Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
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
Audacious Euphony: Chromaticism and the Triad’s Second Nature, Richard Cohn
Audacious Euphony
Richard C ohn
1
1
Oxford University Press, Inc., publishes works that further
Oxford University’s objective of excellence
in research, scholarship, and education.
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
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
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
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.
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
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.
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
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
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
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:
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
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
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
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).
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
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
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.
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
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.
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.
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
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 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
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
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
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
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(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.
19. The general principles are established, in different ways, in Douthett 2008 and in Tymoczko
2011b.
CHAPTER 2 Hexatonic Cycles 37
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
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
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.
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
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.
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
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
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.
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
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 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
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
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
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
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,
Figure 3.9. From Weitzmann’s Der übermässige Dreiklang. Upper- and lower-case
letters are the roots of major and minor triads respectively.
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 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
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.
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
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
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
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
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
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.
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
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
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
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)
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
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).
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
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
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
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
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
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.
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 5.6. 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
Chromatic Sequences
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
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
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.
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
(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
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.
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
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
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.
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.
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.
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
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.
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 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
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
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.
111
CHAPTER 6 Navigating the Universe 113
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.
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
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 →
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
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.
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
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
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
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
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
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
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 6.17. 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
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.
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
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.
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
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
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.
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 7.6. 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.
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 7.9. Prelude to Tristan und Isolde, opening measures. Each filled notehead
is a “spoiler” that lies outside of the majority diminished seventh chord.
(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
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 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
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
Brünnhilde’s Immolation
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.
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
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
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 7.20. 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
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
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.
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
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
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
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.
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
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
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
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.
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
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
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
“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
Horizontal Extensions
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
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
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.
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.
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
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).
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
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
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
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.
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
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
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
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.
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
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.
211
212 Glossary
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
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
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
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
229
230 Index