minor.

Two Early-Century Examples: Beethoven and Schubert

Composers of the early nineteenth century sometimes treated this melodic ful-
crum as a site for motivic play. Consider the initial movement of Beethoven’s
f minor Piano Sonata (Op. 57, “Appassionata”).8 A secondary theme in Aᅈ major
(m. 35) has a consequent phrase that mutates to aᅈ minor (m. 42) and remains in
that key until the end of the exposition at m. 65, featuring Eᅈ/Fᅈ motivic play
throughout those measures.9 The motive is raised to a higher power in the devel-
opment, which begins in aᅈ minor, renotated as gᅊ minor, and progresses to
E major at m. 67, saliently featuring the motion from Dᅊ to E on successive down-
beats. Motion continues around the hexatonic cycle, to e minor (m. 79); skipping
over C major, whose status as global dominant requires it be reserved for a later
moment; and proceeding directly to c minor (m. 83) and Aᅈ major (m. 87). The
entire passage prolongs Aᅈ major by displacing its fifth Eᅈ to its augmented fifth
E and then restoring it.
Similar motivic play of the dominant and its upper neighbor is evident in the
first movement of Schubert’s A major Piano Sonata, D. 959. A score of the exposi-
tion and development is available at Web score 3.3 . The dominant reached at
m. 28 of the exposition is prolonged for more than one hundred measures through
two extended expansions, each initiated by a chromatic sequence that arpeggiates

8. The analysis offered here is based on Proctor 1978, 173–74. See also Bribitzer-Stull 2006, 179–80.
9. These echo the Fᅈ/Eᅈ play at m. 23 (bass) and mm. 26 and 29 (treble), which are in turn echoes of
the Dᅈ/C emanations that conclude the initial f minor theme at mm. 10–15.
 CHAPTER 3 Reciprocity  49

downward through the stations of an E augmented triad. Both arpeggiations

involve B → C displacements that, upon resolution, trigger significant motivic
reverberations. The initial major-third cycle, sketched at figure 3.3(a), culminates
at m. 39 when C is displaced back to B in the bass. The bass then isolates and works
over the B → C displacement throughout the subsequent extension of the B major
local dominant (figure 3.3(b)). Similar bass motivic play occurs locally at the
G major fantasy drift (mm. 65–68, figure 3.3(c)), more structurally at the reanima-
tion of the major-third division (mm. 82–91, figure 3.3(d)), and prior to the final
stabilization of E major, where a C major sforzando (m. 103) is not recuperated
until a medial caesura eight measures later.
The apotheosis of this motive occurs in the development section, whose open-
ing measures (figure 3.3(e)) have been the subject of much marvel by performer/
critics. Charles Fisk describes it in the following evocative terms:

The new theme articulates itself as a fantastical ten-measure period: its first phrase
[mm. 131–35] slips away from C major into B major, while its second [mm. 136–40]
slips just as magically back up to C. An even more ethereal variant of the same
phrase pair immediately follows [mm. 141–50], its sixteenths now spun out into
gossamer webs. For these two periods, the music simply oscillates between C and
B, achieving what [Charles] Rosen characterizes as a stasis with a “physical
effect . . . like nothing in music before.” (2001, 216, quoting Rosen 1980, 287)

The oscillation identified by Fisk persists, indeed, through the remainder

of the developmental core, even after escaping the “poised, transfixed stasis”
(Brendel 1991, 126) of its opening musette. The subsequent ten-measure period
(mm. 151–60) modulates from C major to b minor and back. The C → B melodic
arc is then carried by the phrases of the final extended period (mm. 161–80),
which approach a retransitional E major first from c minor, its hexatonic pole, and
then from a minor, its minor subdominant.
The phrase pairings throughout the developmental core suggest that B acts
as lower neighbor to structural C (Jonas 1982 [1934], 92). The key that jointly
provides a context for both harmonies is e minor (see Schenker 1954 [1906],
p. 226; Hauptmann 1888 [1853], 159–60), whose shadow control is indicated by
the phrygian approach to its dominant (mm. 134, 144) and the deceptive return to
its submediant (mm. 139, 149). E minor in turn substitutes for the E major that
frames the development, which opens with a melodic motion from B4 to C5
(m. 129 bis) and concludes with its reversal (m. 179). B thus performs the role
of lower neighbor to its own upper neighbor, magnifying the melodic fulcrum
introduced with the equal divisions of the exposition.

Three Late-Century Examples: Liszt, Rimsky-Korsakov, Fauré

The give and take between consonant and augmented triads becomes foregrounded
in a number of compositions from the second half of the nineteenth century.
50  Audacious Euphony

Figure 3.3. Excerpts from Schubert, Sonata in A major, D. 959, 1st mvt.
 CHAPTER 3 Reciprocity  51

Figure 3.4. Liszt, A Faust Symphony, 1st mvt., mm. 1–22.

The opening of Liszt’s Faust Symphony, completed in 1854, famously and explicitly
inverts the values classically accorded these sonorities (figure 3.4). The passage
consists of two slow rotations through three segments of material (marked A, B,
and C in the example), each of which extends approximately four measures.
Augmented triads dominate the surface. Moreover, with the exception of mm. 1
and 13, the pitch-class pool for the entire passage draws exclusively on the CEGᅊ
and FACᅊ augmented triads, which combine to form a hexatonic collection. Of
particular interest are the four boxed figures, whose staggered downward motion
tropes a suspension figure, but with consonance and dissonance inverted with
respect to formal function (but not metric location): the position of preparation
and resolution is occupied by dissonant augmented triads; that of the suspension,
by consonant minor triads (Morgan 1976, 60).
It seems likely that Rimsky-Korsakov had the opening of Faust in his ear when
he wrote the opening measures of his Symphony no. 2 (1868), subtitled Antar
(figure 3.5). Antar, like Faust, begins with two slow rotations through a series of
three texturally differentiated segments, each approximately four measures long.
The second rotation transposes the first segment by a minor sixth (down in Antar,
up in Faust); the final two segments are then transposed upward by major third. As
both compositions combine two augmented triads into a hexatonic collection,
their second rotations recirculate the same tones as their respective antecedents.
52  Audacious Euphony

Figure 3.5. Rimsky-Korsakov, Symphony no. 2 (Antar), 1st mvt., mm. 1–24.

Looking back two decades later at the kuchist movement of which he had been
principal in the 1860s, Rimsky wrote that “Liszt was extreme, so was Berlioz,
so was Wagner. And so were we” (Taruskin 1996, I: 70). The opening of Antar
suggests that Rimsky was understating his capacity to nuance that extremism to
artistic ends. Whereas Faust’s opening overturns the asymmetric consonance/
dissonance binary with one swipe of the hand, Antar’s balances an exquisitely
fine point between its terms. Although Antar is stricter than Faust in its
hexatonicism—the passage contains not a single pitch foreign to the collection—
its augmented triads are less apparent. The segment labeled A presents three
minor triads, and the segment labeled C selects one of them for prolongation.
The segment labeled B, by contrast, prolongs the FACᅊ augmented triad, embel-
lishing two of its components with an escape tone that very tentatively suggests a
reconstitution of one of the minor triads.
Whereas figure 3.5 symmetrically segments the opening of Antar into six units,
based on thematic and textural rotation, figure 3.6 asymmetrically partitions the
same music on the basis of harmonic content, splitting the A material into two
segments and fusing the B and C material into a single one. In the first rotation,
 CHAPTER 3 Reciprocity  53

Figure 3.6. Analysis of the opening of Antar.

the A material consists of a box and an oval that respectively enclose fᅊ minor →
d minor and bᅈ minor → fᅊ minor. The remaining material echoes and expands the
initial fᅊ minor → d minor progression, interpolating an augmented triad between
them. The role of the augmented triad, on this interpretation, is to connect the two
more stable consonant triads that flank it, grossly distending a progression that
would have otherwise been at home in the eighteenth century. In effect, the aug-
mented triad staggers the simultaneous semitonal motions of the opening pro-
gression: first Fᅊ → F in the bassus and then, four measures later, Cᅊ → D in the
cantus.10 In transposing its predecessor downward by minor sixth, the second
rotation inverts the function of the two triadic pairings. The bᅈ minor → fᅊ minor
at the interior of the previous rotation is now positioned at the head, and it is this
unit that is subsequently expanded through the same passing augmented triad.
Conversely, the fᅊ minor → d minor that dominates the first rotation is tucked into
the interior of the second one. As a result, the series of four minor triads that opens
the composition, fᅊ minor → d minor → bᅈ minor → fᅊ minor, is expanded in
the progression from one rotation head to the next (fᅊ minor → d minor, m. 1; bᅈ
minor → fᅊ minor, m. 13) and also in the progression from one expansion
(fᅊ minor → Faug → d minor, mm. 4–12) to the next (bᅈ minor → Aaug → fᅊ minor,
mm. 16–24).
Where figure 3.6 presents the augmented triad as prolonging a motion between
its flanking consonant triads, figure 3.7 inverts those roles. The opening gesture in
the first bassoon (= cantus) is a hexatonic spiral that, on the basis of parallelism,
suggests three semitonal pairs: Cᅊ → D, A → Bᅈ, F → Fᅊ. Assuming that we are
inclined to hear parallel passages in parallel ways, we are encouraged to hear the
melodic gesture as unfolding an augmented triad.11 But which one? As the metric
grid does not lock in until m. 4, it is unclear whether the first or second component
of each pair is the accented one, and hence whether it is the FACᅊ or BᅈDFᅊ

10. On staggered semitones in Liszt, see Satyendra 1992, 102–3. A more complete interpretation
would acknowledge the tentativeness of d minor at m. 8. D falls back to Cᅊ throughout the segment
labeled C, at the same time as A escapes to Bᅈ, suggesting a bᅈ minor triad and delaying the ultimate
consolidation of d minor until m. 11, when the sustained Bᅈ finally resolves to A.
11. The “parallel passages in parallel ways” dictum was stated by Gottfried Weber (1846 [1817–21],
365) as “What the ear has once heard in a certain passage, it will not only expect again, on the
recurrence of the same passage, but will sometimes even perceive beforehand,” and reappears
prominently in Lerdahl and Jackendoff 1983.
54  Audacious Euphony

Figure 3.7. Alternative analysis of the opening of Antar.

augmented triad that is unfolded. The second bassoon (= bass) presents a

similar but complementary problem. It also unfolds the hexatonic collection in
semitonal pairs: Fᅊ → F, D → Cᅊ, Bᅈ → A. Here, too, the floating metrics defeat any
assignment of priority to a component of each pair, and hence to one of the two
augmented triads. Moreover, simultaneous tones in the cantus and bassus belong
to different augmented triads. Even if our ears locked into a particular metric ori-
entation, they would be receiving conflicting information from the outer voices.
Only the inner voice of the opening segment, sounded by the timbrally distinct
horn, has a clear commitment to one of the augmented triads: it sounds A → F →
Dᅈ → A. This ever-so-slight tipping of the balance, in an otherwise austere equilib-
rium, is subtly confirmed by the pitches that are held invariant in the first three
measures as their respective triads are registrally redistributed.
The segment labeled B in figure 3.5 provides clarity, first to the bassus and then
the cantus. In the bass, the arrival of pedal F2 at m. 4 stakes down the metric grid
clearly for the first time, conferring the accent of the Fᅊ → F onto its second term.
Applying this information in retrospect to the opening gesture causes us to hear
the second bassoon line in terms of Fᅊ → F, D → Cᅊ, Bᅈ → A, emphasizing the same
augmented triad sounded in the horn. The subsequent escape-tone figures in the
cantus similarly disambiguate the hexatonic spiral of the previous measures, by
tracing the same melodic course an octave lower. What was metrically flat and
amorphous in the first segment becomes shaped in the second gesture, clearly
thrusting the accentual weight onto the first term of each pair: Cᅊ → D, A → Bᅈ. The
timbral continuity of the bassoon helps to forge this connection and to project this
weighting retrospectively onto the cantus of mm. 1–3, which now is interpreted in
terms of Cᅊ → D, A → Bᅈ, F → Fᅊ. This analysis of the second segment leads us to
hear the opening segment, in each of its three melodic parts, as projecting FACᅊ,
even though its constituent tones are not sounded simultaneously before their
prolongation at mm. 4–8. Through this lens, each of the minor triads sounded in
the opening segment results from displacement of a component of the augmented
triad. This same hearing then extends naturally to the third segment of each rota-
tion, which alternates between two minor triads, in 63 and 64 inversion, respectively,
neither of which projects convincingly as an object of prolongation.
The passage excerpted as figure 3.8, from Gabriel Fauré’s Requiem Mass of
1877, provides an instructive comparison. Like much of the d minor Introit/Kyrie
 CHAPTER 3 Reciprocity  55

Figure 3.8. Fauré, Requiem Mass, Introit, mm. 50–61.

movement from which it is drawn, this passage is animated by the tonic’s relation-
ship to an F major triad that consistently functions as dominant. The melodic
fulcrum of this relationship is the Cᅊ that mediates between C and D, and hence
the harmonic fulcrum is the FACᅊ augmented triad that appears at m. 60, just
prior to the cadence. This same augmented triad, spelled variously, also appears
three times in the interior of the phrase, each time as the second component of a
two-chord unit whose first component is a consonant triad: F major at m. 52,
fᅊ minor at m. 54, and bᅈ minor at m. 56. Other features confirm the two-bar
groupings throughout the passage: with the exception of the final cadential pair-
ing, the first measure of each pair descends a melodic fourth on its final beat, and
the second measure of each pair supports an A4 reciting tone.
Our preference for analyzing parallel passages in parallel ways presents us with
a choice similar to the one that we faced in our analysis of Antar: the Exaudi’s
harmony is structured either by the connections between the initial, consonant
measure of each pair or by those of its terminal, dissonant ones. The first of these
options does not present a very coherent species of diatonic tonality: F major is
embellished by fᅊ minor and bᅈ minor before resolving as dominant of d minor. bᅈ
minor is easily reconciled as the minor subdominant of F major. What remains
intractable is the fᅊ minor triad. Perhaps it functions as iii of a D major that other-
wise has no presence in the passage (or elsewhere in the movement)? This feels a
little desperate and, moreover, does not address the enharmonic metamorphosis
of Cᅊ into Dᅈ as fᅊ minor is displaced by bᅈ minor at m. 56. The second option
understands this bouquet of harmonies in terms of the augmented triad to which
each one leads. This alternative places FACᅊ at the conceptual center of the pas-
sage, assigning it the role of a switching station through which the various conso-
nant triads are threaded. We are aware that the augmented triad plays this role
because Fauré shows us, by leading each chord in and out of the switching station,
thereby isolating each semitonal displacement. There is no consistent diatonic
explanation that accounts for the simultaneous presence of this particular group of
triads in a single phrase. What draws them together is their shared status as single
semitonal displacements of FACᅊ.
56  Audacious Euphony

The same can be said of the opening measures of Antar, where the same collec-
tion of minor triads is no more tonicizing than in Fauré (van den Toorn 1995,
127–28). This is so even though their mutual relationship to the FACᅊ augmented
triad only unfolds slowly, across the entire introduction. The augmented triad can
function as a switching station whether it has the presence of chronological medi-
ator, as in the Exaudi, or chronological consequent, as in Antar, or no role at all,
as in many of the pieces examined in chapter 4. The center of a circle is equally
orienting to a set of dancers, whether marked by a pole, a hole, or the imagination
of the dancers.12

Reciprocity in Weitzmann’s Der übermässige Dreiklang

Carl Friedrich Weitzmann was the first theorist to recognize the compositional
dynamic documented here. Weitzmann’s 1853 monograph on The Augmented
Triad tells three genesis stories about its protagonist, each of which involve the
verb entstehen or its nominal equivalent, Entstehung. Saslaw translates the verb as
“arise,” and the noun as “origin.” An alternative translation, generate/generation,
emphasizes organicist implications that may or may not be nested within
Weitzmann’s conception. Because such an implication is not guaranteed, Saslaw
has done well to avoid them; but nor is it precluded, and it will serve my interest to
pursue it.
The first story occurs in his chapter 2, “Preparation, Origin [Entstehung], and
Introduction of the Augmented Triad,” which offers “a primer as to how this
strange chord could come to life [in Leben treten könne], prepared through major
and minor triads and their inversions” (Weitzmann 2004 [1853], 166; my transla-
tion). The primer presents sixteen ways to connect a consonant triad to an aug-
mented one via semitonal voice leading. The second story occurs in chapter 6,
“Natural Origin [Entstehung] of the Augmented Triad Most Important to Each
Key.” Weitzmann combines f minor and C major triads into a pentachord, FAᅈCEG,
from whose interior he extracts the augmented triad: F[AᅈCE]G. “From the con-
nection of these two nebenverwandt chords arises [entsteht] the augmented triad
most important to the two keys represented by them” (184–85). Weitzmann
explains that even though E is foreign to the key of f minor, and Aᅈ to the key of
C major, each arises as that key’s most important neighbor tone.13 These two tales
relate to each other as specific to general. The first account concerns how an
augmented triad comes into being at a particular moment in a particular composi-
tion. The second deals with the augmented triad’s position in a musical system,

12. The absence of the perfectly even chord about which the nearly even ones circulate is a theme
of Tymoczko 2011b, which shows that it is productive to think of pentatonic and diatonic
collections as circulating about perfectly even, and thus microtonal, collections. See also Douthett
2008.
13. Saslaw translates Nebenton as “secondary tone,” emphasizing that these neighbor tones are
chromatic to the respective keys.
 CHAPTER 3 Reciprocity  57

apart from its particular instantiations. The first is in the sense of “Isaac was born
of Abraham”; the second, in the sense of “invention is born of necessity.”
In chapter 7, Weitzmann explores the augmented triad’s Mehrdeutigkeit, in the
sense that interested Vogler fifty years earlier. He notes that once enharmonic vari-
ants are taken into account, the AᅈCE triad also arises in four other keys, besides
the f minor and C major already explored: “So we find the augmented triad AᅈCE
and its enharmonic equivalents in the nebenverwandt keys F minor and C major,
further in the relative keys of each, in Aᅈ major and A minor, finally in the
nebenverwandt keys of the latter, in Dᅈ minor and E major” (186–87). Although
Weitzmann’s ordering has transformational implications that we will consider in
chapter 4, for him that ordering evidently held no value except as an aid to memory.
On a subsequent page, he lists the same six keys in the format reproduced here
as the first block of table 3.1, writing that CEGᅊ and its enharmonic equivalents
“can appear as the most important [augmented triad] of the following keys listed
under them” (188–89). The remaining eighteen triads are grouped into three anal-
ogous clusters, each headed by an augmented triad and listing the keys in which it
is “the most important.”
Having created this list, Weitzmann’s discourse begins to project a subtle
inversion. Until now, he has viewed the augmented triad as a serf in the employ
of the particular consonant triad from which it arises. But now, having observed
that each augmented triad has multiple patrons, he begins to wonder what life
is like from its point of view. “The closest relatives of an augmented triad,”
he writes, “are thus the major triads on its bass tone, third, and fifth, [plus] the
minor triads to whose roots each of [the augmented triad’s] three voices forms
the leading tone. . . . Its more distant relatives are the minor versions of the
just-designated major chords and vice versa” (188–89). Several chapters later,

Table 3.1. Weitzmann’s grouping of the consonant triads as displacements

of augmented triads

I. {C, E, Gᅊ} (and its enharmonic transformations)

1. C major 2. E major 3. Aᅈ major
4. a minor 5. cᅊ minor 6. f minor
II. {Dᅈ, F, A} (and its enharmonic transformations)
1. Dᅈ major 2. F major 3. A major
4. bᅈ minor 5. d minor 6. fᅊ minor
III. {D, Fᅊ, Aᅊ} (and its enharmonic transformations)
1. D major 2. Gᅈ major 3. Bᅈ major
4. b minor 3. eᅈ minor 4. g minor
IV. {Eᅈ, G, B} (and its enharmonic transformations)
1. Eᅈ major 2. G major 3. B major
4. c minor 5. e minor 6. gᅊ minor
58  Audacious Euphony

Figure 3.9. From Weitzmann’s Der übermässige Dreiklang. Upper- and lower-case
letters are the roots of major and minor triads respectively.

Weitzmann graphically portrays these relationships in a diagram that is repro-

duced here in translation as figure 3.9.
Each augmented triad is presented at the center of a cluster of consonant triads;
major and minor triads are indicated by large- and small-case roots, respectively.
And here is where Weitzmann’s third genesis tale involving the augmented triad
can be found:

From the following augmented triads . . . arise [entstehe] the [consonant] triads
indicated by the letters next to them. . . . The chords placed immediately next to the
augmented triad are attained through the half-step progression of one of their
voices; the [chords] further away [are attained] through the half-step progression
of two of their voices. (202–5)

With this, Weitzmann turns back the flow of his second genesis narrative.
At the systematic level, it is the augmented triads that are the sources, and the
consonant triads the products.
The first genesis narrative nonetheless remains intact. Immediately following
the passage just quoted, Weitzmann presents seven full pages of examples, com-
prehensively enumerating the ways that an augmented triad can resolve. It is
always the dissonance that is resolving to the consonance, never the other way
around. In a moment-to-moment sense, the relation of consonant triad to disso-
nant augmented triad continues to be diodic. But in a systematic sense, Weitzmann
is able to entertain the possibility that the relation is reciprocal.
These passages from Weitzmann’s treatise are so rich in implication that they
guide the work presented in the next three chapters of this book. Chapter 4 consid-
ers the internal structure of the six-triad pools that are clustered in table 3.1, from
the standpoint of the Tonnetz graphics and triadic transformations introduced in
chapter 2. Chapter 5 uses figure 3.9 as a stage from which to extend the Tonnetz,
and its attendant transformations, so that it breaks out of the augmented triad
boundaries that confined them in chapter 2. Chapter 6 uses that extended universe
as a playing board, or map, upon which to present pan-triadic analyses of extended
passages from the Romantic repertoires, and to assess and categorize those
passages on the basis of the voice-leading strategies that they execute.
C HA P T E R
Four
Weitzmann Regions

This chapter explores a second preliminary model of the triadic universe, based on
table 3.1’s six-triad groups, which we will refer to as Weitzmann regions. The initial
model studied in chapter 2, based on minimal-work voice leading, also features
six-triad groups, the hexatonic cycles. Both models partition the twenty-four
triads into eight triplets, each of which contains three major-third–related triads of
the same species, and each of which is paired with a triplet featuring triads of the
opposite species. Where hexatonic and Weitzmann regions differ is in how those
triplets are paired. Hexatonic regions pair major triads with minor triads built on
the same root, or on roots four semitones away. Weitzmann regions pair triads
whose roots lie an odd number of semitones apart.
The first half of this chapter studies the internal structure of Weitzmann
regions, focusing on their voice-leading properties, graphic representations, and
the transformations that connect the triads within a region. The second half sam-
ples the diverse ways that composers of the long nineteenth century, from C. P. E.
Bach in 1763 to Richard Strauss in 1949, first dipped their paddles into sectors of
a Weitzmann region and ultimately ran their canoes directly down its rapids.

The Structure of a Weitzmann Region

The Weitzmann regions have a quite different internal structure from the hexa-
tonic regions studied in chapter 2. The six triads of a hexatonic region naturally fall
into a cycle, an inherently graded space in which triadic distance correlates with
voice-leading work. On that same basis, a Weitzmann region is a flat, uniform
space: all of its triadic pairs stand exactly two voice-leading units apart. The triads
can be cyclically ordered in several ways, but there is no ordering that is more
natural than the others, from the standpoint of voice leading.
Figure 4.1 uses two graphic metaphors to suggest the voice-leading structure
of a Weitzmann region. The image at (a) is of a water bug with augmented-triad

59
60  Audacious Euphony

Figure 4.1. Two portraits of the structure of a Weitzmann region.

body and six consonant-triad feet, the three on each side representing the modally
matched subregions. All journeys from one foot to another are routed through the
body. Some pairs of feet are located on a single side, others directly opposite, and
others across the body at an angle. Each of these three pair types requires a two-
stage connection: one from the source foot into the body, and one from the body
to the destination foot.
Figure 4.1(b) reimagines the water bug as an oval hallway with three rooms on
each side. This metaphor encourages one to inhabit the space and travel through
it. Moving between rooms, one might stay on the same side, or move directly
across, or move across at an angle. Each of these three trajectories can be experi-
enced in various ways. A firefighter would likely execute each of them with a single
kinetic impulse and experience each as a single Gestalt of essentially equivalent
magnitude, despite small differences. A professor might experience the hallway as
a marked intermediate point that articulates the journey into two stages (“I came
into the hallway for a reason. . . . Oh, yeah, to go to class”). A retiree might be
inclined to linger in the hallway for a while and see who else might be passing
through. In the same way, the augmented triad may or may not be marked as such:
some music will linger there, some will invite passing notice of its features, and
some will rush quickly through the passageway without registering any impres-
sion of it.
Figure 4.2 sets this abstract structure to music, modeling the voice leading
between C major and the five other triads in its region. The first stage of each route
is the same: G → Gᅊ/Aᅈ converts C major to C augmented. What distinguishes
them from each other is what happens at the second stage: which tone moves
by semitone, and in which direction. The progressions at (a) and (b), connecting
two major triads, are equivalent to connections between legs on a single side of the
water bug. The two stages have identical magnitude but opposite direction, preserv-
ing the center of balance discussed in chapter 2. Although these progressions are
coextensive with the LP and PL transformations through a hexatonic cycle, it is not
clear that these labels are pertinent, as they suggest that each motion is implicitly
routed through a minor triad, not an augmented one. I will nonetheless retain the
labels despite their isolation from their etymological source, since ordinary lan-
guage works that way as a matter of course. (We say that this book is written in the
English language even though no tongue is involved in producing it.)
 CHAPTER 4 Weitzmann Regions  61

Figure 4.2. Voice leading from C major to the remaining five triads of its Weitzmann
region.

The progression at (c), C major → a minor, resembles a straight line across the
body. Here the two stages are consolidated in a single voice that moves by whole
step, rather than divided among two voices moving by semitone as in all of the
other intraregional progressions. The progressions at (d) and (e) connect C major
to the two remaining minor triads in the region, f minor and cᅊ minor. As in
a hexatonic region, the three cross-modal progressions shift the bug’s center of
balance from one side to the other. The distance of that shift is by two units, more
unsettling than the analogous hexatonic transformations (L, P, and H), which
shift that center by a single unit.

Weitzmann Transformations and N/R Cycles

Although the triads of a Weitzmann region have no natural cyclic ordering on the
basis of voice-leading proximity, they do fall quasi naturally into a cycle on the
basis of their historical origins in classical syntactic routines. This ordering was
already suggested by Weitzmann in his initial, heuristically motivated description
of the regions (quoted on p. 57 of this book). The chain suggested by that account
is presented as figure 4.3. Diachronically, the account radiates outward from C
major and f minor and terminates at the boundary triads, E major and dᅈ minor.
The latter two form a direct connection that links the chain into a cycle. This is not
a connection that Weitzmann indicates. It does not suit his purposes to do so,
since his interest is in the contents of a region rather than its internal structure,
and he has already generated all six triads without this final link in the cycle.
Although Weitzmann conceived of nebenverwandt and relative as bilateral
relations between keys, his conception easily converts into a set of idealized
voice-leading actions on triads, each of which is associated with a triadic
transformation.1 The nebenverwandt transformation (abbreviated N) takes any

1. I noted in chapter 1 that triads and tonics are interchangeable in much nineteenth-century theory.
The transformational conception, although not explicitly present in Weitzmann’s 1853 monograph,
62  Audacious Euphony

Figure 4.3. A graphic depiction of the N/R chain described in Weitzmann 2004
[1853], extended to a cycle.

triad to its nebenverwandt chord, which Weitzmann defines as the inversion about
the root of a major triad, or about the fifth of a minor triad.2 The relative transfor-
mation (abbreviated R) takes a consonant triad to its relative major or minor.
Figure 4.3 accordingly can be characterized as an N/R cycle, analogous to the L/P
alternation that generates the hexatonic cycle.
Complete N/R cycles appear in at least four passages from Schubert’s
instrumental music, always traveling in the “authentic” direction (clockwise on the
figure 4.3 cycle). Figure 4.4, from the first movement of his Fourth Symphony of
1816, is the chronologically earliest of the four, and the only one to restrict its
lexicon to consonant triads.3 Each major triad is prolonged as a local tonic, which
each minor triad serves as subdominant. In the remaining three passages, these

Figure 4.4. Schubert, Symphony no. 4, 1st mvt., mm. 86–106.

is not egregiously anachronistic: Oettingen 1866 interpreted the nebenverwandt in explicitly

transformational terms (Mooney 1996, 71).
2. Weitzmann’s teacher, Moritz Hauptmann, defined this relation as the inversion about the tone that
participates in both the major third and the perfect fifth of its triad (1888 [1853]). Morris 1998 pro-
poses the label L′ (L-inverse) for this transformation, whose semitonally moving and stationary
voices swap those of L.
3. Langlé 1797 contains, among a number of four-voice synthetic examples, a progression that embeds
an N/R cycle (p. 73, no. 26).
 CHAPTER 4 Weitzmann Regions  63

Figure 4.5. Schubert, Symphony no. 9, 1st mvt., mm. 304–15.

functional roles are exchanged: minor triads are local tonics, and the major triads
that precede them are embedded within dominant sevenths.
These passages include the famous trombone solo from the first movement
of the Ninth Symphony, partly presented at figure 4.5. The trombones first carry
solo material at mm. 197–219 of the exposition, through a partial N/R chain. The
passage excerpted here, from near the end of the development, presents the first
of two passes through a complete N/R cycle. The addition of the sevenths puts
all twelve pitch classes in play, enabling Schubert to present complementary
whole-tone scales in the treble register, chaining pungent minor-second suspen-
sions with minor-third resolutions (Krebs 1980, 87).4 Liszt’s appropriation of the
progression in the first movement of his Faust Symphony of 1854 (Cohn 2000)
resembles figure 4.4: it moves in the authentic direction, employs no chordal
dissonances, and accents the major triads as local tonics. In several other composi-
tions, Liszt runs the cycle in the reverse direction, exploring its plagal capacity:
minor subdominants move to their major tonics, and minor tonics, to their major
dominants.5

4. Whole-tone scales are also evident in the two remaining Schubert passages, from the Octet (Taruskin
1996, I: 261) and the c minor Piano Sonata, presented below as figure 4.17(c). See also Borodin’s
Second Symphony, 1st mvt., mm. 41ff.
5. “O Lieb, so lang du lieben kannst” of 1845, transcribed as the Liebestraum no. 3 for piano, and the
piano etude “Vision” of 1852 (see Ahn 2003, 69).
64  Audacious Euphony

In focusing on the combination of N and R, we have neglected the Weitzmann

region’s remaining mode-reversing relation, C major → cᅊ minor at figure 4.2(e).
Like most juxtapositions of chromatically related triads, this one lacks a canonical
name. Sigfrid Karg-Elert (1930) called this the same-third (Terzgleich) progression,
and David Lewin (1987) proposed Slide, both names of which (following Kopp
1995, 289) suggest letter S.6 This direct pairing is not classically normative, because
S-related triads share no membership in a diatonic scale. In the eighteenth
century, S-related triads are less likely to be juxtaposed directly than to substitute
for one another, as alternative modal expressions of a mediant or submediant
degrees. It is only with Schubert that the progression achieves prominence, both
as an indirect relationship between tonics and as a direct juxtaposition between
triads.7 S-progressions become more frequent and fluent in Liszt, Wagner, and
Prokofiev and have become a staple of film music (Lehman 2010). The Slide
transformation nonetheless retains a maverick status among the Weitzmann
transformations, similar to H within the hexatonic group.
Figure 4.6(a) portrays a Weitzmann region on the Tonnetz. The structure
resembles a stalk with six leaves. The thickening of the stalk enhances the presence
of the augmented triad as an entity, no longer the mere boundary of chapter 2.
Figure 4.6(b) superimposes transformational arrows that model the five progres-
sions studied in connection with figure 4.2, as they act on C major. Four of the
arrows pass through a vertex that contains the tone common to the connected
triads. The fifth arrow, labeled R, passes across an edge that connects the two tones
common to C major and a minor.
Here we encounter one of the benefits of treating the augmented-triad axis as
an object rather than as a boundary. The latter interpretation conveys the impres-
sion that R involves less distance than the other Weitzmann-internal relations
(Tymoczko 2009b). Indeed, it suggests that R covers the same distance as L and P,
the other common-tone maximizers. From the standpoint of voice-leading mag-
nitude (“work”), this is an illusion. R is a more distant relation than L and P, just
as surely as a whole step is larger than a semitone, and is neither closer nor farther
than the other four relations internal to a Weitzmann region. Considering the
augmented-triad stalk as an object dispels that illusion.
This status can be rendered unmistakably clear by expanding the stalk of
figure 4.6(a) into the gray parallelogram of figure 4.6(c). Like the consonant triads,
the augmented triad is a two-dimensional polygon with boundaries and area
conveying its status as a surface in its own right, rather than just a boundary
between regions. All voice-leading distances are now “true.” The image is
consistent with the hallway metaphor of figure 4.1(b), and with the notion of a

6. Karg-Elert’s treatment of the common-third progression has had a particular influence on harmonic
theory in Russia, as documented in Segall 2011.
7. Direct juxtapositions occur in the Characteristic Allegro for Piano Four Hands (Lebensstürme),
m. 260 (Rings 2006, 202), and mediated by an augmented triad at the opening of the Sanctus from
the Aᅈ major Mass. Prominent structural S-pairings occur in Lebensstürme, the opening two move-
ments of the Bᅈ major Piano Sonata, and the second moment of the String Quintet. Rings’s (2006)
analysis of Lebensstürme engages many of the issues treated here.
 CHAPTER 4 Weitzmann Regions  65

Figure 4.6. Three depictions of a Weitzmann region on the Tonnetz.

Weitzmann region as a free-commerce zone rather than one that is governed by a

natural internal order.

Remarks on the Tonnetz

These multiple depictions of the augmented triad represent a tension that has been
worked over repeatedly in the last chapter and a half. On the uninflected Tonnetz
of Arthur von Oettingen and Hugo Riemann, the major-third axis is simply
another boundary that, like those generated by the remaining consonant intervals,
contributes to the definition of triads. The area that it attains in figure 4.6(c)
suggests its capacity to represent a musical object that might be prolonged for a
significant span of compositional time. The ambiguous status of R in figure 4.6
manifests this tension and clarifies what is at stake. On the common-tone basis on
which much nineteenth-century harmonic theory rests, R joins its common-tone–
maximizing brethren, L and P, at the first rank. The uninflected Tonnetz captures
66  Audacious Euphony

this view, bestowing equivalent status on the three axes generated by the conso-
nant intervals.8 By contrast, from the viewpoint of voice-leading work, measured
in semitonal units, L and P alone are the primary triadic relations; R joins the
remaining Weitzmann transformations at the second rank. The heavily adapted
Tonnetz of figure 4.6(c) captures this second view. A compromise position is sug-
gested by the mild inflection of the major-third axis of figure 4.6(a), which is sus-
ceptible to either interpretation. This flexibility may or may not be a good thing,
depending on one’s interpretive values and goals.
There are good reasons to prefer voice-leading measures to common-tone–
based measures and, accordingly, to ship the Tonnetz off to the museum, as
Tymoczko 2009b recommends. Figure 4.7 models the distribution of the twenty-
three types of triadic pairings (i.e., connecting some fixed triad to each of the
remaining triads) under three measurements: (a) number of common tones;
(b) Tonnetz distance, measured as edge-traversals that represent double-common-
tone retention; and (c) voice-leading work. Method (c) creates the most categories
and most evenly distributes the twenty-three progressions among the available
categories.
Voice-leading distance is also superior to Tonnetz distance because it provides
more intuitively satisfactory results, in those cases where their results disagree.
Tymoczko (2009c, 2010) notes that modally matched fifths such as C major → F
major are closer on the Tonnetz than are nebenverwandt relations, such as C major
→ f minor, even though the latter progression involves less voice-leading work. He
notes further that these discrepancies become more acute when one moves from
the specific case of triadic relations in a twelve-note universe to generalized chordal
relations in an n-note universe. In part for these reasons, he uses fused-triad graphs
such as figure 2.1, where points represent chords rather than their pitch-class
constituents.
There are six reasons why I will nonetheless continue to cultivate the Tonnetz
as a primary (but not exclusive) mode for representing the triadic universe. First,
when the augmented-triad axis is interpreted as an object as in figure 4.6(c),
distances on the Tonnetz align with those on the fused-triad graphs. The discrep-
ancies that Tymoczko identifies are thus neutralized, and the Tonnetz can be
treated as a true map of voice-leading proximity. Second, although elsewhere

Figure 4.7. Triadic distributions for three distance measures.

8. I adopted that view in Cohn 1997, and it has been developed since by a number of theorists, such as
Capuzzo 2004, Goldenberg 2007, and McCreless 2007.
 CHAPTER 4 Weitzmann Regions  67

I have explored nineteenth-century triadic practice against a general backdrop of

hypothetical possibilities (Cohn 1997), the primary aims of the present work are
more historical and analytical. The limitations of the Tonnetz in the general case
are not as limiting for my specific purposes. Third, the “unconformed” (or flat)
Tonnetz fixes the directional axes, facilitating comparisons between passages and
compositions. Although fused-triad graphs can also be fixed in this way (Tymoczko
2011b, 416–17), their directional axes lose legibility because some of their compo-
nents recede into a third dimension. Fourth, chapter 6 shows that there are advan-
tages to maintaining individual tones as primary objects, rather than prepackaging
them into triads. The Tonnetz can track continuities among individual pitch
classes, yielding analytic information that would otherwise be difficult to recover,
and provides locations for other pitch-class combinations that arise in nineteenth-
century music. These include consonant dyads and dissonant seventh chords that
would otherwise need to be referred to a consonant triad via expansion or con-
traction, forcing interpretations that might be underdetermined or even arbitrary.
Fifth, the Tonnetz maintains contact with a certain mode of historical thinking,
recovery of which is part of the project of this book. Finally, because the Tonnetz’s
historical origins are associated with ideas about tuning, it presents an opportunity
to explore the interplay of the triad’s two forms, and hence of the two syntaxes to
which they respectively give rise. This exploration is undertaken in chapter 8.
There will nonetheless be circumstances where the fused-triad graphs
prove useful. They have their own more recent history, with which we will also
want to maintain contact. Their elimination of the atomic pitch-class level makes
them trimmer and more geometrically flexible. As shown in chapter 2, they wrap
more easily into a cycle. Accordingly, fused-triad graphs are substituted when
surfaces are composed exclusively of triads, and when focusing the eye on cyclic
closure is more central to the analysis than tracking particular pitch classes or
voices.

Historical Origins of Weitzmann Regions

I am aware of no eighteenth-century compositions that explore all six triads of a

Weitzmann region. Its origins can nonetheless be traced to classical routines, nota-
bly involving a i–III–Vᅊ complex. In the most typical case, relative major divides
the motion from minor tonic to major dominant. (Particularly in the eighteenth
century, such progressions are more likely to occur as structural pillars—middle-
grounds—rather than as direct successions. See Schmalfeldt 2011, chap. 8.)
Expressed in transformational language, and as depicted in figure 4.8, N is exe-
cuted by combining R with LP. In a more direct variation, the order of the final
two triads can be exchanged. Minor tonic proceeds to its major dominant via N;
after a fermata, an LP disjunction, with characteristically paradoxical contrary
motion, brings new music in the relative major, characteristically at a faster tempo
(LaRue 2001 [1957]). That the three chords form an unordered space during the
later part of the eighteenth century is underscored by C. P. E. Bach’s assertion that
68  Audacious Euphony

Figure 4.8. Minor tonic, major mediant, and major dominant, as a subset of a
Weitzmann region.

direct motion from C major to E major can be executed by anyone “who knows
that E is the dominant of a and that a minor is very closely related to C major”
(Kramer 1985, 552).
There are several ways to expand this core progression by adding a fourth
triadic member of the Weitzmann region. In a famously perplexing passage from
C. P. E. Bach’s f minor Piano Sonata of 1763, Aᅈ major (III) is connected to C major
(Vᅉ) through an anomalous Fᅈ major, the tonic’s Slide relation. Although the con-
temporaneous critic Johann Nikolaus Forkel found no beauty in this passage, he
nonetheless defended it on the grounds that it was written by a great composer and
ipso facto must contain something of virtue (Kramer 2008, 11–12). Forkel and his
contemporaries were likely intrigued and undone by the diatonic indeterminacy
of this interpolated chord, which can be understood as Fᅈ major only in retrospect
and as E major only in prospect. Once enharmonic distinctions are neutralized in
chromatic space, the progression is understood in terms of two parallel PL motions
that substitute for the single LP of figure 4.8.9 Figure 4.9 depicts these local pro-
gressions as downward motions on the Tonnetz and recuperates dominant to its
proper location by means of a broken arrow that recognizes the cyclical structure
underlying the planar representation.
A more common way to expand the i–III–Vᅊ core is to juxtapose the relative
major with its own minor subdominant (ivᅈ of III). Figure 4.10 shows the basic
complex of relations. C minor is flanked by its major dominant and relative major.
The latter in turn is flanked by its minor subdominant. The transformational
symbols indicate that the inner two chords are R-connected, the outer adjacencies
are N-connected, and the progression is a contiguous segment of an N/R cycle
(figure 4.3). The remote outer triads are connected by S, but the identity of their
common third is obscured by the enharmonic distinction indicated by the broken
arrow. The Bᅉ/Cᅈ enharmony is essential: no renotation will bring them into
orthographic conformance without sacrificing locally consonant and diatonic
connections, each of which is fully salient to perception. The relation of these two

9. Recall from chapter 2 that PL and LP progress two stations around a six-station hexatonic cycle,
but in opposite directions. A four-station journey in one direction reaches the same destination as a
two-station journey in the other; hence, PL × 2 = LP.
 CHAPTER 4 Weitzmann Regions  69

Figure 4.9. Tonnetz model of C. P. E. Bach, Sonata in f minor, H. 173 (Wq. 57/6),
1st mvt., development.

Figure 4.10. A double-agent complex about C minor/Eᅈ major.

tones creates an opportunity for a composer to gently tease the enharmonic seam
or, if so inclined, to plunge directly through it.
Two Chopin compositions from 1830 illustrate the two options. Figure 4.11(a)
presents mm. 8–16 of the first Mazurka (Op. 6 no. 1, fᅊ minor). After tonicizing
the relative major at m. 12, Chopin leads three times to its subdominant: twice as
a minor triad and once as major. The replacement of Fᅊ for Fᅉ prevents the latter
from coming into direct contact with its enharmonic equivalent Eᅊ when the
global dominant arrives at the end of that measure.10 Figure 4.11(b) presents the
opening phrase of the first Nocturne (Op. 9 no. 1, bᅈ minor). After attaining
the relative major at m. 5, Chopin again leads three times to its subdominant, twice
as a minor triad and once as a major. Here, though, the order is reversed: Dᅈ’s
major subdominant precedes the twice-iterated minor version. The direct motion
from gᅈ minor to F major enacts a Slide transformation that transforms Bᅈᅈ3

10. The enharmonic duality of Eᅊ and F (circled in the score) inspired a significant set of analytic
musings in Hyer 1994. I cannot, however, follow Hyer in hearing d minor as a local tonic, because
of the Bᅉs, and because his suggestion is inconsistent with the relative stability of the later and lower
member of motivic semitones (F → E, Fᅊ → Eᅊ) throughout the Mazurka. Similar progressions arise
in the finale of Chopin’s b minor Piano Sonata (compare the Bᅈ at m. 12 with the Aᅊ at m. 16) and
in his f minor Etude Op. 25 no. 2, where accented Fᅈ in the tenor register at mm. 16–17 anticipates
the E4 when dominant arrives on the downbeat of m. 19.
70  Audacious Euphony

Figure 4.11. Two early Chopin examples.

directly into A3 in the same voice (as indicated by the arrow); the two notated
pitches are metrically accented and analogously positioned in the arpeggiated
figuration.11 What is entered as the third of a minor triad, pressing downward by
semitone, exits as the third of a major triad, pressing upward by semitone.
Figure 4.12 presents Tonnetz models of both phrases. In the Mazurka (a), the
motion to d minor at the bottom of the figure is undone, and relative major pro-
ceeds to dominant. The minor subdominant is thus a sideshow that delays the

11. Compare Smith 1986, 123–24. I consider the Eᅈ that is sounded with the gᅈ minor triads at both
mm. 6 and 7 to be a dissonant under-seventh that transforms at m. 8 into a dissonant over-seventh.
Justification of this position is deferred until chapter 7.
 CHAPTER 4 Weitzmann Regions  71

Figure 4.12. Tonnetz models of figure 4.11.

structural motion to the goal dominant. In the Nocturne (b), the S transformation
is modeled on the Tonnetz as a downward motion, enharmonically reinterpreting
the common tone and landing on the dominant at the bottom of the strip. The
identification of that dominant with the one adjacent to the tonic at the top of the
strip is indicated by the broken line at the right side of the figure.
72  Audacious Euphony

The Double-Agent Complex

Developing an idea that was prevalent at the turn of the twentieth century,
Daniel Harrison (1994) has suggested that the melodic energy invested in thirds
bestows agency upon their constituent harmonies: triadic major thirds strive
upward, bearing dominant energy, while triadic minor thirds press downward
with subdominant “attitude.” These appellative tendencies are intensified when
the thirds are chromatic inflections that displace their diatonic equivalents. In the
Chopin Nocturne, the downward agency of Bᅈᅈ is converted instantly into the
upward agency of A. I find that I need to sing the pitch continuously in order to
convince myself that the two pitches are identified; and even then, it feels like a
trick.
The redirection of an energized tone through enharmonic transformation is
one of Vogler’s (1802) principal genera of Mehrdeutigkeit. In the most familiar
case, the root of a diminished seventh chord (FᅊACEᅈ) is reinterpreted as a
seventh (ACEᅈGᅈ), redirecting the energy from g minor to its relative major, Bᅈ
major. In an analysis of two passages from the first movement of Beethoven’s
Fifth Symphony, Lewin recognized the rhetorical potential of this redirection and
vested it inside an anthropomorphic parable:

[In the exposition] you enter . . . wearing your Fᅊ cloak, as leading tone to G; but
you abruptly hurl the cloak away and reveal yourself in a suit underneath as Gᅈ,
upper neighbor to F. . . . [At the parallel moment of the recapitulation] everyone is
waiting for you throw off your Fᅊ cloak and reveal yourself as Gᅈ. You throw off
your Fᅊ cloak all right, but now you are wearing an Fᅊ suit beneath it! You resolve
as a leading-tone to G. (2006, 107)

Lewin’s story is a Cold-War allegory, a musical version of a trope that arises in

Fleming, LeCarré, Fowles, and dozens of dime-store novels, movies, and television
scripts from the era from the end of the war to the fall of the wall. It is the story
of the double agent. All is not what it seems. Or perhaps it is what it seems, in
which case it is not what it seemed not to seem. And so forth, in a recursive fractal
explosion, an eternal ricochet, a toggle that is turtles all the way down.
Some double agents are gnarl-visaged trench-coat lurkers, and others sparkle
and glisten in honeyed tones and décolletage. The two Chopin passages suggest
that musical double agency is found not only in the dissonant harmonies of Vogler
and Lewin but in acoustically perfect beauties as well. Both passages present the
components of the double-agent complex in a canonical order, and in a canonical
tonal context. Consideration of three further pairs of excerpts will demonstrate
that nineteenth-century composers found the complex remarkably malleable,
both in the ordering of its elements and in the tonal environments in which they
operate.
The first pair, consisting of Wagner’s Tarnhelm motive (1853; figure 2.4) and
the opening phrase of Liszt’s Il Penseroso (c. 1840; figure 4.13) from volume 2
of the Années de pèlerinage, demonstrates the permutational possibilities of
the double-agent complex. The Tarnhelm swaps the order of two of Chopin’s
 CHAPTER 4 Weitzmann Regions  73

Figure 4.13. Liszt, Il Penseroso, mm. 1–9.

components: the relative major follows rather than precedes its minor subdomi-
nant. This produces a direct LP progression from gᅊ minor to e minor, creating the
uncanny effect discussed in chapter 2, and temporarily destabilizing gᅊ minor’s
status as tonic. E minor’s tonic potential is suggested by the opening melodic
gesture and by the half-cadential rhetoric of the B open fifth. Wagner often
exploits this potential by adding an under-seventh F to the initial triad and resolv-
ing it as an augmented sixth. B major also accrues potential tonic value, which is
activated on those occasions when the second phrase is omitted. At the same
time, the salience of the underlying i–III–Vᅊ core is sufficient to prompt a tonally
well-formed hearing in gᅊ minor (Darcy 1993, 169–70).12
The opening phrase of Il Penseroso omits the relative major altogether, only
to restore it as the goal of its second phrase. The composition opens with an LP
progression, as in the Tarnhelm (which it precedes by a decade); after acquiring
an Fᅊ under-seventh, a minor proceeds directly to dominant Gᅊ major through
a Slide operation that reinterprets C as Bᅊ (as in the Chopin Nocturne).13 The
subdominant potential of a minor is exercised in the composition’s second
phrase, which ends by tonicizing the relative major, represented as an open fifth as
in the Tarnhelm. (In the event, Liszt fills the open fifth as e minor at the beginning
of the third phrase.)

^
12. The passages that fit this model problematize Kevin Swinden’s assertion (2005, 261) that ᅈ1 always
^ ^
notationally stands in for the leading tone. Swinden acknowledges that although “bass line (ᅈ)6–(ᅈ)3
. . . may support such structures, I have yet to find convincing examples to demonstrate this
pattern.” He cites as typical the Tarnhelm progression (see figure 2.4 in chapter 2 of this book),
^
“where ᅈ1 was treated as a leading tone (F ) in the key of Gᅊ minor.” Yet the subsequent
^ ^
progression of e minor to a BFᅊ dyad presents an instance of what he was unable to find: a 6 → 3 bass
^ ^
in gᅊ minor, supporting a ᅈ1 → (ᅈ)7 soprano.
13. Rings 2011, 78, has some pertinent comments on the affective paradox that this respelling
represents.
74  Audacious Euphony

Figure 4.14. Two passages from Schubert’s Impromptus, Op. 90.

The double-agent complex is as tonally adaptable as it is permutationally mal-

leable. In each of the examples examined so far, tonic has been situated on the
interior minor triad of the four-triad complex. (It is interior with respect to the
layout of figure 4.10, where the R-related triads are interior and the S-related
ones exterior.) Figure 4.14 presents excerpts that tonicize exterior members of
the complex. The excerpts are related by source and by formal function: they begin
the codas of Schubert’s second (Eᅈ) and third (Gᅈ) Impromptus (D. 899, Op. 90,
1828), compositions that have been substantively linked in other ways (Fisk 2001,
46–47). The Eᅈ Impromptu is a sectionalized ternary composition, both of whose
principal sections terminate with a scalar eᅈ minor leading to a marcato Gᅈ major
(figure 4.14(a)). The four measures that begin the b minor interior section are
replicated at the beginning of the coda. Upon its initial presentation, this phrase
 CHAPTER 4 Weitzmann Regions  75

Figure 4.15. Tonnetz models of figure 4.14.

elicited a V → i consequent in b minor. In the coda, the i → V antecedent in b minor

is answered by V → i in eᅈ minor. The upward transposition by major third across
the caesura, transforming dominant Fᅊ major to dominant Bᅈ major via LP, is
undone by a downward transposition four measures later, transforming tonic
eᅈ minor to tonic b minor, also via LP (figure 4.15(a); compare Rings 2007, 60).
The double agency of Eᅈᅈ qua D is quite piquant in the measure following the half
cadence, as indicated by the arrows in figure 4.14(a).
The third Impromptu’s coda enters its double-agent complex from the opposite
side, with Gᅈ-major tonic playing the role of exterior major triad (figure 4.15(b)).
The coda begins by transforming Gᅈ major into a local dominant of cᅈ minor,
exactly as in the Second Impromptu (figure 4.14(b)). Cᅈ minor proceeds to its
relative major, Eᅈᅈ major, which functions as a German sixth chord, immediately
restoring Gᅈ major as tonic. (The trio of triads is as at figure 4.8, but here a major
tonic arpeggiates down to its subdominant, instead of a minor triad arpeggiating
up to its dominant.) The second phrase of the coda retraces the same path but
goes a step farther: D major acts as dominant to a fleetingly tonicized g minor,
transposing the earlier Gᅈ → cᅈ and completing the four-triad complex.14
Figure 4.14 suggests the tonal adaptability of the double-agent complex, which can
in principle center around any of its four triadic constituents.15 The final two examples,
presented at figure 4.16, demonstrate that the double-agent complex can also be
present where none of its constituent chords serves as tonic. Figure 4.16(a) presents
some measures from the first song of Schumann’s Dichterliebe, “Im wunderschönen
Monat Mai.” The complex is centered on local tonics b minor and D major,

14. This passage is analyzed in Riemann 1877.

15. Only three of the four possibilities are exemplified here. The fourth possibility, oriented toward the
interior major triad, is represented by the development section of Mozart’s Piano Sonata in Bᅈ,
K. 333, which makes much of the enharmonic relation between Gᅈ qua subdominant agent and
Fᅊ qua dominant agent of an unrealized g minor tonic.
76  Audacious Euphony

Figure 4.16. Two nontonic examples of a double-agent complex.

neither of which is a candidate for global tonic. The same complex appears,
untransposed and in the same register, at mm. 3–5 of “Ich will meine Seele
tauschen,” the fifth song of Dichterliebe, which is in b minor (Komar 1971, 75).
Brahms’s song “Von ewiger Liebe” (Op. 43 no. 1), also in b minor, presents an
even more tonally remote version of the double-agent complex. After the first
couplet tonicizes the mediant, the second couplet, presented as figure 4.16(b),
begins with its modal variant, d minor. A double-agent complex sets the couple’s
journey through the smokeless darkness, “Nirgend noch Licht und nirgend
noch Rauch.” The parallel halves of the text are set by a parallel set of motions
from a tonic to its major dominant: first d minor → A major, and then fᅊ minor →
Cᅊ major, each harmony extending for one measure. The four-measure unit thus
connects the tonic’s minorized mediant, whose F agent discharges downward, to
the dominant of its dominant, whose Eᅊ agent presses upward (see Karg-Elert
1930, 271).

Expanded N/R Chains

Although the four-element double-agent complex attained a life of its own in the
nineteenth century, it also began to expand in the direction of the fully ramified
 CHAPTER 4 Weitzmann Regions  77

Figure 4.17. Three expanded N/R chains.

six-element N/R cycle, and complete Weitzmann region, that we find already in
Schubert (see figures 4.4 and 4.5). Figure 4.17 models two chromatic excerpts from
late-eighteenth-century pieces, both of which embed a double-agent complex
within a five-triad N/R chain. As it happens, both progressions feature the same
set of triads, presented in the same order, although the progressions are embedded
in compositions whose tonics are a tritone apart! The first excerpt is from Jiří
Benda’s Piano Sonata in a minor (1783), of which Rey Longyear and Kate
Covington (1988) write that “in the eighteen measures of the second theme-group
the composer starts in C major (III), pauses on a half cadence in E minor, then
goes to such remote keys as C minor and Aᅈ major before the sudden modal muta-
tion into C major for the closing theme-group” (454). Figure 4.17(a) models the
passage beginning from the B7 caesura, whose resolution to e minor occurs in 63
position, precipitating a series of harmonies over a G pedal, and connecting
e minor to its Slide-related Eᅈ major through an alternation of N and R. The entire
78  Audacious Euphony

N/R chain-segment connects B major to Eᅈ major, neither of which acquires an

easy role in the overall C major prolongation.
Figure 4.17(b), adapted from an analysis of James Webster (1991, 321), models
a passage from the opening Adagio of Haydn’s Symphony no. 99 (1793) in Eᅈ major.
Haydn comes upon B major by renotating the upper neighbor of Eᅈ’s dominant.
The five-chord progression issuing from B major is the same as Benda’s, although
the final connection between C minor and tonic Eᅈ major is divided by a retro-
gression to a prolonged G major, acting as the former’s back-relating dominant.
The culmination of this process of expansion can be seen at figure 4.17(c), from
Schubert’s Piano Sonata in c minor (1828), where the sequential progression of
Haydn’s Symphony is compressed and extended to cover the entire six-triad
Weitzmann region as a fully ramified N/R chain, such as that presented earlier in
connection with figures 4.4 and 4.5.
In sketching the expansion from the four-chord double-agent complex, which
might merely tease the enharmonic seam, to the six-chord N/R chain, which
necessarily plunges through it, we have purchased a view onto an incremental
historical process that moves from the circumscribed realm of classically determi-
nate diatonic tonality into the symmetric chromatic universe of twelve equally
tempered tones. This opens a file drawer into which material will be inserted in the
coming chapters. In the final chapter of this book, I shall attempt to organize those
contents into a historical framework.

Weitzmann Regions without Sequences: Wagner and Strauss

The flat-terrain status of a Weitzmann region dictates that not all of its progres-
sions will be sequential or cyclic. One of many possible nonsequential paths is
presented in figure 4.18, a passage from act 1 of Parsifal. The passage is locally in b
minor, which in the larger context represents the minor dominant within a tonally

Figure 4.18. Wagner, Parsifal, act 1, mm. 404–13.

 CHAPTER 4 Weitzmann Regions  79

closed section in e minor. Beginning at m. 405, a locally subdominant e minor

spins into its Weitzmann region, sounding its minor triads in succession, followed
by its major ones. An augmented sixth is attached to the final G major and resolves
to the Fᅊ major dominant, reintegrating into the b minor orbit, preparing an
e minor cadence some seven measures later.

Figure 4.19. Strauss, “Frühling” (Four Last Songs), mm. 4–33.

80  Audacious Euphony

We conclude with an excerpt at the near end of the long nineteenth century,
composed nearly two centuries after C. P. E. Bach’s sonata. The opening stanza of
“Frühling,” from Richard Strauss’s Orchesterlieder (1949), is organized around a
Weitzmann region, although Strauss’s characteristically restless harmonic practice
requires the analyst to wrestle the interpretation to the ground a bit. Figure 4.19
provides a sparse score of the opening thirty-three measures, and figure 4.20 pro-
vides a series of analytical snapshots. The score is available at Web score 4.19 .
The alternation between c minor and aᅈ minor (considered earlier in connection
with figure 2.14) moves at m. 10 to the latter’s relative major, B major, combining
LP and R to form S, and exhausting half of the Weitzmann region (figure 4.20(a)).
After visiting several harmonies outside the region (A major and Bᅈ major,
not shown), the first stanza concludes with a melismatic setting of “Vogelsang”
(birdsong) that uses the region’s remaining components, cadencing on Eᅈ major
(figure 4.20(b)). Here the previous transformational progression is reversed:
a broad S-gesture, connecting e minor to Eᅈ major, is bisected by G major,
the former’s relative major, so that here S is formed by combining R with PL. In
figure 4.20, (a) and (b), summarizing the harmonic motion so far, make evident
one additional feature: the six bass pitches form a whole-tone collection (Kaplan
1994), permuting the whole-tone scale used by Schubert in his C major Symphony
and Octet (see figure 4.5).
The relative major now achieved, an orchestral interlude progresses to domi-
nant using a double-agent juxtaposition much like the opening of Chopin’s bᅈ
minor Nocturne (see figure 4.11(b)). After a return trip to tonic, Eᅈ major pro-
ceeds to its minor subdominant, at m. 24. The latter progresses immediately to the
cadential dominant, threading the seam and juxtaposing double agents Cᅈ and
B (figure 4.20(c)). The cadential dominant discharges onto C major, moving
into a different Weitzmann region for the start of the second verse at m. 29.
From a transformational standpoint, this second region is established in the same
way as its predecessor: with an S gesture connecting C major to cᅊ minor, routed
through the latter’s Relative major, E major (figure 4.20(d)).
Figure 4.21 summarizes the first thirty-three measures of “Frühling” on
the Tonnetz. The crooks overlaying the figure indicate the appearances of the
R + LP = S motive, as it is presented at the opening of the first stanza, retrograded

Figure 4.20. Sparse-score model of “Frühling”.

 CHAPTER 4 Weitzmann Regions  81

Figure 4.21. Tonnetz model of Frühling.

at the close of the first stanza, fragmented into its component transformations in
the interlude, and inverted at the opening of the second stanza. The figure presents
the image of a continuous cascade along the triads of the region structured
by GBEᅈ, including two enharmonic transformations, followed by a shift to the
region centered on CEAᅈ.
With this observation, we have ruptured the boundaries of the individual
Weitzmann region and started to explore the ways that adjacent regions connect
and collaborate. In doing so, we have trespassed into territory to be explored in
chapters 5 and 6.
This page intentionally left blank
C HA P T E R
Five
A Unified Model of Triadic Voice-Leading Space

How Hexatonic and Weitzmann Regions Interact

We have studied the internal structures of the hexatonic and Weitzmann regions
and characteristic nineteenth-century passages that remain within one of them.
The challenge now is to traverse the boundaries of the individual regions, in order
to view the larger triadic universe as a connected system under efficient voice
leading. The four hexatonic regions, and their constituent transformations, are
closed, mutually disconnected systems (in a mathematical sense, they are groups),
as are the four Weitzmann regions. On their own, the regions are static, like a
right and left leg hopping in place. In collaboration, Weitzmann and hexatonic
transformations gain the power of perambulation, unifying the triadic universe.
Producing a connected universe from its constituent parts makes good on the
aphorism “any chord can go to any other chord.” Yet it does so without unbridling
the twenty-four triads into an aleatoric torrent.1 The connected universe affords a
method of evaluating the voice-leading distance between any pair of consonant
triads, of recognizing patterns of motion, and of judging the coherence of progres-
sions on the basis of their voice-leading properties. Once we understand the
mechanisms of coordination, and the anatomy of the larger universe that ensues,
we will be in a position to explore progressions that are unidirectional, moving up
or down through registral space, or clockwise or counterclockwise through
pitch-class space.
The point of entry is a figure from Weitzmann’s Der übermässige Dreiklang,
presented as figure 3.9 and reproduced here with overlay as figure 5.1. Weitzmann
created this figure in order to portray the first- and second-order relations of
each of the four augmented triads. In the first tier of triads, highlighted in this
version of the figure, are those that can be reached by displacing one of the aug-
mented triad’s tones by one semitone; together, these six triads constitute its

1. The aphorism is attributed variously to Weitzmann, Liszt, and Reger; the metaphorical gumbo is
cooked to a recipe supplied by Harrison 1994, 1–7.

83
84  Audacious Euphony

Figure 5.1. Figure 3.9 with Weitzmann regions overlaid. Upper- and lower-case
letters are the roots of major and minor triads respectively.

Weitzmann region. The second tier contains those triads that can be reached by
displacing two of its tones in the same direction.
Although Weitzmann has nothing further to say about this figure, several
aspects of it merit our attention. First, those consonant triads that displace the
augmented triad downward are consistently positioned to its left, and conversely,
those that displace its tones upward are positioned to its right, conferring a direc-
tional consistency: within each quadrant of the figure, rightward motion in graphic
space is associated with upshifting in pitch space.2 Second, each cluster of triads
appears twice in the figure: those that appear to the right of a given augmented
triad are identical in content to those that appear to the left of its rightward neigh-
bor, and vice versa. Pruning the redundancies leaves only those twenty-four triads
highlighted by the overlay. Third, each group of consonant triads that separates
adjacent augmented ones forms a hexatonic region (although its six components
are not arranged into a cycle). Fourth, the entire figure is implicitly cyclic: the same
consonant triads appear at its left and right margins.
Taken together, these observations imply figure 5.2. Its cardinal points contain
the four augmented triads, each of which is connected to the six triads of its
Weitzmann region: minor triads to its clockwise side, major to its counterclock-
wise side. Left-to-right ordering in figure 5.1 corresponds to clockwise ordering in
figure 5.2. Intersecting the Weitzmann regions are the hexatonic regions, shown as
textured pools. The figure suggests that the Weitzmann and hexatonic regions are
in a figure–ground relation. For the hexatonic mariner, the Weitzmann regions
form bridges. For the Weitzmann landlubber, the hexatonic regions are rivers to
cross. To circumnavigate the entire complex, one need only locate an appropriately
amphibious vehicle.
Figure 5.3 gives structure to the hexatonic pools by connecting their constitu-
ent triads into the familiar hexatonic cycles. The cosmetic distortions of the hexa-
tonic cycles align modally matched triads on the same side of the figure, in
proximity to the augmented triad that they displace. Figure 5.3 is a version of Cube
Dance, a graph created by Jack Douthett in 1992 and published in Douthett and
Steinbach 1998. Cube Dance is a “true” model of voice-leading distance between
triads: “Every distance can be interpreted as representing voice-leading size”
(Tymoczko 2009b, 271).3 Directional motion on the graph correlates consistently

2. Upshift and downshift, from Lewin 1998, refer to directed voice leading in pitch or pitch-class space.
An ordered pair of chords upshifts if, in the transition from one to the other, more voices move up
than down. The same pair downshifts when the order is reversed. The terms may apply either to
register-specific pitch sets, such asᇳC4, E4, G4ᇴ→ᇳC4, F4, A4ᇴ , or to pitch-class sets under idealized
voice leading, such as C major → F major.
3. This observation first appeared in Tymoczko 2006, which positions Cube Dance as a contiguous
sector of the continuous space representing all three-chords. See also Tymoczko 2011b, chap. 3.
 CHAPTER 5 A Unified Model  85

Figure 5.2. Four Weitzmann water bugs in union with four hexatonic pools.

with melodic direction: clockwise and counterclockwise motion denote upshifting

and downshifting, respectively.
Since Cube Dance includes the four hexatonic cycles and the four Weitzmann
regions as contiguous subgraphs, it models all of the passages that are internal to
them, as studied in chapters 2 and 4, respectively. But it also supplies, for the first
time, a way to model progressions that move between regions, which is the usual
case. To get a preliminary sense of how Cube Dance models a composition
that crosses boundaries between regions, consider the Adagio opening of the
Overture to Schubert’s Die Zauberharfe from 1821. Schubert later used the same
composition as the Overture to Rosamunde, under which name it is normally per-
formed today. The score is available at Web score 5.4 . After an eight-measure
fanfare, the oboe sounds an antecedent phrase in c minor, followed by a conse-
quent in its relative major. After an Eᅈ major cadence, Schubert pulsates on this
chord and then drops G to Gᅈ and pulsates further on eᅈ minor. The bass now
slips down through D (forming a transient augmented triad) to Dᅈ, supporting a
cadential 64 that resolves in classical fashion to Gᅈ major. After a tonally closed
period in Gᅈ major, the same sequence of events occurs twice in transposition:
pulsations on Gᅈ major and fᅊ minor, transient motion through FACᅊ, cadence in
A major; pulsations on A major and a minor, transient motion through AᅈCE,
cadential 64 in C major. In the event, the latter initiates a deceptive motion to Aᅈ
major, and a standing-on-the-dominant, before proceeding with a C major Allegro.
86  Audacious Euphony

Figure 5.3. Jack Douthett’s Cube Dance. The four hexatonic cycles are portrayed
by the circuits of unbroken lines; the four Weitzmann regions by the broken-line
“water bugs.”

Figure 5.4 charts the path of prolonged harmonies on Cube Dance, together
with the transient augmented triads. The figure brings out the modulatory pro-
gression’s semitonal logic, its arc of motion away from and then back toward an
origin, and its constant downshifting.
Figure 5.5 shows the same passage on a Tonnetz that expands and unites the
segments presented in chapters 2 and 4. Like those earlier graphs, perfect fifths rise
to the right, major thirds rise toward the northeast, minor thirds “rise” toward the
southeast, and major and minor triads are represented, respectively, by up-tipped
and down-tipped triangles. The figure conjoins adjacent hexatonic systems at their
shared major third axes. Figure 5.6 presents a complementary conception of how
this Tonnetz is assembled, from the standpoint of the four Weitzmann regions.
If the major-third axes are interpreted as boundaries, then the Tonnetz is a
relation graph of the twenty-four triads under maximal pitch-class intersection
(adjacent triads share two tones). If interpreted as locations in their own right, as
in figure 4.6(c), each axis represents an augmented triad, and the Tonnetz becomes,
like Cube Dance, a model of the twenty-four consonant and four augmented triads
under single semitonal displacement.
 CHAPTER 5 A Unified Model  87

Figure 5.4. Schubert, Overture to Die Zauberharfe, opening measures, portrayed

on Cube Dance.

Figure 5.5. Schubert, Overture to Die Zauberharfe, portrayed on the Tonnetz.

88  Audacious Euphony

Figure 5.6. Connected Tonnetz assembled from the four Weitzmann regions.
Broken edges indicate where corresponding points are glued together.

Under this second interpretation, the Tonnetz shares the characteristic that
Tymoczko (2009c) demonstrated for Cube Dance: “Every distance can be inter-
preted as representing voice leading size.” Unlike Cube Dance, the Tonnetz suffers
from a Bering Strait distortion, and in two distinct dimensions: locations at the left
side of the graph replicate those at the right, and locations at the bottom of the
graph replicate those at the top. Were we freed from the dimensional constraints
of the printed page, these distortions could be corrected by gluing together
pitch-class identities in one dimension to create a cylinder and then in a second
dimension to wrap the cylinder into the shape of a doughnut, or what topologists
call a torus. Although we can avoid these distortions by reverting to Cube Dance,
we would lose some analytically valuable information and historical connections,
as discussed in chapter 4.
The progression from Schubert’s overture is depicted in figure 5.5 as a direct
motion along the main diagonal. As such, it resembles the hexatonic progressions
that move along the opposite diagonal. This resemblance, however, occludes a
distinction emphasized in chapter 2. Vectors that parallel the major-third axis
preserve a constant center of balance, as upshifting and downshifting voices cancel
out. Any vector orthogonal to that axis represents voice leading that either
cascades or escalates. One might imagine figure 5.5 as a staircase seen from far
above. Each hexatonic strip is a tread, and each augmented triad is a riser. From
our distant vantage point, step height is equalized, and we must imagine which
directions correspond to up and down.
The downward direction of the vector in figure 5.5 is consistent with the down-
ward direction of Schubert’s voice leading. The rightward direction of that vector,
however, is problematic from that same standpoint. Passage of time is associated
with left-to-right motion by virtue of European orthographic convention, with
clockwise motion by virtue of clock conventions, and with upward motion by
virtue of organismal physiology. As a result of cognitive blending beneath the
 CHAPTER 5 A Unified Model  89

threshold of awareness, upshifting in pitch space is spontaneously identified with

rightward motion across the page, as in figure 5.1, and with clockwise motion
about a circle, as in Cube Dance. The Tonnetz confounds those intuitions, by
identifying rightward motion with downshifting in pitch or pitch-class space. To
reengage those default intuitions, it is tempting to recast the Tonnetz by inverting
it around its vertical axis. I resist this temptation in order to retain the orientation
familiar to veteran Tonnetz surfers. It will also serve the project of the final
chapters, when historical issues come to the fore.4
Cube Dance and the Tonnetz present us with two essentially equivalent modes
for representing triadic progressions from the standpoint of their voice-leading
properties. Each graph has its own heuristic advantages. The Tonnetz is more com-
pact, shows individual pitch classes rather than fusing them into triads, preserves
connections with historical modes of thought, is more amenable to tracing
common tones, and consistently correlates planar directions with transformations.
Cube Dance is a more direct model of true voice-leading distance, both because it
does not require an exercise of imagination across the Bering Strait and because it
gives the augmented triads explicit locations. We shall move back and forth
between these two models in response to the characteristics of the particular
composition at hand.

Chromatic Sequences

In chapters 2 and 4 we studied sequences that transpose by a series of major thirds.

These sequences exhibit balanced voice leading and are internal to a hexatonic or
Weitzmann region. We are now in a position to study those sequences that move
parsimoniously between adjacent regions, in some cases circumnavigating the
triadic universe. The transpositional values that generate those sequences repre-
sent an odd number of semitones. Transposition by semitone (T1, T11) engages all
three voices in parallel motion. The remaining transpositions (T3, T5, T7, and T9)
feature two voices moving in similar motion, one by semitone and the other by
whole step. Thus, all odd-value transpositions involve three units of semitonal
work, ranking them next on the scale of parsimony, after the hexatonic and
Weitzmann transformations already studied.
In incrementing from two to three units of voice-leading work, we have crossed
over a threshold. Just as a three-semitone motion in a single voice is characteristi-
cally conceived as a leap that implicitly combines and elides across two distinct

4. The historical orientation arises because the designers of early Tonnetze were concerned with acous-
tic ratios rather than melodic motion. Euler, Oettingen, and Riemann conceived C → B as a rising
major seventh, conjoining a 3:2 fifth and a 5:4 major third, and it is this upward motion that projects
from left to right on their Tonnetze, in accordance with European orthographics and organicist/
teleological conditioning. Interpreting C → B as a melodic motion reverses the directional flow, so
that upshifting now travels the Tonnetz from right to left. Incidentally, this suggests that these his-
torical theorists thought of voice leading in idealized terms, rather than in terms of the pitch-class
paths emphasized in Tymoczko’s work.
90  Audacious Euphony

steps, so we will conceive of three units of semitonal work as implicitly combining

and eliding across two distinct parsimonious transformations. The hexatonic and
Weitzmann regions furnish a roster of such transformations. Any pair of hexa-
tonic transformations, such as L with P, preserves hexatonic region, and similarly,
any pair of Weitzmann transformations, such as N with R, preserves Weitzmann
region. Motion between adjacent regions combines one transformation each from
the hexatonic and the Weitzmann group of transformations, such as L with R.
To pursue the exploration of these combinations, we define hexatonic-group
(or H-group) transformations as those that preserve hexatonic region but shift
Weitzmann region; conversely, a Weitzmann-group (W-group) transformation
preserves Weitzmann region and shifts hexatonic region. An odd-valued transpo-
sitional sequence is generated by alternating two transformations, one from each
group. As there are three hexatonic-group transformations (L, P, and H) and
three Weitzmann-group transformation (R, N, and S), there are 3 × 3 = 9 combi-
nations that generate an odd-valued chromatic sequence through parsimonious
voice leading.
Table 5.1 presents these nine combinations. The numbers indicate the interval
of transposition, in chromatic semitones (up or down, depending on the order
of the operations and the mode of the initial triad). Two transformations from
each group lie within the norms of classical syntax: L and P from the hexatonic
group, and R and N from the Weitzmann group. It is as a result of their cross-
pollination, L or P combined with R or N, that the earliest odd-transposition
sequences arise in the eighteenth century. These 2 × 2 = 4 combinations are
more darkly shaded in table 5.1. The remaining five combinations involve a
maverick transformation, S or H. The three gray-shaded ones arise in music of
Schubert and Liszt, to be documented below; I have yet to discover examples
of the remaining two.

Minor-third (T±3) sequences

The minor third transpositions in Die Zauberharfe result from the alternation of R
(from the Weitzmann group) and P (from the hexatonic group). When modeled

Table 5.1. Combination table of W- and H-group operations

R N S
L T±5 T±1 T±3
P T±3 T±5 T±1
H T±1 T±3 T±5
Each number indicates the number of semitones that a triad is transposed when the H-group operation
at the head of its row is combined with the W-group operation at the head of its column. The transpo-
sition may be up or down, depending on the order of the operations and on whether the initial triad is
major or minor.
 CHAPTER 5 A Unified Model  91

Figure 5.7. Schubert, Drei Klavierstücke, D. 946, no. 2, mm. 46–52.

on the Tonnetz (figure 5.5), the bounding axes are generated by minor thirds,
forming diminished seventh chords that combine to form an octatonic collection.
The P/R cycle thus can be considered as a primordial triadic generator for octa-
tonic systems, analogous to the role of L and P with respect to a hexatonic
system.5
A minor-third sequence is also formed by two other combinations of H- and
W-group operations: N with H, and S with L. Since both pairings involve one rare
operation, they occur less frequently. Figure 5.7 illustrates a T3 sequence formed
by an N/H chain, from a late Schubert piece published posthumously as the second
of the Drei Klavierstücke. In this passage, the classically normative N transforma-
tion occurs within each segment, and the maverick H occurs across phrasing
boundaries. In later compositions, these functions are swapped: Liszt’s Malediction
for piano and string orchestra (seven measures after rehearsal B), Rimsky-
Korsakov’s orchestral tone poem Skazka (see Taruskin 1996, I: 271), and Bruckner’s
choral motet “Ecce Sacerdos” (from rehearsal 2) all feature H within phrases, and
N between them.
The alternation of S and L generates an extended downshift sequence
(offset by upward registral transfers) in Liszt’s Grande fantaisie symphonique
for piano and orchestra (“Lélio Fantasy,” 1834; figure 5.8). The opening of the
Sanctus from Schubert’s Aᅈ major Mass follows a similar progression in the
upshifting direction (with falling bass), although the insertion of a back-relating
dominant after S makes the voice leading less efficient than in the Liszt passage
presented here.

5. This generative parallelism is but one reflector of a broader structural parallelism that relates
hexatonic and octatonic organizations of chromatic space. See Cohn 1991 and 1997, Lerdahl 2001,
258, and Tymoczko 2011b, 125. Siciliano 2005b and Goldenberg 2007 present further examples of
octatonic R/P chains, in Schubert and Smetana, respectively.
92  Audacious Euphony

Figure 5.8. Liszt, Grande fantaisie symphonique über Themen aus Berlioz’ “Lélio,”
mm. 195–202.

Perfect-fifth (T±5) sequences

Chromatic sequences that modulate by perfect fourth or fifth become common

quite early, due to the classically privileged position of that interval. Lowinsky
1989 [1967] calls attention to a number of hexachord fantasias that upshift by per-
fect fourth, the earliest of which originate in Elizabethan-era England. The most
common T5 sequence results from the alternation of L with R. It is found in a
circle diagram of Werckmeister (1698), where it served as a map of modulations
between keys (tonics) rather than direct progressions between chords. It was
recognized as a more local harmonic map by Georg Vogler in 1778 (Brower 2008,
99n41) and Honoré de Langlé in 1797 (Damschroder 2008, 80). Carl Stein
recommended that piano students learn to play the entire twenty-four-triad cycle,
“as it forms the groundwork on which may be constructed an almost infinite number
of passages and variations,” noting the progression’s symmetry and its “double
union of intervals—two of them always remaining undisturbed” (1888, 38).6
Eventually, the L/R chain became the central component of Moritz Hauptmann’s
influential conception of tonal space (1888 [1853]).
Although the cycle is too long to be circumnavigated in a single gesture, a
remarkable passage from the Scherzo of Beethoven’s Ninth Symphony comes close.
A reduced score is available at Web score 5.9 . The progression begins on
C major and traverses nineteen triads before terminating at A major, completing
more than two tours of Cube Dance (Cohn 1991, 1992, 1997). On the Tonnetz,
it requires more space than can legibly fit on a page, moving from right to
left along the horizontal axis. Figure 5.9 presents an example of more modest
dimensions and more leisurely tempo, from the first movement of Brahms’s
D major Symphony, Op. 73. The passage presents the liquidating consequent of
the movement’s fᅊ minor theme, in preparation for the arrival of the dominant

6. Cohn 1997 attributes the quote directly to Logier 1827, but it was introduced by Stein in his 1888
adaptation of Logier’s text for an American readership.
 CHAPTER 5 A Unified Model  93

Figure 5.9. Brahms, Symphony no. 2, 1st mvt., mm. 102–18.

(mm. 106–18). It exits the fᅊ minor diatonic system at Gᅉ. The subsequent Cᅉ and
Fᅉ bring us into another realm entirely, as chromatic alterations of root and fifth.
By this point, we are in free fall in the bass, and “free rise” in the upper voices,
without any way of predicting the duration of this sequential process.
Figure 5.10 depicts this progression on the Tonnetz as a leftward drift (see also
Web animation 5.10 ). The sequence terminates at its ninth station, d minor.
Figure 5.11 traces the same passage on Cube Dance, starting and ending at 4:00
and traveling clockwise. The representation suggests that d minor brings a mea-
sure of closure to an R/L chain initiated from fᅊ minor, by virtue of the return to
the original radius.
A second way to effect a sequence by perfect fourths is through the alternation
of N and P. Figure 5.12 indicates that the progression is equally effective in both
directions. For Daniel Harrison (1994, 33–34), this pair of progressions typifies a
set of dualisms that includes major/minor, plagal/authentic, dominant/subdomi-
nant, and upward/downward leading-tone energy. Thus, in the authentic direction
of progression (a), minor tonics are preceded by their major dominants, whose
agents discharge upward; in the plagal direction of progression (b), major tonics
are preceded by their minor subdominants, whose agents discharge downward.

Figure 5.10. Tonnetz model of figure 5.9. An animation, with recorded performance,
is at Web animation 5.10.
94  Audacious Euphony

Figure 5.11. Cube Dance model of Brahms excerpt (figure 5.9).

Both directions are well represented in the literature; Harrison points to passages
from J. S. Bach’s g minor Organ Fantasy and Schubert’s d minor String Quartet as
exemplary.

Semitone (T±1) sequences

The most common, and chronologically earliest, option involves the alternation of
N and L. The ascending progression at figure 5.13(a) appears more frequently; one

Figure 5.12. T5 and T7 sequences generated by N/P alternation.

 CHAPTER 5 A Unified Model  95

Figure 5.13. T1 and T11 sequences generated by N/L alternation.

example, from among many, is discussed below (see figure 5.21). Although the
descending progression is less frequent, it was used by Rameau in one of the most
famously expressive passages in the history of pre-Classical opera (Rehding
2005).7
More rare is the progression at figure 5.14, excerpted from Liszt’s Il Penseroso,
which connects Gᅈ major to D major through a series of semitonal descents, first
lowering the third of each major triad and then lowering the root and fifth of each
minor triad. Like the Brahms passage illustrated in figure 5.11, the entire S/P
chain executes a complete tour around Cube Dance, elliptically connecting two
PL-related harmonies that stand directly adjacent in the same hexatonic and
Weitzmann region.8

Transformational Substitutions

A composer who pursues a chromatic sequence beyond several stages, and lacks a
programmatic motivation for doing so, risks being perceived as mechanical, lack-
ing invention, and so forth. Some of the excerpts that have arisen in this exposition
preserve their stake to more aesthetically positive qualities by varying the pace

Figure 5.14. T11 sequence generated by P/S alternation in Liszt’s Il Penseroso,

mm. 17–20.

7. Langlé 1797, 82–83, includes synthetic models of both progressions among his tours de l’harmonie.
Ahn 2003, 105–7, shows that both directions are represented in Liszt’s “Wilde Jagd” (Études d’exécution
transcendante, 1852).
8. Gollin 2000, 306, identifies a similar progression in Prokofiev’s War and Peace.
96  Audacious Euphony

with which the stations of the sequence are delivered. But a composer who wishes
to circumnavigate the pan-triadic universe can introduce variation in the progres-
sion of triads, breaking the sequential pattern while maintaining the voice-leading
trajectory and its pacing. This can be done by substituting, for any triad in a
sequence, one of its major-third transpositions.9 Describing the same phenome-
non in transformational terms, we can say that transformations within the same
group can freely substitute for one another without perturbing the voice-leading
trajectory. Represented on Cube Dance, the directional trajectory continues but
traverses different nodes. Represented on the Tonnetz, vectors maintain their
angle, but crooks displace their position on the plane.
Beyond introducing variety to the transformational and transpositional pal-
ette, these substitutions can also provide a shortcut to a tonal goal that would be
remote were the sequence to strictly run its course. If an H/W-group pairing
induces a chromatic sequence that transposes by semitone or fifth, then the triadic
chain must traverse all twenty-four triads before closing back to its triad of origin.
Major-third substitution allows a goal triad to move forward in the chain by eight
or sixteen stations, expediting its arrival without interrupting or reversing the
voice-leading trajectory of the passage.
Figure 5.15 presents an excerpt from the Die Erlöseten des Herrn fugue, from
Brahms’s German Requiem, as a clear introduction to transformational substitu-
tions. The initial fourfold setting of “wird weg” progresses from bᅈ minor to B
major (an enharmonic proxy for Cᅈ major) through an upshifting L/R chain, each
chord progressing to its diatonic submediant. At “müssen,” N substitutes for R,
taking B major to its minor subdominant, e minor, rather than its submediant,
gᅊ minor. The L/R chain immediately resumes, carried in the voices by a stretto of the
fugal incipit, and is carried through six further stations, terminating at Bᅈ major,
the movement’s tonic. The connection from bᅈ minor to Bᅈ major, through an
unimpeded L/R chain, would require seventeen stations; the N-for-R substitution
allows the target to arrive after nine moves.
Figure 5.16 depicts the progression as a leftward motion on the Tonnetz, inter-
rupted by a diagonal jog that transfers files, marking the substitution of e minor for
gᅊ minor, and N for R (see also Web animation 5.16 ).10 The fault lines, at a 45°
angle, show the orthogonal cut against the major-third regions, upshifting across
all four major third alleys and restoring the original.
In Chopin’s g minor Ballade, Op. 23 (1835), a mediant chain runs in the oppo-
site direction and requires two N-for-R substitutions to achieve tonal closure. The
extended Eᅈ major prolongation in the middle of the Ballade is dominated by the
earworm for which the composition is famous, sounded in Eᅈ major at m. 68, in
the tritone-related A major at m. 106, and again in Eᅈ major at m. 167. In the
abstract, an antipodal progression of this type can be conceived as a reversal, or as
the continuation of a trajectory that leads around the back side of the chromatic

9. For further discussion and examples, see Cohn 1998b. Tymoczko 2011b, 87–88, 282–83, describes
the same phenomenon using a different geometric layout, which resembles Cube Dance but has an
additional virtue: by projecting into a third dimension, Tymoczko’s continuous space consistently
affiliates root interval with direction along a single axis.
10. For a different view of this passage, see Brown, Dempster, and Headlam 1997, 173–75.
 CHAPTER 5 A Unified Model  97

Figure 5.15. Brahms, Ein deutsches Requiem, 2nd mvt., mm. 261–71.

universe, executing a two-stage cycle of tritone transpositions. The latter concep-

tion is more pertinent in this case: although not transpositionally related, the two
tritone modulations downshift along parallel tonal trajectories. Initial Eᅈ major
proceeds through a downshifting L/R chain, each triad progressing to its diatonic
mediant (mm. 90–93). The chain’s terminus, d minor, is the minor subdominant
of A major, which it reaches after an extended dominant. A similar L/R chain
connects A major to gᅊ minor (mm. 115–24), although its internal terms are

Figure 5.16. Tonnetz model of figure 5.15. An animation, with recorded

performance, is at Web animation 5.16.
98  Audacious Euphony

Figure 5.17. Tonnetz model of Chopin Ballade, Op. 23, mm. 68–167.

permuted, with E major preceding cᅊ minor rather than following it. gᅊ minor is
accorded an under-seventh at m. 124, attaining subdominant function in Eᅈ major,
to which it leads after a prolonged Bᅈ dominant. Figure 5.17 sketches the progres-
sion as a rightward motion on the Tonnetz but with two northeast-jogging N-for-R
substitutions, the first taking d minor to A major rather than the expected F major,
and the second performing the same function at a tritone transposition. The two
file transfers move the tonic forward by sixteen stations in the L/R chain while
preserving its overall downshift trajectory.11
N-for-R substitutions also guide the opening of the development from
Bruckner’s Third Symphony, which connects f minor to a minor through two
transpositionally related phrases, each modulating up a whole step. Each phrase
begins with a slowly evolving series of soundsheet arpeggios that upshift across
four stations of an L/R chain, and ends with an N substitution that carries a major
dominant to a minor tonic. Described in this way, the transformational progres-
sion should be identical to that of the Chopin Ballade (although retrograded on
the Tonnetz, since it begins with a minor rather than major triad). Yet the phrases
are transpositionally related by whole step rather than tritone, because a PL trans-
formation (outfitted as motion from a major tonic to its German sixth) is interpo-
lated prior to the N substitution. As shown in figure 5.18, PL’s balanced voice
leading causes the progression to linger within a Weitzmann region (Gᅈ major →
D major → g minor), putting the upshifting juggernaut on pause while a second file
transfer is executed, to supplement the one provided by the N substitution. As with
the Chopin Ballade, the combined phrases execute a complete traversal of the four
major-third alleys (this time moving from right to left).
A fourth passage based on an L/R chain, from Chopin’s f minor Fantasy,
Op. 49, presents a variation on the phenomenon we have been studying. A reduc-
tion is given as figure 5.19; the score is available at Web score 5.19 . The transi-
tion from the opening Lento to the agitato core arpeggiates upward through a
series of overtone-distributed triads whose highest pitches are circled by inverse
cambiatas. An antecedent phrase downshifts through six links of an L/R chain,
connecting f minor to Bᅈ major. The goal triad acquires a seventh and resolves
across the fermata to its nebenverwandt, eᅈ minor, from which a consequent phrase

11. Schenker 1954 [1906], 300, ventures a similar reading. For discussion and an alternative reading,
see Brown, Dempster, and Headlam 1997, 167–69.
 CHAPTER 5 A Unified Model  99

Figure 5.18. Tonnetz model of Bruckner, Symphony no. 3, 1st mvt., opening of the
development.

L/R-chains back to tonic f minor. Like the previous two passages, N is interpolated
within an L/R chain. What is different about this passage is that the interpolated
N functions not to continue the downshift trajectory but rather to temporarily
reverse it: Bᅈ major → eᅈ minor is an upshift amidst a cascade of downshifts. This
circumstance results from N’s interpolation not between two H-group progres-
sions, as in the previous examples, but rather between two operations from its own
W-group. Consequently, the rapid turnover of H- and W-regions, characteristic of
all of the chains we are studying in this section, is slowed by a momentary tarrying
within a single W-region, via the R/N-chain segment that connects g minor to
Gᅈ major.
Figure 5.20 represents the passage on a Tonnetz (see also Web animation
5.20 ). As in the previous three examples, there is a horizontal migration by
five alleys, interrupted by a file transfer that expedites the return of the tonic. What
is new here is the sharp angle of the jog, representing a temporary retreat before
further advance. It is also worth noting here, anticipating a topic of chapter 6, that
the internal six triads of the progression constitute and exhaust a pitch retention
neighborhood, represented by the hexagon about Bᅈ.

Figure 5.19. Chopin, Fantasie, Op. 49, mm. 43–68.

100  Audacious Euphony

Figure 5.20. Tonnetz model of figure 5.19. An animation, with recorded

performance, is at Web animation 5.20.

One final example, from Liszt’s Grande fantaisie symphonique for piano and
orchestra (“Lélio Fantasy”) of 1834, demonstrates that H/W substitutions are not
limited to L/R chains.12 Measures 379–429 feature an oboe melody that is trans-
posed successively upward by semitone. Figure 5.21 presents a synopsis of the
underlying harmonies, which execute a segment of an N/L chain, transposing
from Dᅈ major to E major. The latter’s nebenverwandt, a minor, proceeds not to the
projected F major via L but rather to C major via R. As with the Chopin Fantasy,
this amounts to a temporary retreat, briefly colonizing a Weitzmann region. From
C major, the N/L chain now proceeds on its course, arriving not at the projected
F major but rather on the tonic Dᅈ major. Figure 5.22 summarizes the progression
on the Tonnetz.
Having examined these five passages, this is a good moment to stand back and
evaluate the Tonnetz as a mode of representation. Its directional vectors reflect the
tight patterning of chromatic sequence, as well as the disruption of those sequences
when substitutions occur. The superposition of the 45° fault lines encourages the
tracking of voice-leading trajectories. But a limitation of the Tonnetz is beginning
to come into view. The Tonnetz makes a two-dimensional presentation of an
underlying phenomenon that, under certain conditions, inhabits three or more
dimensions. Those conditions are met when we accept enharmonic equivalence,
as we must in the case of chromatic sequences. At that point, the Tonnetz disguises
cyclic closure. Each of the four passages studied in this section proceeds through
exactly five hexatonic regions, each represented by an augmented-triad-bounded
alley, and exactly five Weitzmann regions, each represented by an augmented-triad

Figure 5.21. Liszt, Grande fantaisie symphonique über Themen aus Berlioz’ “Lélio,”
mm. 379–439.

12. See also Brahms’s “Walpurgisnacht.”

 CHAPTER 5 A Unified Model  101

Figure 5.22. Tonnetz model of figure 5.21.

stalk. Each type of region is represented by four species, one per augmented triad,
and so a progression through five regions closes a cycle. But this cyclic closure is
not represented on the Tonnetz.
This is the domain where Cube Dance excels. Figure 5.23 gives a Cube Dance
representation of the passage from the Liszt Fantasy. The graph shows the passage’s
consistent clockwise = upshift trajectory, the temporary downshift to C major, and
the recontinuation as C major progresses toward cyclic closure at the return of Dᅈ
major. What the graph fails to honor is the sequential nature of the upshift pro-
gression. The progression could have selected one triad at random from each trio
of T4-related triads that constitute the intersection of an H- and W-region; the
cyclic progression would have looked just as orderly. This suggests that Cube
Dance, with its cyclic explicitness, is superior to the Tonnetz, to the extent that our
interest focuses on tracking the voice-leading trajectory of a triadic passage, at the
expense of the particular progression of triads that realizes that trajectory. The
decision between the two modes of representation is analogous to the more famil-
iar one of choosing to track a pitch succession in pitch or pitch-class space. The
former tracks register, the particular choice of an octave location for the presenta-
tion of a pitch class, whereas the latter makes cyclic positioning explicit. Both have
their uses for particular purposes, and one would not wish to dispense with either
of them.
The remaining work of this chapter, and that of chapter 6, focuses more on
voice-leading trajectories, cyclic closure, and other sorts of patterns that occur
against a circularized enharmonic backdrop and less on the particular triads that
are selected to carry out these trajectories and patterns. Our first task will be to
retune our theoretical and representational apparatus, to make it as sensitive as
possible to these new considerations as they come to the fore.
102  Audacious Euphony

Figure 5.23. Cube Dance model of figure 5.21.

Voice-Leading Zones

I now make explicit two intertwined ideas that have floated close to the surface in
the analytical work of the previous section. The first idea is about triadic objects:
underlying the claim that T4-related triads can freely substitute for each other in
the fulfillment of a voice-leading trajectory is the idea that these triads are equiva-
lent, that they belong to the same equivalence class of objects.13 The second, cor-
responding idea is about actions on those objects: underlying the claim that the
three members of the hexatonic group of transformations, and correspondingly,
the members of the Weitzmann group, can freely substitute for each other in the
fulfillment of a voice-leading trajectory is the idea that these transformations are
equivalent, that they too belong to the same equivalence class of transformations.
It is important to proceed very carefully here, so as not to convey the wrong
idea. When we say that two objects or transformations are equivalent, we are

13. The equivalence of T4-related triads was first advanced by the French theorist Camille Durutte in
his Esthétique musicale: Technie ou lois générales du système harmonique of 1855, whose relevance
to this work was brought to my attention by Levenberg 2008. The same triplets arise in Carol
Krumhansl’s research (1990, 43), where their clustering arises through dimensional scaling of data
concerning how listeners assess proximity between keys. See Krumhansl 1998, 257, for a discussion
of how those data might apply to triads.
 CHAPTER 5 A Unified Model  103

saying that they are so with respect to some well-defined context, not with respect
to every conceivable context. For example, we might say that all of the even num-
bers are equivalent with respect to their ability to divide integrally in half. This is
not to say that I would be just as happy with two dollars as with two million, or
would walk just as well with ten legs as with two. Analogously, when we say that
C major and E major are equivalent, we are saying that they are so with respect
their capacity to function in a certain voice-leading context, not that they are
equally good final chords for the Jupiter Symphony.
Recognition of the equivalence status of these relations leads naturally to a
model of voice-leading zones,14 equivalence classes of trichords with the capacity
to share a center of balance, in the sense that the latter term was cultivated in
chapter 2. Consonant triads share a voice-leading zone if they are transpositionally
related by major third. On the Tonnetz, zone-equivalent triads occupy the same
hexatonic strip, and their triangles are equivalently oriented (up-tipped or down-
tipped). On Cube Dance, zone-equivalent triads are positioned on the same radius.
The twenty-four consonant triads thus partition into eight zones, coextensive with
the “triplets” referenced at the beginning of chapter 4. The augmented triads
occupy the remaining four zones. Thus, the objects on Cube Dance occupy twelve
distinct voice-leading zones.
The number twelve is fortuitous, tapping as it does into well-consolidated intu-
itions about the reckoning of time. We can take advantage of these intuitions by
labeling the twelve voice-leading zones as the stations of the clock. Figure 5.24
does so, by superimposing a clock face over Cube Dance. The clock face is altered
in several insignificant ways. Adopting a convention familiar to music theorists,
zero substitutes for twelve at the top of the figure. Multiples of three, adjacent to
the augmented triads, are presented in a special typeface. And all twelve numbers
are underlined, a convention that distinguishes zone labels from numbers as they
might be used for other purposes in the exposition.
Naming is its own virtue, since it allows for unambiguous reference. But the
assignment of particular labels to particular zones here is not arbitrary; rather, it
identifies several ways that the structure of the zones is isomorphic to the structure
of the numbers modulo 12 (and, transitively, to the structure of the pitch classes).
Those isomorphisms will be useful, to the extent that they encourage certain
realizations about structure to emerge as by-products of our intuitions about num-
bers (or the pitch class universe). The assignment of these particular numbers to
these particular sets of triads is useful in two different ways, the first of which is
self-evident, the second perhaps less intuitive on first encounter.
First, with a single exception, assigning numbers to voice-leading zones allows
voice-leading distance to be modeled as subtraction, modulo 12.15 Consider, for
example, the voice-leading distance between d minor and G major, located in
zones 4 and 8, respectively. Identifying their zones is sufficient to identify the
voice-leading distance between them, by subtracting modulo 12 and grabbing the

14. Cohn 1998b develops the same model under the term sum classes.
15. The exception is the hexatonic pole. Its maverick status is related to its contrary motion, a property
that it alone possesses, among relations between triads in different zones.
104  Audacious Euphony

Figure 5.24. The twelve voice-leading zones, depicted on a clock face. Each zone
contains the triad(s) on its radius. The pitch classes of each triad sum, modulo 12,
to the number indicated on its radius.

absolute value. As |8 – 4| = |4 – 8| = 4, modulo 12, motion between the two triads,

in either direction, requires four units of voice-leading work. We can verify this
easily at a piano, or on a staff, or by counting edges on Cube Dance. But the iso-
morphism of voice-leading zones and the numbers modulo 12, together with our
intuitions about the structure of numbers, allows us to short-circuit all of these
procedures. Selecting now any of the eight remaining pairs of triads from these
same voice-leading zones, we know that they are separated by four voice-leading
units. What is proper to one pair is ipso facto true of the other eight. Behold the
beautiful cognitive economy of equivalence!
Second, and perhaps counterintuitively, the number assigned to each triad is
the sum of its pitch classes, modulo 12 (Cohn 1998b; Tymoczko 2011b, 89).16
Knowing this allows us to assign zone labels to triads without looking them up on
a table or a diagram. For example, d minor ({DFA} = {2, 5, 9}) belongs to zone 4 by
virtue of 2 + 5 + 9 = 16 = 4modulo 12; similarly, G major ({GBD} = {7, 11, 2}) belongs

16. Pitch-class sums were first explored, in a completely different context, by Babbitt 2003, 299 (initially
published 1972). Their pertinence here was first suggested to me by Jack Douthett.
 CHAPTER 5 A Unified Model  105

to zone 8 by virtue of 7 + 11 + 2 = 20 = 8modulo 12. Readers who worry that these

sums require tedious calculation or brute-force memorization will be gratified to
learn that they can be determined in a musically engaging way with low cognitive
overhead. Mentally fix the C augmented triad at 0, piggybacking on the C = 0
convention of atonal theory. Then secure the remaining augmented triads, in
ascending order, to the ascending multiples of three. Each of the remaining eight
zones is now adjacent to one of these four cardinal points. To ascertain a particular
consonant triad’s zone, one need only determine which augmented triad it dis-
places, and in which direction. If you have made it this far in this book, you should
by now be familiar with thinking of consonant triads in this way. D minor is at zone
4 because it upshifts FACᅊ at zone 3; G major is at zone 8 because it downshifts
GBEᅈ at position 9.
Having established a system of classifying and labeling consonant triads on the
basis of their positions within a voice-leading system, the next step is to establish
corresponding equivalences for the transformations that connect them. We
define three such transformation classes: an H class, consisting of the three mem-
bers of the hexatonic group (L, P, and H); a W class, consisting of the three
members of the Weitzmann group (R, N, and S); and an E class, consisting of
three transformations that map triads within their own zone (LP, PL, and identity
operation E).17 An H-class operation maps a consonant triad in zone X into the
unique zone X ± 1, and a W-class operation maps a consonant triad in zone Y
H
into the unique zone Y ± 2. Thus, we write 1 ← → 2 to summarize a range of
propositions that include the following:

• To get from a triad in zone 1 to some triad in zone 2 requires some H-group
transformation.
• To get from a triad in zone 2 to some triad in zone 1 requires some H-group
transformation.
• Any H-group transformation applied to a triad in zone 1 produces a triad
in zone 2. And
• Any H-group transformation applied to a triad in zone 2 produces a triad
in zone 1.
W
Similarly, we interpret 2 ← → 4 as making corresponding claims about the role
of the W-group transformations, in moving between triads in zones 2 and 4, and
so forth.
With the H- and W-group transformation classes as basic elements, we can
establish classes of compound transformations that connect triads from nonadja-
cent zones. Triads that are transpositionally related by an odd value, three voice-
leading units apart, are connected by a compound HW or WH transformation; in
effect, this is the generalization that underlies all of the work with sequences and

17. H and W are a species of exchange operations, which are introduced in Lewin 1987, appendix B,
and developed in Lewin 1995 and Cook 2001.
106  Audacious Euphony

transformational substitutions in this chapter. Triads that are separated by four

units of work, as in the {d minor, G major} example considered above, are con-
nected by HWH; by five units of work, by WHW; and finally, triads that are
the maximum of six units apart (transposition by major second or tritone) are
connected either by HWHW or by its inverse, WHWH.
The final sentence of the preceding paragraph consolidates a large number of
specific instances: the three-term compounds cover twenty-seven specific sets of
transformations, and the four-term compounds cover eighty-one. Thus the gener-
alizing power is along the lines of a claim about numbers such as odd + odd =
even, which consolidates in one statement an infinite variety of propositions
covered by it. For readers mystified by this mode of discourse, be assured that the
remainder of the book proceeds independently of these abstractions.

Remarks on Disjunction and Entropy

The model proposed here is premised on the proposition that conjunct voice lead-
ing is privileged in nineteenth-century music. Yet such a proposition is risky, if
assigned universal value. The inversion of asymmetric (“violent”) binaries is
among the cardinal features of European society and culture after the French revo-
lution. If nineteenth-century composers accorded privilege to smooth voice lead-
ing, they also accorded privilege to the contravention of privilege. One need not
search far for examples. Since the Carolingian era, the asymmetry of consonance
and dissonance constituted an absolute hierarchy; as we have seen in connection
with the opening of the Faust Symphony (figure 3.4), by the middle of the nine-
teenth century it became possible to stabilize a dissonant chord and destabilize its
consonant neighbors. In classical tonality, a dissonant chord requires resolution to
a key; in Tristan und Isolde, “the various would-be ‘keys’ come to sound . . . as an
agglomeration of phenomena accessory to the Chord-as-Ding-an-sich” (Lewin
2006, 220). Since the sixteenth century, the chromatic is an ornament to the
diatonic; in the nineteenth century, the diatonic often becomes a subset of the
chromatic. It would be naive to fantasize that the privileges accorded to melodic
conjunction are immune from analogous reversals.
Moreover, one of the central tropes of Romantic aesthetics is a fascination with
the remote, unattainable, and inexplicable. Already quite early in the century,
composers were moved to find ways to depict these tropes through music (Kramer
1994; Hoeckner 1997). A late manifestation of this concern is revealed in a letter
of Gustav Mahler, concerning the C major finale of his First Symphony:

Again and again, the music had fallen from brief glimpses of light into the darkest
depths of despair. Now, an enduring, triumphal victory had to be won. As I discov-
ered after considerable vain groping, this could be achieved by modulating from
one key to the key a whole tone above (from C major to D major, the principal key
of the movement). . . . My D chord . . . had to sound as though it had fallen from
heaven, as though it had come from another world. (qtd. in Buhler 1996, 127)
 CHAPTER 5 A Unified Model  107

The passage turns our initial assumption on its head. Mahler was privileging not
conjunction but its opposite.
Fortunately, a pan-triadic model that privileges conjunct voice leading brings
with it, as a residual benefit, the ability to model disjunction as well. When we were
exploring the hexatonic and Weitzmann regions as autonomous entities, the
best emblems of disjunction were the maverick transformations H and S, which
represent diametric opposition within the normative hexatonic and N/R cycles.
With the drawing together of these components in this chapter, other possibilities
emerge.
Imagine that you are a triadic composer who wishes to jump as far away from
C major as possible in a single stroke. But, unlike Mahler, you have at your disposal
a model of the triadic universe as a metric space under voice-leading proximity, in
the form of figure 5.24. Locate C major at zone 11, note zone 5’s diametric relation,
and consider the options: transpose either by tritone or by major second. No need
to vainly grope.
Imagine now that you wish to embellish this progression with the addition of
two further triads. Your first impulse might be to select from zones 2 and 8, midway
between 5 and 11. The outcome is a set of four modally matched triads whose
roots are drawn from the four distinct augmented triads (and hence form one
of the all-combinatorial tetrachords). But perhaps modal uniformity lacks suffi-
cient lexical diversity for your purposes. You would prefer to mix the two triadic
species in equal proportion while still maintaining a sense of distance. One set of
options is illustrated in figure 5.25: two zone-diametric triads are combined with
their hexatonic poles. These progressions uniquely feature minimal pitch-class
intersection: the four triads have no pitch-class overlap, and together they use all
twelve tones. In the excerpt at figure 5.25(a), from a J. S. Bach Prelude (Dahlhaus
1967b, 87), each major triad is followed by its hexatonic pole; the pair is then
transposed down by a zone-diametric whole step. The four-chord complex rotates

Figure 5.25. Two examples of triadic disjunction.

108  Audacious Euphony

through voice-leading zones 1, 8, 7, and 2, in that order; its transposition in the

following measure rotates through the remaining voice-leading zones. Bach’s four-
chord complex became a trope of disjunction in the long nineteenth century:
examples occur in such diverse venues as Mozart’s c minor Fantasy, Mendelssohn’s
Variations sérieuses, Chopin’s G major Nocturne (mm. 25–27), Schumann’s d
minor Piano Trio, Wagner’s Parsifal, Debussy’s Prelude to the Afternoon of a Faun,
Elgar’s King Olaf, and Hugo Distler’s motet Fürwahr, er trag unsere Krankheit.
Under permutation, the same complex occurs in Chopin’s G major Nocturne
(mm. 1–12), Liszt’s Die Legende vom heiligen Stanislaus, Strauss’s Salome, and
Kodály’s Mountain Nights.18 In the excerpt presented as figure 5.25(b), from
Parsifal, a pair of H-related triads is transposed down by the other zone-diametric
interval, a tritone. Together, these progressions exhaust (to within transposition
and reordering) the possible ways of distributing the pitch-class aggregate among
four distinct consonant triads.
All of the examples explored so far have an aspect of sequential patterning. If
we want to forgo consistency as well as conjunction, then we want to project not
maximal disjunction but maximal variety along the conjunct/disjunct spectrum.
Here we can take advantage of the homology between figure 5.24 and the universe
of twelve tones, about which we know a good deal. If we consider the voice-leading
distance between two triads in terms of an interval, then we are in a position to
define the total interval content of an ensemble of triads and to represent that
information in an array. To achieve maximum entropy, the triadic pairs should be
distributed as evenly as possible among the six intervals. The optimal way to do
this is to find an ensemble of triads whose voice-leading zones are homologous to
one of the all-interval tetrachords, of which there are two types, prime form [0146]
and [0137]. Both tetrachords are octatonic subsets, and the eight voice-leading
zones occupied by the consonant triads have an octatonic structure. Thus, there
exist triadic quartets whose constituent pairs are maximally distributed across the
six interval classes.
Taking inspiration from an aspect of Elliott Carter’s compositional practice,
Robert Morris (1990) has shown that any nonintersecting union of a tritone with
a minor third creates an all-interval tetrachord. Piping this observation through
our homology, we learn that a maximally entropic ensemble of four triads consists
of two zone-diametric (T2- or T6-related) triads of one mode, combined with two
triads of the opposite mode that are related to each other by an odd transposition
(and thus three zones apart, as established earlier in this chapter). Figure 5.26 pres-
ents two options, one for each species of all-interval tetrachord. Figure 5.26(a)
models mm. 8–12 of the finale of Beethoven’s “Tempest” Sonata, which moves
through aᇳiv, ᅈII, V6, iᇴprogression in d minor. The voice leading zones, {g = 7,
Eᅈ = 8, A = 2, d = 4}, correspond to prime form [0146], one of the all-interval tet-
rachords. Figure 5.26(b) models the opening measures of Götterdämmerung,
which toggles between eᅈ minor and a series of foils: first Cᅈ major, then dᅈ minor,
and finally D major, with dominant seventh, in a quotation of the Fate motive.

18. Sources that identify and analyze these passages include Hull 1915, Karg-Elert 1930, Lendvai 1983,
Cohn 1996, and Cohn 2004.
 CHAPTER 5 A Unified Model  109

Figure 5.26. Two “all-interval” quartets of triads. (a) is aᇳi, iv, ᅈII, V6ᇴprogression
in d minor. (b) alternates eᅈ minor with Cᅈ major, then dᅈ minor, and finally D
major (with a seventh), in a quotation of the Fate motive.

These voice leading zones, {eᅈ = 7, Cᅈ = 8, dᅈ = 1, D = 5}, correspond to prime

form [0137], the other all-interval tetrachord. Arrowed lines connect order-adjacent
triads, and the remaining connections are indicated by broken lines.
Disjunction and entropy are the province of poststructural approaches to
nineteenth-century analysis, and it is the practitioners of those approaches who
are best qualified to weave those appraisals into significant critical arguments. The
comments provided here suggest, perhaps surprisingly, that atonal pitch-class
theory has the capacity to stimulate, undergird, or nuance those attributions, when
its conceptual framework is analytically directed toward the model of the triadic
universe proposed in these pages.
This page intentionally left blank
C HA P T E R
Six
Navigating the Triadic Universe:
Three Compositional Scripts

Having established the voice-leading structure of the triadic universe, explored

ways to represent that structure, and identified vehicles for navigating it, we are
now in a position to investigate how nineteenth-century composers created com-
positional strategies in response to it. In the compositions explored in chapter 5,
varied repetition was the guiding principle, generating chromatic sequences
and deforming them via transformation-class substitutions. In the compositions
analyzed in this chapter, repetition plays a less systematic role. Compositional
material is still repeated, varied, and transformed, but the periodicity, flow,
contrast, and trajectory of events acquire the flexibility associated with sonata,
symphony, and Lied, genres that furnish most of the excerpts investigated here.
The analyses of this chapter are organized around three scripts. Some compo-
sitions establish neighborhoods and enact pitch retention loops. They visit a small
set of adjacent voice-leading zones, and create continuity through retention of one
or more focal tones. Some progress through a cycle. The analytical focus is on
Schubert’s song “Der Doppelgänger” and the first-movement development of
Brahms’s Second Symphony. Other compositions execute a classical departure →
return trope, using novel resources to define a route, a destination, and the nature
of reversal. In one variation familiar from the eighteenth century, the return over-
shoots the origin, requiring a final recuperation. I shall analyze two Schubert
songs, the well-known “Auf dem Flusse” and the little-known “Liedesend’” as well
as two instrumental compositions: the well-known first movement of Schubert’s
Bᅈ major Piano Sonata, and a little-known Organ Kyrie of Franz Liszt. The final
script, continuous upshift, is also a variation on a classical trope, in nuce the
Mannheim Rocket, in elaboration the Sturm und Drang developmental core. After
reviewing some sequential models from Beethoven, I examine the upshifting
development of Dvořák’s “New World” Symphony, which treats thematic materials
more freely.
These analytical vignettes make no claim to completeness. My goal for
them has been to establish that pan-triadic voice-leading models can make
contributions to interpretation. But they do not constitute interpretation in and

111
 CHAPTER 6 Navigating the Universe  113

Neighborhoods and Pitch Retention Loops

Example 6.1 reproduces a Tonschema (tonal scheme) from Czech theorist Ottokar
Hostinský’s 1879 Die Lehre von den musikalischen Klängen. The figure is a varia-
tion on Riemann’s Tonnetz; its diagonal positioning of major and minor thirds
is echoed in the Tonnetze presented in this book. Hostinský grows the tonal
network from the C circled at its center. Radiating from C are six edges that
connect to the six tones with which it is consonant. Six further edges, forming a
hexagon, bind those six tones directly to each other by consonances. The hexagons
and radii together form six triangles, representing the six triads that include C (as
root, third, or fifth of a major or minor triad). Following Igor Popovic (1992, 101),
we will call this complex of tones and triads C’s neighborhood.
Each of these six tones generates its own neighborhood, whose bounding
hexagon overlaps with the one surrounding C, producing a third tier of tones,
initiating a process that radiates outward as far as one pleases to imagine. It is not
difficult to imagine what might have stimulated a late-century theorist to explore
the structure of pitch-class neighborhoods. Much music of the nineteenth century
is concerned with kaleidoscopic pan-triadic harmonizations of a static pitch. This
concern is well documented in the case of Italian opera, which William Rothstein
(2008) posits as historically antecedent to central European developments.1
German Lieder that exploit this technique to obvious programmatic effect include
Schubert’s “Zügenglöcklein,” D. 871 (1826), and Peter Cornelius’s “Ein Ton,” Op. 3,
no. 3 (1854).2
Schubert’s late Schwanengesang in b minor, “Der Doppelgänger,” presents
an example of an entire composition based on a neighborhood centered on Fᅊ
(Saslaw and Walsh 1996, 231). A score of the song is available at Web score 6.2 .
The song opens with the serial presentation of four tones of that region,ᇳB, Aᅊ,
D, Cᅊᇴ , in dyadic partnership with a drone Fᅊ. This segment develops into an
ostinato, sometimes substituting A for Aᅊ as a fifth tone of that neighborhood.

Figure 6.1. Hostinský’s Tonnetz.

1. Webster 1991, 15–17, indicates another antecedent in Haydn’s “Farewell” Symphony (1772). Siciliano
2002 and Clark 2011a both devote considerable attention to pitch retention schemes in the music of
Schubert.
2. I am grateful to Christoph Hust for calling my attention to these songs. See also Siciliano 2005b
regarding Schubert’s “Trost.”
114  Audacious Euphony

Each dyad is susceptible to triadic completion in two distinct ways, and the song
engages all of these possibilities (Code 1995; Kurth 1997). The song’s dramatic
wave crests at the completion of Fᅊ’s pitch-class neighborhood, when Dᅊ is
introduced at m. 47 as the root of a minor triad. When that Dᅊ is presented in
conjunction with B, as components of a B major triad, Fᅊ’s neighborhood is
completed at the dyadic and triadic level as well, marking the end of the song.
Figure 6.2 presents this narrative as a series of cumulative snapshots. Web ani-
mation 6.2 unites them as a continuous presentation. Figure 6.2(a) shows the
situation at m. 1. By m. 4 (figure 6.2(b)), four of the tones in Fᅊ’s neighborhood
have been presented, each as a component of a dyad. At m. 7 (c), the singer arpeg-
giates b minor, filling the BD dyad. At mm. 10–11 (d), A is sounded for the first
time, as a component of fᅊ minor and D major. At m. 12, a fourth triad, Fᅊ major,
completes the first couplet. At m. 46 (e), the neighborhood’s final tone, Dᅊ,
is sounded, as part of a dᅊ minor triad. At the final cadence (f), BDᅊ is sounded
within a B major triad, completing the neighborhood’s dyadic and triadic
“aggregates.”
Figure 6.2(g) sketches some events omitted from the above narrative, which
transpire mostly to the left (subdominant) flank of the central hexagon. (1) At the
end of the first two couplets (mm. 12–14 and 22–24), E is outposted as a lone
representative of b minor’s subdominant region; its appearance locally is as the
seventh of the dominant.3 (2) E returns at the end of the third couplet (mm. 32–33)
as a component of a CE dyad, again disconnected from the central hexagon, and
again sounding against dominant pitch classes as part of a French sixth chord. (3)
At the end of the fourth couplet (m. 42), E returns as part of a complete C major
triad, still disconnected from the Fᅊ neighborhood, and sounding against a single
dominant pitch, Aᅊ, as part of a German sixth chord. (4) The climactic dᅊ minor
alternates with its nebenverwandt at mm. 48–50, briefly cultivating new territory
for the first time on the dominant (northeast) side of the central complex. (5) The
enharmonic reinterpretation of C at m. 51, as part of a G major German sixth,
causes a return to the subdominant region and begins to close the triadic gap that
separates the C major complex from the central neighborhood. (6) This gap is
completely filled with the sounding of an e minor triad, initially over a dominant
pedal point at m. 54, ultimately over a tonic pedal, as part of the final plagal motion
at m. 61.
These analytical remarks, in addition to illustrating pitch retention strategies
via neighborhoods, also showcase three virtues of the Tonnetz as a mode of descrip-
tion and analytical representation. First, by representing triads as assemblies of
atomistic pitch-class components, the Tonnetz adapts well to surfaces that feature
dyads and individual tones. Rather than compelling us to decide whether the ini-
tial dyad signifies b minor or B major, it simply documents its presence and invites
us to infer its plausible completions. Second, the Tonnetz is neutral with respect
not only to triadic constituency but also to tonal (key) centricity. Although, as we

3. The idea that a subdominant pitch represents a subdominant function, even when attached
to a dominant triad, is a central element of Riemann’s theory of function and its development in
Harrison 1994.
Figure 6.2. Stroboscopic Tonnetz portrait of “Der Doppelgänger.” An animation,
with recorded performance, is at Web animation 6.2.
116  Audacious Euphony

shall see in chapter 8, it can be adapted to express tonal interpretations, it is not

compelled to do so. Thus, for example, we need not submit to arbitration the ques-
tion of whether the final harmonies of “Der Doppelgänger” move from minor
subdominant to B major tonic (L. Kramer 1986; Kurth 1997) or to major domi-
nant from e minor tonic (R. Kramer 1994; Code 1995; Schwarz 1997) or form
some liminal blend of the two (Saslaw and Walsh 1996). Finally, as an interpreta-
tive consequence of these two levels of descriptive neutrality, the Tonnetz facili-
tates tracking the accumulation of dyads and pitch classes, encouraging us to
observe how they might form systematic aggregations independently of their tri-
adic constituency. In the case of “Der Doppelgänger,” for example, we were able to
watch Fᅊ’s neighborhood build, and come to completion, at the level of individual
tones, dyads, and triads and to trace the filling of a problematic gap opened up by
the outlying E first introducted at m. 12.
Our discussion so far has not presupposed any particular ordering of a neigh-
borhood’s six triads. But the triadic adjacencies on Hostinský’s Tonschema visu-
ally emphasize the relationships between those triads that hold two tones in
common, suggesting a canonical ordering that makes a cyclical tour of the neigh-
borhood. The transformational vehicles for such a tour are the three common-
tone maximizers, P, R, and L, placed in any order and then repeated in the same
order. Taking advantage of an acronymous fortuity, we will call such a tour a pitch
retention loop. Figure 6.3 models the cantabile section of Verdi’s “Ah sì, ben mio”
(Il Trovatore, act 3). The first of the cantabile’s three quatrains modulates from f
minor to Aᅈ major. The second quatrain begins in aᅈ minor, sounds Fᅈ major
through its consequent phrase, and repeats the text of that consequent over dᅈ
minor. The final quatrain prolongs Dᅈ major. A final L would return the initial
triad; instead, this transformation indicates the progress of the cantabile across its
entire span.

Figure 6.3. Tonnetz model of Manrico’s aria from Verdi’s Il Trovatore.

 CHAPTER 6 Navigating the Universe  117

A more ambitious example of the structuring force of pitch retention loops

occurs in the first movement of Brahms’s Second Symphony. The final part of the
development, which begins at m. 246 and elides into a characteristically underar-
ticulated recapitulation fifty-six measures later, is initiated by the reentry of the
timpani, which has been absent since the middle of the exposition. A score of the
extended passage is available at Web score 6.4 . As is normal for a symphony in
D major, the timpani sounds only two pitches during the retransition: tonic D at
its beginning and near its end, and dominant A through its middle. The neat con-
ventionality of this description masks an anomaly: with one underemphasized
exception, these pitches serve not as roots of tonic and dominant triads (or their
dissonant extensions) but rather as thirds and fifths that disperse inventively across
their respective neighborhoods.
The passage proceeds in a consistent four-bar hypermeter, articulated by har-
monic changes and by the periodic rotation of three thematic/textural modules
whose first appearances are transcribed in figure 6.4. A is a dyadic figure derived
from the opening theme, sounding heterophonically and in double hemiola
throughout the orchestra. B is a metrically malleable violin theme accompanied by
tremolo strings. The lyrical, metrically stable C is the only module to feature sig-
nificant internal harmonic progressions. The triple rotation of these modules
establishes a twelve-bar hypermeter as well, which is sustained even after the
rotation is abandoned.
Five twelve-bar hypermeasures are indicated in figure 6.5, which presents the
timpani in metric reduction. Double bars indicate the boundaries of four-bar
hypermeasures. (The timpani’s characteristic rhythm is given for each hypermeasure,
so that the durational values do not always add up to equal units as in regular score
notation.) My discussion takes this reduction quasi-literally: I use “measure” (in
quotes) to refer to what is notated as four bars of the score, reserving “hypermea-
sure” for Brahms’s twelve-bar units. The first two hypermeasures cycle through the

Figure 6.4. Brahms, Symphony no. 2, 1st mvt., mm. 246–57, with three themes
designated.
118  Audacious Euphony

Figure 6.5. Brahms, Symphony no. 2, 1st mvt. thematic rotations in mm. 246–302.

three modules in order. The third hypermeasure disrupts this pattern, omitting the
initial dyadic segment and filling out the hypermeasure by repeating the final lyri-
cal module, B → C → C. The fourth hypermeasure is dominated by the previously
absent A module, performed by the entire orchestra for its first two “measures”
and by the brass alone for the final one.
The boundaries of the fifth hypermeasure are ambiguous, as indicated by the
overlap in figure 6.5. Continuing to project a twelve-bar periodicity, one expects a
new hypermeasure at m. 294. This expectation is rewarded by the timpani roll that
has initiated each previous hypermeasure, and is further supported by the timpa-
ni’s return to the tonic after an absence of twenty-four measures. On this hearing,
the fifth hypermeasure dissolves one “measure” early, with the arrival of the reca-
pitulation at m. 302. There is, however, motivation to hear this fifth hypermeasure
as a complete twelve-bar unit, beginning already at m. 290. The downbeat of that
measure is a “loud rest,” marked especially by the absence of the timpani, which
has sounded at every other “downbeat” of the retransition but is tacit throughout
this “measure.” Four other factors reinforce the accentual status of m. 290. A root-
position tonic chord sounds for the first time since m. 50 (and indeed will not
sound again until m. 477!). The B module, absent since m. 273, returns in the
violin and continues into the subsequent “measure,” so that m. 294 sounds as con-
tinuation rather than a new event. The subito piano at m. 290 creates an accent of
change, and the counterpointing of two melodic modules creates a textural accent.
These factors conspire to suggest an equivocal reading and to assign a transitional
role to the “measure” beginning at m. 290.
These considerations create a context for analyzing the harmonic design of
these measures, which revolve kaleidoscopically about the timpani pedals without
privileging the tonic and dominant triads that they traditionally imply. One can
get an immediate sense of how Brahms withholds the harmonic definition that
D → A → D might otherwise signal in a D major context by examining the
 CHAPTER 6 Navigating the Universe  119

harmonies that occur when the timpanist smacks a new drum. The initial D, at m.
246, is the fifth of a G major triad and, moreover, one that initially sounds like
dominant of C by virtue of its phrygian approach. The switch to the A drum at m.
262 occurs, ironically, at the moment that D sounds as a harmonic root. When the
timpani returns to D at m. 294, it is a diminished fifth above the Gᅊ sounding in
the bass.4
Figure 6.6(a) models mm. 246–70 on a Tonnetz that combines two overlapping
hexagons. The first twelve-bar hypermeasure charts a progression around part of
the D retention loop, G major → g minor → Bᅈ major.5 The harmonically mobile C
module hurries through d minor to D major at m. 257 (shown by dotted arrows)
before returning to Bᅈ major at the beginning of the second twelve-bar unit (258).
The following “measure” (262) revisits d minor but over an A in the timpani.
Preparing now to exit the D neighborhood, we can note that four of its triads have
been prolonged for a full four-bar “measure.” Tonic D major has appeared briefly
in the anacrustic m. 257, and its relative minor not at all.
The initial progression d minor → F major that enters the A neighborhood at
the onset of the C module (m. 266) is a transposition of the corresponding moment
twelve measures earlier (g minor → Bᅈ major), as is the continuation a minor →

Figure 6.6. Tonnetz models of Brahms Symphony no. 2, 1st mvt.

4. In this respect, the entry of the timpani on D at m. 294 associates with the first timpani entrance of
the movement, at m. 32, where a solo D tremolo, initially understood to opaquely represent an
eagerly anticipated tonic, becomes absorbed into a diminished seventh chord in the low brass, with
Gᅊ as the lowest pitch of all. See Schachter 1983a.
5. The augmented triad at m. 252, and at analogous points in the B module, is omitted for visual clarity
but may easily be imagined along lines sketched out in connection with figure 4.6(c).
120  Audacious Euphony

A major (267–69). This transpositional logic projects a return to F major at the

opening of the third hypermeasure, corresponding to the return to Bᅈ major at
m. 258. Instead, Brahms continues a counterclockwise tour of the A neighbor-
hood, establishing fᅊ minor at m. 270. To this point, the motion through the loop
is entirely “stepwise.” One further step around the loop completes the tour of the
A neighborhood and, at the same moment, returns to the tonic triad.
Brahms forestalls this outcome by reversing course, as shown at figure 6.6(b):
he toggles between fᅊ minor and A major for six measures (mm. 274–279) and
returns to F major at m. 281. The fourth hypermeasure, beginning at m. 282, is
dominated by the dyadic A module that went missing in the previous rotation. In
its two previous occurrences, the major-third dyad in the strings sounded as part
of a major triad completed by the winds. At m. 282, only the FA dyad is sounded.
This dyad is initially heard to represent F major, the harmony of the previous mea-
sure. But the remaining events in this rotational unit complicate this interpreta-
tion. At m. 286 the same music repeats, but F is replaced by Fᅊ. If we hear FA as
root and third of F major, then we have motivation to hear FᅊA as root and third
of Fᅊ minor, by virtue of the parallelism. But a Slide relation from F major to
fᅊ minor is not consistent with the double common-tone retention that character-
izes the rest of the passage. Moreover, at m. 290, FᅊA is completed by D. If we
heard fᅊ minor at m. 286, then we hear an implicit Leittonwechsel, with D emerg-
ing from a phantom Cᅊ. But perhaps it is more plausible to retrospectively inter-
pret FᅊA as completing a D major that was already implicitly present. And if so,
perhaps we have a motivation, on analogous grounds, to retrospectively interpret
FA in term of d minor, if not when it is first sounded at m. 282 then perhaps at
some point in the middle of its span.
These questions are ultimately not subject to arbitration. But, as we noted in
connection with our discussion of “Der Doppelgänger,” one of the virtues of the
Tonnetz is that it assigns a determinate but triadically neutral location to a dyad, at
the edge that separates the two viable triadic interpretations. Accordingly, we can
avoid arbitration entirely, in favor of a Solomonic option that splits the difference.
The graphic solution, which I have adopted in figure 6.6(b), is to open up the
dyadic edge into an oval, so that it becomes a location in its own right, not just a
boundary between two adjacent triangles.
Although the arrival of D major at m. 290 marks the return of the tonic, the
release of the A pedal in the timpani, and the completion of the A neighborhood,
its rhetorical force is compromised by the subito piano and the absent timpani,
which has attacked every previous hyperdownbeat. Accelerating the periodicity of
harmonic change, Brahms now completes the tour of the D retention loop that had
been aborted at m. 262, traversing b minor (with Gᅊ under-seventh) at m. 294 and
arriving back at G major (with E added sixth, 296), from which the retransition
had originated in m. 246. After the G major → g minor at m. 298 (with added
under-seventh) reprises the progression with which the retransition began, Brahms
concludes with a plagal nebenverwandt from minor subdominant to tonic (in 64
position) for the beginning of the recapitulation.
Figure 6.7 provides a synopsis of the sixty measures (see also Web animation
6.7 ). Larger dots mark harmonies that arrive on hypermetric downbeats, most
 CHAPTER 6 Navigating the Universe  121

Figure 6.7. Synoptic Tonnetz model of Brahms Symphony no. 2, 1st mvt., mm.
246–302. An animation, with recorded performance, is at Web animation 6.7.

of which are prolonged for four measures. Harmonies that arrive elsewhere in the
hypermeasure are shown selectively, with smaller dots. The tour sounds every
triad in the double neighborhood, and no triads outside of it. The figure commu-
nicates the comprehensive logic of the entire passage as a counterclockwise tour of
the space, from subdominant to subdominant, with some back-filling that extends
the retention loop around the dominant A and delays the premature return to the
tonic region.

Departure → Return Scripts

The path → reversal sequence is a fundamental template of musical form. Its clas-
sical prototype moves from tonic to dominant and back. Tonics and dominants
can be realized in many ways, not all of which will project an equally strong sense
of reversal. But if both departure and return realize least-motion voice leading,
then each upper voice is structured as a palindrome. The sense of departure →
return is particularly strong for a performer, for whom the complementarity of
downshifting and upshifting involves a kinetic aspect.6
The situation just described is not specific to the relationship between a tonic
and its dominant. Reversing the order of any series of harmonies reverses their
individual voices. The complementarity of departure and return thus is reflected in
a complementarity of upshifting and downshifting, as those terms apply both to
the individual voices and to their aggregation into triads. Once fifth relations lose
their prototypical status, the specific case of I → V → I is absorbed into the general
case of downshift/upshift complementarity. This suggests that a departure → return
script can be fulfilled by any triadic progression whose voices move in a uniform
direction. A return script is fulfilled not only by a literal reversal of that triadic
progression but also more abstractly by a progression that moves through the
same voice-leading zones in reverse order, freely substituting same-group
transformations and same-sum triads.

6. Depending on the mode of performance, this kinesis involves a shift of bow angle, hand position,
throat physiology, embouchure, or some combination thereof.
122  Audacious Euphony

To illustrate, compare the opening of “Auf dem Flusse,” the seventh song
from Schubert’s Winterreise, with the beginning of its final stanza. A score of
the opening and closing stanzas of the song are available at Web score 6.8 .
The first quatrain, “Der du so lustig rauchtest . . . ,” is set as a period whose
antecedent leads from e minor to a half-cadence, and whose consequent prolongs
and fully cadences in the tonally distant dᅊ minor. The antecedent contains
only the e minor tonic and its N-related major dominant. Across the phrase bound-
ary, the root of B major drops a semitone to the fifth of dᅊ minor, effecting a
leading-tone exchange. Because N and L represent different transformation
classes (Weitzmann and Hexatonic respectively), their alternation produces a
unidirectional downshifting, from zone 10 to 5. This motion is reversed at the
cadence, when Aᅊ major upshifts to dᅊ minor. (This upshift is idealized; as is
standard at the full classical cadence, voices move down, forgoing common-tone
preservation and voice-leading parsimony in order to effect melodic resolution.)
This reversal triggers an unraveling of the previous transformations: the leading-
tone exchange is undone at the midpoint of the two-measure interlude, and the
opening nebenverwandt is reversed at the onset of the second stanza, which repeats
the music of the first.
The final stanza, “Mein Herz, in diesem Bache” is delivered at the rate of one
couplet per period, rather than one quatrain as previously, and the entire stanza is
given two complete settings. Consequently, this single quatrain of text is accorded
four full periods, as many as the four previous quatrains combined. The initial set-
ting of the first couplet at mm. 41–47 replicates that of the opening quatrain,
cadencing in dᅊ minor. Schubert connected the initial stanza to its successor
through a two-measure interlude (mm. 12–13), prolonging dᅊ minor for a mea-
sure and then converting it to a dominant through Aᅊ → B. At m. 48, Schubert
compresses the interlude into a single measure: one beat of dᅊ minor and one beat
of semitonal ascent to the dominant of the new tonic. But Schubert assigns the
mid-measure semitonal ascent to the “wrong” voice:” Fᅊ instead of the Aᅊ of
m. 12. Accordingly, dᅊ minor leads not to B major via L but rather to Dᅊ major
via P. And when that chord acts as a dominant, it leads not to tonic e minor but
rather to gᅊ minor. As at the end of the first stanza, Schubert has upshifted through
zone 8 to the home zone 10. But the P-for-L substitution causes the triadic repre-
sentatives of those zones to transpose upward by major third. It is the job of the
second couplet, then, to undo this diversion. It does so by directly juxtaposing
the dominants of gᅊ minor and e minor, paving the way for the tonic cadence at
the end of the final stanza’s first full setting.
Figure 6.8 presents a paradigmatic analysis (Agawu 2009) that facilitates com-
parison of these progressions. The first line shows what the two settings have in
common, in an abstract sense: both use W → H → W to downshift zones 10 → 5 and
the same combination to upshift back to zone 10. (Recall that W and H, respec-
tively, signify Weitzmann and hexatonic classes of transformations, not the trans-
formations themselves.) The first and second stanzas (“I, II”) realize both legs
of this journey in terms of N → L → N. The first couplet of the fifth stanza (“V:1”)
realizes the departure leg in the same way but diverts the course of the return leg
through a P-for-L substitution, symbolized here by a diagonal arrow that throws
 CHAPTER 6 Navigating the Universe  123

Figure 6.8. Voice-leading zones in Schubert’s “Auf dem Flusse.”

the progression off its horizontal course, and requires rectification through the
rectilinear N → PL → N combination (“mm. 50–55”).
The second setting of the final stanza, beginning at m. 55, likewise sets each
couplet to its own period, and each of these spins a new variation on the down-
shift/upshift script. Schubert accelerates the harmonic rhythm, and hence the rate
of downshifting, so that zone 5 is reached already by the half-cadence of each four-
measure antecedent, rather than just before the full cadence of each eight-measure
period as previously. We will take the final period (mm. 64–72) first, as presenting
the simpler case. Whereas at m. 48 Schubert diverted the harmonic flow through
an H-class substitution during the upshift leg, here he does the same during the
downshift, substituting for the anticipated L the third and final constituent of the
H-class group, the “shocking” hexatonic pole (H) (Newcomb 1986, 164). Thus, B
major proceeds to g minor rather than to dᅊ minor, and its nebenverwandt pro-
ceeds to D major rather than to Aᅊ major as in the first two stanzas. As at m. 48,
a substitution for L has created a transposition upward by major third but this
time at the end of an antecedent phrase rather than at its beginning.
In the setting of the previous period (mm. 55–62), the antecedent phrase is
identical, except that G major substitutes for g minor. The rare juxtaposition of
124  Audacious Euphony

two major triads (mm. 56–57) prolongs zone 8 for an additional measure and
causes the omission of zone 7 altogether. The half cadence is followed, as in the
earlier stanzas, by a leading-tone exchange, taking the root of D major to the fifth
of fᅊ minor and driving us into the as yet uncharted depths of voice-leading zone
4 and then fleetingly to 2 (Cᅊ major) in preparation for the cadence. The two-
measure interlude reclaims tonic e minor on schedule by the unparsimonious, but
conventionally diatonic, upshift from fᅊ minor to B major.
Leaving aside the tonally stable E major third and fourth stanzas, the chro-
matic sections of “Auf dem Flusse” all involve a departure → return script, with
each departure realized as a downshift from e minor and each return as an upshift
to that tonic or its zonal affiliate, gᅊ minor. Scanning down a column of chords in
figure 6.8 brings to light substitutional relations within voice-leading zones, all
involving transposition by major third. Scanning down a column of transforma-
tions, we see that all of those chordal substitutions result from transformational
substitutions within the Hexatonic class: P and H both stand in for L. The
Weitzmann-class transformations are perpetually represented, without substitu-
tion, by the ubiquitous nebenverwandt.
Richard Kramer has written that “the modulatory adventure of Auf dem Flusse
is extreme. No other song in Winterreise is endowed with a tonal graph anywhere
near as complex” (1994, 159). Encountering this characterization without knowl-
edge of the song, one might imagine that “Auf dem Flusse” is deeply entangled in
the enharmonic ambiguities characteristic of Schubert’s most tonally adventurous
compositions. Against this expectation, it comes as a surprise to realize that “Auf
dem Flusse” achieves the status that Kramer assigns it without ever approaching,
much less threading, an enharmonic seam. Except for the diatonically conven-
tional E major setting of the third and fourth stanzas, “Auf dem Flusse” lives
entirely “downstream” from its tonic. The song is nebenverwandt-saturated: every
minor triad is followed or preceded by its major dominant, and with the single
exception of G major (m. 58), every major triad is adjacent to some minor tonic to
which it is dominant. The subdominant, accordingly, is absent to a remarkable
degree: aside from a tonicized fᅊ minor at m. 62, whose subdominant status is real-
ized only in retrospect, only eight chords have subdominant value, and none of
these has a duration of more than a single beat.
The previous paragraph brings to the fore the potential affiliation of down-
shifting and upshifting, respectively, with dominant and subdominant. To what
extent does the voice-leading dualism of down and up interact with the Riemannian
dualism of subdominant and dominant? It is tempting, for example, to hear gᅊ
minor at m. 50 as a tonic substitution, the Leittonwechsel of the tonic’s modal vari-
ant. By extension, its Dᅊ major predecessor functions as dominant, and by further
extension, so too do the other tonicized leading tones at mm. 12 and 21, as David
Lewin suggests (2006, 115). Projecting the logic, Aᅊ major represents double-
dominant Fᅊ major, as the modal variant of its Leittonwechsel, and so too does the
fᅊ minor tonicized at m. 62. This leads to a hypothesis that these functional assign-
ments extend to entire zones, so that 10 always represents tonic, 7 and 8 dominant,
and 5 double-dominant. But this proposition immediately attracts counterexam-
ples. Lewin, for example, hears gᅊ minor (m. 50) as the relative of the dominant,
 CHAPTER 6 Navigating the Universe  125

rather than as the tonic representative suggested by the hypothesis. And the paral-
lelism, with respect to phrase positioning and text, of G major (m. 58) and g minor
(m. 66) with e minor at m. 45 inclines us to hear G in terms of tonic function, not
the dominant of our hypothesis. The hypothesis that chords whose roots share an
augmented triad always share harmonic function, which I advanced in Cohn 1999
and critique in chapter 8, is as Procrustean as Ernő Lendvai’s theory of axis tonality
(1971), which makes an analogous claim about roots that share a diminished
seventh chord.
One of the benefits of voice-leading zones is that they supervene upon such
functional hearings, suggesting but not mandating them. Downshifting and
upshifting may be interpreted in terms of plagal and authentic motions, respec-
tively. But if they make claims that are counterintuitive or uncomfortable, the
functional scaffold can be shed, leaving the zones to carry the weight of the struc-
tural claims. One advantage of this view of dualism is that it is immune to dis-
missal on the grounds that it is based on the metaphysical or spurious physical
claims discussed in chapter 2. It merely explores the ramifications of the following
aggregation of circumstances: that pitch motion exists on a linear continuum,
which our culture metaphorically projects onto a vertical axis, as higher and lower;
that pitch-class motion exists on a cyclical continuum; thus, that motion along
either continuum can occur in two opposite directions; and that those motions
have experiential correlates in the aural perception of sound and the kinetics of its
production. This is a dualism that even the most skeptical realist can abide!7
These considerations come to the fore in the first movement of Schubert’s Bᅈ
major Piano Sonata, whose entire score is available at Web score 6.9 . In its
general outlines, the tonal structure of the movement conforms to classical prin-
ciples: an exposition that establishes tonic, proceeds through secondary thematic
material, and concludes in the dominant; a development that concludes with a
retransitional dominant; and an ordered recapitulation of the thematic material of
the exposition, tonally adjusted to conclude in the tonic. This skeleton, though,
supports a number of tonal events that are anomalous from a classical standpoint.
Intervening between the first theme and its perorational counterstatement is a
passage that prolongs Gᅈ major. The second theme is presented in fᅊ minor and
tonicizes A major before arriving at the dominant, the four paragraphs of the
development depart respectively from cᅊ minor, A major, Dᅈ major, and d minor,
and the recapitulation of the second theme is given in b minor and D major. Of the
eight tonic triads just listed, only one (d minor) is diatonic to Bᅈ major, and only
two (Gᅈ major, Dᅈ major) are diatonic to bᅈ minor, and hence derivable via first-
order modal mixture. The remaining five keys are diatonically indeterminate
and involve some degree of enharmonic paradox along lines already explored in
chapter 1, in connection with a segment of this movement.

7. The upshift/downshift duality is also strongly related to the duality of strong and weak harmonic
progression, a distinction made by Arnold Schoenberg and developed by Nicholas Meeus (2000).
Strong progressions, whose roots descend by third or fifth, or ascend by second, all upshift under
idealized voice leading. The only complication here is the equivocal voice-leading status of modally
matched chords whose roots are a major second apart (IV → V in major; ii → iii in minor), which
involve six units of voice leading in either direction.
126  Audacious Euphony

Figure 6.9. Three models of Schubert’s Sonata in Bᅈ major, D. 960, 1st mvt.

Figure 6.9 presents the main tonal events of the movement, together with
three analytic underlays.8 The first of these analyses assigns function according
to the hexatonic constituency of a triad. Triads in Bᅈ major’s hexatonic system are
assigned tonic function; similarly, subdominant and dominant function are
bestowed on triads hexatonically associated with Eᅈ and F major, respectively.
The structural transition from tonic to dominant, on this reading, occurs at m. 58,
when A major displaces fᅊ minor. (Measures 68–71 are heard as transiently revisit-
ing tonic amidst an overriding dominant prolongation; see Webster 1978–79.) The
first three paragraphs of the development begin with three hexatonic associates of
the dominant, cᅊ minor, A major, and Dᅈ major, respectively. The analysis inter-
prets the rhetorically punctuated and extensively prolonged d minor as a prema-
ture tonic return, overriding the spineless and opiated dominant seventh that
precedes the recapitulation. The recapitulation is entirely in the tonic, except for a
8. This first analysis is equivalent to one proposed in Cohn 1999; the second two analyses develop
observations made near the end of that article, using the technologies and representational modes
introduced in chapter 5.
 CHAPTER 6 Navigating the Universe  127

transient tonicization of A major, representing dominant, and the sounding of the

fᅊ theme in b minor, representing subdominant. The large-scale functional scheme
is consistent with the Classical sonata, including the recapitulatory subdominant
overshoot characteristic of many of Schubert’s first movements. Only two ele-
ments, both associated with d minor, are anomalous: the transient revisitation of
tonic after dominant has been established in the exposition, and the premature
arrival of a tonic representative at the end of the development.
The principal thematic materials of the movement emphasize major thirds that
in combination outline a series of augmented triads (Cohn 1999, 224–25). If we
affiliate these augmented triads with the consonant triads that they displace, then
we hear the functional progression in a different way. The structural motion to
dominant occurs already at the fᅊ minor second theme (m. 48); it need not wait for
the A major cadence ten measures later. The retransitional d minor at m. 171 is
similarly reevaluated: its FA major third affiliates that triad with dominant, not
tonic. These motivic considerations suggest that the harmonies be reinterpreted
according to their Weitzmann-region constituency, leading to my second analysis.
This analysis conforms more closely to the classical ideal for coordinating themes
and structural harmonies: dominant arrives with the second theme and is pro-
longed for the remainder of the exposition and the entire development (except for
its cᅊ minor opening, which uniquely inhabits the distant region of the double
dominant). Tonic returns at the beginning of the recapitulation and maintains its
functional hold throughout, with the exception of some quite transient toniciza-
tions. According to this interpretation, the b minor theme in the recapitulation
represents tonic, by virtue of the common D third.
Using voice-leading zones, the third analysis creates a hybrid of the first
two. The exposition is interpreted as a two-phase descent in voice-leading space.
The fᅊ minor second theme displaces Bᅈ major and its Gᅈ surrogate downward by
one voice-leading degree, 5 → 4. The F major third theme then displaces fᅊ minor
downward by two voice-leading degrees, 4 → 2. The nadir of this downshift is
reached with cᅊ minor at the opening of the development. The trajectory is reversed
when A major displaces cᅊ minor, restoring zone 2. The motion through d minor
to Bᅈ major (retransition, recapitulation) represents a compensatory upshift
through zones 4 and 5. According to this interpretation, the b minor theme in the
recapitulation constitutes an overshoot to sum class 7.
We have the option of interpreting this voice-leading scheme in terms of
Riemannian functions; indeed, Schubert gives us many reasons for doing so. Zones
5 and 2 may be assigned tonic and dominant value, respectively; zone 4 may be
interpreted as occupying an intermediate space between the two functions,
and zones 1 and 7 may be heard as overshooting toward the double-dominant
and subdominant sides, respectively. But we need not make this move if aspects
of it make us uncomfortable. If we are attracted to functions primarily because
they encourage us to bundle musical events into coherent trajectories, such
as the departure → overshoot → return scheme documented here, then they are
dispensable. The zones provide another, more theoretically neutral way of docu-
menting those same trajectories. We may not like that neutrality. We may like
the added value that the language of functions provides, and may be prepared to
128  Audacious Euphony

Figure 6.10. Liszt, Missa pro organo lectarum, Kyrie.

articulate its nature. But we are not stuck with the functional apparatus if we do
not find added value in it or, worse, if we find that it bears implicit baggage that
detracts value.
To focus the point, we now analyze two pan-triadic compositions that execute
a departure → return scheme but where functional ascriptions sit far less comfort-
ably than in Schubert’s Piano Sonata. Figure 6.10 presents a brief wordless setting
of the Kyrie from a Liszt Organ Mass of 1879. Each of its three phrases, corre-
sponding to the Kyrie → Christe → Kyrie sequence of the Latin Mass ordinary,
contain three parallel segments, corresponding to the standard threefold repeti-
tion of each textual unit, followed by a fourth, presumably wordless, cadential
segment. The reduced score notates each of these twelve segments as a measure.
The composition is entirely triadic and begins and ends in Bᅈ major. Aside
from the two interior cadences, there is little to suggest diatonic scale degrees or
Riemannian functions. There is nonetheless an exquisite patterning, the door to
which is opened by an observation of Ramon Satyendra (1992, 98–101): leaving
aside the first two neighboring cadential gestures, each individual voice constitutes
a palindrome whose axis is the double bar about which the key signature converts
from flats to sharps. Yet, because the firing order of the voices is staggered, the
palindrome evident in each voice individually does not result in a palindromic
structure for the triads that those voices comprise.
Figure 6.11 normalizes the voice leading in order to illustrate the trajectory
of the individual voices. The arcs suggest which triads would be identical, if the

Figure 6.11. Palindromic model of figure 6.10.

 CHAPTER 6 Navigating the Universe  129

triadic palindrome were strict. The deviations occur in the third and fourth triads,
where Dᅈ major → f minor is mirrored by cᅊ minor → A major, its retrograded
transposition down a diminished fourth. The palindrome thus is proper not to the
progression of the triads themselves but to that of voice-leading zones that they
represent. The progression of zones illustrates what is already evident to the organ-
ist’s hand: that all voice leading shifts downward in the first two phrases and
upward in the final phrase. Figure 6.12 displays this departure → return scheme on
Cube Dance.
Schubert’s 1816 song “Liedesend’” provides a second, more complex example
of a departure → return scheme realized as a palindrome of voice-leading zones.
A score of the entire song is available at Web score 6.12 . What is at issue here
is not the local harmonic progressions, which define tonics using standard dia-
tonic conventions, but rather the progression of tonics, which George Grove
selected as an example of Schubert’s inscrutable modulatory practice (Clark
2011b). The song is sectionalized and episodic: each of the nine strophes has its
own mood and tempo, and six of them terminate with a fermata. The modulatory
scheme is evidently no less episodic. Seven of the nine strophes are preceded by a
change of key signature, and the song closes in a key that is first introduced only at
the beginning the poem’s final couplet. Suzannah Clark’s recent analysis (2011b,
304–11) suggests that the song is tonally structured by a series of common-tone

Figure 6.12. Cube Dance model of figure 6.10.

130  Audacious Euphony

Figure 6.13. Voice-leading zones in Schubert’s “Liedesend.”

modulations. This observation can be extended by charting the progression of

voice-leading zones.
Figure 6.13 presents the tonal structure, as a palindromic motion through
voice-leading zones. The nine strophes are indicated by Roman numerals and
represented by their beginning and ending tonics. Where successive tonics lack
common tones, they are bridged over by a prolonged dominant that shares a tone
with both. Minor tonics arise three times: at the song’s beginning, conclusion, and
palindromic center. Connecting the c minor beginning to the f minor center is a
series of major triads that cascade from zone 8 down to 11; connecting the center
to the e minor conclusion is a series of major triads that upshift through the same
zones in reverse order. The zonal palindrome is ruffled only by the D major con-
clusion of the eighth stanza, which introduces a local downshift against the pre-
vailing upshift. This perturbation defers the arrival at the goal zone 10, while
reprising and ultimately reversing, at the start of the song’s final stanza, the exact
triadic succession that occurred at the conclusion of its initial one. Figure 6.14
plots the same design on a sparse Tonnetz, painted over with major-third-
generated fault lines. Broken lines indicate a new tonic that is asserted rather than
approached through a modulatory process. Asterisks indicate tonally closed verses.
The graph highlights the downshift/upshift scheme, as well as the common-tone
continuities that Clark observes.
Like the surface harmonies of Liszt’s Kyrie, the tonics in “Liedesend’” do not
convey a strong sense of harmonic function. Any attempt to organize the tonics
about either the opening c minor or the closing e minor would be tendentious to
an extreme degree, an act of desperation in support of a monotonality that is being
 CHAPTER 6 Navigating the Universe  131

Figure 6.14. Sparse Tonnetz model of Schubert’s “Liedesend’.”

clung to at all costs. The palindromic voice leading suggests that the modulatory
scheme need not be deemed arbitrary even if its logic is not monotonal.

Continuous Upshifts

Chapter 5 studied sequential passages that lead in a uniform direction through a

series of adjacent voice-leading zones. Our concern here is with passages that do
so without benefit of patterned repetition. I document continuous upshifting in
two quite distinct environments: as a slow unfolding of structural chords, mostly
tonics, across Schumann’s song cycles, and as a characteristically energized surge
through the development section of first movements of Beethoven and Dvořák.
Fred Lerdahl (2001, 138) plots the key sequence of Robert Schumann’s cycle of
sixteen songs, Dichterliebe, as a coherent path through Weber’s graph of regions.9
The key signatures of songs 1–8 decrement from three to zero sharps, monotoni-
cally every two songs were numbers 4 and 5 swapped. The scalar downshift accel-
erates in songs 9–11, which increment one flat per song. Beginning with the
reversal from three flats to two at song 12, the pattern in key-signature space is
fractured. Lerdahl notes that the keys of the final songs nonetheless fulfill the pro-
jected path by substituting parallel or minor-third related keys, which are privi-
leged in Weber’s space and Lerdahl’s.
The continuous downshifting in scalar space through the first eleven songs of
Dichterliebe is realized in triadic space as a phased upshifting that continues
straight through to the end of the cycle.10 Figure 6.15 traces the progression of the

9. Much ink has been spilled concerning whether the cycle is “organically unified,” and on what
grounds. Ferris 2000, 26–38, summarizes positions; Perrey 2002 and Hoeckner 2006 are more
recent contributions.
10. Komar 1971, 78, observes a pattern of upward displacements between adjacent tonics.
132  Audacious Euphony

Figure 6.15. Tonnetz model of the tonal plan for Schumann’s Dichterliebe. The
numbers, corresponding to the ordinal position of each song in the sixteen-song
cycle, are positioned at their beginning and ending triads. Arrows lead the eye at
points where the right-to-left is perturbed or where a song ends in a different key
than it began.

boundary harmonies, defined as those that begin and end songs. Boundary har-
monies are tonics except when a song begins and/or ends with a dominant prolon-
gation (songs 1 and 9) or when its conclusion modally inflects the opening tonic
(songs 9 and 16). Numbers refer to order position of the sixteen songs. Arrows
guide the eye when a song references multiple structural harmonies or when the
prevailing flow is reversed.
Dichterliebe executes two revolutions about Cube Dance. The initial point of
cyclic renewal occurs at the beginning of song 9, the exact halfway point in the
cycle, where a prolongation of A7 matches that of Cᅊ7 in song 1. In this respect, the
first nine songs of Dichterliebe echo Schumann’s Op. 24 Liederkreis, also setting
poems of Heine, whose ninth and final song completes a similar upshifting rota-
tion about voice-leading space (Hoeckner 2001). This suggests dividing the set
into equal halves, a possibility supported by a circumstance of compositional gen-
esis: Schumann initially composed Dichterliebe as a twenty-song set but withdrew
the original pairs at order positions 5–6 and 15–16, the median points of each
hypothesized half (Hallmark 1975).
Table 6.1 documents the progression of the songs through the two rotations about
the voice-leading cycle, with the withdrawn songs restored (shaded in the table). The
pace of upshifting through the first rotation does not precisely match that of
the second, but some parallels are worth noticing. Each half upshifts 2 → 8 and
then introduces a unique downshift (onset of songs 5 and 12). The final four songs
(in each group of eight) upshift 7 → 1. In the initial conception, the final five songs
(in each group of ten) upshift from zone 5.
E. T. A. Hoffmann’s Kreisleriana, with which Schumann was famously absorbed,
provides a possible source for his interest in the patterned upshifting and zonal
completion that guides the key sequences of his two Heine cycles. In the vignette
titled “Kreisler’s Musico-Poetic Club,” Kreisler slowly improvises a series of ten
triads (some with sevenths) and reports the sensations and images that each one
evokes (Hoffmann 1989 [1813–14], 131–36). Kreisler’s progression begins with a
downshift Aᅈ major → aᅈ minor and ends with the zone-identical downshift
C major → c minor. The six interior harmonies execute an upshift rotation about
voice-leading space, connecting E major → a minor → F major → Bᅈ major → Eᅈ
major → G major.
 CHAPTER 6 Navigating the Universe  133

Table 6.1. Voice-leading zones in Schumann’s Dichterliebe

Songs 1–8 Songs 9–16
Song no. Triad(s) Zone(s) Zone(s) Triad(s) Song no.
1 Cᅊ+ 2 2, 4, 5 A+, d–, D+ 9
2 A+ 2 7 g– 10
3 D+ 5 8 Eᅈ+ 11
4 G+ 8 5 Bᅈ+ 12
a Eᅈ+ 8 7, 8 g–, G+ A
b g–, D+ 7, 5 5 Bᅈ+ B
5 b– 7 7 eᅈ– 13
6 e– 10 8 B+ 14
7 C+ 11 11 E+ 15
8 a– 1 1, 2 Cᅊ–, Dᅈ+ 16
The published songs, numbered 1–16, are supplemented by four that Schumann composed for the cycle
but omitted, which are inserted here in the initial position that he intended them and labeled a, b, A, B.
Triads are those prolonged at the beginnings and endings of songs, whether or not they are local tonics.

The interior of a Classical sonata-form development, which Caplin 1998 calls

its core and Hepokoski and Darcy 2006 its central action zone, characteristically
projects tonal instability and mounting tension, frequently through one or more
sequences that sustain an upshift trajectory. Figure 6.16 models two upshifting
cores from first movements of middle-period Beethoven sonatas. The first, from
the “Tempest” Sonata (d minor, Op. 31 no. 2), contains an intensely compressed
upshift bracketed by two briefer downshifts (indicated by lines beneath the reduced
score). Ametric strummed arpeggios precede that core, leading to Fᅊ major. The
propulsive core begins with a “lights out” downshift from Fᅊ major to fᅊ minor.11
Like Kreisler’s slow and dreamy improvisation, the core of the “Tempest” develop-
ment begins and ends with two zone-identical downshifts, fᅊ minor → Cᅊ major
(99–106) and d minor → A major (117–21), flanking an upshifting passage that
orbits voice-leading space.
A related script is executed over a broader span in the development of
the “Waldstein” Sonata, Op. 53 (figure 6.16(b)).12 The principal theme in F major
(m. 90) begins a downshift that culminates four measures later at D7, as dominant
of g minor. A T5 chain from m. 96, at a four-measure pace, accelerates eightfold at
m. 104, prolonging bᅈ minor and Gᅈ major. After a brief downshift, the sequence

11. A 2009 collection of essays on the “Tempest” Sonata edited by Pieter Bergé, Jeroen D’hoe, and
William E. Caplin contains the following characterizations of the fᅊ minor triad at m. 99: “electrify-
ing, wrenching shock, jolt” (Burnham, 42–44), “surprising harshness” (Burstein, 70), “brutally
negat[ing] explosion” (Hatten, 170), “punched with a vengeance” (Hepokoski, 201), “shocking”
(Kinderman, 220). The critical consensus is as stunning as the moment itself.
12. The Tonnetz interpretation is based in part on Lubin 1974, 109ff.
134  Audacious Euphony

Figure 6.16. Voice-leading zones in development sections of two Beethoven

piano sonatas.

continues at its four-measure pace. Beginning at m. 120, the upshift stalls in a

Weitzmann region, connecting Bᅈ major → b minor through an N/R chain and
reaching the retransitional dominant at m. 130. These two passages suggest that
extended upshifts are ingredients of the adrenal concoction with which Beethoven
so often injected his developments.
The opening movement of Dvořák’s e minor Symphony (1895), “From the New
World,” contains a development section with an even more relentless upshifting
trajectory, carried by a free circulation of thematic materials that are not organized
into sequential repetitions as in the classical prototype. This ninety-six-measure
development is the longest continuous segment of music analyzed in this book
and, as such, constitutes a fittingly weighty close to the six-chapter arc that
constitutes the core presentation of pan-triadic syntax. A score of the entire
development is available at Web score 6.17 .
As a backdrop to our study of this development, I begin with some observations
about the entire movement. One of its peculiarities is its avoidance of the dominant.
B is never tonicized, as major or minor. On those rare occasions when cadential
dominants appear, their characteristic effect is undercut by weak bass degrees
(mm. 237, 269), deceptive resolution (m. 408), or brief duration (mm. 56, 431). The
sole exception is the two-measure caesura that connects the initial Adagio to the
exposition (mm. 22–23). Throughout the movement, the off-tonic role is appropri-
ated by four upper mediant triads (Kopp’s term, 2002), presented in upshifting order.
 CHAPTER 6 Navigating the Universe  135

Figure 6.17. Dvořák, Symphony no. 9, 1st mvt., tonal arc of the second theme.

Figure 6.17 tracks the second theme as it appears in exposition and recapitula-
tion. It first appears in g minor (m. 91), and then in G major at m. 137, the key in
which the pastoral third theme closes the exposition. The corresponding measures
of the recapitulation transpose this music up by chromatic semitone. Aᅈ major, a
notational proxy for Gᅊ, is the hexatonic pole of the tonic, an extraordinary key in
which to close a recapitulation. In figure 6.18, the keys of the second theme climb
the Tonnetz, through the gray band. This progression can be captured as a climb in
voice-leading space, 7 → 8 → 10 → 11, as part of a P/S chain. The beginning of the
coda takes this onward to 2 (A major and then F major) before the latter moves as
a Neapolitan to the fleeting structural dominant at m. 408.
The relaxed upshift through the movement’s pastoral breathing points is
mirrored by an intense upshift across its breathless development section, tracing

Figure 6.18. Tonnetz analysis of figure 6.17.

136  Audacious Euphony

Table 6.2. Triads and voice-leading zones of the development of Dvořák’s

“New World” Symphony, organized around the four thematic entries
Measure Theme Triads Zones
189 3 E+ 11
209 3 F+, fᅊ– 2, 4
233 1 eᅈ–, e– 7, 10
257 1 fᅊ–, A+, Bᅈ+, B+, e– 4, 2, 5, 8, 10

two complete rotations from the G major of the end of exposition to the e minor
of the beginning of the recapitulation. The entire development is in a clear four-bar
hypermeter. Hortatory thematic incipits structure the development into four para-
graphs, the first two dominated by the pastoral third theme and the final two by its
principal theme. These two themes are linked by their #MRRR#M rhythm, associated
with the Hungarian choriambus, and are related by contour inversion.
Table 6.2 indicates the harmonic platforms from which these choriambic vol-
leys are launched. The series of initial stations, E major → F major → eᅈ minor → fᅊ
minor, does not present a very coherent image of purposeful tonal progression.
Interpreting these triads and their successors as representatives of voice-leading
zones, one begins to get a sense of the prevailing upshift winds that blow across
these plains.
Figure 6.19 segments the development into four-bar units. Ovals enclose
hypermeasures that do not represent voice-leading zones; either a diminished
seventh chord is prolonged, as at m. 213, or there is rapid planing between disso-
nances, as at m. 253. The first paragraph presents E major (zone 11), then moves to
cᅊ minor (1), and progresses through its hexatonic region to F major (2), from
which the second paragraph is initiated and proceeds to fᅊ minor (4). Voice-leading
zone 5, represented by a Bᅈ dominant seventh, ushers in the third paragraph.
The third paragraph, which is articulated at m. 233 with the return of the prin-
cipal theme, executes an extended N/L chain that carries straight through to the
start of the fourth paragraph at m. 257. A series of semitonally ascending minor
tonics appear every eight measures: eᅈ minor (233) → e minor (241) → f minor
(249) → fᅊ minor (257). With a single exception, the four-bar median of each eight-
bar span is marked by the arrival of a dominant seventh chord of the subsequent
minor tonic. The exception occurs in the third span, which alone lacks a thematic
incipit. The Dᅈ major dominant projected for m. 253 arrives two bars early. This
buys Dvořák four bars in which to further accelerate the harmonic rhythm in
anticipation of the fourth paragraph at m. 257, where thematic activity resumes.
Dvořák fills these four measures by synopsizing the harmonic course of the prior
twenty measures, planing the series of minor triads on eᅈ minor, e minor, and
f minor, and accompanying each with an under-seventh (see Bass 2001, 51). The fᅊ
minor with which the fourth paragraph begins is rhetorically marked not only by
the return of the principal theme but also by virtue of fulfilling two trajectories
right on time: a slow one initiated from m. 233 and following an eight-bar period-
icity, and a fast one initiated from m. 255 and following a one-bar schedule. The final
 CHAPTER 6 Navigating the Universe  137

Figure 6.19. Dvořák, Symphony no. 9, 1st mvt., voice-leading zones in the
development section (mm. 177–273).

paragraph then reprises the first part of the N/L chain with one addition, A major,
and one deletion, eᅈ minor.
Figure 6.20(a) presents the entire development of the “New World” Symphony
on the Tonnetz. An animated version of this figure, synchronized with a recorded
performance, is presented at Web animation 6.20 . Starting at the right edge, the
initial gesture, covering the precore and the onset of the first paragraph, travels
along an octatonic R/P chain. The second arrow, moving southwest from cᅊ minor,
indicates the hexatonic progression that connects the first paragraph to the open-
ing of the second. A disjunctive move to fᅊ minor, at the end of the second para-
graph, commences an intensive southwest slide along the N/L chain that dominates
the third paragraph. A broken arrow emerging from the graph’s southwest fringe
repositions fᅊ minor back at the top of the slide, to prevent a drift outside the frame
but also to emphasize how the fourth paragraph reprises the opening of the third.
Figure 6.20(b) extracts the right side of the figure and overlays a different
interpretation on the same data. The first paragraph is interpreted here as execut-
ing a four-station tour about an E retention loop. The second paragraph is initiated
138  Audacious Euphony

Figure 6.20. Dvořák, Symphony no. 9, 1st mvt., Tonnetz model of the development.
Animation, with recorded performance, is at Web animation 6.20.

at the moment that E is displaced to F. Bridging the two paragraphs is a second

pitch retention loop about A, beginning at m. 201 (A major) and extending to
the diminished seventh chord at m. 227. This reading draws attention to the rela-
tively slow pace of pitch-class turnover and upshift pacing through the opening
two paragraphs, as compared with the accelerated upshifting of the final two
paragraphs.
The two components of figure 6.20 constitute two “performances” upon a
single pattern of black dots. Both are “note accurate,” but they shape the data in
different ways and draw breaths at different articulation points of the score. The
two graphs illustrate the interpretive leeway available to the pan-triadic analyst,
using the representational “instrument” of the Tonnetz. To what extent do these
acts of “performance” relate directly to musical performances, as traditionally con-
strued? To what extent are they metaphoric and suggestive, inviting imaginative
translation? That is something that only performers can determine; I merely sug-
gest here that both possibilities are available for contemplation.
C HA P T E R
Seven
Dissonance

Four Eighteenth-Century Approaches to Dissonance

Aside from the augmented triad, this book has largely ignored dissonant harmo-
nies, or swept them out of the frame with a series of deferrals. Readers would be
justified in suspecting the waving of hands and in wondering whether the neglect
of seventh chords, and other larger-cardinality harmonies, is simply a flaw in pan-
triadic theory. Even if classical models strain in response to the highly chromatic
repertory of late Romanticism, they do have the virtue of supplying a first-level
descriptor for most dissonant harmonies, and of making claims about their behav-
ioral tendencies. Why should such models be exchanged for one that appears to
write dissonant harmonies out of existence?
Now is the moment when I hope to reward readers for their patience concern-
ing these matters. I propose here not a unified theory of chordal dissonance but
rather a set of responses that extend pan-triadic theory in several orthogonal
directions. This flexibility is consistent with the precedent of classical theory.
Jean-Philippe Rameau, the theorist for whom dissonant harmonies first arose
as categorical possibilities, generated them through a series of ad hoc, even
mutually incompatible, methods (Christensen 1993, 98–100). Rameau’s various
approaches can be classified into four strategies—deletion, reduction, substitution,
and combination—all of which endured in the nineteenth century and three of
which survive into present-day harmonic theory and pedagogy. It is important
to distinguish here between a strategy, as a general set of assumptions, and the
several methods by which it might be implemented. The methods that I propose
are distinct from Rameau’s, which in turn are distinct from the familiar methods
of modern harmony texts.
In the Traité of 1722, Rameau principally generated dissonant harmonies from
a consonant triad by adding one or more thirds above its fifth or below its root. The
task for the analyst is to identify the consonant triad en route toward locating its
root. This procedure is simple if the triadic subset is unique, as with dominant and

139
140  Audacious Euphony

half-diminished sevenths and minor ninths. But with diminished seventh chords,
which include no consonant triads, or with major and minor sevenths or major
ninths, which include several, the process does not yield determinate solutions,
and Rameau addresses them through a set of ad hoc responses to be considered
shortly. This problem was obviated around 1800, when diminished and augmented
triads were added to the roster of basic sonorities (Damschroder 2008, 17–19).
Any stack of thirds could now be generated upward from a root, and any seventh
or ninth chord could be unambiguously reduced to its triadic base, be it dissonant
or consonant. What was retained through this change of method was the strategy
of reduction to a subset, a strategy that continues to underlie the ubiquitous prac-
tice of labeling dissonant harmonies as extensions of the lowest three members of
their third-stack.
In his ad hoc responses to harmonies lacking a unique root, Rameau put two
other strategies into play. The Traité interpreted the diminished seventh chord as
resulting from the displacement of the root of a dominant seventh chord, thereby
appropriating to harmonic theory a substitution strategy familiar from theories of
counterpoint and of rhetorically based melodic analysis (Christensen 1993, 100).
Substitutional strategies were subsequently invoked in Johann Philipp Kirnberger’s
notion of the accidental seventh chord, whose dissonant tone substitutes for its
consonant resolution; by François-Joseph Fétis (2008 [1844]), who considers chro-
matic passing and neighboring chords to be substitutes for, and intensifiers of,
diatonic ones; and in harmonic pedagogy influenced by Heinrich Schenker.1
Rameau advanced an alternative approach to diminished seventh chords in his
1737 Génération harmonique, where he generated them from the combination of
two triads, the major dominant and minor subdominant, each of which is repre-
sented by only two of its constituent tones. A similar strategy is evident in his 1760
Code pratique, which viewed major and minor seventh chords as combining two
complete triads. Although generation of dissonant harmonies as the combination
of consonant ones is only occasionally sighted in contemporary harmonic theory
(e.g., Straus 1982; Harrison 2002a), it was quite common in the nineteenth cen-
tury: it underlay Moritz Hauptmann’s and Hermann von Helmholtz’s analyses of
major and minor sevenths, Helmholtz’s analysis of minor triads as clouded conso-
nances representing three distinct major triads, Hugo Riemann’s theories of dis-
sonant harmony (Gollin 2011), Georg Capellen’s 1908 theory of hybridization
(Doppelklang), and Schenker’s analysis of the French sixth chord.2
The fourth historical strategy, deletion, is implicit in eighteenth-century mod-
ulatory theory, which interprets one level of a composition’s structure in terms of
a succession of keys that stand in one-to-one relationship with the eponymous
consonant triads. Thus, a succession of keys lends itself to interpretation (if not yet
representation) as a succession of individual harmonies. This conception becomes

1. For example, Aldwell and Schachter 1989, Gauldin 2004. Schenker, in his later writings, denied
harmonic status to substitutional dissonances, but modern Schenkerians have conceded this point,
perhaps in deference to the universal and evidently irreversible harmony centricity of music theory
pedagogy in the academy.
2. Hauptmann 1888 [1853], 67; Helmholtz 1885 [1877], 294, 341; Gollin 2011; Bernstein 1993, 89;
Schenker 1954 [1906], 278.
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  141

fully implicit in conjunction with Rameau’s theory of imputed dissonance, which

held that only tonic triads are consonant, all others bearing an acoustically sup-
pressed sixth or seventh (Christensen 1993, 129). The implication is that a triadic
progression is coextensive with its progression of tonics. If one wishes to model a
composition as a succession of consonant triads, one’s job is not to assign triadic
constituencies to dissonant harmonies, whether their dissonance is acoustically
realized or virtually imputed, but rather to delete them from the analysis. Rameau’s
views on the relationship of modulation and consonant succession were adopted,
although inflected somewhat differently, by Riemann (Mickelsen 1977, 66; Gollin
2000, 236). Schenker adopted a different version of the deletion strategy by stipu-
lating that consonances alone are capable of prolongation. Unlike Rameau or
Riemann, it was immaterial to Schenker whether a consonant triad carried a local
status of tonic, as he rejected the notion of local tonics altogether (Schachter 1987).
What is significant is Schenker’s treatment of nonadjacent consonant triads as if
directly adjacent, by suppressing the dissonant harmonies between them. More
significant yet is Schenker’s translation of this strategy into a mode of representa-
tion that resembled a musical score: his graphs represent nonadjacent consonant
triads as register-specific pitches that participate in linear strands, or voices, whose
coordinated behavior is available for study by analogy with the behavior of voices
in a direct succession of consonances such as one might find in a part-song or
chorale.
The four strategies play different roles in this chapter. I shall say little about
deletion, which tacitly underlies many analyses already presented and to which
Anglophone readers are habituated by Schenkerian practice. A second strategy
that has already been in evidence is the reduction to a subset, which I have implic-
itly invoked whenever I have considered a dominant seventh chord to represent
a major triad, ignoring its seventh, or, more controversially, considered a half-
diminished seventh chord as a minor triad, ignoring its putative root. The second
part of this chapter develops this way of thinking about dissonances that have a
unique consonant subset.
The third part develops the substitution strategy in a way that inverts Rameau’s
method. Taking a cue from Benjamin Boretz’s analysis of the Tristan Prelude, it
interprets dominant sevenths, as well as their half-diminished inversions, as semi-
tonal displacements of fully diminished seventh chords. The two chord types con-
stitute a system of nearly even tetrachords, and motion between them benefits
from the voice-leading efficiency that accrues to such a system. Many, but not all,
of the terms, concepts, and modes of representation that have been developed for
triads can therefore be adapted for the tetrachordal case.
Although the fourth strategy, combination, has not figured heavily in my
thinking, it is worth taking a moment here to briefly sketch some ways that it
might be integrated it into the theory developed here. The analytic output of that
theory has been communicated, to a large degree, by means of the Tonnetz, which
represents triads and their progressions in compact, efficient, and legible ways.
Those qualities do not, for the most part, transfer to dissonant harmonies, most of
which are represented on the standard Tonnetz by shapes that are irregular, sprawl-
ing, or even disconnected. Edward Gollin (1998) solves this problem by extending
142  Audacious Euphony

Figure 7.1. Franz Schreker, Kammersymphonie, at rehearsal 13.

the Tonnetz into a third dimension, according the interval of the minor seventh
its own axis, and representing dominant and half-diminished seventh chords
as tetrahedra. This ingenious solution is limited, as a mode of exploration and
communication, by the dimensional constraints of the printed page.
There is, however, a genus of dissonances that do benefit from compact and
determinate locations on the Tonnetz. These are the tetrachords that combine two
edge-adjacent triads, forming a parallelogram from their two triangles. Of the
three dissonances generated in this manner, two are familiar, and their generation
by combination was identified by Rameau 1760 and Hauptmann 1853. The chord
of the major seventh is formed by combining two complete L-related triads to
form a parallelogram bisected by their shared minor-third edge. Similarly, the
chord of the minor seventh is formed by R-related triads straddling a major-third
edge. The third harmony, which does not arise to prominence until the twentieth
century, is the “split third” [0347]-type tetrachord formed by the union of two
P-related triads. This is the “alpha” sonority that Lendvai 1971 locates prominently
in the music of Bartók. To briefly suggest its descriptive potential, figure 7.1(a)
presents a brief progression from Franz Schreker’s Kammersymphonie of 1916 (see
Harrison 1994, 22). The Cᅊ split third fills a gap that is left along the major third
axis by the juxtaposition of the two previous triads, the hexatonic poles A major
and f minor, with both of which it shares two common tones. Figure 7.1(b)
indicates this hybrid chord as a parallelogram.

Reduction to a Triadic Subset

The strategy and the method are Rameau’s: a dissonant harmony reduces to a con-
sonant triad when the latter is uniquely contained in the former. Figure 7.2 shows

Figure 7.2. Six chords with uniquely embedded consonant triads.

 CHAPTER 7 Integrating Dissonant Harmonies into the Model  143

how this method assigns analyses to six chords, presented in inversionally related
pairs. Standard third-stacking theory roots all six chords on B, but only the first
chord of each pair is so rooted according to the approach pursued here. Concerning
the rare sonorities presented here as (d) and (f), there is little received opinion
against which the analysis proposed here can deeply cut. But the analysis of (b) as
d minor chord with added sixth or under-seventh, rather than as a half-
diminished seventh chord rooted on B, is a hard sell in a musical culture
habituated to the view that root generation by stacked thirds is bestowed by nature,
if only implicitly via the standard names for chords. Carl Dahlhaus writes that
“the notion that B is the ‘characteristic dissonance’ of the subdominant seventh
chord b-d-f-a is frustrated by musical reality, in which it is evident that a is the
chord’s actual dissonance” (1990 [1967], 56), and Harrison 2011 suggests that the
recourse to subposed dissonances in such cases exemplifies “analytic difficulties,”
presumably by virtue of the prima facie untenability of the concept.
Although the analysis of figure 7.2(b) as d minor was first suggested by Rameau,
it is most frequently associated with harmonic dualism, as developed by Arthur
von Oettingen and Riemann in the second half of the nineteenth century. For
Riemann, the DFA chord is generated downward from A. When B is added to that
chord, it functions as a characteristically dissonant under-seventh (Rehding 2003,
94–96). The interpretation of the half-diminished seventh chord as a minor triad
with under-seventh or added sixth was hardy enough to survive well into the
twentieth century, even among theorists otherwise uncommitted to a dualist
theory of harmony (see chapter 3, note 6). Ultimately, that interpretation was over-
come by a renewed third-stacking monolithy, to the point that its peremptory dis-
missal became self-evident. Recent empirical work has motivated a revival of
Riemann’s proposal, grounding it in perceptual experiments rather than physical
or metaphysical speculation (Parncutt 1989, 149).
A consideration of some passages from Wagner will suggest that late-nineteenth-
century dualists might have been partly motivated by an empiricism of a different
sort. Figure 7.3(a) presents the final cadence from Tristan und Isolde, connecting a
minor subdominant to a major tonic. On its final beat, the subdominant receives
an added major sixth in the highest voice, temporarily creating a “ø7” sonority. But
the “seventh” of that chord, B, does not resolve, nor does it feel any pressure to do
so. It is the putative root, Cᅊ, that is transient: as an under-seventh, it passes step-
wise upward to the third of the resolution chord, just as a dominant seventh typi-
cally behaves, but in the reverse direction. The minor-to-major plagal progression
with rising under-seventh, which is a mirror image of the major-to-minor authen-
tic progression with falling seventh, has its origins in the sixteenth century and
serves as a final cadence in Chopin, in Brahms, and in classic film music, among
other places.3
A dismissal of this Cᅊ as naught but a passing tone courts the danger of
missing something, in an opera that derives so much expressive bang from the
3. Thomas Tallis’s Third Tune for the Psalter of Archbishop Parker (1557, familiar from the Vaughan
Williams Fantasy) concludes with such a progression: a minor → E major, with passing cantus Fᅊ.
A similar progression is evident a century later in Monteverdi’s Incoronazione di Poppea (act 1,
scene 12, cadence of the second quatrain).
144  Audacious Euphony

Figure 7.3. Under-sevenths in Wagner.

syntactic possibilities of the ø7 chord. Act 3 of Tristan (figure 7.3(b)) opens with a
bᅈ minor triad to which is appended a G. There is little temptation to hear this as
a passing formation, as the G does not evidently issue from F, and it endures for
three very slow beats. Figure 7.3(c) shows a similar example from Parsifal. Here
the foreign tone (again a G) is heard as an outlier by virtue of registral distribution
and orchestration: bᅈ minor sounds in the brass, whereas G is registrally and tim-
brally isolated in the timpani and pizzicato strings. Moreover, by Parsifal’s third
act, we have often heard the horn music, affiliated with the title character.
Characteristically that motive is fast, loud, major, and consonant in keeping with
its brash protagonist. Here it is slow, muted, minor, and muddied by dissonance, to
suggest Parsifal’s desultory wandering.
Wagner often treats minor triads and their ø7 supersets as interchangeable in
his late music. In figure 7.3, (d) and (e) present two versions of the Communion
theme from Parsifal, the first triadic, the second with a timbrally isolated under-
seventh. Wagner’s treatment of the Tarnhelm motive in the Ring tetralogy is analo-
gous. In its initial and most characteristic form (figure 2.4) it juxtaposes gᅊ minor
and e minor, but Wagner frequently adds an under-seventh to the initial chord
(Rothfarb 1988, 154; Lewin 1992). The final two examples, figure 7.3, (f) and (g),
juxtapose the beginnings of two parallel phrases from Parsifal; but for their initial
bass pitch, the phrases are identical to within transposition. Each of these
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  145

examples suggests that Wagner’s musical reality stretched beyond the limits that
frustrated Dahlhaus’s putative listener.

Hexatonic Poles in Parsifal

I now bolster the claim advanced in the previous section by exploring the case of
hexatonic poles in Parsifal.4 The opera’s first chromatic progression occurs at m. 28
of the Prelude, when Aᅈ major → e minor deforms the diatonic theme associated
with the ritual of Holy Communion. Its last chromatic progression occurs eighteen
measures before its final measure, when Kundry is “de-souled” to a hexatonic dis-
tortion of the music associated with the Holy Grail (similar to figure 5.25(b)).
Hexatonic poles similarly bookend the second half of act 2. Kundry opens that
scene by twice singing Parsifal’s name, once arpeggiating cᅈ minor (m. 739), the
second time its hexatonic pole, Eᅈ major (m. 751). She closes her role in the scene
by twice cursing Parsifal to a life of perpetual wandering; her two iterations of
“Irre” are set to d minor and Gᅈ major. Twelve measures later, Klingsor hurls the
spear in bᅈ minor and Parsifal catches it in D major.
Hexatonic poles also play a role in the large-scale tonal structure of each of
Parsifal’s three acts. The first act begins in Aᅈ major, and returns to that key for the
Communion service near the end of the act. Its hexatonic pole, e minor, opens the
Amfortasklage, the psychological climax of the act. The second act is tonally closed
in b minor; its pole, Eᅈ major, is the principal key of the flower maidens’ music.
The third act begins in bᅈ minor, reaches D major at the Good Friday meadows
music (m. 676), and returns to bᅈ minor with the choral music that leads into the
final scene (m. 862). That final scene begins in e minor (m. 918), and culminates
in the Aᅈ major that opens the shrine (m. 1088) and closes the opera. The four
structural keys of act 3 retrograde the Grail-theme distortion at figure 5.25(b).
Hexatonic poles can thus be said to have motivic value in Parsifal, in the
dual sense of unifying the opera and marking its individuality. This abstract
idea of motive is more consistent with twentieth-century uses of the term by
such theorists as Schenker, Rudolf Réti, and Hermann Keller (surveyed in Cook
1987; Dunsby 2002) than with the well-known nineteenth-century Leitmotif tradi-
tion of Wagnerian analysis, where motives depend for their identity on melodic,
rhythmic, and/or timbral features.
This motivic network is expanded considerably by taking into consideration
hexatonic poles whose triadic constituents are uniquely embedded into dissonant
formations. Consider the two passages excerpted in figure 7.4, the first opening
Amfortas’s lament, the second associated with Parsifal’s memory of his mother.
Both passages involve hexatonic poles, with their characteristic contrary semitonal
motion in three voices. In the Amfortasklage, dominant Gᅊ major is accompa-
nied by an over-seventh. In the Herzeleide motive, subdominant cᅊ minor is

4. An expanded version of this Parsifal analysis is found in Cohn 2006; some of its ideas originate in
Cook 1994.
146  Audacious Euphony

Figure 7.4. Hexatonic poles in Parsifal, with supplementary dissonance.

accompanied by an under-seventh. A stacked-third approach to harmonic struc-

ture might grant the hexatonic pole in the first case and admit it as a member of
the motivic network documented above but would deny it in the second case on
the basis that the root progression is different.
And what is the reward for seeing a new motivic connection? Music theorists
treasure them in themselves, because they allow us to see a composition as a tightly
knit and unified web. But to stop there is to invite the charge of complicity in (or
worse, fetishization of) the ideology of unity. There it is again, and there it is again,
and again . . . who cares? There’s a good response to this question, but it is rarely
made explicit: the value of motivic connections is not only what they are but what
they allow the analyst to do. To see a connection is to open a door; one still needs
to walk through it.
In order to do so, we place our observations about hexatonic poles into an
interpretative framework that engages the core of the Parsifal story. The opera’s
two principal Leitmotive, one associated with a ritual (Communion) and the other
with a solid object (Grail), are theologically related through the blood of the savior,
which the latter contains and the former symbolically transmits into the Christian
body. The central problem of the opera is that Amfortas is unable to perform the
Communion ritual because the process of ingesting the blood inflicts agony on his
wounded body beyond his ability to bear it. Object and ritual are thus bound
together with fluid and sensation into a causal chain: the Grail is the source, the
Communion ritual the agent, the blood transmission the action, and the agony of
sin the result.
Admitting that consonant triads may be embedded into dissonances enables
us to see that Wagner uses hexatonic poles in Parsifal not only to deform the
Communion and the Grail but also to depict the other two components of this
schematic knot, the blood and the agony. Figure 7.5(a) sets the first textual refer-
ence to the pain of Amfortas. At the inception of the text (“The pain soon returned,
but more intensely”), V 24 in d minor (over a tonic pedal) is displaced by iiø7 in
c minor. Each dissonant chord embeds a unique consonant triad, A major and
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  147

Figure 7.5. The pain and the blood in act I of Parsifal.

f minor, whose connection executes the semitonal displacements characteristic of

the hexatonic pole. The same two triads are concealed in the music of figure 7.5(b),
which occurs at the first mention of the Holy blood in act 1 and each of the three
times that the blood is referred to in act 3.
The four-stranded knot is drawn most tightly in the final scene of the first
act, when the Grail is uncovered and the Communion ritual is performed.
Amfortas, “shot through with pain,” imagines the savior’s blood surging out of the
chalice, entering his body through his wound, flowing through his heart, and min-
gling with his own blood “defiled by shame.” The music most characteristic of
Amfortas’s lament, the Sündenqual (“pain torment,” figure 7.3, (f) and (g)), receives
its first explicit verbal association just after Amfortas enters the scene, to choral
singing of “Den sündigen Welten, mit tausend Schmerzen, wie einst sein Blut
geflossen” (“as once his blood flowed, for the sinful world of a thousand pains”).
Figure 7.6 analyzes the Sündenqual progression as a sequence of hexatonic poles
concealed behind a persistent descending fifth motion in the bass that feigns
a red-herring tonality (see Lendvai 1988, I: 142–43). The dominant and ø7
chords paired in each leg of the sequence are those of the “pain” progression at
figure 7.5(a).
The four components of Parsifal’s central problem are thus bound by a single
musical signifier, the H progression. Far from arbitrary, this signifier relies on a
homology between the structures of hexatonic poles and of the circumstances that
148  Audacious Euphony

Figure 7.6. Parsifal, act 1, mm. 1369–71 (“torment of sin” motive, P/V score, p. 83).

they portray. At many levels, Parsifal straddles the border between life and death.
Christian doctrine of transubstantiation holds the blood to be alive, even though
its organic source, the redeemer, has been dead for a millennium. This is the logic
of the uncanny, as theorized by Ernst Jentsch and Freud: what ought to be dead is
unaccountably living. This same logic is played out in four of the opera’s six named
characters. Amfortas teeters on the brink of death, and ancient Titurel inhabits a
tomb. Kundry, shielded from death by a curse, is older yet; Wagner describes her
death as the de-souling (Entseelung) of a zombie. And Klingsor conjures botanical
abundance from a desert wasteland: from death springs life.
A homologous logic is played out in the hexatonic pole progression, with the
life/death duality mapped onto consonance and dissonance. When a consonant
triad progresses to its hexatonic pole, its root is displaced down to the raised
seventh degree, while its fifth is displaced upward to the flatted sixth degree.
The resulting interval ought to be a dissonant diminished seventh. But if we
perceive this new chord as a triad, then we are perceiving the resulting interval as
a consonant major sixth. What ought to be dissonant is unaccountably consonant,
in a dynamic that I sketched in chapter 2 (see pp. 21–22). Alfred Lorenz writes, of
figure 5.25(b), that “during the lingering on the notes that are initially understood
as dissonant, the chord cleanses itself, without any motion, into the most radiant
beauty” (1933, 89).
There is much more that could be said about Parsifal from this standpoint
(see Cohn 2006), but my primary purpose here is not to interpret that opera.
The topic is pertinent to this chapter insofar as it illustrates the analytic and inter-
pretative profit that can be made available by viewing dissonant harmonies in
terms of their consonant-triadic subsets. This analytic move broadens the motivic
network, enriching our understanding of how Parsifal’s musical relations organize
and interpret its web of extramusical symbols. And these results, to the extent that
they satisfy our desire to interpret this most complex of nineteenth-century
masterpieces, help to justify the approach toward analysis of dissonant harmonies
advocated in this section.

The Tristan Genus as Nearly Even Tetrachord

Reduction to a consonant subset, however, hardly provides a universal solution for

all dissonances that contain a unique consonant subset. Consider the common
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  149

Figure 7.7. A low-work tetrachordal progression embedding a high-work triadic

progression.

classical prototype at figure 7.7(a). Taken on its own terms, the progression involves
only two units of voice-leading work: two upper voices descend by semitone;
two lower ones are stationary. The progression between its constituent triads at
figure 7.7(b), however, involves five units of work, the maximum possible for two
triads of opposite species. The “supplementary” dissonances smooth the voice
leading of a progression that is otherwise disjunct. These considerations provide
a motivation for seeking an alternative approach, one that takes dissonant harmo-
nies as they are, rather than reducing them to something simpler and more familiar.
This section responds by advancing a model of dissonant harmony based on an
analogy with the triadic case, rather than an extension of it.5
Nineteenth-century music contains many compositions, or extended passages,
whose surfaces are dominated not by consonant triads but rather by dominant
(“V7”) and half-diminished (“ø7”) seventh chords, or their enharmonic equiva-
lents. Figure 7.8 presents three sequential templates. Figure 7.8(a), which involves
only V7 chords, arises already in the eighteenth century. Embellished by chromati-
cized voice exchanges, the progression becomes known as the “omnibus” or “devil’s
mill” (Wason 1985; Telesco 1998). Figure 7.8(b) and (c), which alternate V7 and ø7
chords, also serve as occasional late-eighteenth century templates and are fre-
quently elaborated in music of Chopin (Tymoczko 2011b, 284–92).6 Each of the
three sequences transposes by an odd value and features double semitonal motion
against two stationary voices. But there is a difference: the first sequence features
contrary motion; the remaining two, similar motion. Thus, figure 7.8(a), which
transposes V7 chords by minor thirds, shares some of the anomalous features that
accrue when triads are transposed by major thirds (Tymoczko 2011b, 97). As we
shall soon see, this is related to the structure of V7 chords: the minor third has this
special status because it divides the octave into as many equal parts as a seventh
chord has tones. (This formulation is italicized to emphasize its parallelism to the
one on p. 19 of this book.)
Not all passages that harness these parsimonious features are sequential.
In Tristan und Isolde, Wagner famously blew open the gates to a compositional
field cohabited by V7 and ø7 chords whose interactions were disciplined neither

5. This analogy was initially reported in Cohn 1996 (40n39). Its details were initially worked out in five
papers presented at Buffalo conferences in 1993 and 1997 (Lewin 1996; Callender 1998; Childs 1998;
Gollin 1998; Douthett and Steinbach 1998) and are adapted and transformed in broadly synthetic
writings of Jack Douthett (2008) and Dmitri Tymoczko (2006, 2011b). The work presented in this
section is indebted to all of these writings.
6. Tymoczko (2011b, 284–93) presents a number of examples of seventh chords in T11 and T5 sequences,
as well as in T2 and T8 sequences with which they share properties.
150  Audacious Euphony

Figure 7.8. Three sequential templates for Tristan-genus seventh chords.

by tonality nor by patterned sequential repetition. Succeeding generations of com-

posers romped on this Tristan field for a half-century and more, using semitonal
voice leading to connect ø7 chords directly to each other, and to V7 chords, in a
manner that defied diatonic/tonal coherence more frequently than not.7
The coast of Cornwall is littered with the detritus of ill-equipped music-theo-
retic vessels that perished in search of a suitable tonic at which to moor.8 After
touring the wreckage, Benjamin Boretz approached those shores with a different
conveyance. Figure 7.9 presents the opening six chords of the Prelude in three
pairs, each of which consists of a ø7-type chord linked to a dominant seventh.
Of the first pair, Boretz writes that “these two chords . . . share a common relation
to . . . D–F–Gᅊ–B; each . . . contains just three of its four pitch classes, with one
pitch ‘contrapuntally’ displaced by a semitone; for only the Dᅊ ‘spoils’ the first
chord of m. 2, and when it ‘resolves’ to D, the F of the complex is ‘displaced’ to E”

Figure 7.9. Prelude to Tristan und Isolde, opening measures. Each filled notehead
is a “spoiler” that lies outside of the majority diminished seventh chord.

7. Bass 2001 provides a number of examples of parsimonious motion among ø7 chords in

early-twentieth-century music.
8. The history of Tristan analyses is told in Motte 1976, Wason 1985, Bailey 1985, and Nattiez 1990.
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  151

(1972, 163).9 Boretz’s observation is striking in light of its affinity to two distinct
episodes in the history of music theory. Like Rameau in the Traité, Boretz claims that
fully diminished and dominant seventh chords stand in a substitutional relation. But
Boretz reverses Rameau’s generative arrow. Whereas Rameau values the dominant
seventh as progenitor by virtue of its relative consonance, Boretz bestows primacy
on the diminished seventh by virtue of its status as an equal division of the octave.
The second affinity is with Carl Friedrich Weitzmann’s Der übermässige Dreiklang:
Boretz’s reversal, for the case of tetrachords, is analogous to Weitzmann’s reversal for
the case of major and augmented triads, as discussed at the end of chapter 3.
Nor does the analogy end there. The perfectly even diminished seventh chord,
like the perfectly even augmented triad, may be displaced upward as well as down-
ward, yielding a second species of nearly even tetrachord: the half-diminished as
well as dominant seventh chord. Together, these two species of chords, which
I shall designate the Tristan genus, forms a parsimonious voice-leading system
analogous to the pan-triadic system.10
The following exposition develops the system of nearly even tetrachords by
analogy with the trichordal case, up to the limits of that analogy’s productive
power. Table 7.1, which aligns analogous terms from the two domains, can serve
as a guide to this development, as it unfolds in the pages ahead. Because the anal-
ogy relies on generalization, the exposition is more technical and abstract than
most other parts of this book.
Boretz’s observations about the remaining chord pairs in the opening of the
Tristan Prelude already mark out some of the lines along which this analogy devel-
ops. The description that he provides for chords (1) and (2) (with reference to
figure 7.9) applies with equal force to chords (3) and (4). The only change is in the
identity of the tones that “spoil” the BDFGᅊ chord: Dᅊ and E in the first pair, and
Fᅊ and G in the second. The same description likewise applies to chord (5), where
C is the spoiler. But it does not apply to chord (6), whose B root is the single
remnant of the former complex and now plays its own role as spoiler of a new

9. Boretz’s 1970 dissertation, “Metavariations,” was published in installments in volumes 8–11 of

Perspectives of New Music. The Tristan analysis appears in volume 11, no. 1 (1972), to which my
page references apply. Metavariations was republished by Open Space Press in 1995. Related views
of the Tristan Prelude are elaborated in Bass 2001 and Chafe 2005, and of related passages from the
opera in Lerdahl 2001, 302. See also Morgan 1976, on the Prelude to act 3 of Parsifal. Only Chafe
acknowledges, en passant, the powerful precedent of Boretz’s analysis.
10. This is not the Tristan constellation of hexachords developed in Soderberg 1998. The Tristan genus
is equivalent to set-class 4–27, prime form [0258], a system of classification that honors inversional
equivalence. That equivalence relationship is a by-product here, as in the triadic case discussed in
chapter 2 (see p. 38); what is primary is the relationship of single semitonal displacement that this
group of chords holds with respect to the equal division of its cardinality.
In using traditional names for seventh chords, I do not wish to take on board the functional
agency that classical theory bestows on them. Nor do I wish to deny the force of this agency, which
is metaphorically figured variously as energy, appellation, attraction, magnetism, charge, desire,
and so forth, but only to neutralize it at this stage of theoretical development. It is not a necessary
feature of Tristan-genus chords, to the same degree, in each and every one of its manifestations in
the universe of compositions that use them. For the same reason, despite my reservations about the
universal value of third-stacking, I will continue to take advantage of its ability to furnish familiar
and unambiguous chord labels.
152  Audacious Euphony

Table 7.1. Comparison of nearly even chords of cardinality three (triads)

and of cardinality four (Tristan genus)
Genus Species, n = 3 Species, n = 4

(1) Perfectly even n-note chord Augmented triad Diminished seventh

chord
(2) Downward displacement Major triad Dominant seventh
of (1) chord
(3) Upward displacement Minor triad Half-diminished
of (1) seventh chord
(4) Union of (2) and (3) Consonant triads Tristan genus
(5) Set of displacements of an Weitzmann region Boretz region
instance of (1)
(6) Geometric representation Weitzmann water bug Boretz spider
of (5)
(7) Bridges between Hexatonic region Octatonic region
adjacent (5)’s
(8) voice-leading zones 1 2 4 5 7 8 10 11 1 3 5 7 9 11
(9) Transformations within (5)
(a) One voice by whole step R R*
(b) Two voices by semitone N, S S3(2), S3(4), S6
(10) Transformations within (7)
(a) n – 2 voices by semitone L, P S2, S4, S5
(b) n voices by semitone H Octatonic pole
(“maverick”)
(11) Voice-leading map of Cube Dance 4-Cube Trio
unified system Tonnetz 3-D Tonnetz
(Gollin 1998)

diminished seventh chord complex centered on ACDᅊFᅊ. Boretz interprets the

shift between majority diminished seventh chord complexes as the first of several
interregional modulations that, he argues, provides the Prelude with its harmonic
scaffold (169).

Boretz regions

The triadic case presents a model for fashioning Boretz’s observations into a sys-
tematic framework for exploring voice-leading relations among members of the
Tristan genus. More than a century earlier, Weitzmann placed each nearly even
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  153

(consonant) triad adjacent to the perfectly even augmented triad that it minimally
displaces. The six consonant triads that cluster about an augmented triad, by virtue
of their mutual adjacency to it, are second-order adjacent to each other. Chapters 3
and 4 explore how those clusters form a tight voice-leading system and propose that
the augmented triad to which they are adjacent asserts a virtual power over that
system even when absent from the musical surface. By analogy, a consequence of
Boretz’s observation that eight members of the Tristan genus minimally displace a
fully diminished seventh chord is that those eight chords are second-order adjacent
to each other. Thus they, too, form a tight voice-leading system even when, as in the
Tristan Prelude the diminished seventh chord is absent from the surface.11
As the first five chords of the Tristan Prelude are single semitonal displace-
ments of a single equal division, BDFGᅊ, we will say that they share membership
in a Boretz region, a tetrachordal analogue to the Weitzmann regions of triads. This
region contains eight members: the four ø7 chords that displace B, D, F, or Gᅊ
upward, and the four V7 chords that displace one of those tones downward. The
two remaining diminished seventh chords have their own Boretz region, each with
a cluster of eight Tristan-genus chords. Table 7.2 classifies the members of the
Tristan genus by Boretz region (compare with table 3.1).

Boretz spiders

Figure 7.10 models voice leading within the third Boretz region, which furnishes
the initial five chords of the Tristan Prelude.12 As with the Weitzmann water bug

Table 7.2. The contents of the three Boretz regions, modeled after
Weitzmann’s grouping of his four triadic regions (table 3.1)
I. {C, Eᅈ, Fᅊ, A} and its enharmonic transformations
1. Ddom7 2. Fdom7 3. Aᅈdom7 4. Bdom7
5. Eᅈø7 6. Fᅊø7 7. Aø7 8. Cø7
II. {Cᅊ, E, G, Bᅈ} and its enharmonic transformations
1. Cdom7 2. Eᅈdom7 3. Fᅊdom7 4. Adom7
5. Cᅊø7 6. Eø7 7. Gø7 8. Bᅈø7
III. {D, F, Aᅈ, B} and its enharmonic transformations
1. Cᅊdom7 2. Edom7 3. Gdom7 4. Bᅈdom7
5. Dø7 6. Fø7 7. Gᅊø7 8. Bø7

11. In Tristan, the virtual presence of the diminished seventh chord is corroborated by Wagner’s initial
sketch of the Prelude (Bailey 1985, 131). Moreover, diminished seventh chords often replace
Tristan-genus chords in contexts where the latter normally appear, as in m. 68 of the Prelude
(Mitchell 1967, 190–91).
12. Tymoczko 2011b, 371, has a formally identical graph but with the additional feature that axes
correspond to motion in the three musical voices.
154  Audacious Euphony

Figure 7.10. A Boretz spider.

(figure 4.1), any two constituent chords are separated by two units of voice leading.
When tetrachords are of the same species, as in figure 7.8(a), voices move in con-
trary motion, as the resolution of the “spoiler” into the diminished seventh body is
offset by the displacement of a different voice out of that body, on the same flank.
When they are of opposite species, the spoiler resolves into the body and a new
spoiler emerges on the other side. The semitones are either distributed between
two different voices, as in chords (1) → (2) of figure 7.9, or concentrated into a
single voice moving by whole step, as in chords (2) → (3).
Back-and-forth motion across a Boretz spider, as across a Weitzmann water
bug, toggles downshift with upshift, balancing between adjacent zones in voice-
leading space. Figure 7.11 presents the Tristan progression’s idealized voice lead-
ing. Arrows indicate direction of motion. The first five chords alternately upshift
and downshift, moving back and forth across the Boretz spider and balancing
between adjacent zones in voice-leading space. Zone labels beneath the score,

Figure 7.11. Idealized voice leading of the Tristan Prelude opening, with
transformations labeled above and voice-leading zones indicated below.
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  155

derived via the summing protocol presented in chapter 5, indicate an alternation

between zones 3 and 1, balancing about the BDFGᅊ diminished seventh chord,
whose corresponding integers sum to 2.13

Boretz-group transformations

The transformational labels above the score in figure 7.11 are adapted from Childs
1998 and Douthett and Steinbach 1998. “S” here stands for similar motion (it no
longer refers to Slide, as in the triadic case). The two superscripts classify the inter-
val of the stationary and moving dyad, respectively. In the initial progression, the
stationary BGᅊ and moving FDᅊ belong to interval classes 3 and 2, respectively,
hence the S3(2) label. R*, like triadic R, moves a single voice by whole step. In addi-
tion to the three cross-species Boretz transformations used in this excerpt, there is
a fourth, S6, which holds the tritone dyad and moves the interval class 5 dyad; we
will encounter this in figure 7.14 below.

Octatonic regions

Regions that combine the half-diminished seventh chords of one system with the
same-rooted dominant seventh chords of the neighboring system form bridges
between adjacent Boretz regions. As the eight chords of this bridging region draw
their tones from an octatonic scale, they are referred to as octatonic regions. Their
role with respect to the Boretz regions is exactly analogous to the role of the hexa-
tonic regions with respect to the Weitzmann regions, although their individual
transformations lack the minimal work of the hexatonic transformations L and P.
The progression from chord (5) → (6) (in figure 7.9) is one of four octatonic trans-
formations that can bridge adjacent Boretz regions.

Octatonic-group transformations

Adrian Childs (1998) dubs the particular transformation from chord (5) to chord
(6) an octatonic pole because it plays the same maverick role as the hexatonic
pole progression does in triadic space. Sigfrid Karg-Elert’s 1930 treatise identified
their affinity, calling them both Kollektivwechsel, or “collective exchange” (285).
Figure 7.12 documents the analogy: both progressions combine inversionally
related species; involve upshifting in all-but-one voice, offset by downshifting in
the remaining voice; and use all of the tones of their eponymous scales.14

13. The labels for zones acquire meaning only within the limited context of the tetrachordal system of voice
leading. Some of the labels are identical to those used for voice-leading zones in triadic space. These are
“false friends”: the triadic and tetrachordal systems are not in communication with each other.
14. Lendvai (1983, 510; 1988, 139) identifies prominent octatonic poles in the Ring and in Boris
Godunov. Cohn 1996, 26–28, observes that hexatonic and octatonic poles are directly juxtaposed in
the first movement of César Franck’s Piano Quintet. See also Cook 2005.
156  Audacious Euphony

Figure 7.12. Hexatonic and octatonic poles.

Figure 7.13. The four octatonic-bridge transformations from Dø7.

The three remaining octatonic transformations, presented in figure 7.13, share

the feature of double semitone voice leading with the Boretz-region transforma-
tions. Adapting from Childs’s labeling protocol, these S transformations bear only
a single superscript, indicating the interval classification of the stationary dyad.
The octatonic transformations lack the minimal-work feature of their hexatonic
analogues because their chords contain more tones. After the held dyad is
accounted for, there are two voices, not one, that are left to migrate between equal
divisions, tipping the majority status from one to the other.

Brünnhilde’s Immolation

Lewin 1996 observes that the opening music of Brünnhilde’s Immolation

(Götterdämmerung, act 3) follows a voice-leading script similar to the opening of
the Tristan Prelude. Figure 7.14 presents the harmonies, beginning with the tempo
change (p. 318 of the Schirmer piano/vocal score) and extending through the first
ten measures of Brünnhilde’s aria. As with the Tristan opening, the progression
alternates V7 and ø7 chords, here occupying the Boretz region centered on EGAᅊCᅊ.

Figure 7.14. Brunnhilde’s Immolation, from Götterdämmerung, act 3.

 CHAPTER 7 Integrating Dissonant Harmonies into the Model  157

Figure 7.15. Three Boretz spiders in union with three octatonic pools.

The octatonic bridge into a new region involves two chords rooted on Cᅊ and,
unlike the Tristan excerpt, sustains the double semitone voice leading of its prede-
cessors. The consecutive upshifts, Eᅈ7 → Cᅊø7→ Cᅊ7, enter a new voice-leading zone.
As in Tristan, this marks the moment when the seventh chords hearken to their
appellative properties after ignoring them for some measures.

4-Cube Trio

Figure 7.15 portrays the entire system of interlocking Boretz and octatonic regions.
As in the triadic case (figure 5.2), the spiders and pools are subject to figure–
ground reversals. The Tristan and Götterdämmerung passages foreground Boretz-
region membership, but one can also conceive of passages where octatonic regions
act as the primary macroharmonies, and the Boretz spiders constitute interre-
gional bridges between them.
Because the octatonic transformations involve two voices moving by semitone
in parallel motion, they jump voice-leading zones, leaving a gap that is not present
in the triadic case. That gap can be bridged using either minor seventh or French
sixth chords. In figure 7.16, known as 4-Cube Trio, both of these chord types
appear simultaneously as octatonic bridges at the 12:00, 4:00, and 8:00 positions.15

15. Douthett initially presented this figure in 1993 under the name Power Towers. Douthett and
Steinbach (1998, 256) reassigned that name to a version of the graph lacking the French sixth
chords. In a footnote (262, n. 12), Douthett and Steinbach attach 4-Cube Trio to a prose description
158  Audacious Euphony

Triangular nodes indicate minor seventh chords, and stars indicate French sixth
chords. As both fulfill the bridging function independently, either can be removed
without disconnecting the graph. The labels for these chords are omitted in
figure 7.16 to avoid clutter but may be inferred for each node from its incident
Tristan-genus chords, which occupy positions associated with the six odd
voice-leading sums. The set of nodes is completed by the spiders’ heads at 2:00,
6:00, and 10:00, representing the three diminished seventh chords.

Figure 7.16. Jack Douthett’s 4-Cube Trio.

of the original figure. (The graph links three four-dimensional cubes, or tesseracts.) Dmitri
Tymoczko (2011b, 106) independently rediscovered a version of 4-Cube Trio, which emerged as a
subgraph of his continuous four-dimensional voice-leading space. By virtue of that status, Tymoczko
shows that 4-Cube Trio, like Cube Dance, is an accurate model of voice leading among its included
chords (in the sense that the most direct edge distances between chords represent their most effi-
cient voice leadings). Peck and Douthett 2011 describes other features of 4-Cube Trio and clarifies
in what sense it is cubic.
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  159

Circumnavigating the Tristan-Genus Universe

4-Cube Trio, like Cube Dance, is a true model of voice leading (Tymoczko 2011b,
106). Voice-leading distance between any two chords is calculated by determining
the shortest path between them and counting the edges. 4-Cube Trio also serves as
a space for circumnavigation of the universe of nearly even tetrachords. As in the
triadic case, the simplest path is through transpositional sequences that perpetu-
ally upshift or downshift. If the sequence uses only members of the Tristan genus,
then only odd zones of voice-leading space are engaged. The diminished seventh,
minor seventh, and French sixth chords in the even zones have virtual status,
in the same sense that the augmented triads have virtual status in sequences of
consonant triads.
As with the triadic case, not all transpositional values induce uniform voice
leading. If two Tristan-genus chords are transpositionally related by minor third
or tritone—that is, by exactly those intervals that are internal to a diminished
seventh chord—then the voice leading is balanced, featuring contrary motion,
as in the examples from the Tristan Prelude and Brünnhilde’s Immolation
(figures 7.8(a), 7.9, and 7.14). It is the remaining transpositional values, involving
intervals absent from a diminished seventh chord, that produce sequences with
circumnavigatory powers.
Table 7.3 presents a composition table of the Boretz and the octatonic transfor-
mations. Each transpositional value is produced by combining the Boretz
transformation to its left with the octatonic transformation heading its column.
As with table 5.1, the direction of transposition is determined by the order
in which the transformations are applied and by the species of the transformed
chord. Each of the four transposition values is produced by four distinct transfor-
mation pairs. Table 7.3 thus documents sixteen distinct transformation pairs.
When a particular pair is selected, and its components presented in alternation, a
Tristan-genus sequence tours the odd voice-leading zones and circumnavigates
4-Cube Trio.

Table 7.3. Combination table for the four Boretz-region and four
octatonic-region operations, modeled after Table 5.1
S2 S4 S5 Octatonic pole
S3(2) T±1 T±5 T±4 T±2
S 3(4) T±5 T±1 T±2 T±4
S6 T±2 T±4 T±1 T±5
R* T±4 T±2 T±5 T±1
Each transposition is the product of the Boretz transformation at the head of the row and the octatonic
transformation at the head of the column. The transposition may be up or down, depending on the
order of the operations and on whether the initial triad is dominant or half-diminished.
160  Audacious Euphony

Figure 7.17. Roulade from Chopin’s Prelude, Op. 45.

In figure 7.8, (b) and (c) exemplify two of these sixteen possibilities: the T11
sequence is generated by S2 from the octatonic group and S3(2) from the Boretz
group; the T5 sequence substitutes S3(4) as its Boretz-region transformation. To
give some sense of the variety of transformation combinations available even
within a single transpositional value, compare figure 7.8(b) with figure 7.17, an
extended downshifting roulade in Chopin’s Op. 45 Prelude that circumnavigates
4-Cube Trio multiple times by alternating S4 and S3(4) (Childs 1998).
As in the triadic case, class substitutions can introduce variety while still main-
taining an overall voice-leading trajectory. Figure 7.18 illustrates with a passage
from Chopin’s f minor Mazurka, Op. 68 no. 4.16 Tymoczko (2011b, 284–86) shows
that the initial two phrases of this Mazurka execute one of many possible realiza-
tions of a script that connects a series of dominant seventh chords, each a semitone
lower than its predecessor (T11), through intermediate ø7 chords. He also shows, in
a companion analysis, that the e minor Prelude, Op. 28 no. 4, similarly connects a
series of dominant seventh chords transpositionally related by descending perfect
fifth (T5) and conjectures that the former piece is “a virtual rewriting of the latter.”
The passage at figure 7.18, which precedes the reprise of the Mazurka’s initial
phrase, supports that conjecture. The final harmonies in mm. 33–37 form a T5
chain of dominant seventh chords, as in the e minor Prelude. Locally, the progres-
sion alternates S2 and S3(4), realizing the template given earlier as figure 7.8(c).
Leading into m. 38, S4 substitutes for S2, so that Aø7 sounds in place of projected Eᅈø7.
This substitution triggers a conversion of the T5 sequence to one generated by T11,

Figure 7.18. Chopin, Mazurka, Op. 68 no. 4, mm. 33–40, with voice-leading zones
and Tristan-genus transformations indicated.

16. Any claims about this remarkable piece must be tempered by the knowledge that its text is notori-
ously unstable. Its editions, which disagree in many details, have been patched together from a
nasty set of scratchings from Chopin’s pen. See Kallberg 1985.
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  161

now realizing the figure 7.8(b) template. The semitonal transpositions continue
until Dᅈ7, the German sixth chord in f minor, is reached, at which point the reprise
begins. As with the transformational substitutions in chapter 5 (beginning with
figure 5.15), the substitutions do not affect the voice-leading trajectory, which
continues to downshift at a uniform pace throughout the passage.
Zone 1 plays a central articulating role in this six-measure passage and, indeed,
throughout the Mazurka. It is initially represented by G7, functioning as f minor’s
double dominant, in m. 33. The return of zone 1, represented by Bᅈ7 at the third
beat of m. 36, prompts an acceleration of surface rhythm, as diminished seventh
chords are interpolated on the second beat of each measure beginning at m. 37.
The next return of zone 1, represented by Dᅈ7 on the third beat of m. 39, breaks the
T11 sequence in order to reprise the Mazurka’s initial phrase. In that reprise (not
shown), zone 1 returns immediately in the form of a G7 chord, which triggers the
T11 sequential descent that Tymoczko traces in that phrase. That descent is aborted
with the appearance of yet another zone 1 representative, the E7 at the third beat of
measure 5. In the consequent phrase, that same chord triggers an extended dream
fantasy in A major (mm. 12–18), the sole moment in the Mazurka when a single
triad, f minor’s hexatonic pole, is stabilized for an extended period.
The interpolated diminished seventh chords at mm. 37–39 (figure 7.18) show
how even-numbered stations can be engaged in the midst of a passage otherwise
monopolized by Tristan-genus chords, and bridge gaps in the cumulative voice-
leading descent. Even-numbered stations are integrated more comprehensively in
the much-studied e minor Prelude, whose score is presented as Web score 7.19 .
The Prelude is a slowly unfolding parallel period both of whose twelve-measure
phrases are constructed of four-measure segments. Figure 7.19 models the first
seventeen measures in a 4:1 reduction that represents each measure as a beat, and
each segment as a measure. The first phrase reaches its structural subdominant at
m. 9 and then prolongs dominant for three measures. The second phrase roughly
inverts those proportions. Subdominant arrives already at the end of its fourth
measure followed by eight measures of dominant prolongation prior to the final
cadence at m. 25. What has caught the interest of literally dozens of publishing
analysts, and many more teachers of analysis, are the connections between the
opening tonics of each phrase and their structural subdominants. The voice lead-
ing of both connecting gestures is exclusively semitonal and distributed among the
four voices so as to combine into species familiar from the classical harmonic bes-
tiary: dominant sevenths, minor sevenths, half- and fully diminished sevenths,
and French sixths. It is this feature that prompts Tymoczko (2011b, 286) to specu-
late that the Prelude and Mazurka realize the same abstract script. Framing these
chromatic downshifts are progressions featuring whole steps. These “leaps” in
chromatic space, indicated by diagonal lines in figure 7.19, mark the moments
when tonic first loses and then regains its focus.
The chromatic linear progressions are well formed from a Schenkerian stand-
point, and accordingly, it is attractive to interpret the entire passage as completely
secured by its diatonic frame and unilaterally oriented toward e minor. On Carl
Schachter’s reading (1988), the events internal to the chromatic downshift bear no
harmonic value and colonize no Stufen. Taking a more harmonic approach,
162  Audacious Euphony

Figure 7.19. Chopin, Prelude, Op. 28 no. 4, diatonic frame.

Fred Lerdahl hears the downshifting tritones summoning a series of alternative

tonics, none of which materializes (2001, 109; see also Tymoczko 2011b, 287
n. 21). Figure 7.20 presents a more complete 4:1 reduction of the opening phrase,
omitting only the upper neighbors in the cantus. Diagonal lines indicate tritones.
Those tones disposed to leading tone status are selected for annotation. The tri-
tones initially summon tonics distributed along a line of fifths: first E and then A
and D. Each leading tone defies the summons and continues downward in a
triumph of gravity and inertia over magnetism (using terms cultivated by Larson
1994). Near the beginning of the second segment, an FᅊC tritone summoning G,
the next projected tonic, sustains into the following measure. There it is joined
by another tritone, DᅊA, which potentially interrupts the line of fifths, circling
back toward the initial E minor tonic. Like their predecessors, first Dᅊ and then
Fᅊ resist the call, descending instead to Dᅉ and Fᅉ, and the leading-tone-seeking
energy is transferred to another diminished seventh two measures later, at the
end of the second segment. Here finally, the call to resolution is heeded: as if
awakened from its omnitonal haze by the whole-step motion in the bass, the
Gᅊ remembers its monotonal responsibilities, and its rise to A triggers the reorien-
tation to tonic. Attention to these leading-tone potentials leads easily to assignment
of harmonic status for each chord, as dominants of unrealized tonics. Like the
linear analysis advanced by Schachter, the harmonic one is well formed, provided
that one allows tonality to structure one’s hearing even in the absence of sounded
local tonics.
Robert Gauldin (2004, 715) suggests a third hearing that combines aspects of
the previous two, as a linear span harmonically supported not only at the boundar-
ies but also by an interior structural pillar: E7 on the downbeat of m. 4, whose bass

Figure 7.20. Chopin, Prelude, Op. 28 no. 4, mm. 1–9, with tritones indicated and
leading tones identified.
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  163

Figure 7.21. Chopin, Prelude, Op. 28 no. 4, mm. 1–9, adapted from Gauldin 1997.

E is a plausible component in the Bassbrechung of the tonic triad, and whose GᅊD
tritone reappears at the end of the second segment.17 Figure 7.21, adapted from
Gauldin, presents this hearing as a well-formed linear graph. The first four-measure
segment converts the e-minor tonic to a V7 of the subdominant, and the second
segment prolongs that dissonant harmony by opening up a chasm of arpeggiation,
and filling it chromatically within each individual voice. The harmonic stasis of the
second segment suggests figure 7.22 as an underlying model of the first phrase,
suturing the end of the first segment to that of the second.
The second phrase, modeled at figure 7.23, amply rewards this hypothesis. Its
opening segment retraces the course of the first phrase but arrives at E7 a half-
measure early. Chopin uses the two netted beats to break the constraints that the
Prelude has husbanded with such consistent dedication. While the melody leaps a
diminished seventh to the upper octave on beat 3, the lower three voices jump a
minor third, a Gulliver stride in the Lilliputian world of this Prelude. The location
where the chasm of arpeggiation is hurdled corresponds to that in the first phrase
where the hypothesized jump of figure 7.22 was declined, in favor of the four-
measure bridge that traverses the chromatic span.
Figure 7.24 presents the first phrase in 4:1 reduction and addresses the details
and pacing of its downshift, which covers a total of fifteen semitones: four in each

Figure 7.22. Chopin Prelude, first phrase, with m. 4 sutured to m. 8.

17. Schachter 1995, 150, considers the possibility that the E at m. 4 is structural but rejects it since its
dominant seventh resolves locally to a dissonance.
164  Audacious Euphony

Figure 7.23. Chopin Prelude, second phrase.

lower voice and three in the cantus. An initial segment, leading from B7 at m. 2
into E7 on the downbeat of m. 4, descends by four cumulative semitones in the
span of two measures.18 A second, quicker segment, extending from the third beat
of m. 4 to the following downbeat, drops five further semitones in the span of a
single measure, omitting the odd voice-leading zones. The final segment slows the
pace of downshifting, requiring three measures to descend the four units from
zone 8 to zone 4 before accelerating in m. 8 in preparation for the tonal reorienta-
tion at the following downbeat.
Figure 7.25 wraps the circumnavigatory path about 4-Cube Trio (see also Web
animation 7.25 ).19 The path intersects itself in the motion from 3 → 2, repre-
sented as Bø7 → Bo7. The two instances of this progression mark the two incremen-
tal accelerations of the outer phases, in preparation respectively for the arrival of
the structural V7 of iv and its resolution to iv6 five measures later. The consequent

Figure 7.24. Chopin Prelude, first phrase, with downshift segmented into three
phases.

1st segment 2nd segment 3rd segment

18. I follow the tradition of considering the tenor E as a suspension that stands in for Dᅊ. A more literal
reading of the harmony would add EFᅊAB as a harmony at zone 6.
19. This figure closely tracks Tymoczko 2011b, 287–89, using a simplified geometry.
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  165

Figure 7.25. Chopin Prelude modeled on 4-Cube Trio. Animation, with recorded
performance, is at Web animation 7.25.

phrase is also represented on figure 7.25; divergences from its antecedent are
indicated by broken lines. This phrase, too, begins with an incrementally down-
shifting segment to an e minor seventh chord at zone 0, with some changes in
details due to a permuted firing order, followed by a leap to EGAᅊCᅊ at zone 10.
After the voices of that chord are exchanged, further leaps to zones 7 and 5 prema-
turely terminate the downshifting segment of the consequent phrase, as both of
the chords represented by these zones discharge standard tonal responsibilities
with respect to e minor.
The focus of this analysis of the Chopin Prelude, as with those that pertained
exclusively to progressions limited to chords of the Tristan genus, has been on
cultivating ways of documenting and exploring motion among standard dissonant
harmonies, when those motions do not hearken to the call of the tonal forces to
which those harmonies are normally subjected under the terms of classical tonal-
ity. But in each of these compositions, the summons is eventually heard, and dis-
sonant sonorities revert to their inborn behaviors, sometimes retrospectively
projecting those behaviors onto their predecessors. This suggests that the denizens
of 4-Cube Trio, like those of Cube Dance, are overdetermined creatures, homoph-
onous diamorphs through which composers and listeners can rapidly and
166  Audacious Euphony

radically exchange syntactic expectations. Ways to conceive and represent these

syntactic exchanges are taken up in chapter 8, in the tetrachordal as well as the
triadic case.

Scriabin’s Mystic Species and Generalized Weitzmann Regions

The analogy between triads and Tristan-genus chords is so rich and multileveled
because the two are specific manifestations of a general phenomenon: the voice-
leading efficiency that exists between nearly even chords of uniform size. The
following paragraphs sketch the general case, necessarily resorting to an abstract
mode of discourse, including a number of technical terms that are italicized at
point of presentation.
Given some nonprime universe of nq tones (n and q are integers, n > 2, q > 1),
there exist q perfectly equal divisions of chord size n.20 Each perfectly even divi-
sion serves as the core for 2n nearly even chords: n upshifters that result from
upward semitonal perturbation, and n downshifters that result from downward
perturbation. Each such region plays a role corresponding to that of a Weitzmann
region (for q = 4, n = 3) or a Boretz region (for q = 3, n = 4); I shall refer to it
as a generalized Weitzmann region (GWR). Any two chords that share a GWR
are exactly two voice-leading units apart, whether related to each other by
transposition or inversion.
Two GWRs are adjacent if their cores are transpositionally related by semi-
tone. If so, one GWR of the pair will be lower, and the other higher. Given two
adjacent GWRs, the n upshifters of the lower GWR and the n downshifters of the
higher GWR combine to form a bridging region, which plays a role corresponding
to that of the hexatonic regions (for q = 4, n = 3) or octatonic regions (for q = 3,
n = 4). Each nearly even chord is connected to n – 1 opposite-mode members of
its bridging region by semitonal displacement in n – 2 voices (the other two voices
remaining fixed) and to the remaining opposite-mode member of its bridging
region by displacing n – 1 voices in one direction and the remaining voice in the
opposite direction (generalized H, or Karg-Elert’s collective exchange; see Cook
2005, 131). The union of each chord with its collective exchange is equal to the
union of the two adjacent cores of their respective GWRs.
Earlier I noted that for every chord size there is a unique nearly even chord
species, with superior voice-leading properties, but that, even among this parsi-
monious company, the potential of the nearly even trichord holds a particularly
privileged status. The general formula just given shows why this is so. For every
perfectly even division, there is a nearly even chord species whose members

20. I place to the side those nearly even chords whose size is prime relative to that of the chromatic
system, for example, the diatonic and pentatonic scales within a 12-tone chromatic universe, or the
triad within a 7-tone diatonic universe. Nearly even chords of this type are prime generated
(Pressing 1983; Lewin 1996) and maximally even (Clough and Douthett 1991) and stand near the
center of what Tymoczko calls type-2 lattices (2011b, 107–12). Because they do not displace
perfectly even chords, they do not participate in the GWRs described in this section.
 CHAPTER 7 Integrating Dissonant Harmonies into the Model  167

partition into GWRs. Intra-regional progressions always involve two units of

voice leading, no matter the size of the chord or the universe. Accordingly, there is
nothing special about the 3-in-12 case. It is in the motion within the bridging
regions that variety is located. Leaving aside the collective exchange as a special
case, in all other motions through a bridging region, the number of voices is two
less than the number of tones in the chord.
Thus, from the standpoint of minimal voice leading, the ideal situation occurs
when n = 3. It is only then that n – 2 = 1, that is, that a bridging motion involves
only a single unit of voice-leading work. In this trichordal case, the bridges between
two Weitzmann regions are the hexatonic transformations of chapter 2, among
which are the minimal-work L and P transformations. In the tetrachordal case
studied earlier in this chapter, where n = 4, the octatonic bridges between Boretz
regions (S2, S4, S5) involve n – 2 = 2 units of voice-leading work, and hence no
minimal-work transformations.
As the cardinality of the nearly even species grows further, so too does the
number of moving voices, and the bridging transformations become increasingly
less efficient. Consider the nearly even six-tone chord species, Forte-class 6–34
with prime form [013579], whose members minimally perturb one of the two
whole-tone scales. Both of its GWRs consist of six downward perturbations that
produce Wozzeck chords (featured in the Berg opera; see Perle 1967) and six
upward perturbations that produce mystic chords (featured in many late composi-
tions of Scriabin). Motion between any pair, selected from these twelve, involves
two units of voice-leading work, just as in a Weitzmann or Boretz region.
Figure 7.26 models the opening of Scriabin’s Feuillet d’album, Op. 58 (1910),
which James Baker (1986, 129) calls “a study of the properties of 6–34, with
which Scriabin was then preoccupied.”21 The score of the opening measures is
available at Web score 7.26 . Figure 7.26(a) presents the three transpositions
of the mystic chord as they appear at the opening of the first three phrases.
Thin lines indicate pitch-class prolongations through registral displacement; the
thicker crosses indicate the progressive voice leading from one chord to the next.
The first chord would be a whole-tone collection but for the substitution of Dᅊ for
D in the “alto” voice. This substitution is recuperated in the second chord by Eᅈᅈ
in the bass but is offset by E → F, which takes Dᅊ’s place as the alto. In the third
chord, F is recuperated to E in the bass voice, while a new spoiler, Cᅊ, appears in
the alto.
Scriabin makes these idealized voice leadings salient through registral trans-
fers and motivic play (see figure 7.26(b)). At m. 4, the Dᅊ5 spoiler and soon-to-be-
spoiled E4 of the first chord exchange registers, triggering a semitonal motion that
resolves Dᅊ back into the source whole-tone collection and then spoils that collec-
tion by E5 → F in anticipation of the second mystic chord. A similar tenor/alto
voice exchange occurs between the spoiler F5 and the soon-to-spoiled C/Bᅊ at
mm. 8–9. In this case, the spoiling motion is delayed until after the appearance of
the remaining voices on the downbeat of m. 9, and the motion from Bᅊ to Cᅊ

21. See also Pople 1989, who charts the interplay between the mystic and whole-tone collections. For
related approaches to Scriabin’s late music, see Reise 1983 and especially Callender 1998.
168  Audacious Euphony

Figure 7.26. Scriabin, Feuillet d’Album, Op. 58.

transforms a whole-tone collection on beat 2 to a mystic chord on beat 3. In each

of the four semitonal transitions, the direction of semitonal motion echoes the
direction of registral displacement.
Both of the progressions illustrated here are in and out of a single equal divi-
sion and thus are analogous to transformations within a Weitzmann or Boretz
system. Bridging progressions, which would travel between two systems, would
involve a transfer from one majority whole-tone scale to the other, and would be
far less parsimonious, because of the number of tones included in each chord. Two
tones, the spoiler and a soon-to-be-spoiler, would be held invariant, and each of
the remaining four voices would be displaced by semitone, migrating from one
whole-tone scale to the other. For example, the first mystic chord in Feuillet could
be inverted into a Wozzeck chord by holding C and Dᅊ invariant and raising the
remaining pitches by semitone.
Whether or not some past composer has made, or some future composer could
make, interesting music with such routines is an open question. Even were the
question closed with a resounding no, the exploration is still worthwhile for the
light that it sheds on the special capacities of triads, as selected from all possible
nearly even chord types.
C HA P T E R
Eight
Syntactic Interaction and the Convertible Tonnetz

Everywhere, Romanticism exploits the ability to hear one and the same phenomenon
in two and more ways; it is fond of this coexistence and its indefiniteness.
—Ernst Kurth, Romantische Harmonik und ihre Krise in Wagners “Tristan”

The pan-triadic model developed in this book has left classical tonality to the side,
for strategic reasons detailed in the introduction. Yet because the two syntaxes are
so deeply intertwined in most nineteenth-century compositions, the terms and
concepts of classical tonality have surreptitiously crept into virtually every analysis
in the book. If the syntaxes operate independently of one another, then the chal-
lenge is to model their intertwining without collapsing them into each other. As
Steven Rings has emphasized in a recent assessment, such a model must do more
than dump transformational labels into the same cage with Roman numerals or
Schenkerian representations; it needs to study how they get along within those
confines. Otherwise, he writes, “the theoretical divide is thus reified in analytical
practice, resulting in a curiously bifocal view of chromatic harmony, one in which
the triad seems not so much ‘overdetermined’ as dichotomous” (2007, 34). In this
chapter I review some ways that recent theorists of chromatic harmony, including
David Lewin, Brian Hyer, myself, Steven Rings, and Candace Brower, have met
this challenge, before developing a model that builds on aspects of that earlier
work but overcomes some of its limitations.

Some Previous Proposals

Figure 8.1 presents two models of syntactic interaction, both crafted in response to
late piano compositions of Schubert. Figure 8.1(a) is a version of a figure from my
1999 analysis of the first movement of the Bᅈ Sonata. In figure 6.9 of this book,
I presented three interpretations of that movement; this figure is associated with
the first of those. The figure positions the four hexatonic cycles as cross sections of

169
170  Audacious Euphony

(b) adapted from Rings 2007

Figure 8.1. Two models of syntactic interaction.

a cylinder (Cohn 1999) and connects each triad horizontally to its perfect-fifth
transpositions. The two ends of the cylinder glue together with a one-third twist,
forming two continuous circles of fifths, composed of major and minor triads,
respectively. I suggest there that the cycle that includes Bᅈ major has tonic function
and that the fifth relations above and below bear dominant and subdominant
function, respectively.
Figure 8.1(b), adapted from Steven Rings’s 2007 analysis of the Eᅈ
Impromptu for Piano, D. 899 no. 2, positions each triad at the intersection of a
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  171

“vertical tonal-functional” and a “horizontal hexatonic” axis. His model is implic-

itly cylindrical and hence isomorphic to mine. What distinguish his figure from
mine are the arrows along the vertical axis, accompanied by transformational
labels adapted from David Lewin (1982, 1987). The arrows are single-headed
because, unlike the Hexatonic- and Weitzmann-group transformations used in
this book, these transformations only work in a single direction; they are not their
own inverses. DOM is equivalent to T5; Lewin writes that the source chord
“becomes the dominant of ” the target chord (1987, 176). (According to this defini-
tion, G major is not the dominant of C minor, even though musicians frequently
refer to that relation in those terms.) SUBD, its inverse, is equivalent to T7 = T–5;
the source chord “becomes the subdominant of ” the target chord (177). Rings
emphasizes that these transformations capture not only the syntactic flow but also
the semantic charge that is associated with classical tonality. This charge roughly
stands in for what musicians elsewhere capture through such terms as summon-
ing, leading, gravity, magnetism, attraction, desire, and so forth. Rings writes that
the dimensions “mark out a unified space in which we can map progressions that
exploit both the triad’s tonal-gravitational properties and its triadic-transforma-
tional potential, without privileging one at the expense of the other” (52).
Three considerations limit the capacity of figure 8.1 for generalization beyond
the compositions to which they initially responded. The first is their orientation to
a hexatonic configuration of chromatic space, which limits their applicability to a
repertory that often organizes its chromatic progressions along a minor-third/
octatonic axis.1 A second limitation is that Rings arrows implicitly commit to the
presence of tonal-gravitational forces whenever motion occurs along the fifths
axis, excluding the possibility that those forces are slackened in the case of
sequences, whether diatonic as François-Joseph Fétis argued, or chromatic as in
the L/R chains studied in chapter 5. That possibility can only be honored if the
Rings arrows function as a default component that can be deactivated under
appropriate circumstances, rather than as an inherent property of the model.
Finally, both models relegate third relations to the chromatic axis and thus
have difficulty capturing the common tonal procedure of dividing a diatonic fifth
progression into two constituent thirds. Lewin proposed dividing DOM into two
MED transformations, each taking the root down a diatonic third; he writes that
the source chord “becomes the mediant” of the target chord (1987, 176). Inversely,
SUBD, the motion from a subdominant to its tonic, divides into two SUBM trans-
formations, each taking the root up a diatonic third. To integrate these transfor-
mations on figure 8.1(b), the mediants would need to be placed halfway between
two DOM- or SUBD-related triads, as in figure 8.2.

1. Minor third relations are not significant in the sonata movement, but they occur prominently in the
Impromptu. Rings responds by shearing the figure, derigging major third L relations and rigging up
minor third R ones in their place. Rings ingeniously gets this contraption to carry interpretive hay
by asserting that the moment of realignment in the model coincides with a moment of disjunction
in the Impromptu, reflecting (or creating?) a semantic bang. He thereby converts an awkwardness in
the model into a profit in the interpretation, in the manner of Lawrence Kramer’s hermeneutic win-
dows (1990). But the particular case resists generalization. Where such semantic compensations are
lacking, the deficit in the model will remain a deficit in the net reckoning.
172  Audacious Euphony

Figure 8.2. Rings model with mediants interpolated.

This solution introduces an interpretive complication: c minor → Aᅈ major

projects both eastward into a hexatonic cycle and southward toward subdominant
f minor.2 Depending on one’s analytical aims, this complication need not be
a disadvantage, and indeed might even be an advantage if there were an analyti-
cal motivation to conceive of these two projections of c minor → Aᅈ major as rep-
resenting a crisp phenomenological distinction. Lewin 1992 asserts just such a
distinction in an analysis of Parsifal, claiming that one hears c minor → Aᅈ major
and its transpositions in terms of L when the progression is affiliated with
Amfortas’s suffering, but otherwise in terms of MED. But this offers little guidance
in the general case. More commonly, the distinction will lack a difference, forcing
an arbitrary choice, or will involve an ambiguity that might benefit from cultiva-
tion rather than premature resolution.3
To indicate the kinds of difficulties that can arise from such a crisp distinc-
tion, consider figure 8.3, which presents a synopsis of the first forty-three
measures of Liszt’s Consolation no. 3 in Dᅈ major. A complete score is given at
Web score 8.3 , and a recording is embedded in Web animation 8.9 . The
music consists of two sentences, both of whose presentation phrases cadence
in f minor. The initial sentence classically continues through ii6 to a perfect cadence
in Dᅈ major, executing an expanded cadential progression (Caplin 1998, 61).
The second sentence continues instead to a cadence in A minor, after which a
second continuation returns to Dᅈ major, completing a major-third division
(see figure 2.11(b)).

2. An additional problem is that the figure ceases to function as a product network, with all the techni-
cal advantages that such a status entails (Lewin 1987, 206; Rings 2007, 52). In order for the graph to
maintain its status as a product network, the newly added mediant chords would need to form L and
P relations with their neighbors along the rows. Instead, they participate in R/S cycles, creating a
transformational path through a Weitzmann region rather than a hexatonic one.
3. Lewin concedes as much when he suggests that a diatonic hearing of Amfortas’s pain figure is
“latently possible” (1992, 55n4).
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  173

Figure 8.3. Synopsis of Liszt Consolation no. 3. See Web score 8.3 for a complete
score.

From the standpoint of a final-state hearing, the double positioning of f minor

on figure 8.2 is well motivated. The initial sentence moves northward, engaging f
minor as a mediant on the way to a dominant. After returning to tonic at its initial
position, the second sentence moves westward, engaging f minor as a Leittonwechsel
that initiates a hexatonic journey. Yet this conception ignores a significant aspect
of in-time experience. Arriving at the second f minor cadence, one has no reason
to be aware of having embarked on a westward journey through chromatic space.
Indeed, the principle of “parallel passages in parallel ways” (see chapter 3, note 11)
suggests rather a retracing of the northward path toward dominant, as at m. 7. The
continuation phrase forces a retrospective reevaluation of that position; we realize
that we were migrating leftward, not upward. This reevaluation depends on iden-
tifying f minor on the vertical axis of figure 8.2 with its associate on the horizontal
axis. But the model presents us with no means for establishing that identity: the
two f minors occupy different positions, and our phenomenological journey from
one to the other involves a magical wormhole for which the model has no explicit
account.
An influential paragraph from a 1984 article by Lewin will help identify the
problem and suggest a solution.

The nature and logic of Riemannian tonal space are not isomorphic with the nature
and logic of scale-degree space. The musical objects and relations that Riemann iso-
lates and discusses are not simply the old objects and relations dressed up in new
packages with new labels; they are essentially different objects and relations,
embedded in an essentially different geometry. That is so even if in some contexts
the two spaces may coexist locally without apparent conflict; in this way the surface
174  Audacious Euphony

of a Möbius strip would locally resemble the surface of a cylinder to an ant who had
not fully explored the global logic of the space. (345)4

Is the geometry that Lewin envisions compatible with my cylinder and Rings’s
grid? Standing at a triad, one inhabits two distinct spaces, represented by the
intersecting axes, “without apparent conflict.” Yet Lewin’s thrice-iterated conjunc-
tion of “objects and relations” suggests that the intersection of the spaces includes
not only triads but also relations that pair them. He imagines the intersecting
space as a surface rather than a set of discrete points. Paths intersect not only
at points where they cross but also at segments where they merge. The grids of
figure 8.1 show the triadic objects coexisting without apparent conflict, but the
forced assignment of diatonic third relations to one axis or the other in figure 8.2
precludes the possibility that a triadic relation can coexist simultaneously in two
spaces.
Brian Hyer’s 1989 dissertation developed a geometry capable of simultane-
ously modeling both of Lewin’s spaces, while situating each object and relation in
a unique location. Hyer positions each triad as a point and connects it to its L, P,
and R associate, as well as directly to its modally matched fifth. This Tonnetz
models chromatic space by identifying (“gluing”) enharmonically and syntonically
equivalent points at opposite ends of the plane. Each such dimensional folding
individually creates a cylinder like figure 8.1(a). As Hyer phrased it in a subse-
quent article, “the [transformational] group as a whole disperses the functional
‘significance’ of [a single] triad among the harmonic consonances woven together
to form its algebraic fabric; there is no one triad that forms a tonic for the group as
a whole” (1995, 127).
Hyer’s Tonnetz, however, has the capacity to change shape in response to how
the listener hears the relations among its objects. If the triads are heard to collabo-
rate in the definition of some tonic, then the glue loses its bond. “To assert a given
triad as a tonic . . . forces us to imagine transformational relations with regard to
the tonic, and to calculate them in scale degrees rather than generic semitones, in
effect decircularizing [the Tonnetz], extending its [axes] in all directions” (1995,
127). Converting from a circular to a planar geometry “impos[es] a sense of per-
spective on the surrounding terrain, a point of view from which all the other triads
appear to be near, more or less remote, or over the horizon” (127–28). Inversely,
“when it becomes strained to hear relations between triads with respect to a given
tonic triad, then we in fact no longer hear that triad as a tonic. At that moment . . .
the circularized form of the lattice comes back into play” (Hyer 1989, 215).5
Hyer’s convertible Tonnetz is ordered up to Lewin’s blueprint in almost every
respect. Each triad occupies a unique position, as does each direct triadic relation.

4. The italicized passages in the original 1984 publication were romanized in the 2006 reprint (194).
The relevance of this passage to the present situation hinges on the interpretation of “Riemannian
tonal space,” whose domain of reference was mobile in Lewin’s writings of the 1980s. It is nonetheless
clear that the Riemann/scale degree distinction has strong affinities with binary relations that Lewin
elsewhere cultivates in terms of chromatic/diatonic and atonal/tonal.
5. The distinction between the circular and planar interpretations is equivalent to the conforming/
nonconforming distinction in Harrison 2002a.
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  175

The interpretation of each relation is contingent on whether the space is closed or

open. The structure of the space cannot be inferred from the standpoint of a single
triad, or even of a direct relation such as the Dᅈ major and f minor of figure 8.3.
That structure is cylindrical when the space is closed, exactly as in Lewin’s meta-
phor. Only one detail is astray: where Lewin envisions a Möbius strip, Hyer con-
structs a plane. To bring the vision to full realization requires us to imagine Hyer’s
plane closed into a loop, with a half-twist. Candace Brower’s 2008 consideration of
the Tonnetz as a model of diatonic space provides a motivation for exercising our
geometric imagination in exactly this way.

The Diatonic Tonnetz

Rather than accepting the Tonnetz as a space given all at once, Brower generates it
in stages from a seed node. The stages roughly retrace the development of pitch
systems across historical time (Brower 2008, 71) and encourage us to explore with
some precision the interplay of forces that precipitate and shape this development.
Brower recognizes a centrifugal force that operates on purely tuned triads, trans-
posing them by purely tuned intervals. This is balanced by a centripetal force that
relies on our psychophysiological propensity to ignore pitch differences beneath a
threshold of just-noticeability, treats similarity as equivalence, and induces sys-
temic closure through symmetric completion (68). Brower’s tale is thus a variation
on Plato’s allegory of musical forces whose reconciliation provides a model for the
ideal civic society (McClain 1978).
Although motion from a tone to its incremental approximation masquerades
as stasis under our notational system, it can be conceived in terms of voice leading
through a comma. Accordingly, Brower’s distinction between pure and flexible
tuning can be recast in terms of a tension that has been central to this book from
the outset, between acoustic consonance and melodic proximity. Accordingly, in
retracing the stages of Tonnetz generation, the following exposition periodically
pauses to evaluate the respective contributions of consonance maximization and
voice-leading minimization. Although at times these forces collude so deeply as to
be indistinguishable, we will occasionally find seams that allow them to be pried
apart.
At the first stage, shown at figure 8.4(a), a purely tuned C major is transposed
by its most acoustically powerful interval, yielding F major to the left and G major
to the right (Brower 2008, 71).6 Because the interval of transposition is internal to
the structure of the triad, where it appears only once, neighboring triads share a
single point. Alone among the initial tones of the triad, E is unshared at this stage
of development. Yet E stands in the same fifth relation with two neighbors on its

6. This derivation of the diatonic scale is appropriated from Hauptmann and Riemann; see Harrison
1994, 44–45. Brower’s Tonnetz rotates mine by 90°, to bring absolute direction in correspondence
with the intuition that dominants are above their tonics. I have not followed her in this, not only
because of historical precedent but also because her orientation causes upshift voice leading to flow
down the page, and vice versa.
176  Audacious Euphony

Figure 8.4. Genesis of a diatonic Tonnetz.

rank, A to its left and B to its right. Honoring this acoustic potential leads to the
closure of two local triangles and creates a trapezoid about the periphery of the
seven-tone figure, as shown at figure 8.4(b).
The relation of the two minor triads inverts the relation of adjacent major
triads, suggesting the possibility of replicating the entire three-triad complex on
the minor side. This suggestion can be fulfilled by adding either b minor to the
right or d minor to the left (figure 8.4(c)). The latter solution is preferred because
it evidently replicates a tone already present, thereby creating a first stage of sys-
tematic closure. From the standpoint of pure intonation, the closure is an illusion:
if each edge represents a justly tuned interval, then the two D tones differ by a
syntonic comma. Honoring the comma was a cardinal priority for many theorists
of the mid-nineteenth century and, indeed, continues to be so in some musical
microcultures (Perlman 1994; Duffin 2007). Yet there is a strong incentive to
accept some nonidentity as a quasi identity, since to refuse to do so at every oppor-
tunity is to invite a boundless proliferation. Accepting this identity creates a first
stage of systemic closure, completing the portrait of the diatonic collection as a
parallelogram whose interior is tiled into six consonant triads, as at figure 8.4(d).
I will use the term syntonic image to characterize the relation between the circled
tones occupying its remote corners. The parallelogram encapsulates its contents
into a microecological hothouse, activating the internal dynamics signified by
Rings arrows.
Under Renaissance modality, the six encapsulated triads are in principle
equally stable, and any of them can serve as finalis. Under classical tonality, only
the central triads, sheltered from the parallelogram’s unstable corners, enjoy this
privilege. In a process elegantly described by Gottfried Weber (1846 [1817–21]) as
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  177

Figure 8.5. T → S → D → T on the Tonnetz.

“Attunement of the Ear to a Key,” and recently revived in a more overtly psycho-
logical framework by Fred Lerdahl (2001, 193ff.), the blank ear takes a triad
as a metonym for a diatonic system that remains a hypothesized default, subject
to confirmation. Hearing a consonant triad, one provisionally situates it inside
its appropriate parallelogram. Standard confirmation comes in the form of a
tonic → subdominant → dominant → tonic (abbreviated T → S → D → T) progres-
sion represented on the parallelogram by two leftward arrows, one initial and one
terminal, linked by a longer rightward arrow connecting its two sides (Brower
2008, 71).
There are two distinct ways to interpret the motion of the arrows in figure 8.5.
On the planar Tonnetz, subdominant and dominant balance about tonic, and the
rightward arrow represents an overcompensation. Riemann imagined the sub-
dominant as a “stretched bow which slings the arrow beyond the mark” (Riemann
1893, qtd. in Rings 2007, 50). The slung arrow may be interpreted to represent the
“vault” across the syntonic comma. Yet Riemann took equal temperament as an
empirical given (Rehding 2003, 50), and so that vault was executed not in the
acoustic signal but rather in the imagination.7 Once the acoustic identity of the
two D tones is recognized, it is a small step to honor their unification in concep-
tion, leading to a more conjunct relation between subdominant and dominant.
This conjunction is represented on the Tonnetz by a geometric manipulation that
is difficult to represent or imagine in two dimensions but can be reproduced in
three dimensions with appropriate implements. First cut out the parallelogram of
figure 8.4(d). Then loop it and glue together its remote corners. D is now at the
center of two orthogonal line segments that share points of terminus: a seven-
point coil of fifths, FCGDAEB, and a three-point line of minor thirds, FDB.
The second stage of closure involves accepting the pseudo-fifth BF as if it were
perfect, creating a direct connection between its components. The result is that the
coiling line of fifths becomes a circuit. Since the coiling process positioned B two
ranks above F on the FDB line of minor thirds, they can be brought into direct
contact only by crooking that line at its center, fixing the DF edge, and folding the
BD edge so that B comes alongside F. B now intersects the coil of fifths as it is pro-
jected beyond its F terminus. The new edge that connects B and F locally closes a
triangle that represents the diminished BDF triad and globally closes the chain of
triangles into a loop, representing the seven diatonic triads arranged into a con-
tinuous series of diatonic mediants (Agmon 1995). Because the DF edge is fixed
and the BD edge inverts, every intermediate point of the parallelogram downfolds

7. On the relation of the ideal and the material in Riemann’s treatment of intonation, see Harrison 1994,
237, and 2002a, 136.
178  Audacious Euphony

by a distance proportional to its proximity to the mobile BD edge. The front side
of one end of the strip is attached to the back side of its opposite end, transforming
the planar Tonnetz into a Möbius strip (Brower 2008, 73–74). This half-twist is the
equivalent of Oedipus at the crossroads: it fulfills Lewin’s prophecy in all its details
(except for the ants, which retain their arbitrary-but-evocative status). The right-
ward arrow on figure 8.5, connecting the subdominant to dominant, can now be
interpreted as continuing in a uniform direction around the backside of the strip,
rather than reversing direction across its anterior plane.
Either interpretation of syntonic images, as fence posts bounding a flat field
or as linking elements around the backside of a Möbius strip, is consistent with
the image of diatonic space as an encapsulated microecology. But what a life is
lived within those confines! To the classically attuned ear, the identification
of the tonic immediately evokes a hierarchization of the constituent tones and
chords, intuitions about their context-free proximity, a syntax that governs com-
ponent ordering, and the semantics of Rings arrows, which symbolize a rich net-
work of overlapping metaphors that have been canonized to the point of quasi
naturalization: energy, gravity, attraction, magnetism, discharge, orientation,
center, departure and arrival, return, passing, neighboring, leading, deception,
completion/incompletion, suspension, finality, stability/instability, and so forth.
Music theorists have an overlearned sense of how all this works, no matter
what tradition of classical theory we were raised in or project in our writings and
teachings.
So far, deriving the Tonnetz “from scratch” does not require any appeal to
melodic proximity, except at the point where the syntonic images are granted
identity. Yet as a reified object detached from its mode of genesis, the diatonic
Tonnetz possesses several salient properties that can be conceived and represented
in voice-leading terms. Triads that share an edge are related by a single unit of
voice-leading work, as measured on the diatonic Möbius strip. The double
common-tone feature lacks significance; it follows from the definition of edge
adjacency, and hence is tautologous. What is significant is that the mobile
voice moves by the minimal diatonic interval of a step, a result of the nearly even
status of diatonic triads in seven-tone diatonic (not twelve-tone chromatic)
space.8
A second salient voice-leading feature is the correspondence between directed
motion on the Tonnetz and in voice-leading space: leftward motion (D → T → S)
is entirely upshifting. How this feature is realized in registral pitch space
depends on which of the two interpretations of the Tonnetz is adopted. The voice-
leading correlate of the planar conception, with its disjunctive overshoot from
subdominant to dominant, is well embedded in the experience of the fumbling
student of keyboard harmony, for whom the downshift from subdominant to
dominant is a challenge that must be negotiated without a common-tone anchor
(figure 8.6(a)). The Möbius conception, which conjoins subdominant to dominant
around the back side, likewise has a voice-leading correlate in the perpetually

8. Compare chapter 2, p. 37 of this book.

 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  179

Figure 8.6. Two voice leadings for T → S → D → T.

upshifting figure 8.6(b), which reconstructs the double emploi of Jean-Philippe

Rameau.
There is little reason to doubt the deep connection between eighteenth-century
diatonic tonality and the acoustic capacities of the triad. Nonetheless, there exist
segments of diatonic music that force acoustic properties into remission and tame
the semantic power of Rings’s arrow. What alerted Fétis in the 1840s to the special
nature of these sequential moments was the behavior of the pseudo-fifth, whose
appellative powers are neutralized by the “symmetry of movement and succession”
(2008 [1844], 30). The specific identities of intervals dissolve into their generic
categories, and the contours of the Möbius strip disindividuate into a strangely
shaped, greased racing oval.
Although what remits Fétis’s “law of tonality” is his “law of uniformity,” there
is another, silent partner in the success of these progressions: the nearly even status
of the diatonic triad, which ensures that any two of its species are step connected
(Agmon 1991). The diatonic Tonnetz is already a robust space, dually determined
by the properties of the overdetermined diatonic triad that generates it. Here the
potential reciprocity of acoustics and voice leading, and their analytic inseparabil-
ity, can already be glimpsed even before the limits of the parallelogram are
breached.

Horizontal Extensions

Incremental extensions: Modulation to closely related keys

The encapsulation of the diatonic Tonnetz as a Möbius strip depends on the accep-
tance not only of the syntonic comma as a perfect unison but also of the pseudo-
fifth as a perfect fifth. No amount of tempering will cure the latter case; one is
asked to simply accept it axiomatically. Renaissance musicians often opted instead
to extend the chain of perfect fifths outside the parallelogram. The result is a his-
torical expansion of the diatonic Tonnetz along the fifth axis, at first incremental,
ultimately compounded in a process whose documentation was a lifelong project
of Edward Lowinsky (1946, 1989 [1967]; see also Brower 2008, 83). By the classical
180  Audacious Euphony

era, the tracks, in all their potentially infinite extension, were a thing made rather
than a thing in the making. Although eighteenth-century composers had the
option of traveling the length and breadth of the extended line of fifths, they usu-
ally declined to do so. A narrower band of fifth space, closed by the pseudo-fifth,
was better suited to the cultivated microecology of the hothouse and the aesthetic
and semantic attributes of classical tonality that arose within it.
A limited degree of mobility along those tracks nonetheless suited the ends of
classical tonality, as it allows the diatonic capsule to be embedded into a larger
“modulatory” shell that replicates its properties and dynamic features on a broader
scale. Sharps to the right and flats to the left temporarily displace their diatonic
letter-name equivalents, shifting the parallelogram along a horizontal track. The
principle of close relation dictates that such a shift be limited to a single degree on
either side of the tonic. With each shift comes a new pair of syntonic images, a new
pair of potential tonics, and a reversal of Rings arrows and the syntactic and
semantic forces they represent. These shifts might be characterized in terms of
secondary (applied) chords, tonicizations, or modulations, depending on their
duration, the presence or absence of cadential signifiers, and the proclivities of the
characterizer.
In order to demonstrate the plotting of fundamental diatonic progressions and
close modulations on the Tonnetz, figure 8.7 analyzes the opening sentence of
Liszt’s Consolation no. 3, as a series of snapshots of a listener’s awareness as the
music unfolds. These images are fused in Web animation 8.9, an animated version
with embedded sound file . Figure 8.7(a) depicts the opening three measures,
which arpeggiate a Dᅈ major triad in the manner of a nocturne. In addition to
presenting the triad as a triangle, the figure also presents the diatonic collection in
which any acculturated listener imagines it contained, a collection of tones inferred
but as yet unsounded.
Figure 8.7(b) represents the four-chord basic idea of mm. 3–6, in terms of the
T → S → D → T paradigm introduced in connection with figure 8.5. North American
harmony pedagogy would likely analyze this same progression in terms of T → [D]
→ D → T, positioning the {G, Bᅈ, Dᅈ, F} chord to the right of the dominant. The
subdominant interpretation offered here follows the embedded-triad protocol
introduced in connection with figure 7.3, maintaining the Dᅈ and F common
tones in a single location rather than displacing them a syntonic comma to the
southeast. Only the G, treated as a supplementary under-seventh, is located to the
dominant side, dislocated from the remaining tones of the chord on the plane.
This interpretation is consistent with the Riemannian approach to dissonant har-
monies, treating them as a mixture of dominant and subdominant components.
Figure 8.7(b) places G within a star and connects it by a dotted line to the triad that
it supplements. (This is an ad hoc solution that functions well here but will create
clutter in many other cases.) The initial displacement of tonic thus shifts in both
directions on the Tonnetz: leftward at the level of the triad (Aᅈ → Bᅈ) and rightward
at the level of the diatonic collection (Gᅉ for the hypothesized Gᅈ of figure 8.7(a)).
Both involve an upshift in pitch-class space. The subsequent motion to the domi-
nant seventh chord involves a compensation at both levels. The harmony moves to
the right, or dominant, side of tonic, as indicated by the overshifting arrow. At the
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  181

Figure 8.7. Liszt Consolation no. 3, first sentence. See Web score 8.3 for a score.
Animation, with recorded performance, is at Web animation 8.9. (a)–(b) The basic
idea, tonicizing Dᅈ major. The starred tones are supplementary dissonances for the
triads to which they are attached. (c) Repetition, modulating to f minor.
(d) Continuation from f minor to Dᅈ: ii6. (e) Cadence.
182  Audacious Euphony

same time, the dissonance, Gᅈ, starred at the left edge of the figure and connected
to the Aᅈ major that it supplements, replaces the Gᅉ that supplemented the prior
harmony. Thus, the collection shifts left as the harmony shifts right, both motions
involving a downshift in pitch-class space. The cadential tonic at m. 6 is then
reached via triangulation from both the dominant and subdominant side. Although
both dissonant chords internal to the basic idea individually bear dominant and
subdominant elements, in combination they balance more strongly to the domi-
nant side, marking out territory to be pursued when the basic idea is repeated.
Figure 8.7(c) models the subsequent tonicization of f minor, fulfilling the sub-
dominant potential of the basic idea’s initial dissonant harmony. The supplemen-
tary dissonances here are not starred, to enhance the legibility of the figure, but are
easily imagined. The cadence in f minor shifts the syntonic images rightward, from
Eᅈ to Bᅈ. The initial dissonance on bᅈ minor comes into relief here, as the only
diatonic triad that bears subdominant function in both keys. The figure also
includes an initial vertical extension to the rank above, in order to acknowledge
the honorary-diatonic status of E as f minor’s leading tone and as the third of its
major dominant, substituting in both capacities for Eᅈ. Such vertical extensions
are theorized in the next section of this chapter and exemplified by the second
sentence of Liszt’s Consolation.
Figure 8.7(d) presents the beginning of the first sentence’s continuation phrase
at mm. 12–13, where f minor is transformed into a diminished seventh chord over
a common F bass, which presses toward eᅈ minor qua diatonic ii6. The diminished
seventh chord presents a particular challenge to the graphic apparatus, as it mixes
subdominant and dominant tones that are not adjacent on the plane.9 Again, I
have adopted an ad hoc solution, forming a quadrilateral bounded by DF above
the tonic and AᅈCᅈ on the left edge, and inserting a black dot to represent the
temporary occupation of that space. Figure 8.7(e) presents the expanded cadential
S → D → T progression.
As an analytic statement about Liszt’s Consolation, figure 8.7 is far from ideal.
Its four components cumulatively consume considerable space in order to make
some fairly rudimentary claims about music well understood using other repre-
sentational modalities that are more familiar, more economical (in the case of
chord and function labels), and less abstract (in the case of Schenkerian graphs).
The Tonnetz representations do have some added value, allowing one to track the
role of common tones in forging moment-to-moment connections, redefine a
single triad in multiple tonal contexts, and observe the interaction of dominant
and subdominant regions (or, more neutrally, upshift and downshift voice lead-
ing). But these considerations hardly justify the investment in initial pedagogical
overhead and ongoing production costs. A more significant motivation will emerge
as we study the interaction of these diatonic Tonnetz representations with their
expansion into chromatic pan-triadic space.

9. The analysis of a diminished seventh chord as mixture of subdominant and dominant derives from
Rameau (see p. 140 of this book). Its Tonnetz depiction is inspired by Oettingen 1866, 265. See also
Harrison 1994, 65–69.
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  183

Recursive extensions along the L/R chain

The familiar processes described in the preceding section are subject to recursive
application, precipitating more ambitious excursions away from the global tonic.
A pair of shifts in a single direction already exits the system of closely related keys,
tonicizing a chord that is chromatic with respect to the initial tonic. Sequential
tonicization of fifth-related, modally matched tonics, as in many classical develop-
ment sections, compounds this process. The moving window begins to resemble a
local train, migrating incrementally in a uniform direction and locking temporar-
ily into a series of adjacent stations. The track on which this journey takes place is
the L/R chain studied in connection with figure 5.9.
Sequential migrations along the L/R track can also occur in the absence
of cadential activity to lock in the series of stations. In such cases, the window
slides transiently through a series of keys, each closely related to its neighbors
in the chain. An early example is the Adagio from J. S. Bach’s d minor keyboard
Toccata (BWV 913, pre-1708; a score is available at Web score 8.8 ). After
establishing g minor as tonic, the Toccata moves incrementally leftward,
substituting flats for naturals in key signature order until six flats are in play. It then
reverses course, replacing the flats with naturals in the order that the latter were
cast out. This rightward motion overshoots, replacing all six flats with their cor-
responding naturals, ultimately terminating at d minor’s dominant in preparation
for the final fugue. Each added flat or natural signifies a new minor tonic, but none
are cadentially confirmed. Some approaches are abandoned at the cadential 64 ;
others, at its dominant resolution. Still others make it as far as an imperfect
cadence, which is immediately undone; the sensation is of a housefly landing and
immediately taking off again. In such situations, the incrementally migrating
window is a surveillance engine that slows to inspect each station but alights
at none.
In still other passages, a series of mediant-related triads acts as a slippery slope
for an extended excursion into chromatic space. The prototype is mm. 159–71 of
the Scherzo from Beethoven’s Ninth Symphony (see p. 92 of this book, and Web
score 5.9 ). As the triads fly by, they have the potential to be gathered into bou-
quets of six and constituted as a diatonic space. But after several hypothesized
tonics fail to materialize, we give ourselves over to the momentum of the journey.
Once the pattern is broken, our tonic-seeking radar might reawaken, stimulating
us to gather up the most recent few triads and kindle a cadential fire that activates
their tonal potential in retrospect. I imagine this a bit like an express-train traveler
who stops wondering what it might be like to disembark at each local station
and focuses on the somatic sensations of the ride, until the slowing of the wheels
reengages images of the external world.
Here is where the Tonnetz earns its value despite the extra weight it requires us
to port about. Exiting the initial diatonic space, the parallelogram slides incremen-
tally left or right. Flying past multiple diatonic spaces, its diagonal edges ease
open, and the horizontal track converts to an alley of indefinite extension. Rings
arrows retract into the body of the moving vehicle. Approaching a new diatonic
station, doors flip closed, Rings arrows reemerge, and we psychologically reorient
184  Audacious Euphony

ourselves to the new tonal environment, reengaging the whole suite of syntactic
expectations and semantic predications of the classical microecology.
Brower (2008, 76) implies that the acoustic properties of triads underlie the
horizontal expansion of the diatonic Tonnetz, through recursive iterations of the
same process that generates the subdominant and dominant triads from the tonic.
A complementary view is suggested by Milton Babbitt (2003, 195), who notes that
maximization of scalar common tones presents a sufficient motivation for modu-
lating by fifth, independent of that interval’s acoustic properties. The common-
tone relations among the triads of an L/R chain constitute yet a third layer of
overdetermination. Ranging along such a chain, we can forget about diatonic con-
stituency and acoustic privilege and simply enjoy the smooth ride provided by
parsimonious voice leading and the variety provided by the extensibility of the
chain to twenty-four stations before replication occurs. In these progressions,
the nearly even triad of pan-triadic syntax takes on progenitive powers equivalent
to and inseparable from those of its coextensive image, the consonant triad of
classical harmonic theory.

Vertical Extensions

The L/R chain cannot capture one chromatic extension that arose at an early his-
torical stage, that flared up prematurely in connection with figure 8.7(c), and that
is familiar to any harmony student from an early stage of tutelage: the substitution
of the semitone below tonic for the whole tone, when the latter is proper to the
diatonic scale. Figure 8.8(a) locates this substitution on the Tonnetz, with respect
to A minor: Gᅉ is replaced by the Gᅊ directly above it on the dominant side of
tonic, converting the minor dominant east of tonic to the northeast-positioned
major dominant, the nebenverwandt of the minor tonic. A dual echo that initially
arose in the sixteenth century (see chapter 7, note 3) and became increasingly
prevalent between 1725 and 1825 involves the substitution of the semitone above
the fifth scale degree for the whole tone, again when the latter is proper to the
diatonic scale. Figure 8.8(b) locates this dual substitution with respect to C major:
Aᅉ is replaced by Aᅈ directly beneath it on the subdominant side of tonic, convert-
ing the major subdominant west of tonic to the southwest-positioned minor
subdominant, the nebenverwandt of the major tonic.
Since Schenker’s time, it has been standard to explain the phenomena por-
trayed in figure 8.8 by invoking modal “borrowing” or “mixture.” One says that,
lacking a semitonal discharge to tonic, minor borrows that capacity from parallel
^
major. Lacking a semitonal discharge to 5, major borrows an asset from minor’s
private stash. This lovely allegory of civic reciprocity in the late-Platonic republic
has three limitations, qua explanation: it is presumptuous, logically unparsimoni-
ous, and ahistorical. It presumes that, from the white-note perspective, there
already exist fully ramified three-sharp and three-flat scales. But one searches
in vain for those collections in figure 8.8, as one does in early polyphony. It is
unparsimonious because it borrows three tones, only to discard two. Why order a
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  185

Figure 8.8. Diatonic collection with nebenverwandt substitutions.

three-course dinner when the tone for which you hunger is available à la carte?
The substitutions in figure 8.8(a) are sufficiently explained by a voice-leading prin-
ciple articulated by Marchettus of Padua in the early fourteenth century without
any reference to scales: a semitone may substitute for a whole tone. As I showed in
chapter 2, in the general case the displacement of a single tone by a single unit
requires an exchange of chord type, and thus a violation of harmonic consistency.
It is only when the chord is nearly even and its cardinality is odd that such a
semitonal substitution preserves chord type. It was only in the fifteenth century,
a century after Marchettus, that harmonic consistency became a compositional
desideratum and that the Gᅉ → Gᅊ substitution at the pitch-class level became
staged as an e minor → E major substitution at the triadic level.
In our genesis story, this is the moment when the near-evenness of the conso-
nant triad becomes crucial, and its acoustic properties become epiphenomenal
by-products. The e minor → E major substitution reduces voice-leading work while
maintaining harmonic consistency, a circumstance enabled by the triad’s near
evenness in chromatic space. True, the major-for-minor substitution is also an
acoustic gain, as it replaces a minor triad with a more purely tuned major triad.10
But this gain is more than offset by an immense loss at the level of the scale: the Gᅊ
enters into a howlingly dissonant relation with all of the natural-scale tones with
which it is not directly connected on the Tonnetz. It is in the rectification of this

10. Local acoustic considerations nonetheless have considerable historical force in motivating major-
for-minor substitutions, independently of any gain in voice-leading parsimony. This is most clear in
the case of the tièrce de Picardie, and its nineteenth-century expansion as the aspera → astra plot
prototype. Moreover, although e minor → E major and F major → f minor register equivalent voice-
leading gains, the former arises at an earlier historical moment, a circumstance that can only be
explained in terms of an acoustic bonus.
186  Audacious Euphony

anomaly that acoustic factors once again assert their powers. C and F are replaced
by Cᅊ and Fᅊ, directly above them on the Tonnetz. These substitutions raise the
acoustic value of Gᅊ’s neighborhood through westward extension of a line of fifths
and create the conditions for their own encapsulation in an A major region. The
process just described is susceptible to exponentiation, pushing ever northward
across historical time. An inverse procession to the south is initiated with the
pitch-class substitution A → Aᅈ, triadically staged as F major → f minor.

The Convertible Tonnetz

Such diachronic fantasies are of little consequence to composers and listeners

of the classical and later eras. For them as for us, the universe is given all at once,
as a set of opportunities to be taken or declined, depending on composerly
temperament and the perceived tastes and tolerances of patrons and listeners.
The Tonnetz becomes a black-box theater, whose walls can be erected, dismantled,
and moved without fuss. One is free to establish an encapsulated diatonic space,
activating the Rings arrows with their associated tensions and attractions. One
may also navigate purposefully along one of its bounded hexatonic or octatonic
alleys, or over a set of connected diatonic spaces without alighting, or by establish-
ing residence in the encapsulated space of a pitch retention loop. One may peram-
bulate more freely, meandering from triad to triad across shared edges. Adding a
supplementary dissonance to any triad, one may transfer into an orthogonal space
consisting of Tristan-genus objects, and wander that space as if oblivious to
the semantic values of classical tonality and the syntactic obligations that they
impose.
Yet as we navigate the unbounded Tonnetz, we are alert to the possible reasser-
tion of classical syntax on the triadic and Tristan-genus objects that are traversed.
When several triads are sounded in a contiguous horizontal segment of the space,
we formulate a hypothesis rooted in the syntactic mechanisms of diatonic tonality.
Confirmation reorients our attention from the far-flung space toward a well-de-
fined sector of it, and ultimately to a particular location in that space, the cadenced
triad. The process might be likened to entering the four walls of our home, or
sprouting landing gear and coming within view of a planet or an airport, or exiting
the freeway and entering a fortified village or gated community. This last image
suggests that the Tonnetz resembles a hybrid automobile: one engine for the high-
way of chords, and one for touring the encapsulated neighborhoods, with all of
their tensions and fraught attractions. At any moment, one may brake on a sub-
dominant, reverse into a dominant seventh along the curb, pull forward to a tonic,
and kill the engine.
Figure 8.9 analyzes the second sentence of Liszt’s Consolation as a sequence of
six stroboscopic snapshots, arranged chronologically from top to bottom. (Web
animation 8.9 merges them into a single animation, with sound file embedded .)
The first image, labeled “first sentence,” presents a synopsis of the composition’s
first twenty-three measures, treating f minor as the mediant degree within a
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  187

Figure 8.9. Liszt Consolation no. 3, second sentence: a stroboscopic Tonnetz

portrait. Animation, with recorded performance, is at Web animation 8.9.
188  Audacious Euphony

well-ordered Dᅈ major region. The second image synopsizes the presentation

phrase of the second sentence. This time, as in figure 8.7(b), f minor is treated as a
modulatory goal, sliding the diatonic capsule to the right and engaging a new set
of syntonic images at Bᅈ. What is new in figure 8.9 is the sprouting of a hexatonic
alley, indicating the future trajectory of the phrase.
The remaining components of figure 8.9 show the progressive ascent on the
hexatonic elevator. The triad sublimates its earlier role as acoustic generator,
becoming the optimizer of voice-leading parsimony. Yet these triads are not only
objects in pan-triadic voice-leading space. Each one is tonicized through the
resources of classical tonality. F minor becomes F major at m. 31 through their
shared dominant seventh, and F major becomes a minor at m. 35 by interpreting
the former as bearing subdominant function and sounding the latter’s dominant.
Every tonicized triad thus carries a dual role, as a voice-leading object in Dᅈ hexa-
tonic space and as an acoustic object in its own local diatonic space. Every event in
mm. 31–35, for example, has a determinate role within the diatonic pitch space of
a minor. Within that context, there exists no functional or enharmonic ambiguity.
All of the modal degrees of a minor are present. A minor serves as a field of attrac-
tion for its surrounding pitch classes, bestowing leading-tone charges upon Gᅊ, F,
and so forth. We are fully comfortable with mm. 31–35 as a diatonically determi-
nate microcosm whose perceptions are organized in relation to a minor. What we
do not understand is the diatonic function of that tonic itself, in relation to the
^ ^
larger Dᅈ major macrocosm: whether it is rooted on its 5 or 6; whether it functions
as subdominant by virtue of carrying the subdominant agent or as dominant by
virtue of carrying the dominant agent; whether it is consonant by virtue of its local
tonic status or dissonant by virtue of carrying the two agents characteristic of a
diminished seventh chord (see p. 22). A minor, like F major before it and A major
after it, is a diatonically organized, sealed capsule that bobs on the pan-triadic sea.
Its interior is an ecological microsystem within which the fraught tensions and
attractions, gravities and magnetisms of tonality are active. On its exterior, it is
tethered to the broader universe by different forces and relations.
Figure 8.9 thus presents a more complex picture of the passage than our earlier
consideration of the passage in connection with figure 2.11(b). There, we consid-
ered the same music as passing through a hexatonic cycle whose stations are “buff-
ered” by diatonic detritus. We are now in a position to give the latter their due.
Each station of the hexatonic cycle brings along its own entourage of objects, inter-
nal syntactic relations, and external semantic effusions, all representing and
expressing the principles of classical tonality.
Most of the analyses offered throughout this book might plausibly benefit from
reconsideration in light of the hybrid model of the convertible Tonnetz offered in
this chapter. In all of them, familiar tonal terms have made unmarked cameo
appearances, as so many faces in the crowd of words. Readers should now be in a
position to translate these informal remarks, cast as they are in the pidgin dialect
that we use when we talk about classical tonality—that familiar merger of language
and perspectives of Schenker, Rameau, Riemann, Fétis, Kurth, Tovey, Rosen,
Krumhansl, Hepokoski . . .—into a more formal apparatus along the lines of
figure 8.9.
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  189

Such a hybrid apparatus is even relevant in those passages where pan-triadic

syntax does its best to cleanse our ears of the tonal semiotic. Even at its most pan-
triadic, a composition can immediately invoke classical syntax, as Liszt does at the
end of each presentation phrase of his Organ Kyrie (see figure 6.10). Every major
and minor triad, whether bobbing on its isolated tub or swept into the currents of
parsimonious voice leading and sequential replication, is capable of diatonic
encapsulation. Our filled ears are as subliminally alert to this potential as the blank
ear that attunes a triad to a key.
Although the convertible Tonnetz is specific to the nearly even trichord, the
process of convertibility that it models is also pertinent to that other nearly even
chord-class that doubles as a participant in classical syntax: the Tristan genus.
These affinities are already implicit in several formulations from chapter 7: in
Brünnhilde’s Immolation, the Cᅊ7 chords “hearken to their appellative properties
after ignoring them for some measures” (see p. 157 of this book). Similarly, the Gᅊ
in Chopin’s e minor Prelude, “awakened from its omnitonal haze by the whole-step
motion in the bass, . . . remembers its monotonal responsibilities” (see p. 162 of
this book).11 The dimensional limitations of the Tonnetz prevent us from proceed-
ing by direct translation from the trichordal to the tetrachordal case. But this is a
heuristic barrier, not an ontological one. If it were possible to assign members of
the Tristan genus distinct locations on the Tonnetz, or if their position on Edward
Gollin’s tetrahedron (1998) were legible on the planar page, then their participa-
tion in a voice-leading system, and their potential to revert to diatonic syntax,
could be traced as in the triadic case. In order to imagine this participation,
we need only contemplate the triadic case and boot it through the metaphorical
conduit.

Two Analytical Vignettes: Wagner and Brahms

Figure 8.9 interlocks and overlays diatonic and pan-triadic syntaxes in the projec-
tion of an essentially sequential motion through a hexatonic cycle. This chapter
concludes with two analytic vignettes that navigate more diverse, less patterned
routes through pan-triadic space, and use the convertible Tonnetz to show how
those navigations interact with classical diatonic syntax.

The Faith Proclamation from Parsifal

Figure 8.10 presents eleven measures from the Prelude to act 1 of Parsifal. This
music, known as the Faith Proclamation, is structured as a sentence. The basic
idea, beginning on Aᅈ major, is transposed up a minor third, beginning on Cᅈ
major. Lewin (1987, 161) and Lerdahl (2001, 127) both hear the presentation

11. See also the treatment of the Tristan chord in Lewin 2006, which views the chord perched on the
cusp of a double syntax consistent with the model developed here.
190  Audacious Euphony

Figure 8.10. The Faith Proclamation, Parsifal, act 1 Prelude, mm. 45–55.

segments as half-cadential motions from tonic to dominant. But there are three
contextual factors that support hearing them as motions from a local subdomi-
nant to its tonic. Globally, plagal cadences play a significant role in the opera’s
tonal logic and semiotic network (Lewin 1984). Locally, the basic idea follows a
tonicization of Eᅈ major through a plagal extension of the Dresden Amen. And
associatively, the melodic journey—anacrustic tonic, climax on a metrically
^ ^
accented 4, inverse-arch, metrically accented cadential 3—has already been pre-
sented as a motivic topos at the conclusion of the Prelude’s opening phrase (Spear
motive, mm. 4–6). I would, nonetheless, not argue too hard for a plagal hearing as
“correct.” The relationship between Aᅈ major and Eᅈ major is ultimately underde-
termined, a Hauptmannian antithesis without synthesis, such as discussed in
chapter 3. In such situations, it is advantageous to have a conceptual and represen-
tational system that does not require a determination. The Tonnetz of figure 8.11
provides such an analysis.
The continuation phrase, itself a nested sentence, begins up another minor
third, on Eᅈᅈ major, and devolves into a plagal drift, indicated on figure 8.11 as a
motion through an L/R chain.12 The extent of that drift is not predictable; this is
one of those triadic chains that “can be arrested at any point or . . . can just as easily
go on in perpetuity” (Mitchell 1962, 9). The first sign of an impending diatonic
encapsulation occurs in the leftward motion from eᅈ minor to Cᅈ major, which
stanches the perpetual downshift that has pervaded the entire sentence. The
reverse hemiola, which elongates the counting pulse, reinforces the sense that

12. Wagner writes the chord as D major for pragmatic notational reasons; as there is no motion through
the enharmonic seam, the entire passage could have been written using flats, as in figure 8.11.
 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  191

Figure 8.11. A Tonnetz analysis of figure 8.10.

some force is being applied to stop the flow. The appearance of the first accented
dissonance in the passage, an Fø7 acting as subdominant with characteristic domi-
nant dissonance, clinches the sense of eᅈ minor’s impending encapsulation.
“Stopping the flow” has a particular meaning in Parsifal, an opera whose cen-
tral event is the healing of Amfortas’s bleeding wound. The force that stops the
flow is figured here as eᅈ minor. The hermeneutic interpretation can be extended
if we notice the role of eᅈ minor in the presentation phrase, indicated by the gray
triangle in figure 8.11. Each of its three double-common-tone associates is sounded
in turn; the music perambulates about eᅈ minor without knowing how to respond
to it, just as the goose-headed Parsifal cluelessly gawked at the Communion ser-
vice in act 1. Only after an evidently aimless journey does Parsifal stumble back
upon that eᅈ minor thing in act 3, recognize its value, and use it to stanch the
plagal drift. All of this, of course, is foreshadowing; following the eᅈ minor cadence,
the plagal flow is temporarily reversed by a Cᅈ major sounding of the Grail theme
at m. 56 but then devolves into an even more extended plagal drift that extends all
the way to the chromatically tortured reprise of the Communion theme at m. 79
(Murphy 2001).

The finale of Brahms’s First Symphony

Figure 8.12 presents a reduced score of the opening of the recapitulation of

the finale of Brahms’s First Symphony. (An orchestral score is available at Web
score 8.12 .) The principal theme is presented by the violins in C major, diverted
into a chromatic episode at m. 204, and presented again in a modified C major
counterstatement at m. 220. Both phrases of the chromatic episode begin with
partial thematic statements in the winds, in Eᅈ major (m. 204) and in B major
(m. 212), and are followed by four-measure liquidations in the horns and pizzicato
strings. The tonal domiciles of these four thematic statements are traced on
figure 8.13(a) by a series of black dots, ordered from top to bottom on the Tonnetz.
For clarity, the graphs encapsulate only the C major regions, but the reader
should imagine the Eᅈ major and B major thematic statements defined by similar
parallelograms, as they are in Web animation 8.13 .
192  Audacious Euphony

Figure 8.12. Brahms, Symphony no. 1, 4th mvt., mm. 186–221.

 CHAPTER 8 Syntactic Interaction and the Convertible Tonnetz  193

Figure 8.13. Two Tonnetz analyses of figure 8.12. Animation, with recorded
performance, is at Web animation 8.13.

The C major region of mm. 186–200 is inflected by only two brief and easily
contextualized chromatic tones. The flatward inflections that appear beginning
in m. 201 coalesce around Eᅈ major at m. 204, projecting an R/P chain on the
southeast diagonal that is taken to the next step by eᅈ minor at m. 207. The horn
fifths at m. 208 initially project Bᅈ major, as eᅈ minor’s nebenverwandt to the
northeast, but the Dᅈ at mid-measure defines a southward move from hypothe-
sized Bᅈ major to realized bᅈ minor. After three measures of prolongation,
bᅈ minor is connected to Cᅈ major by a westward motion along an L/R chain.
Each chord of this chain bears latent potential to blossom into a diatonic region,
but only the final one exercises that capacity, by virtue of its positioning at a
hypermetric downbeat.
The liquidation of the B major thematic statement, beginning at m. 215, is a
nearly exact transposition of the Eᅈ major liquidation eight measures earlier.
The parallel shapes on figure 8.13(a) bring these out nicely. The only significant
difference is that the fᅊ minor triad that begins the pizzicato unison passage is
194  Audacious Euphony

abbreviated, affording the westward L/R chain a bonus measure with which to
overshoot G major, so that tonic C major rather than dominant G major arrives at
the next hypermetric downbeat.
Because figure 8.13(a) crawls fairly low to the ground, its explanatory power is
somewhat attenuated. One stumbles through the darkness, eventually blundering
into the light of C major, blinking and disoriented. Figure 8.13(b) proposes an
alternative hearing that remains oriented to C major throughout, organizing the
progression of harmonies about an L/P (hexatonic) chain that unfolds at the mid-
dleground. Eᅈ major and B major are the most salient stations of that chain, by
virtue of the thematic statements at mm. 204 and 212, respectively.13 Eᅈ minor,
which arises at m. 207 through the substitution of Gᅈ for G and returns at m. 211,
is prolonged in the interim through a retention loop about Bᅈ, before progressing
to Cᅈ-qua-B major when the retained Bᅈ discharges as a leading tone. B minor
arises through a similar substitution, is prolonged through a retention loop around
Fᅊ, and proceeds to G, the actual dominant, which in the event is overshot on the
way to tonic.
From the standpoint of C major, the entire chromatic episode can be seen to
fulfill a dominant function, with the entire alley acquiring the function of its most
diatonic member. By extension, the pitch retention loops visit the region of the
double dominant. But the work carried out in chapter 6 suggests that such an
interpretation is not mandatory, if all it does is provide a language with which to
express the intuition that the passage executes a departure–return scheme. The
voice-leading zones provide an alternative means for capturing the same intuition.
They suggest that the motion from C major to eᅈ minor represents a downshift,
that the motion through the hexatonic alley balances along the fulcrum, that the
retention loops temporarily move to a lower voice-leading zone, and that only in
the final overshoot from G major to C major is the initial downshift reversed.
In this Brahms excerpt, as in the Faith Proclamation, diatonic encapsulations
emerge from triadic drifts along the L/R chain. In the Wagner case, the forces that
halted the rightward drift and erected the diatonic walls included temporary rever-
sal of the drift, accretion of supplementary dissonances, change in harmonic
rhythm, and the flanking of a triadic goal by its syntonic images. Brahms invokes
rather different resources to halt his leftward drift. There are no supplementary
dissonances, no reversals, and no syntonic images. His walls are erected instead
through the sounding of a diatonic and triadic theme on a hypermetric downbeat.
Where Wagner methodically stakes his tent using structural means, Brahms simply
plants his feet using what Harrison (1994, 81) refers to as rhetorical ones.

13. The hexatonic connection of Eᅈ major to G major, along the same path as figure 8.13(b), arises in
the final section of Brahms’s Alto Rhapsody, a c minor composition from six years earlier that also
transforms to C major in its final section.
C HA P T E R
Nine
Double Syntax and the Soft Revolution

A Summary Example from Schubert

We end, as we began, with an excerpt from Schubert’s instrumental music, whose

consideration will bring together the central analytic concerns of this book
and provide a focus for the central topic of this chapter. The passage, from the
scherzo of the C major Symphony, begins at m. 93 with a waltz fragment that is
presented first in the dominant and then in the tonic and ends with a dominant
prolongation that precedes the reprise of the Scherzo at m. 153 (figure 9.1). Its
interior, from m. 109, forms a chromatic tunnel that recedes from C major in
stages, loses contact with it entirely as the hypermeter broadens to six-bar units
at m. 117, and then gradually reattunes to that key as four-bar units are restored
at m. 141.
The initial chromatic events are the Aᅈ and Dᅈ neighbors that sound in mm.
109 and 110. These receive subsequent harmonic support respectively from
f minor at m. 113 and Dᅈ major four bars later. To a monotonal ear, these tones
are products of phrygian mixture. To one with more modulatory flexibility they
represent a tonicization of f minor, the minor subdominant and nebenverwandt.
The honorary diatonic status of the nebenverwandt relation in chordal space
papers over what is already a fairly significant journey through key space,
from zero flats to four. Exiting the tunnel, C major comes back into view when
A major (m. 129) proceeds to d minor (135). When the latter assumes its diatonic
role as a supertonic, the former is heard as its applied dominant, also evincing a
nebenverwandt.
Approached from either portal, the cᅊ minor triad at m. 123 is inscrutable.
Entering from Dᅈ major, it is heard as dᅈ minor, modally inflecting a chord that
already presses the limits of modal mixture in C major.1 Exiting through A major,

1. Here we encounter a peculiarity in the theory of modal mixture. Phrygian occupies the extreme end
of modal space, involving four flatwise substitutions on major. The other extreme, lydian, requires
four sharpwise substitutions on minor. The current textbook sanctioning of phrygian but not

195
196  Audacious Euphony

Figure 9.1. Schubert, C major Symphony, 3rd mvt., mm. 101–53.

it is heard as cᅊ minor, a local diatonic mediant. But there exists no enharmonic

construal of the triad in m. 123 that renders it diatonic in either C major, or in the
locally tonicized f minor and d minor that directly flank it, or even in the parallel
modes of any of those three keys.

lydian mixture rests on Schenker’s whims and tastes a century ago rather than that of systematic
consistency and completeness.
 CHAPTER 9 Double Syntax and the Soft Revolution  197

Riemann’s function theory, because it is relatively free of diatonic constraints,

is helpful here, providing a well-formed and plausible reading of cᅊ minor as
the Leittonwechsel of A major, the dominant of the supertonic.2 One benefit
of this approach is that it calls attention to a symmetry in the passage: cᅊ minor’s
protensive relation to an emerging supertonic inverts Dᅈ major’s retensive relation
to a receding tonic. Transformations express that symmetry: C major is trans-
posed up a semitone to Dᅈ major through the transformational sequenceᇳN, Lᇴ .
cᅊ minor is transposed to d minor by the same transformations applied in reverse
order,ᇳL, Nᇴ .
The Tonnetz at figure 9.2(a) provides a geometric expression of that same
symmetry. An initial southwest arrow exits the C major capsule along an NL axis,
connecting to Dᅈ major. A second arrow along a parallel axis charts the reentry
into the C major capsule, connecting from cᅊ minor to d minor, which dis-
charges its supertonic function according the syntactic demands of life within the
capsule.
Riemannian theory is thus able to assign every harmony in the passage a
function with respect to the C major tonic. But there still remains the problem of
how to hear the enharmonic Dᅈ major → cᅊ minor junction across the barline that
precedes m. 123. The dotted arrow that connects them in figure 9.2(a), which
seems to travel a great distance, belies the aural perception that these two triads are
proximate, just as Schubert’s conversion from sharps to flats confuses the visual
perception that the triads share individual tones. Honoring those perceptions,
while retaining fidelity to some classical theory of tonality, will never remove the
problem; it will only displace it to some further outpost. If Dᅈ at m. 117 becomes
“Cᅊ” at 123, which becomes “Cᅊ” at 129, then “Cᅊ” at 129 is heard as Dᅈ and
cannot function as the leading tone of the supertonic at m. 135. Conversely, if Gᅊ
at m. 123 is inherited from “Aᅈ” at m. 117, which is what becomes of “Aᅈ” at
m. 113, then “Aᅈ” at m. 113 is heard as Gᅊ and cannot function as the flatted sub-
mediant of the tonic triad at m. 109. This circle can never be squared by diatonic
logic. For chromatic triadic progressions that move through the enharmonic seam,
seven-tone diatonic space provides no adequate stick for measuring distances.
The solution comes easily once we relinquish the diatonic gauge for a chro-
matic one. Cᅊ and Dᅈ become notational variants of a single object. Like the
Morning Star and the Evening Star, they achieve identity by intension as well as
extension, by sense as well as by reference. Equivalence of root distance then fol-
lows automatically: the unified Cᅊ/Dᅈ object is heard as equally distant from the
F that precedes it and the A that follows it. What is compromised in this move is
the syntactic position and semantic force of that tone. From a protensive phenom-
enological perspective, it can be heard as subdominant agent of f minor, and from
a retensive one as dominant agent of d minor. But, because there is no reason to
prefer one of these vistas over the other, the sound notationally indicated by Cᅊ at
m. 129 is essentially neither Cᅊ nor Dᅈ. The ability of that sound to coexist at once
in both diatonic spaces results from its position in chromatic space, equidistant

2. Schubert also transposes a minor tonic down by semitone in “Auf dem Flusse,” studied in chapter 6.
McCreless (1996, 88–89) identifies a similar modulation in Beethoven’s Trio, Op. 1 no. 3.
198  Audacious Euphony

from F and A. Although the chord still retains its forked syntactic and appellative
residues in diatonic space, what comes to the fore as syntactic driver is the mini-
mal change relation between adjacent triads: each motion requires only a single
semitonal motion.
Once attuned to the minimal change at the center of the passage, that attune-
ment disperses toward its exterior. We might then hear C major → f minor (m. 109)
and A major → d minor (m. 135) as double semitonal upshifts. (This perception in
no way overrides or otherwise conflicts with our simultaneous perception of their
classical relation to C, measured in diatonic fifths.) This might then lead us to
notice that these two double upshifts are bridged by a cumulative semitonal upshift,
as f minor progresses through the hexatonic alley to A major. Pressing one stage
further on both flanks, to G major → C major at m. 105 and d minor → G major at
m. 141, we might now notice their participation in the same upshift, although
taking larger strides in voice-leading space.
Figure 9.2(b) records this second hearing on the Tonnetz. No diatonic region
is encapsulated. All triadic locations are governed by common-tone relations.

Figure 9.2. Two Tonnetz analyses of figure 9.1.

 CHAPTER 9 Double Syntax and the Soft Revolution  199

A wave of voice-leading efficiency crests at the hexatonic alley, whose balanced

voice leading temporarily suspends the pervasive upshift across the entire pas-
sage and then recedes in the rearview mirror. The voice-leading perspective of
figure 9.2(b) is inverse to the monotonal one of figure 9.2(a): the one focuses as the
other blurs. What we initially conceived as a tunneling into and out of diatonic/
tonal darkness becomes, from the standpoint of voice leading, a transient emergence
into the hexatonic light.
The central claim of this book is that these two incommensurate ways of
measuring triadic distance emerge respectively from two independent proper-
ties of consonant triads. Listening to the passage with reference to C major, as it
disappears into chromatic obscurity, crosses the enharmonic crevice, and reemerges
back into diatonic clarity, we are hearing in terms of a system of distance relations
governed by the degree to which harmonic roots are consonantly related to each
other and to C major. That system is predicated on the membership of C major and
its satellites in a family of objects, the consonant triads, that uniquely optimize
acoustic consonance. Listening to the same passage attuned to voice leading, we
hear an elimination of chordal dissonance and the elimination of whole steps in
individual voices (m. 109), then a reduction from two moving voices to one
(m. 117), a central passage that both maximizes common-tone retention and min-
imizes voice-leading sizes (mm. 123–40), and an increase in voice-leading work
synchronized with a restoration of chordal dissonance (m. 141). Attuning to these
processes, we are hearing trichords as participants in a system predicated on their
nearly even status, as minimal perturbations of the augmented triad that divides
the octave into three equal parts.

Double Syntax and Its Skeptics

My characterization of sense-making in response to the music of figure 9.1 relies on

an assumption of double syntax that some scholars have found implausible.
Objections to that assumption come in two forms: an ontological/aesthetic one that
ascribes immanent properties to compositions, and an epistemological/cognitive
one that describes mental capacities in response to them. The ontological objection,
characteristic especially of American music theory before 1990, is that music of
high aesthetic value is organically unified and hence cannot be generated from mul-
tiple sources. This objection is sometimes framed with an appeal to the alleged
unity of natural language. For example, Charles J. Smith, responding to work of
Gregory Proctor (1978) that had elements that could be characterized in terms of
double syntax, wrote that Proctor’s approach “suggests the separateness of chro-
matic and diatonic tonalities, since it treats them as virtually distinct languages.
I cannot believe that any chromatic master conceived of his musical terrain as so
partitioned; I hear no grinding of gears as one area is left and the other entered”
(Smith 1986, 109). Claims of double syntax sit uneasily with postwar music
theory’s commitment to the idealist notion that a good composition resembles an
organism in its indivisibility. Although that metaphor has become epistemically
200  Audacious Euphony

and historically bracketed in the last twenty years, and to a certain extent has
receded altogether from the explicit discourse of American theorists, I suspect that
it nonetheless continues to subliminally work on the instinctive sensibilities of
music theorists.
Smith’s objection can be reframed so as to transfer its weight from immanent
properties to human listening capacities. Thus, Carl Dahlhaus writes that “the
supposition that the musical hearing switches between tonal and non-tonal appre-
hension during a composition or a phrase would be problematic” (1967, 100–1;
my translation). Although directed toward an earlier historical moment when “non-
tonal apprehension” is of a quite different nature than that which concerns us here,
Dahlhaus’s admonition is framed as a universal claim about human capacities.
A more explicit psychological turn is evident in a similar formulation of Fred
Lerdahl, who also shares Smith’s perception of a seamless unity in the music itself:
“The argument that listeners switch between two systems . . . is implausible as a
psychological position. It is problematic in particular for late tonal music, which
moves smoothly between diatonicism and chromaticism even within a single
phrase” (2001, 85).3
Neither Dahlhaus nor Lerdahl presents evidence in support of his judgment,
implicitly relying instead on its appeal to some species of common sense. That spe-
cies may nonetheless be broadly enough shared that it is able to bear the weight of
the objection. Dahlhaus, Smith, and Lerdahl evidently voice their objections to
double syntax independently of one another and converge on those objections
at three distinct historical moments, from three distinctive eras and intellectual
traditions. This convergence suggests that the double syntax hypothesis touches a
broad cultural nerve, clashing not only with idiosyncratic habits of thought sedi-
mented within the microecology of some sequestered musicological terrarium but
also with ideas that are deeply lodged in the experiences and sensibilities of the
late-twentieth-century Euro-American music scholar and composer. Accordingly,
it seems likely that the responses of Dahlhaus, Smith, and Lerdahl represent those
of a much broader population of music scholars, who perhaps share the impulse to
reject the claims advanced in this book on the grounds that they are untenable on
a priori grounds, which is to say, independently of any evidence that might be
brought to bear in support of them.
My response to these objections takes three forms. First, the door that Smith
attempted to close with his appeal to the autonomy of natural languages I will
reopen through an appeal to multilingualism, a branch of linguistics developed in
the quarter-century since Smith lodged his objection. Second, by reflecting on
some quotidian modes of experience, and speculating on the mental operations
that underlie them, I seek to establish that there are other sorts of common sense
than those to which Dahlhaus and Lerdahl implicitly appeal when they dismiss
double syntax on the basis of some tacit theory of cognitive capacities. Finally, and
most speculatively, I suggest that double syntax provides a framework through

3. Lerdahl’s diatonicism/chromaticism dichotomy comes in the context of a critique of neo-Riemannian

theory and thus is equivalent to both the classical/pan-triadic distinction of this book and to
Dahlhaus’s tonal/non-tonal distinction.
 CHAPTER 9 Double Syntax and the Soft Revolution  201

which to achieve a perspective on some aspects of the historical developments that

connected the First and Second Viennese Schools, such that we can see how the
one in some sense “became” the other. It thus suggests an approach to a problem
that is often swept under the rug for lack of alternatives.

Code Switching and Double Determination

The recent evidence from linguistics provides ample motivation for questioning
Smith’s assumption, characteristic of its era, that natural languages are rigidly dis-
crete, autonomous systems, compartmentalized and segregated in the minds of
speakers and hearers. It turns out that most people have the ability “to alternate
between different linguistic varieties at the drop of a hat, and to make use of
several simultaneously when it suits them to do so” (Gardner-Chloros 2009, 177).
Although this ability might seem a special case or odd trick within the demo-
graphic pool that stocks the Society for Music Theory with members (i.e.,
Anglophones above the poverty line), for most outside that pool it is an unculti-
vated norm of linguistic behavior from birth. Penelope Gardner-Chloros observes
that “if you add together people who live in multilingual areas of the world . . . ;
people who speak a regional language or dialect on top of a national language . . . ;
and migrants and their descendants . . . , you are left with small islands of mono-
lingualism in a multilingual sea” (2009, 5–7).
Linguists refer to language juxtaposition as code switching, a term that is used
by consensual inertia even though it is widely acknowledged to be misleading in
several dimensions (Gardner-Chloros 2009, 11–12). The fault lines along which
competing theoretical models of code switching fracture may have familiar reso-
nances for music theorists. The “classical” model developed in the early 1990s
maintains that interacting languages are asymmetrically related: one is the “matrix”
into which the lexical items of the other are inserted (Winford 2009). More recent
evidence suggests that many code switchers hold equivalent knowledge of the two
languages, shuttling between them as equals. A second question pertains to the
distinctiveness of the languages when placed into dialogue with each other. The
early view is that the juxtaposed languages have autonomous status and thus that
the “switch” between them crisply snaps off at a precisely locatable juncture
between words. Katherine Woolard polemically responds that “alternation between
rigidly discrete systems [is] a mythic ground of analysis that is as empirically
and theoretically ill-founded as the idealized speaker/hearer” (1998, 6), and a
considerable body of evidence has accumulated in support of the complexity of
interlinguistic transition (Gardner-Chloros 2009, 45).
Of particular relevance to the approach presented here is the phenomenon
that Pieter Muysken identifies as congruent lexicalization. Where language popu-
lations overlap, share a significant geographic boundary, or associate through
trade, colonization, or migration patterns, there often exist words that have similar
form and meaning in both languages. These cognate terms, or homophonous
diamorphs, are thought to “cause, or at least facilitate, a codeswitch from one
202  Audacious Euphony

language to the other” (Broersma and de Bot 2006, 2; see also Muysken 2000, 123).
Their function thus resembles that of pivot chords in theories of modulation: a
dual-language zone intercedes between language 1 and language 2, so that “it is
very possible to know that a definite codeswitch has occurred even when we are
unable to say at what point it occurred” (Woolard 1998, 7–8).
The topic becomes acutely relevant to music when viewed in terms of the
mental processes of the individual language user, equivalent to the level at which
both Dahlhaus and Lerdahl register their skepticism. On a hot summer day in
Alsace, Gardner-Chloros (2009, 45) recorded an office worker complaining on the
phone about a malfunctioning air-handling system. “The whole extract was deliv-
ered at high speed and the fifteen or so switches [between French and Alsatian] in
no way interrupt the flow.” Moreover, exchanges of this type tend to occur beneath
the horizon of consciousness: “Speakers are on the whole not very aware of their
code-switching behaviour, and tend to be surprised at their own performance if
you play it back to them” (121). The behavior of listeners has been less studied;
although some neural researchers have begun to measure electrical activity in
response to bilingual speech, the evidence is as yet preliminary (140). Yet it seems
safe to infer that, like any other competent language user, a “code switcher” would
spontaneously adjust her elocutionary pace to the perceived tolerances of her lis-
tener. If the Alsatian office worker is code switching rapidly in an effective conver-
sation, it is likely that her interlocutor is receiving her switches with as few hitches
as she has in creating them.
Evidence from bilingualism, then, suggests that the human brain has ample
potential for coprocessing distinct languages within compressed time spans, with-
out cognitive hiccups. That potential is amplified when the interacting languages
possess intersecting lexical units, which serve as bridges that facilitate the exchanges
between the constituent languages. Such a unit performs distinct roles in the dis-
tinct languages, and functions as a seam through which a speaker slips from one
language to another, bringing along her appropriately acculturated interlocutor
without stumble. It is overdetermined, functioning both with respect to the
language that precedes it and that which follows. Moreover, the entire process
is cognitively opaque, transacted beneath the horizon of awareness.4
This evidence does not in itself mandate any conclusions about music, which
is no language. Although the overlap is significant enough to have fueled 2,500
years of metaphorical cross-appropriation, there are also significant blocks in the
metaphorical pipeline. In this connection, it is often noted that music is a form of
art practiced by specialists, not of universally practiced discourse. An aesthete
might then assert that the existence of some quotidian ability or tendency may
have no implications for our understanding of the ineffably sublime masterwork.
Such claims come less easily, however, if one’s concerns are with music as received
rather than produced. Recent work at the music/language boundary suggests that

4. There would seem to be a number of other ways that code-switching research could stimulate
the thinking of musical scholars, particularly given the intensive interest in hybridity among
ethnomusicologists and in cross-national encounters among historical musicologists. Evidently,
only Mark Slobin (1992) has mined this field; the theoretical and empirical explosion in bilingual
studies suggests that a return visit might be productive.
 CHAPTER 9 Double Syntax and the Soft Revolution  203

“as cognitive and neural systems, music and language are closely related” (Patel
2008, 417). If one has the impulse to dismiss the double syntax hypothesis on the
basis of hypothesized cognitive capacities and proclivities in response to acousti-
cally delivered data, the evidence from bilingualism provides a strong motivation
for reevaluating that impulse and the hypotheses that underlie it.

Cognitive Opacity

There is reason, moreover, to believe that overdetermination is cognitively opaque

not only in the limited case of language but also more broadly in our quotidian
interactions with the world about us. Consider, for example, how we conceive our
interactions with garments, which perform (“optimize”) at least three distinct
functions: they protect the body (from natural elements for survival), conceal it
(from socially undesirable exhibiting), and adorn it (to signal social position
and in support of sexual selection). Usually a dress or sweater performs multiple
functions simultaneously, conflating them in our awareness. But one can imagine
situations that isolate these functions from each other. Alone in your remote
wood-heated cabin in December, you pull on a rugged old sweater. You have no
need to conceal or adorn your body; there is no social function to be performed.
You are simply protecting your body from the cold. Asleep in the city on a warm
June night, you are woken by your dyspeptic dog; you throw on an oversized
T-shirt and grab his leash. You have no need for allure, prestige, or warmth. You
simply want to conceal your body from voyeurs. Back at the cabin in August, you
anticipate an intimate evening by slipping into an item recently ordered from an
apparel catalogue. You have no need to conceal your body from your partner or to
protect it from the elements. You simply want to project an allure. In each instance,
you go through the same sequence of actions and experience the same set of
sensations: open the drawer, locate the item, check the label for orientation,
raise arms, locate headhole, feel the garment slip down your torso, smooth,
straighten, tuck. As you proceed through this familiar series of actions and
experiences, you may feel no motivation to keep these functions mentally com-
partmentalized, to introspect on which you are fulfilling or in what combination.
Dressing may feel pretty much like a unitary act. Although we may use the same
object to fulfill quite distinct, nonoverlapping functions, in our minds the func-
tions can merge into a single set of subroutines. We move between them without
“grinding the gears.”
The clothing analogy may lack force in one respect: when double syntax is
claimed for music, the functions are not merely swapped, one for the other.
Something potentially more disjunctive occurs: there is a sense in which their
terms exchange function. When the triads at the boundaries of the Schubert
Scherzo (figure 9.1) are diatonically encapsulated by C major and ordered
according to eighteenth-century norms, they are referred to C for their mean-
ing. Equivalently, depending on theoretical tradition, one might say that each
triad is heard in terms of, is oriented toward, functions with respect to,
204  Audacious Euphony

is generated/controlled by, or is hierarchically subordinate to C. I will use the

expression “triad (3) → tonic (1)” to generalize across each of these expressions
as a neutral party; the parenthesized numbers refer to cardinality, to prevent
confusion between tonic-as-tone and tonic-as-triad.
The formulation lends itself to recursive expansion. When C is tonic, the
triadic tones orient the diatonic ones, which in turn orient the chromatic ones,
leading to the expanded expression chromatic (12+) → diatonic (7) → triad (3) →
tonic (1).5 By contrast, when the triads at the interior of that same passage are
released from the diatonic capsule and the enharmonic seam breached, they are
referred to the entire chromatic collection for their meaning. Each individual tone,
including C, now “takes its meaning from” (functions with respect to, orients
toward, etc.) the triad of which it is a constituent, which in turn “takes its meaning
from” its nearly even status with respect to the chromatic collection. Accordingly,
the arrows flip; we now have chromatic (12) ← triad (3) ← tone (1). It is as if the
arrow that initially points from chromatic (12+) to tonic (1) is attached to the system
at the point of the triad (3), where it rests on a spring-loaded pivot and is locked into
position by the diatonic (7). When the diatonic lock is released, the spring uncoils
and the direction of the arrow instantly reverses. It requires some diatonic reconsti-
tution to reload the spring, and some cadential labor to lock it back into its initial
position, pointing away from the chromatic (12) toward the tonic (1).
The metaphor of arrow reversal helps explain why double syntax encounters
such a deep vein of resistance in the musical case. Its advocates are not merely sug-
gesting that listeners exchange modes of organizing relations “on the fly.” They are
suggesting that all of the relations of subordination, causality, and orientation
undergo a reversal. It is as if at one moment the person wears the garment and at
the next the garment wears the person.
A second round of introspection suggests, however, that causal inversion also
has a potential for cognitive opacity. On Tuesday, you drive your car to town for a
tune-up. On Wednesday, you drive your car to town so that you can get your hair
cut. In the first case, you take the car. If the car could get to the shop without you,
you would gladly send it. In the second case, the car takes you. If the barber were
near enough, you would leave the car in the garage. Comparing these two situa-
tions, the arrow of causality is reversed. Yet, from a cognitive standpoint, one
hardly notices the difference. Find your keys, unlock the car, warm it up, buckle
the seat belt, down with the handbrake, down with the clutch, car into reverse,
open garage door—it’s all the same. These reversals are insidious; they float below
the threshold of consciousness unless one exerts some effort through analysis. Do
you want money to get power, or power to get money? Do you seek wealth to buy
a yacht, or a yacht to demonstrate your wealth? Do you smile because you are
happy, or to elicit a mirroring response that will make you happy?
Even if these introspective musings identify some intersubjective tendencies,
as I suspect they do, one is not perforce compelled to accept their application

5. Compare the cone in Krumhansl 1990, 128, replicated in Lerdahl 2001, 46. The reference to “12+”
chromatic elements indicates that, when enharmonic distinctions are honored, there is no crisp
cutoff in the number of distinct chromatic pitch classes.
204  Audacious Euphony

is generated/controlled by, or is hierarchically subordinate to C. I will use the

expression “triad (3) → tonic (1)” to generalize across each of these expressions
as a neutral party; the parenthesized numbers refer to cardinality, to prevent
confusion between tonic-as-tone and tonic-as-triad.
The formulation lends itself to recursive expansion. When C is tonic, the
triadic tones orient the diatonic ones, which in turn orient the chromatic ones,
leading to the expanded expression chromatic (12+) → diatonic (7) → triad (3) →
tonic (1).5 By contrast, when the triads at the interior of that same passage are
released from the diatonic capsule and the enharmonic seam breached, they are
referred to the entire chromatic collection for their meaning. Each individual tone,
including C, now “takes its meaning from” (functions with respect to, orients
toward, etc.) the triad of which it is a constituent, which in turn “takes its meaning
from” its nearly even status with respect to the chromatic collection. Accordingly,
the arrows flip; we now have chromatic (12) ← triad (3) ← tone (1). It is as if the
arrow that initially points from chromatic (12+) to tonic (1) is attached to the system
at the point of the triad (3), where it rests on a spring-loaded pivot and is locked into
position by the diatonic (7). When the diatonic lock is released, the spring uncoils
and the direction of the arrow instantly reverses. It requires some diatonic reconsti-
tution to reload the spring, and some cadential labor to lock it back into its initial
position, pointing away from the chromatic (12) toward the tonic (1).
The metaphor of arrow reversal helps explain why double syntax encounters
such a deep vein of resistance in the musical case. Its advocates are not merely sug-
gesting that listeners exchange modes of organizing relations “on the fly.” They are
suggesting that all of the relations of subordination, causality, and orientation
undergo a reversal. It is as if at one moment the person wears the garment and at
the next the garment wears the person.
A second round of introspection suggests, however, that causal inversion also
has a potential for cognitive opacity. On Tuesday, you drive your car to town for a
tune-up. On Wednesday, you drive your car to town so that you can get your hair
cut. In the first case, you take the car. If the car could get to the shop without you,
you would gladly send it. In the second case, the car takes you. If the barber were
near enough, you would leave the car in the garage. Comparing these two situa-
tions, the arrow of causality is reversed. Yet, from a cognitive standpoint, one
hardly notices the difference. Find your keys, unlock the car, warm it up, buckle
the seat belt, down with the handbrake, down with the clutch, car into reverse,
open garage door—it’s all the same. These reversals are insidious; they float below
the threshold of consciousness unless one exerts some effort through analysis. Do
you want money to get power, or power to get money? Do you seek wealth to buy
a yacht, or a yacht to demonstrate your wealth? Do you smile because you are
happy, or to elicit a mirroring response that will make you happy?
Even if these introspective musings identify some intersubjective tendencies,
as I suspect they do, one is not perforce compelled to accept their application

5. Compare the cone in Krumhansl 1990, 128, replicated in Lerdahl 2001, 46. The reference to “12+”
chromatic elements indicates that, when enharmonic distinctions are honored, there is no crisp
cutoff in the number of distinct chromatic pitch classes.
 CHAPTER 9 Double Syntax and the Soft Revolution  205

to music. We are still free to insist, as a metaphysical priority, that music is differ-
ent, that our responses to it are isolated not just from language but from all other
mental functions, such that whatever we say about the functioning of the brain
does not apply to its functioning in response to music. But such a reversal flies in
the face of evidence that causal reversal is cognitively opaque in music, too. At the
opening of Parsifal, G is a neighbor to Aᅈ; soon enough, we recognize that Aᅈ has
become a neighbor to G. Eᅈ major is dominant of Aᅈ; soon enough, we recognize
that Aᅈ has become the subdominant of Eᅈ. Arsis and thesis flip, so that strong
beats become weak and vice versa. A cunning composer or crafty performer knows
how to take advantage of the opacity of causal reversal, to flip the arrow of causal-
ity without rocking a listener’s boat. If musical cognition is not entirely remote
from other sorts of mental processing (Zbikowski 2002), then we cannot dismiss
the possibility that switching between musical syntaxes might transpire beneath
our horizon of awareness, especially when the switching is routed through an
overdetermined object such as a triad or a dominant seventh chord.

The Soft Revolution

In this music of Schubert, so gorgeous and so tonally secure to the foregrounded

ear, the position of the tonic as functional center is placed into question for the first
time, like a soft revolution.
—Diether de la Motte, Harmonielehre

In the age of Mozart, musical pitches are primarily organized by diatonic tonality.
In the age of Webern, there exists some species of music whose tones are orga-
nized in some other way. How does music that is heard to be organized by
diatonic tonality become music that is heard to be organized in some other
way? Either that “other way” is conjured tabula rasa from what was absent from
diatonic tonality, or it is conjured partly by what was present in it. The first possi-
bility is no possibility at all; it is inconsistent with everything else we know about
historical process and human cognition. That leaves the second possibility: that
there is some aspect of diatonic tonality that was reshaped, recontextualized,
developed in some direction that had been hitherto inconceivable (compare
Morgan 1998, 3–4).
Where is the thread that unravels the garment and then gets rewoven? Not,
evidently, in the sound combinations, which are complementary: acoustic, root-
indexing harmonies in the age of Mozart; anything but, in the age of Webern. Nor,
evidently, in the relations between sounds: in the former, harmonies gain their
systemic function in their relations to tonics; in the latter, harmonies gain their
systemic function according to their properties and potentials within the chro-
matic system. Neither lexicon nor syntax furnishes a lever that can trigger a phased
process of change, such as would be consistent with the evolutionary model
implied by the second possibility. The burden of historical continuity can be
distributed onto the backs of motive, rhythm, and form. But the primacy of
206  Audacious Euphony

pitch renders such a move artificial: motivic relations rely on similarity and equiv-
alence judgments that take shape in part within a set of systematic assumptions;
rhythm is heard in the context of a meter that is established in part by rate of
harmonic change; formal articulation in part depends on what counts as closure
and repetition. All such considerations in some measure presuppose some prior
organization of pitch relations.
A theory of double syntax suggests a way out of this impasse. It allows that at
the same time that triads are root-indexing consonances, they have the capacity to
serve as something else entirely: nearly even chords participating in a system of
tight voice leading. Under cover of a unified lexicon, a composer learns to hear a
different set of syntactic possibilities, and to train listeners to that hearing. This
step can be seen as the initial, evidently innocuous stage in a process that eventu-
ates in the music that arises and characteristically represents the terminal end of
the long European century.
One way of conceiving of this process, following Patrick McCreless’s appro-
priation of evolutionary theory, is to say that the triad is preadapted for this second
role. In a passage that McCreless (1996, 108–10) quotes from evolutionary biolo-
gist Stephen Jay Gould, preadaptation “asserts that a structure can change its func-
tion radically without altering its form as much. We can bridge the limbo of
intermediate stages by arguing for a retention of old functions while new ones are
developing” (Gould 1977, 108). For example, fish jaws “were well designed for
their respiratory role; they had been selected for this alone and ‘knew’ nothing
of any future function. In hindsight, the bones were admirably preadapted to
become jaws. The intricate device was already assembled, but it was being used for
breathing, not eating” (108).
McCreless suggests (1996, 110) that classical tonality is preadapted to chro-
matic space, but it is not clear from his account what aspect of classical tonality
assumes the role of the fish’s jaw. That burden falls variously on “structural semi-
tone harmonic relations” (110), directed linear motion through chromatic space
(98–99), and sequential transposition in chromatic space, on the basis that each
initially arises in passages that are fully reconcilable to the seven-tone diatonic and
is subsequently reused by composers in ways that can only be understood within
the twelve-tone chromatic. Yet none of these phenomena is ever intrinsic to a dia-
tonic system, in the way that the jaw is proper to the fish. They arise only when a
diatonic system is supplemented by chromatic tones, at which point the evolution-
ary train has already left the station.
The work presented in this book suggests a different candidate: it is the conso-
nant triad, fully native to a nonchromaticized diatonic space, that plays the crucial
role. Triads are preadapted to play the role of voice-leading optimizers by virtue of
their near evenness, which manifests in both diatonic 7-space and chromatic
12-space. Triads first assume that role in the diatonic sequence, when, as described
by François-Joseph Fétis (2008 [1844]), they cease to exploit the orienting, or
“summoning,” capacity of that scale’s tritone and instead exercise their capacity,
described in Agmon 1991, for intertriadic stepwise voice leading within a quasi-
flat diatonic space. Moving through a Baroque ritornello, from presentation to
Fortspinnung to cadence, the arrow of orientation first flips outward, then back in,
 CHAPTER 9 Double Syntax and the Soft Revolution  207

as triads move with ease between their dual natures in diatonic space. This dynamic
is replicated and intensified in chromatic space, which presents a new habitus in
which intertriadic voice leading proves capable of even greater parsimony. Moving
through the Schubert Scherzo of figure 9.1, the arrow of orientation undergoes a
similar journey, only this time without need for sequential repetition to sustain it.
Further evolutionary stages can be provisionally proposed, at the risk of
bleaching the nuance out of a complex set of musical transformations through a
few broad strokes. At the second stage, it is the syntax, in the form of a prioritiza-
tion of voice-leading efficiency, that is firmly planted, while the lexicon is
exchanged. If consonant triads can serve as objects in a system of relations defined
by common-tone preservation and voice-leading work, so too can other nearly
even chords, the members of the Tristan genus, beginning with the omnibus pro-
gressions that arose already in the eighteenth century (Telesco 1998), and eventu-
ating in late Wagner. We are now at one further level of remove on the historical
continuum: the syntactic arrow has been reversed, and a hitherto secondary
member of the lexicon has been promoted to primary status.
A third step involves another lexical transformation, but this time at the scalar
rather than the chordal level. The most characteristic harmonic progressions to
exploit the syntax for which the triad was preadapted are the L/P-chaining major
third cycle, and the R/P-chaining minor third cycle. These progressions respec-
tively create six-tone (hexatonic) and eight-tone (octatonic) scales that take their
place as substitutes for the seven-tone diatonic gamut. An alternative route to
the octatonic scale is through the parsimonious organization of Tristan-genus
seventh chords. Although arising initially as secondary by-products of harmonic
progressions that exploit the capacity for efficient voice leading among nearly even
trichords and tetrachords, eventually these scales become things in themselves.6
Individual tones now relate directly to these new scales, without the intervention
of nearly even trichords or tetrachords. Thus arises the scalar tonality associated
with the tradition that extends from Liszt through Rimsky-Korsakov to the early-
twentieth-century Russian and French composers, recently explored in chapter 9
of Tymoczko 2011b. The triad can now be replaced by other dissonant formations,
octatonic and hexatonic subsets. At the fourth and final stage, hexatonic and
octatonic scales are absorbed into a broader atonality. Chords and scales are
transformed into the sets and set classes of Second-Viennese-School atonality,

6. Richard Taruskin (1985, 1996, 2005) also emphasizes that octatonic collections initially arise as
by-products of other procedures of diatonically contextualized chromaticism. Taruskin’s claim that
Liszt is the first composer to lay out the octatonic scale in a single voice and to “categorically express”
circles of minor thirds as harmonic progressions (2005, 429) is supportable only on an idiosyncratic
reading of “categorical.” Theorist Honoré de Langlé (1797) included a circle of minor thirds with
octatonic bass among his sequential tours de l’harmonie. Thomas Christensen (1992, 9) suggests an
even earlier precursor from mid-seventeenth-century Spain, although it is more likely a modulatory
blueprint than a directly sounded progression. The finale of Beethoven’s “Acht Variationen über
Tändeln und Scherzen” of 1799 presents a complete series of minor-third transpositions (Seidel
1963, 201). See also figure 5.4 in chapter 5, from an 1821 composition of Schubert’s. The passage from
the Scherzo of Beethoven’s Ninth Symphony (1822), referred to on p. 88 and presented as Web score
5.9, lays out octatonic collections in each of its constituent voices, including the melodically prominent
top voice, as a by-product (Cohn 1991). Blum 1986 raises similar questions about Taruskin’s claim.
208  Audacious Euphony

which take their meaning directly, like the consonant triad at stage 1, in relation to
the chromatic 12-gamut.
In each of the latter stages, the arrow of orientation remains sprung, pointing
from smaller-cardinality entities toward larger ones. All of the stages are in that
sense “atonal,” although only gradually do they begin to take on the sonic proper-
ties that we associate with the prototypical atonality of Schoenberg and Webern.
What distinguishes the later stages from the earlier ones is the number of degrees
of remove from diatonic tonality. At the moment that diatonic scales, triads, and
Tristan-genus seventh chords lose their privileged lexical positions, the ontoge-
netic road of return becomes bumpy. Once composers begin to use [0148]-type
chords in a hexatonic context, or [0146]-type chords in an octatonic one, the
imposition of a diatonic cadence often sounds like the grinding of gears that Smith
describes.
The lexical substitutions of the later stages are relatively easy to effect. It is the
syntactic substitution at the first stage that provides the key that unlocks the cap-
sule of diatonic tonality. The central argument of this book is that the consonant
triad has two natures, and that those two natures lead to two syntaxes. The central
argument of this chapter is that the juxtaposition of syntaxes is not as problematic
as is sometimes thought, particularly when smoothed over by lexical continuity.
The central suggestion of this penultimate section is that lexical continuity can
bridge over a syntactic impasse along the historical axis, as well as within a phrase
of Schubert.

On Musical Overdetermination

Since no later than the early eighteenth century, Western musicians and listeners
have, as a first-level default, interpreted musical events with respect to a tonic that
is expressed through the coordination of consonant triads and diatonic collec-
tions. Independently, composers cultivated a taste for economical voice leading
and harmonic consistency and sought harmonic structures that could respond to
them both simultaneously. What they were seeking was sitting right in front
of their ears, preadapted and ready to serve. The simultaneous optimization of
efficient voice leading and harmonic consistency led composers to discover new
syntactic possibilities in the familiar sounds of diatonic tonality and provided a
forum for exploring the parallactic clash between old and new ways of construing
musical distance, at a historical moment when old and new were being placed into
acute tension in every domain of European culture.
The triad’s double determination can be seen as a fortuitous curiosity, a quirk,
a lucky draw to a straight flush. And yet, the triad is not the only homophonous
diamorph in Western music, nor is the nineteenth century the only moment when
preadaptation contributed to fundamental syntactic change. For it turns out that
the diatonic scale, the triad’s necessary partner in the production of classical tonal-
ity, is also doubly determined. There are many plausible reasons why the diatonic
collection became privileged in the Carolingian era, but it was surely not because
 CHAPTER 9 Double Syntax and the Soft Revolution  209

of its position within an equally tempered chromatic space first theorized five
centuries later and first realized three further centuries down the road. Yet once
that space arose, diatonic collections had properties that proved both unique and
functionally powerful: “The major scale is a maximal structure possessing [the
property of unique multiplicity] in the usual equal-tempered division of the
octave” (Babbitt 2003[1965]).7 Although the diatonic scale evidently originated at
a moment when modulation was inconceivable, it was preadapted to serve as a
modulatory vehicle capable of achieving maximum variety and depth. Thus, the
diatonic scale’s overdetermination played a central role in the new modulatory
system of early-eighteenth-century harmonic tonality, just as the consonant triad’s
overdetermination played a central role in the early nineteenth century’s discovery
of an alternative to it.
The story has one further chapter. The double determination of the three-
element triad and the seven-element scale are matched by a double determination
of the twelve-element equal-tempered chromatic gamut.8 The chromatic collec-
tion arose in stages as musicians rejected the tritone as a surrogate for the perfect
fifth. Projecting further along the chain of just fifths, the next opportunity for sys-
temic closure arises at the twelfth generation. Acceptance of the octave-adjusted
Pythagorean comma as a suitable pseudounison halts the proliferation tones,
leaving a chromatic gamut of twelve tones. The chromatic collection thus has
twelve elements through an arithmetic fortuity: it is when m = 12 that 3m suitably
approximates 2n for some value of n.
Once temperament distributes the twelve tones equally about the cyclic octave,
another property of that number comes forward as a stylistic determinant: twelve
is an abundant number, divisible by 2, 3, 4, and 6, that is, by every nonunit interval
besides its just-fifth generator. The system of twelve tones is thus preadapted for
equal division. The projection of just fifths “knew nothing” of the abundant divis-
ibility of the number 12. Had the eleventh or thirteenth powers of 3 been suitable
approximations of some power of 2, we would have inherited a chromatic system
incapable of equal division: no whole-tone scales, no augmented triads, no dimin-
ished seventh chords.
Accordingly, the richness of nineteenth-century chromatic syntax results from
overdetermination and preadaptation at two distinct levels of structure. (One
might say, with only a slight abuse of language, that its double determination is
itself doubly determined.) One level is the one emphasized throughout this
book, which is also a central theme of Tymoczko’s Geometry of Music: the acousti-
cally consonant triads are preadapted to serve as optimal voice-leading objects
in chromatic space by virtue of their near evenness. But the second level of

7. By “unique multiplicity,” Babbitt refers to the number of instances of each interval class: six perfect
fourths, five whole steps, four minor thirds, and so forth, what Gamer 1967 refers to as a deep scale.
Common-tone retention under transposition is determined by the multiplicity of the interval of
transposition. Consequently, a scale with maximum intervallic multiplicity has a maximum diversity
of common tones under its various transpositions. For any particular diatonic collection, there is
another diatonic collection that shares n tones with it, where n ranges from 2 to 6. Among seven-tone
“scales,” only the chromatic heptachord shares that potential.
8. For a more detailed account, see Balzano 1980.
210  Audacious Euphony

overdetermination, the preadaptation of the fifth-generated series of pitch classes

to the multiple even divisions of chromatic space, is just as essential. Had the series
of just fifths achieved pseudoclosure at a prime number of generations, then the
chromatic system would lack perfectly even divisions. Without perfectly even
divisions, the 3-in-12 nearly even consonant triads would behave like 7-in-12 dia-
tonic collections in chromatic space, or 3-in-7 diatonic triads in diatonic-scalar
space. Without perfectly even trichords to minimally perturb, the consonant triads
would group neither into hexatonic systems nor Weitzmann regions. Without
diminished seventh chords, the Tristan-genus tetrachords would group into nei-
ther octatonic systems nor Boretz regions. Under incremental voice leading, all
nearly even chords of a given cardinality would form a single exhaustive cycle,
rather than partition into equivalent co-cyclic regions.
The munificent economy of Tymoczko’s distributive god is thus spread across
multiple sectors of the musical firmament. Through their serendipitous coordina-
tion, the several layers of overdetermination generate much of the structural intri-
cacy and affective multiplicity of much of the music at the heart of the repertory of
concert halls and operatic stages. The tale of this book accordingly spins a varia-
tion on a tale twice told elsewhere. The script that unifies these variants is the
primal tension and collaboration between physical forces of the external world, on
the one hand, and relations internal to the musical chord or scale, on the other. To
understand these forces, as they play out in music, requires a structural conception
of musical materials and relations, such as is characteristic of (but not limited to)
what we think of as atonality. It requires a conception of the cognitive mechanisms
associated with the homophonous diamorph, to model how the grammars come
into contact and thread one into the other. And it requires a theory of classical
diatonic tonality, but that alone is not sufficient.
G L O S S A RY

The following technical terms used in this book either are not common to
standard music-theory pedagogy or are used here in nonstandard ways. A term
used in the definitions is italicized if defined elsewhere in the glossary.

Augmented triad. A perfectly even three-note chord.

Balanced voice leading. Under idealized voice leading, at least one voice moves up
and another voice moves down.
Boretz region. A collection of eight Tristan-genus chords—four half-diminished
seventh chords and four dominant seventh chords—each of which is a single
semitonal displacement of the same diminished seventh chord.
Chain. A series of triads derived by alternating two distinct triadic transforma-
tions.
Consonant triad. A nearly even three-tone chord, either major or minor (equiv.
harmonic triad; in German, Klang).
Cube Dance. A connected graph whose nodes are the twenty-four consonant
triads and the four augmented triads, and whose edges represent minimal
voice-leading work, or single semitonal displacement.
Diminished seventh chord. A perfectly even four-note chord.
Double syntax hypothesis. The hypothesis that the mind is capable of organizing
musical patterns, simultaneously or in immediate succession, in two distinct
and incompatible ways.
Downshift voice leading. A motion between two chords, under idealized voice
leading, where all moving voices proceed downward.
4-Cube Trio. A connected graph whose nodes are the twenty-four Tristan-genus
chords, the four diminished seventh chords, the six French sixth chords, and
the twelve minor seventh chords, and whose edges represent single semitonal
displacement.
Generalized Weitzmann region (GWR). A collection of all of the nearly even
chords related to a single perfectly even chord by a single semitonal displacement.
Weitzmann regions and Boretz regions are examples.
Generating interval. An interval that forms a chord by recursive application. In a
whole-tone scale, the whole tone is the generating interval; in the diatonic
scale, the perfect fourth or fifth is the generating interval.
Hexatonic cycle. An arrangement of six consonant triads such that each is
adjacent to those two triads to which it relates by single semitonal displacement.

211
212  Glossary

H(exatonic)-group transformation. Any transformation that maps a consonant

triad within its hexatonic region and outside of its Weitzmann region. The three
H-group transformations are the Leittonwechsel (L), the parallel (P), and the
hexatonic pole (H).
Hexatonic pole (H). A relationship between two triads that share membership in
a hexatonic cycle but that share no tones in common; or the transformation
that produces one of these triads from the other.
Hexatonic region. An unordered collection of consonant triads containing the six
members of a hexatonic cycle.
Hexatonic scale. The union of two adjacent augmented triads. In scalar order, the
step intervals alternate semitones and minor thirds.
Homophonous diamorph. A sound that has meanings or functions in two
distinct systems.
Idealized voice leading. The ordered dyads between two chords when their tones
are paired one to one and the voice-leading work is as small as possible. The
idealized voice leading from C major to F major is (C, C), (E, F), (G, A), no
matter how those chords are registrally realized or how the actual instruments
or singers move between their constituent tones.
Leittonwechsel (L) (English: leading-tone exchange). The relation between two
consonant triads that share a minor-third dyad, for example, C major and e
minor; or the transformation that produces one of these triads from the
other.
Nearly even. A collection of tones that are distributed as evenly as possible with-
out being perfectly even.
Nebenverwandt (N) (English: next related). The relation between a major triad
and the minor triad whose root lies a perfect fourth above it, for example,
C major and f minor; or the transformation that produces one of these triads
from the other.
Neighborhood. A collection of six consonant triads that share a single tone.
Octatonic region. A collection of eight Tristan-genus chords—four half-diminished
seventh and four dominant seventh chords—that draw their tones from the
same octatonic scale.
Octatonic scale (or collection). The union of two distinct diminished seventh
chords. In scalar order, the step intervals alternate semitones and whole
steps.
Pan-triadic. Any composition, or segment thereof, that consists exclusively or
predominately of major or minor triads without determining a tonal center.
Parallel (P). The relation between two consonant triads that share a root, for
example, C major and c minor; or the transformation that produces one of
these triads from the other.
Perfectly even. A collection of tones that divides the octave into equal parts.
Pitch retention loop. A cyclic ordering of a neighborhood such that each pair of
adjacent triads shares two tones.
Relative (R). The relation between two consonant triads that share a major-third
dyad, for example, C major and a minor; or the transformation that produces
one of these triads from the other.
 Glossary  213

Single semitonal displacement. A relation between two chords of cardinality n

that share n – 1 common tones and whose remaining tones are a semitone
apart; or, the motion between two such chords.
Slide (S). The relation between two consonant triads that share a common third,
for example, C major and cᅊ minor; or the transformation that produces one
of these triads from the other.
Syntonic image. The relation between two tones that are separated by a syntonic
comma in just intonation. Typically one of these tones is two 3:2 perfect fifths
(minus an octave) above the tonic (3/2 × 3/2 × 1/2 = 9/8) and the other is a 4:3
perfect fourth minus a 6:5 minor third above the tonic (4/3 ÷ 6/5 = 10/9). The
difference between the tones is 9/8 ÷ 10/9 = 81/80, slightly larger than one-fifth
of an equally tempered semitone.
Tonnetz. An graph whose nodes are pitch classes and whose edges represent
consonant intervals.
Tristan genus. The union of the twelve dominant and twelve half-diminished
seventh chords.
Upshift voice leading. A motion between two chords, under idealized voice
leading, where all moving voices proceed upward.
Voice-leading work. The sum of the magnitude of all moving voices of two chords
connected by idealized voice leading.
Voice-leading zone. A collection of chords whose pitch classes sum to a constant
value. Consonant triads share a voice leading zone if they are transpositionally
related by major third, in which case they share membership in both a
hexatonic region and a Weitzmann region. Tristan-genus chords share a voice
leading zone if they are transpositionally related by minor third or tritone, in
which they share membership in both an octatonic region and a Boretz region.
W(eitzmann)-group transformations. Any transformation that maps a consonant
triad within its Weitzmann region and outside of its hexatonic region. The
three W-group transformations are the relative (R), nebenverwandt (N), and
slide (S).
Weitzmann region. A collection of six consonant triads, three major and three
minor, each of which is a single semitonal displacement of the same augmented
triad.
This page intentionally left blank
B I B L IO G R A P H Y

Abbate, Carolyn. 1991. Unsung Voices. Princeton, NJ: Princeton University Press.
Adorno, Theodor. 1964. Moments musicaux. Frankfurt: Suhrkamp.
Agawu, Kofi. 1989. “Schubert’s Harmony Revisited: The Songs ‘Du liebst mich nicht’ and
‘Dass Sie hier gewesen.’” Journal of Musicological Research 9: 23–42.
——. 2009. Music as Discourse: Semiotic Adventures in Romantic Music. Oxford: Oxford
University Press.
Agmon, Eytan. 1986. “Diatonicism, Chromaticism, and Enharmonicism: A Study in
Cognition and Perception.” Ph.D. diss., City University of New York.
——. 1991. “Linear Transformations between Cyclically Generated Chords.” Musikometrika
3 (Quantitative Linguistics 46): 15–40.
——. 1995. “Functional Harmony Revisited: A Prototype-Theoretic Approach.” Music
Theory Spectrum 17, no. 2: 196–214.
Ahn, So-Yung. 2003. “Harmonic Circles and Voice Leading in Asymmetrical Trichords.”
Ph.D diss., State University of New York at Buffalo.
Aldwell, Edward, and Carl Schachter. 1989. Harmony and Voice Leading. 2nd ed. New York:
Harcourt Brace.
Anson-Cartwright, Mark. 1996. “Chord as Motive: The Augmented-Triad Matrix in
Wagner’s Siegfried Idyll.” Music Analysis 15, no. 1: 57–71.
Babbitt, Milton. 2003 [1965]. “The Structure and Function of Music Theory.” In The Collected
Essays of Milton Babbitt, ed. Stephen Peles, with Stephen Dembski, Andrew Mead, and
Joseph Straus. Princeton, NJ: Princeton University Press. Originally published in College
Music Symposium 5: 49–60.
Bailey, Robert. 1985. “An Analytical Study of the Sketches and Drafts.” In Prelude and
Transfiguration from Tristan und Isolde, by Richard Wagner, ed. Robert Bailey. Norton
Critical Scores. New York: W. W. Norton.
Baker, James M. 1986. The Music of Alexander Scriabin. New Haven, CT: Yale University
Press.
Balzano, Gerald. 1980. The Group-Theoretic Description of 12-Fold and Microtonal
Systems.” Computer Music Journal 4, no. 4: 66–84.
Bass, Richard. 2001. “Half-Diminished Functions and Transformations in Late Romantic
Music.” Music Theory Spectrum 23, no. 1: 41–60.
Benjamin, William. 1976. “Interlocking Diatonic Collections as a Source of Chromaticism
in Late Nineteenth-Century Music.” In Theory Only 1: 31–51.
Bergé, Pieter, Jeroen D’hoe, and William E. Caplin, eds. 2009. Beethoven’s Tempest Sonata:
Perspectives of Analysis and Performance. Leuven, Belgium: Peeters.
Bernstein, David W. 1986. “The Harmonic Theory of Georg Capellen.” Ph.D. diss., Columbia
University.

215
216  Bibliography

——. 1993. “Georg Capellen’s Theory of Reduction: Radical Harmonic Theory at the Turn
of the Twentieth Century.” Journal of Music Theory 37, no. 1: 85–116.
Berry, David Carson. 2004. “The Meaning(s) of ‘Without’: An Exploration of Liszt’s Bagatelle
ohne Tonart.” 19th-Century Music 27, no. 3: 230–62.
Blum, Stephen. 1986. Letter to the editor. Journal of the American Musicological Society, 39,
no. 1: 210–15.
Boatwright, Howard. 1966. “Analysis Symposium: W. A. Mozart, Menuetto in D major for
Piano (K. 355).” Journal of Music Theory 10, no. 1: 22–31.
Boretz, Benjamin. 1972 [1970]. “Meta-Variations, Part IV: Analytic Fallout (I).” Perspectives
of New Music 11, no. 1: 146–223. Originally published in “Meta-Variations: Studies in
the Foundations of Musical Thought.” Ph.D. diss., Princeton University.
Boulez, Pierre. 1971 [1963]. Boulez on Music Today, trans. Susan Bradshaw and Richard
Rodney Bennett. Cambridge, MA: Harvard University Press. Originally published as
Musikdenken Heute 1. Mainz: B. Schott’s Söhne.
Brendel, Alfred. 1991. Music Sounded Out: Essays, Lectures, Interviews, Afterthoughts.
New York: Farrar Straus Giroux.
Bribitzer-Stull, Matthew. 2006. “The Aᅈ-C-E Complex: The Origin and Function of
Chromatic Major Third Collections in Nineteenth-Century Music.” Music Theory
Spectrum 28, no. 2: 167–90.
Broersma, Mirjam, and Kees de Bot. 2006. “Triggered Codeswitching: A Corpus-Based
Evaluation of the Original Triggering Hypothesis and a New Alternative.” Bilingualism:
Language and Cognition 9, no. 1: 1–13.
Brower, Candace. 2008. “Paradoxes of Pitch Space.” Music Analysis 27, no. 1: 51–106.
Brown, Matthew. 2005. Explaining Tonality: Schenkerian Theory and Beyond. Rochester,
NY: University of Rochester Press.
Brown, Matthew, Douglas Dempster, and Dave Headlam. 1997. “The ᅊIV (ᅈV) Hypothesis:
Testing the Limits of Schenker’s Theories of Tonality.” Music Theory Spectrum 19, no. 2:
155–83.
Buhler, James. 1996. “‘Breakthrough’ as Critique of Form: The Finale of Mahler’s First
Symphony.” 19th-Century Music 20, no. 2: 125–43.
Bukofzer, Manfred. 1940. “Popular Polyphony in the Middle Ages.” Musical Quarterly 26,
no. 1: 31–49.
Burton, Deborah, and Gregory Harwood. Forthcoming. Francesco Galeazzi, Theoretical-
Practical Elements of Music: An Annotated Translation and Introduction. Studies in the
History of Music Theory and Literature, vol. 5. Urbana: University of Illinois Press.
Originally published as Elementi teorico-pratici di musica con un saggio sopra l’arte di
suonare il violino analizzata ed a dimostrabili principi ridotta, vol. 2. Rome: Michele
Puccinelli, 1796.
Callender, Clifton. 1998. “Voice-Leading Parsimony in the Music of Alexander Scriabin.”
Journal of Music Theory 42, no. 2: 219–33.
Caplin, William E. 1998. Classical Form: A Theory of Formal Functions for the Instrumental
Music of Haydn, Mozart, and Beethoven. New York: Oxford University Press.
Capuzzo, Guy. 2004. “Neo-Riemannian Theory and the Analysis of Pop-Rock Music.” Music
Theory Spectrum 26, no. 2: 177–99.
——. 2006. “Pat Martino’s The Nature of the Guitar: An Intersection of Jazz Theory and Neo-
Riemannian Theory.” Music Theory Online 12, no. 1. http://mto.societymusictheory.org/
issues/mto.06.12.1/mto.06.12.1.capuzzo.pdf
Chafe, Eric. 2005. The Tragic and the Ecstatic: The Musical Revolution of Wagner’s Tristan
und Isolde. Oxford: Oxford University Press.
Childs, Adrian P. 1998. “Moving beyond Neo-Riemannian Triads: Exploring a
Transformational Model for Seventh Chords.” Journal of Music Theory 42, no. 2: 181–93.
 Bibliography  217

Christensen, Thomas. 1992. “The Spanish Guitar and Seventeenth-Century Triadic Theory.”
Journal of Music Theory 36, no. 1: 1–42.
——. 1993. Rameau and Musical Thought in the Enlightenment. Cambridge: Cambridge
University Press.
Cinnamon, Howard. 1986. “Tonic Arpeggiation and Successive Equal Third Relations
as Elements of Tonal Evolution in the Music of Franz Liszt.” Music Theory Spectrum
8: 1–24.
Clampitt, David. 1997. “Pairwise Well-Formed Scales: Structural and Transformational
Properties.” Ph.D. diss., State University of New York at Buffalo.
——. 1998. “Alternative Interpretations of Some Measures from Parsifal.” Journal of Music
Theory 42, no. 2: 321–34.
Clark, Suzannah. 2011a. Analyzing Schubert. New York: Cambridge University Press.
——. 2011b. “On the Imagination of Tone in Schubert’s Liedesend (D473), Trost (D523), and
Gretchens Bitte (D564).” In The Oxford Handbook of Neo-Riemannian Music Theories,
ed. Edward Gollin and Alexander Rehding. New York: Oxford University Press.
Clough, John. 1998. “A Rudimentary Geometric Model for Contextual Transposition and
Inversion.” Journal of Music Theory 42, no. 2: 297–306.
Clough, John, and Jack Douthett. 1991. “Maximally Even Sets.” Journal of Music Theory
35, nos. 1–2: 93–173.
Code, David Løberg. 1995. “Listening for Schubert’s ‘Doppelgängers.’” Music Theory Online
1, no. 4. http://mto.societymusictheory.org/issues/mto.95.1.4/mto.95.1.4.code.html
Cohn, Richard. 1991. “Properties and Generability of Transpositionally Invariant Sets.”
Journal of Music Theory 35, nos. 1–2: 1–32.
——. 1992. “Dramatization of Hypermetric Conflicts in the Scherzo of Beethoven’s Ninth
Symphony.” 19th-Century Music 15, no. 3: 22–40.
——. 1996. “Maximally Smooth Cycles, Hexatonic Systems, and the Analysis of Late-
Romantic Triadic Progressions.” Music Analysis 15, no. 1: 9–40.
——. 1997. “Neo-Riemannian Operations, Parsimonious Trichords, and Their Tonnetz
Representations.” Journal of Music Theory 41, no. 1: 1–66.
——. 1998a. “Introduction to Neo-Riemannian Theory: A Survey and a Historical
Perspective.” Journal of Music Theory 42, no. 2: 167–80.
——. 1998b. “Square Dances with Cubes.” Journal of Music Theory 42, no. 2: 283–96.
——. 1999. “As Wonderful as Star Clusters: Instruments for Gazing at Tonality in Schubert.”
19th-Century Music 22, no. 3: 213–32.
——. 2000. “Weitzmann’s Regions, My Cycles, and Douthett’s Dancing Cubes.” Music Theory
Spectrum 22, no. 1: 89–103.
——. 2004. “Uncanny Resemblances: Tonal Signification in the Freudian Age.” Journal of the
American Musicological Society 57, no. 2: 285–323.
——. 2006. “Hexatonic Poles and the Uncanny in Parsifal.” Opera Quarterly 22, no. 2:
230–48.
——. 2011. “Tonal Pitch Space and the (Neo-)Riemannian Tonnetz.” In The Oxford Handbook
of Neo-Riemannian Music Theories, ed. Edward Gollin and Alexander Rehding.
New York: Oxford University Press.
Connolly, Daniel K. 1999. “Imagined Pilgrimage in the Itinerary Maps of Matthew Paris.”
Art Bulletin 81, no. 4: 598–622.
Cook, Nicholas. 1987. Guide to Musical Analysis. London: Dent.
Cook, Robert. 1994. “Alternative Transformational Aspects of the ‘Grail’ in Wagner’s
Parsifal.” Paper presented at Music Theory Midwest, Bloomington, IN.
——. 2001. “Transformational Approaches to Romantic Harmony and the Late Works of
César Franck.” Ph.D. diss., University of Chicago.
——. 2005. “Parsimony and Extravagance.” Journal of Music Theory 49, no. 1: 109–40.
218  Bibliography

Copley, R. Evan. 1991. Harmony: Baroque to Contemporary, vol. 1, 2nd ed. Champaign, IL:
Stipes.
Dahlhaus, Carl. 1967a. Untersuchungen über die Entstehung der harmonischen Tonalität.
Kassel: Bärenreiter.
——. 1967b. “Zur chromatischen Technik Carlo Gesualdos.” Analecta Musicologica
4: 77–96.
——. 1980a [1974]. Between Romanticism and Modernism, trans. Mary Whittall. Berkeley:
University of California Press. Originally published as Zwischen Romantik und Moderne.
Munich: Musikverlag Katzbichler.
——. 1980b. “Harmony.” In The New Grove Dictionary of Music and Musicians, ed. Stanley
Sadie, vol. 8. London: Macmillan.
——. 1989 [1980]. Nineteenth-Century Music, trans. J. Bradford Robinson. Berkeley :
University of California Press. Originally published as Die Musik des 19. Jahrhunderts.
Wiesbaden: Akademische Verlagsgesellschaft Athenaion.
——. 1990 [1967]. Studies on the Origin of Harmonic Tonality, trans. Robert Gjerdingen.
Princeton, NJ: Princeton University Press. Originally published as Untersuchungen über
die Entstehung der harmonischen Tonalität. Kassel: Bärenreiter.
Damschroder, David. 2008. Thinking about Harmony: Historical Perspectives on Analysis.
Cambridge: Cambridge University Press.
——. 2010. Harmony in Schubert. Cambridge: Cambridge University Press.
Darcy, Warren. 1993. Wagner’s Das Rheingold. Oxford: Clarendon; New York: Oxford
University Press.
Douthett, Jack. 1993. “Pn-Relations and Graphs in Music.” Paper presented at the first
Buffalo Music Theory Working Group on Neo-Riemannian Theory, Buffalo, NY.
——. 2008. “Filtered-Point Symmetry and Dynamical Voice Leading.” In Music Theory and
Mathematics: Chords, Collections, and Transformations, ed. Jack Douthett, Martha M.
Hyde, and Charles J. Smith. Rochester, NY: University of Rochester Press.
Douthett, Jack, and Peter Steinbach. 1998. “Parsimonious Graphs: A Study in Parsimony,
Contextual Transformations, and Modes of Limited Transposition.” Journal of Music
Theory 42, no. 2: 241–64.
Draeseke, Felix. 1987 [1861]. “Die sogenannte Zukunftsmusik und ihrer Gegner.” In
Schriften 1855–1861, ed. Martella Gutiérrez-Denhoff and Helmut Loos. Bad Honnef:
Gudrun Schröder Verlag. Originally published in Neue Zeitschrift für Musik 55, nos.
9–10. Translated excerpt in “The Evolution of Beethoven’s Late Style: Another ‘New Path’
after 1824?” by William Kinderman. Beethoven Forum 8: 71–99, 2000.
Duffin, Ross. 2007. How Equal Temperament Ruined Harmony (and Why You Should Care).
New York: W. W. Norton.
Dunsby, Jonathan. 2002. “Thematic and Motivic Analysis.” In The Cambridge History of
Western Music Theory, ed. Thomas Christensen. Cambridge: Cambridge University
Press.
Durutte, Camille. 1855. Esthétique musicale: Technie ou lois générales du système harmo-
nique. Paris: Mallet-Bachelier.
Engebretsen, Nora. 2002. “The Chaos of Possibilities: Combinatorial Group Theory
in Nineteenth-Century German Harmony Treatises.” Ph.D diss., State University of
New York at Buffalo.
——. 2008. “The ‘Over-determined’ Triad as a Source of Discord: Nascent Groups and
the Emergent Chromatic Tonality in Nineteenth-Century German Harmonic Theory.”
In Music Theory and Mathematics: Chords, Collections, and Transformations, ed. Jack
Douthett, Martha M. Hyde, and Charles J. Smith. Rochester, NY: University of Rochester
Press.
 Bibliography  219

Ferris, David. 2000. Schumann’s Eichendorff Liederkreis and the Genre of the Romantic Cycle.
Oxford: Oxford University Press.
Fétis, François-Joseph. 1832. “Cours de philosophie musicale et d’histoire de la musique,”
lesson 8. Revue musicale 6, no. 25: 196–98.
——. 2008 [1844]. Complete Treatise on the Theory and Practice of Harmony, trans. Peter M.
Landey. Hillsdale, NY: Pendragon. Originally published as Traité complet de la théorie et
de la pratique de l’harmonie. Paris: Maurice Schlesinger.
Feurzeig, Lisa. 1997. “Heroines in Perversity: Marie Schmith, Animal Magnetism, and the
Schubert Circle.” 19th-Century Music 21, no. 2: 223–43.
Feynman, Richard P. 1965. The Character of Physical Law. Cambridge, MA: MIT Press.
Fisk, Charles. 2000. “Current Chronicle.” With response by Richard Cohn. 19th-Century
Music 23, no. 3: 301–2.
——. 2001. Returning Cycles: Contexts for the Interpretation of Schubert’s Impromptus and
Last Sonatas. Berkeley : University of California Press.
Forte, Allen. 1972. “Sets and Nonsets in Schoenberg’s Atonal Music.” Perspectives of
New Music 11, no. 1: 43–64.
——. 1983. “Motive and Rhythmic Contour in the Alto Rhapsody.” Journal of Music Theory
27, no. 2: 255–71.
——. 1987. “Liszt’s Experimental Idiom and Music of the Early Twentieth Century.”
19th-Century Music 10, no. 3: 209–28.
Galeazzi, Francesco. 1796. Elementi teorico-pratici di musica con un saggio sopra l’arte di
suonare il violino analizzata ed a dimostrabili principi ridotta, vol. 2. Rome: Michele
Puccinelli.
Gamer, Carlton. 1967. “Some Combinational Resources of Equal-Tempered Systems.”
Journal of Music Theory 11, no. 1: 32–59.
Gardner-Chloros, Penelope. 2009. Code-Switching. Cambridge: Cambridge University
Press.
Gauldin, Robert. 2004. Harmonic Practice in Tonal Music. 2nd ed. New York:
W. W. Norton.
Gessele, Cynthia M. 1994. “‘Base d’harmonie’: A Scene from Eighteenth-Century French
Music Theory.” Journal of the Royal Musical Association 119, no. 1: 60–90.
Goldenberg, Yosef. 2007. “Schenkerian Voice-Leading and Neo-Riemannian Operations:
Analytical Integration without Theoretical Reconciliation.” Journal of Schenkerian
Studies 2: 65–84.
Gollin, Edward. 1998. “Some Aspects of Three-Dimensional Tonnetze.” Journal of Music
Theory 42, no. 2: 195–206.
——. 2000. “Representations of Space and Conceptions of Distance in Transformational
Theories.” Ph.D. diss., Harvard University.
——. 2006. “Some Further Notes on the History of the Tonnetz.” Theoria 13: 99–111.
——. 2011. “On a Transformational Curiosity in Riemann’s Schematisirung der Dissonanzen.”
In The Oxford Handbook of Neo-Riemannian Music Theories, ed. Edward Gollin and
Alexander Rehding. New York: Oxford University Press.
Gould, Stephen Jay. 1977. Ever Since Darwin: Reflections in Natural History. New York:
W. W. Norton.
Grave, Floyd K., and Margaret Grave. 1987. In Praise of Harmony: The Teachings of Abbé
Georg Vogler. Lincoln: University of Nebraska Press.
Hallmark, Rufus. 1975. “The Genesis of Dichterliebe: A Source Study.” Ph.D. diss., Princeton
University.
Hantz, Edwin. 1982. “Motivic and Structural Unity in Liszt’s ‘Blume und Duft.’ ” In Theory
Only 6, no. 3: 3–11.
220  Bibliography

Harrison, Daniel. 1994. Harmonic Function in Chromatic Music. Chicago: University of

Chicago Press.
——. 2002a. “Dissonant Tonics and Post-tonal Tonality.” Paper presented at the annual
meeting of the Music Theory Society of New York State, New York, NY.
——. 2002b. “Nonconformist Notions of Nineteenth-Century Enharmonicism.” Music
Analysis 21, no. 2: 115–60.
——. 2011. “Three Essays on Neo-Riemannian Theory.” In The Oxford Handbook of
Neo-Riemannian Music Theories, ed. Edward Gollin and Alexander Rehding. New York:
Oxford University Press.
Hauptmann, Moritz. 1888 [1853]. The Nature of Harmony and Meter, trans. W. E. Heathcote.
London: S. Sonnenschein. Reprint, New York: Da Capo, 1991. Originally published as
Die Natur der Harmonik und der Metrik. Leipzig: Breitkopf und Härtel.
Helmholtz, Hermann von. 1885 [1877]. On the Sensations of Tone as a Physiological Basis
for the Theory of Music. 2nd English ed., trans. Alexander J. Ellis. London: Longmans,
Green. Reprint, New York: Dover, 1954. Originally published as Die Lehre von den
Tonempfindungen als physiologische Grundlage für die Theorie der Musik. 4th German
ed. Braunschweig: F. Viewig. 1st German ed., 1863.
Hepokoski, James, and Warren Darcy. 2006. Elements of Sonata Theory: Norms, Types,
and Deformations in the Late Eighteenth-Century Sonata. Oxford: Oxford University
Press.
Hoeckner, Berthold. 1997. “Schumann and Romantic Distance.” Journal of the American
Musicological Society 50, no. 1: 55–132.
——. 2001. “Poet’s Love and Composer’s Love.” Music Theory Online 7, no. 5. http://mto.
societymusictheory.org/issues/mto.01.7.5/mto.01.7.5.hoeckner.html
——. 2006. “Paths through Dichterliebe.” 19th-Century Music 30, no. 1: 65–80.
Hoffmann, E. T. A. 1989 [1813–14]. E. T. A. Hoffmann’s Musical Writing: Kreisleriana, The
Poet and the Composer, Music Criticism, ed. David Charlton. Cambridge: Cambridge
University Press.
Hook, Julian. 2002. “Uniform Triadic Transformations.” Journal of Music Theory 46,
nos. 1–2: 57–126.
Horton, Julian. 2004. Bruckner’s Symphonies: Analysis, Reception, and Cultural Politics.
Cambridge: Cambridge University Press.
Hostinský, Otakar. 1879. Die Lehre von den musikalischen Klängen: Ein Beitrag zur
aesthetischen Begründung der Harmonielehre. Prague: H. Dominicus.
Hull, A. Eaglefield. 1915. Modern Harmony: Its Explanation and Application. London:
Augener.
Hyer, Brian. 1989. “Tonal Intuitions in ‘Tristan und Isolde.’” Ph.D. diss., Yale University.
——. 1994. “Chopin and the In-F-able.” In Musical Transformation and Musical Intuition,
ed. Michael Cherlin and Rafael Atlas. Dedham, MA: Ovenbird.
——. 1995. “Reimag(in)ing Riemann.” Journal of Music Theory 39, no. 1: 101–38.
Jonas, Oswald. 1982 [1934]. Introduction to the Theory of Heinrich Schenker: The Nature of
the Musical Work of Art, trans. John Rothgeb. New York: Longman. Originally published
as Einführung in die Lehre Heinrich Schenkers. Vienna: Universal.
Kallberg, Jeffrey. 1985. “Chopin’s Last Style.” Journal of the American Musicological Society
38, no. 2: 264–315.
Kant, Immanuel. 1982 [1787]. Critique of Pure Reason, trans. Wolfgang Schwarz. Aalen,
Germany : Scientia. Originally published as Kritik der reinen Vernunft, 2nd ed. Riga:
J. F. Hartknoch.
Kaplan, Richard. 1994. “Tonality as Mannerism: Structure and Syntax in Richard Strauss’s
Orchestral Song ‘Frühling.’” Theory and Practice 19: 19–29.
 Bibliography  221

Karg-Elert, Sigfrid. 1930. Polaristische Klang- und Tonalitätslehre (Harmonologik). Leipzig:

F. E. C. Leuckart.
Katz, Adele. 1945. Challenge to Musical Tradition. New York: Alfred A. Knopf.
Kirnberger, Johann Philipp. 1982 [1771 (vol. 1) and 1776 (vol. 2, part I)]. The Art of
Strict Musical Composition, trans. David Beach and Jurgen Thym. New Haven, CT:
Yale University Press. Originally published as vol. 1 and vol. 2, part I, of Die Kunst des
reinen Satzes in der Musik. Berlin and Königsberg.
Klumpenhouwer, Henry. 1994. “Some Remarks on the Use of Riemann Transformations.”
Music Theory Online 0, no. 9. http://mto.societymusictheory.org/issues/mto.94.0.9/
mto.94.0.9.klumpenhouwer.art.html
Klumpenhouwer, Henry. 2002. “Dualist Tonal Space and Transformation in Nineteenth-
Century Musical Thought.” In The Cambridge History of Western Music Theory, ed.
Thomas Christensen. Cambridge: Cambridge University Press.
Komar, Arthur. 1971. “The Music of Dichterliebe: The Whole and Its Parts.” In Dichterliebe, by
Robert Schumann, ed. Arthur Komar. Norton Critical Scores. New York: W. W. Norton.
Kopp, David. 1995. “A Comprehensive Theory of Chromatic Mediant Relations in
Mid-Nineteenth-Century Music.” Ph.D. diss., Brandeis University.
——. 1997. “Pentatonic Organization in Two Piano Pieces of Debussy.” Journal of Music
Theory 41, no. 2: 261–87.
——. 2002. Chromatic Transformations in Nineteenth-Century Music. Cambridge: Cambridge
University Press.
Kramer, Lawrence. 1986. “The Schubert Lied: Romantic Form and Romantic Consciousness.”
In Schubert: Critical and Analytical Studies, ed. Walter Frisch. Lincoln: University of
Nebraska Press.
——. 1990. Music as Cultural Practice, 1800–1900. Berkeley : University of California
Press.
Kramer, Richard. 1985. “The New Modulation of the 1770s: C. P. E. Bach in Theory, Criticism,
and Practice.” Journal of the American Musicological Society 38, no. 3: 551–92.
——. 1994. Distant Cycles: Schubert and the Conceiving of Song. Chicago: University of
Chicago Press.
——. 2008. Unfinished Music. Oxford: Oxford University Press.
Kraus, Joseph C. 1990. “Mozart’s Chromatic Third Relations: Evidence from the Late
Quartets and Quintets.” Journal of Musicological Research 9, no. 4: 229–54.
Krebs, Harald M. 1980. “Third Relation and Dominant in Late 18th- and Early 19th-Century
Music.” Ph.D. diss., Yale University.
Krumhansl, Carol. 1990. Cognitive Foundations of Musical Pitch. Oxford: Oxford University
Press.
——. 1998. “Perceived Triad Distance: Evidence Supporting the Psychological Reality of
Neo-Riemannian Transformations.” Journal of Music Theory 42: 265–81.
Kunze, Stefan. 1970. “Über Melodiebegriff und musikalischen Bau in Wagners Musikdrama,
dargestellt an Beispielen aus Holländer und Ring.” In Das Drama Richard Wagners
als musikalisches Kunstwerk, ed. Carl Dahlhaus. Studien zur Musikgeschichte des 19.
Jahrhunderts, 23. Regensburg: Gustav Bosse.
Kurth, Ernst. 1923. Romantische Harmonik und ihre Krise in Wagners “Tristan.” 3rd ed.
Berlin: Max Hesses.
——. 1991. Selected Writings, ed. and trans. Lee Rothfarb. Cambridge: Cambridge University
Press.
Kurth, Richard. 1997. “Music and Poetry, a Wilderness of Doubles: Heine—Nietzsche—
Schubert—Derrida.” 19th-Century Music 21, no. 1: 3–37.
Langlé, Honoré de. 1797. Traité d’harmonie et de modulation. Paris: Boyer.
222  Bibliography

Larson, Steve. 1994. “Musical Forces, Step Collections, Tonal Pitch Space, and Melodic
Expectation.” In 3rd International Conference on Music Perception and Cognition:
Proceedings, ed. Irène Deliège. Liège, Belgium: European Society for the Cognitive
Sciences of Music.
LaRue, Jan. 2001 [1957]. “Bifocal Tonality: An Explanation for Ambiguous Baroque
Cadences.” Journal of Musicology 18, no. 2: 283–94. Originally published in Essays on
Music in Honor of Archibald Thompson Davison. Cambridge, MA: Harvard University
Department of Music.
Lehman, Frank. 2010. “Neo-Riemannian Analysis of Film Music: Puzzles and Prospects.”
Paper presented at the Yale Graduate Music Symposium, New Haven, CT.
Lendvai, Ernő. 1971. Béla Bartók: An Analysis of His Music. London: Kahn & Averill.
——. 1983. The Workshop of Bartók and Kodály. Budapest: Editio Musica Budapest.
——. 1988. Verdi and Wagner, vol. 1, trans. Monika Palos and Judit Pokoly. Budapest:
International House.
Lerdahl, Fred. 2001. Tonal Pitch Space. Oxford: Oxford University Press.
Lerdahl, Fred, and Ray Jackendoff. 1983. A Generative Theory of Tonal Music. Cambridge,
MA: MIT Press.
Lerdahl, Fred, and Carol Krumhansl. 2007. “Modeling Tonal Tension.” Music Perception
24, no. 4: 329–66.
Levenberg, Jeffrey. 2008. “Generally Duruttesque: Harmonic Generations and
Transformations.” Paper presented at New England Conference of Music Theorists,
Cambridge, MA.
Lewin, David. 1967. “A Study of Hexachord Levels in Schoenberg’s Violin Fantasy.”
Perspectives of New Music 6, no. 1: 18–32.
——. 1969. “Behind the Beyond: A Response to Edward T. Cone.” Perspectives of New Music
7, no. 2: 59–69
——. 1982. “A Formal Theory of Generalized Tonal Functions.” Journal of Music Theory
26, no. 1: 23–60.
——. 1984. “Amfortas’s Prayer to Titurel and the Role of D in Parsifal: The Tonal Spaces of
the Drama and the Enharmonic Cᅈ/B.” 19th-Century Music 7, no. 3: 336–49.
——. 1987. Generalized Musical Intervals and Transformations. New Haven, CT: Yale
University Press.
——. 1992. “Some Notes on Analyzing Wagner: The Ring and Parsifal.” 19th-Century Music
16, no. 1: 49–58.
——. 1995. “Generalized Interval Systems for Babbitt’s Lists, and for Schoenberg’s String
Trio.” Music Theory Spectrum 17, no. 1: 81–118.
——. 1996. “Cohn Functions.” Journal of Music Theory 40, no. 2: 181–216.
——. 1998. “Some Ideas about Voice-Leading between Pcsets.” Journal of Music Theory
42, no. 1: 15–72.
——. 2006. Studies in Music with Text. Oxford: Oxford University Press.
Longyear, Rey M., and Kate R. Covington. 1988. “Sources of the Three-Key Exposition.”
Journal of Musicology 6, no. 4: 448–70.
Lorenz, Alfred. 1933. Der musikalische Aufbau von Richard Wagners Parsifal. Das Geheimnis
der Form bei Richard Wagner 4. Berlin: M. Hesse.
Louis, Rudolf, and Ludwig Thuille. 1982 [1913]. “An Annotated English Translation of
Harmonielehre of Rudolf Louis and Ludwig Thuille,” trans. Richard Isadore Schwartz.
Ph.D. diss., Washington University. Originally published as Harmonielehre. 4th ed.
Stuttgart: Carl Grüninger.
Lowinsky, Edward E. 1946. Secret Chromatic Art in the Netherlands Motet, trans. Carl
Buchman. New York: Columbia University Press.
 Bibliography  223

——. 1989 [1967]. “The Musical Avant-Garde of the Renaissance; Or, The Peril and Profit
of Foresight.” In Music in the Culture of the Renaissance and Other Essays, vol. 2.
Chicago: University of Chicago Press. Originally published in Art, Science, and History
in the Renaissance, ed. Charles S. Singleton. Baltimore, MD: Johns Hopkins University
Press.
Lubin, Steven. 1974. “Techniques for the Analysis of Development in Middle-Period
Beethoven.” Ph.D. diss., New York University.
Masson, Charles. 1694. Nouveau Traité des Regles pour la Composition de la Musique.
Reprint, with an introduction by Imogene Horsley, New York: Da Capo, 1967.
McClain, Ernest G. 1978. The Pythagorean Plato: Prelude to the Song Itself. Stony Brook,
NY: N. Hays.
McClary, Susan. 1994. “Constructions of Subjectivity in Schubert’s Music.” In Queering
the Pitch: The New Gay and Lesbian Musicology, ed. Philip Brett, Elizabeth Wood, and
Gary C. Thomas. New York: Routledge.
——. 2010. “In Praise of Contingency: The Powers and Limits of Theory.” Music Theory
Online 16, no. 1. http://mto.societymusictheory.org/issues/mto.10.16.1/mto.10.16.1.
mcclary.html
McCreless, Patrick. 1996. “An Evolutionary Perspective on Nineteenth-Century Semitonal
Relations.” In The Second Practice of Nineteenth-Century Music, ed. William Kinderman
and Harald Krebs. Lincoln: University of Nebraska Press.
——. 2007. “Elgar and Theories of Chromaticism.” In Elgar Studies, ed. J. P. E. Harper-Scott
and Julian Rushton. Cambridge: Cambridge University Press.
McKinney, Timothy R. 1993. “Melodic Pitch Structures in Hugo Wolf ’s Augmented-Triad
Series.” Indiana Theory Review 14, no. 1: 37–94.
Meeus, Nicholas. 2000. “Toward a Post-Schoenbergian Grammar of Tonal and Pre-tonal
Harmonic Progressions.” Music Theory Online 6, no. 1. http://www.mtosmt.org/issues/
mto.00.6.1/mto.00.6.1.meeus_frames.html
Meyer, Leonard B. 1989. Style and Music: Theory, History, and Ideology. Philadelphia:
University of Pennsylvania Press.
Mickelsen, William C. 1977. Hugo Riemann’s Theory of Harmony. Lincoln: University of
Nebraska Press.
Mitchell, William J. 1962. “The Study of Chromaticism.” Journal of Music Theory 6, no. 1:
2–31.
——. 1967. “The Tristan Prelude: Techniques and Structure.” In The Music Forum, vol. 1, ed.
William J. Mitchell and Felix Salzer. New York: Columbia University Press.
Moomaw, Charles Jay. 1985. “Augmented Mediant Chords in French Baroque Music.”
Ph.D. diss., University of Cincinnati.
Mooney, Michael Kevin. 1996. “The ‘Table of Relations’ and Music Psychology in
Hugo Riemann’s Harmonic Theory.” Ph.D. diss., Columbia University.
Morgan, Robert P. 1976. “Dissonant Prolongation: Compositional and Theoretical
Precedents.” Journal of Music Theory 20, no. 1: 49–91.
——. 1998. “Symmetrical Form and Common-Practice Tonality.” Music Theory Spectrum
20, no. 1: 1–47.
Morris, Robert D. 1990. “Pitch-Class Complementation and Its Generalizations.” Journal of
Music Theory 34, no. 2: 175–245.
——. 1998. “Voice-Leading Spaces.” Music Theory Spectrum 20, no. 2: 178–208.
Motte, Diether de la. 1976. Harmonielehre. Kassel: Bärenreiter.
Murphy, Scott. 2001. “Wayward Faith: Divergence and Reconciliation of Melodic Sequence
and Harmonic Cycle in Some Measures from the Prelude of Wagner’s Parsifal.” Paper
presented at the annual meeting of the Society for Music Theory, Philadelphia, PA.
224  Bibliography

Muysken, Pieter. 2000. Bilingual Speech. Cambridge: Cambridge University Press.

Nattiez, Jean-Jacques. 1990. Music and Discourse: Toward a Semiology of Music, trans.
Carolyn Abbate. Princeton, NJ: Princeton University Press.
Newcomb, Anthony. 1981. “The Birth of Music Out of the Spirit of the Drama: An Essay in
Wagnerian Formal Analysis.” 19th-Century Music 5, no. 1: 38–66.
——. 1986. “Structure and Expression in a Schubert Song: Noch einmal Auf dem Flusse
zu hören.” In Schubert: Critical and Analytical Studies, ed. Walter Frisch. Lincoln:
University of Nebraska Press.
Oettingen, Arthur von. 1866. Harmoniesystem in dualer Entwickelung: Studien zur Theorie
der Musik. Dorpat and Leipzig: W. Gläser.
Oster, Ernst. 1966. “Analysis Symposium: W. A. Mozart, Menuetto in D major for Piano
(K. 355).” Journal of Music Theory 10, no. 2: 32–52.
Parly, Nila. 2009. “Flying a Wagner Kite: Subjunctive Performances of a Rheingold Scene
Based on a Dramaturgical Sketch by Carolyn Abbate.” Cambridge Opera Journal 21,
no. 2: 159–80.
Parncutt, Richard. 1989. Harmony: A Psychoacoustical Approach. Berlin: Springer.
Patel, Aniruddh. 2008. Music, Language, and the Brain. New York: Oxford University
Press.
Peck, Robert, and Jack Douthett. 2011. “Wreath Products and n-Cube Symmetry:
A Music-Theoretical Application.” Paper presented at the annual meeting of the
American Mathematical Society, New Orleans.
Perle, George. 1967. “The Musical Language of Wozzeck.” In The Music Forum, vol. 1, ed.
William J. Mitchell and Felix Salzer. New York: Columbia University Press.
Pople, Anthony. 1989. Skryabin and Stravinsky, 1908–1914: Studies in Theory and Analysis.
New York: Garland.
Perlman, Marc. 1994. “American Gamelan in the Garden of Eden: Intonation in a
Cross-Cultural Encounter.” Musical Quarterly 78, no. 3: 510–55.
Perrey, Beate Julia. 2002. Schumann’s Dichterliebe and Early Romantic Poetics. Cambridge:
Cambridge University Press.
Popovic, Igor. 1992. “Common Principles in Music-Theoretical Systems.” Ph.D. diss.,
Yale University.
Pressing, Jeff. 1983. “Cognitive Isomorphisms in World Music: West Africa, the Balkans,
Thailand and Western Tonality.” Studies in Music 17: 38–61.
Proctor, Gregory. 1978. “Technical Bases of Nineteenth-Century Chromatic Tonality:
A Study in Chromaticism.” Ph.D. diss., Princeton University.
Prout, Ebenezer. 1903. Harmony: Its Theory and Practice. 16th ed. London: Augener.
Rameau, Jean-Philippe. 1737. Génération harmonique, ou traité de musique théorique
et pratique. Paris: Prault fils.
——. 1760. Code de musique pratique, ou méthodes pour apprendre la musique. Paris:
Imprimerie royale.
——. 1971 [1722]. Treatise on Harmony, trans. Philip Gossett. New York: Dover. Originally
published as Traité de l’harmonie réduite à ses principes naturels. Paris: J.-B.-C. Ballard.
Ratner, Leonard. 1992. Romantic Music: Sound and Syntax. New York: Schirmer Books.
Rehding, Alexander. 2003. Hugo Riemann and the Birth of Modern Musical Thought.
Cambridge: Cambridge University Press.
——. 2005. “Rousseau, Rameau, and Enharmonic Furies in the French Enlightenment.”
Journal of Music Theory 49, no. 1: 141–80.
Reise, Jay. 1983. “Late Skriabin: Some Principles behind the Style.” 19th-Century Music 6,
no. 3: 220–31.
 Bibliography  225

Riemann, Hugo. 1877. Musikalische Syntaxis: Grundriß einer harmonischen Satzbildungs-

lehre. Leipzig: Breitkopf und Härtel.
——. 1880. Skizze einer neuen Methode der Harmonielehre. Leipzig: Breitkopf und Härtel.
——. 1890. Katechismus der Harmonielehre (theoretisch und praktisch). Leipzig: Max
Hesse.
——. 1897. Katechismus der Musik (Allgemeine Musiklehre). 2nd ed. Leipzig: Max Hesse.
Rings, Steven. 2006. “Tonality and Transformation.” Ph.D. diss., Yale University.
——. 2007. “Perspectives on Tonality and Transformation in Schubert’s Impromptu in Eᅈ,
D. 899.” Journal of Schenkerian Studies 2: 33–63.
——. 2011. Tonality and Transformation. New York: Oxford University Press.
Rosen, Charles. 1980. Sonata Forms. New York: W. W. Norton.
——. 1995. The Romantic Generation. Cambridge, MA: Harvard University Press.
Rothfarb, Lee A. 1988. Ernst Kurth as Theorist and Analyst. Philadelphia: University of
Pennsylvania Press.
Rothstein, William. 2008. “Common-Tone Tonality in Italian Romantic Opera: An
Introduction.” Music Theory Online 14, no. 1. http://mto.societymusictheory.org/issues/
mto.08.14.1/mto.08.14.1.rothstein.html
Rouget, Gilbert. 1985. Music and Trance: A Theory of the Relations between Music and
Possession. Chicago: University of Chicago Press.
Salzer, Felix, and Carl Schachter. 1969. Counterpoint in Composition. New York: McGraw
Hill.
Santa, Matthew. 2003. “Nonatonic Systems and the Parsimonious Interpretation of
Dominant-Tonic Progressions.” Theory and Practice 28: 1–28.
Saslaw, Janna K., and James P. Walsh. 1996. “Musical Invariance as a Cognitive Structure:
‘Multiple Meaning’ in the Early Nineteenth Century.” In Music Theory in the Age of
Romanticism, ed. Ian Bent. Cambridge: Cambridge University Press.
Satyendra, Ramon. 1992. “Chromatic Tonality and Semitonal Relationships in Liszt’s Late
Style.” Ph.D. diss., University of Chicago.
Schachter, Carl. 1983a. “The First Movement of Brahms’s Second Symphony: The Opening
Theme and Its Consequences.” Music Analysis 2, no. 1: 55–68.
——. 1983b. “Motive and Text in Four Schubert Songs.” In Aspects of Schenkerian Theory,
ed. David Beach. New Haven, CT: Yale University Press.
——. 1987. “Analysis by Key: Another Look at Modulation.” Music Analysis 6, no. 3:
289–318.
——. 1988. Review of Heinrich Schenker, Counterpoint. Musical Times 129, no. 1748: 524–29.
——. 1995. “The Triad as Place and Action.” Music Theory Spectrum 17, no. 2: 149–69.
Schenker, Heinrich. 1954 [1906]. Harmony, trans. Elisabeth Mann Borgese. Chicago:
University of Chicago Press. Originally published as Harmonielehre. Part I of
Neue musikalische Theorien und Phantasien. Stuttgart: J. G. Cotta.
——. 1979 [1935]. Free Composition, trans. Ernst Oster. New York: Longman. Originally
published as Der freie Satz. Vienna: Universal.
——. 2005 [1924]. “Beethoven’s Sonata in F minor, op. 57.” Translated by Robert
Snarrenberg. In Der Tonwille, vol. 2, ed. William Drabkin, trans. Ian Bent, et al.
Oxford: Oxford University Press. Originally published as “Beethoven: Sonate opus 57.”
In Der Tonwille, vol. 7. Vienna: Universal.
Schmalfeldt, Janet. 2011. In the Process of Becoming: Analytic and Philosophical Perspectives
on Form in Early Nineteenth-Century Music. New York: Oxford University Press.
Schubert, Peter. 2002. “Counterpoint Pedagogy in the Renaissance.” In The Cambridge
History of Western Music Theory, ed. Thomas Christensen. Cambridge: Cambridge
University Press.
226  Bibliography

Schwarz, David. 1997. Listening Subjects: Music, Psychoanalysis, Culture. Durham, NC:
Duke University Press.
Sechter, Simon. 1853–54. Die Grundsätze der musikalischen Komposition. 3 vols. Leipzig:
Breitkopf und Härtel.
Segall, Christopher. 2011. “The Common Third Relation in Russian Music Theory.”
Paper presented at the annual meeting of the Music Theory Society of New York State,
Buffalo, NY.
Seidel, Elmar. 1963. Die Enharmonik in den harmonischen Großformen Franz Schuberts.
Heidelberg: Heidelberger Reprographie A. Grosch.
Sessions, Roger. 1950. Musical Experience of Composer, Performer, Listener. Princeton, NJ:
Princeton University Press.
Shamgar, Beth. 1989. “Romantic Harmony through the Eyes of Contemporary Observers.”
Journal of Musicology 7, no. 4: 518–39.
Siciliano, Michael. 2002. “Neo-Riemannian Transformations and the Harmony of Franz
Schubert.” Ph.D. diss., University of Chicago.
——. 2005a. “Toggling Cycles, Hexatonic Systems, and some Analysis of Early Atonal
Music.” Music Theory Spectrum 27, no. 1: 221–248.
——. 2005b. “Two Neo-Riemannian Analyses.” College Music Symposium 45: 81–107.
Slatin, Sonia. 1967. “The Theories of Heinrich Schenker in Perspective.” Ph.D. diss.,
Columbia University.
Slobin, Mark. 1992. “Micromusics of the West: A Comparative Approach.” Ethnomusicology
37, no. 2: 1–87.
Slonimsky, Nicholas. 1937. Music since 1900. New York: W. W. Norton.
Smith, Charles J. 1986. “The Functional Extravagance of Chromatic Chords.” Music Theory
Spectrum 8: 94–139.
Smith, Peter H. 2006. “You Reap What You Sow: Some Instances of Rhythmic and Harmonic
Ambiguity in Brahms.” Music Theory Spectrum 28, no. 1: 57–97.
Sobaskie, James William. 1987. “Associative Harmony: The Reciprocity of Ideas in Musical
Space.” In Theory Only 10, nos. 1–2: 31–64.
Soderberg, Stephen. 1998. “The T-Hex Constellation.” Journal of Music Theory 42, no. 2:
207–18.
Somer, Avo. 1995. “Chromatic Third-Relations and Tonal Structure in the Songs of Debussy.”
Music Theory Spectrum 17, no. 2: 215–41.
Sorge, Georg Andreas. 1980 [1745]. “Georg Andreas Sorge’s Vorgemach der musicali-
schen Composition: A Translation and Commentary,” trans. Allyn Dixon Reilly. Ph.D.
diss., Northwestern University. Originally published as Vorgemach der musicalischen
Composition. Lobenstein: the author.
Stein, Carl. 1888. Logier’s Comprehensive Course in Music, Harmony, and Practical
Composition. Enl. rev. American ed. New York: Carl Fischer.
Stein, Deborah. 1985. Hugo Wolf ’s Lieder and Extensions of Tonality. Ann Arbor, MI: UMI
Research Press.
Straus, Joseph. 1982. “Stravinsky’s ‘Tonal Axis.’” Journal of Music Theory 26, no. 2: 261–90.
——. 1987. “The Problem of Prolongation in Post-tonal Music.” Journal of Music Theory 31,
no. 1: 1–21.
——. 1990. Remaking the Past: Musical Modernism and the Influence of the Tonal Tradition.
Cambridge, MA: Harvard University Press.
Subotnik, Rose. 1987. “On Grounding Chopin.” In Music and Society: The Politics of
Composition, Performance, and Reception, ed. Richard Leppert and Susan McClary.
Cambridge: Cambridge University Press.
Swinden, Kevin J. 2005. “When Functions Collide: Aspects of Plural Function in Chromatic
Music.” Music Theory Spectrum 27, no. 2: 249–82.
 Bibliography  227

Taruskin, Richard. 1985. “Chernomor to Kashchei: Harmonic Sorcery; Or, Stravinsky’s

‘Angle.’” Journal of the American Musicological Society 38, no. 1: 72–142.
——. 1996. Stravinsky and the Russian Traditions: A Biography of the Works through Mavra.
2 vols. Berkeley : University of California Press.
——. 2005. Music in the Nineteenth Century. Oxford History of Western Music, vol. 3. New
York: Oxford University Press.
Tchaikovsky, Peter Ilich. 1976 [1872]. Guide to the Practical Study of Harmony, trans.
Emil Krall and James Liebling. Bonn–Bad Godesberg: Forberg-Jurgenson. Originally
published as Rukovodstvo k prakticheskomu izucheniyu garmoniy. Moscow.
Telesco, Paula J. 1998. “Enharmonicism and the Omnibus Progression in Classical-Era
Music.” Music Theory Spectrum 20, no. 2: 242–79.
Temperley, David. 2001. The Cognition of Basic Musical Structures. Cambridge, MA: MIT
Press.
Tischler, Hans. 1964. Practical Harmony. Boston: Allyn & Bacon.
Todd, R. Larry. 1988. “The ‘Unwelcome Guest’ Regaled: Franz Liszt and the Augmented
Triad.” 19th-Century Music 12, no. 2: 92–115.
Tymoczko, Dmitri. 2005. “Voice Leadings as Generalized Key Signatures”. Music Theory
Online 11, no. 4. http://www.mtosmt.org/issues/mto.05.11.4/mto.05.11.4.tymoczko.pdf
——. 2006. “The Geometry of Musical Chords.” Science 313: 72–74.
——. 2009a. Comment on Kyle Gann, “So I’m Neo-Riemannian: Who Knew?” PostClassic.
http://www.artsjournal.com/postclassic/2009/03/so_im_neo-riemannian_who_knew.
html#comment-17813. Retrieved June 18, 2010.
——. 2009b. “Generalizing Musical Intervals.” Journal of Music Theory 53, no. 2: 227–54.
——. 2009c. “Three Conceptions of Musical Distance.” In Mathematics and Computation
in Music, ed. Elaine Chew, Adrian Childs, and Ching-Hua Chuan. Heidelberg:
Springer.
——. 2010. “Geometrical Methods in Recent Music Theory.” Music Theory Online 16, no. 1.
http://www.mtosmt.org/issues/mto.10.16.1/mto.10.16.1.tymoczko.html
——. 2011a. “Dualism and the Beholder’s Eye.” In The Oxford Handbook of Neo-Riemannian
Music Theories, ed. Edward Gollin and Alexander Rehding. New York: Oxford University
Press.
——. 2011b. A Geometry of Music: Harmony and Counterpoint in the Extended Common
Practice. New York: Oxford University Press.
Van den Toorn, Pieter. 1995. Music, Politics, and the Academy. Berkeley : University of
California Press.
Vogel, Martin. 1993 [1975]. On the Relations of Tone, trans. Vincent Jean Kisselbach, ed.
Carl A. Poldy. Bonn: Verlag für Systematische Musikwissenschaft.
Vogler, Georg Joseph. 1802. Handbuch zur Harmonielehre und für den Generalbass. Prague:
K. Barth.
Waller, Derek. 1978. “Some Combinatorial Aspects of Musical Chords.” Mathematical
Gazette 62, no. 419: 12–15.
Wason, Robert W. 1985. Viennese Harmonic Theory from Albrechtsberger to Schenker and
Schoenberg. Ann Arbor, MI: UMI Research Press.
Weber, Gottfried. 1846 [1817–21]. Attempt at a Systematically Arranged Theory of Musical
Composition, vol. 1. Translated from the 3rd edition by James F. Warner. Boston:
J. H. Wilkins and R. B. Carter. Originally published as Versuch einer geordneten Theorie
der Tonsezkunst. Mainz: B. Schott.
Webster, James. 1978–79. “Schubert’s Sonata Form and Brahms’s First Maturity.” Part I,
19th-Century Music 2, no. 1: 18–35; part II, 19th-Century Music 3, no. 1: 52–71.
——. 1991. Haydn’s ‘Farewell’ Symphony and the Idea of the Classical Style. Cambridge:
Cambridge University Press.
228  Bibliography

Weitzmann, Carl Friedrich. 2004 [1853]. “Two Monographs by Carl Friedrich Weitzmann,
I: The Augmented Triad.” Translated with an introduction by Janna Saslaw. Theory
and Practice 29: 133–228. Originally published as Der übermässige Dreiklang. Berlin:
T. Trautweinschen.
Werckmeister, Andreas. 1698. Die nothwendigsten Anmerkungen und Regeln, wie der Bassus
continuus oder General-Baß wol könne tractiret werden. Aschersleben: Gottlob Ernst
Struntz. Reprint, Michaelstein: Kultur- und Forschungsstatte Michaelstein, 1985.
Westergaard, Peter. 1996. “Geometries of Sounds in Time.” Music Theory Spectrum 18,
no. 1: 1–21.
Wimsatt, William C. 1981. “Robustness, Reliability, and Overdetermination.” In Scientific
Inquiry and the Social Sciences, ed. Marilynn B. Brewer and Barry E. Collins.
San Francisco: Jossey-Bass.
Winford, Donald. 2009. “On the Unity of Contact Phenomena and Their Underlying
Mechanisms: The Case of Borrowing.” In Multidisciplinary Approaches to Code Switching,
ed. Ludmila Isurin, Donald Winford, and Kees de Bot. Amsterdam: John Benjamins.
Woolard, Kathryn A. 1998. “Simultaneity and Bivalency as Strategies in Bilingualism.”
Journal of Linguistic Anthropology 8, no. 1: 3–29.
Zarlino, Gioseffo. 1968 [1558]. The Art of Counterpoint, trans. Guy A. Marco and Claude
V. Palisca. New Haven, CT: Yale University Press. Originally published as part III of
Le istitutioni harmoniche. Venice.
Zbikowski, Lawrence. 2002. Conceptualizing Music: Cognitive Structure, Theory, and
Analysis. Oxford: Oxford University Press.
INDEX

Abbate, Carolyn, 4n4, 22 Piano Trio in c minor, Op. 1 no. 3,

Adorno, Theodor, 10 197n2
agent, 47, 68–78, 93, 188, 197 Sonata for Violin and Piano in
Agmon, Eytan, 22, 37, 177, 179, 206 F major, Op. 24 (“Spring”), 2nd
Ahn, So-Yung, 95n7 mvt., 26–28, 45n4
all-interval tetrachord, 108–9 Symphony no. 5 in c minor, Op. 67,
Aristotle, 41n24 1st mvt., 72
arrow reversal, 204–5 Symphony no. 9 in d minor, Op. 125,
augmented triad, 19, 33–37, 43–46, 2nd mvt., 92, 183
56–58, 64–65, 83–86 Benda, Jiří
Piano Sonata in a minor, 1st mvt.,
Babbitt, Milton, 104n16, 184, 209 77–78
Bach, C. P. E. Berg, Alban
and relatedness of C major and Wozzeck, 167
E major, 67–68 Bering Strait problem, 14, 28, 88–89
Piano Sonata in f minor, H. 173 Blum, Stephen, 207n6
(Wq. 57/6), 1st mvt., 68–69 Boatwright, Howard, 44
Bach, J. S. Boretz, Benjamin, 141, 150–51
Fantasy and Fugue for Organ in g Boretz-group transformation, 155, 159
minor, BWV 542, 94 Boretz region, 141, 152–56, 159
Prelude in a minor, BWV 889 Boretz spider, 153–55
(Well-Tempered Clavier, vol. 2), Bot, Kees de, 201–2
107–8 Brahms, Johannes
Toccata in d minor, BWV 913, 183 Alto Rhapsody, Op. 53, 43
Baker, James M., 167 Concerto for Violin and Cello in a
balanced voice leading, 19–21, 159 minor, Op. 102, 1st mvt., 29–31
Balzano, Gerald, 209n8 Deutsches Requiem, Ein, Op. 45,
Bass, Richard, 136, 150n7, 151n9 2nd mvt., 96–97
Beethoven, Ludwig van Symphony no. 1 in c minor, Op. 68,
Acht Variationen über “Tändeln und 4th mvt., 191–94
Scherzen” (WoO 76), 207n6 Symphony no. 2 in D major, Op. 73,
Piano Sonata in d minor, Op. 31 no. 2 1st mvt., 92–94, 117–21
(“Tempest”), 1st mvt., 133–34; 3rd “Von ewiger Liebe,” Op. 43
mvt., 108–9 no. 1, 76
Piano Sonata in C major, Op. 53 “Walpurgisnacht,” Op. 75 no. 4,
(“Waldstein”), 1st mvt., 133–34 100n12
Piano Sonata in f minor, Op. 57 Brendel, Alfred, 49
(“Appassionata”), 1st mvt., 45n4, 48 Bribitzer-Stull, Matthew, 48n8

229
230  Index
bridging region, 84, 155–56, 166–67
Broersma, Mirjam, 201–2
Brower, Candace, 175–78, 184
Brown, Matthew, 11, 96n10, 98n11
Bruckner, Anton
“Ecce Sacerdos,” 91
Symphony no. 3 in d minor, 1st mvt.,
98–99
Callender, Clifton, 167n21
Capellen, Georg, 7, 140
Caplin, William E., 133
Capuzzo, Guy, 34n18, 66n8
Chafe, Eric, 151n9
chain
minor-third (T±3), 90–92
perfect-fifth, (T±5), 92–94, 160–61
semitone (T±1), 94–95, 160–61
see also L/R chain; N/H chain;
N/L chain; N/P chain; N/R chain;
S/L chain; S/P chain; sequence,
chromatic
Childs, Adrian P., 155, 160
Chopin, Frédéric
Ballade in g minor, Op. 23, 96–98
Etude in f minor, Op. 25 no. 2, 69n10
Fantasy in f minor, Op. 49, 98–100
Mazurka in fᅊ minor, Op. 6 no. 1, 69–71
Mazurka in f minor, Op. 68 no. 4,
160–61
Nocturne in bᅈ minor, Op. 9 no. 1,
69–71
Nocturne in G major, Op. 37 no. 2, 108
Piano Sonata in b minor, Op. 58,
4th mvt., 69n10
Prelude in e minor, Op. 28 no. 4,
160–66
Prelude in cᅊminor, Op. 45, 160
Christensen, Thomas, 139–41, 207n6
Clark, Suzannah, 113n1, 129–30
Clough, John, xi, 29n11, 34n17
code switching, 201–3
cognitive opacity, 202–5
coherence, 15
collective exchange, 155–56, 166–67
combination, see dissonance, strategies
for analyzing
common tones, 5–8, 17, 29, 34–36,
129–30, 209n7; see also pitch
retention loop
consonant triad, 1n1
contrary motion, 19–24, 103n15, 149,
154, 159
Cook, Robert, 31, 105n17, 145n4
Copley, R. Evan, xiv
Cornelius, Peter
“Ein Ton,” Op. 3 no. 3, 113
Covington, Kate R., 77
Cube Dance, 84–89, 101, 103–4, 152
Dahlhaus, Carl, 7n10, 10, 13, 143, 200
Darcy, Warren, 22, 133
Debussy, Claude
Prelude to the Afternoon of a Faun, 108
deletion, see dissonance, strategies for
analyzing
Dempster, Douglas, 96n10, 98n11
determination, double, see syntax,
double
developmental core, 25–26, 45, 49,
133–34
diatonic encapsulation, see encapsulation,
diatonic
diminished seventh chord
as dissonant sonority, 140, 182
as perfectly even chord, 150–55
disjunction, 4, 22, 106–9
dissonance, strategies for analyzing
combination, 140, 141–42
deletion, 140–41
reduction, 140–43
substitution, 140–41, 150–51
Distler, Hugo
Fürwahr, er trag unsere Krankheit, 108
DOM, 170–71
double-agent complex, 68–78
double syntax, see syntax, double
Douthett, Jack
and near evenness, 36n19
and Cube Dance, 84, 86
and Tristan-genus chords, 149n5, 155,
157–58n15, 158
role in developing pan-triadic
theory, xi
downshift voice leading, 84–85, 121,
154–55, 160–66
Draeseke, Felix, 12
dualism
harmonic, xiii, 32, 37–38, 46n6, 143
melodic, 32, 37–38, 93, 124–25

Durutte, Camille, 102n13
Dvořák, Antonín
Symphony no. 9 in e minor, Op. 95
(“From the New World”), 1st mvt.,
134–38
efficient voice leading, see voice leading
work, minimal
Elgar, Edward
King Olaf, 108
encapsulation, diatonic, 176–78, 189,
190–94
Engebretsen, Nora, 5
enharmonicism, 9, 11, 18, 23, 68–69, 72,
174–75
essential, 9
entropy, 106–9
equivalence class, 102–3
Euler, Leonhard, 28n10, 89n4
evolutionary theory, 205–7
Fauré, Gabriel
Requiem, Introit, 54–56
Fétis, François-Joseph
and chromaticism, 140
and species of tonality, 9–10, 13–14
and summoning, 206
uniformity and symmetry in, 47,
171, 179
Feurzeig, Lisa, 23
Feynman, Richard P., 41n24
Fisk, Charles, 49, 73
4-Cube Trio, 152, 157–59
Franck, César
Piano Quintet in f minor, 155n14
Freud, Sigmund, 41n24, 148
function, harmonic, 11, 114n3, 124–28,
170–71
fused-triad graph, 66–67
Galeazzi, Francesco, 5
Gamer, Carlton, 209n7
Gardner-Chloros, Penelope, 201–2
Gauldin, Robert, 162–63
generalized Weitzmann region (GWR),
166–67
geometric models, 14, 173–75
Goldenberg, Yosef, 66n8, 91n5
Gollin, Edward, 30n13, 95n8, 140–42,
189
Index  231
Gould, Stephen Jay, 206
Grieg, Edvard
Peer Gynt, 21
GWR, see generalized Weitzmann region
H, see hexatonic pole
Hantz, Edwin, 43n1
Harrison, Daniel
and atonal analysis of nineteenth-
century music, 13
and harmonic progression, 83n1
and dualism, xv, 46n6, 93–94
and Riemann’s treatment of
intonation, 177n7
and leading-tone agency, 47, 72
and rhetoric, 47, 194
and Riemannian function, 11, 114n3
and subposed dissonances, 143
and the Tonnetz, 28n10, 174n5
and transformational theory, 39
Hauptmann, Moritz, 5, 62n2, 92, 140,
142, 175n6
Haydn, Franz Joseph
Symphony no. 98 in Bᅈ major,
4th mvt., 26, 28
Symphony no. 99 in Eᅈ major,
1st mvt., 78
Headlam, Dave, 96n10, 98n11
Helmholtz, Hermann von, 5, 140
Hepokoski, James, 133, 133n11
hexatonic cycle, 17–24, 84–86, 169–70
in Beethoven, Piano Sonata in f
minor, Op. 57 (“Appassionata”),
1st mvt., 48
in Beethoven, Sonata for Violin and
Piano in F major, Op. 24
(“Spring”), 2nd mvt., 26–27
in Brahms, Concerto for Violin and
Cello in a minor, Op. 102, 1st mvt.,
29–31
in Brahms, Symphony no. 1 in
c minor, Op. 68, 4th mvt., 193–94
in Haydn, Symphony no. 98 in
Bᅈ major, 4th mvt., 26
in Liszt, Consolation no. 3 in
Dᅈ major, 30, 186-88
in Mozart, Symphony in Eᅈ major,
K. 543, 4th mvt., 25–26
in Schubert, Piano Trio in Eᅈ major,
Op. 100, 1st mvt., 27
232  Index
hexatonic cycle, (continued)
in Schubert, Mass in Eᅈ major, D. 950,
Sanctus, 31–32
in Wagner, Parsifal, 31–32
hexatonic-group transformations, 90–91,
105–6
hexatonic pole (H), 31, 103n15, 107–8
in Bach, J. S., Prelude in a minor,
BWV 889 (Well-Tempered Clavier,
vol. 2), 107–8
in Chopin, Mazurka in f minor,
Op. 68 no. 4, 161
in Dvořák, Symphony no. 9 in
e minor, Op. 95 (“From the
New World”), 1st mvt., 135
in Schubert, Piano Sonata in A major,
D. 959, 1st mvt., 49
in Schubert, “Auf dem Flusse”
(Winterreise, D. 911), 123
in Schubert, Mass in Eᅈ major, D. 950,
Sanctus, 31–32
in Wagner, Parsifal, 31–32, 107–8,
145–48
hexatonic region, see hexatonic cycle
hexatonic scale, 20, 207
hexatonic transformations, see hexatonic-
group transformations
Hoffmann, E. T. A., 1, 132
homophonous diamorph, 40, 165–66,
201–2
Hostinský, Otakar, 28n10, 113, 116
Hyer, Brian, 29, 69n10, 174–75
hypermeter, 117–21, 136
idealized voice leading, 6, 19, 89n4
Jackendoff, Ray, 53n11
Jentsch, Ernst, 148
Kant, Immanuel, 46
Karg-Elert, Sigfrid, 64, 155, 166
Katz, Adele, 11
Kirnberger, Johann Philipp, 1, 46, 140
Klumpenhouwer, Henry, 38–39
Kodály, Zoltán
Mountain Nights, 108
Kollektivwechsel, see collective exchange
Komar, Arthur, 131n10
Kopp, David, xi, 11, 12n12, 30n13, 37, 64
Kramer, Lawrence, 4n4, 171n1
Kramer, Richard, 68, 124
Krause, K. C. F., 5
Krumhansl, Carol, 102n13, 204n5
Kurth, Ernst, 10–11, 12, 22n7,
46n6, 169
L, see Leittonwechsel
Langlé, Honoré de, 62n3, 92, 95n7,
207n6
Larson, Steve, 162
Leittonwechsel (L), 29–30, 38, 65–66
Lendvai, Ernő, 22n6, 125, 142, 155n14
Lerdahl, Fred
and Chopin’s e minor Prelude, 162
and diatonic space, 177
and double syntax, 200
and hexatonic versus octatonic
space, 91n5
and structural parallelism, 53n11
and the Faith Proclamation from
Parsifal, 189–90
and the key sequence of
Dichterliebe, 131
and Tristan und Isolde, 151n9
Lewin, David
and classical versus Wagnerian
tonality, 106
and directed voice leading, 84n2
and enharmonic transformation, 72
and exchange operations, 105n17
and harmonic function, 124–25
and Riemannian versus scale-degree
space, 173–74, 178
and syntax in music, 13
and triadic transformations, 29, 64,
171–72, 189–90
and Tristan-genus chords, 156,
189n11
role in developing pan-triadic theory,
xi, xiii
Liszt, Franz
“Blume und Duft,” 43
Consolation no. 3 in Dᅈ major, 30,
172–73, 180–82, 186–88
Faust Symphony, 1st mvt., 51, 63
Grande fantaisie symphonique über
Themen aus Berlioz’ “Lélio” (“Lélio
Fantasy”), 91–92, 100–102
Legende vom heiligen Stanislaus,
Die, 108

Liebestraum no. 3 (“O Lieb, so lang du
lieben kannst”), 63n5
Malediction for piano and string
orchestra, 91
Missa pro organum lectarum, Kyrie,
128–29, 189
Penseroso, Il (Années de pèlerinage,
vol. 2), 72–73, 95
Transcendental Etude no. 6, “Vision,”
63n5
Transcendental Etude no. 8, “Wilde
Jagd,” 95n7
Longyear, Rey M., 77
Lorenz, Alfred, 22n7, 46n6, 148
Louis, Rudolf, 22n6, 46n6
Lowinsky, Edward E., 92, 179
L/P chain, see hexatonic cycle
L/R chain, 92, 183–84
in Bach, J. S., Toccata in d minor,
BWV 913, 183
in Beethoven, Symphony no. 9 in d
minor, Op. 125, 2nd mvt., 92, 183
in Brahms, Ein deutsches Requiem,
Op. 45, 2nd mvt., 96–97
in Brahms, Symphony no. 2 in D
major, Op. 73, 1st mvt., 92–94
in Brahms, Symphony no. 1 in c
minor, Op. 68, 4th mvt., 193–94
in Bruckner, Symphony no. 3 in d
minor, 1st mvt., 98–99
in Chopin, Ballade in g minor, Op. 23,
96–98
in Chopin, Fantasy in f minor, Op. 49,
98–99
in Wagner, Parsifal, 190
Lubin, Steven, 133n12
Mahler, Gustav
Symphony no. 1 in D major, 4th mvt.,
106–7
Marchettus of Padua, 7, 185
Martino, Pat, 34n18
Marx, Adolph Bernhard, 1, 5
Masson, Charles, 7n11
maximal evenness, 37
McClary, Susan, 4n4, 112
McCreless, Patrick, 66n8, 197n2, 206
McKinney, Timothy R., 43n1
MED, 171–72
Meeus, Nicholas, 125n7
Index  233
Mehrdeutigkeit, 44, 57, 72
Mendelssohn, Felix
Variations sérieuses, Op. 54, 108
meter, 53–54; see also hypermeter;
rhetoric
minimal perturbation, see near evenness
Mitchell, William J., 11, 153n11, 190
mixture, modal, 184, 195
Möbius strip, 173–74, 178–79
Monteverdi, Claudio
Incoronazione di Poppea, L’, 143n3
Moomaw, Charles Jay, 43–44
Mooney, Michael Kevin, xi
Morgan, Robert P., 46, 151n9, 205
Morris, Robert D., 62n2, 108
Motte, Diether de la, 205
Mozart, Wolfgang Amadeus
Fantasy in c minor, K. 475, 108
Minuet in D major, K. 355/576b
(attrib.), 44
Symphony in Eᅈ major, K. 543, 4th
mvt., 25–26, 28
Mussorgsky, Modest
Boris Godunov, 155n14
Muysken, Pieter, 201
mystic chord, 166–67
N, see nebenverwandt
near evenness, 34, 166–67, 204
of consonant triads, 34–37, 40, 179,
184–85, 207–8, 210
of Tristan-genus chords, 151–52, 189
relation to acoustic consonance, 40,
40n23
relation to efficient voice leading,
34–37, 40
nearly even, see near evenness
nebenverwandt (N), 47–48, 61–62,
122–24, 184–85
neighborhood, 113–14, 116
neo-Riemannian theory, xiii–xiv
Newcomb, Anthony, 123
N/H chain, 91
N/L chain, 94–95
in Dvořák, Symphony no. 9 in e minor,
Op. 95 (“From the New World”),
1st mvt., 136–37
in Liszt, Grande fantaisie symphonique
über Themen aus Berlioz’ “Lélio”
(“Lélio Fantasy”), 100
234  Index

N/P chain, 93–94 Prokofiev, Sergei

N/R chain, 61–62, 76–78 War and Peace, 95n8
in Beethoven, Piano Sonata in prolongation, 22n7, 33, 45–46, 141
C major, Op. 53 (“Waldstein”), Prout, Ebenezer, 22n6
1st mvt., 134 Puccini, Giacomo
in Benda, Piano Sonata in a minor, Tosca, 21
1st mvt., 77–78
in Chopin, Fantasy in f minor, R, see relative
Op. 49, 99 R*, 155
in Haydn, Symphony no. 99 in Rameau, Jean-Philippe, 1, 95, 139–43,
Eᅈ major, 1st mvt., 77–78 151, 179, 182n9
in Liszt, Faust Symphony, reciprocity, 46–48
1st mvt., 63 reduction, see dissonance, strategies for
in Schubert, Piano Sonata in c minor, analyzing
D. 958, 4th mvt., 77–78 Rehding, Alexander, 95, 177
in Schubert, Symphony no. 4 in Reise, Jay, 167n21
c minor, D. 417, 1st mvt., 62 relative (R), 62
in Schubert, Symphony no. 9 in rhetoric, 47, 194
C major, D. 944, 1st mvt., 63 Riemann, Hugo
and dissonance, 140, 141, 180
octatonic pole, see collective exchange and dualism, 32, 38, 143
octatonic region, 155, 157, 159 and enharmonicism, 22n6
octatonic-group transformations, and harmonic distance, ix, 1, 5
155–56, 159 and harmonic function, 11, 114n3,
octatonic scale, 155, 207 124, 127–28, 177, 197
octatonic transformations, see octatonic- and intonation, 89, 177
group transformations and neo-Riemannianism, xiii
Oettingen, Arthur von, 29, 32, 38, 47, and the Tonnetz, 28n10, 65, 113,
62n1, 65, 89n4, 143, 182n9 175n6
omnibus progression, 149, 207 and Wechsel, 29
overdetermination, 40–41, 165–66, 179, Riemannian tonal space, 173–74,
202–3, 208–10 Rimsky-Korsakov, Nikolai
Antar (Symphony no. 2), 1st mvt., 21,
P, see parallel 51–54, 56
palindrome, 121, 128–31 Skazka, 91
pan-triadic, xiv Rings arrows, 170–71, 176–83
paradigmatic analysis, 122 Rings, Steven, 6n8, 15, 64n7, 73n13,
parallel (P), 29–30, 38, 65–66 169–71
Parly, Nila, 22 Rosen, Charles, 49
parsimony, voice-leading, see voice- Rothstein, William, 113
leading work, minimal Rouget, Gilbert, 23
Patel, Aniruddh, 203
Peck, Robert, 158n15 S (similar-motion transformations on
pitch retention loop, 116–21 Tristan-genus chords), 155
Plato, 175 S2, 160, 167
Popovic, Igor, 113 S3(2), 155, 160
poststructuralism, 4, 109 S3(4), 155, 160
preadaptation, 206–210 S4, 160, 167
Proctor, Gregory, 6n7, 26n9, 33n16, S5, 167
48n8, 199 S6, 155

Salzer, Felix, 11
Saslaw, Janna, 47, 56, 56n13
Satyendra, Ramon, 25, 43n1, 53n10, 128
Schachter, Carl, 11, 21, 119n4, 161, 163n17
Schenker, Heinrich
and consonance versus dissonance,
22n7, 45
and equal division of the octave,
26n9
and keys versus chords, 1
and mediant relationships, 98n11
and mode mixture, 184, 196–97n1
approaches to dissonance, 140, 141
Schmalfeldt, Janet, 67
Schoenberg, Arnold
and strong versus weak harmonic
progressions, 125n7
and the law of least motion, 7n11
String Trio, Op. 45, 21
Schreker, Franz
Kammersymphonie, 142
Schubert, Franz
“Auf dem Flusse” (Winterreise,
D. 911), 122–25, 197n2
Characteristic Allegro for Piano Four
Hands (“Lebensstürme”), D. 947,
64n7
“Doppelgänger, Der” (Schwanengesang,
D. 957), 113–16
Impromptu in Eᅈ, Op. 90 (D. 899)
no. 2, 74–75, 170–71
Impromptu in Gᅈ, Op. 90 (D. 899)
no. 3, 74–75
Klavierstuck in Eᅈ major, D. 946 no. 2
(Drei Klavierstücke), 91
“Liedesend’,” D. 473, 129–31
Mass in Aᅈ major, D. 678, Sanctus,
64n7, 91
Mass in Eᅈ major, D. 950, Sanctus,
30–32
“Nacht und Träume,” D. 827, 21
Piano Sonata in c minor, D. 958,
4th mvt., 77–78
Piano Sonata in A major, D. 959,
1st mvt., 48–50
Piano Sonata in Bᅈ major, D. 960,
1st mvt., 2–4, 6–8, 64n7, 125–27,
169–70
Piano Trio in Eᅈ major, Op. 100,
1st mvt., xi, 27–28
Index  235
Rosamunde Overture, see Zauberharfe,
Die, Overture
“Stadt, Die” (Schwanengesang,
D. 957), 47
String Quartet in d minor, D. 810
(“Death and the Maiden”), 94
String Quintet in C major, D. 956,
2nd mvt., 64n7
Symphony no. 2 in Bᅈ major, D. 125,
4th mvt., 45
Symphony no. 4 in c minor, D. 417,
1st mvt., 62–63
Symphony no. 9 in C major, D. 944,
1st mvt., 63; 3rd mvt., 195–99
“Trost,” D. 523, 113n2
Zauberharfe, Die, Overture
(Rosamunde Overture), 85–88
“Zügenglöcklein,” D. 871, 113
Schumann, Robert
Dichterliebe, Op. 48, 131–33; no. 1,
“Im wunderschönen Monat Mai,”
75–76; no. 5, “Ich will meine Seele
tauschen,” 76
Piano Trio no. 1 in d minor,
Op. 63, 108
Scriabin, Alexander
Feuillet d’album, Op. 58, 167–68
Sechter, Simon, 46
Seidel, Elmar, 45n3, 207n6
semantics, 14, 21–22, 171, 171n, 178–80,
184, 186, 188
sequence
diatonic sequence, 10, 47, 207
chromatic, 10, 19, 47, 89–90
see also chain
Sessions, Roger, 13
Shamgar, Beth, 4
Siciliano, Michael, 34n18, 113n1
single semitonal displacement, 18,
34–35, 37, 55, 151n10, 153;
see also Leittonwechsel;
parallel
S/L chain, 91–92
slide (S), 64
Slobin, Mark, 202n4
Slonimsky, Nicolas, xiv
Smith, Charles J., 199–201
Smith, Peter H., 47n7
Soderberg, Stephen, 151n10
Sorge, Georg Andreas, 44
236  Index
S/P chain, 95
in Dvořák, Symphony no. 9 in
e minor, Op. 95 (“From the
New World”), 135
in Liszt, Il Penseroso, 95
spoiler tone, 150–54, 167–68
Stein, Carl, 92
Steinbach, Peter, 84, 155, 157–58n15
Strauss, Richard
“Frühling” (Four Last Songs), 32–33,
79–81
Salome, 21, 108
SUBD, 170–71
SUBM, 171–72
substitution, see dissonance, strategies for
analyzing
substitution, transformational, see
transformational substitution
sum, pitch-class, 104–5; see also
voice-leading zone
Swinden, Kevin J., 73n12
syntax, 12–15, 39–40
double, 11–12, 169–72, 195–201, 205–8
syntonic comma, 175–77
syntonic image, 176–78
T (transposition), see chain
Tallis, Thomas
Psalm tune no. 3 for Psalter of
Archbishop Parker, 143n3
Tarnhelm, 21–23, 72–73
Taruskin, Richard, xv, 21, 23, 207n6
Tchaikovsky, Peter Ilich, 5
third relations, 19, 171
Thuille, Ludwig, 22n6, 46n6
Todd, R. Larry, xi, 43n1
Tonnetz, 27–30
and Cube Dance, 86–89, 100–101
and diatonic progressions, 175–79
comparison to other geometrical
models, 65–67
convertible (circular versus planar),
174–75
flexibility of, 114–16
spatial orientation of, 89, 175n6
three-dimensional, 141–42
torus, 88
transformation, 29–32, 38, 39, 90,
155–56, 159; see also
transformational substitution
compound, 30, 60–65,
equivalence among, 102–6
transformational subsitution, 95–96,
121–24, 160–61
Tristan genus, 148–52
Tymoczko, Dmitri
and dualism, xiii, 37–38
and measurements of voice-leading
distance, 6n8, 7, 66, 84, 88, 159
and near evenness, 34, 36n19, 40n23,
56n12, 166n20
and octatonic triadic systems, 91n5
and representation of
transformational substitution,
96n9
and scalar tonality, 207
and triads as homophonous
diamorphs, 40, 209–10
and Tristan-genus chords, 149nn5–6,
153n12, 158n15, 160–61, 164n19
and voice-leading direction, 19n3, 32,
89n4
uncanny, 21–24
uniformity, law of, 10, 47–48
upshift voice leading, 84–85, 121, 131–38,
154–55, 166
Verdi, Giuseppe
“Ah sí, ben mio” (Il Trovatore), 116
Vogler, Georg Joseph, 44–45, 72, 92
voice-leading work, 6–8
minimal, 6–8, 17–19, 33–37, 38, 151,
166–67
voice-leading zone, 102–6, 154–55, 159
Wagner, Richard
Götterdämmerung, 21, 108–9, 156,
Parsifal, 21, 30–32, 43, 78–79, 107–8,
144–48, 172, 189–91, 194
Ring des Nibelungen, Der, 21, 22–23,
72–73, 155n14
Siegfried Idyll, 43
Tristan und Isolde, 143–44, 149–55
Walküre, Die, 4n4, 11, 43
Waller, Derek, 18n1
Wason, Robert W., 45n3
Weber, Gottfried, 7n11, 8, 53n11, 131,
176–77
Webster, James, 78, 113n1

Weitzmann, Carl Friedrich
and the emancipation of dissonance,
7n11
and the relationship between
consonant and augmented triads,
56–58, 83–84, 151–53
metaphorical presentation of the
augmented triad by, 43
nebenverwandt in, 46–47
triads versus tonics in, 61
Index  237
Weitzmann-group transformation,
90–91, 105–6
Weitzmann region, 59–61
graphical representation, 64–65
Werckmeister, Andreas, 92
Wolf, Hugo
“Das Ständchen,” 45n4
Zarlino, Gioseffo, 7
zone-diametric triads, 107–8