Harmony in Mendelssohn and Schumann

This innovative book continues David Damschroder’s radical

reformulation of harmonic theory, presenting a dynamic
exploration of harmony in the compositions of Mendelssohn and
Schumann, two key figures of nineteenth-century classical music.
This volume’s introductory chapters creatively introduce the basic
tenets of the system, with reference to sound files rather than notated
music examples permitting a more direct interaction between reader
and music. In the Masterpieces section that follows, Damschroder
presents detailed analyses of movements from piano, vocal, and
chamber music, and compares his outcomes with those of other
analysts, including Benedict Taylor, L. Poundie Burstein, and Peter
H. Smith. Expanding upon analytical practices from the eighteenth
and nineteenth centuries, and strongly influenced by Schenkerian
principles, this fresh perspective offers a stark contrast to
conventional harmonic analysis – in terms of how Roman numerals
are deployed and how musical processes are described in words.

d a v i d d a m s c h r o d e r is Professor of Music Theory at the

University of Minnesota. His current research focuses on
harmony in tonal music, a project that began with a careful
examination of historical analytical practices, the basis for his
Thinking About Harmony: Historical Perspectives on Analysis
(Cambridge, 2008). He has since published studies on harmony
in selected composers including Haydn, Mozart, Beethoven,
Schubert, and Chopin, and he is the author of Tonal Analysis:
A Schenkerian Perspective (W. W. Norton, 2017).
Harmony in Mendelssohn and
Schumann

david damschroder
University of Minnesota
University Printing House, Cambridge CB2 8BS, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,
New Delhi – 110025, India
79 Anson Road, #06–04/06, Singapore 079906

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781108418034
DOI: 10.1017/9781108284110
© David Damschroder 2018
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2018
Printed in the United Kingdom by Clays, St Ives plc
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloging-in-Publication Data
Names: Damschroder, David.
Title: Harmony in Mendelssohn and Schumann / David Damschroder.
Description: Cambridge, United Kingdom ; New York, NY : Cambridge
University Press, 2017. | Includes bibliographical references and index.
Identifiers: LCCN 2017033145 | ISBN 9781108418034 (alk. paper)
Subjects: LCSH: Mendelssohn-Bartholdy, Felix, 1809–1847 – Harmony. |
Schumann, Robert, 1810–1856 – Harmony. | Music – 19th century – Analysis, appreciation.
Classification: LCC MT90 .D35 2017 | DDC 781.2/5–dc23
LC record available at https://lccn.loc.gov/2017033145
ISBN 978-1-108-41803-4 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
 Contents

Preface [page vii]

part i methodological orientation: harmonic

analysis through listening [1]

1 Foundational diatonic processes [3]

2 Surges [16]

3 IV5–6 V [24]
4 I5–6 II [31]
5 Surging 6-phase chords [42]

6 Surges with ninths [50]

7 Colorful variants of II [60]

8 III along the path from I to V [69]

part ii masterpieces [79]

9 Mendelssohn: Octet in E♭ Major (op. 20), movement 1 in response

to Greg Vitercik and Benedict Taylor [81]
10 Mendelssohn: Song without Words in F Major (op. 85, no.1) in
response to Allen Cadwallader [116]

11 Schumann: “Warum?” from Phantasiestücke (op. 12) in response

to L. Poundie Burstein [127]
12 Mendelssohn: Song without Words in A♭ Major (op. 53, no.1) in
response to Yosef Goldenberg [136]

13 Schumann: Three songs from Liederkreis (op. 39) in response to

Charles Burkhart and David Ferris [146]

v
vi Contents

14 Schumann: “Im wunderschönen Monat Mai” from

Dichterliebe (op. 48, no. 1) in response to Deborah Stein [171]

15 Schumann: Sonata in A Minor for Violin and Piano

(op. 105) in response to Peter H. Smith [183]

Notes [232]
Select bibliography [271]
Index of Mendelssohn’s works [277]
Index of Schumann’s works [278]
Index of names and concepts [279]
Preface

How did composers working just before or during the first half of the
nineteenth century conceive of and apply harmony? My emerging answer
to that question has by this point devoted due attention to the four
superlative composers active in or near Vienna in the decades around
1800 (Harmony in Haydn and Mozart, Harmony in Beethoven, and
Harmony in Schubert), as well as to a wondrous expatriate Polish pianist/
composer in Paris a bit further into the century (Harmony in Chopin). Now
it is time to assay what Mendelssohn and Schumann were accomplishing in
various German locales.
Though textbook and treatise authors of the era were active in building
analytical systems to make sense of the contemporary harmonic practices
(generally employing Roman numerals, as I relate in Thinking About
Harmony), their budding efforts have been extensively transformed in
my writings, in part because I incorporate notions proposed by later
thinkers (Heinrich Schenker in particular) and in part through my
willingness to jettison aspects of conventional modern harmonic analysis
stemming from those early efforts in favor of fresh and (I trust) improved
ways of proceeding that may more fully unlock for us the processes these
composers were pursuing. Though my work is intended mainly for
graduate students and professional musicians, I hope that my
reconfigured harmonic theory also will be introduced at the foundational
level of instruction. Harmony in Beethoven offers an inviting
Harmonielehre that might aptly supplement any of the standard
undergraduate harmony texts, giving initiates who may have become
complacent or indifferent an eye-opening exposure to a new way of
thinking about the topic.
As a complement to the Harmonielehre, this volume opens with
Harmonic analysis through listening (chapters 1 through 8), written in
a way that should be accessible to undergraduates while concurrently
offering more seasoned readers plenty of rewarding content. It is curious
yet true that, though music enters our consciousness through the ears,
almost all of what you might draw upon to assist in developing your
analytical capacities is absorbed through the eyes. I have taken the vii
viii Preface

unusual step of here introducing the foundations of my harmonic

perspective through the act of listening to music. In contrast to passive
reading about harmonic analysis, you will be invited to ponder questions
that get to the heart of how a passage you listen to is conceived
harmonically and realized in its details, and then to compare your
responses to my suggestions. These featured excerpts may be accessed via
audio examples available on this volume’s web page (www.cambridge.org
.9781108418034). Rhythmic grids annotated with location symbols
provide the means of identifying discrete moments within each sounding
excerpt, so that both the questions and the responses can be crafted with
specificity. (It is important that you peruse my proposed responses in the
endnotes even when you are sure that your own are correct, since I may
introduce terminology or a strategy that will be needed in future
responses.) I have indicated each question’s level of challenge by the
number of bullets (•, ••, or •••) to the left of its identifying number, so
that you have the option of pursuing only the easier questions during a first
pass through the chapters, returning to the others later. Each listening
excerpt is introduced by some preliminary commentary and the vocal
generation of its essential features, focusing on the harmonic
relationships that will be featured in what follows. Though you might be
taken aback by my request that you proceed through eight chapters of
a “scholarly” book in this interactive mode, I hope that you will find it to be
a transformative experience, touching on vital components of musical
perception and comprehension that are difficult to access through more
conventional modes of author–reader interactions.
This book’s centerpiece is a Masterpieces section offering detailed
analyses of compositions by Mendelssohn and Schumann, featuring the
tools that I advocate for harmonic analysis. Roman numerals generally are
displayed in the context of Schenkerian graphs, which provide insights
regarding harmony’s hierarchical organization as well as on a range of
other parameters. (During this portion of the book, some prior exposure to
Schenkerian analysis is assumed, either through my Tonal Analysis:
A Schenkerian Perspective, or by some other means.) As has been the case
also in the earlier analytical volumes of my Harmony Project, I guide you
through a direct comparison of each analysis with an interpretation by
another prominent analyst (or sometimes two others), thereby deepening
your perceptions regarding these works and highlighting what is at stake in
the analytical process. (These alternative analyses all appear in publications
that should be available at any collegiate music library. The critiques are set
off from the main flow of my analyses by shading.) Consequently, my
 Preface ix

Harmony in . . . volumes not only point a way forward for the study of
nineteenth-century music but also together comprise a unique and wide-
ranging assessment of the state of tonal analysis in English-language
scholarship over the past fifty or so years. Since my Masterpieces
chapters already offer substantial analytical challenges, I have not
endeavored to extend the project’s purview even further to incorporate
assessments of publications in other languages. (Such works occasionally
have been addressed in the endnotes.) Likewise some worthy analytical
publications were passed over because they focus on issues only marginally
related to harmony, making the sort of comparative analysis pursued here
unworkable.
Though I completed my first two Cambridge books without a clear sense
of what ultimately was to emerge, or even awareness that something
warranting being called a Harmony Project was in the works, by now my
six monographs from Cambridge together constitute a bountiful and
unified body of analytical commentary on this important repertoire.
I intend next to explore harmony in music after 1850, leading in due
time to Debussy.
I thank the University of Minnesota for granting me a sabbatical leave
permitting a year of uninterrupted work on this volume and for the
support of an Imagine Fund award that both covered the costs associated
with the music examples and sound files and allowed me to acquire books
and to visit major research libraries. I am grateful to the New York Public
Library, Astor, Lenox and Tilden Foundations, for allowing me to purchase
on microfilm and to make references to the Oster Collection: Papers of
Heinrich Schenker. As in the earlier volumes of my Harmony Project,
Peter Smucker has provided expert setting of the music examples.

Conventions regarding note relations, chords, keys,

and Roman numerals

Pitch simultaneities (such as C-E-G) are indicated using hyphens (-), while
pitch successions (such as C–E–G) are indicated using dashes (–).
Direction may be indicated in melodic succession: ascending as C<E<G,
descending as G>E>C. A black arrow may be used to indicate
a descending-fifth relationship that is or emulates a V(7)–I succession,
whereas an outline arrow may be used to indicate a succession from
a chord of the augmented-sixth type; for example, C➔F–D➔G➔C;
C–A♭–D⇨G➔C.
x Preface

Keys and chords are distinguished as follows: C Major (with a capital M)

is the key of C Major; C major (with a small m) is a C major chord.
Unless another analyst’s methodology is being discussed, Roman
numerals are presented in capital letters regardless of a chord’s quality,
modified by one or more accidentals if the chord is altered. Thus C Major:
I II V I and not I ii V I; and A Minor: I II V♯ I♯ (closing on a major tonic),
not i ii° V I. An accidental to the left of the numeral corresponds to the
chord’s root, to the right corresponds to its third. If the chordal fifth,
seventh, or ninth is altered, the analytical symbol will incorporate the
5♮
corresponding Arabic numeral, as in C Minor: II ♯ . (Arrow notation –
here II➔ – offers an attractive, though less precise, alternative to the
complete analytical symbol.) The bullet symbol (•) indicates an absent
root. For example, B-D-F in C Major will be analyzed as V•7 (or, with less
precision, as V➔).
Likewise a progression of chordal roots generally is presented in capital
letters (C–D–G–C), though on occasions when quality is a factor in the
discussion a capital letter may refer to major quality, a small letter to minor
quality, and a small letter followed by a degree circle (°) to diminished
quality; for example, C–a–F–d–b°–G–e–C.
A bracket is used to connect the analytical notation for two musical
events that normally would follow one another but that in the context
————
under discussion occur at the same moment; for example, C| F♯ B | E when
an F♯-A♯-C♯ chord sounds with, rather than before, root B in a descending
circle of fifths.
Parentheses around a pitch in an analytical example indicate that it is not
actually present in the score, though it is understood. Parentheses around
analytical notation may refer to the expansion of a deeper-level harmony (for
example, when I is expanded by I IV V I) or to the harmonic assertion of a
voice-leading phenomenon (for example, when the 6 phase of
a I5–6, as in C-E-G to C-E-A, asserts the harmonic role of VI). Open
parentheses designate a voice-leading transition between two harmonies. For
example, I ( ) IV indicates that the chords between I and IV (perhaps a circular,
parallel, or sequential progression) do not themselves participate in the
harmonic progression, but instead serve to connect the harmonies I and IV.
When a score’s chordal spellings do not coincide with the structurally
appropriate spellings (for example, the substitution of easier-to-read F♯-
A-C♯ for cumbersome G♭-Bº-D♭), I generally will use the structurally
appropriate spellings in my examples and commentaries, often placing
the enharmonic spellings within square brackets to assist readers in
locating the pitches in question within the score.
 Preface xi

I pay very close attention to hierarchies among pitches and chords.

To alert readers to various hierarchical relationships I often will
underline some pitch names to indicate their hierarchical prominence.
For example, C<E D>B C above bass C–G–C conveys the relationship
between two unfolded strands: a more prominent outer strand E>D>C,
and a subordinate inner strand C>B<C.
Because diverse musical contexts are analyzed using graphs, it is difficult
to pin down precise guidelines for how their notation should be crafted and
read. Many styles of “Schenkerian” notation have appeared since the
publication of Schenker’s Free Composition (hereafter abbreviated as FC),
which itself does not present a single normative style. I regard the creation of
a reductive graph as an art, endeavoring to use notation that is as clear and
informative as possible. In general, open noteheads in my graphs represent
deeper structural or harmonic events than filled-in noteheads, while notes at
the endpoints of beams or slurs are deeper than internal notes. Notes
connected to a beam by a stem are more integral to the structure than
those that are not. Occasionally annotations using abbreviations (including
those pertaining to form borrowed from James Hepokoski and Warren
Darcy’s Elements of Sonata Theory, Oxford University Press, 2006)
indicate functions of individual pitches or formal events, as follows:
ant. anticipation
C closing zone
CP chromatic passing note
CV chromatic variant
EEC essential expositional closure
ESC essential structural closure
HC half cadence
IAC imperfect authentic cadence
IN incomplete neighboring note
MC medial caesura
N neighboring note
P an individual pitch: passing note
P form: primary-theme zone
PAC perfect authentic cadence
prg. progression
S secondary-theme zone
susp. suspension
TR transition
W wobble
xii Preface

Of course, the graphs often will incorporate Roman numeral harmonic

analyses, and in this regard I sometimes depart from Schenker’s practice.
Because it is innovative, I document my Roman numeral usage very
carefully as the chapters unfold.
Because measure numbers are a pervasive feature in my close analyses,
I have developed an abbreviated style of reference, in the form measurebeat.
For example, the symbol 23 indicates the third beat of measure 2. Generally
2
the word “measure” will not precede the number. I regard measures in 2
6
and 8 as containing two beats. A measure designation such as 14/16 means
that a given chord is prolonged from measure 14 through measure 16, with
contrasting content occurring between statements of the chord, whereas
the designation 14–16 indicates a continuous prolongation of a single
chord without significant internal contrast. The symbol 15|16 indicates
measure 16 along with its upbeat. When an x appears among the measure
numbers in a music example, it signifies either that the example’s content
at that point does not actually sound but instead is suggested by the context
or that the example displays a hypothetical continuation that the composer
does not in fact pursue.
part i

Methodological orientation: harmonic

analysis through listening
1 Foundational diatonic processes

Though any of the twelve pitch classes may sound during a tonal
composition, certain pitches will be more prominently featured since
in all cases a key will be established. For example, in A Major (the key
of our first example, below) the pitches A, C♯, and E are especially
prominent, as members of the tonic triad. Their most frequent inter-
actions will be with their diatonic cohorts B, D, F♯, and G♯. The other
five pitch classes (whose spellings will vary according to usage) may be
deployed to add greater emphasis or variety (topics explored beginning
in chapter 2), or instead to fulfill subordinate embellishing or connec-
tive roles. During a harmonic progression different pitches in turn
come to the fore. For example, in the succession from V to I, the
dominant’s E, G♯, and B at first will supersede even the tonic pitch A in
prominence. With the resolution to I, A and C♯ take over from G♯ and
B as highlighted pitches.
Our first example lacks even that level of harmonic variety. Because of its
unique location at the end of a ninety-measure composition, the tonic
harmony is prolonged throughout. In that context the listener will focus
especially on the deployment of the tonic triad’s three pitches (A, C♯, and
E) and on how Mendelssohn connects those pitches. Of course the pitch A,
as the key’s tonic, will sound – either in the melody or in the bass, or in
both – when the greatest level of stability is desired. Though Mendelssohn
here limits himself to one harmony, he projects the tonic triad’s pitches in
multiple registers.

1.1 Mendelssohn: Song without Words in A Major

(op. 62, no. 6), mm. 83–90

As preparation for your exploration of audio example 1.1, sing arpeggia-

tions of A, C♯, and E in all of its inversional arrangements, outlined below.
Use a pitch pipe, an online pitch pipe, or a piano to establish the pitch A in
your mind.
3
4 Harmony in Mendelssohn and Schumann

Start by singing
A < C♯ < E

high in your vocal range. (Use any solfege system you are comfortable with;
or sing the letter names, substituting the single-syllable German “Cis” for
the multi-syllable “C-sharp.”) Then, starting lower, sing

E < A < C♯
followed by
C♯ < E < A
and finally

A < C♯ < E
low in your vocal range. You might also sing these pitches according to
their scale degrees numbers:

^1 < ^3 < ^5
^5 < ^1 < ^3
^3 < ^5 < ^1
^1 < ^3 ^5

For future reference, here are syllables for sharp and flat pitches (depart-
ing from standard German practice for B since Germans instead use H) for
readers who prefer to sing using note names:

Cis Dis Eis Fis Gis Ais Bis (raised)

Ces Des Es Fes Ges As Bes (lowered)

Example 1.1 The work’s concluding A Major tonic harmony, enlivened by

arpeggiations and local chromaticism.

•1. Indicate the scale degrees that sound at the five spots marked R.1
•2. Indicate the scale degrees that sound at the top of the three chords marked Z.2
••3. Though three distinct intervals occur at the spots marked W, they are all related.
Building upon your response to question 1, indicate which two scale degrees
sound in each instance, as well as what intervals those pitches form in each case.
Briefly explain how these intervals relate.3
••4. Among the eight pitches marked X or Y, three are chromatic. Which three?4
 Foundational diatonic processes 5

1.2 Mendelssohn: Song without Words in A Minor

(op. 53, no. 5), mm. 0|1–62

Over the past several centuries a large number of musicians have

helped in establishing procedures for analyzing tonal music. Not
surprisingly, there remain many areas of disagreement. Most analy-
tical observations depend upon some foundational premises (devel-
oped, one might hope, through careful study of compositions that
share specific stylistic traits). When practitioners who uphold con-
trasting premises study the same composition, alternative interpreta-
tions will likely emerge – sometimes even in coming to terms with
the most elementary of musical utterances. Consider the tonic triad
in A Minor: A-C-E. It may happen that the prolongation of the tonic
is temporarily called into question through stepwise upward motion
in two voices, resulting in the chord A-D-F. What if, immediately
thereafter, A-C-E is restored? Though figured bass numbers may
5
6
5
provide a neutral accounting of this motion, as 3
4
3 , a consideration
of the internal chord’s genesis and role would lead to considerable
controversy among analysts. Some would interpret the pitch D as a har-
monic chordal root, generator of the concurrently sounding F and A. From
that perspective, the passage might be analyzed as I IV I. Others would
interpret the pitch D instead as a melodic embellishment of the tonic’s third:
D as neighbor to C (and concurrently F as neighbor to E). From
that perspective the tonic root A does not yield to any other harmony
during A-D-F but instead perseveres despite the concurrent embellishment
of both the chordal third and fifth. A single analytical label might be
displayed:
I

perhaps fortified by Arabic numerals that provide a visual representation of

the embellishment

5





6





5






I
3





4





3

In this case the symbol 64 emerges through a linear initiative, not through
chordal inversion.
Likewise the A-C-E triad might be embellished by lower neighbors,
as in G♯-B-E, or by more complex combinations of neighbors, as in
G♯-B-D-E or G♯-B-D-F. Though it may not always be possible to
6 Harmony in Mendelssohn and Schumann

draw a precise line between melodic embellishment and harmonic

progression, in the perspective advocated in this study many
local chordal interactions of this sort will not be interpreted
harmonically.
Audio example 1.2 offers opportunities to come to terms with
local embellishing chords that are not harmonically motivated.
None of the questions posed ask for a Roman numeral analysis,
because that analysis would amount to no more than a prolonged
I. Though this excerpt offers more chordal variety than does audio
example 1.1, ultimately both in their own ways project a single
harmony, the tonic. (Whereas audio example 1.1 occurs at the end
of a composition, audio example 1.2 serves as an introduction.) To
prepare for what you will hear, sing in turn two models of
neighboring-note embellishment, beginning on a tonic A low in
your vocal range:

A < C < E
A < D < F
A < C < E
A < C < E
A < B < D
A < C < E

Then sing a combination of both neighboring-note embellishments:

A < C < E
A < D < F
A < B < D
A < C < E

Finally sing a model in which three neighboring notes emerge

concurrently:

A < C < E < A

A < B < D < G♯
A < C < E < A
 Foundational diatonic processes 7

Example 1.2 The composition’s introduction, bounded by A Minor tonic chords.

•1. An A Minor tonic chord sounds during seven of the fourteen R segments.
Which seven?5
•2. Downward stems are attached to fourteen of the noteheads in the grid above.
Exactly how many of them correspond to the bass pitch A?6
••3. Inner-voice quarter notes C and E (both members of the tonic triad) sound with
bass A at the downward-stem location within R1. In the span from R2 through
R13, only two other three-note combinations sound at downward-stem loca-
tions. Name those two combinations and indicate which locations correspond to
each. Describe how those combinations relate to the tonic triad.7
•••4. Throughout most of the excerpt, the upper-voice melody (corresponding to the
notes with upward stems in the grid) follows the contour of the inner-voice
pitches explored in question 3. Yet at three locations, upward motion among the
inner voices coordinates with downward motion above. Name those locations,
and describe the chord (the same in each instance) formed by the five sounding
pitches above the bass.8

1.3 Schumann: “Die alten bösen Lieder” from Dichterliebe

(op. 48, no. 16), mm. 15|16–193

Analyzing the individual chords that constitute a progression warrants a

careful consideration of their interactions with the chords that both pre-
cede and follow them. Many – perhaps all – chords commonly used in
tonal compositions are susceptible to a range of deployments. There may
be a tension between or among competing interpretations, leaving the
listener in a state of uncertainty until further progress through the work
provides, ex post facto, the context that helps clarify the roles the individual
chords play. A chord in 64 position offers special challenges for listeners, as
well as special opportunities for composers. When, for example, a B-E-G♯
chord sounds in an E Major context, the BE fourth formed with the bass
may in fact represent the inversion of the tonic triad’s EB fifth, in which case
8 Harmony in Mendelssohn and Schumann

the chord (in second inversion) may convey a tonic function. Yet that
sonority often functions in a quite different way, one in which B (not E) is
asserted as the root, while the E and G♯ above delay (as suspensions,
accented passing notes, or neighboring notes) the arrival of dominant
chord members D♯ and F♯, respectively. In a “left-to-right” hearing of a
passage, one may reach a spot for which multiple equally viable continua-
tions could be imagined. Only by taking what follows into account can one
develop a well-considered conviction regarding how the chord in question
might best be interpreted.
The vocal melody in audio example 1.3 makes a surprising descending
leap immediately after completing an ascending arpeggiation of E Major’s
tonic-triad pitches, from ^1 to ^8. To prepare for that encounter, sing
^1 < ^3 < ^5 < ^8

in E Major and then descend at random to ^1, or to ^3, or to ^5 . Repeat the

exercise in other major keys as well. Just before proceeding to audio
example 1.3, perform the exercise again in the key of E Major.

Example 1.3 An internal phrase that briefly tonicizes C♯ Minor’s mediant key, E Major.

•1. In order, name the four pitches that are emphasized at locations R1 through R4.
What is their relationship to the local tonic, E Major?9
•2. Within its context (as conveyed in question 1), what figured bass numbers would
correspond most aptly to what sounds at the downbeat of R3 (mentally filling out
the chordal texture appropriately)?10
•3. The chord at X reprises one of the four R chords. Which one?11
••4. Especially since it is preceded by the chord at X, the chord at S1 initially might be
interpreted as a restoration of one of the earlier R chords. Which one? If no more
than a restoration is being accomplished, what harmony would be projected at
that point? Through the reiteration of S1 at S2 and then the perhaps surprising
continuation at S3, another interpretation of S1 emerges. Discuss this context in
terms of which pitch is asserted as the chordal root, and how that assertion affects
the harmonic analysis.12
••5. Though the voice prolongs one pitch, the keyboard part begins a vigorous activity
at T. What pitches other than the tonic’s E, G♯, and B sound during this passage?
What is their relationship to tonic pitches? Given the harmonic activity prior to T,
would it seem more appropriate to deploy one or, instead, five Roman numerals
to account for what transpires during the excerpt’s final measure?13
 Foundational diatonic processes 9

1.4 Mendelssohn: Song without Words in F♯ Minor

(op. 30, no. 6), mm. 0|1–222

During a parallel period (one of music’s most straightforward organiza-

tional schemes), two phrases sound in succession, with only the latter
reaching the full closure generated by a perfect authentic cadence. In this
scenario the first phrase (the antecedent) will come across as an incomplete
musical utterance, thereby motivating the second try that will occur during
the phrase that commences immediately thereafter (the consequent). Often
the harmonic trajectory will be I–V followed by I–V–I (half cadence, then
perfect authentic cadence). (Though usually at least one other harmony
comes between I and V, in audio example 1.4 that is not the case. Also,
because the example is in the key of F♯ Minor and leading tone E♯ is
incorporated within the dominant, that chord’s analytical symbol will
appear as V♯.) The broad melodic trajectory supported by the period’s
harmonic progression often will take the form of ^3>^2, ^3>^2>^1. Because a
single harmony sometimes will span multiple measures, the melodic pre-
sentation of ^3 or ^2 might not coincide with the onset of its supporting
chord. For example, a gradual arpeggiation up to ^3 might be a salient
feature of an initial tonic harmony’s presentation.
Though a parallel period may serve as an independent musical entity,
such a construction often serves as a component of a broader musical
composition. The parallel period featured in audio example 1.4 constitutes
the first part of a song, here preceded by a six-measure introduction.
The dominant’s dissonant impact is here heightened through the incor-
poration of the chordal seventh and ninth. As preparation for your listen-
ing, sing the following alternation between tonic and dominant chordal
arpeggiations:
F♯ < A
C♯ < E♯ < G♯
F♯ < A
C♯ < E♯ < G♯ < B
F♯ < A
C♯ < E♯ < G♯ < B < D
C♯ < E♯ < G♯ < B
F♯ < A
10 Harmony in Mendelssohn and Schumann

Example 1.4 An introduction followed by a parallel period in F♯ Minor.

••1. The downbeat eighth note during R introduces the root of the F♯ Minor tonic
harmony. How are the three members of that harmony deployed during the
remainder of R?14
•2. As your answer to question 1 should indicate, the tonic harmony’s F♯ and A are
stated prominently during R. They are then repeated during the beats that
follow. What two pitches emerge during X1 (both in the melody and in the
accompaniment) to serve as their immediate successors?15
•3. Though the two passages are very similar, the bass note during X2 (within the
period’s antecedent phrase) is not the same as that during X1 (within the
introduction). Name the bass pitch deployed in each instance, and indicate
the term that often is used to describe how that of X1 functions.16
••4. The harmony introduced during X1 is intensified through the introduction of a
dissonant pitch at Y1 (repeated at Y2) and another at Z. Name, in order, these
two dissonant pitches and indicate what intervals they form against dominant
root C♯.17
••5. Provide a detailed account of how the melody during W1 relates to R.18
•6. Has the harmony changed at S, or instead has the preceding chord’s inversion
changed? Indicate both what harmony is projected at S (using a Roman
numeral) and that chord’s inversion (using conventional figured bass
numbers).19
•••7. The two measures that intervene between X1 and the onset of W1 do not have
counterparts between X2 and the onset of W2. Explain what Mendelssohn does
instead in the latter case.20
••8. Whereas W1 spans four measures, W2 spans five. Which one of W2’s measures
most deserves to be regarded as an addition to what was presented in W1? What
is the highest-sounding pitch during that measure, and what chromatic pitch is
deployed?21
••9. A suspension occurs at T, with resolution during X3. Using conventional figured
bass numbers (e.g., 9–8, 7–6, etc.), indicate the type of suspension that
Mendelssohn deploys. Which of X3’s three pitches serves as the resolution?22
••10. A I–V♯–I harmonic progression often supports the descending melodic line
^3 >^2 >^1 . Precisely where do that line’s A and G♯ sound during the antecedent
phrase? (Indicate locations within W1 and X2, respectively.) Likewise, where do
those pitches sound during the consequent phrase, preceding the tonic arrival
(supporting ^1 ) at U? (Indicate locations within W2 and X3, respectively, noting
that in this instance a descending registral shift occurs.) Given this state of
affairs, what cadence type concludes each phrase?23
 Foundational diatonic processes 11

1.5 Mendelssohn: Song without Words in E Minor

(op. 62, no. 3), mm. 0|1–82

From acoustics one learns that a sounding pitch C generates a range of partials:
C<C<G<C<E<G<. . . . That natural phenomenon helps explain why our musi-
cal forebears latched onto C-E-G as a foundational triadic entity. Once har-
mony became a sophisticated component of musical thinking, the trajectory
from the tonic to the dominant and back (C<G>C in C Major) became
ubiquitous, confirming the dominant root’s prominence among the tonic’s
partials. That root (^5 ), which of course generates its own set of partials
(G<G<D<G<B<D<. . .), typically will be enhanced by ^2 and ^7 . The leading
tone’s upward resolutional tendency (often strengthened by a concurrent 8–7
motion, G to F, resulting in the sounding of dissonant BF or BF ) helps explain why
the trajectory from V to I usually transpires directly, without intervening
harmonies. In contrast, the path from the tonic to the dominant often engages
at least one other harmony – II and IV being the most common choices,
especially in a major key.
This intervening chord may assist in harmonizing one of the founda-
tional melodic pitches. For example, the ubiquitous melodic descent from
^3 to ^2 during I to V might be supported in part by IV7 emerging against
prolonged soprano E, or by II supporting the onset of D. When the melody
unfolds pitches of the tonic and the dominant – for example, C<E followed
by D>B – II may be deployed as support for the melody’s D, with V delayed
until the B. Though soprano D and bass G do not actually sound together
in that context, one may regard the D of the D>B third as still in force once
bass G and the melody’s B arrive.
Recognizing basic pitch patterns that recur again and again in tonal
music is an important component of informed listening. In the case just
described the foundational CE to DG expands to become CE (I) through DF (II in
first inversion) to GB (dominant). Reinforce this conception by singing the
following arpeggiations first in C Major and then in other major keys:

C < E
G < D

C < E
F < D
G < B

Then shift into minor, lowering the tonic’s third by a half step but retaining
the leading tone:
12 Harmony in Mendelssohn and Schumann

C < E♭
G < D

C < E♭
F < D
G < B♮

Conclude by singing this pattern in the key of E Minor, as preparation for

audio example 1.5.
Example 1.5 A four-measure introduction followed by the first stanza’s first phrase in
E Minor.

•1. Which two pitches of E Minor’s tonic triad are featured melodically during Q?24
•2. Though in a higher register, the melody during R should come across as a
response to that of Q. What harmony is being arpeggiated? What three pitches
(in order) are deployed? Disregarding the registral contrast, what relationship is
formed by the highest sounding pitch of Q and the highest sounding pitch of R?
(Convey your answer as two scale degrees.)25
•3. Because Q and R are exactly equivalent rhythmically and both begin with an
upbeat (triplets), what occurs at S will come across as an upbeat onset of the
melody that appears over the next four measures. Yet harmonically it is closely
related to what has just occurred during R. Name the two pitches that sound at S
and indicate how that moment relates to what has preceded it.26
••4. The onset pitches of all seven of the melody’s locations marked as T are
members of the tonic triad. (The non-tonic context of T7 will be explored
below.) Which two consecutive T pitches might aptly be interpreted as a direct
outgrowth of the introduction’s Q material? Which other T pitch presents one of
Q’s notes in a higher register?27
•5. Compare the melodic embellishment that occurs between T2 and T3 with that
between T3 and T4. Name the pitches employed.28
•••6. Had the tonic harmony been maintained for a bit longer, the melody pitch at T7
could have served as the culmination of a three-measure ascending arpeggiation
of I. Yet concurrently with its arrival the harmony changes, converting T7 into
an embellishment of U1. How does T7 to U1 relate to content from the introduc-
tion? What bass pitch sounds against the melody’s two pitches? Ultimately what
harmony is projected during that beat? (Provide a Roman numeral and, if the
chord is inverted, figured bass numbers.)29
•7. If the melodic pitches of U1 and X had sounded where T7 and U1 occur (that is,
shifted one eighth note earlier), that beat’s harmony would have been enriched
by the addition of its dissonant seventh. Yet because T7 delays the arrival of U1,
X is pushed over to the next downbeat, thereby delaying the arrival of U2. This
situation is so common that chords like that at X have been given a special name.
What is that name? What type of cadence concludes the phrase?30
 Foundational diatonic processes 13

1.6 Schumann: “Die alten bösen Lieder” from Dichterliebe

(op. 48, no. 16), mm. 1–73

The notion of harmonic progression is closely allied with the melo-

dic filling-in of the span between the tonic and dominant roots in
the bass (usually ascending, as in C up to G in C Major, though
sometimes descending). When D is placed between C and G, some
manifestation of the supertonic likely will emerge. With F, the sub-
dominant and the supertonic (inverted) both are viable. In either
case the ascending fifth is segmented into a second plus a fourth, the
latter of which may be further filled in diatonically. For example, a
C<D<G bass leaves room for the supertonic to shift to its first
inversion prior to the arrival of V, while C<F<G likewise leaves
room for C to proceed to E for an inverted tonic prior to the
succession to IV or II.
Since these basic paths may be enhanced by further linear initia-
tives, such as adding a passing note within the third (resulting in
C<D<E or D<E<F) or adding a chromatic passing note within the
second (resulting in C<C♯<D or F<F♯<G), it is important to become
well acquainted with the more foundational trajectories first. To that
end, sing each of the following bass motions both in C Major and in
C Minor and in transpositions to other keys. (Sing in whole notes
except for the boxed pitches, which should be sung as half notes to
reinforce their interrelationship.) Sing in C♯ Minor just prior to
proceeding to audio example 1.6.
14 Harmony in Mendelssohn and Schumann

Example 1.6 A brief introduction followed by the first vocal phrase in C♯ Minor.

•1. Describe the four principal characteristics of the first chord (R): its soprano pitch,
its bass pitch, its quality, and the appropriate figured bass numbers.31
•2. The grid shows that the melodies during S1 and S2 are slightly at variance in terms
of rhythm. How do they differ in terms of pitch?32
••3. Whereas S1 is too sparse to convey any overt chordal contrast, the fourth and fifth
vocal pitches within S2 are supported by a chord that differs from what sounds
before and after. (Allow half a measure for each chord to fully form, disregarding
the neighboring-note embellishments.) What three pitch classes are members of
the perimeter chords? What five pitch classes (segmented into four plus one)
sound during the middle chord?33
•4. Compare the perimeter chord of S2 and the chord at T. (Provide both Roman
numerals and figured bass numbers for both chords.)34
•5. What pitch sounds in the vocal line at U? What two harmonies (Roman numerals
only) that precede the dominant might most likely be used to support that pitch in
a progression in C♯ Minor? Which one does Schumann deploy here?35
•6. Analyze the two chords that create the phrase-ending cadence (W and X). What
cadence type occurs?36

1.7 Schumann: “Ich will meine Seele tauchen” from

Dichterliebe (op. 48, no. 5), mm. 8|9–122

Contrasting the situation with I and IV, the I and II triads share no pitch
classes. Sing an arpeggiation of the succession I–II in the key of B Minor:

B < D < F♯
C♯ < E < G

The strong contrast between chords can be eased if the supertonic’s

seventh is incorporated. Sing

Of course, II7 introduces a dissonant seventh (B) that normally would

resolve downward by step. Fortunately that resolution pitch (here A♯) is a
member of the major dominant triad that generally follows the supertonic.
Sing
 Foundational diatonic processes 15

Example 1.7 The beginning of the binary form’s A2 section in B Minor.

••1. Which letter marks the location of a vocal pitch that serves as a chordal seventh?
Is this dissonance suitably prepared and resolved?37
•2. What is the relationship between the keyboard bass and the vocal melody in the
region from U to W? What harmony or harmonies are projected?38
••3. In coordination with other initiatives, the broad melodic trajectory during the
excerpt’s harmonic progression (which is stated twice during the example)
engages a foundational B<C♯<D melodic ascent against bass D>C♯>B. (The
initial bass pitch D is well prepared by the presentation of the dominant in its
4
2
position immediately preceding the onset of the example.) Which three letters
in the grid correspond to the onsets of melodic pitches B, C♯, and D, and which
three correspond to the onsets of bass pitches D, C♯, and B?39
•4. For good reason (as explored in question 1), the pitch A♯ sounds at the onset of
the V♯ harmony at T1. What occurs melodically and harmonically during the
passage from T1 to T2.40
 2 Surges

The very word chromatic implies a special coloration extending beyond a

key’s diatonic pitches. Often chromatic shifts emerge late in an individual
harmony’s prolongation, as the succession to its successor draws near. An
individual harmony’s particular evolution (over the course of two or more
chords) gives the listener a strong hint of what will come next. In a surge, a
harmony targets a specific successor through an appropriate utilization of
chromaticism and dissonance. A surge may occur only when the roots of two
successive harmonies form a descending perfect fifth (or ascending perfect
fourth). In this chapter we explore one straightforward surge configuration in
two of its most common locations within tonal space: a surging I➔ targeting
IV (as in C-E♮-G-B♭ in C Major or C Minor); and a surging II➔ targeting V
(as in D-F♯-A♮-C). The voice leading from the surging chord to its successor
will mimic that which transpires between V7 and I – including various
inversional alternatives such as a surging chord in 42 position resolving to a
chord in 63 position. (More intense surges and supersurges will be explored in
later chapters.) In the perspective proposed here, the surging chord should
not be interpreted as an actual dominant or labeled as V7. Instead it is a phase
of I or of II that takes on the same outer form as V7. Depending on its location
within tonal space and the mode, the chordal third and/or fifth may need to
be raised chromatically, while the seventh may need to be lowered chroma-
tically in order for the chord to precisely emulate V7.
It is important to maintain an awareness of where each root resides in
relation to the prevailing tonic during a surge. Such information is lost in the
common alternative method of analysis in which the surging tonic and
supertonic chords instead are labeled in the keys of the chords to which
they resolve (as is conveyed by the “applied dominant” symbols V7/IV and
V7/V). A juxtaposition of labels such as I followed by V7/IV conceals the fact
that chromaticism is at play, taming this potent shift analytically to become a
succession of diatonic chords in two related keys. In contrast, the surge labels
♭
that will be deployed here (for example, I8– 7 or I ➔) succeed in conveying
how a single harmonic function undergoes an internal evolution that makes
the succession to its successor seem inevitable. In practice, composers exer-
cised considerable freedom in this regard: I may proceed to IV without
16 surging, and I➔ may target IV without a prior diatonic, consonant I.
 Surges 17

2.1 Mendelssohn: Song without Words in E♭ Major

(op. 38, no. 1), mm. 1–5

Thinking in E♭ Major (the key of audio example 2.1), tonic E♭-G-B♭ will
surge toward the subdominant when D♭ is added to the triadic pitches.
Note in particular that the surging tonic’s third G and seventh D♭ form a
diminished fifth that resolves into the subdominant’s CA ♭ third (or, through
inversion, a D♭ to G augmented fourth that resolves into a AC ♭ minor sixth).
Pay special attention to this relationship (highlighted by boxes) when
singing the arpeggiations indicated below. Of course, an identical set of
relationships prevails when V7 resolves to I. Though not specially marked,
the last two chords of this exercise present the very model upon which a
surging tonic is based.

Example 2.1 The first phrase of composition in E♭ Major.

••1. The melody deploys five different pitches during the span from R1 through R4.
Name those pitches. How do they relate to the bass of measures 1 and 2? Also,
how does the melody during the R2 and R3 beats relate to the bass during those
same beats?1
•2. Omitting consideration of non-harmonic pitches and of inversion, what is the
principal difference in pitch content comparing R1 in measure 1 and R2 through
R4 in measure 2? Name the pitch classes that are engaged at these two locations
and provide a harmonic analysis that conveys their roles within the phrase.2
••3. Compare the melody during the three beats of S1 and S2 with that of the two beats
of R1 and R2. Also compare those three S beats with the concurrently sounding
bass pitches and with the melody during T2 through U.3
•4. A low E♭ sounds in the bass at the downbeat of each of the example’s measures.
That pitch conflicts with the harmony being projected during measure 4. (Thus
the E♭ should be interpreted as a pedal point.) Omitting consideration of E♭, what
four pitch classes during measure 4 join forces to project a harmony? Provide an
appropriate harmonic analysis for T1 through U. What type of cadence occurs?4
18 Harmony in Mendelssohn and Schumann

2.2 Schumann: “Es treibt mich hin” from Liederkreis

(op. 24, no. 2), mm. 66|67–74

Though the accidentals required will vary from one key to another (♮, ♯, or Ü
for a raised pitch; ♮, ♭, or º for a lowered pitch), the means of creating the
particular type of surge for I➔ featured in this chapter will always be the
same, depending on the key’s mode: in a major key, a chromatically lowered
minor seventh is added to a diatonic major triad; in a minor key, a diatonic
minor seventh is added to a triad whose third is chromatically raised
(converting it from minor to major quality). These adjustments result in a
chord containing a diminished fifth (or its inversion, the augmented fourth)
that resolves into the subdominant’s root and third. In the key of B Minor,
diatonic B-D-F♯ requires two adjustments to evolve into this form of I➔
surge: the chordal third D must shift to D♯, and seventh A must be added.
Sing the following arpeggiated chords, paying special attention to the
diminished fifth of I➔ and its resolution during IV. (These pitches are
displayed within boxes.)

Example 2.2 The coda of a song in B Minor.

•1. Schumann conveys most of the chords within the excerpt as a melodic anticipation at
the end of a measure followed by the full chord’s gradual emergence during the first
two eighth notes of the next measure. Instances of this process beginning at S1 and at
S2 project the tonic harmony. How do those two statements of the tonic relate?5
•2. The chords initiated at R1 and at R2 embellish the tonic chords S1 and S2. Which
four pitch classes are used in their formation? Indicate appropriate figured bass
numbers for both.6
•3. Compare the chords that emerge at S2 and at S3, indicating what two shifts
transpire in the pitch content, what general term is used to indicate this transfor-
mation, what harmony is being targeted, and how the two chords may be analyzed
using a Roman numeral and other symbols.7
•4. Provide a careful harmonic analysis for the passage from T through W. Indicate
what type of cadence occurs?8
 Surges 19

2.3 Mendelssohn: Song without Words in A Major

(op. 53, no. 6), mm. 106–113

Because the roots of I and IV span the interval of an ascending fourth

(or descending fifth), they relate exactly as do the roots of V and I and
thus are susceptible to connection via a surge. In a major key the tonic
chord’s major quality already makes the succession to IV a favored
option (engaging the half-step ascent from the tonic’s third to the
subdominant’s root). With the addition of a minor seventh, a dissonant
diminished fifth will resolve to the root and third of IV. When singing
the following arpeggiated chords, be particularly attentive to the inter-
actions among the boxed pitches, emphasizing the C♯<D and G♮>F♯
resolutions.

Another type of tension (followed by resolution) results from embel-

lishments such as passing notes and suspensions, a classic instance being
the cadential 64 resolving into a dominant 53 . Yet if the dominant will be
introduced in an inversion, the notes of a cadential 64 may occur in an
“inversion” as well (or, one may say that the notes have been unfurled
into 63 or 53 position while retaining their tendencies toward stepwise
downward resolution). That is the case in the relationships formed by
the boxed pitches in the following arpeggiated chords. It is as if E were the
low note of both the fourth and fifth chords, inducing 6>5 and 4>3
resolutions.
20 Harmony in Mendelssohn and Schumann

Certainly responses to tonal musical phenomena will vary among ana-

lysts. To some, this passage would contain a succession from IV to I (in first
inversion) prior to the dominant’s arrival. Yet many analysts would hear
IV proceeding instead to V, announced by the sounding of root E (in the
chordal interior) but not fully in place until accented passing note C♯
resolves to B and suspension A resolves to G♯.

Example 2.3 An eight-measure phrase in A Major at the onset of the coda.

•1. The tonic chord at the onset of R1 concludes the broader composition’s
main harmonic trajectory, sealed by a PAC in A Major. The coda here
begins with a tonic-reinforcing progression. In that context, what Roman
numerals pertain to the chords at S1 and at U1–U2? Given the excerpt’s
initial tonic and these two chords, what harmony would one expect to hear
between S1 and U1–U2? At what point does that harmony sound in its root
position?9
•2. Assess the transformation of the tonic between R1 and R3. Spell out the chords at
the endpoints of that trajectory (one with three pitch classes, one with four) and
propose a suitable harmonic analysis to convey the progression from these two
chords to S1.10
••3. The tonic prolongation during the excerpt’s first four measures reaches
its principal interior chord at R2, which targets the tonic restoration at R3
(explored in question 2). Name the four pitches that constitute R2, carefully
distinguishing between the pitches that congeal to form this chord and
the tonic pedal A that persists throughout the passage. Also indicate in
order the three pitches that come between R1’s initial bass A and R2’s
bass E.11
•4. Though the chord at S2 is closely related to its predecessor at S1, something has
changed. What?12
•••5. The notion of interval inversion sometimes creates unexpected compositional
opportunities. Whereas D<E would be a conventional bass trajectory during the
harmonic succession from IV to V, a D>E inversion may substitute. Yet if that
happens, the large interval of a seventh often would be broken up into smaller
components. First, name in order the three pitches that sound in the bass
between D and E in the vicinity of S1 through T3. Second, how might one
interpret the chord at T1 if bass E were sounding already by that point? Third,
can you hear the passage as a succession from IV to (embellished) V even if bass
E is delayed until T3?13
•••6. Your answer to question 5 should have revealed the deployment of chro-
maticism in the bass during the span from T1 to T2. Another chromatic
pitch emerges in that vicinity as well. Name that pitch and explain its
role.14
•7. How is the phrase’s concluding tonic chord embellished in the melody
at U1?15
 Surges 21

2.4 Schumann: “Ein Jüngling liebt ein Mädchen” from

Dichterliebe (op. 48, no. 11), mm. 4|5–82

The tonic triad’s root and third may be embellished by concurrent lower
and/or upper neighboring notes. Use the following series of arpeggiations
in E♭ Major to explore how this works.

E♭ < G < B♭
D < F < B♭
E♭ < G < B♭
F < A♭ < B♭
E♭ < G < B♭
D < F < A♭ < B♭
E♭ < G < B♭
Though the I and II triads have no pitches in common, the voice leading
between those harmonies often does achieve a sense of connection through
the deployment of the supertonic’s seventh, which is also the root of I. In
the arpeggiations below, note how at first the high E♭ is a stable chordal
element, but that upon the arrival of root F it attains a downward resolu-
tional tendency, fulfilled by the dominant’s third, D. (Focus especially on
the boxed E♭>D descending second.)

The same trajectory may be projected more intensively if the supertonic

surges: the raising of the supertonic’s third to A♮ creates an EA♭♮ diminished fifth
that resolves to the dominant’s DB♭ third. Two sets of arpeggiations are
provided below. In the first, the diatonic II7 evolves into II➔ (via the shift
from A♭ to A♮, displayed within boxes); in the second, the II➔ surge emerges at
the onset of the supertonic function, with no diatonic predecessor.
22 Harmony in Mendelssohn and Schumann

Example 2.4 The E♭ Major song’s first vocal phrase, which follows after an
introduction that traverses an identical chordal path.

•1. Hear how the chord during R2 restores the pitches that sounded at R1, resulting in
the prolongation of the tonic harmony into the excerpt’s third measure. At the
midpoint (S), common tone B♭ persists against three pitch classes that serve as
neighbors to tonic chord members. What neighboring note sounds in the low bass?
What other neighboring notes sound in the piano accompaniment (one doubled by
the voice)? What figured bass numbers correspond to the chord at S?16
•2. Considering only the melody, the measure labeled R1 projects a chordal skip
downwards from one tonic chord member to the next, whereas the measure of R2
and T pursues that agenda in the ascending direction. In order, name the tonic-
chord pitches involved.17
••3. The last of the four pitches described in question 2 is not supported by the tonic
harmony (as was the second). Instead, another chord in which it serves as a
member emerges during T. Is this chord consonant or dissonant? Is it diatonic
or chromatic? Is it surging? If so, what harmony is being targeted? Suggest both an
appropriate Roman numeral (in the key of E♭ Major) and figured bass numbers for
this chord. (As is always an option, here the chordal fifth is omitted. A suspension
delays the full sounding of the chord until the measure’s final eighth note.)18
•4. Spell and analyze the chord that sounds at U. Which pitch sounds in the bass, and
which pitch is sung?19

2.5 Schumann: “Aus alten Märchen” from Dichterliebe

(op. 48, no. 15), mm. 8|9–162

Though its wide range may be a challenge for some voices (in which case a few
notes may be imagined rather than sung), the following arpeggiation outlines
essential features of the excerpt by Schumann that will be explored below.
Note especially that the pitches forming the descending span ^5 >^4>^3>^2 in E
Major are displayed within boxes. Though one should not accent these notes
in performance, do consider how that basic line serves as a unifying force
within the excerpt, even if on occasion another chord member reaches higher.
 Surges 23

Example 2.5 The E Major song’s first vocal phrases, which follow after an introduction
that traverses the same chordal path.

••1. The excerpt may be broken down into five segments, as follows:

– the prolongation of the initial tonic, with emphasis on melodic pitch B (= ^

5 ).
– an interior melodic space supported by non-tonic harmonies, between the
initial melodic B’s prolongation and the arrival of G♯ (= ^3 ).
– the return of the tonic, now supporting G♯.
– the upward arpeggiation of an octave (G♯<B<E<G♯) in the piano, in part
matched by the vocalist, supported by the tonic harmony.
– the emergence of the phrase’s melodic goal F♯ (= ^ 2 ), eventually supported by
the dominant root B.
20
Indicate which of the letters in the grid correspond to each of these five segments.
B
•2. How is the tonic’s E fifth conveyed melodically during the span from R1 through
21
R3?
•3. Some analysts would label the chord at S1 as IV (despite the lack of a dominant
successor). Instead, develop an interpretation that regards it as a melodic
embellishment of the tonic. Supply an analytical symbol that uses both Roman
and Arabic numerals.22
•4. Provide a harmonic analysis for the four chords from R4 through U1, making
sure to allow each chord to fully form through the sounding of its bass note late
within each beat.23
•5. Though the tonic harmony prevails throughout the passage from U2 through
both U3 and U4 to goal U5, its inversion shifts several times. Indicate the
appropriate figured-bass numbers for each of these four locations.24
•••6. Though the foundational harmonic trajectory of chords U5 and X is among tonal
music’s most common, the intervening chords W1 through W3 are susceptible to
two highly contrasting analytical interpretations. The matter hinges on whether the
chord at W1 is interpreted as the onset of the harmony that comes to rest at X or
instead as an unfurled predecessor of the chord that sounds in a more normative
configuration at W2. (W1, W2, and X all deploy the same three pitch classes.) In such
cases it is useful to write down analytical notation for both hypotheses, so that their
merits and deficiencies may be thoroughly examined. To come to terms with the
first of two hypotheses for this passage, analyze the chords from W1 through X as a
foreground harmonic progression in the tonicized key of B Major. (Four Roman
numerals will be required.) Then shift to the hypothesis that the chord at W1 is
unstable, initiating a tendency that is repeated and more normatively conveyed by
the chord at W2, which resolves conventionally as embellishment of the chord at
W3. In this reading the chord at W3 would be regarded as the principal harmonic
link between U5 and X. (The chord at S3 functions in a manner similar to that at S1:
as an embellishment of the chords that surround it.) To come to terms with this
hypothesis, supply a harmonic analysis that deploys exactly three Roman numerals
(all in E Major) for the passage from U5 through X.25
 3 IV5–6 V

Whether accomplished vibrantly through a surge or more modestly as a

consonant diatonic succession, a progression’s passage from an initiating
tonic to the subdominant represents significant progress, in that IV is a
mere step away from the dominant that most progressions sooner or later
will attain. In this scenario the composer’s harmonic concern shifts from
generating the momentum to launch the progression on its ascending
root trajectory to a nuanced steering through a more delicate operation:
that of leading from IV to V, a chord that shares no pitch classes with IV.
Without intervention a chord rooted on ^4 instead might project a ten-
dency to proceed upwards another fourth, to the subtonic. (In C Major or
C Minor, a C ➔ F opening gesture could in fact proceed as F ➔ B♭ ➔ E♭.
Mediant E♭ might then serve as an intermediary between the initial
tonic and an upcoming dominant.) Generating effective harmonic pro-
gressions occasionally involves mitigating certain potentialities for a
given chord so that it may function effectively within a specific desired
trajectory.
Two factors often help steer IV toward V: stifling its surge potential
and introducing a common tone with V. This may be accomplished by
raising IV’s fifth up a step in a 5–6 shift. (In C Major, that would result by
either juxtaposing F-A-C and F-A-D or by adding D as a fourth pitch
class against a prolonged F-A-C.) If nothing further transpires before the
dominant sounds, then the added pitch may be regarded as a non-
harmonic intermediary between IV and V (in the manner of an anticipa-
tion) rather than as a supertonic root. In chapter 5 a more assertive
role for that pitch, warranting the deployment of a II numeral, will be
explored.

3.1 Schumann: “Es treibt mich hin” from Liederkreis

(op. 24, no. 2), mm. 59–66

Thinking hierarchically, some musical sonorities result from a collection

24 of passing and/or neighboring notes that, though they congeal into a
 IV5–6 V 25

chord, do not project a distinct harmonic function. In such cases supply-

ing a Roman numeral impedes comprehension, because the chord’s
context does not fulfill the implications of the analytical symbol.
Consider how a tonic harmony may be prolonged through the juxtaposi-
tion of its pitches in two different positions. Sing an arpeggiation of
B Minor’s tonic harmony first in its 63 position and then in its 64 position,
as follows:

D < F♯ < B
F♯ < B < D
Next add passing notes at the edges of the texture, diatonically filling in
the bass D<F♯ and soprano B<D thirds.

Next create a full-fledged passing chord to come between the two

presentations of the tonic.

As the tonic proceeds to its successor – here the subdominant – a

similar juxtaposition of pitches in two distinct inversions may tran-
spire. Sing the following arpeggiation of five chords representing the
harmonic succession from I (the first three chords) to IV (the last two
chords).

Finally make an added-sixth addition to the IV harmony and then

proceed to V♯, noting how the common tone C♯ serves both as IV’s
added sixth and as V♯’s fifth.
26 Harmony in Mendelssohn and Schumann

Example 3.1 The song’s final vocal phrase, in B Minor.

••1. Often one harmonic function will be conveyed through a set of interrelated
chords, some of which will project pitches from that harmony while others
serve as connectors or as embellishment. The trajectories of the exposed outer
voices from R1 through R2 and between T1 and T2 reveal two contrasting ways in
which such projections may be accomplished. First, affirm the tonic presence at
R1 and R2 by indicating the figured bass numbers that reveal how pitch classes B,
D, and F♯ are arranged at those two points. Second, indicate how the vocal
melody and the piano bass relate during this passage. Do they work in parallel
or in contrary motion? What vertical intervals are formed by these two lines at
those two locations? Then make a similar assessment of the passage from T1 to
the downbeat of T2 (at this point not considering the further development later in
the T2 measure). What figured bass numbers should be applied? What Roman
numeral? How do the outer voices relate?1
•2. One of the most fascinating (and sometimes challenging) aspects of harmonic
analysis is that the same chordal constructions may function in contrasting
manners, depending upon the tonal context. For example, the chord at R2 often
sounds in a context in which the bass note F♯ is asserted as the dominant root,
causing the B and D above to resolve downwards to the dominant’s third and
fifth. That is not the case at R2, where F♯-B-D instead projects the tonic harmony
in its second inversion. Likewise the chord at S might have played a significant
harmonic role if what follows next had in fact projected a dominant harmony.
Since that is not the case, the chord at S serves more humbly as a passing chord
connecting two tonic chords. Name the four pitch classes that sound at S, and
indicate the figured bass numbers that correspond to that chord.2
•3. From question 1 we come to understand that the parallel motion that has
prevailed between the vocal melody and the piano bass comes to an end through
the melody’s ascent against the bass’s descent between T1 and T2. Yet something
quite interesting happens during the latter measure: the vocal line leaps down-
wards to the pitch that would fulfill the parallel-motion trajectory. Name that
pitch, and explain its role within the chord in which it resides.3
•4. Assess the contrasting roles of the same pitch classes at R2 and at U1. Then analyze
the passage from U1 through W, indicating what cadence type is deployed.4
 IV5–6 V 27

3.2 Mendelssohn: Song without Words in G Minor

(op. 19, no. 6), mm. 30–34

When IV serves as the principal intermediary between I and V♯, both a

surge (between I and IV) and a 5–6 shift (between IV and V♯) may
transpire. Sing the following chordal arpeggiations as presented, and
then sing each line backwards (that is, as G>D>B♭, etc.). The subdomi-
nant’s 6-phase pitch, A, is displayed within a box.

Example 3.2 A phrase near the end of a composition in G Minor.

•1. Compare the chordal pitch content during R1 with that during R2 and also with
that during R4.5
•2. From the answer to question 1, it is apparent that a tonic function is asserted from
R1 through R4. Yet often during such expanses some embellishment provides a
contrast to the persistent sounding of the prolonged chord. What three non-tonic
pitches sound at S1 (and again at S2)? (Hint: all three pitches serve as neighbors to
tonic chord members.)6
•3. Over the course of R1 through R4 it becomes clear what chord Mendelssohn
intends for the location marked T1. How does the chord at T1 fulfill the resolu-
tional tendencies that precede it? Provide a harmonic analysis for the passage
from R1 through T1.7
•4. Any chord is susceptible to a transformation resulting in a surge targeting a
successor rooted a perfect fifth lower. How might the chord at T1 have been so
transformed? In this case Mendelssohn elects not to proceed in that direction
(which would launch a circle of fifths), but instead treats the chord at T1 as the
principal intermediary between I and V♯ within a progression in G Minor.
Comparing the chords at T1 and T2, explain what Mendelssohn does to assure
that the progression will lead straightforwardly to V♯. Provide suitable analytical
symbols for what happens from T1 to T2.8
••5. What two pitches sound on the beat at U1? What harmony emerges over the
course of that beat? What type of cadence is projected between U2 and W?9
28 Harmony in Mendelssohn and Schumann

3.3 Schumann: “Am leuchtenden Sommermorgen” from

Dichterliebe (op. 48, no. 12), mm. 22–61

The foundational state of a 5–6 shift model would proceed from a chord in
5 6
3 position to one in 3 position. Sing the following arpeggiated chords,

during which IV5–6 in B♭ Major conforms to that model.

Chordal unfurlings add variety to harmonic progressions. The unfurling

of IV’s 6-phase chord (such as E♭<G<C in B♭ Major) into 53 position
(C<E♭<G) is particularly attractive, since it both stabilizes the 6-phase
moment and promotes a more emphatic bass motion to the dominant
that usually will follow (here as C to F). Sing the following arpeggiations,
being especially attentive to the bass E♭<C and soprano B♭<C trajectories.

Significant intervals within a composition often will be filled in by

diatonic or chromatic passing notes. Connective chords may emerge at
such locations. Retaining an emphasis upon the E♭>C bass and B♭<C
soprano, sing the following arpeggiations, which include a connective
chord between IV’s 5-phase and unfurled 6-phase chords.
 IV5–6 V 29

Example 3.3 A composition in B♭ Major’s first vocal phrase.

•1. What four pitch classes are deployed during the beat marked Q to lead decisively
into the tonic harmony at R?10
•2. Does the tonic prolongation during the measure labeled R include the addition
of a minor seventh (A♭), inducing a surge toward IV?11
•3. What is the relationship between the chords that sound during the beats labeled
S1 and S2?12
•••4. The piano sounds a chord with bass E♭ and soprano B♭ at S1. First indicate the
pitches to which those lines move at S2. Then name the connective pitches that
sound in those voices at T. Finally, indicate which two other pitch classes sound
within the chord at T.13
•5. Provide a harmonic analysis of the phrase’s concluding chords (U and W),
indicating whether they fulfill the expectations established by the harmonic
trajectory from R through S2.14

3.4 Schumann: “Ich grolle nicht” from Dichterliebe

(op. 48, no. 7), mm. 1–42

An important factor in hearing how harmonic progressions work is to be

able to come to terms with various layers of embellishment or alteration
that modify a foundational progression. In the arpeggiations that follow,
such a foundational progression gradually is transformed through the
incorporation of multiple musical devices. Begin by singing the founda-
tional progression, as follows:
C < E < G
F < A < C
G < B < D
C < E < G
Now more gradually phase in the subdominant, with a descent in thirds
between the tonic and subdominant roots, as follows:

C < E < G
A < C < F
F < A < C
G < B < D
C < E < G
30 Harmony in Mendelssohn and Schumann

Next expand the subdominant through the addition of its 6 phase, as

follows:

C < E < G
A < C < F
F < A < C
F < A < C < D
G < B < D
C < E < G

Finally modify IV’s 6-phase chord through the incorporation of chro-

matic pitch A♭ (borrowed from C Minor), as follows:

Example 3.4 The opening phrase of a song in C Major.

•1. From bottom to top, what three pitches sound at the onset of Q? Which of the
intervals formed by those pitches is featured (filled in by a passing note) in
the melody during R?15
•2. What harmony emerges at S1? How does the piano bass during the span from S1
to S2 relate to the vocal melody that began at R?16
••3. The chord that sounds at S3 has the important task of supporting the word “Herz”
(heart), poetically regarded as the organ responsible for love. In this case that
heart is breaking (“bricht” at location U). Two separate musical devices are
deployed by Schumann to help convey that unfortunate state – one represented
by the soprano pitch at S3, the other by the fact that that pitch persists at T1
despite the altered harmonic context. Before assessing those distinctive features,
explain how the chord at S3 may be interpreted as representing a frequently
encountered phenomenon that comes between IV (at S1 and S2) and V7 (at T1
and T2). Then indicate what special and less frequently encountered devices
Schumann deploys at S3 and at T1 to heighten the emotional impact of the
passage.17
4 I5–6 II

Though not emphasized in the historical literature on harmony, it appears that

one of the salient features of harmonic practice by Mendelssohn and
Schumann is, as it were, the interchangeability of parts during the progression
from I to V. The ascending-fifth root trajectory’s division into a fourth
followed by a second or, instead, into a second followed by a fourth not only
involves the selection of different harmonies to intervene between the peri-
meter points, but also reveals how a range of related procedures may work in
the same way regardless of where within tonal space they are deployed. Thus
the same sort of surge that transpires between I and IV when IV is selected may
emerge between II and V when II is selected. Likewise the IV-to-V and I-to-II
trajectories bring on the same sorts of 5–6 shifts and corollary connective
chords. Consequently this chapter, which assesses how II may serve as the
principal intermediary between I and V, offers little in the way of new ideas,
but instead considers how what we explored in the context of IV in chapter 3
may be called into service when the fourth and second that together constitute
the ascending fifth from I to V are presented in the opposite order. It is very
important to keep in mind that this II is not derived from IV. Instead of
conceiving of the tonal system symmetrically, with the dominant a fifth above
and the subdominant a fifth below the tonic, the tonal system is here regarded
as unidirectional, with the principal harmonic path of tonal music being the
ascent of a fifth from the tonic root to the dominant root and then back again.
Along that path, II and IV are equally viable and distinct intermediary options.

4.1 Schumann: “Ich wandelte unter den Bäumen”

from Liederkreis (op. 24, no. 3), mm. 13|14–223

A phrase’s initial tonic harmony often is given a chance to settle in before the
harmonic progression gets underway in earnest. Embellishing chords that
feature neighboring notes of tonic chord members or passing notes that
connect two different tonic chord members may be deployed. Sing the follow-
ing arpeggiated chords, noting how the internal chords embellish the perimeter
tonic arpeggiations. Project bass A♯ as the lower neighbor to the B Major tonic
root and soprano C♯ as a passing note between the tonic chord’s B and D♯. 31
32 Harmony in Mendelssohn and Schumann

The following arpeggiations project the basic progression I II V in B

Major. That progression accommodates variable levels of dissonance. In
this case the tonic’s octave (B) is retained against supertonic root C♯,
resulting in II7. (This B resolves appropriately to A♯ at the onset of V.)
Likewise the supertonic third E is retained against dominant root F♯,
resulting in V7. One further dissonance emerges as well: whereas in con-
ventional voice leading the supertonic’s fifth (G♯) would descend to the
dominant’s root (F♯), in this case the voice leading does not succeed in
achieving that goal, descending only as far as G♮, the minor ninth of V♮ 79 .

It is very common for the succession from I to II to transpire as I5–6 II.

Incorporate I6 into your arpeggiations according to the grid provided
below.

Almost any whole step motion in tonal music might on occasion be filled
in melodically as two half steps. Sing the arpeggiations below, in which the
tonic’s 5–6 shift (F♯<G♯) transpires with an internal chromatic passing
note (F♯<FÜ<G♯).
 I5–6 II 33

Example 4.1 The second of a song in B Major’s four stanzas.

•1. What is the relationship between the chords at R1 and the onset of R2? How does
the chord at Q2 serve as a connector between those two points? How do the
chords at Q1 and Q2 relate?1
•2. What interval is traversed in the vocal melody from R2 through the onset of R3?
How does this melody relate to that which transpires from R1 through R2?2
••3. How is the tonic chord transformed between R2 and R3? What role does chro-
maticism play in this process?3
•4. All of the pitch classes that sound during the measure of R4–5 match those that
sound at the onset of R3 except for one: the last vocal pitch (doubled by the
piano), at R5. What is that pitch, and what dissonant interval does it form against
the initial G♯?4
•5. What harmonic labels correspond to the chords at S and at T? (Hint: allow the
chord during the measure labeled as T time to fully take shape. Five different
pitches sound in an ascending arpeggiation in the chordal interior.)5

4.2 Schumann: “Berg’ und Burgen schau’n herunter”

from Liederkreis (op. 24, no. 7), mm. 4|5–12

Though a neighboring note sometimes transpires quickly as a local embel-

lishment, often such notes congeal within a chordal formation. Such
chords may serve as embellishments of other chords, rather than asserting
a harmonic function. When singing the arpeggiated chords below, inter-
pret B and D as upper neighbors to the tonic root and third, respectively,
and later G♯ and B as lower neighbors.

Whereas the tonic harmony’s 5–6 shift (E to F♯ in A Major) often

facilitates the traversal of the harmonic succession from I to II (chords
whose triads share no common tones), the inclusion of the supertonic’s
seventh (A) adds another link between those two chords. Sing the
34 Harmony in Mendelssohn and Schumann

following arpeggiated chords, which convey the progression I5–6 II♯37 V,

noting the F♯ and A common tones between I6 and II♯37 .

The trajectory from I6 to II ♯7 may undergo evolution through the addi-

tion of pitch E (a passing note between F♯ and D♯) and the chromatic shift
of C♯ to C♮ (a borrowing of A Minor’s tonic third). Pay special attention to
these initiatives while singing the following arpeggiations:

Example 4.2 A song in A Major’s first phrase, which follows after a four-measure
introduction.

E
••1. What harmony is conveyed at R1, R2, and R3? Against a persistent A fifth (with
pedal function), other pitch classes take on chordal roles during the beats
marked Q1, Q2, and Q3. Carefully explain each chord’s role. (Explore the multi-
ple phases of Q2.)6
•2. What change in pitch content emerges in the vicinity of R4? How might this
change be conveyed using the analytical notation of harmonic analysis?7
•3. First, what harmonies sound at the phrase’s perimeters (R1 and T)? Given that
trajectory and the shift addressed in question 2, what chordal root would you
expect to find in the vicinity of location S? Is the chord that emerges there
consonant or dissonant? Diatonic or chromatic? Provide a suitable analytical
label for this chord in the context of A Major.8
•••4. Though Schumann might have proceeded directly from the chord at R4 to
that at S, in this case the former chord evolves in two distinct ways (at R5)
before its successor is introduced. Explain how the chord at R5 relates to
that at R4.9
 I5–6 II 35

4.3 Mendelssohn: Song without Words in D Major

(op. 85, no. 4), mm. 2–5

Composers often construct multifaceted harmonic progressions by

expanding upon foundational progressions in various ways. To establish
an orientation for one such expansion, sing the following arpeggiations,
which project a I5–6 II V harmonic trajectory.

D < F♯ < A
D < F♯ < B
E < G < B
A < C♯ < E < A
The second chord among the following arpeggiations unfurls I6 into its 53
position. The second and third chords develop the soprano melody into a
multi-chord arpeggiation of the tonic’s three members: A<D<F♯ (displayed
within boxes), supported by I5–6. That ascending initiative is then matched
by a descending initiative during the fourth through sixth chords: E>C♯>A
(likewise within boxes), supported by II V.

Further refinement results from connecting I5 and the unfurled I6 by

means of a connective chord (with passing notes C♯ and E). In addition, II
may be presented in its first inversion to bring the bass in close proximity to
the following dominant’s root. Sing the following arpeggiations, focusing
especially on these new initiatives:
36 Harmony in Mendelssohn and Schumann

Example 4.3 A composition in D Major’s first phrase, which follows after a one-
measure introduction.

•1. The phrase’s initial tonic harmony extends from R1 through R4. The outer voices
at two locations therein create an
D
F♯
 FD♯ voice exchange. Which locations? A
connective chord comes between those two points. What pitch classes constitute
that chord, and what is that chord’s figured bass?10
•2. The voice exchange addressed in question 1 projects a D<F♯ melodic third. Two
other intervallic unfoldings precede that third in the melody: one at R2 and
another at Q1. Name the pitches that participate in those unfoldings, and
comment on dissonance and its resolution among the three unfoldings.11
•••3. The three unfoldings addressed in question 2 together project a basic melodic
line from A through G to F♯ over the course of R1 through R4. An arpeggiation to
F♯ an octave higher then ensues. What pitches sound within that F♯<F♯ octave?
Analyze (not necessarily with Roman numerals) the chords that support those
pitches. Finally, the attainment of the arpeggiation’s goal F♯ coincides with a
change of harmony that destabilizes the F♯, resulting in an immediate descent to
E. Where does this occur, and what harmony transpires at that location?12
••4. An E>A melodic trajectory concludes the excerpt. List all nine pitches that appear
in the melody during locations S and T (through the end of the excerpt) and
indicate how each either serves as a member of the prevailing harmony (which
harmony?) or performs an embellishing role (what type of embellishment?).13

4.4 Mendelssohn: Song without Words in A♭ Major

(op. 38, no. 6), mm. 1–53

During a I IV V I progression certain initiatives may emerge in predictable

ways. I might surge as the succession to IV draws near, while a 5–6 shift
may transpire between IV and V. Sing the following set of arpeggiations, in
which those optional features are incorporated. Pay close attention to how
the 8>♭7 of I (A♭>G♭) leads downwards to the subdominant’s third (F) and
how IV’s 6-phase pitch (B♭) is retained as the dominant’s fifth. (These
pitches appear within boxes.)
 I5–6 II 37

Some chords work at a very local level, embellishing or connecting more

foundational chords. The second and fourth chords in the following set of
arpeggiations play such roles. Note especially how the emergence of a tonic
surge may be accomplished by moving each triadic pitch up a third by step:
that is, root to third (A♭<B♭<C), third to fifth (C<D♭<E♭), and fifth to
seventh (E♭<F<G♭).

To prepare for a more elaborate progression that follows, first sing some
foundational arpeggiations that project I5–6 II V, noting how the tonic’s 6-
phase chord may emerge when the arpeggiated tonic’s fifth, E♭, ascends to
F (as shown within boxes).

When a phrase proceeds from I to V, it is very common for the

foundational melodic trajectory to transpire as ^3>^2. That second often
will be embellished by means of local descents of a third, here C>A ♭
(supported by I5–6) followed by B♭>G (supported by II V). In the set of
arpeggiations that follows, both of those thirds are filled in by a passing
note. (Focus on the melodic trajectory of the boxed notes, observing
that the initial C is reinstated after the C>B♭>A♭ descent prior to the
shift to B♭.)
38 Harmony in Mendelssohn and Schumann

Example 4.4 The first phrase of a composition in A♭ Major.

••1. A melody often will weave back and forth between two individual strands. In this
excerpt’s melody, the pitches at R1 and at U2 not only participate in establishing
the tonic harmony but also serve as the starting points for two distinct strands that
persist throughout the excerpt. Name these two third-related pitches, and then
indicate what other third-related pairs of pitches continue these two strands at
locations W1–W2, R2–R3, and S–T2. (The pitches in question will not necessarily
sound at the onset of the region marked by the letter designation.)14
••2. The only pitch to sound in the bass during the span from P through R2 is A♭,
which serves as a pedal point. Consequently the harmonic progression that
transpires there will come across as working closer to the musical surface than
does the broad phrase-defining progression projected by S and T1–2. Provide a
harmonic analysis of that local progression, focusing on the chords at six
locations: R1, U2, W1, W2, X, and R2.15
••3. Though the chords at both Q1 and Q2 serve as connectors between two tonic
chords, their manners of embellishment are contrasting. Explain what occurs in
both cases, indicating which pitch classes form the connective chords and how
those pitches function.16
•4. The phrase’s foundational harmonic progression transpires after the tonic
prolongation of P through R2 concludes. How does the chord at R3 relate to
that at R2? What Roman numerals correspond to the harmonies of S and T2?17
•••5. The chord at location Q3 serves as a local connector between the chords at R2
and R3. What pitch classes sound within this connective chord? The chord on
the beat at location T1 often sounds near a cadence. Explain how it usually
functions, and comment on the less common turn of events near the end of the
T1 beat, just prior to the resolution to the chord at T2.18

4.5 Mendelssohn: Song without Words in A Major

(op. 102, no. 5), mm. 1–7

As an initial orientation, sing arpeggiations of the progression I5–6 II➔ V in

A Major, as follows:
 I5–6 II 39

A < E < A < C♯

A < F♯ < A < C♯
B < F♯ < A < D♯
E < G♯ < B < E
By presenting the tonic’s 6-phase chord in a 53 unfurling (as in A-C♯-E
to F♯-A-C♯-F♯), the bass attains some space (A>F♯) that may be filled in
by a passing note. In the set of arpeggiations that follows, that passing G♯
is complemented by chromatic passing note E♯, which connects 5-phase
E and 6-phase F♯. These arpeggiations also feature a further expansion of
the 6-phase chord through the addition of bass passing note E (thus
producing a chord in 42 position), which motivates the presentation of
6
II➔ in 5 position.

Example 4.5 A composition in A Major’s opening phrase.

•1. How do the tonic chords at the downbeats of R1 and R2 differ? How do those at
the downbeats of R1 and R3 differ?19
••2. Name the five bass pitches that sound during the span from R3 through T. (Each
is a half note, sounding for one full measure.)20
•3. What is the relationship between the tonic harmony (R1 through R3) and the
chord at S1?21
•••4. The sonority at X incorporates both a retained tonic pitch A in the alto register and
three other pitches that form a connective chord that targets the upcoming chord at
S1. (One of X’s pitches, in the tenor register, is chromatic.) Name those three
pitches.22
•5. What figured bass numbers pertain to the chord at S2?23
•6. What consonant, diatonic chord often would serve as the successor to the chord
at S1? In what two ways does the chord of T represent an evolved state of that
successor? Provide a suitable analytical symbol for the chord at T.24
•7. What figured bass numbers pertain to the chord at T?25
••8. Explain the rationale for Mendelssohn’s deployment of the pitches D♯ and D♮ in
the vicinity of U1 and U2.26
40 Harmony in Mendelssohn and Schumann

4.6 Mendelssohn: Song without Words in E Major

(op. 38, no. 3), mm. 7–14

As an initial orientation, arpeggiate the progression I5–6 II V (with unfurled

I6), as follows:

E < G♯ < B
C♯ < E < G♯
A < C♯ < F♯
B < D♯ < F♯
Now add a surging connective chord between I5 and I6.

Finally, arpeggiate the initial tonic harmony as a chord in 63 position, and

descend by step (parallel 63 chords) until the connective chord is reached, as
follows:
 I5–6 II 41

Example 4.6 The song in E Major’s first phrase, which follows after a six-measure
introduction.

••1. Four chords stand out as projectors of the excerpt’s foundational harmonic
progression. Analyze the progression represented by the chords at the following
locations: R3, U, W, and X1. Also comment on the melodic trajectory projected by
those chords (taking into account that a descending registral shift occurs).27
•2. Near the end of the introduction that precedes this excerpt, the pitches C♮, D♯, F♯,
and A sound in the range from one to two octaves above Middle C. Explain how the
soprano pitches that sound at locations R1, R2, and R3 relate to those earlier pitches.28
•3. A root-position tonic chord sounds at locations R1, R2, and R3, each with a different
soprano pitch. Which of those three chords sounds at R4? What happens in the
meantime, at S? Carefully describe the relationship between chords R3 and R5.29
••4. From question 1 we understand that the chord at U is closely related to the initial
tonic of R1 through R5. Often a connective chord will intervene between those
two more foundational chords. That connective chord might relate back to the
initial tonic, or it might instead target the succeeding chord. The second alter-
native prevails with the chords at T5 and U. Name the pitch classes employed in
both chords, and explain how the former may be said to “target” the latter.30
••5. In questions 3 and 4, the roles of chords R5 and T5 within the broader harmonic
trajectory have been assessed. Both of those chords sound in first inversion. What
voice-leading procedure does Mendelssohn deploy in the vicinity of locations T1
through T4 to connect those two more foundational points?31
5
•6. The harmony that concludes the phrase’s progression (at X1) is introduced in its 3
position, suitable for a cadence. (What type of cadence occurs?) Yet by the
measure’s end this chord has been fortified by two dissonant pitches that help
motivate the return of the tonic harmony for the phrase that begins just after this
excerpt. Those two pitches are featured in the melody at X2 and X3. Name the two
pitches, and indicate what intervals they form with the chordal root.32
 5 Surging 6-phase chords

Many thousands of musicians have created harmonic analyses over the

past several centuries. As numerous textbooks created during this period
attest, a wide range of analytical symbols – not necessarily the Roman
numerals that are deployed here – have been called into service. Certainly
disagreements are to be expected in this undertaking, especially since the
composers whose music we analyze generally did not give us much gui-
dance regarding how they thought about harmony.
The analytical perspective advocated here is notable for its sparse
deployment of Roman numerals. Perhaps already these procedures have
proven to be insight-enhancers for some readers, though others might be
resisting them. (Many of the chords that are described as “connective,”
“embellishing,” or “passing” would be awarded a harmonic label by most
modern analysts.) Likewise in a diatonic context a 5–6 shift here generally
will not trigger a new Roman numeral: adding an A to a C-E-G tonic chord
in C Major does not necessarily imply that the root has shifted from C to
A. Instead, the A often will be interpreted as a dissonant addition to the
stable tonic pitches, in the manner of an anticipation of the following II
chord’s fifth. Yet if certain other factors come into play, then the balance
between the maintenance of C as root and a shift to A will tip in favor of the
latter, in which case a two-tiered analytical symbol such as

C Major: I5−−−−−−−−−−−−−−−−6 II
(= VI➔)

will be employed. The factor that triggers the VI➔ interpretation is, of
course, the emergence of the chromatic pitch C♯. Chords that are spelled as
C♯-E-A or C♯-E-G-A and that resolve to D-F-A project a strong sense of a
descending-fifth root succession (A to D). Consequently the VI➔ label (or,
with greater precision, VI♯ or VI♯37 ) will be introduced within parentheses
below the Arabic 6 of the I5–6 label. Such surges of the tonic 6-phase chord
offer composers a wider range of intensity with which to work in formulat-
ing their harmonic trajectories. (Surges of even greater intensity will be
explored in chapters 6 and 7.) Likewise IV5–6 might evolve in a way that
42
 Surging 6-phase chords 43

would make the symbol II➔ (or II♯ or II♯37 ) appropriate. In this context II
emerges out of IV, contrasting the supertonic deployments explored in
earlier chapters.

5.1 Mendelssohn: Song without Words in B♭ Major

(op. 85, no. 6), mm. 14–171

First sing a foundational set of arpeggiations corresponding to the progres-

sion I5–6 II V I, as follows:

In the following set of arpeggiations, the tonic’s 6-phase chord (which

in this case incorporates 6 in addition to 5) evolves in two ways: first,
through an unfurling that places G in the bass; second, through the
chromatic shift of B♭ to B♮, thereby generating a surge targeting the II
chord that follows. The succession from II to V is facilitated by the
emergence of II’s first inversion, placing bass E♭ and F in a stepwise
relationship.
44 Harmony in Mendelssohn and Schumann

Example 5.1 The first stanza’s final phrase in a composition in B♭ Major.

•1. What bass pitch sounds on the downbeat of each of the excerpt’s measures: at R1,
S1, T1, and U? Though the harmonies are not necessarily fully in place at these
downbeats, what foundational harmonic progression is suggested by these bass
notes?1
•2. Each of the excerpt’s first three measures engages in a different sort of
internal initiative. Perhaps the most straightforward is that which tran-
spires between S1 and S2. How are those chords related? Also explain how
both F♯ and F♮ may occur in close proximity in the melody during this
measure.2
•3. The internal initiative between T1 and T2 likewise is frequently encountered. How
are those chords related?3
••4. In relation to the chordal root B♭, what Arabic numerals correspond to the
soprano pitches occurring on the beats at R1 and R2? What bass note sounds at
R2? What chromatic event occurs during the beat of R2? Drawing upon these
individual pitch initiatives, provide a coherent narrative concerning the harmo-
nic trajectory from R1 through S1.4
•5. What cadence type concludes the excerpt, at U?5

5.2 Mendelssohn: Song without Words in A♭ Major

(op. 53, no. 1), mm. 33–72

A 5–6 shift applied to a subdominant harmony may come about not

only through the literal ascent from a fifth to a sixth above the
subdominant root, but also by means of a descent from that root’s
octave to the sixth. The following set of arpeggiations emphasizes that
melodic trajectory.

The expanded set of arpeggiations below develops the progression in

two ways: first by moving from the tonic root position to first inversion,
 Surging 6-phase chords 45

with parallel tenths in the outer voices; and second by allowing IV’s 6-
phase chord to surge in its approach to V. The latter deploys chromatic D♮
to fill in a D♭<E♭ major second in the bass.

Further refinements in the progression result from the deployment of a

connective chord between the root-position and first-inversion tonic chords
at the onset (retaining the use of tenths between the outer voices) and
allowing the tonic to surge in its approach to the subdominant. In addition,
a passing note (C) fills in the descent from 8 to 6 during IV5–6. Sing the
following set of arpeggiations, keeping in mind how these various evolutions
and expansions relate to the more foundational progressions you sang
earlier.
46 Harmony in Mendelssohn and Schumann

Example 5.2 A composition in A♭ Major’s first phrase, which follows after a brief
introduction.

••1. First, compare the chords at locations R1 and R2. Second, assess the role of the
chord prolonged for a full measure at Q. Third, explain how the chord at R3
relates to what has preceded it.6
••2. Explain how the particular inversion of the chord at R2 affects the inversion of the
resolution chord at S1. If the chord during R2 is to resolve appropriately, what
relationship must prevail between the melody’s two pitches during S1? Is there
any precedent within the excerpt to help the listener in appropriately interpreting
that relationship?7
••3. Name the six melodic pitches that sound during S1 through S3. Which two of
those six pitches play the most foundational role? If that is so, then suggest a
harmonic label (with appendages to the right of the Roman numeral) that
conveys what transpires during these beats.8
••4. Though the chord that congeals during S3 is incomplete (with the pitch D♭ that
sounded during S1 and S2 absent), it is diatonic. At S4 the situation changes.
(Again one pitch, now B♭, is absent.) Explain the chordal evolution at this point.9
•5. As was the case also at R2, S1, and S3, the melody’s second pitch at T serves as the
chord member. What harmony emerges during T?10

5.3 Schumann: “Allnächtlich im Traume” from Dichterliebe

(op. 48, no. 14), mm. 0|1–111

As an initial orientation, sing a set of arpeggiations that corresponds to the

harmonic progression I IV7 V I, as follows:

B < F♯ < B < D♯

E < G♯ < B < D♯
F♯ < A♯ < C♯
B < D♯ < F♯ < B
Now expand these arpeggiations to include a I➔ IV surge, the addition
of IV’s 6 phase (asserted as II➔), and the filling-in of a C♯>B>A♯ third in
the melody between II➔ and V, as follows:
 Surging 6-phase chords 47

Finally, further develop the initial tonic by ascending gradually from B to

D♯ in the melody, using a procedure often referred to as reaching-over:
descend from B to A♯, supported by an embellishing chord; then, without
changing chords, allow C♯ to reach over A♯; then descend from C♯ to B for the
tonic’s restoration; finally, without changing chords, allow D♯ to reach over B.

Example 5.3 A song in B Major’s opening stanza.

•1. The tonic harmony’s root and fifth (B and F♯) are maintained as pedal points
throughout R1, Q, and R2. What pitches sound above those pedal points during
Q? What ascending melodic interval (a component of the tonic triad) is traversed
in the piano melody (for the most part doubled by the voice) during this
passage?11
•2. What pitch sounds in the melody at S1? What harmony supports that pitch? What
sort of chordal evolution occurs between S1 and S2? What descending melodic
interval is traversed between S1 and T, complementing the ascending interval
discussed in question 1?12
•3. Compare the chords at R2 and R3. How is the tendency introduced at R3 fulfilled
by the succession to the chord at U1?13
••4. Two initiatives transpire concurrently at U2 – one diatonic, one chromatic.
Explain how these initiatives suitably connect the chords at U1 and W1–2.14
••5. Explain how the melody pitch that sounds on the beat at W1 functions in the
context of the progression from U2 through W2. Especially consider issues
relating to register (where within the vocal range a pitch sounds), taking into
account that the text being projected at this point reads as follows: “und lautauf-
weinend stürz’ ich mich zu deinen süßen Füßen” (“and sobbing audibly I throw
myself upon your charming feet”).15
48 Harmony in Mendelssohn and Schumann

5.4 Mendelssohn: Song without Words in B Minor

(op. 30, no. 4), mm. 111–1171

First sing a set of arpeggiations that projects I IV V♯37 I in B Minor, as

follows:

B < D < F♯ < B

E < G < B < E
F♯ < A♯ < C♯ < E
B < D < F♯ < B < D
Now sing arpeggiations that introduce a surge: I ➔ IV. In a temporary
major-mode inflection, allow the subdominant to be projected as a chord
of major quality (E-G♯-B).

In the next set of arpeggiations, concurrently rescind IV’s major quality

(by lowering G♯ to G♮) and incorporate the 6-phase pitch C♯. Also embel-
lish the dominant by means of a cadential 64 chord.
 Surging 6-phase chords 49

Example 5.4 A phrase near the end of a composition in B Minor.

•1. Though several B major chords sound in the measures preceding this excerpt,
the cadence at location U confirms that B Minor prevails as the tonal center. In
that context, propose precise Roman numeral labels for the chords at R and at
the onset of S1.16
•••2. Name the five bass pitches that sound during the span from S1 through T1. One
of tonal music’s most intriguing features (and greatest analytical challenges)
concerns the variable ways in which individual pitches may function. Consider
the bass pitch at Q2. One analytical hypothesis (which will be countered by a
better one in question 3) would propose that Q2 serves as the onset of the
dominant. If that is so, then how might one convey the relationship between S1
and Q1 leading up to Q2? What pitch sounding at Q2 conflicts with a dominant
reading, and how might it behave if the dominant interpretation is to be
successful? How could the bass pitch at S2 be understood as conforming to
this dominant reading?17
•••3. Several factors argue against the hypothesis explored in question 2. An alter-
native hypothesis would propose that the bass pitch at Q2 serves as a passing
note: that the line ascending in half steps from S1 through S2 is an indivisible
unit, followed by the arrival of the dominant root at T1 (with the dominant
harmony fully falling into place during T2). Strong support for that reading
comes from a melodic line in the alto register that coordinates with the bass line
from S1 through S2. What pitches constitute that line, and how does it relate to
the bass? What four pitch classes (two of which sound at Q2 but then are
followed by rests) should be understood to constitute the chord at S2? How
does the chord at S2 relate to that at S1? Finally, how should the chord at Q1 be
interpreted in the context of this hypothesis?18
••4. In its context the downbeat melody pitch at T1 would generate the expectation of
a specific pitch at T2. Name those two pitches. Something vastly different occurs
on the downbeat at T2. What pitch sounds at that point? What is its role within
the harmonic context?19
 6 Surges with ninths

Because music is an art that deeply touches the emotions, a tonal compo-
ser’s palette of available chordal constructions offers many options. The
chords with intense potency generally will contain at least one chromatic
pitch or at least one dissonant pitch. Or, both of those characteristics may
prevail concurrently. Since by their very nature surges are among the key
emotive players in a harmonic progression, the level of potency conveyed
in their forward thrust may be adjusted with great precision. In earlier
chapters we saw how a minor triad may be chromatically adjusted through
the raising of its third and how a minor seventh may be added. Now we up
the ante: a minor or major ninth may intensify the surge’s drive toward its
successor.
The behavior of a chordal ninth differs to some extent from that of a
chordal seventh. Whereas the seventh, once integrated into a chord as a
member (rather than as a fleeting passing note), generally will prevail until
that chord resolves, the ninth may resolve within the chord: that is, 9–8
(a descending resolution by step within the chord) is an option unavailable
to 7. This gives rise to the interpretation of the ninth as a displacement
of an octave doubling of the root (8). In the context of V proceeding to I in
C Major, the dominant G-B-D may be fortified not only by seventh F, but
also by ninth A or A♭. When a ninth sounds, the chord might be presented
as B-D-F-A(♭) – that is, with A(♭) instead of, rather than in addition to, root
G. Yet G remains the unsounded root of B-D-F-A(♭). A bullet symbol (•)
will be deployed in the harmonic analyses to indicate such an absent root
(here as V(♭ )). Any surge may evolve so as to emulate this potently
7
9

dissonant dominant.
Thus during the succession I5–6 II in C Major the tonic 6-phase chord
(C-E-A) might transpire in a context in which the pitch A is asserted as
an intermediate root, fostering a local succession from VI to II. To surge,
VI➔ must sound with a raised third (A-C♯-E). The forward thrust is
enhanced when the minor seventh is added (A-C♯-E-G). Yet four new
possibilities emerge upon the addition of a ninth: that ninth might be

50
 Surges with ninths 51

major (A-C♯-E-G-B) or minor (A-C♯-E-G-B♭), and its addition might

coordinate with the omission of the chordal root (C♯-E-G-B or C♯-E-
G-B♭). Likewise during IV5–6 V, the surging subdominant 6-phase chord
(= II➔) might appear as D-F♯-A, D-F♯-A-C, D-F♯-A-C-E, D-F♯-A-C-E♭,
F♯-A-C-E, or F♯-A-C-E♭. (To confirm your understanding, also spell out
all the surge spellings for I➔ targeting IV.) Whereas the Arabic numer-
als, accidentals, and bullet symbol to the right of a Roman numeral are
capable of indicating which of these options is deployed with precision,
the convenient right-pointing arrow refers to this general class of chords
without distinguishing among the variants.
Though in many cases only one surging chord will sound, certainly it is
possible that multiple variants will be juxtaposed; for example, with the
9
7—–
gradual addition of the seventh and then the ninth to a major triad (II♯3—— ),

or with the softening of a chord’s impact through a 9–8 resolution within

9–8
the chord (II 7— ).
♯3—

6.1 Schumann: “Und wüssten’s die Blumen” from Dichterliebe

(op. 48, no. 8), mm. 34–37

Sing a set of arpeggiations in which the melody’s ^3>^2>^1 trajectory in A

Minor is embellished by an incomplete upper neighbor (D), supported by
the subdominant.

Now expand the phrase’s initial tonic through a C<D>C neighboring

motion in the melody. Lead from C to D via a chromatic C♯, with upward
propulsion generated by C♯’s dissonant interactions with G and with B♭,
and then proceed from D back to C, with downward propulsion generated
by D’s dissonant interaction with G♯.
52 Harmony in Mendelssohn and Schumann

Example 6.1 The final measures of a song in A Minor.

•1. The excerpt deploys only four bass pitch classes. That which sounds at R1 is
notated as a half note (maintained as a pedal point against the chordal progres-
sion in the upper voices), thus persisting until its reiteration at R5. The second
sounds briefly during U, the third in multiple registers at W1–2, and the fourth at
X. For now omitting consideration of what happens between R1 and R5, provide a
harmonic analysis for this foundational progression.1
••2. This excerpt concludes the song’s coda, which occurs after the singer has com-
pleted delivering the text. Two measures before the excerpt begins, what sounds
at R1 had joined with the singer’s final pitch (an A) to project a PAC’s concluding
minor tonic (in this case with omitted third and fifth). Both immediately after
that PAC and at this excerpt’s onset, a sudden build-up of tonal energy results
from the introduction of pitches that dissonate against root A or with one
another. Name the melodic pitch that sounds at the onset of locations R2, R3,
and R4, and then provide a complete list of the intervals that these pitches form
against the bass and with one another.2
••3. Provide a detailed harmonic analysis for the progression that transpires above
bass pedal point A over the span from R1 through R5.3

6.2 Mendelssohn: Song without Words in A Major

(op. 53, no. 6), mm. 11–18

When a broad melodic contour proceeds directly from ^3 to ^2 (leading even-

tually to ^1), II often will support ^2 initially. Though this ^2 may be maintained
during the following V, a local descent to ^7 is a frequently encountered
alternative. Sing the following set of arpeggiations, based on that trajectory.
 Surges with ninths 53

A wondrous though challenging aspect of tonal harmony is the notion that

a pitch might convey multiple meanings, with only the continuation following
a chord clarifying what structural role has been fulfilled. Consider the shift
from an A-C♯-E tonic to what might be spelled as A-C♯-E-G♮-B♭. Given that
spelling, we would regard the chord to be functioning as I➔, targeting IV. In
that case, major third C♯ would ascend to D, while minor seventh G♮ would
descend to F♯ and minor ninth B♭ would descend to A. Yet an alternative
spelling for the chord (used by Mendelssohn in the example to be explored
below) would be A-C♯-E-G♮-A♯. In that case the A serves as a pedal point in
conflict with the emergence of an F♯➔ chord (tonic A’s asserted and surging
6 phase) projected by its upward-resolving major third (A♯), a neutral fifth
(C♯), and downward-resolving minor seventh (E) and minor ninth (G♮).
Sing the following set of arpeggiations twice. First sing the first two
chords as notated but conceive of the second chord as being spelled using
a B♭ rather than an A♯. Thereafter improvise a continuation that pro-
ceeds to IV, followed by V and I. Then sing the set of arpeggiations as
notated, noting the vastly different outcome if B♭ is interpreted instead
as A♯. (A C♯-to-B suspension resolution transpires as well.)

In the set of arpeggiations that follows, a dominant (which one might

imagine was attained as the goal of the phrase preceding the excerpt)
lingers at the onset of the new phrase, preceding the A tonic’s arrival.
This dominant is embellished by two upper neighboring notes: A embel-
lishing the dominant’s third, G♯, in the bass; and F♯ embellishing the
dominant’s root, E, in the soprano.
54 Harmony in Mendelssohn and Schumann

Example 6.2 A composition in A Major’s second phrase: a consequent phrase

cadencing on I.

•1. Preceding this excerpt an antecedent phrase cadences on an E-G♯-B dominant

harmony during its eighth and final measure. In a modest blurring of the
boundary between phrases (but with Q1 clearly positioned as the first of eight
consequent-phrase measures), that dominant persists at the consequent phrase’s
onset. Where does tonic A-C♯-E arrive? What are the appropriate figured bass
numbers for the dominant that precedes that tonic at the phrase’s onset?4
•••2. Though listeners will expect the G♯ leading tone to resolve to tonic A, the pitch A
that sounds at Q2 should not be interpreted as that resolution. Instead, V7 persists
through Q3, with the chord at Q2 embellishing – rather than resolving – the
dominant. Spell out the pitches that constitute the dominant harmony at Q1 and
Q3 (omitting the persistent local neighboring embellishment provided by A♯), and
then assess what happens at Q2. (Hint: only the outer voices shift at that point.)5
••3. The consonant tonic chord that finally arrives at R1 is prolonged through R2. Some
analysts would interpret the intervening chords at S1 and S2 as local embellishment
of the tonic, whereas others would deploy a V label, proposing a local I V I harmonic
progression as a means of projecting the phrase’s initial tonic. Again not considering
the persistent neighboring embellishment (now D♯), what two pitches sound in
addition to triadic E-G♯-B during S1–2, and what roles do they perform?6
••4. For the moment omitting consideration of the chord at R3, what harmonic pro-
gression concludes the excerpt (from T1 through W)? Note that a chromatic shift
occurs at T2. After spelling out the chords of T1 and T2 (again omitting considera-
tion of the persistent neighboring embellishment, now C♯), assess how this shift
participates in a broader voice-leading initiative that transpires in the alto register.7
••5. The harmonic trajectory from the chord at R2 to the chord at T1 encourages a
specific connective voice-leading operation. In general, what pitch might be
added to the three chord members of R2 in that process? Now spell out that
proposed chord, arranging its four pitch classes in a stacked-thirds formation.
Next raise one of those pitches so as to create a surge, targeting the chord at T1.
Finally, assess how the chord during R3 extends beyond that level of intensity.8
 Surges with ninths 55

6.3 Schumann: “Die Rose, die Lilie, die Taube, die Sonne”
from Dichterliebe (op. 48, no. 3), mm. 0|1–42

When singing the set of arpeggiations below, note how the I II V harmonic
progression deploys II in its first inversion, so that the bass pitches of II and
V form a stepwise relationship (G<A).

Because the roots of II and V are fifth-related, the supertonic may surge
in its approach to the dominant, as is conveyed in the following set of
arpeggiations. In this case that surge is very potent, since both a dimin-
ished fifth ( DG ♯ ) and a minor seventh ( FG♯♯ ) push toward inward resolution
C♯
(to V’s A third and E
A fifth, respectively).

Composers have many options for expanding a progression’s initial

tonic harmony. In the set of arpeggiations below, the common voice-
8
7
6
5
leading pattern 3


4
3 is deployed. (In your performance, focus on how

the melody’s C♯ and B serve as passing notes and G serves as a neighboring

note.)
56 Harmony in Mendelssohn and Schumann

Example 6.3 A song in D Major’s first phrase.

••1. The passage from R1 through R5 is unified by a text that relates four things (rose,
lily, dove, sun) that the protagonist once loved, but that now pale in comparison
to the person described in the phrase that ensues immediately after this example.
The chord at Q twice serves to loop the progression back to the beginning again,
with success in completing the opening trajectory only on the third try, at R5.
Compare the melodic interval formed by the vocal pitches at R1 and R2 with the
melody’s continuation from R2 through R3 and R4 to R5. Then propose a rationale
for the chordal progression that transpires at those locations.9
•2. The structural relationships discussed in question 1 argue in favor of regarding
the chord at R5 as a restoration of the tonic from R2. (Some analysts instead might
propose that the chord at R4 extends to the chord at S1.) What three bass pitches
sound from S1 through T? Which two are diatonic? Assuming a three-measure
tonic prolongation at the phrase’s onset, what Roman numerals correspond to
those diatonic chords?10
••3. Keeping in mind that chordal sevenths and ninths often emerge as passing notes,
consider the relationship between the consonant, diatonic chord at S1 and the
dissonant, chromatic chord at S2. Begin by spelling the chord at S1 as a root-
position triad. (In Schumann’s realization, that chord sounds in its first inversion
at S1.) What happens to the triadic third (sounding in the bass) at S2? What two
pitch classes sound above that third? What has happened to the chordal root?
Finally, suggest analytical symbols for the chords at S1 and S2.11

6.4 Mendelssohn: Song without Words in C Minor

(op. 38, no. 2), mm. 8b|9–162

Sing the following set of arpeggiations, which conveys how I II V 6
5

4
3 may
^
support the stepwise descent from 5 to 2.^

Now sing the expanded set of arpeggiations below, in which the

tonic’s 6-phase chord (E♭-G-C) is represented by surging E♮-G-B♭-D♭
and the II-to-V succession is enhanced by the addition of II’s minor
seventh (thus II8–7).
 Surges with ninths 57

Now let’s explore how one might prolong the initial tonic. One of the
most fruitful ways to proceed is projected in the following set of arpeggia-
tions, in which an embellishing 64 chord transpires.

That embellishment may be given a higher degree of stability by

an unfurling into 53 position, as is conveyed in the following set of
arpeggiations.

E♭ < G < B♭ < E♭

A♭ < C < E♭
E♭ < G < B♭ < E♭
The additional layer of arpeggiation conveyed in the following set of
arpeggiations is potentially confusing. The added long horizontal lines
are intended to help guard against relating the added embellishing
chord’s G, B♭, and E♭ to the initial or concluding tonic chord. In this
context those pitches embellish the 64 embellishment; they do not restore
the initial tonic.
58 Harmony in Mendelssohn and Schumann

Example 6.4 The first phrase of an internal section, in tonicized E♭ Major, from a
composition in C Minor.

••1. How do the chords labeled as Q1 or Q2 relate to the chords labeled as R1 or R2?12
••2. How does the chord at S relate to the Q1 and Q2 chords that surround it?13
•3. Four melodic pitches sound over the course of R2 and R3. Name those pitches.
What interval is formed by the first and last of those pitches? How do the first
three melodic pitches during T1 relate to that preceding outlined interval?14
••4. What harmonic course is pursued over the span from R2 through T1?15
•5. Compare the chordal pitch content of T1 and T2. Which chord is consonant?
Which is dissonant? How might one display what transpires using Roman and
Arabic numerals?16

6.5 Mendelssohn: Song without Words in E Major

(op. 67, no. 6), mm. 91–99

When singing the following set of arpeggiations, focus especially on the

two trajectories that lead away from the tonic in the bass: either an
ascending second (E<F♯) or a descending fifth (E>A).

The ascending second (E<F♯) often is enhanced by a 5–6 shift whose

6-phase chord (E-G♯-C♯) may be asserted as VI➔, surging in the evolved
state (E♯-G♯-B-D♮). That evolution leads to the chromatic bass E<E♯<F♯.
The descending fifth (E>A) offers another surge opportunity, during
which a diminished fifth or augmented fourth will target resolution on
the following chord’s root and third (here the surging tonic’s DG ♯ to the
subdominant’s CA ♯ ). Note the occurrence of both of these initiatives in the
following set of arpeggiations.
 Surges with ninths 59

Example 6.5 A phrase near the end of a composition in E Major.

•1. Keeping in mind that an E pedal point persists at the bottom of the texture
throughout the example, what is the relationship among the following three
chords: R1, R3, and S2?17
••2. Whereas the chord at S2 is a frequently encountered option following I, the chord
of S1 is equally viable. (Note that it contains a chromatic pitch.) How does that
chord function within its local E Major context? What chord emerges at R2 to
effectively lead from I to the chord at S1?18
•3. Whether preceded by S1 or by S2, the harmony represented by T1 and T2 is a vital
component of any harmony progression. Indicate the pitch content of the
T chords and comment on the immediate successors, R3 and U.19
 7 Colorful variants of II

Though we already have encountered a range of supertonic variants

(including surges with or without ninths), two additional categories of
supertonic were frequently deployed by Mendelssohn, Schumann, and
their contemporaries. In the first, a surging chord’s fifth will conform to
the minor- rather than the major-mode key signature. Thus, whereas in C
Major II➔ might sound as
D F♯ A C, we encounter instead D F♯ A♭ C.

Applying this conversion to other common surge configurations,

F♯ A C becomes F♯ A♭ C and
F♯ A C E♭ becomes F♯ A♭ C E♭.

Though the A♭ places these chords closer to the realm of C Minor than to C
Major, in practice they may be deployed in either mode. Conventionally
these entities have been referred to as augmented sixth chords (even
though foundationally a diminished third – AF ♯♭ – prevails in their formula-
tion) and have been given quaint nicknames – French, Italian, and
German – that lack the precision of Roman numerals in positioning chords
within tonal space. Here the II numeral will be retained for our analyses,
sporting either an array of Arabic numerals with accidental(s) – as in II 7♯
for D-F♯-A♭-C in C Minor – or an outline arrow – as in II⇨ – designating a
configuration that will be referred to as a supersurge.
The second colorful supertonic variant featured in this chapter is spelled
using the lowered second scale degree, for example D♭-F-A♭ in C Major or
C Minor (often called the Neapolitan chord or the Phrygian II). In a minor
key the diatonic supertonic is of diminished quality (D-F-A♭) and thus is
not suitable as the goal of a surge (easily generated via I5–6: C-E♭-G to C-E♭-
A♭, to which a G♭ might be applied to generate VI➔). Given that context
(or even without it), the lowering of D to D♭ is a surprisingly common
event in music of the nineteenth century. Some typical deployments will be
explored below.

60
 Colorful variants of II 61

7.1 Mendelssohn: Song without Words in A Minor

(op. 38, no. 5), mm. 0|1–4

Given the locations of the tonic and dominant chords’ pitch classes
within tonal space, it is very common for 4–3 or 6
5 4
3 suspensions to

transpire against the arrival of the dominant root. Both are deployed in
turn in the excerpt by Mendelssohn to be explored below. The following
sets of arpeggiations for singing convey these two contexts. (Given the wide
range, it may be necessary to imagine rather than sing the low E.)

The first of the progressions above may be abridged through the omis-
sion of the suspension’s preparation phase. Sing the following set of
arpeggiations, in which soprano A arrives not against the tonic root, but
instead as the dominant already is taking shape.

A < C < E
E < B♮ < D < E < A
E < B♮ < D < E < G♯

In the reworking of the second set of arpeggiations below, the second

chord’s A is supported not by the tonic harmony but instead by pitch
classes F, C, and D♯. Whereas D♯-F♯-A-C (explored in chapter 6) would be
analyzed as a surging II➔, the use to the minor mode’s diatonic ^6 (F♮) gives
greater potency to the chord, to the extent that it here is awarded a different
analytical name and symbol: supersurge and II⇨. (Many analysts tem-
porarily abandon the tonal focus provided by Roman numerals, referring
to this chord as a German augmented sixth, with symbol Ger+6. That
practice will not be followed here. My use of II ultimately derives from
an eighteenth-century precedent, displayed in an example labeled as 1.4 in
62 Harmony in Mendelssohn and Schumann

my Thinking About Harmony.) As you sing, pay special attention to how

the supersurge’s potent DF ♯ dissonance resolves, noting that although that
resolution highlights the dominant’s root (E – a half step below F and a half
step above D♯), that harmony’s other member here do not arrive until the
following chord.

Example 7.1 The composition in A Minor’s opening measures.

•1. The chord at S1 embellishes the tonic harmony stated at R1 and R2. What pitches
serve as chord members of S1?1
•2. One of the pitches sounding on the beat at T1 conflicts with the presentation
7
of V♯ 3 . Name that pitch, and indicate when and to what pitch it resolves.2
••3. Compare the chords at T1 and T3. What name is often applied to the latter?
Comment on the resolution of its non-dominant pitches.3
•••4. Regard the prevailing harmony during R3 to be I. What two embellishing pitches
sound in the melody during that beat, and what sort of embellishing role do they
fulfill? Compare that context with what happens at location U. What harmony is
projected at that point, and in what evolved form? Carefully explain the surpris-
ing relationship between R2 and U.4

7.2 Schumann: “Am leuchtenden Sommermorgen”

from Dichterliebe (op. 48, no. 12), mm. 16–201

Whereas conventional harmonic analysis relegates supersurging chords to

a category devoid of scale-degree orientation within a key (deploying
instead a set of nicknames – German, French, and Italian), here such
chords are deemed to serve as flamboyant evolved states of normative
harmonic entities. Regardless of what configuration II may take, the
progression I5–6 II V is among the most utilized in tonal music. The
supertonic attains a special fascination because it may transpire as II, as
II➔, or as II⇨.
 Colorful variants of II 63

As you sing the set of arpeggiations below, note especially how the
perimeter pitches of the third chord (which functions as II⇨) form a
compound perfect fifth ( DG ♭♭ ). A simultaneous direct descent by half step
to the root and fifth of V ( CF ) does occur occasionally, though more often a
cadential 64 would be deployed to prevent parallel fifths. In this case another
strategy of fifths-prevention is called into service: a descent from D♭ to C
during the supersurging supertonic’s prolongation. (In that D♭ is the
chordal ninth, it may freely descend to C within the chord, a privilege
not available to the seventh, B♭.)

The expansion to the set of arpeggiations below offers a surprise:

whereas B♭ Major’s diatonic 6-phase chord is spelled as G-B♭-D, that
chord’s quality here temporarily shifts to major (G-B♮-D). (The pitch B♮
may be referred to as a wobble.) Note that eventually B♭ is restored, prior to
the onset of II⇨.
Also of interest is the potent connective chord that comes between the B♭
and G chords. Observe how two concurrent linear trajectories (B♭>A♭>G
and F<F♯<G) produce an augmented-sixth interval in a context that does
not match any of the conventional types.
64 Harmony in Mendelssohn and Schumann

Example 7.2 A progression that extends from the end of a phrase cadencing on I in B♭
Major through to V.

•1. The tonic chord that inaugurates this example is a restatement (later in the song) of
the tonic chord that concludes audio example 3.3. After that earlier passage, the
progression proceeds immediately to a chord like that at location U1. Here, instead,
the chord of S1 emerges. What is the relationship between the chords at R and at S1?
What special feature of the latter makes it a magical moment within the song?5
•2. The chord at location T may be interpreted as an embellishing chord, which one might
presume would come between two statements of the chord introduced at S1. What
pitches are deployed in this embellishing chord? What surprise awaits listeners at S2?6
••3. First assess where within tonal space the progression has landed at location
W. Then, also considering what has transpired between R and S2, speculate
regarding what sort of chord might be expected at location U1. Finally, pinpoint
exactly which pitch classes Schumann deploys for that chord and suggest a
suitable analytical symbol.7
••4. An individual harmonic function may be projected through a succession of
multiple chords that subtly shift over time. A common example is the surge,
wherein a consonant and diatonic chord may become dissonant and chromatic as
the approach of its successor draws near. Though less often encountered, a
reduction of dissonant intensity may occur as well. The most common instance
of this phenomenon is the resolution of a chordal ninth within the chord (as 9–8).
Explain how this notion pertains to Schumann’s writing in the vicinity of U1–2.8

7.3 Mendelssohn: Song without Words in C Minor

(op. 38, no. 2), mm. 56–592

As you sing the set of arpeggiations below, note especially how the succes-
sion from IV to V♮ in C Minor is navigated by means of a 5–6 shift (C to D).

Note two refinements in the set of arpeggiations below. First, the tonic
surges in its approach to IV. Second, IV’s 6-phase chord is chromatically
altered: F-A♭-D♭ rather than F-A♭-D. That shift (a wobble) is short-lived:
the emerging dominant restores D♮.
 Colorful variants of II 65

In the next set of arpeggiations, passing chords have been incorporated

within the initial tonic prolongation and during IV5–6. Note how the
diatonic pitch E♭ within the melody’s F>E♭>D♭ third (in the vicinity of
IV5–6) – rather than E♮, which would correspond to the B♮ of the preceding
tonic prolongation – facilitates D♭’s arrival.

In the final set of arpeggiations, below, the cadential 64 chord is treated to

an embellishing chord that incorporates the upper neighbors of all three
chord members (against a prolonged G in the bass).
66 Harmony in Mendelssohn and Schumann

Example 7.3 A passage near the end of a composition in C Minor.

•1. How do the chords at R1, R2, and R3 relate?9

••2. How do the chords at S1 and S2 relate? What non-diatonic pitch does
Mendelssohn employ during S2?10
••3. How does the role of the chord at T2 within its context differ from that of chord T1
within its context?11
••4. The chord at U1 calls for a specific successor. Does the chord at U2 or, instead, the
chord at U4 fulfill that expectation? What role does the other of those two chords
perform?12
•5. Atypically this example has been cut off just before the phrase’s cadence for
inclusion here, since Mendelssohn concurrently begins a new initiative in the bass
at that moment. Without such an initiative, what chord would one expect to hear
on the downbeat after U4? What pitch would reside in the soprano, and what
pitch in the bass?13

7.4 Mendelssohn: Song without Words in E♭ Major

(op. 38, no. 1), mm. 25–28

In the set of arpeggiations below, the melody’s B♭ (which sounds at the onset)
is prolonged through the end of a tonic expansion that encompasses a
motion through II and V♯ back to I, prior to a descent to A at the cadential
dominant. That prolongation is achieved by means of a filling-in of the
tonic’s B♭>G third, followed by a restoration of the B♭ (as marked by boxes).
Focus attention on that foundational structure as you sing the arpeggiations,
noting that the internal A is itself prolonged via its own local descent:
A>G>F♯. (In this way the A extends its reach from II into the domain of V♯.)
 Colorful variants of II 67

Though the pitch A is an essential component in the stepwise filling-in of a

B♭>G third, it may be lowered temporarily to A♭ in what might be described as
a wobble. Diatonic II (C-E♭-A) may be replaced by ♭II (C-E♭-A♭), with A♮
restored upon the succession to V♯37 (D-F♯-A-C). (Even if the dominant’s
fifth is omitted, the listener will understand that A♭ has reverted to A♮.)
Though A♭ sounds instead of A♮ in the melodic line of the arpeggiations
below, ultimately A♮ should be understood to prevail (by the time of the
C<D<F♯ arpeggiation), prior to that line’s continuation to G.
The set of arpeggiations below also projects a harmonic collision near
its end. Whereas the G
B♭
 BG♭ voice exchange may be regarded as a tonic
expansion, the internal pitches C♯ and E♮ work with B♭ and G to
project II➔.

The next set of arpeggiations offers one further refinement – the expan-
sion of the initial tonic by means of an embellishing chord. As you sing the
F♯
arpeggiations, feel the tension of the C augmented fourth resolving into
G
the restored tonic’s B♭ sixth.
68 Harmony in Mendelssohn and Schumann

Example 7.4 An internal phrase in tonicized G Minor from a composition in E♭ Major.

••1. Name the three pitches that are engaged in a stepwise melodic descent during R1.
Then name the three pitches that are engaged in a stepwise melodic descent from
S2 through U2. The perimeter pitches of the first third both belong to the
harmony that is projected at R1. What is that harmony? In contrast, the perimeter
pitches of the second third belong to no harmony in the vicinity. That is because
one of those two pitches is a wobbly note and will be understood to revert to its
diatonic state during U2. (Though that pitch is not present in the melody, it
sounds an octave lower.) Which pitch may be described as a wobbly note? What
diatonic pitch is restored? What harmony prevails at the point of restoration?14
•2. Without the wobble, what pitches might have constituted the harmony at S1–2,
and what Roman numeral would apply? With the wobbly note that Mendelssohn
deploys, what pitches and Roman numeral prevail instead?15
•3. What pitch classes are called into service for the embellishing chord at location
T. Comment on the resolution of its outer-voice dissonance (from the latter
chord of T to R2).16
••4. Given that V♯ is fully in place by the end of the U2 beat, what of interest happens
during U3?17
••5. How does the melody of the R3 and S3 beats relate to that of the R1 beats? How do
the harmonies of those beats relate?18
•6. How does the phrase cadence?19
8 III along the path from I to V

Considering the three internal diatonic triads along the ascending path
from I to V, II and IV have much in common – they share pitch classes, and
a surge opportunity always will be available in their deployment (I➔ IV
or II➔ V). In contrast, III blends in with both I and V, since it shares
two common tones with each. Though a I–III–V trajectory may guide a
progression at a foundational level, generally other content (such as a
segment of the circle of fifths, to be explored below) will come between
I and III and between III and V.
Alternatively III may be so closely allied with the preceding I that it
might best be analyzed as an element of that tonic’s prolongation.
Consider, for example, the progression
C < E < G < C
E < G < B
C < E < G < B♭

The internal E-G-B chord has little chance of coming across as an inde-
pendently asserted harmony in the context of the broader I ➔ initiative.
Especially in a minor key the diatonic circle of fifths offers an exemplary
pathway between I and III. For example, in the key of C Minor the E♭ major
mediant chord may come about through a trajectory that extends from the
tonic C minor chord through F minor and B♭ major chords. The B♭ chord’s
major quality helps make the E♭ arrival seem like a goal, and often an A♭
will join the triadic B♭-D-F in creating a surge. For that matter, the C minor
and F minor chords might be transformed into C-E♮-G-B♭ and F-A♮-C-E♭
as well, so that the circular progression proceeds as C ➔ F ➔ B♭ ➔ E♭.

8.1 Mendelssohn: Song without Words in E Minor

(op. 102, no. 1), mm. 5|6–93

A segment of the circle of fifths serves to connect tonic E Minor and a

tonicization of the mediant key, G Major, in the set of arpeggiations below.
69
70 Harmony in Mendelssohn and Schumann

That tonicization, which is grounded on a G–D–G bass arpeggiation, is

enhanced through the deployment of IV between I and V.

Because the circle of fifths juxtaposes chords whose roots relate as

descending perfect fifths, its potency often is enhanced through surges:
E ➔ A ➔ D ➔ G. The pitches that form the surge-igniting diminished
fifths are displayed within boxes in the set of arpeggiations below.

Example 8.1 A composition in E Minor’s second phrase.

•1. A circle of fifths transpires during this excerpt’s opening measures. Beginning
with tonic root E, what other roots are engaged over the course of its trajectory?
(Consider carefully where the circle of fifths ends and a tonicizing harmonic
progression begins.) What Roman numeral (in the context of E Minor) corre-
sponds to the circle’s goal chord?1
••2. Three subtly different means of igniting a surge are juxtaposed during this
excerpt’s opening measures. Compare what happens at R1–2, S1–2, and T.2
•3. What two features of chordal construction shift over the course of Y1 to Y2?3
•4. Provide a harmonic analysis for the tonicization of G Major from U1 through U2.4
 III along the path from I to V 71

8.2 Mendelssohn: Song without Words in A Minor

(op. 62, no. 5), mm. 42–121

Though the circle of fifths is an effective way to proceed from I to III, a

less involved trajectory that uses only the latter of the circle’s internal
chords may occur; for example, instead of A D G➔ C, one may establish
A Minor as the tonic and then proceed directly to G➔ leading to C. (In
this context, the G chord serves as an embellishing chord.) In the set of
arpeggiations below, that course is accomplished in the context of
ascending parallel tenths in the outer voices (boxed). This example
provides an especially clear instance of how the mediant may emerge as
an internal component of a stepwise ascending bass line connecting the
tonic and dominant roots. In this context, the A and C are closely related,
with the chord on D (here II7 in its first inversion) serving as the principal
intermediary between I and V♯.

In the following set of arpeggiations, the progression from the

initial tonic to the first dominant chord is more fully developed.
5
6
5
Once the tonic is established by means of a 3
4
3 embellishment, an
evolved state of II➔ emerges against tonic root A in the bass. This collision
of two different harmonies is especially effective in this case because the A
and C at the top of the texture perform a double duty: as a doubling of the
root and the third of I, and as the seventh and the ninth of II.
72 Harmony in Mendelssohn and Schumann

Example 8.2 A composition in A Minor’s opening phrases, which follow after a brief
introduction.

••1. How are the tonic pitches A, C, and E deployed (in two strands) in the upper
register during the first beat (quarter note followed by eighth note) at R1 (and
reprised at R2)? How does what happens during S1 relate to that trajectory?5
•2. Whereas the left-hand piano part during T repeats the A, C, and E that sounded
during R2, thus continuing the projection of the tonic harmony, the right-hand
part introduces two chromatic pitches, which join with A in forming another
harmony that sounds at the same time as the tonic below it. (This may be
described as a collision.) Name those two chromatic pitches, and indicate what
harmony comes between the I of R1 through R2 and the V♯ of U1.6
•3. What earlier location, within the excerpt’s first phrase, corresponds to R3? What
harmony was projected at that point? How does the content of R3 (both melodic
and harmonic) compare with that earlier location?7
•4. What is the relationship between the chords at R4 and R5? Comment on how the
chord at S2 prepares for the arrival of the chord at R5.8
•5. The chords at R4 and R5, assessed in question 4, both support an E (= ^ 5 in A
Minor) in the soprano melody. The melody then descends by step – through D, C,
and B – to reach A supported by I for a PAC to conclude the excerpt at R6. What
harmonies support those three interior pitches of the descending fifth? Comment
on the relationship between the bass’s linear trajectory from R4 through U4 and
the soprano melody’s stepwise descent.9
 III along the path from I to V 73

8.3 Schumann: “Und wüssten’s die Blumen” from

Dichterliebe (op. 48, no. 8), mm. 0|1–81

As you sing the set of arpeggiations below, note how the pitch B is deployed
in both the second and third chords (proceeding from the supertonic to the
dominant).

Now assess how a different character is achieved by allowing diatonic B

to wobble to B♭, so that ♭II fulfills the supertonic function. Observe that the
wobble’s extent is limited to that chord only: B♮ is restored for the domi-
nant. Also observe how, in the bass trajectory from tonic A up to dominant
E, the mediant pitch – C – is deployed between the tonic root (A) and the
supertonic third (D).

In the next set of arpeggiations, the initial tonic and lowered supertonic
harmonies are granted more space for elaboration, which in both cases
involves a stepwise descent from chordal third to root in the melody
against a bass arpeggiation of the triad’s three members (as in the tonic’s
C>B>A against A>E>C). Because the concluding bass pitch of the lowered
supertonic is D, the dominant is introduced in an inversion that places D in
the bass, later displaced by B.
74 Harmony in Mendelssohn and Schumann

In the final transformation a passing chord comes between the two

manifestations of the first dominant, while two passing chords come
between the internal tonic and the mediant, which now is presented in
its first inversion.
 III along the path from I to V 75

Example 8.3 The opening measures of a composition in A Minor.

•1. Compare the pitch content of the passages labeled R1 and S1 according to two
parameters: diatonic versus chromatic; major versus minor triad formation.
What foundational harmonies are projected? What harmony might one expect
to follow after S1? How does the chord at T1 fulfill that expectation?10
•2. Carefully explain the relationship between the chords at locations T1 and T2.
What role does the chord at location X1 perform within the measure’s broader
trajectory?11
••3. What is the relationship between the chords at R3 and R4? (Regard the chords at
X2 and X3 as connective chords.)12
•4. Provide a Roman numeral analysis for the chords at S3, T2, and U.13

8.4 Schumann: “Ich hab’ im Traum geweinet” from

Dichterliebe (op. 48, no. 13), mm. 22|23–38

The first set of arpeggiations for what will eventually become an extra-
ordinarily chromatic presentation deploys chromatic pitch G♮ in the sur-
ging I➔ that targets IV. The DG ♭♮ diminished fifth and its AC ♭♭ resolution are
highlighted by boxes.

In the expanded set of arpeggiations below the tonic proceeds to the

mediant via a circle-of-fifths trajectory: E♭ A♭➔ D♭ G♭ (marked by boxes).
A juxtaposition such as the goal G♭ chord and the E♭➔ chord that follows
often is challenging to assess analytically. Does the latter chord reinstate
the initial tonic (now with surging force) so that, broadly, I ➔ IV tran-
spires? (In that scenario, the G♭ chord provides a consonant context for the
introduction of D♭, which will become the tonic surge’s dissonant seventh
upon the re-emergence of root E♭.) Or is that latter chord instead a local
surging 6-phase connection internal to the stepwise trajectory III5–6 IV? (A
careful study of Schumann’s more richly detailed score, from which these
arpeggiations are derived, would support the I ➔ IV reading.)
76 Harmony in Mendelssohn and Schumann

Two new initiatives are incorporated within the following set of arpeggia-
tions. First, the opening tonic is prolonged through 5
6
5
3
4
3 neighboring

embellishment. Second, the G♭ mediant is prolonged through ascending

motion from its 53 position to its 63 position, filled in diatonically. That is,
the G♭<B♭ third is filled in by A♭, the B♭<D♭ third is filled in by C♭, and the
D♭<G♭ fourth is filled in by E♭ and F. (The way in which boxes are applied
below distinguishes these lines from one another.) Because this is a linear
initiative within a hierarchically conceived framework, that trajectory’s
internal chords (G♭-B♭-E♭ and A♭-C♭-F) are not asserted as harmonies.

In the final set of arpeggiations, the initial tonic’s 64 embellishment is

unfurled, with A♭ sounding in the bass. Within its context this A♭ does not
 III along the path from I to V 77

serve as a chordal root, but instead is a doubling of one of the tonic’s

neighboring notes. Soon thereafter, however, A♭ re-emerges as the root of
the circle of fifths’s internal A♭➔ chord.
Whereas a diatonic expansion of the mediant was developed in the
set of arpeggiations above, the set below offers a richer alternative. Now
each chordal interval is filled in chromatically, while a D♭ pedal point
persists throughout. Because a perfect fourth, a major third, and a
minor third are involved, the ascending motions commence at different
moments. Observe how, among the boxed pitches, Eº in the upper
strand begins the initiative (since that strand has the greatest distance
to traverse, a perfect fourth), with G♮ (within a major third) launching
the lower strand on the next chord, and C♭ (within a minor third) on
the following chord. Once all three lines are in motion, they proceed in
lockstep until the goal chord is attained. It would be foolhardy to label
each individual chord within this trajectory with a Roman numeral.
Only the endpoints, which project the mediant’s pitch classes (G♭, B♭,
and D♭), are players in the broader scheme of things.
Once the inverted G♭ chord arrives, a reversal of direction leads to I➔. A
passing chord (A♭-D♭-F♭) connects those two more foundational chords.
78 Harmony in Mendelssohn and Schumann

Example 8.4 The third, final stanza of a song in E♭ Minor.

•1. For reasons that will be explored later, the chord at R3 contains an unexpected
5
pitch. For now thinking of R3 as if it were a diatonic 3 chord above the sounding
bass, explain what transpires from R2 through R3, indicating the roots and any-
thing notable about the construction of the internal chords at T1 and T2.14
••2. Compare the chords at S and at T1 within their contexts.15
••3. Having heard a succession of chords whose roots relate by descending perfect
fifth (as explored in question 1), listeners should expect to hear a chord spelled as
G♭-B♭-D♭ at R3. Schumann elides that stable moment, using the chord at R3 to
initiate a chromatic ascending trajectory (eventually in three voices) to connect
the elided G♭-B♭-D♭ mediant and its first inversion, B♭-D♭-G♭. First, what note
sounds along with members of the G♭ mediant chord at R3 to convey that this
upward initiative has already begun? Second, which chord among the successors
(T3, T4, T5, R4, or T6) projects the first-inversion G♭ chord?16
•4. The bass pitches of the chords from R5 through X1–3 fall within a stepwise
ascending trajectory. Name these bass pitches and indicate what harmonic
progression they support, prior to the tonic resolution at R6.17
part ii

Masterpieces
9 Mendelssohn: Octet in E♭ Major (op. 20),
movement 1
in response to greg vitercik and benedict taylor

The analytical literature in English devoted to Mendelssohn is somewhat

less robust than that for Schumann. Thus the two book-length studies of
Mendelssohn’s music featured in this chapter hold especially prominent
places. In The Early Works of Felix Mendelssohn: A Study in the Romantic
Sonata Style, Greg Vitercik offers an array of analyses.1 Though in other
chapters his examples are graphs in the Schenkerian sense, his presentation
on the first movement of Mendelssohn’s Octet instead features figures
consisting of rows of Roman numerals annotated by measure numbers,
sometimes with further annotations addressing form. These are comple-
mented by a detailed textual commentary. Even if his and my examples
contrast visually, readers should be able to correlate the two easily.
Benedict Taylor’s Mendelssohn, Time and Memory: The Romantic
Conception of Cyclic Form is a recent addition to the analytical literature.2
The chapter in which the Octet’s first movement is discussed is titled
“Cyclicism and the development of material across the Octet.” Though
considerations of harmony (my chapter’s focus) are not at the center of his
purview, his exploration of the exposition offers many ideas that should be
intriguing to ponder in juxtaposition with my reading.
My conception of form has been influenced deeply by James Hepokoski
and Warren Darcy’s Elements of Sonata Theory.3 The Octet’s formal orga-
nization will be assayed below using their terminology, which I find insight-
ful especially when integrated with my harmonic perspective, thus giving my
presentations a character and scope distinct from what appears in their book.
As does Vitercik, I draw upon Mendelssohn’s 1825 manuscript version
of the movement to vivify some of my points. That document is readily
available for study in a facsimile reprint.4

The exposition’s primary-theme zone (P, mm. 1–37)

For a movement as long as the Octet’s Allegro moderato, a primary-

theme zone (P) devoted to the initial presentation of the tonic key may 81
82 Harmony in Mendelssohn and Schumann

contain multiple parts and fill a considerable time span. Mendelssohn

juxtaposes two distinct tonic-establishing trajectories: from measure 1
through measure 21, and with fresh melodic content and a more austere
mood from measure 21 through measure 37. In the former, multiple
departures from and returns to the tonic root E♭ in the bass transpire
during a unified push toward the IAC at 211. In 9.1 tonic root E♭ is
articulated in measures 1–2, 9–11, 13–15, and finally 21–22. The latter
coordinates with a downward linear trajectory from the Kopfton (adum-
brated in the passage leading to 91): B♭>A♭>G. Though the descent is
spread over multiple octaves (connected by a flourish of passing notes), a
high G emerges during measure 22, at the onset of the second broad
tonic-prolonging trajectory.

Example 9.1 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm. 1–22.

What comes between I and V during the progression of measures 1

through 9 serves as a model of exemplary tonal construction that
Mendelssohn will develop skillfully as the movement proceeds. Clearly he
well understood how the addition of 6-phase C to an E♭ major tonic triad
may serve effectively in guiding the progression toward the F supertonic.
Here that C emerges in the bass, attained through an E♭>D>C passing
motion from the initial tonic root. The supertonic that follows in measures
5 through 7 does not surge. In fact, the shift of its C to C♭ provides a
temporary hint of E♭ Minor’s mellower supertonic.5 Another descending
 Mendelssohn: Octet in E♭ Major (op. 20) 83

third in the bass (F>E♭>D) leads to the dominant’s presentation in its first
inversion, with leading tone D resolving to tonic root E♭. The deployment of
D below dominant root B♭ not only softens the resolutional strength at this
early point in the tonic expanse but also introduces a melodic trajectory that
Mendelssohn will develop in what ensues: note the sforzando-enhanced
D<E♭ second of measures 12–13 and especially the non-conforming D<E♮
second of measures 16–17. What is Mendelssohn up to?
Despite contrasts in spelling, the same four pitch classes are deployed at
the downbeats of measures 10, 11, 12, 14, 15, and 16. Mendelssohn’s
composition here demonstrates the principle of Mehrdeutigkeit (multiple
meaning), a technique for which the diminished seventh sonority is espe-
cially well suited.6 A common-tone resolution is pursued in measures 10,
11, 14, and 15: F♯ serves as a lower neighbor to the tonic’s G, A♮ as a lower
neighbor to its B♭, and C as an upper neighbor to its B♭, all against a
retained root E♭. In measure 12 several parameters change. Note especially
that the sforzando has been shifted to beat 3 and that G♭ appears in place of
F♯. Here A♮-C-E♭-G♭ functions as II➔, targeting the dominant.7 Though
the bass sounds a C (corresponding to 41–2), the succinct harmonic pro-
gression has skipped over the tonic’s 6 phase altogether, proceeding
directly to this inverted supertonic (corresponding to measures 5 and 6,
now surging).
At first measure 16 will seem to be following the course introduced
during measure 12. This ruse requires the use of B♭ as a chord member at
163, rather than the B♮ shown within parentheses in 9.1. (The B♮ would
affirm the chord’s internal location among diminished seventh chords,
demarked by long slurs in the example.) Mendelssohn ultimately trans-
forms the passage into an extended elaboration upon the tonic 6-phase
chord. Note that the sforzando and the C<D slur of measure 12 are missing
in measure 16, that new accent marks commencing on C have been added,
and especially that E♮ sounds instead of E♭ at 171. My parenthetical B♮ (a
note that might have sounded instead of B♭) in 9.1 helps clarify that a
parallel progression of diminished seventh sonorities (with “wrong” note
B♭ sounding instead of B♮) transpires from 161 through 192. Though the
C-E♭-F♯[G♭]-A♮ chord of 161 might have resolved directly to C-E♭-G (in
the manner of the chords of measures 10, 11, 14, and 15, now embellishing
the tonic’s 6-phase chord), Mendelssohn instead pushes the diminished
seventh sonority gradually upwards to B♮-D-F-A♭, which in fact does
resolve to C-E♭-G. Consequently he compensates for the lack of a
6-phase tonic sonority for 121’s bass C by flamboyantly leading to such
84 Harmony in Mendelssohn and Schumann

a chord at 193. Yet in this case the continuation is to IV (supporting the

Kopfton’s upper neighbor, C, replicating the C of measures 5 and 6),
instead of to II or II➔. A V8–7 I cadence (notably not coinciding with the
endpoint of a descending fifth-progression from Kopfton ^5 ) concludes
the trajectory.
Listeners will not know for certain whether what commences during
measures 21 and 22 represents a new phase of P or instead the onset of
the transition (TR). Only the return to the E♭ tonic chord in measures
29 and 37 confirms that this material resides within P. Considering how
tonal music tends to work, the most likely structural content for this
region would be a middleground fifth-progression descending from
Kopfton ^5 to tonic E♭, thereby reaching a point of stability through a
PAC, contrasting the IAC at 211. Mendelssohn’s two related progres-
sions during these measures appear to be pursuing that course, though
complications emerge in both. In the first (through measure 29), not
all of the pitches literally sound in the soprano, thus triggering some
imaginative analytical filling-in; and though the second (through
measure 37) appears headed for goal ^1, the concluding progression
from V to I takes an unusual turn, softening the cadential potency
usually associated with a PAC.
Because these related trajectories both feature deviations from the
conventional path, 9.2a shows a somewhat idealized foundational
progression that will serve as a starting point for our discussion of
Mendelssohn’s more vivid and sometimes stupefying realizations. The
tonic chord’s initial high BG♭ third is unfolded as G<B♭ during measures
22 through 24, in coordination with the shift from the tonic’s 5-phase to
its surging 6-phase chord. (An embellishing chord sounds in between.)
The latter chord’s dissonant outer-voice BE ♮♭ helps motivate the onset of
the fifth-progression’s downward course: from B♭ to A♭. Yet note that
the melody’s ensuing unfoldings of two strands (A♭ followed by G, F
followed E♮ from 254 through 271) eventually lead to a continuation
only of the lower of those lines: from E♮ through E♭ to D during
measures 27 and 28. G’s successors G♭ and F sound only an octave
lower, in the Violin II line. In 9.2a those pitches are displayed in their
normative soprano positions. (Parentheses are not applied because
these pitches do sound as proposed during the second pass through
the phrase, during measures 31 and 32.) The graph also suggests that
foundationally the supertonic might be prolonged via a voice exchange
♭
(F 
A F
A♭ ) followed by the onset of a surge. In Mendelssohn’s score the
 Mendelssohn: Octet in E♭ Major (op. 20) 85

latter is integrated with the former, so that the melody’s G♭ serves as G’s
successor. At the phrase’s end (which in this first pass through the material
forgoes the dominant prolongation proposed in the model), the sounding
D<E♭ achieves the same E♭ that would have served as the goal had G♭ and F
actually transpired in the melody. Consequently, despite the significant
lapses (and keeping in mind that the next phrase makes good on these
points), one may project a viable descending fifth-progression, concluding
in a PAC.

Example 9.2 Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1 (a) analysis of mm. 21–
37; (b) two models for common-tone embellishing chords.

The high B♭ that comes between F and E♭ in the Violin I line during
measure 28 reminds listeners that a B♭-to-E♭ fifth-progression is a
means of prolonging the initiating B♭. Thus the A♭ that follows at 294
86 Harmony in Mendelssohn and Schumann

should be interpreted as a downward step from Kopfton B♭. Though the

tonic’s 6-phase chord evolves into a surge during 241–2, a simple C
(embellished by D♭ and B♮) between the tonic and supertonic chords
suffices during measure 29. This time the melody that follows favors
the upper of the two strands: after upper-strand A♭>G and interior-
strand F>E♮ again occur, a G♭ (at 314) continues the course from G. Yet
now the resolution of the dominant reached at 321 goes awry. The
graph in 9.2a suggests that the chords containing the pitches E♭, G, and
B♭ during measures 33 through 35 “ought” to come between two
iterations of the dominant, followed by a conventional resolution to
the tonic. (It would be worth the reader’s time to skip ahead for a
moment to the coda, where multiple iterations of bass G<A♭<(A♮)<B♭
beginning in measure 295 lead, like the G<A♭<B♭ of measures 33 and
34, to a cadential 64 that does yield to the dominant 53 for a conventional
PAC close. Also compare with what happens in the region between
measures 101 and 112 from the exposition S indicated by the hairpin
symbol in 9.5, in the tonicized dominant key.) Yet as the dynamic level
lowers to piano, an A♭-C♭-E♭-F chord emerges during measure 36. One
might contemplate three different viable continuations. First, a low-
ering of that chord’s E♭ to D would produce a highly evolved state (and
unusual inversion for a cadential moment) of the dominant. But D
does not emerge. Second, the chord replicates one sounded earlier –
during measure 7 – and thus it might herald a restored V (as during
measure 8). But that does not happen either. Third, the chord’s direct
resolution to the E♭ tonic relates closely to the tonic embellishments
of measures 10, 11, 14, and 15. In 9.2b that procedure (Model 1) is
juxtaposed with what occurs during measures 36 and 37 (Model 2).
I regard this as the most apt analytical response. Mendelssohn has
allowed what initially seemed to be an embellishment between two
dominants to anticipate the phrase’s closing tonic. In retrospect,
measures 34 through 38 should congeal as a chromaticism-enhanced
3




4
3 tonic prolongation. Though attentive listeners at first probably
I 8
♭7
6
5

would regard bass B♭ during measures 34 and 35 as a reinstatement of the

dominant root from 321, what follows likely would trigger a revision: that
the B♭ followed by the low E♭ of measure 37 arpeggiate the fifth and root of
an asserted tonic.
 Mendelssohn: Octet in E♭ Major (op. 20) 87

As far as it goes, Vitercik’s harmonic analysis of measures 1 through 21

is exemplary (see his fig. 2). We agree that measures 1 through 9 present
an effective (though conventional) harmonic progression, pursuing a
route through ii to reach V and I. Yet his cursory treatment of measures
9 through 21 – displayed as (I – I) in his fig. 3 and as I – – in his fig. 4 –
contrasts my hierarchical presentation in 9.1. I take measure 8’s bass D,
supporting a dominant chord in 65 position, to indicate a tentativeness in
the progression to that point, offering a motivation for continuing
efforts that eventually are rewarded by the arrival of a more definitive
bass B♭ in measure 20. That B♭ is attained only after further reinforce-
ment of D, at 123 and (in what is transformed into a contrasting context)
at 163. Thus the C>A♭<B♭ bass of measures 19 and 20 is for me a victory
within a subtle struggle, which Vitercik does not probe. Contrasting
his front emphasis, my graph is weighted toward the end: its beamed
E♭<B♭>E♭ bass arpeggiation connects pitches from measures 1, 20,
and 21.
Taylor likewise appears to favor the cadence of measure 9 over that of
measure 21, in that only the former E♭ root is displayed in his ex. 2.1
(though both appear in the more detailed ex. 2.5). His inclusion of just
one pitch – F, annotated by the parenthesized label (Fm) – between the
E♭ roots of measures 1 and 9 seems misguided to me, especially since his
caption for ex. 2.1 is “harmonic reduction of exposition.”8 I suggest that,
instead, the principal chord between the two E♭ chords is that with bass
D (measure 8), which both Vitercik and I label as V. It even appears that
Taylor contradicts himself, since his ex. 2.5 shows slurred treble
clef F<A♭ against slurred bass clef F>D>B♭ between those two E♭s.
(These pitches congeal as B♭-D-F-A♭.) Especially given his (and
A♭
Mendelssohn’s) vertically aligned D compound diminished fifth
(which resolves to the tonic’s third), “(Fm)” seems an inapt synopsis
G
E♭

of what occurs harmonically between the two tonic chords.

Our contrasting hierarchical assessments persist as P continues, as
evidenced by a comparison of my 9.1 and Taylor’s ex. 2.5. Observe how
I prioritize the dominant through the flagged bass D in measure 12,
matching the flagged D in measure 8. Mendelssohn’s score emphasizes
both Ds: a crescendo reaches forte at the D of measure 8, while that at
123 is marked sforzando. Taylor instead deploys a slur to connect the
88 Harmony in Mendelssohn and Schumann

C of 121 and the E♭ of 131. (That notation conveys that C outranks

D hierarchically. Though he does not explicitly mention it, that C
corresponds to an F chord: not the Fm of measures 5 and 6, but a
surging F➔, spelled as A♮-C-E♭-G♭.) Then he abruptly reverses the
hierarchy: Mendelssohn’s dominant root B♭ (instead of D) in measure
20 is granted a stem, so that E♭<B♭>E♭ (exactly in agreement with my 9.1)
is conveyed as the principal bass trajectory of measures 13 through 21.
Though Taylor’s presentation does not aspire to the same level of
Schenkerian sophistication as does mine, the persistent G open note-
heads on the upper staff of his ex. 2.5 convey a “prolongation” of ^3
(p. 65). In that regard, note the extraordinary misalignment of upper-
and lower-staff pitches in the vicinity of measures 19 and 20 in that
example. The pitch B♭ (in my reading an important reinstatement of ^5 )
arrives on the downbeat of measure 20, exactly in alignment with bass
B♭, not off to the right, as in Taylor’s representation. Likewise in his ex.
2.3a Taylor picks out several metrically weak Gs for analytical emphasis,
passing over the downbeat B♭s of measures 2 and 4 and B♭’s neighbor, C,
at the downbeat of measure 6. Several accent marks in the Violin I line
help shape the melodic contour.9 The tonic fourth B♭<E♭ (measures 2–3,
repeated an octave higher without accents during measure 4) is echoed
by the supertonic’s C<F (measures 6–7, whose upper-octave repeat is
integrated with the chromatic shift of C to C♭ over the measures 7|8 bar
line), with a likewise accented A♭ (now off the beat during 81) leading to
the cadential G (unaccented but on a downbeat, at 91). This trajectory
projects ^5 – not ^3 – as the foundational pitch, with a third-progression
to an interior G.
Vitercik’s Roman numeral analysis for measures 21 through 37
appears in both his fig. 1 and fig. 3.10 I applaud his interpretation of
E♮-G-B♭-D♭ (measures 23–24) as a chord rooted on C. His VI and my
VI♮3 labels convey the same perspective.11 My label (fully constituted
in 9.2a) provides further information about the chord’s construction,
while its physical location reveals the chord to be a highly evolved
manifestation of the tonic’s 6 phase.12 The ii label that follows does not
tell the whole story. Vitercik incorrectly describes the passage as
“sitting on the minor supertonic throughout mm. 24–27” (p. 73).
The last of those measures instead concludes with an evolution of the
supertonic into A♮-C-E♭-G♭.13 Though his Roman numeral usage can
 Mendelssohn: Octet in E♭ Major (op. 20) 89

display a diatonic chord (ii) or a surging chord (VI), it is not equipped

to convey how one harmonic function progresses through multiple
stages over the course of its prolongation. That is particularly critical
here, since he offers multiple hypotheses regarding where the phrase is
heading: the supertonic, the dominant, or the tonic (p. 73). How the
supertonic behaves (as well as the particular dissonances that sound
against the onset of dominant root B♭) needs to enter into that discus-
sion. Finally, I do not think Vitercik intends for readers to interpret the
measure numbers in his fig. 3 literally. In my practice, placing the
number 28 below a V numeral and the number 37 below a I numeral
would indicate a nine-measure dominant prolongation. Here it is
probably instead a rather awkward way of conveying how two succes-
sive iterations of a progression both lead to a cadence on the tonic: his
V (28) is from the first iteration, while his I (37) is from the second. (He
says nothing about the extraordinary writing during measures 34
through 36.)
Before we explore Taylor’s analysis of measures 21 through 37, as
conveyed in his ex. 2.5, readers may want to supplement that example’s
four measure numbers in order to better facilitate comparison with my
9.2a and with the score, as follows:

21 22 23 25 27 28 29 31 32– 33– 34 36– 37

and also restore the natural sign that was inadvertently omitted to the
left of the bass A notehead at measure 27.
Taylor’s supertonic focus persists through the inclusion of just one
internal chord between the tonic E♭s of measures 9 (21) and 37 in his
ex. 2.1: a C notehead annotated as V/Fm. That chord sounds in both
of the phrases during this region – at measures 25/27 and 31. Yet
again I suggest that Taylor has gravitated toward a chord that is
hierarchically dependent upon others in the vicinity and thus unsui-
table as the sole representative of what happens between the two
tonic E♭s. The C-rooted chords at the downbeats of measures 25, 26,
and 27 each are preceded by an F minor chord, which both Vitercik
and I interpret as the supertonic. And an F chord (in an evolved state)
follows after the last of those C-rooted chords: A♮-C-E♭-G♭ at 274.
Consequently the focus on C rather than F seems to me a misreading
of the passage. (Observe how C appears as internal to a bass slur in
90 Harmony in Mendelssohn and Schumann

my 9.2a. This hierarchical relationship prevails even when the surge-

inducing A♮ substitutes for diatonic A♭ during measures 27 and 31.)
Besides his topsy-turvy hierarchical assessment, Taylor is inconsis-
tent in his choice of chords for inclusion within his ex. 2.5. Whereas
in the first of the two phrases the F chord before the C chord of
measures 25/27 is omitted, but the F➔ chord that follows is included
(if we correct bass A♭ to A♮ during measure 27), in the second phrase
neither of those F chords is displayed; that is, what I suggest should
be projected as an F–C–F prolongation of the supertonic includes the
internal C chord but excludes both perimeter F chords. Taylor
instead proceeds directly from an F-targeting C chord to a B♭ chord
(measures 31–32). It is uncertain whether he interprets the B♭ chord
as a dominant, for three reasons: first, he slurs through rather than to
bass B♭ in his ex. 2.5 (in contrast to his adequate treatment of the
earlier phrase’s equivalent B♭, in measure 28), contrasting my
stemmed and beamed bass B♭ in 9.2a; second, his presentation
appears without any Roman numeral, leaving one guessing regarding
his reading of the harmonic trajectory; and third, his summary
“harmonic structure” for measures 1 through 37 (his fig. 2.1) reads
as follows: “E♭ V/f E♭.”14
Finally, a small but important point needs to be raised. Both
Taylor and I are intrigued by the sonority projected during the
second half of measure 36. I feature it as Model 2 in 9.2b and
clarify how it expands the movement’s already rich deployment of
common-tone chordal resolutions. In my understanding of that
notion, such chords essentially are linear – not harmonic – phe-
nomena. Thus I reject Taylor’s proposed label for the chord: ii 7♭ 5
(p. 65). The pitch F is not the chord’s root (as the ii numeral indicates).
E♭ is, and has been projected as a root since measure 34, the onset point
♭7
6
5
for an 8

3




4
3 tonic expansion that incorporates one extra pitch near

its end (along with the chromatic lowering of C to C♭): the F that
converts the 64 embellishment into a chord in 42 position. With or with-
out that F, that chord remains an embellishment. The F, A♭, and C♭ are
all non-harmonic pitches that eventually fall in line with the assertion of
E♭ as root. (The chord with C♭ has been unfurled, so that A♭ is the lowest
sounding pitch.)
 Mendelssohn: Octet in E♭ Major (op. 20) 91

The exposition’s transition (TR, mm. 37–68)

Not surprisingly, Mendelssohn’s exposition is anchored by P in E♭ Major

and by S (the secondary-theme zone) in the tonicized dominant key, B♭
Major. The TR that comes between those zones often will reach an MC
(medial caesura) on II♮, preceding V’s arrival at the onset of S. In this case,
in a less common though equally viable strategy, the transitional energy
proceeds all the way to V (at 681), followed immediately by the onset of S
(682) on the same harmony. (The swiftness of S’s onset may appear to
contradict the very notion of caesura that defines this moment. As we shall
see later, during its recapitulation equivalent an extra measure is inserted
before the onset of S.) Given that ^5 has been established as the movement’s
Kopfton during P, it would be typical for the Urlinie to descend through ^4
and ^3 to ^2 during TR. Here E♭ Major’s ^2 – F – will then persist, serving
locally as ^5 during the B♭ Major tonicization of S and beyond.)
Though TR begins with material similar to that from the onset of P
(observe how the E♭>D>C bass of measures 1 through 4 takes on a
chromatic character with the E♭>D>D♭>C of measures 37 through 41),
I will show later how the E♭ C➔ F trajectory (which notably does not
proceed to dominant B♭) is subsumed within a broader, inexorable circular
progression. Consider first a diatonic model for a descending circle of fifths
connecting tonic E♭ and dominant B♭ [9.3a]. Though the retention of the
pitch A♭ (diatonic in the principal key of E♭ Major but not in the key of B♭
Major to which the exposition is heading) results in an exemplary stepwise
descent from ^5 through ^4 and ^3 to ^2, the progression has awkward
moments, especially as it proceeds through the diminished triad D-F-A♭
and as the goal B♭ root is preceded by bass A♭ rather than by its leading
tone, A♮. The alternative path proposed in 9.3b is more felicitous on both
counts. The diminished triad is done away with entirely, while the third,
fifth, and seventh chords are targeted by surges, as the arrows in the
annotation below the example convey. One might forgive the divergence
from the diatonic pitch collection in the broad melodic descent (with A♮
substituting for A♭) in order to sail through tonal space along this elegant
and appealing route. Mendelssohn elects to transform the progression even
further, producing a vigorously forward-driving trajectory [9.3c]. Observe
how its second chord integrates components of the corresponding chords
from the other models: diatonic E♭ from 9.3a and chromatic A♮ and C♯
from 9.3b, with seventh G added to produce a supersurging chord. The D
chord to which it resolves now sports an F♯, thereby attaining a surge
92 Harmony in Mendelssohn and Schumann

character, targeting G. With its B♮, this G chord surges mildly toward
C. At that point, however, a significant makeover transpires: instead of
the C-E♭-G of the earlier models, Mendelssohn deploys E♮-G♭-B♭-D♭,
another supersurging chord. (Mendelssohn spells the chordal ninth, D♭,
enharmonically as C♯ because it proceeds upwards to neighbor D♮, a
member of the 64 embellishment of the F chord that follows.) This is, in
fact, an ideal location for a supersurge: since the key of B♭ Major will be
tonicized during S, a C⇨ F➔ B♭ trajectory commendably projects the
sense of II⇨ V➔ I, featuring the supertonic’s frequent evolution into a
supersurge.15

Example 9.3 Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1 (a) model of a diatonic
circle-of-fifths connection between tonic E♭ and dominant B♭; (b) a surge-intensified
variant of the circle-of-fifths connection; (c) analysis of measures 37–68.

Our discussion above suggests that making creative adjustments to a

circle of fifths – with customization to fit the specific context in which it is
to be deployed – is a procedure that might distinguish a composer of merit.
By traversing an awkward E♭-to-A♮ diminished fifth first, the remainder of
the progression may proceed memorably and without a hitch. This strategy
places emphasis on setting the circular progression in motion: both over-
coming the inertia of the E♭ tonic triad and directing it toward A♮ rather
than the more obvious choice of A♭. For this purpose it is surprisingly
 Mendelssohn: Octet in E♭ Major (op. 20) 93

efficient and effective merely to add a pitch to the tonic triad. Observe in
9.4a how placing a C♯ atop E♭-G-B♭ creates a sonority that may be inter-
preted as an A♮⇨ supersurge, targeting D. (The same added note under-
stood as D♭ would instead generate an E♭➔ surge targeting A♭, thereby
giving attentive listeners a sense of uncertainty that is resolved only upon
hearing what comes next.) As with any supersurging chord, the chordal
ninth may be incorporated (here B♭) or the root may sound (here A♮).
Those alternatives are juxtaposed in 9.4a.

Example 9.4 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm. 37–52.

An appealing way to reach this added C♯ is through an ascending

passing motion (B♭<C<C♯), as proposed in 9.4b. (That example incor-
porates A♮, rather than B♭, in the supersurging chord, since that turns
out to be Mendelssohn’s choice.) The further elaboration in 9.4c shows
a consonant 53 chord as support for passing note C, deploying material
(now expanded) related to what supported the neighboring note C of
measure 6. Mendelssohn consequently establishes a parallelism
between the onsets of P and of TR even though the F minor chords
ultimately play contrasting roles in these two contexts. Though some
readers might question the assessment of chordal hierarchies conveyed
in 9.4c, the proposed E♭ A♮⇨ D➔ trajectory is affirmed by the circular
progression’s unimpeded continuation in the measures that follow, as
documented in 9.3c.
94 Harmony in Mendelssohn and Schumann

Though I have reservations about Vitercik’s reading of the material

I interpret as TR, I should begin by clarifying that we are not quite so
far apart as a cursory comparison of his fig. 5 and my 9.3c and 9.4c
would suggest. His presentation focuses on a melodic neighboring
initiative, and thus the pitches that he singles out (annotated by measure
numbers) are as follows:
F – (F♯) – G – (F♯/G♭) – F
45 54 55 58 59

Had his presentation instead focused on chordal roots, he might have

displayed

D G C F

exactly as in my reading.16 The discrepancies in our annotating mea-

sure numbers have two causes. The first – his 54 versus my 52–53
(which should be understood as the formation stage of a chord that
persists until the next number, 57) – comes from his melodic focus:
though root D sounds at 521, F♯ does not emerge in the bass until 544.
The second – his 55 versus my 57 – stems from what I regard as a
misreading on his part. Especially given Mendelssohn’s careful slur-
ring of the Violin I, Viola I, and Violoncello II lines, the content
of measure 55 certainly should be understood as falling within the
orbit of the preceding and following D chords, in what I might label
8
♮7
♮6
♭6
5
as D ♯3






4




♯3 . Though the G-B♮-D chord of measure 57 is intro-

duced as if it also were embellishing the D chord, what follows confirms

that its G is asserted as a root, moving the progression beyond the
D chord.17
If one also considers the B♭ goal chord of measure 68, then a con-
spicuously long string of chords related by descending fifth transpires
between measures 52 (54) and 69. Vitercik’s F before his (F♯) – my D –
does not accord with that trajectory, whereas my A♮ does. That A♮
directly precedes the D chord, whereas his reading leaves eight measures
unaccounted for.18 Beyond that, the juxtaposition of neighboring-note
embellishments that is the centerpiece of his argument – a “parallelism . . .
suggesting a process of growth and transformation” between E♭–F–E♭
and F–G–F (p. 75 and fig. 5) – seems to me a far stretch. By consulting
my 9.2a one will observe that I regard B♭ – not F – as the principal root
between the E♭s of measures 21 and 29/37. (I also challenge Taylor on
 Mendelssohn: Octet in E♭ Major (op. 20) 95

this point, as the commentary above confirms.) In applying labels to

individual chords in his fig. 6, Vitercik produces something that seems
to me incoherent as a harmonic progression:

I – VI – ii – VI/ii – ii/ii – V
37 41 45 52 55 68
I propose that the chord at 514 needs to be incorporated, and am
mystified by the juxtaposition of ii/ii and V.19 (Compare with my
9.3c, which incorporates C and F chords between my equivalents of
his ii/ii and V.)
Taylor’s ex. 2.1 incorporates much of the content featured in my
examples: C➔ F (measures 41–45) from 9.4c, and both D➔ G (mea-
sures 52–57) and F➔ B♭ (measures 59–68) from 9.3c. Yet the two
supersurging chords (A♮⇨ and C⇨) are conspicuously missing.
Conventional harmonic analysis has done a disservice to this class of
chords by instituting labels that are incommensurate with a root-
oriented perspective regarding chordal progression.20 Too often, ana-
lysts tacitly demote the Ger+6, Fr+6, and It+6 chords to a subordinate
hierarchical level. Though the roots featured in Taylor’s reduction all in
fact play roles at one structural level or another in my reading, I am not
persuaded by his conception of what lies behind the broad coherence
that Mendelssohn achieves. His “up a step” notion (p. 61) accounts for
Fm and for Gm but leaves a gap (where A♭M would go) prior to the goal
B♭. My circular conception, which places the Fm arrival at a subordinate
level, displays a circular trajectory from tonic E♭ to dominant B♭ without
a gap (9.3c).

The exposition’s secondary-theme zone and closing zone

(S and C, mm. 68–129b)

A harmony often will be prolonged through the traversal of linear progres-

sions that fill in its intervals. Broadly E♭ Major’s B♭ dominant harmony
(tonicized as I in B♭ Major) pervades S in part because an F-to-B♭ fifth-
progression transpires over S’s entire duration, as beamed in 9.5. Such lines
may occur also at the foreground level, as in measures 67 through 75. From
9.3c one may understand that the soprano pitch at TR’s MC (681) is F. Yet
even as that F is being established an unfolded E♭>C third (674) begins a
downward trajectory, complemented by S’s opening B♭<D third (682–4).
96 Harmony in Mendelssohn and Schumann

The upper pitches of those unfoldings initiate a descending line from

the foundational F through E♭ to D. Embellished by upper neighbor E♭
(at 691, reverting to D at 693; and more briefly at 702), D eventually
yields to C during 703 and then B♭ at 711. This fifth-progression’s lower
third is repeated during measures 72 through 75, where another interval
of the prolonged B♭ major harmony is filled in as well, as the sixth-
progression B♭>A♭>G>F>E♭>D, split between the Violin IV and Viola
I parts. That line’s A♭ interacts with the upper line’s D, intensifying the
rise to neighbor E♭.

Example 9.5 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm. 68–113.

At the onset of P the pervasive arpeggiation of the tonic’s E♭, G, and B♭

chord members in the Violin I line soon competes with an alien element:
the C that emerges at 41. In that case the tonic’s characteristic 5–6 shift
transpires against a retained 5 (so that E♭, G, B♭, and C all sound
together), and the 6-phase chord is unfurled, with C residing in the
bass. The supertonic (the most likely continuation after I5–6) follows,
beginning at 51. Though the procedure expands to fill many more mea-
sures, the same initiative plays out at the onset of S: B♭ Major’s tonic B♭
chord is introduced in measure 68, unfurled 6-phase bass G arrives at 761,
and the supertonic C chord emerges between 851 and 861. In this more
luxuriant setting the 6-phase chord evolves into a dynamic, supertonic-
targeting surge. Diatonic G-B♭-D is transformed into G-B♮-D during
measure 77, and both seventh F and ninth A♭ are introduced during
measure 84.
 Mendelssohn: Octet in E♭ Major (op. 20) 97

It is critical to perceive that the Violin I’s F at 842 is a restoration of the F

that sounded at the juncture between TR and S (as the initiating pitch of the
descending fifth-progression discussed above). At the onset of the tonic’s
6-phase chord, Mendelssohn temporarily focuses on interior pitches. By
the time F is reinstated, the surging G➔ chord is ripe for resolution. The F’s
tendency to descend to E♭ is fulfilled an octave lower (in the Viola I line)
over the measures 84|85 and 85|86 bar lines, followed at 862 by its sounding
in the upper register (in the Violin IV line).
The supertonic’s expected successor is V, of course. Though those two
chords are shown in direct succession in 9.5, Mendelssohn’s foreground
realization offers a more elaborate trajectory. Focusing on my graph’s C
and F chords, imagine filling in the G<A♮ whole step in the alto register
with A♭, a chromatic pitch in B♭ Major. An unfurled A♭-C-E♭ chord in fact
sounds at 871, and by the time bass C is restored three measures later the F
dominant has emerged. The connection between A♭ and C is accomplished
through an ascending 5–6 sequence, as follows:
m. 87 88 89 90
5—6
A ♭ B♭5—6 C(5)—♮6
A♭ ( ) F➔

The F’s arrow of course targets the B♭ tonic, which we would expect to
support the middleground ^3 (resolving dissonant E♭) within the descend-
ing fifth-progression. The stem of bass B♭ in the middle of 9.5 sports a
prong, indicating its role as the end of one B♭–F–B♭ bass arpeggiation and
the onset of another. Yet note that an x appears above this chord among the
measure numbers, conveying that it does not in fact sound within the
composition. Sidestepping the stability of an internal tonic, Mendelssohn
forges a shortcut: instead of F<B♭>G, as graphed, F♯ is inserted after F
(inducing a D➔ surge) so that the more direct route F<F♯<G may tran-
spire. Consequently that 6-phase chord’s 5-phase predecessor has been
elided.
Though this time the G chord does not evolve into a surge, it
nevertheless proceeds (as before) to II. In the score, ninth D surmounts
the supertonic’s E♭, G, and B♭ chord members. Though listeners might
expect a descent from D to root C during the supertonic prolongation,
the C is postponed until 983 and 1013. To convey the foundational
structure as clearly as possible, that C appears beginning at measure
94 in 9.5.
Though Mendelssohn did not invent the device, certainly at an early age
he came to understand how the ^2>^1 dominant-to-tonic close of an
98 Harmony in Mendelssohn and Schumann

extended harmonic progression may be feigned at a point where a cadence

would be expected. Given the proportions of this exposition’s S, the
dominant root F achieved at 961 might reasonably yield to tonic root B♭
at 1021. The dominant’s F, A, C, and E♭ chord members have congealed
during both measures 98 and 101. Given this context, the chord of measure
102 is extraordinary: though it contains the tonic triad’s members (B♭, D,
and F), it also sports an A♭ (with a AD ♭ compound diminished fifth sounding
at the outer edges of the texture), and the dynamic level has risen. On the
one hand, it may seem that the expected ^2>^1 motion has occurred (in the
Violin III line); yet, on the other hand, the “resolution” chord is anything
but stable. My reading of Mendelssohn’s intentions is conveyed in 9.5: the
melodic B♭ at measure 102 does not serve as the goal of the broad descend-
ing fifth-progression, but instead is the internal pitch of a local C>B♭>A
third-progression that helps prolong dominant F-A-C. This B♭ holds out
against a chordal progression that first proceeds to restore the dominant
(via bass D<E♭<E♮<F during measures 102 through 109) and then to
embellish it (via bass F<G>E♭<F during measures 111 and 112). (This
dominant is the equivalent of the dominant that failed to materialize in
the similar structure at measures 34 through 37, as discussed above.) The
fourth beat of measure 112 projects the A♮ that completes the C>B♭>A♮
third-progression. The F-A♮-C-E♭ dominant harmony resolves to the tonic
(supporting the broad fifth-progression’s goal pitch, B♭) on the following
downbeat, which serves as the EEC (Essential Expositional Closure), con-
cluding S.
The exposition’s remaining measures constitute C (the closing zone),
during which the descending fifth-progression of S is reprised in the
upper register. (See 9.6.21) Unfolded thirds serve again as a key ingre-
dient: the E♭>C and B♭<D of measures 116 and 117–118, respectively,
correspond to those of measures 67 and 68. Contrasting the diversion
from bass F through F♯ to G during S (measures 91–92), the dominant
chord that supports ^4 resolves to the tonic, supporting ^3. After some
repetition, the progression concludes with V7 and I, which coordinate
with the completion of the fifth-progression (^2 and ^1). Whereas the
exposition’s first ending adds an A♭ to the B♭-D-F chord, decisively
cancelling its local impact as a tonic in B♭ Major so that it may more
broadly project V➔, targeting the movement’s E♭ tonic for the exposi-
tion’s repeat, the second ending restores the F (= ^2) from which the fifth-
progression descended, as the movement proceeds to the development.
 Mendelssohn: Octet in E♭ Major (op. 20) 99

Example 9.6 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm. 113–124.

Taking into account the differences in how we deploy Roman

numerals, the harmonic analysis in Vitercik’s fig. 7 corresponds
closely to that in my 9.5 for the passage from measure 68 through
measure 90. Yet some of the finer points of his reading warrant
scrutiny. Does the submediant actually arrive in measure 75, as he
proposes? Though the pitches G, B♭, and D all sound at 753, I
regard them as participating in an embellishment of D➔, with the
submediant arriving on the following downbeat.22 Likewise measure
84 is not the arrival point of ii.
It is common practice for an analyst to reduce a score’s rich content to a
more manageable collection of foundational pitches, forming a structural
skeleton onto which readers should be able to affix the omitted details.
I have pursued that strategy in 9.5 by showing only a 6-phase tonic, a
supertonic (eventually surging), and a dominant over the span of mea-
sures 92 through 101. The notation in my graph’s bass clef conveys a
hierarchy: the F that arrives in measure 96 is the most foundational pitch
in the vicinity, with E♭, G, and E♮ (in that order) gradually leading toward
the foreground. From that perspective, the continuation of Vitercik’s fig.
7 after the V of measure 90 with only the chord labels V/vi within
parentheses (measure 92) and I 64 (measure 96) seems to me off the
mark. I discuss the chord that he labels as V/vi in my commentary rather
100 Harmony in Mendelssohn and Schumann

than displaying it in my graph, since its role is to thwart the normative

trajectory from V to I. I would not present an analysis that accounts for
bass F and F♯ but not the G to which that F♯ resolves (Violoncello I at
923). Yet that apparently is what is being displayed in his fig. 7.23
Had the chords corresponding to my graph’s bass pitches E♭
and E♮ been taken into account, the conception that the following
bass F (at measure 96) introduces the dominant root might have
gained ascendancy in Vitercik’s conception. I go so far as to omit
the embellishing B♭ and D that initially sound above that F,
projecting (with my annotation “96–101”) that it takes six mea-
sures for the dominant chord’s members to fully fall into place.
Vitercik does the opposite, eliminating the fifth and third above
bass F, thereby elevating the B♭ and D into members of a B♭ tonic
chord. That perspective is reiterated in his fig. 8, where
the harmonic content of measures 96 through 102 is conveyed
using the two symbols I 64 and I 65 .24 Thus my reading of a broad
dominant prolongation – with root F dominating from 961 through
1124 – is contradicted by his tonic assertions. Though our disagree-
ment is more global, it includes his accepting what I refer to in my
commentary above as a “feigned” dominant-to-tonic close (measure
102) as a genuine tonic resolution.
Though I have found many correspondences among Taylor’s dia-
gram of the “harmonic structure” (fig. 2.1), his textual commentary
(a full paragraph on page 61), and my reading in 9.5, my concern
regarding his hierarchical choices persists. For example, his diagram
shows exactly three symbols for measures 84 through 93, as follows:
V/c V/A♭ g

I am curious why V/c is displayed but not c; and why V/A♭ is

displayed but not A♭.25 (He does mention both omitted chords in
his textual commentary.) In my presentation above, I convey how
the A♭ chord resides between the preceding c and the following
F.26 The latter chord is both stemmed and beamed in my 9.5, and
yet it is addressed neither in Taylor’s diagram nor in his
commentary.
 Mendelssohn: Octet in E♭ Major (op. 20) 101

The development (measures 127b–215)

The background B♭ dominant harmony (supporting ^2) that was prolonged

during the exposition’s S and C persists through the development. Though
locally stable for many measures, ultimately the B♭ chord takes on surge
characteristics (as V7), targeting the restored tonic that will inaugurate the
recapitulation.
My reading of the development proposes that the same sort of broad E♭–
B♭–E♭ fifth-arpeggiation that extends over the entire composition is repli-
cated more locally during the development, as B♭–F–B♭. This relationship
shapes the foundational model shown in 9.7. Observe how the F chord at
first retains A♭ (diatonic in E♭ Major), though a surge, wherein A♮ replaces
A♭, emerges prior to the resolution back to B♭. This chordal trajectory
supports the melodic third-progression D>C>B♭, which transpires below a
prolonged F (= ^2). As we shall see, Mendelssohn gives unexpected empha-
sis to that third’s internal C. Our discussion of how 9.7 is filled out will be
divided into five parts, each with its own example: the prolongation of the
initial B♭ chord [9.8], the trajectory connecting the B♭ chord and the F
chord [9.9], the initial prolongation of the F chord [9.10], the continued
prolongation of the F chord, including its surge and resolution to B♭ [9.11],
and finally the stabilization of the restored B♭ chord followed by the
incorporation of seventh A♭ prior to tonic E♭’s return at the onset of the
recapitulation [9.12]. Though Mendelssohn deploys multiple registers,
these models convey the essential structural features in a central register
throughout.

Example 9.7 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm.
127b–185.
102 Harmony in Mendelssohn and Schumann

Example 9.8 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm.
127b–141.

Example 9.9 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm.
127b–155.

Example 9.10 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm.
151–164.
 Mendelssohn: Octet in E♭ Major (op. 20) 103

Example 9.11 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm.
151–185.

Example 9.12 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1, mm.
185–205.

The very word “development” suggests that materials from the exposition
will recur in a new guise, allowing listeners to deepen their understanding of
various notions through fresh elaborations. The harmonic progression I5–6 II
V I was featured several times during the exposition (as displayed in 9.1, 9.2a,
and 9.5). Mendelssohn’s reliance on that trajectory persists during the devel-
opment. Its incorporation within measures 127b through 141 [9.8] resembles
that of measures 1 through 9, in that both deploy an upper neighbor to the
local tonic’s fifth (B♭<C>B♭ and F<G>F). The tonic 6-phase chord’s surge is
like that introduced in measure 77 [9.5]. Note that in some instances the
supertonic surges whereas in others it does not. (It does during the
104 Harmony in Mendelssohn and Schumann

development, measure 138.) To my ears the F>E♭>D♭ melodic third of

measures 139 through 141 recalls the content of measures 67 and 68, where
E♭>C and B♭<D are unfolded. The wobbly D♭ is of only local significance,
setting the stage for an upcoming event that will be described below.
Looking again at 9.7, observe how the departure from the initial B♭ chord
features a bass ascending fifth (B♭<F), whereas the most conspicuous event
among the upper pitches is the emergence of C, a fourth below the F that is
the focus of attention during the initial B♭-chord prolongation [9.8].
Mendelssohn elects to thrust that C into the limelight by placing it a fifth
above the initial F. Two complementary slurs enhance 9.9: the ascending
fifth from B♭ to F in the bass, and the ascending fifth from F to C in the
soprano. An intermediate arrival point rooted on D♭ comes between those
perimeter chords. The circle of fifths is deployed in making these connec-
tions. First, B♭ E♭➔ A♭➔ D♭ (with roots that are diatonic in B♭ Minor, which
helps explain why the initial B♭ chord’s D wobbles to D♭); then D♭ G⇨ C➔
F. Note especially that a chord rooted on G (supersurging D♭-F-A♭-B♮) is
deployed, instead of one rooted on G♭.27 A variance from the progression of
perfect fifths is essential if dominant root F – rather than F♭ – is to be
attained. (Compare with the opening measures of Beethoven’s Piano Sonata
in A♭ Major (op. 26), mvmt. 3, which I explore in Harmony in Beethoven, 3.6,
to see what happens when only perfect fifths are deployed in this context.)
Mendelssohn treats this ascending trajectory’s three principal chords in the
same way: the melody’s F>E♭>D♭ (beamed at the right edge of 9.8) coordi-
nates with the restoration of the initial B♭ chord; A♭>G♭>F (beamed in 9.9)
coordinates with the internal D♭’s arrival; and C>B♭>A♭ (also beamed in 9.9
and continuing onwards through G to F) coordinates with the F goal.
Through this expanded content, the voice leading of D to C (the first two
pitches of the beamed third-progression in 9.7) is filled in chromatically, as
D>D♭>C, with C temporarily positioned at the top of the texture.
What follows after that initiative has run its course helps to reestablish the
preeminence of the melodic pitch F, in that an F-to-F octave-progression
transpires during measures 155 through 164 [9.10]. Mendelssohn’s harmo-
nic progression is audacious. (Note the consecutive diminished seventh
chords during measures 159 and 160.) I propose that he has deformed a
more normative progression in which the soprano pitches D♮ followed by D♭
in my model might have proceeded instead as D♭ followed by C, correspond-
ing to IV or IV7 (with bass B♭) and V♮ (with bass C). Mendelssohn passes
over diatonic B♭-D♭-F-(A♭) in favor of B♮-D♮-F-A♭ (which I interpret as
eliding IV’s 5-phase chord in favor of its 6 phase, asserted as II➔), while E♮-
G-B♭-D♭ is an evolved state of the C-E♮-G-B♭ dominant seventh.
 Mendelssohn: Octet in E♭ Major (op. 20) 105

The reading of B♭–F–B♭ as the development’s principal trajectory [9.7] is

supported by the fact that both the B♭ and F chords are expanded via a
tonicizing I5–6 II V I harmonic progression. Yet whereas the tonicization of
B♭ Major, shown in 9.8, may deploy diatonic roots throughout, the toni-
cization of F Minor, shown in 9.11, is complicated by the fact that the F
chord’s 6-phase pitch (D♭) and the supertonic root (G) do not form a
perfect fourth/fifth. This characteristic of minor keys is responsible for the
higher incidence of chromaticism in that context, in the service of restoring
the perfect relationship: either as D♭ to G♭ (introducing the lowered super-
tonic chord), or as D♮ to G, which Mendelssohn pursues here. (One must
read the score with care at this point. Since the E♭ Major key signature is
retained during the F Minor tonicization, Mendelssohn was able to write D
noteheads at 1791 without adding naturals to the left. I have used the
natural symbol in 9.11 as a reminder of D♮’s status as a chromatic pitch.)
The circle of fifths proves useful again in connecting third-related roots.
Whereas proceeding through three fifths (as in B♭ E♭ A♭ D♭ or D♭ G C F in
9.9) results in the traversal of an ascending third, proceeding through four
fifths (as in 9.11’s F B♭ E♭ A♮ D♮) results in the traversal of a descending
third, exactly what is called for in the context of I5–6 with an unfurled 6-
phase chord. Because that third is minor (landing on D♮, not D♭), one of the
internal perfect fifths must be trimmed by a half step. Mendelssohn takes
care of that matter by proceeding from E♭ to A♮, rather than to A♭. Because
the tonic’s 6-phase chord surges (as ♮VI➔ in measure 180), the supertonic
that it targets is adjusted to become a consonant chord (with fifth D♮ rather
than diatonic D♭). Though the dominant seventh that follows might have
resolved to an F-A♭-C tonic (in F Minor), Mendelssohn concurrently
transforms that chord (now thinking more broadly in E♭ Major) into a
surging II➔. For the time being the B♭ dominant chord that concludes this
progression at measure 185 is consonant, though the development has one
more important goal to accomplish (as shown in 9.7): the conversion of the
long-prolonged dominant into a surging, tonic-targeting entity through
the addition of its minor seventh, A♭.
Recall that Mendelssohn projected the development’s internal F chord in
two distinct phases. First, it was stabilized [9.10]; and second, it was destabi-
lized at the end of a tonicizing I5–6 II V I harmonic progression [9.11].
Through the latter, F becomes F➔, targeting the restored B♭ dominant. The
material near the end of the development likewise projects two distinct phases
for the B♭ dominant chord. (Both phases are shown in 9.12.) First it is
stabilized (measures 185 through 196), with two and then two more neigh-
boring notes embellishing the B♭ chord’s members. Then, in a masterful nod
106 Harmony in Mendelssohn and Schumann

to continuity, Mendelssohn proceeds through yet another tonicizing I5–6 II V

I harmonic progression that now concludes with B♭ becoming B♭➔, targeting
the E♭ tonic that will emerge at the onset of the recapitulation. Consequently
all three of the principal chords displayed in 9.7 undergo a similar harmonic
elaboration, with the latter two evolving into surges. The B♭ chord’s 6-phase G
emerges in measures 196 in a texture of only two sounding pitches. The II➔
that follows is formed from three pitches. Skipping over the next three
measures for now, observe that V➔ is fully in place at measure 201. The
consonant B♭ chord that might have emerged if the embellishing 64 pitches of
measure 202 had resolved conventionally (downwards to 53 ) is replaced by a
surging B♭➔ chord through the upward continuation from 64 to 75 (with a
ninth added as well). Thus the final structural element conveyed in 9.7 – the
addition of the B♭ dominant chord’s seventh – has at last fallen into place.
The sonority of measure 198 (and prolonged through measure 200, with
the Cello II’s E♭ serving as a passing note) reveals Mendelssohn’s adept
handling of diminished seventh chords. At first one might interpret measure
198 as an embellishment of the preceding C➔ chord. (As an experiment,
play measure 197’s C➔ chord at the piano, then the chord of measure 198,
and then the C➔ chord again. In this context the most suitable spelling for
the diminished seventh chord would be B♮-D-F-A♭.) But since the chord in
fact resolves to an F➔ chord in measure 201, in retrospect one might
perceive the bass F of measure 198 as an asserted root, with the other three
pitches forming a common-tone diminished seventh chord (which might be
spelled as F-G♯-B♮-D, as is conveyed within parentheses in 9.12). To
acknowledge the mehrdeutig mystery of measures 198 through 200, I have
placed the V7 harmonic label at measure 201, though of course the stemmed
bass F of measure 198 begins the process of its assertion.
Coordinating with this harmonic agenda, Mendelssohn deploys a C
neighboring note multiple times. (Three Cs are flagged in 9.12.) When
B♭ is restored as the fifth of the tonic E♭ chord in measure 216, it again will
proceed to neighbor C, even deploying a C>C♭>B♭ chromatic return like
that of measures 204 and 205. (The exposition equivalent of this recapitu-
lation material may be viewed in 9.1, measures 1 through 8.)

Whereas Vitercik and I agree that F Minor offers the development’s

principal contrast to B♭ Major, our readings of the path Mendelssohn
forges between B♭ and F differ. Before exploring those alternative routes,
 Mendelssohn: Octet in E♭ Major (op. 20) 107

I would like to propose that there is a typographical error in his fig. 9.28
The v numeral (in B♭ Major) above measure number 146 does not
concur with what transpires in the score at that point. That blemish
would disappear if the number 149 replaced 146, since a C-F-A♭ chord
sounds at that point, confirmed by root-position F-A♭-C two measures
later. In that way his “prolongation of ii” commencing at measure 137
would continue through the surging C chord of measures 147 and 148.
(To get the gist of his reading, as I think he intends it, locate all of the
chords within a circle of fifths between those points in my 9.8 and 9.9:
C ➔ F➔ B♭ E♭➔ A♭➔ D♭ G⇨ C➔.29)
Alas, his C-to-C prolongation (measures 137–148) cannot coexist
with my B♭-to-B♭ prolongation (measures 127b–141, as projected in
9.8). The question thus becomes whether the supertonic harmony in
measure 137 proceeds (via a surge) directly to the major dominant, as
proposed in 9.8, or instead extends to the minor dominant of mea-
sures 149/151. I contend that all these fifth-related chords congeal
into three segments. First, within the development’s initial harmonic
♭
progression prolonging B♭ Major, II ➔ V➔ I 3 provides a founda-
tional model, as

C➔ F➔ B♭
Then this trajectory is hoisted up a third to become

E♭➔ A♭➔ D♭
The next ascending-third hoist results in

G⇨ C➔ F

As was noted in my commentary, the three arrival points (the

underlined pitch names) coordinate with descending thirds in the
melody: F>E♭>D♭, A♭>G♭>F, and C>B♭>A♭. Thus my reading seg-
ments nine chords into three groups of three, in coordination with
perceptible waves of activity in the score. If my hunch regarding
a typographical error is correct, Vitercik instead is proposing an
undifferentiated eight-chord prolongation from C to C, followed by
a resolution to F.
Though up to now figures brimming with Roman numerals have
complemented Vitercik’s commentary, the latter part of the develop-
ment is addressed with one astonishing assertion: that the movement’s
108 Harmony in Mendelssohn and Schumann

E♭ Major tonic is reinstated at measure 171. He certainly understands

that in its context – still over forty measures prior to the onset of the
recapitulation – this would be a very odd thing for a composer to do.
But instead of reconsidering his proposal, he proceeds to defend
it from several angles: asserting that the cadence at measure 177
“establishes E♭ major with absolute clarity,” suggesting that it is “a
distillation of the tonic-supertonic relation” and that it offers a “neu-
tralization of the modulatory impetus that has shaped both the exposi-
tion and the development” (p. 89), etc. In contrast, even before reading
Vitercik’s analysis closely I had written instructions to my graph-setter
to use “smaller noteheads” for the chords in the vicinity of measures
171 through 178 in 9.11. What for Vitercik is one of the two or three
most defining harmonic events in the entire movement is for me an
internal element along the path from the tonic in tonicized F Minor to
its 6-phase D♮ (prior to the G, C, and F that round out F’s tonicization,
as is conveyed in 9.11). In my reading the development’s F tonicization
is anchored by fifth-related B♭’s prolongation both before (measures
127b through 141) and after (measures 185 through 215). Vitercik
instead subjects the F chord to a non sequitur descent to E♭, an
interpretation that I reject.
Taylor analyzes the development only through the arrival of F Minor.
His first statement on the topic reads as follows: “C minor, the relative
minor comparatively absent from the exposition, will become in turn
the focus of the first part of the development” (p. 61). While it is true
that a C minor chord sounds only a few times during the exposition (see
measures 19 and 51 in the context of E♭ Major and measure 86 in the
context of tonicized B♭ Major in my 9.1, 9.4c, and 9.5), only one C
minor chord – serving not as the relative minor of tonic E♭ but as the
supertonic in tonicized B♭ Major – appears in any of my graphs of the
development (see measure 137 in 9.8). Taylor certainly would respond
that we are in the key of C Minor during measures 131b through 137
(p. 71), thus making it “the focus” at that point and more significant
than the three separate deployments of the chord during the
exposition.30 I do not want to enter into a debate about whether we
are “in” the key of C Minor or “on” B♭ Major’s C minor supertonic
chord. What is important to understand is that “the focus of the first
part of the development” is a broad prolongation of its initial B♭ chord
through measure 141, following exactly the same I5–6 II V I harmonic
 Mendelssohn: Octet in E♭ Major (op. 20) 109

trajectory (conveyed in 9.8) that was and will continue to be traversed

often throughout this movement. Only after completing that trajectory
does Mendelssohn begin the process of moving away from B♭ to the
development’s interior F region.
Taylor’s ex. 2.11a is annotated by three labels – Cm, B♭m 64 – 53 , and
D♭ 64 – during the span of measures 137 through 143. Through compar-
ison with my 9.8 and 9.9 one will note two critical points of dis-
agreement. First, his B♭m 64 label (measure 139) conflicts with my
interpretation of what the composition is doing at that point: an F
dominant chord is being embellished, rather than a B♭ tonic chord
being anticipated. (His 64 and 53 symbols do not convey a resolution of
the former into the latter over root F, but instead acknowledge two
manifestations of a B♭ chord, first in its second inversion and then in its
root position. The same issue recurs with the D♭ 64 label at the right edge
of his example.) Second, I interpret that B♭ chord, once it does arrive
at measure 141, as a restoration (with wobbly third) of the B♭ chord
from the development’s onset. The snippet selected for the example
(which bears the caption “climax of development section”) curiously
extends from the middle of one initiative (the C supertonic chord
within my 9.8) to the middle of another (the A♭ predecessor of
D♭ within my 9.9).

The recapitulation and the coda (measures 216–280

and 280–318)

Whereas an exposition’s P and S will transpire in different keys (in this

movement, the tonic and the dominant), during the recapitulation both
generally will project the tonic. If one considers that a coda (if present) will
add further tonic reinforcement, then a concern might justifiably be raised
that too broad an expanse of tonic will occur starting at the onset of a
recapitulation.31 In this movement that concern is exacerbated by the fact
that the exposition TR’s MC deploys a dominant chord, which in a
transposition to the tonal environment of the recapitulation would add
yet another prominent point of tonic articulation.
My speculation that a composer might be concerned about just how
much tonic reinforcement may reasonably be thrust upon the listener
during a recapitulation and coda may help one come to terms with the
110 Harmony in Mendelssohn and Schumann

considerable contrast in content between this movement’s exposition and

recapitulation. The former contains five principal articulation points: there
are two tonic cadences during P (corresponding to the endpoints of 9.1 and
9.2); a medial caesura on the dominant (the goal of 9.3c); and cadences
reinforcing the tonicized dominant at the ends of S and C (9.5 and 9.6). A
full realization of the exposition content during the recapitulation (wherein
S is transposed into the tonic key) would result in all five of these articula-
tion points projecting a tonic harmony. Followed by a tonic-reinforcing
coda! Had the exposition’s MC instead been on II♯, then there would be
some contrast in the recapitulation context (where the II♯ would be
transposed to V). But that is not the case.
{Curly brackets are used to contain my commentary on the details of
several crucial compositional decisions that come to the fore from the onset
of the recapitulation P through the arrival of its S. How Mendelssohn
resolves these issues in the 1825 manuscript differs markedly from what
appears in the later published score. Because the facsimile of the manuscript
lacks page numbers, let’s first get oriented: beginning with 1 for the page
containing the opening measures, the recapitulation’s P begins on page 21,
measure 4, while its S begins on page 25, measure 5.}
Mendelssohn saves the day by turning his uncommon choice of exposi-
tion MC into an advantage. In that the recapitulation’s MC chord would be
I, we in fact do not even require a TR, since the I at the end of P may serve
in place of TR’s MC. And that’s not all: to reduce the tonic saturation, we
might also eliminate the second of P’s two parts. During the recapitulation
Mendelssohn proceeds so that the endpoint of 9.1 (an E♭ chord) doubles as
the endpoint of 9.3c in a tonic context (also an E♭ chord), followed by the
onset of 9.5 in a tonic transposition.
{Recapitulation material corresponding to the exposition TR does not
appear in the manuscript version. The second part of P (corresponding to
9.2), which begins on page 22, measure 18, is heavily reworked to give it the
more agitated character of a TR, but its cadence (which fulfills the role of
MC) corresponds to the exposition’s measure 37, not to a tonic transposi-
tion of its measure 68.}
This breathtaking adjustment – the elimination of almost fifty mea-
sures – is made less abrupt through Mendelssohn’s adumbration of melo-
dic material from S toward the end of P (beginning in measures 233) and
by the extra measure (in coordination with a new melodic line in the Violin
I part and the instruction to slow down) inserted between P’s tonic cadence
(serving here as a pseudo-MC) and the onset of S. (Compare measure 68
and measures 236–237.)
 Mendelssohn: Octet in E♭ Major (op. 20) 111

{In the manuscript (page 25, measure 5), a cadence on tonic E♭ precedes
the onset of the S melody. In the published score, Mendelssohn places the
first phrase of that melody over dominant root B♭ (above which a 64
embellishment presents the pitch classes of tonic E♭-G-B♭), prior to the
cadence that concludes P. (See measures 233ff.) Though I set the dividing
line between P and S after the tonic root E♭ arrives in measures 236, some
listeners might propose that S commences at measure 233, prior to the P
material’s cadence point, resulting in a dovetailing of the recapitulation’s
two main components.}
With no TR, the background ^5 of P is juxtaposed with what in the
exposition was the ^2 of S, which through transposition into the tonic key
becomes a reiteration of ^5 . Consequently the exposition S’s descending
fifth-progression from ^2 (F to B♭ in 9.5) becomes ^5 > ^4> ^3> ^2> ^1 (B♭ to E♭)
in the tonic key – exactly what we would expect to occur during a
recapitulation that follows after an interruption on ^2, which has held
sway from the exposition TR’s MC through the end of the development.
In comparing 9.1 and P’s recapitulation trajectory, observe how the
stepwise C-to-C bass of measures 16 through 19 is replicated during
measures 227 through 231, now with a somewhat stronger tonic focus
before the 6-phase C-E♭-G chord solidifies. (Mendelssohn is integrating
the bass C<D<E♭ of measures 12–13 and the C<D<E♮ of measures 16–17.)
The A♭ bass that follows at 194 is replaced by chromatic A♮ during measure
232. Could this serve as compensation for the fact that the continuation of
P from the exposition (after the cadence at measure 21) is omitted?
{Among other adjustments, the manuscript’s bass A♭ at page 22, measure
13, beat 3, becomes A♮ during the published score’s measure 232.} At
least that material’s A♮<B♭ approach to the dominant (274 to 281) is
incorporated!
Though an exact transposition from dominant B♭ Major into tonic E♭
Major would have sufficed for S, Mendelssohn’s recapitulation S is sub-
stantially reworked (9.13a), while of course retaining the foundational
descending fifth-progression. What at first appears to be a varied sequen-
tial trajectory to prolong tonic E♭, commencing with E♭5–6 F in measures
240–241, eventually veers off to E♭’s 6-phase C chord (at 2473), though E♭ is
restored (surging) in measure 253. (The local harmonic progression of
these measures is explored in 9.13b.) The omission of IV5 (replaced by an
evolved state of IV’s 6-phase chord, asserted as II➔) in measures 255–256
is reminiscent of what transpired during measure 159 from the develop-
ment (9.10). The bass C<D<E♭ echoes what occurred earlier during
112 Harmony in Mendelssohn and Schumann

P (measures 12–13 and 227–228), displayed here as part of a filled-in B♭<E♭

fourth. Near the cadence the Violin I melody soars upwards to pitches that
serve as cover tones. The melodic G<A♭>G>F (from ^3 through its incom-
plete upper neighbor to ^2) proposed for measures 264 and 265 is spread
between the Violin II and (an octave lower) Viola II parts. (Because the F
does not sound in the upper register, it appears within parentheses in 9.13.)
The ESC (Essential Structural Closure) – which corresponds to the exposi-
tion’s EEC – occurs at measure 266.

Example 9.13 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1 (a) mm.
68/129–266; (b) mm. 237–253.

The recapitulation C is similar to that of the exposition. Again a des-

cending fifth-progression is traversed, with a focus on the third from
 Mendelssohn: Octet in E♭ Major (op. 20) 113

B♭ down to G during measures 268 through 276, and from G down to E♭

during measures 276 through 280.32 That descent’s initial B♭ reinstates the
Kopfton pitch after the background descent has already run its course.
During the coda B♭ again serves as the initiation point for a descending
fifth-progression, though in that case the B♭ is not freshly articulated: from
C’s tonic cadence (with melodic E♭) at 2801, the next melodic event is A♭
(already a step below the initiating B♭) at 2804, which is shown in 9.14a as
emanating from the B♭ earlier in C.

Example 9.14 Analysis of Mendelssohn: Octet in E♭ Major (op. 20), mvmt. 1 (a) mm.
268–313; (b) mm. 280–290.

The F minor supertonic chord introduced at 2804 ultimately evolves into

surging A♮-C-E♭-G♭ (measure 293), though first an excursion to F’s upper-
third chord, A♭ major, is pursued. (Observe that this A♭-C-E♭ chord is
labeled as II 7• in 9.14a.) The process through which supertonic F-A♭-C and
this A♭-C-E♭ chord are connected is outlined in 9.14b. Mendelssohn has
granted himself more space in which to work by inverting the modest
ascending third (F<A♭) into a broad descending sixth (F>A♭). Omitting the
E♭ chord that would evoke the initial tonic, he proceeds downward by step:
from F through D♭, C, and B♭ to A♭. (A comparison of this model with the
114 Harmony in Mendelssohn and Schumann

chords that Mendelssohn actually employed in his composition reveals

some potent dissonances. For example, the model’s F-A♮-E♭-G♭ chord at
2883 sounds without root F but with a B♭ anticipation of the following B♭
chord’s root.) As with most circular progressions, here one of the descend-
ing fifths is trimmed by a half step to prevent a spiral out of the key.
Mendelssohn handles that moment with aplomb. The diminished fifth
occurs between D♭ and G. Observe how, unlike the other roots in the
circle, D♭ is delayed during measure 287. (I have taken the liberty of
spelling Mendelssohn’s D♮ as Eº in 9.14b to convey how the F-A♭-C♭-Eº
chord emanates from root D♭.) By the time D♭ arrives, the circle’s next
chord has emerged. In this instance that is not a problem, because both the
D♭ and the G chords are conveyed by the same four pitch classes: D♭-F-A♭-
C♭ versus B♮-D♭-F-A♭! (This is an example of the dominant-seventh/
augmented-sixth enharmonic equivalence taught to every college music
major.) Consequently I have applied a collision bracket above the D♭ and
G⇨ labels below these chords in the model.
Once bass A♮ resolves to B♭ at measure 294, the piece enters into an
extended prolongational routine. The shared pitch content between the
cadential 64 on which we have just landed and the tonic harmony gives
Mendelssohn the freedom to reiterate I-to-V trajectories (often with an A♭
and/or A♮ coming between the tonic third, G, and the dominant root, B♭),
without actually resolving to the tonic, an event that finally occurs at 3131,
concurrent with the melodic F>E♭ that concludes the coda’s broad fifth-
progression. A few moments later the movement is over.

Beginning at the onset of the recapitulation, Vitercik’s focus is dominated

by comparisons of the score’s content with the corresponding material
from the exposition and by how that content differs from the manuscript
version. As a consequence, harmonic issues are not as central a concern
and are for the most part explored within the prose commentary rather
than with figures like those that complemented his earlier discussions.
Below I quote and assess his interpretation of three passages.
“The dominant pedal finally shows up in m. 251” (p. 93). My reading
in 9.13b proposes that the B♭s of measure 251 are not affiliated with the
dominant, which arrives instead at 2523. Though Mendelssohn’s trans-
fer of the ascending chromatic line from the Violoncello I to Violoncello
II part at B♭ gives it a visual boost in the score, in sound the chromatic
line proceeds through that B♭, not to it.
 Mendelssohn: Octet in E♭ Major (op. 20) 115

“The chord in m. 235 [should read 253] is taken as an enharmonic

alteration of V of ii that leads through II 43 (mm. 255–56) to V 65 (mm.
257ff.)” (p. 94). Vitercik and I agree that the succession from G-B♭-D♭-
E♭ in measures 253 and 254 to C-E♭-G♭-A♮ in measures 255 and 256 is
unusual and consequently warrants some special analytical effort. My
solution is to propose a three-chord trajectory whose middle chord is
elided: from a surging tonic to IV (A♭-C-E♭), which in this case is
replaced by rather than followed by its surging 6-phase chord (rooted
on F). Vitercik’s remark is cryptic. My best guess regarding what
he means is that the G-B♭-D♭-E♭ chord stands for the more highly
evolved state G-B♭-D♭-F♭, which could be reinterpreted enharmonically
as E♮-G-B♭-D♭, a chord that targets II.
“In m. 280, the cello quietly swings the harmony around to F minor . . .
and F minor gradually dissolves back into the tonic under a leisurely
upper-voice ascent from a♭´´ through c´´´ (m. 290) to e♭´´´ (m. 292)”
(p. 95). Admittedly the level of precision that I endeavor to attain in my
analytical writing may seem excessive at times. Yet there is something
important in my distinction between “F minor” and “F Minor.” In my
usage, the former refers to an F-A♭-C chord, whereas the latter refers to
the key of F Minor. In the passage quoted, I am not sure whether chords
or keys are intended. The word “harmony” would suggest an F minor
chord, but dissolving “back into the tonic” seems to refer to the key of E♭
Major. By consulting my 9.14a, one may observe what I think is most
crucial to say about the F minor chord: that it eventually surges and leads
to the dominant. Whether tonicized (which might be Vitercik’s intention:
F Minor reverting to E♭ Major) or not, once root F sounds, root E♭
is an unlikely immediate successor.
In that Taylor’s assessment of the recapitulation and coda does not
engage with any of its harmonic features, we may here bring this chapter
to a close.
 10 Mendelssohn: Song without Words in F Major
(op. 85, no. 1)
in response to allen cadwallader

Professor John Rothgeb’s fall 1976 course on Schenkerian analysis at the

Eastman School of Music was my initial encounter with a subject that has
become a central focus of my career. I often sat beside another graduate
theory major, Allen Cadwallader, who also became enthralled by the topic.
In the intervening forty years we both have published extensively on
Schenkerian topics, and we both have created textbooks on Schenkerian
analysis. One of his essays assays two of Mendelssohn’s Songs without
Words.1 That in F Major (op. 85, no. 1) is explored in this chapter. Whereas
my approach to Schenkerian analysis is infused with a special attention to
harmony, Cadwallader’s essay pays special attention to motives, acknowl-
edged through annotations in his graphs.

Stanza 1 (A1 section)

Mendelssohn’s song is built from three broad melodic descents from ^5 over
the course of four stanzas. Stanza 1 (measures 2 through 9) offers a straight-
forward foundation that is subjected to elaborations and modifications later in
the song. In stanza 1 and in stanzas 2 (measures 10 through 17) and 3
(measures 18 through 25), the line descends by step only as far as ^2, with
closure on ^1 reserved for stanza 4 (measures 30 through 39). This arrangement
suggests an A1 B A2 formal organization. The first and fourth stanzas – A1 and
A2 – together offer a convincing antecedent/consequent binary construction,
into which stanzas two and three – which constitute B – are inserted.
The prong attached to bass F’s stem at measure 5 in 10.1 signals the
deployment of a double bass arpeggiation during stanza 1, with distinct
dominants supporting ^4 and ^2. Kopfton ^5 is prolonged via upper neighbor
D, a member of an embellishing 64 chord, during measures 2 and 3. The
progression proceeds next to V7, with the tonic’s local downward C>A>F
arpeggiation matched by the dominant’s upward E<G<B♭ in the melody.
The BE ♭ diminished fifth resolves to the tonic’s AF third in measure 5, where
116 the first F–C–F bass arpeggiation concludes.
 Mendelssohn: Song without Words in F Major (op. 85, no.1) 117

Example 10.1 Analysis of Mendelssohn: Song without Words in F Major (op. 85, no. 1),
mm. 2–9.

Since dominant root C will serve as the phrase’s harmonic goal (as
support for ^2), an intervening harmony would be welcomed between the
I of measure 5 and the goal V. In this case an inverted II7 sounds during
82. Its bass, B♭, proceeds by step to the dominant’s C. The harmonic
succession from I to II is projected as I5–6 II. The 6-phase chord
(unfurled so that D sounds in the bass) is expanded via its own embel-
lishing chord, thus establishing a relationship between the initial ^5 (C
with neighbor D) and the 6-phase chord’s ^3 (A with neighbor B♭).
Instead of presenting a 64 chord in measure 6 (equivalent to that in
measure 2), Mendelssohn adds E to G and B♭ above bass D, thereby
replicating in a new context the E-G-B♭ dissonance that resolves to F-A
during measure 5. A more decisive incomplete neighbor B♭ in measure 8
follows the neighbor B♭ of measure 6 (repeated in measure 7). It coin-
cides with the supertonic’s arrival and precedes the melody’s descent to
G at 91.

Cadwallader’s equivalent of my 10.1 appears within his ex. 1.4 (through

measure 9). The dimensions of his graph indicate one major difference
in our viewpoints: A1 corresponds to only the first stanza in my account,
but to the first and second stanzas in his. That distinction will be
discussed later, when we explore the trajectory that extends from mea-
sure 10 through measure 25.
118 Harmony in Mendelssohn and Schumann

If Cadwallader were to convey his conception fully, the Roman

numeral VI at measure 6 would read as VI(7), since he proposes that
the Kopfton C extends into measure 8, where a sixteenth-note C des-
cends to B♭ during beat 2. (That C is relocated to beat 1 in Cadwallader’s
graph.) My ear has trouble maintaining such a C against the sounding
B♭ and D of measures 6 and 7. I give greater weight to the upper-
neighbor motive introduced in measures 2 and 3 and here transported
from a C context to an A context. Having achieved A in the melody of
measure 5 via a passing descent from B♭, a mesmerizing neighboring B♭
recurs during beat 2 of the next three measures. Mendelssohn had the
good sense not to present each of those B♭s in exactly the same way: that
of measure 8 thankfully is embellished by a local appoggiatura, as C>B♭.
In my view that C has nothing to do with Kopfton ^5 .
I propose instead that the structural ^4 sounds in measure 4, with ^3
secured already in measure 5. This trajectory corresponds to one that
Schenker shows in Free Composition, fig. 157 (the second staff),
where the pitch that corresponds to Mendelssohn’s B♭ in measure 8
is explicitly labeled as a neighboring note. Because Cadwallader has
rejected the earlier ^4 as a component of the stanza’s principal descent,
he has little choice but to embrace the local A near the end of measure
8 as a structurally deep pitch (thus attached to the beam in his graph).
In my view that A instead is an event of only local impact, filling in
II’s BG♭ third, as is concurrently evoked by the slur in Cadwallader’s
graph.
Cadwallader’s harmonic analysis places I, VI, II, and V all at the
same level. I prefer to convey such a progression as I5–6 II V to
acknowledge II as the principal intermediary between I and V. That
mode of presentation also would prevent readers from entering into an
examination of Cadwallader’s stem lengths, flags, and slurs in an
attempt to come to terms with his assessment of the hierarchical
relationships. (Does the fact that B♭’s stem is shorter than D’s have
any meaning? Is VI or II hierarchically deeper?) In particular, I miss
the S-shaped slur that often would be used to connect bass F at
measure 5 (or measure 2) and bass C at measure 9, intersected by a
slur from flagged B♭ to C.
 Mendelssohn: Song without Words in F Major (op. 85, no.1) 119

Stanzas 2 and 3 (B section)

In that measures 10 through 13 resemble measures 2 through 5,

listeners likely will expect that measure 17 will serve as an important
harmonic arrival point: perhaps a PAC on I, perhaps a repeat of
measure 9’s HC, or perhaps something different. For example, the A
chord (with wobbly third C♯) during 141 could proceed onwards to
dominant C (perhaps with bass B♭ intervening), either as a goal or
continuing to tonic F. Yet the stanza instead refuses to be rushed
along on its course. It turns out that the trajectory initiated by the A
chord’s sounding spills over into the third stanza (beginning in mea-
sure 18), with a cadential arrival postponed until measure 25. In that
luxuriant context the A chord’s presentation expands into a toniciza-
8–7
tion (as A Minor: I➔ IV5–6 V 6–5 I) during measures 14 through 17. The
4–♯3

representation of those measures by only an A minor chord in 10.2

conveys my conviction that diatonic A-C-E-A is the principal event, with
the initiating A-C♯-E serving as a surge targeting A Minor’s diatonic IV
within the tonicization.

Example 10.2 Analysis of Mendelssohn: Song without Words in F Major (op. 85,
no. 1), mm. 10–25.
120 Harmony in Mendelssohn and Schumann

One ought not to assume that a tonicization necessarily signals struc-

tural depth. The A minor chord (transformed into diminished A-C-E♭
during measure 18) ultimately plays a connective role within a 5–6 shift
from F-A-C to D-F♯-A-C (as shown in 10.2, measures 13 through 19),
initiating a sequential progression that will be explored below.2 More
broadly, tonic F Major is being expanded via a harmonic progression
(similar to that which projects the tonicization of A Minor) that proceeds
as I ➔ IV5–6 V7 I between measures 13 and 23. Note especially the voice
exchange in the outer voices from the beginning to the end of that
trajectory: AF  FA . Whereas a tonic 5–6 shift often comes between I and II
(as in stanza 1), it is also common to proceed to the tonic’s first inversion
followed by an inverted supertonic, resulting here in a bass ascent from A
(measure 23) to B♭ (measure 24), which continues to the dominant’s C
(measure 25).
Though the D➔ chord leading to G in measures 19 and 20 initially
might seem like a variant of the D-to-G trajectory of measures 6
through 8, Mendelssohn ultimately pursues a much difference course.
The melody’s B♭ at 201 is not an upper neighbor to A (as was that of
measure 8), but instead a passing note. The tonic expansion’s coher-
ence stems in large part from an ascending sixth-progression (beamed
in 10.2) from A (= ^3) to F, supported in its initial stages by a segment
of an idiosyncratic ascending 5–6 sequence whose 5-phase A♭-C-E♭ chord
is elided.3 (That chord’s C sounds at the end of measure 20, but G does
not ascend to A♭. Thus the chord is displayed within parentheses in the
graph.) Though usually the progression of 5-phase chords guides such a
sequential ascent, here the omission of one of the 5-phase chords
strengthens the bond between the initial 5-phase chord, F-A-C, and the
third 6-phase chord, F-A-C-E♭, resulting in the projection of a tonic
♭
prolongation culminating in a surge (conveyed as I8– 7 within the har-
monic analysis).
From stanza 1 we understand the principal melodic event during
the harmonic succession from I to II to be A (= ^3) followed by its
incomplete upper neighbor, B♭. (See measures 5 through 8 in 10.1.)
Even though a broad tonic prolongation extends from measures 13
through 23, the A<B♭ second recurs during stanzas 2 and 3, with the
ascending sixth A<B♭<C<D<E<F followed by the descending trajectory
F>D>B♭. A local passing note, A, connects incomplete neighbor B♭
and the stanza’s G (= ^2) goal.
 Mendelssohn: Song without Words in F Major (op. 85, no.1) 121

In comparing 10.1 and 10.2, one notes that in broad terms the two
regions are nearly identical. The deployment of open noteheads in both
graphs conveys a sense of redundancy – that the antecedent structure of A1
recurs during B. The contrast that one expects within a ternary form
derives principally from the broad expansion of the tonic during measures
13 through 23, which replaces the 5–6 shift of the A1 presentation. Some
analysts might interpret the song instead as a binary form with a repeated
(and expanded) A1.
A four-measure retransition (26 through 29) follows the dominant’s
arrival. Twice a 64 embellishment of the dominant (deploying wobbly A♭
rather than diatonic A) sounds, reminiscent of the tonic embellishment of
measures 2 and 3. The presentation of this content here sets the stage for
similar material during the coda (measures 40 through 50). Though a B♭
added to dominant C-E-G at 292 appropriately instills a sense that the
initial tonic of A2 is imminent, its sounding in the bass (thus projecting the
dominant in 42 position) may be of concern to attentive listeners, since it
suggests a tonic resolution in 63 position, not the 53 position that initiated A1
(measure 2). One should expect that Mendelssohn has something special in
mind for the onset of A2.

The mediant is perhaps tonal music’s least understood chord.

Cadwallader presents two conflicting views of how its presence within
measures 14 through 17 relates to the perimeter tonic of measure 10 and
dominant of measure 25 (the endpoints of what I regard as the move-
ment’s B section). In his exx. 1.4 and 1.5a, mediant root A is one of only
three beamed open-notehead bass notes over that span: F–A–C. This is a
high-profile role for the mediant, one that confirms its name, since the
chord is interpreted as mediating between the phrase’s tonic and domi-
nant endpoints. (Cadwallader states: “the tonicization of the mediant
articulates the motion from I to V, producing two sections before the
reprise” (p. 15).) Yet in ex. 1.5b a contrasting interpretation of the
mediant appears. Bass A no longer is displayed as an open notehead,
with a downward stem, or attached to a beam. Even the III numeral is
placed within parentheses and moved out of the way. Here IV – not III –
is projected as the principal intermediary between I and V.
♭
Cadwallader’s I8– 7 IV5 harmonic analysis corresponds exactly to what
I show during that region in 10.2.
122 Harmony in Mendelssohn and Schumann

Despite that similarity, my reading of the mediant contrasts both of

Cadwallader’s proposals. In the Harmonielehre segment of my Harmony
in Beethoven I show how the mediant may emerge as an intermediary
between the tonic’s 5- and 6-phase chords, essentially as an upper fifth
(perhaps surging) to the latter. (See 5.5 in that volume, where C➔
comes between tonic A♭-C-E♭ and its 6-phase chord, F-A♭-C.)
Mendelssohn’s A-C-E chord (measure 17) is progressing neither to
the C-E-G dominant chord of measure 25 nor to the F-A-C-E♭ surging
tonic chord of measure 21, but instead to the sequence’s D-F♯-A-C 6-
phase chord of measure 19. An upward trajectory commences with bass
F in measure 13: F D➔ G E♭➔ …, not an A>G>F bass descent, as
Cadwallader displays during measures 17 through 21 in his ex. 1.5b.
Cadwallader and I seem often to be at odds regarding the distinction
between passing and neighboring notes. For example, since in my view
the Urlinie descends from Kopfton C through B♭ to A during measures
10 through 13, I regard the B♮ of measure 16 as a neighbor to A
(reminiscent to the B♭ neighbors during A1) rather than as a passing
note from C. (Cadwallader proposes the maintenance of Kopfton C
throughout the mediant tonicization, despite the C♯ and then D that
sound in lower registers.) Then he asserts that soprano D in measure 22
serves as a neighbor to Kopfton C, contrasting my passing-note con-
tinuation from D through E to F (as beamed in 10.2). With two
crescendo markings over the span from piano at measure 14 to forte at
measure 21, A-C-F at 222 is the final chord before a diminuendo sets in
and before the broad melodic trajectory shifts from ascending to des-
cending. In my view that chord concludes an eleven-measure tonic
prolongation; in Cadwallader’s it is an assemblage of neighboring (N)
and passing (P) notes within the domain of the subdominant. I regard
an F<A bass, projecting the tonic in root position and then first inver-
sion, as an ideal way to set the stage for the onset of a first-inversion
supertonic, here with bass A ascending by step to B♭ over the measures
23|24 bar line.
The stub of a beam that Cadwallader applies to mediant A’s stem in
his exx. 1.4 and 1.5a conveys a conception that allows him to interpret
the juncture between stanzas at measures 17|18 as the dividing point
between the A1 and B formal sections. My interpretation of how that
mediant functions argues against such a division point at that spot. (See
the P and CP annotations within an F>D third during measures 13
 Mendelssohn: Song without Words in F Major (op. 85, no.1) 123

through 19 in my 10.2. I regard that line as indivisible.) Consequently I

propose that the B section commences at the onset of the second –
rather than the third – stanza. In this view the work takes on a more
overtly cyclic character, with each of three sections beginning on ^5 and
descending either to ^2 (A1 and B) or to ^1 (A2).
A common voice-leading pattern transpires during the retransition
(measures 26 through 29). It may be represented in figured bass as

8 – 7 – ♭6 – 5
3 – – – 4 – 3

From this perspective the 7 should be interpreted as an event of local

significance, fully discharged by the continuation of the descending line.
Cadwallader instead endeavors to prolong that 7. (Note the slurred
B♭>A♭>G third in his ex. 1.5a and the arrow transferring the B♭ to the
bass in measure 29.)

Stanza 4 (A2 section)

Outer voices A and C at 301 provide a sense of tonic restoration even if the
chordal interior projects an evolved state of the tonic’s 6-phase chord.
Mendelssohn takes advantage of the fact that F-A-C and F♯-A-C-E♭ share
two pitch classes. A collision of those two distinct structural moments results
from providing a context in which F-A-C is expected (to resolve the V➔ of
292), while the progression concurrently proceeds with VI➔. A bracket, my
symbol for a collision, annotates I5–6 in the analysis shown in 10.3. Though
A2’s melody begins exactly like A1’s, the chordal support takes on a more
8–7
overt harmonic aspect, as I5–6 II V 6–5 …, with the expectation of a resolution
4–3

to I in measure 33 (corresponding to measure 5). That outcome fails to

materialize, due to the persistence of bass C. By backtracking (with measures
34–35 offering a replacement for measures 32–33 as the stanza’s third and
fourth of eight measures), Mendelssohn gives himself a second chance, this
time forging a new path toward the tonic’s 6-phase chord: whereas C➔ F
followed by D transpires during measures 4 through 6, now C➔ is replaced
by A➔ (in the evolved state C♯-E-G-B♭ at 351). Thus the tonic’s 6-phase
chord D-F-A arrives at 352, in the stanza’s fourth measure (taking into
account the backtracking) rather than its fifth measure (as in A1).
124 Harmony in Mendelssohn and Schumann

Example 10.3 Analysis of Mendelssohn: Song without Words in F Major (op. 85,
no. 1), mm. 30–39.

The succession from the tonic’s 6-phase chord to the supertonic during
measure 8 deployed only diatonic pitches. During measures 35 through 37
a surge emerges, incorporating the D chord’s major third F♯ and minor
seventh C (at 371). The two forms of the 6-phase chord are connected by
passing notes: D<E<F♯ in the bass, F<G<A in the chordal interior, and
A<B♮<C in the soprano. Consequently a local descending second (C>B♭)
introduces the incomplete neighbor, in a more potent manifestation of the
C> ♭
B appoggiatura from measure 8. Whereas an unaccented passing note A
comes between incomplete neighbor B♭ and G (= ^2) during measures 24
and 25, in measure 38 that passing note is featured more prominently. (It
“belongs” on the downbeat even if in the melody it is delayed until beat 2.)
The structural ^2 holds forth for only a sixteenth-note value at the end of
measure 38, prior to the PAC on ^1 at 391.
The eleven-measure coda begins in a manner similar to the retransition
of measures 26 through 29, transposed to a tonic context. It offers a more
wholeheartedly harmonic progression (I ➔ IV V7 I) than was the case
earlier. Yet soon that trajectory bows to the interplay of tonic pitches
arranged mostly in ascending arpeggiations.

The Roman numerals in Cadwallader’s ex. 1.4, measures 2 through 9,

convey a root progression of F–D–G–C during stanza 1. One might
reasonably expect that this progression will recur during stanza 4, which
 Mendelssohn: Song without Words in F Major (op. 85, no.1) 125

would continue beyond C to a cadence on F (= I). This is how A1 and A2

sections normally would relate in the context of an interruption. My
10.3 shows how Mendelssohn fulfills those expectations. Cadwallader’s
ex. 1.6 does not effectively convey an initial F projection and looses sight
of the D entirely.
Nineteenth-century photographers experimented with double exposure,
creating artworks that juxtapose two different images. My term for the
equivalent musical device is collision. Whereas normally tonic F-A-C
would be established first, later giving way to F♯-A-C-E♭, the special context
created at the end of the retransition allows the outer-voice CA compound
tenth (which in fact sounds briefly before F♯ and E♭ are added) at 301 to
convey the tonic harmony prior to the emergence of the 6-phase chord’s F♯
and E♭. (Diatonic 6-phase F-A-D has evolved into surging F♯-A-C-E♭.) The
latter has no equivalent during stanza 1. Here Mendelssohn is adding a
local harmonic trajectory between I and V where an embellishing chord of I
transpired earlier. Cadwallader responds to this situation by deploying the
analytical symbol I° 65 . (The symbol is displayed within quotation marks in
his ex. 1.6; without quotation marks in ex. 1.5a; and is replaced by I below
bass F within parentheses in ex. 1.5b and 1.7.) I find that symbol mislead-
ing. The local progression at that point should be understood as I5–6 II (as
shown in my 10.3). The “diminished 65 ” moment corresponds to I6, not I5.
Cadwallader’s commentary proposes that “the diminished six-five chord …
stand[s] for the resumption of structural tonic harmony” (p. 12). I would
rephrase that to indicate that the diminished six-five chord represents the
6-phase intermediary between I and II.
Is it reasonable for a chord that persists over three downbeats (mea-
sures 6 through 8) during A1 to be absent from A2? Though its manner
of attainment is different, D-F-A emerges conspicuously as the goal of a
surge (A➔ D) at 352. The preceding C➔ chord that Cadwallader labels
as V (within brackets) in his ex. 1.6 is replaced by a surging A➔ chord, so
that D arrives where F was expected. This seems to me a straightforward
softening of the boundary that was strongly articulated at the measures
5|6 bar line, giving stanza 4 a more linear bass trajectory. (It converts a
double-arpeggiation bass trajectory into a single-arpeggiation strategy.)
Then, as is possible for any chordal succession by descending perfect
fifth, the D-to-G root succession (first encountered during measure 8)
evolves into D ➔ G. Thus bass E at 362 is a passing note between D and
126 Harmony in Mendelssohn and Schumann

F♯. Cadwallader proposes the opposite: that this E reinstates the C

harmony of measure 34. (Compare the C-to-E bass slur in
Cadwallader’s ex. 1.6 and the D-to-F♯ slur in my 10.3.) My reading of
the structure corresponds closely to that of stanza 1; Cadwallader’s
diverges significantly.
Further complicating matters, Cadwallader proposes a broad FA♯  AF♯
voice exchange during measures 30 through 37. (See the crossing diag-
onal lines in his ex. 1.6.) As my 10.3 reveals, two distinct I5–6 II succes-
sions are at play, the first at the foreground level, the second at the
middleground level. The earlier F♯’s surge function is fully discharged
by the G that follows immediately thereafter (as the (F)<F♯<G slur
during measure 30 in 10.3 conveys). Consequently that F♯ cannot be
prolonged until the bass F♯ of 371.
The II–V–I close of Cadwallader’s graph (ex. 1.6) coincides closely
with my reading. Note that I slur G>F>E as a third-progression in the
alto register, clarifying F’s role as an accented passing note, and that I
regard the dominant harmony’s arrival to coincide with the first – not
the second – bass C.
11 Schumann: “Warum?” from Phantasiestücke
(op. 12)
in response to l. poundie burstein

For generations various New York City institutions of higher education

have been at the forefront in the range and depth of their course offerings
on Schenker-oriented tonal analysis. L. Poundie Burstein has continued in
that tradition as a faculty member of the City University’s Hunter College
and Graduate Center. Like his esteemed predecessors he has contributed
widely to the analytical literature. His article “Unraveling Schenker’s
Concept of the Auxiliary Cadence” offers an analysis of Schumann’s
“Warum?,” a short and technically undemanding movement surrounded
by seven longer and more challenging pieces within Phantasiestücke.1
Despite its unassuming character, “Warum?” brings several fundamental
analytical questions to the fore. Though some issues may be irresolvable,
I trust that new insight both into Schumann’s composition and into the
analytical process may be forthcoming through a comparison of how
Burstein and I grapple with the work’s challenges.

Each cycle’s final four measures

How a listener ultimately might interpret measures 1–4 (which feature a

non-tonic opening, the “auxiliary cadence” that is the focus of Burstein’s
article) likely will be influenced by the fuller contextualization of very similar
content at the ends of more extended progressions that lie ahead (measures
13–16 and 31–34), as well as an additional two times with a contrasting
harmonization at the movement’s close (measures 35–38 and 39–42).
Though none of these five four-measure units begins with a tonic harmony,
I will propose below that the version that transpires during measures 1–4
should be regarded as the most complete of these alternatives.
Particularly in a major key, various manifestations of two cadential
progressions are used much more than any other options in tonal music
of this era:

II V I and
IV V I
127
128 Harmony in Mendelssohn and Schumann

In the former the fifth-relationship between the roots of II and V offers an

opportunity for a surge, as

II➔ V I
That variant occurs during the first three of the five four-measure units under
consideration. The final two, which extend the structural close achieved
during measures 31–34, pursue the alternative option deploying IV.
Perhaps controversially, I interpret these four-measure units as endings of
progressions. Though certainly their prominent D♭<E♭<F third is the move-
ment’s most strongly and persistently projected melodic feature, I suggest
that it transpires on top of what otherwise might be perceived as a conven-
tional descent to ^1. To emphasize that point, I have omitted the D♭<E♭<F
third from 11.1, which integrates content from all three of the progressions
that utilize II➔. (These harmonizations will be explored in detail and
compared presently.) Note how a melodic E♭>D♭ second (= ^2>^1) appears
in conjunction with the II➔ V➔ I cadential harmonic progression in the
graph. (As an orientation to this conception, locate half note E♭ in the score’s
tenor register during measures 13 and 31 and trace its stepwise descent to D♭
two measures later.)

Example 11.1 Analysis of Schumann: “Warum?” from Phantasiestücke (op. 12).

Whereas measure numbers 1, 13, and 31 appear above the graph’s II➔
chord, the number 1 also is placed above the tonic’s 6-phase chord. (In the
later cycles, the 6-phase chord emerges either one or ten measures before
II➔.) I propose that measure 1’s initial DB ♭♭ compound tenth gives that 6-
phase chord time to establish itself before the E♭ and G♮ of II➔ emerge.
Thus those two distinct events within the progression collide. To accom-
modate this special construction, Schumann has elected not to double the
E♭ (which would clash with the projection of the 6-phase chord) in the
chordal interior during measure 1. (Compare with E♭’s prominent
 Schumann: “Warum?” from Phantasiestücke (op. 12) 129

sounding at the end of a quick ascending arpeggiation during 131 and 311.
The E♭ is realized instead in measure 2.) Likewise in the interest of blending
components of two harmonies, the tonic 6-phase chord’s F is omitted. But
that is not all! Looking again at 11.1, note that an F chord (in measures 11
and 19) precedes the B♭ 6-phase chord in the later complete cycles.
Wondrously, the last chord of the preceding movement (“Aufschwung”)
is F-A♭-C-F. Consequently the content of 11.1 is almost fully accounted for
even in the incomplete first cycle, where it appears in a more concentrated
dosage (four measures plus the preceding movement’s final chord) than will
prevail later. Though the “wrong” chord (something other than an initiating
tonic) begins the movement’s progression in what Burstein labels an aux-
iliary cadence, the B♭-D♭-(F) chord of 11 is exactly “right” in the local context
of the juncture between “Aufschwung” and “Warum?”!
Schumann’s construction is reminiscent of Beethoven’s procedure in the
third movement of his Piano Sonata in E Major (op. 109), which I analyze
in Harmony in Beethoven, chapter 12. That movement offers a set of
variations based on a theme that begins with an initial ascent from ^3 to
Kopfton ^5 . A descent from that ^5 to ^1 occurs over the course of each
variation. Notably, a fresh ^3 emerges concurrently with each close on ^1.
Thus every ending is also a new beginning, ready to inaugurate an initial
ascent to restore ^5 in the next variation. (As a result a G♯ sounds con-
spicuously at the end of the movement.) Likewise the cyclic organization of
“Warum?” complements each descent to ^1 with the attainment of a fresh ^3,
ready for the onset of a new descent. Such an infinite loop is an ideal
context for a non-tonic opening (Burstein’s “auxiliary cadence”). We
become cognizant of the loop in the middle of a cycle, just as a D♭<E♭<F
ascent gets underway; and that loop fades from our perception after
measure 42 with a lingering F, which potentially could serve as the starting
point for yet another descent during a hypothetical measure 43 and beyond
(as it does when the repeat sign is observed). In performance “Warum?”
yields ultimately to a movement titled “Grillen,” which reiterates the open-
ing D♭>C<D♭ of “Warum?” in the register just above Middle C and then
ascends by step not just to F, but now all the way up to A♭.

The annotation “aux[iliary] cad[ence]” at the left edge of Burstein’s ex.

17 not only serves to link his analysis with the topic of the article in
which it is embedded, but also invites a consideration of exactly what
constitutes a cadence. Certainly the chordal progression (conveyed by
Burstein via the analytical symbols V 43 /V V7 I) is a factor in creating a
130 Harmony in Mendelssohn and Schumann

sense of cadence. Yet usually a melodic descent (^2>^1, emanating from

a prior ^3 or ^5 >^4>^3) will arise as well when the tonic serves as a
progression’s goal. A comparison of measure 1 with measures 13 and
31 is especially revealing, in that the latter two occur without the added
agenda of a collision (wherein in my view measure 1 has the challenge
of conveying two different chords within the progression, a reading at
odds with Burstein’s), permitting an uninhibited projection of ^2 that is
absent from measure 1. Yet the voice leading in the range above
Middle C during measures 2 and 3 certainly may be construed as ^2
descending to ^1. For me, that event is an important contributor to the
cadence, in this case complemented by a concurrent reinstatement of ^3
(reinstatement in the sense that the incomplete cycle’s melodic descent
stems from the F a fourth above Middle C in the preceding move-
ment’s final measure). My omission of the concurrent D♭<E♭<F ascent
from my graph in 11.1 is admittedly a provocative analytical state-
ment, intended in part to challenge what might appear to be the
obvious reading of the passage, as Burstein fulfills. Yet my visualiza-
tion of a descent complemented by textual commentary acknowled-
ging a concurrent ascent is not matched in Burstein’s presentation: his
visualization of an ascent is not complemented by textual commentary
acknowledging a descent.2
I especially like Burstein’s display of how the dissonant D♭ of measure 1
descends conceptually to C at measure 2, above which an E♭ emerges via
reaching-over, during the ascending third. Though I do not advocate the
use of applied dominant chords (here V 43 /V) in harmonic analysis, espe-
cially in the context of Schenkerian graphs, I acknowledge that it remains a
common practice. Because I regard the alignment of V7 with the melody’s
E♭ in measure 2 as correct, I suggest that its delay until the melody’s next
pitch, A♭, during measures 13 and 31 may be the result of a typesetting
error.3

The cycle from measures 3–4 through measures 15–16

In one sense the tonic harmony and the melodic D♭ (= ^1) just above Middle
C in measures 3 and 4 conclude a cycle. Yet those measures also initiate the
next cycle. The model of 11.1 distills the essence of the three cycles that
utilize II, focusing on what they have in common. (Observe that the chord
 Schumann: “Warum?” from Phantasiestücke (op. 12) 131

of measures 3 and 15 appears at both edges of the graph, since Schumann is

projecting an infinite loop.) Parentheses around accidentals account for
elective surges: F➔ only during measure 11, B♭➔ only during measure 21.
(The supertonic to which the progression leads surges in all three cycles,
during measures 1, 13, and 31.) This model conforms precisely to basic
tenets of harmonic theory as I understand it: II (perhaps surging) often
comes between I and V; the I-to-II succession may be expanded via a 5–6
shift applied to the tonic; and a local mediant (perhaps surging) may come
between the tonic’s 5- and 6-phase chords.4
The second (first complete) cycle’s distinctive features are conveyed in
11.2a. A local tonic expansion transpires during measures 3 through 10,
supporting a descending third from Kopfton F’s upper third, A♭, back to
F. This local progression likewise deploys a tonic 5–6 shift, in this case with
a highly evolved 6-phase chord that is asserted as VI➔. Once the tonic is
restored in measure 10, a surging mediant emerges, targeting the tonic’s 6-
phase chord, which does not surge during this cycle. Instead the melody
descends chromatically from F through F♭ on the way to E♭, guiding the ear
in following the Urlinie descent. The chord at 122 should be read as D♭-F♭-
B♭, with G♮ serving as an anticipation of the upcoming II➔’s third.
Whereas the II➔ of measure 1 sounds in collision with a downbeat bass
B♭ (which I suggest above should be interpreted as a manifestation of the
preceding tonic 6-phase chord), in measure 13 that II➔ sounds with a
downbeat bass A♭, which anticipates the following dominant. Thus a colli-
sion – differently constituted from the one in measure 1 – occurs here as
well.

Example 11.2a–b Schumann: “Warum?” from Phantasiestücke (op. 12) (a) analysis of
mm. 3–13; (b) analysis of mm. 15–31.
132 Harmony in Mendelssohn and Schumann

Though it is likely that no one who might fruitfully read this book would
seriously dispute the claim that Schumann’s music integrates harmonic
and linear forces, the relative weighting of those factors by analysts
might result in disagreements. Burstein sensitively notes a stepwise
descent (B♭>A♭) in the bass at both measures 1–2 and 12–13. (How
those pitches correlate with the prevailing harmonies differs somewhat
in the two instances, with A♭ arriving sooner in the latter case, where
Burstein labels it as a ped[al].) To seasoned analysts, there is something
tantalizing about a D♭ (measure 5) followed, after some time, by a B♭ and
then an A♭. The D♭>A♭ fourth of course corresponds to the I–V har-
monic trajectory foundational to tonal musical utterances. Is that fourth
here fully projected as a linear entity: that is, as D♭>C>B♭>A♭? Burstein’s
answer is yes: only those pitches bear downward stems over this region
in his ex. 17, and a slur connects the D♭ and A♭ endpoints. My answer is
no: bass C in measure 9 is displayed as an internal element of a fore-
ground tonic prolongation in my 11.2a, with a slur connecting the D♭
tonic roots of measures 3 and 10. Whereas for me the chord with bass C
in measure 9 is a genuine dominant, functioning at a more foreground
level than the B♭, E♭, and A♭ chords that follow, the parentheses that
Burstein places around his V6 label at that point take the chord out of
contention as a tonic-seeking entity. In his reading it instead serves as a
passing chord between I and VI.5
Though Schenkerian analysis is necessarily a reductive process, I
suggest that too much has been removed from measure 9 in Burstein’s
graph. He is careful elsewhere to account for chordal dissonances via
figured bass (as 43 or as 7). Yet here only the number 6 is appended to
 Schumann: “Warum?” from Phantasiestücke (op. 12) 133

Roman numeral V in the graph. The score’s pitches instead warrant

analysis as 65 . That analytical symbol draws attention to the GC ♭ dimin-
ished fifth that would lead one to expect an upward – rather than
downward – continuation after the C.6 The graph also presents the
chord at 101 not as D♭-F-A♭ (the normative resolution of a dominant 65 ),
but instead as D♭-B♭-A♭. Though a B♭ sounds at the end of measure 10
(serving as a local passing note between C and A♮ in the alto register), its
jurisdiction does not extend to the preceding downbeat. Finally, though
one must be cautious in using such information (because his theory was in
a state of continual evolution over several decades), one may consult
Schenker’s marked-up score of the movement in the Oster Collection at
the New York Public Library.7 There Schenker labels the chord on the
downbeat of measure 10 as I.8

The cycle from measures 15–16 through measures 33–34

and beyond

Shown in vertical alignment with 11.2a to facilitate comparison with the

preceding cycle, the graph of the third (second complete) cycle (up to II➔)
in 11.2b reveals how Schumann pursues similar goals through contrasting
means. Whereas the preceding cycle dutifully undertakes a local expansion
of the D♭ tonic harmony during measures 3 through 10 before mediant F
emerges, the second cycle charts a direct course from D♭ to F. As often is
the case, a circle of fifths serves as the means of tonic-to-mediant connec-
tion. Because major third D♭<F is being traversed, one of the circle’s fifths
must be of diminished quality. Schumann places it first (D♭ to G), after
which C and then F follow naturally, especially since both the G and C
chords surge. Unlike the surging F chord of measure 11, this F chord is of
minor quality (diatonic in D♭ Major), which allows another sort of rela-
tionship with the earlier cycle to prevail: the transformation of the B♭ chord
in measure 12 (with diminished fifth F♭) is now applied to the F chord
during measure 20 (with diminished fifth C♭, as well as minor seventh E♭).
The melodic A♭ introduced during the preceding cycle (at measure 6) is
short-lived: a descent back to F transpires through measure 10. The A♭ that
emerges in measure 19 has more staying power. It persists through mea-
sure 30. The arrow in 11.2b shows how it eventually resolved to a G♮ in the
134 Harmony in Mendelssohn and Schumann

chordal interior, so that Kopfton F’s successor, E♭, may assume the promi-
nent position at the top of the texture.
Whereas the tonic 6-phase chord of measure 12 proceeds directly to II➔
on the following downbeat, its counterpart in measure 21 is in no hurry to
move on, despite the fact that it is surging (as VI➔). Instead the F-to-B♭
succession of measures 19 through 22 is repeated during measures 23
through 26 (now with an earlier introduction of chromatic C♭). Though
the chord of 262 (intensified during its repetition during 282) likely will
come across as an embellishing chord consisting of neighboring notes to
VI➔ (C♭-E♭-F-Bº resolving back to B♭-D-F-A♭), it also might be under-
stood to perpetuate the F-to-B♭ trajectory, with the enharmonically equiva-
lent spellings F-A♮-C♭-E♭ and A♮-C♭-E♭-G♭ revealing their roles a
supersurging F chords (thus F⇨ B♭).9
Just as the F-to-B♭ root trajectory within the third cycle is repeated, so
also the “cadential” measures 31 through 34 sound twice more before the
movement concludes. As mentioned earlier, the harmonic trajectory shifts
at that point from II➔ V➔ I to IV V➔ I. (See 11.3.) Whereas II often is
approached via a 5–6 shift from I, IV may be approached via a surging
I. The C♭ of 341 provides the principal ingredient for such a realization.10
Though listeners likely will accept these measures as a conclusion for the
movement despite the emphasis on melodic F, it is clear that the title’s
question – Warum? (Why?) – has not been answered.

Example 11.3 Analysis of Schumann: “Warum?” from Phantasiestücke (op. 12), mm.
33–41.

Though I applaud the insight that led Burstein to annotate his ex. 17
with the comment that “13–21 = varied repeat of mm. 1–7,” I suggest
that his assessment of local chordal interactions has prevented him from
fully realizing the degree to which repetition is a factor in Schumann’s
 Schumann: “Warum?” from Phantasiestücke (op. 12) 135

compositional strategy. In particular I propose a greater correlation

between the F chords of measures 11 and 19 (placed in vertical align-
ment in my 11.2a and 11.2b). For me, a surge (as occurs in measure 11
but not measure 19) is a nuance, the heightening of a tendency that is
already in play even without the chordal evolution: a diatonic F minor
chord likely would proceed to a B♭ chord in any event (as it does
between measures 19 and 21), though the shift of its third from A♭ to
A♮ and the addition of its minor seventh E♭ (as in measure 11) make that
succession more dynamic and inevitable. It is possible to read measures
15 through 21 in Burstein’s ex. 17 as the projection of an F chord as the
principal intermediary between D♭ (I) and B♭ (VI). (This viewpoint
corresponds to my reading in 11.2b, where slurs below the bass connect
D♭<F and F<B♭.) Though equivalent slurring is found between measures
10 and 12 in my 11.2a, fostering the notion of “varied repeat” that
Burstein proposes, it is not possible to read the corresponding region of
his graph in that way. Instead, the principal pitch leading to B♭ is the C
of measure 9. The F of measure 11 is subordinate to the broader
descending linear trajectory. My reading reverses those hierarchies, so
that the D♭ chord at 101 restores the tonic.
By coincidence Schumann’s strategy here is very similar to another
passage featured in this volume, from Mendelssohn’s Song without
Words in F Major explored previously in chapter 10. During measures
2 through 6 a broad F–C–F trajectory is followed by a descending third
to D, whereas during measures 30 through 35 the latter F is elided, so
that a more direct course from F to D prevails. (Compare 10.1 and 10.3.)
In similar fashion, during measures 3 through 11 of “Warum?”
Schumann projects a broad D♭–A♭–D♭ trajectory followed by an ascend-
ing third to F, whereas during measures 15 through 19 a more direct
course from D♭ to F prevails.
 12 Mendelssohn: Song without Words in A♭
Major (op. 53, no. 1)
in response to yosef goldenberg

Once in a while a volume extraordinarily rich in Schenkerian graphs is

published. Perhaps the best known such book (after Schenker’s Free
Composition, of course) is Felix Salzer’s Structural Hearing. Another
project of similar scope and format (one volume of text, another of
examples) is Yosef Goldenberg’s Prolongation of Seventh Chords in
Tonal Music.1 I explore one of his Mendelssohn graphs from that project
here.

The introduction and the A1 section

Mendelssohn’s ternary-form song is built using alternating A and B

sections arranged as | A1 |: B A2 :|. The repeat is written out. (The A1
section – but not the A2 section – is subdivided into a1 and a2 regions. The
succession from A1 to B without a repeat of A1 acknowledges that a2
already is to some extent a repeat of a1.) A brief introduction and a more
extended coda frame those central components. Though the introduction
(displayed at the left edge of 12.1a) amounts to little more than a bald
projection of the tonic key’s A♭–E♭–A♭ bass arpeggiation, we might
admire the especially daring sounding of a B♭-D♮-F-A♭ surging super-
tonic chord against a fleetingly maintained E♭ during 21. The bass arpeg-
giation supports an E♭>D♭>C third-progression above.2 Mendelssohn
cleverly deploys that third (transferred up an octave) in retrograde once
the song proper begins: a C<D♭<E♭ initial ascent leads to Kopfton ^5
during measures 3 through 5.

136
 Mendelssohn: Song without Words in A♭ Major (op. 53, no. 1) 137

Example 12.1 Mendelssohn: Song without Words in A♭ Major (op. 53, no. 1) (a) analysis
of mm. 1–11; (b) a possible interpretation of measures 9 and 10 as IV5–6.

(a)

(b)

Mendelssohn shapes the content of A1 (from 33 through 112) as two

four-measure phrases – antecedent and consequent, a1 and a2. The initial
tonic is presented in two phases. First, its triad’s two thirds are projected at
the edges of the texture, with filled-in A♭<C and C<E♭ ascending in parallel
motion. Second, a surge (targeting IV) is ignited by the addition of chro-
matic G♭. Though the unfolded C<G♭ diminished fifth at the bottom of the
texture (measure 5) typically would be complemented by F>D♭ during IV’s
prolongation, in this case the F holds out, leading directly to the dominant
root (E♭) at 71. (Consequently the deployment of bass D♭ during the
following phrase, at 102, is an especially appealing turn of events.) The
subdominant’s 5-phase D♭-F-A♭ (inverted to F-A♭-D♭ in the score) accom-
modates the addition of 6-phase B♭ during 63. The shift of D♭ to D♮ on the
following beat results in a surging II➔. These various harmonic maneuvers
coordinate with a foundational melodic line that descends from Kopfton E♭
through D♭ and C to B♭, as beamed in 12.1a.
138 Harmony in Mendelssohn and Schumann

In its broad outlines, what follows during a2 is very predictable. The

re-emergence of the tonic harmony from 73 through 82 again supports an
ascent to ^5 . The stepwise descent (this time all the way to ^1) doles out the
expected pitches one per beat from 102 through 111 (or will come across as
such if one forgives C’s tardy arrival within an A♭<C unfolding that extends
beyond 103, where it “belongs,” into 104). The chord at 93 is susceptible to
two contrasting interpretations. Clearly it relates somehow to the antece-
dent phrase’s subdominant, though its emergence a half measure earlier
within its phrase than before, at a forte dynamic level, and with an added
high A♭ are such distinctive features that one should not assume it will play
exactly the same role in the consequent phrase. Perhaps the harmonic
implications of IV are realized immediately, through a succession to V8–7
and I during the next two beats. In that reading the melodic F>E♭ second at
both 91–2 and 101 signals a continuity between two statements of the tonic,
with a restoration of tonic root A♭ in the bass (not found at that point
during a1) and with descending A♭>G>F>E♭ complementing the preceding
C<D♭<E♭ (thus a double confirmation of the Kopfton) appearing before the
broader structure leads to the cadence. Some analysts might propose
instead that the subdominant’s bass F at 93 initiates the trajectory toward
dominant root E♭ at 103, an interpretation that would warrant connecting
the subdominant third F and root D♭ in the bass between 93 and 102 (again
accommodating 6-phase B♭, this time without the continuing evolution
into a surge). (I supply a patch, 12.1b, that readers who favor this inter-
pretation may insert into 12.1a.) In this perspective the A♭-C-E♭ at 101
would be interpreted as a connective chord within the prolongation of IV
(supporting melodic E♭ within a stepwise filling-in of IV’s A♭>D♭ fifth),
rather than as a structural return to I. Because both of these interpretations
have merit, I have left 12.1a as neutral as possible in that vicinity,
placing a hairpin symbol between measure numbers 9 and 10 to
acknowledge that some material has been omitted. The II displayed
during measure 10 might be the principal successor of the phrase’s
initial tonic, or it might instead result from a IV5–6 trajectory similar to
that from the a1 presentation. (Several events later in the song offer
evidence that Mendelssohn had the first of these two interpretations in
mind. As will be conveyed in due course, a melodic descent reminiscent
of the A♭>G>F>E♭ fourth occurs during the B section; a reworking in
the structure during the A2 section removes IV altogether, though the
II is retained; and the coda fills in the AE ♭♭ fourth in both ascending and
descending directions.)
 Mendelssohn: Song without Words in A♭ Major (op. 53, no. 1) 139

In that Goldenberg is presenting an analysis of this Song without Words in

order to explore an event from the B section (to exemplify a V7–8–7 motion),
he does not offer a detailed view of what transpires during A1. His graph
merely shows open-notehead bass A♭ (annotated as I in his harmonic
analysis) and open-notehead soprano C, which he curiously indicates as
occurring during measure 1, rather than what I think he actually intends –
measure 3, after the introduction concludes. The interpretation of this
work’s Kopfton as ^3 sidesteps the notion of initial ascent, a common device
that may commence either on ^1 or on ^3 when the Kopfton is ^5 . I assume
that Goldenberg reads my initial-ascent third as C<D♭<E♭, rather than as
C<D♭<E♭. Likewise the D♭>C>B♭ beamed in my graph’s measure 6 would
necessarily be read as D♭>C>B♭, with the underlined B♭ connecting the
earlier (Kopfton) C and the section-spanning third-progression’s conclud-
ing A♭. Yet that reading calls into question the viability of IV (targeted by
the surging I➔) rather than II as the principal intermediary between I and
V. Indeed, Schenkerian analysts do sometimes disagree regarding the
choice of Kopfton, though this movement’s A1 section is not a passage I
would have expected to be controversial.
My one misgiving about my own reading stems from the fact that
Schenker did not include a model like the one I display in my 12.1a
among the twenty sample fifth-progression descents from ^5 in his Free
Composition, fig. 16.3 (He does, in fact, insert an “etc.” at one point to
indicate that his presentation is not to be regarded as exhaustive.) The
second model in line 5 comes close to what I propose, though it reaches V
at the onset of ^2, instead of after ^2 has been established (the defining feature
in the projection of IV’s 6 phase). Fortunately a few pages later Schenker
offers an analysis that proceeds along the lines of my Mendelssohn reading.
Observe in fig. 40, ex. 9, how a descent from ^5 reaches ^4 at the onset of IV
(likewise achieved via a tonic surge, not acknowledged in Schenker’s
harmonic analysis), then through passing ^3 to ^2, which coincides with a
II numeral. (Schenker sometimes writes IV followed by II and sometimes
IV5–6 in such contexts.) A broken tie connects that ^2 and the ^2 of the
following V♯, and both noteheads are attached to the fifth-progression
beam, which concludes with ^1 at the arrival of I.4
Lacking a more detailed analysis, Goldenberg’s and my perspectives
regarding the Kopfton must be left as alternative hypotheses at this
point. I will confirm below that what occurs during the B section may
be interpreted as a logical continuation from a ^5 established during A1.
140 Harmony in Mendelssohn and Schumann

The B section

Working our way into the B section, we should perceive how the C♭
chord at 171–2 offers a moment of internal stability. Crucially, it sup-
ports a melodic E♭ (Kopfton ^5 continuing from A1). Mendelssohn
reinforces that goal point by repeating – and intensifying – the preced-
ing two measures, so that 191–2 = 171–2. A very effective and satisfying
way to connect a key’s tonic and mediant chords is to traverse a
segment of the circle of fifths, here A♭ D♭➔ G♭ ➔ C♭ (as marked in
12.2). In that each descending fifth is perfect, the mediant attained is
rooted on C♭. That major chord’s diatonic status in A♭ Minor (not the
A♭ Major in which this composition is written) results in an outbreak of
minor-mode hegemony that persists for several measures (as the
numerous C♭s, F♭s, and G♭s in 12.2 confirm). As the potency and
forward drive of this circular trajectory sink in, listeners should come
to realize that the E♭➔ chord at the onset of the B section ultimately
had no impact upon what follows, to the extent that its label might
➔
better be displayed as E♭, indicating a back-relating chord still
beholden to the A♭ tonic from A1’s cadence. Composers often prolong
the tonic for a few measures at the onset of a B section. Though this
local E♭➔ chord might have resolved to A♭ to produce such a tonic
extension, here an inverted D♭➔ chord supplants it to initiate the
circular progression.

Example 12.2 Analysis of Mendelssohn: Song without Words in A♭ Major (op. 53, no. 1),
mm. 3–24.
 Mendelssohn: Song without Words in A♭ Major (op. 53, no. 1) 141

Following the playbook developed during A1, a subdominant in first

inversion follows mediant C♭ during the B section.5 (Compare F-A♭-D♭ at
61 and F♭-A♭-D♭ at 193, appreciating how an E♭ embellishes the melody’s D♭
in both contexts.) Then something a bit startling occurs. With bass F♭
descending to the dominant’s E♭, the expected melodic trajectory at the
top of the texture is incompletely presented – for good reason, since the
reconstituted line on display in 12.2 shows a C♭>B♭ half step atop bass F♭>E♭.
Parallel fifths! Mendelssohn got himself into this situation through the bass
succession F>E♭ introduced during measures 6 and 7, now reprised (using
F♭) for the dominant’s attainment during the B section. Yet at first he does so
without the additional content that had achieved ^2 prior to the bass arrival on
E♭ during that earlier passage. (More on that point in a moment.) His initial
response to this voice-leading dilemma is to allow his melody to substitute
inner-strand A♭ for the offensive C♭ at the end of measure 19 (and again
during the passage’s repeat a measure later). Then, in an alternative
approach to the dominant (displayed within parentheses in 12.2), he reverts
to the successful strategy of measure 6, proceeding downwards from D♭
through C♭ to B♭ in the melody before leading from the subdominant (now
in its 6 phase) to the dominant, which in any event no longer presents a
voice-leading challenge since it appears in inversion. (See the arrow con-
necting the bass F♭ to an open-notehead E♭ in the alto register in 12.2.)
Mendelssohn’s final B-section iteration of the dominant includes one small
detail that warrants special mention. Very often the dominant’s seventh will
sound (in this case, the pitch D♭) in the moments before the tonic’s restoration
at the onset of an A2 section. That D♭’s resolution pitch is C, the tonic’s third.
Mendelssohn here sounds an F, the dominant’s ninth, above D♭ during 242.
(Neighbors E♮ and G embellish that F.) Given that the choice of Kopfton is a
matter of analytical concern, I am delighted to hear an F>D♭ third (ninth and
seventh of V) just before the C<D♭<E♭ ascent that restores Kopfton ^5 !

Both Goldenberg and I regard the arrival of a C♭ major chord at 171–2 as

one of the B section’s most significant moments, though our views
regarding how it functions within the broader context are contrasting.
We also agree that a G♭ chord – G♭➔ in my notation, V/ preceding (♭III)
in his – is that C♭ chord’s predecessor, though his reading confuses
matters by connecting bass E♭ at 161 and C♭ at 171 by means of a dotted
beam, above the annotation “aux.” (indicating a local tonicization of C♭
via an auxiliary cadence). In my graph the chord with bass E♭ instead
participates in a prolongation of G♭➔, as a 63 unfurling of a passing 64
142 Harmony in Mendelssohn and Schumann

chord. (To get the gist of my interpretation, play the following three
chords at the piano: G♭-B♭-D♭, G♭-C♭-E♭, G♭-D♭-F♭.)
Our more significant differences become apparent when one proceeds
backwards from the G♭ chord. Whereas Goldenberg’s placement of the
number 10 between the staves of his example suggests that he regards bass
F♭ (during measures 13 and 14) as a significant pitch, for me a more
consequential chord congeals above the F♮ that follows. It is exactly what
I would expect to precede a G♭ chord: a surging D♭➔. Given how things are
shaping up (as a circle of fifths), the logical predecessor of D♭➔ would be
the A♭ established as the tonic during A1.6 This reading necessarily demotes
the B section’s first chord – E♭-G-B♭-D♭ – to a subordinate hierarchical
plane. In Goldenberg’s perspective that chord instead serves as the arrival
point of V7, to be prolonged throughout the entire section. (Observe that
the only open noteheads Goldenberg deploys during the B section – bass E♭
and soprano D♭ – appear in this chord.) Thus the point of his presentation –
showing how a C♭ (♭III) chord may come between two V7 chords (to
demonstrate his model e) – falls apart for me. I contend that the C♭ chord
instead emerges along the path between tonic A♭ and dominant E♭.
The C♭ chord’s arrival coordinates with a fresh sounding of Kopfton E♭.
One of my proposed readings of the passage from 93 through 102 inter-
prets the melody’s A♭>G>F>E♭ as an important reinforcement of Kopfton
E♭ from above.7 Both my 12.2 and Goldenberg’s graph show a similar
G♭>F♭>E♭ descent reinforcing the E♭ of 171–2, though in Goldenberg’s
graph that E♭ is interpreted as a neighbor of Kopfton C’s neighbor, D♭.
Consequently he is contented to sit on the D♭ that emerges during 193–4,
since it corresponds to his long-prolonged soprano pitch, with no expec-
tation that it will resolve to C until the A2 section commences. With E♭’s
featured role in my reading, I instead hear the D♭ of 193–4 as the onset of
an expected descent to B♭, for a form-defining interruption after ^2. Note
that this B♭ (at 202, completing a G<B♭ unfolding) is absent from
Goldenberg’s graph. Then consider how Mendelssohn develops the
notion: after two unsatisfactory E♭>D♭>B♭ descents, the “hole” between
D♭ and B♭ (for which I supply a parenthetical C♭ in 12.2) is filled in by C♭
during measure 22, facilitated by a contrasting harmonic support.
Consequently the B♭ goal of the melodic descent (a very typical goal for
a B section, since it is a defining ingredient of an interruption) has been
presented thrice. The sforzando soprano D♭ at 231 is not a mere con-
tinuation of the three preceding flagged D♭s in Goldenberg’s graph, but
 Mendelssohn: Song without Words in A♭ Major (op. 53, no. 1) 143

instead a fresh emergence of the dominant’s seventh in that register (as a

reaching-over) after the background descent to ^2. Finally, the high F
during 242, whose resolution to Kopfton E♭ I keenly await (and am pleased
to hear in the F>E♭ second of 261–2) is not present in Goldenberg’s graph.

The A2 section and the coda

Due to the written-out repeat of B plus A2, there are two points in the score
where A2’s concluding tonic occurs: measures 30 and 49. Measure numbers
from the earlier trajectory are used in the discussion below. Though
Mendelssohn dispenses with the antecedent/consequent phrase pairing of
A1 during A2, the expected PAC at 281 is thwarted by the melody’s swerve
up to C. Instead of starting the phrase over again, however, he picks up in the
middle. Consequently the ideal four-measure phrase would consist of 243
through 262 followed by 283 through 302. (A hairpin symbol between measure
numbers 26 and 29 in 12.3a acknowledges the omission of the “first try” at
achieving a PAC in my presentation of the phrase’s structure.) There is a
modest sense of speeding up the trajectory during 261–2, where the G♭ that
arrived in the bass at 53 during A1 now is introduced sooner and an octave
higher. During the “first try” continuation, the surge that it generates leads
from the tonic to IV (at 263). During the second (successful) try another G♭
emerges in a different context, coordinating with the introduction of A♮
(during 291) to target the II that had emerged at an equivalent location during
A1. (Compare the chords of 102 and 292, noting that the lowering of F to F♭
occurs only in the latter.) In a wondrous reformulation of the rather plain
content of 263 through 271, Mendelssohn now draws upon the circle of fifths
to bind structurally deeper chords, here the tonic and the mediant. (Observe
how the A♭ and C chords in the second model of 12.3b both support Kopfton
E♭.) Whereas a similar circle during the B section was built entirely from
descending perfect fifths, leading to the lowered mediant, C♭-E♭-G♭ (measures
17), now one of those descending fifths is of diminished quality, so that the
diatonic mediant, C-E♭-G, is attained.8 Besides its welcome contrast to what
was heard at the corresponding location earlier, this material might appeal to
attentive listeners for an altogether different reason as well, in that it is strongly
reminiscent of a passage from Mendelssohn’s “Wedding March” (Incidental
Music to A Midsummer Night’s Dream). Play from 283 through the cadence,
substituting D♮ for F in the melody at the end of measure 28, to hear that
famous tune embedded within this Song without Words!
144 Harmony in Mendelssohn and Schumann

Example 12.3 Analysis of Mendelssohn: Song without Words in A♭ Major (op. 53, no.
1) (a) mm. 24–30; (b) mm. 26–29.

(a)

(b)

The coda first reiterates the A♭>G>F>E♭ fourth, borrowing the chordal
support of measures 28 and 29 (12.3b, second model), a context that
fortifies Kopfton E♭ prior to the descent to ^1. Yet having already proven
that he is adept at ^5 >^4>^3>^2>^1 (as in 12.3a), Mendelssohn playfully
reverses course during the coda by inverting that conventional descending
fifth from ^5 into an ascending fourth: E♭<F<G<A♭!9 Consequently the coda
reverses the adage that what goes up must come back down into what goes
down must come back up. First he descends to E♭: A♭ (493), G (504), F (504),
E♭ (511). The ascent that follows is a more treacherous undertaking, since
Mendelssohn is defying musical gravity. The E♭ is reiterated during a
tonic’s surge (513), followed by F (521) supported by IV and G (523)
supported by V7. But then he drops the ball (or, musical gravity intercedes)
at the cadence: E♭ rather than A♭ emerges at the top of the tonic chord at
531. The same result comes in the second try (533 through 572), where
 Mendelssohn: Song without Words in A♭ Major (op. 53, no. 1) 145

another flaw emerges: the concluding tonic root fails to arrive in the bass.
Mendelssohn persists. There is significant progress at 581: the goal soprano
A♭ emerges for the first time, though the problem in the bass persists. He
then fortifies the dominant during measures 59 and 60, followed by success
at last over the measures 60|61 bar line, where the G<A♭ melodic resolution
sounds against an E♭>A♭ bass.

In that Goldenberg’s reason for exploring this composition relates to

content from the B section, his graph does not continue beyond that
point.
 13 Schumann: Three songs from Liederkreis (op. 39)
in response to charles burkhart and david ferris

Charles Burkhart served for many years as a member of the music theory
faculty at Queens College in New York. Among musicians he did not teach
directly, he is revered especially for several articles that deploy Schenkerian
analytical techniques and for a popular anthology of musical scores. His
contribution to Schumann studies appears as a work titled “Departures from
the Norm in Two Songs from Schumann’s Liederkreis.”1 My analyses of these
works below seek to restore the normalcy that Burkhart has denied them.
David Ferris teaches musicology at Rice University. His work on
nineteenth-century lieder has focused especially on Schumann. His article
“‘Was will dieses Grau’n bedeuten?’: Schumann’s ‘Zwielicht’ and Daverio’s
‘Incomprehensibility Topos’” displays an analytical sophistication far exceed-
ing what usually would be found in a musicological study.2 Though I will
challenge some of his assertions, I do so with respect for the breadth of his
purview and for his willingness to enter seriously into the analytical fray.

“Mondnacht”: introduction and stanzas 1 and 2

(measures 1–44)

Despite some novelties to be explored as this chapter proceeds, the basic

features of “Mondnacht” fit within a conventional framework for how tonal
music is structured. Twice during Schumann’s projection of the poem’s first
stanza (measures 6–21), one encounters a tonic harmony (clearly and con-
sonantly stated in measures 10–12 and 18–20) followed directly by the domi-
nant (measures 13 and 21). After a reiteration of the introduction’s content, the
second stanza pursues a similar course (measures 28–43). Consequently four
rows of measure numbers appear above the first two chords of the graph
displayed in 13.1. This content corresponds to the A1 section of an A1 A2
binary form, generated by the interruption of the Urlinie at ^2. The vocal B at
131 and the three corresponding later locations (each sounding above ^2) is a
reaching-over of a pitch from the dominant chord’s interior. Setting the stage
for the reintroduction of ^3 during A2, this B has no immediate impact. Though
146 often ^3 would descend to ^2 followed by the emergence of such a higher B, in this
 Schumann: Three songs from Liederkreis (op. 39) 147

case these two distinct voice-leading events transpire at the same moment,
acknowledged by the bracket denoting a collision in 13.1.

Example 13.1 Analysis of Schumann: “Mondnacht” from Liederkreis (op. 39, no. 5).

A convincing projection of G♯ as the Kopfton is made manifest only

gradually: note that G♯ emerges during measure 29 where it was absent during
measures 7 and 14, and also an octave higher during measure 36, initiating a
G♯>F♯>E third-progression (beamed in 13.2). That third is reiterated in the
vocal melody during each of the four statements’ penultimate measures (for
example, measure 42), followed by its successor – third FD♯♯ – in the cadential
measures. Whereas during A1 the vocal line leaps up to B instead of leading
downwards from G♯ to F♯, an F♯ is unambiguously presented at the
corresponding location within A2 (at 583, one beat earlier than during A1
because the harmonic progression leads beyond the dominant to the tonic
during A2), where instead the bass generates complications.

Example 13.2 Analysis of Schumann: “Mondnacht” from Liederkreis (op. 39, no. 5),
mm. 36–43.
148 Harmony in Mendelssohn and Schumann

The exemplary structure projected in 13.2 comes about with the help
of a modest analytical intervention. Whereas the local harmonic progres-
sion during measures 36 through 40 is displayed as a complete and
conventional realization of I5–6 II V7 I (with, as is common, the tonic’s
6-phase chord asserted as VI➔), in Schumann’s score the initial 5-phase
tonic is absent. In Schenkerian analysis an event of this sort is called an
auxiliary cadence. Its impact on the broader structure here is limited, for
two reasons: first, soprano G♯ is introduced at the onset, supported by a
chord derived from the tonic (its 6 phase); and second, the local progres-
sion concludes on a 5-phase tonic that spreads out broadly over three
downbeats – with bass E>E>E – compensating for its earlier omission.
The emblematic G♯>F♯>E third-progression of measures 36 through 40
and its reiteration during measure 42 make G♯ the only viable choice for
Kopfton.
By withholding the 5-phase tonic onset Schumann invites listeners
to contemplate a sense of the infinite, as projected by the word
Himmel (heaven) in measure 9. The text’s heaven-to-earth trajectory
is suggested by the E>B>E>B>E>B bass of measures 10 through 12.
In this context (weighted metrically so that the E pitches are strong
and the B pitches are weak), the tonic chord alternates between its
foundational 53 state and 64 unfurlings. The pitch B is not asserted as a
dominant root until the phrase’s final downbeat (where the down-
ward cascade in the bass finally subsides). The four-time presentation
of similar material (most fully realized as displayed in 13.2) during
A1 enhances this sense of the infinite. Even the introduction partici-
pates in the ongoing repetitive agenda: its II–V succession relates to
that of measures 8 and 9. (Note how an F♯>E>D♯ third is featured in
both contexts: in the soprano during the introduction, in the bass
during the stanza.) In experiencing the work, one perceives how the
introduction’s foundational II–V expands dynamically, enhanced at
the onset by a VI➔ surge targeting II and in the continuation by the
progression’s onward journey from V➔ to I, which arrives on a
hypermetric downbeat.3
A proposal for how the introduction’s structure may have developed
in Schumann’s mind is conveyed in 13.3. Model 1 shows the direct
connection of II and V, incorporating one passing note (E). Two
further enhancements emerge in Model 2: the shift from A to A♯
against F♯>E above, thereby igniting a II➔ surge; and the juxtaposi-
tion of the supertonic chord’s first-inversion and root-position
 Schumann: Three songs from Liederkreis (op. 39) 149

configurations, thereby introducing melodic thirds A>F♯ and C♯>A♯.

In Model 3 those thirds are filled in by passing notes, in coordination
with the early introduction of E above. Though the three members of
E Major’s tonic triad sound concurrently in this model, the context
does not support their analytical interpretation as an asserted
I. Finally, in Model 4 the role of the initiating high C♯ as an interior
chord member temporarily sounding at the top of the texture is
clarified, while the filled-in thirds now feature chromatic pitches. In
fact, the lower line (from A to F♯) fails to reach its goal during the
supertonic prolongation: though G♮ might have descended to F♯ prior
to II➔’s resolution to V, it in fact does not do so. Consequently a
more complex evolution of the surging supertonic sounds in place of
the tamer choice proposed in Models 2 and 3. By this means the
dominant arrival results from the resolution of three tendency notes:
an upward resolution of the chordal third (A♯ to B) and downward
resolutions of the chordal seventh and lowered ninth (E to D♯ and G♮
to F♯).

Example 13.3 Genesis of Schumann: “Mondnacht” from Liederkreis (op. 39, no. 5),
mm. 1–3.

Before considering Burkhart’s harmonic perspective in detail, some

attention to metrical issues is in order. My 13.2 displays eight
measures of content, which in my view constitute two four-bar
150 Harmony in Mendelssohn and Schumann

hypermeasures, notably with a tonic arrival on a hypermetric

downbeat (401). Comparing that reading with the diagram
Burkhart presents on page 148, it is apparent that our views con-
flict: whereas I display 36–43, he indicates the partitioning 37–44.
(Likewise with the preceding three sets of measure numbers; the
two sets that follow will be assessed later.) In my reading the
piano’s C♯ just after the onset of measure 6 helps shape a chord,
prolonged for nearly two measures, that inaugurates the first
phrase. Bass E♯ from 71 “belongs” at 61, as in fact occurs eight
measures later – at 141, the onset of the second phrase. Burkhart
instead explicitly assigns measure 14 to the first phrase, declaring it
to be an “anticipation of the start of the second phrase” in measure
15 (p. 149).
The divergence in our metrical readings helps explain how two
seasoned Schenkerian analysts could propose such wildly different
harmonic perspectives regarding this work. Given his partitioning,
Burkhart may with some justification propose that the C♯ 65 chord of
measure 14 is an offshoot of the “B major chord begun in measure 13”
(p. 149), whereas in the context of my partitioning it is a fresh
beginning that has nothing to do with the preceding B dominant.
Instead its B, its implied G♯ (compare with measure 29), and its raised
E suggest a derivation from the E-G♯-B tonic triad. In my score I have
placed vertical lines just after the first sixteenth-rest or -note of
measures 1, 3, and 6, thereby partitioning the introduction and first
phrase into four components that achieve progressively higher levels
of complexity:

V
II V
II V
I6 (= VI➔) II V7 I ...

From this perspective, the V chord that concludes the first three trajec-
tories and the VI➔ chord (Burkhart’s C♯ 65 chord) that initiates the
fourth are unrelated. (The next trajectory, corresponding to the second
phrase, begins on – rather than just after – the downbeat in measure 14.)
Beyond this, Burkhart appears not to have taken into account the fact
that the phrase’s evolution persists right through its fourth statement,
beginning in measure 36. My ear focuses on soprano F♯ during
 Schumann: Three songs from Liederkreis (op. 39) 151

measures 1–2 and 3–4 and F♯>E during measures 8–11, 16–19, and 30–
33. Given that the work is in E Major, those pitches seem curiously
fragmentary. Consequently I am delighted by the high G♯ that emerges
in the piano during measure 36. Now the F♯>E second may be under-
stood within the broader context of a G♯>F♯>E third (measures 36–41).
(That is why my 13.2 presents an analysis of the fourth phrase, rather
than one of the earlier ones, as Burkhart does.) Alas, Schumann’s filling-
in of the GE ♯ third does not coordinate well with Burkhart’s conception of
a broad dominant prolongation or his reading of the Kopfton as ^5 . My
contrasting tonic prolongation and ^3 Kopfton project a structure much
more in line with normative tonal conventions. If my reading is to
prevail, Burkhart would need to retract a number of potent words that
appear at the onset of his essay: “drastic,” “strange,” “less natural,”
“subverting,” and “unusual.” He applies these words to Schumann’s
composition. I instead would apply them to Burkhart’s analysis.
My revisions to the part of Burkhart’s ex. 2 that corresponds to my A1
would focus on the harmonic impact of measures 10–12 (tonic, not
dominant) and the continuation from the II of measure 8 (to the V of
measure 9, not to that of measure 13). Whereas I endorse the succession
from GE ♯ to FD♯♯ leading up to measure 13, I regard the G♯ as a pitch
prolonged over seven of the phrase’s eight measures, explicitly presented
(in two registers) during the fourth phrase, measures 36 through 42.
Burkhart’s interpretation of the introduction corresponds for the most
part to what I display in 13.3. My most urgent request for revision would
be to remove the ^5 that annotates measure 5’s soprano B in his exx. 5b and
5c. Instead F♯ is the passage’s most prominent melodic pitch. A corre-
sponding F♯ will be incorporated within a G♯>F♯>E third-progression
once the A1 phrase has sufficiently evolved. Whereas the arrow in mea-
sures 1/3 of his example 5c correctly displays the high C♯ as a note that
belongs below F♯, the B of measures 3/5 likewise should be interpreted as
interior to F♯ (prolonged by means of the F♯>E>D♯ third-progression).

“Mondnacht”: stanza 3 and coda (measures 44–68)

Returning to 13.1, one may observe how the set-up for continuation estab-
lished as early as measure 13 ultimately is carried out. The restoration of
Kopfton ^3 from above (as B>A>G♯) inaugurates the A2 section. The critical
152 Harmony in Mendelssohn and Schumann

factor thereafter is that the progression must proceed beyond the dominant
(supporting ^2) to a tonic that coordinates with the continuing downward
melodic trajectory to goal ^1. Because Schumann is working within what at
first will appear to be a four-bar hypermeasure (measures 56 through 59,
ultimately extended to 61), the sounding of ^2 shifts from the downbeat of
the fourth measure (as in measure 43) to the last beat of the third measure (58),
thus making the fourth downbeat available for ^1. Despite the melody’s striving
for an on-time arrival, the bass B>A>G♯ (a reprise of the descending third that
reinstated ^3 at the onset of A2) during measures 58 and 59 complicates matters,
resulting in an expansion of the final tonic from one to three measures. (This
structure, reminiscent of final cadences by J. S. Bach, is displayed in 13.4. The A
neighbor of G♯ in measure 60 echoes that of the Kopfton in measure 51.)

Example 13.4 Analysis of Schumann: “Mondnacht” from Liederkreis (op. 39, no. 5),
mm. 47–61.

The ascending bass trajectory from I to V may be divided into a second

plus a fourth in either order. The two progressions I5–6 II V and I IV5–6 V
both get the job done, and both deploy (in E Major) an F♯-A-C♯ chord
immediately preceding V. Whereas the former route is pursued during A1
(with, as 13.2 shows, the surge of an asserted VI➔ serving as I6), Schumann
shifts to the latter route during A2 (where, as 13.4 shows, a surge transpires
as I➔ targets IV). By this means the latter part of the tonic prolongation
(commencing in measure 53) still corresponds to what happened earlier,
while concurrently the section offers some novelty, especially welcome
after the uncommonly repetitive A1. Despite this change in harmonic
course, again the progression lacks an initial consonant tonic chord
(E-G♯-B): a surge already is underway at the onset of the tonic restoration
 Schumann: Three songs from Liederkreis (op. 39) 153

(with D♮ at 471 serving as I➔’s minor seventh). Because IV replaces II as

the principal intermediary between I and V, the middleground G♯>F♯>E
third-progression of A1 is replaced by an upper-neighbor embellishment of
^3, since A – and not F♯ – is a member of IV. That G♯<A>G♯ trajectory is
covered by a coordinating B<C♯>B motion, so that the pitch A is not
overemphasized (in both outer voices) at the onset of IV.
The brief coda, which begins just after the tonic resolution at 611,
projects echoes of Kopfton G♯, along with reiterations of the B>A>G♯
line that reinstated G♯ at the onset of A2.

My reading of the hypermetric units within A2 likewise contrasts that

presented in Burkhart’s diagram on page 148. Through comparison with
measure 10 and its replicates during A1, measure 56 should be interpreted
as a hypermetric downbeat (with the unit extending through measure 61,
resulting from the addition of two measures to accommodate the embel-
lishment of the final tonic). Acknowledging that the bass F♯>E>D♯ of
measures 8 and 9 expands to three measures (53 through 55) would
support the proposal that measures 51 through 55 form an expanded
five-bar hypermeasure. In the reading that is emerging, both hypermetric
downbeats serve as venues for the introduction of an important harmony
(IV in measure 51, I in measure 56). Extending backwards, the preceding
four measures (47–50) would form a hypermeasure, again confirmed by
the harmonic analysis, which proposes that exactly those four measures
correspond to the prolongation of A2’s restored (and surging) tonic.
Consequently measures 44 through 47 might be interpreted as a three-
measure extension of A1’s concluding hypermeasure, allowing the V
attained at 431 some space to unfold, with the bass fifth B>A>G♯>F♯
leading to tonic root E on the next hypermetric downbeat. Thus I suggest
that my partitioning

40–46 47–50 51–55 56–61

I V8–7––– I––––––––––––––––––––––––––––
(= I––––––––––––––––––––––––V7 I)
(= I––––––––➔ IV5–6 V7 I)

should replace Burkhart’s

37–44 45–52 53–60

A comparison of Burkhart’s ex. 2 and my 13.4 reveals our contrasting
conceptions. For me, root E at 471 is one of the composition’s five most
154 Harmony in Mendelssohn and Schumann

significant bass pitches (warranting inclusion within my sparse 13.1). The

chordal dissonance at that point relates to events closer to the musical
surface; foundationally the chord represents a consonant E-G♯-B tonic
triad.4 Burkhart instead treats the E chord as a local embellishment of an
A that falls within a B>A>G>F♯ fourth connecting the root of V and a (II)
that comes between that preceding and a following V chord. (Note the
broad B-to-B slur in the bass of Burkhart’s exx. 2a and 2c.) Though my
interpretation likewise relates the V harmony that concludes each phrase
within A1 and the V harmony of measure 58, it is a relationship of an
entirely different sort: as the dominant attained in each half of an
interruption-generated binary form. Within that perspective, the latter
dominant serves as the internal component of an E–B–E bass arpeggia-
tion, as conveyed in 13.1 and fleshed out in 13.4.
As stated above, the IV of measure 51, targeted by the tonic’s surge,
emerges on a hypermetric downbeat. A descent in thirds transpires in
the bass thereafter: A > G♯> F♯ > E > D♯. Burkhart’s reading focuses on a
dominant-prolonging B-to-F♯ fourth (slurred in his exx. 2a and 2c),
whereas mine focuses on this A-to-D♯ fifth, which coordinates with the
conventional pursuit of a IV5–6 V harmonic succession, whose V in this
instance is built above the chordal third D♯, rather than above root B.
The fact that both Burkhart and I propose A (which he labels as ^4 in
his examples 2a and 2c and I display as a prolongation of the A
neighbor (N) introduced at 511) as the principal melodic pitch for
measure 53 does not bring our two views closer, since our readings of
the voice leading are contrasting. Whereas he displays the B of mea-
sure 52 as a passing note between C♯ and A, I interpret it as a local
upper neighbor of A, below a prolonged C♯ that persists until the
descent to B at 551.

“Schöne Fremde”: stanza 1 (measures 1–7)

As in “Mondnacht,” Schumann uses interruption to shape a binary form in

“Schöne Fremde”: each of the first two stanzas proceeds from I to V (thus
resulting in two iterations of the A1 section), followed by the third stanza (in
which the text responds to the question posed during the second stanza),
which proceeds along a somewhat different route (featuring IV rather than
II➔ as the principal intermediary between I and V) to reach a tonic close.
Compare this structure (13.5) with that of “Mondnacht” (13.1).
 Schumann: Three songs from Liederkreis (op. 39) 155

Example 13.5 Analysis of Schumann: “Schöne Fremde” from Liederkreis (op. 39, no.
6), mm. 1–24.

The B Major tonic root emerges at 41, bringing to a close a circular

trajectory (D♯ G♯➔ C♯ F♯➔ B) that evokes the movement of treetops,
likened by the text to the gods making their rounds. (See 13.6a.)
Though beginning a composition on the tonic’s upper-third chord is
not common, the second pass through A1 (to be explored below)
offers a clarifying modification, while Kopfton ^3 (= D♯) is suitably
introduced at 13. Two preliminary approaches to the dominant
(deploying a descending parallel progression of 63 chords, as shown
in 13.6b) transpire during the span from 42 through 61. Then a more
definitive dominant arrival is attained by way of II➔ between 62 and
71. Whereas the trajectory of 13.6b descends into the texture’s interior
without engaging the higher strands emanating from the tonic’s D♯
(the Kopfton) and B, Schumann’s alternative trajectory accounts for
both: the D♯>C♯>B third-progression of measures 1 through 4 is
complemented by a C♯>B>A♯ third-progression over the course of
II➔ V. The alliance between D♯ and C♯ is reinforced by the reinstate-
ment of D♯ during 62. Observe how 13.6b’s stepwise fourths
B>A♯>G♯>F♯ and D♯>C♯>B>A♯ both recur in 13.6a, now exuding a
hierarchical shape – as B>A♯>G♯, F♯ (in the tenor register) and as D♯,
C♯>B>A♯ (in the soprano register) – due to the assertion of II➔
between I and V.
156 Harmony in Mendelssohn and Schumann

Example 13.6 Schumann: “Schöne Fremde” from Liederkreis (op. 39, no. 6) (a) analysis
6
of mm. 1–7; (b) parallel 3 chords in measures 4 and 5.
(a)

(b)

Though Burkhart applies a I numeral to the B-D♯-F♯ chord at 41, in his

conception it serves as what would be referred to in my terminology as
an unfurled 64 embellishment of a preceding dominant. (Note the num-
bers 4 and 6 above the I numeral in his ex. 8b.) I instead grant this B-D♯-
F♯ the status of a foundational tonic (the onset of the Ursatz). My
argument against Burkhart’s reading will sidestep measure 4 itself,
focusing instead on what seems to me a misguided attempt to read
measures 1 through 3 as a prolonged V. (Observe how only F♯, A♯, and E
and only Roman numeral V appear during that span in his ex. 8b.)
Burkhart develops a context for an A♯>G♯>F♯ third spanning mea-
sures 1 through 3. (This third appears under a slur, annotated “3rd ↆ,”
in his ex. 8b.) That conception cuts off the crucial arrival point at 41,
which is an essential component of my reading, as follows:
m. 1 2 3 4
A♯ G♯ G♯ F♯ F♯

D♯ B♯ C♯ A♯ B

5 5 5
 Schumann: Three songs from Liederkreis (op. 39) 157

The analytical tug-of-war should be between the two endpoint

chords: D♯ minor and B major. If the former prevails, then the
upper A♯>F♯ third-progression is the “leading” line (a filling-in of
the D♯ triad’s upper third), whereas if the latter prevails, the lower
D♯>B third-progression prevails (a filling-in of the B triad’s lower
third). I contend that the latter chord overpowers the former (the
song is in B Major, after all), and thus the A♯>F♯ span above simply
“follows” the trajectory forged by the lower line in upper fifths. In
this context the A♯-to-F♯ third is not a meaningful interval at a deep
level. Burkhart takes an opposing view: “phrase 1 rapidly traverses
the third a♯1–f♯1 (measures 1–3), leaving, as it were, nowhere to go
except back to a♯1, a move finally made at the phrase’s end (measure
7)” (p. 159). Thus his conception seems wrong to me on two counts:
first, his assumption that A♯>F♯ must lead (rather than follow); and
second, that A♯>F♯ must be interpreted as a third-to-root span of
the F♯ dominant, rather than as a fifth-to-third span of the D♯
mediant. In addition, through comparison with my 13.6a it would
seem that he is focusing on events of the chordal interior, neglecting
the prominent D♯>C♯ trajectory above, which, though a straightfor-
ward and often encountered foundational structure for the A1 sec-
tion of a composition, conflicts with his dominant-prolongation
conjecture.
The parenthesized bass F♯ (representing the dominant root) at the
onset of Burkhart’s ex. 8c suggests that the dominant is the begin-
ning of something, whereas for me it serves instead as the end
of something. As also was the case in “Mondnacht,” multiple
dominant-root arrival points dot the tonal landscape. The first two
are followed by a new beginning. Though less intensely in “Schöne
Fremde” than in “Mondnacht” (which even begins with a dominant
“arrival” followed by a new beginning in measure 1), repeated
approaches to the dominant are presented, with arrivals at 71 (adum-
brated in measures 5 and 6), 151, and 231 (which, as the principal
dominant within A2, resolves to I). The fact that the dominant is still
in force at 74 (where the first word of the second stanza emerges)
and during measures 16|17 and 18 (the onset of the third stanza)
does not make the dominant the structural starting point of those
latter stanzas. It appears that Burkhart is using those fuzzy bound-
aries to justify what seems to me a bizarre imagining of dominant
harmony at the song’s onset.
158 Harmony in Mendelssohn and Schumann

“Schöne Fremde”: stanza 2 (measures 7|8–16)

In “Mondnacht” a B>A>G♯ third-progression in E Major facilitates the

restoration of Kopfton ^3 for the A2 section. (See 13.1.) In “Schöne Fremde”
an equivalent line (F♯>E>D♯ in B Major) is deployed in the bass of
measures 7 and 8 to inaugurate the second, somewhat altered presentation
of A1. One might regard that D♯ as the actual sounding of the Kopfton (in a
novel register); or, one might add a D♯ parenthetically in the register where
it first occurred during measure 1 (and where ^2 will emerge during
measure 18). The graph presented in 13.7 is left in a state of incompletion
at that point, allowing you to gravitate toward whichever course you prefer.

Example 13.7 Analysis of Schumann: “Schöne Fremde” from Liederkreis (op. 39, no. 6),
mm. 8–18.

One may note with delight that two features from the first presenta-
tion of A1 are integrated in an unexpected and beautiful way during
this second presentation. I propose that a principal reason why the
Kopfton is not dynamically presented in the obligatory register in the
vicinity of measure 8 is that Schumann was intent upon projecting
the interior B>A♯>G♯>F♯ fourth-progression (featured in 13.6b) now
in a tonic context. (The initiating chord at 81–2 is a broadening of that
at 42. The descending fourth that initially takes place over four beats
now fills over six measures and, notably, is followed immediately by a
restoration of a four-beat presentation, from 142 through 151.) To
support that span, he expands the five-chord circular progression of
measures 1 through 4 into a full eight-chord traversal: from tonic B
through tonic B (as B➔ E A♯➔ D♯ G♯➔ C♯ F♯➔ B). Because the
pattern, as introduced in the song’s opening measures, requires every
 Schumann: Three songs from Liederkreis (op. 39) 159

other chord to be surging, the initial B tonic possesses a dissonant

minor seventh, A♮. Though again there is a danger that the stepwise
descent to F♯ might leave the upper strands in the lurch, Schumann is
careful to reinstate B (at 142) after the fourth-progression has run its
course. B’s successor A♯ and Kopfton D♯’s successor C♯ belatedly
emerge during measure 18, after A2 commences. (The dominant har-
mony of A1 persists beyond the boundary between A1 and A2, with the
tonic harmony postponed until measure 19 and its root, B, in the bass
until measure 20.) During measure 14 (and also measure 15) the II➔
harmony, spelled as C♯-E♯-G♯-B during measure 6, sounds in the more
highly evolved state E♯-G♯-B-D♮. This supertonic occurs above a
retained tonic root and fifth, FB ♯ . Thus I and II➔ collide, as indicated by
the bracket in 13.7. Though the Urlinie’s D♯>C♯ is not conveyed here with
the same prominence as during the first presentation of A1, now that line
features a chromatic element in its foreground realization, as D♯>D♮>C♯
(discernible in the score just above Middle C between 142 and 151).

Though I hear a tonic restoration at 81, I acknowledge that some

listeners might reasonably propose a broad descent by step from domi-
nant root F♯ at 71 through E (83), D♯ (103), and C♯ (123) to tonic root B
at 141. On p. 163, Burkhart describes what transpires here as “prolong-
ing the opening V chord by means of a complete journey through the
diatonic circle of fifths.” His “complete journey” is from F♯ to F♯ though
the “diatonic” key is B. In my 13.7 the “complete journey” coincides
with the key: B to B. His melodic “falling fifth” – C♯ to F♯ – conflicts with
the motivic fourth – B to F♯ – that my reading highlights. My broken tie
from B at 81 to B at 142 is not viable in the context of his dominant-
prolongation reading.
Again Roman numeral I emerges in the midst of Burkhart’s broader
V prolongation (at 141 in his ex. 8b). I instead interpret V as the end of
something (the close of the first statement of A1, with interruption at ^2),
followed by a new beginning on I. Though he labels the C♯-E♯-G♯-B
chord of measure 6 as II♯, the similarly spelled and correspondingly
deployed E♯-G♯-B-D♮ chord of measure 14 is acknowledged by note-
heads in his graphs and by Arabic ♯7 and ♮6 (ex. 8b), but not by a Roman
numeral. Is the surging supertonic somehow suppressed during this
second statement of A1?
160 Harmony in Mendelssohn and Schumann

“Schöne Fremde”: stanza 3 and coda (measures 16|17–30)

Whereas the local tonic prolongations during the two halves of

“Mondnacht” shift from a deployment of I5–6 II to that of I➔ IV, the
principal progressions (incorporating the structurally deepest domi-
nants) in “Schöne Fremde” shift from the use of II➔ during A1 to IV5–6
during A2. Consequently it is not surprising that the tonic restoration of
measures 19 and 20 is not a consonant I but instead a surging I➔.
(That is, the tonic’s pre-surge stage has been elided. Compare with
“Mondnacht,” measures 47–50.) Again the Kopfton is not strongly articu-
lated at the onset. (Note its placement within parentheses during 13.8.)
Despite IV’s deployment, its 6-phase evolution is such that the same
highly evolved surging II➔ chord that emerged during measure 14 recurs
at 224 (E♯-G♯-B-D♮). (The text’s evocation of happiness is underscored by
the fact that I➔ is surging from the onset, IV is of major quality, and IV’s
6-phase II evolves into surging II➔.) This time the dominant that follows
resolves to I, resulting in a PAC.
The coda that follows is notable for the recurring sounding of a D♯ that
echoes the earlier Kopfton. Even the bass D♯ of measure 8, proposed above
as a low-register representative of the Kopfton, is acknowledged by the
coda’s Kopfton reminiscence at 293.

Example 13.8 Analysis of Schumann: “Schöne Fremde” from Liederkreis (op. 39, no. 6),
mm. 20–24.
 Schumann: Three songs from Liederkreis (op. 39) 161

Given how my perspective regarding the tonic has conflicted with

Burkhart’s in the two statements of A1, we do not need to belabor the
fact that, in his ex. 8b, I endorse the Roman numeral I at measure 19 but
reject the broad slur that connects the V numerals of measures 17 and
23. In a point of agreement, we both propose the onset of that tonic as a
consonant chord prior to the emergence of seventh A♮, even if in the
score (measure 19) that initial consonant state is elided.
As a Schenkerian analyst, I would be very concerned if I had shaped
my reading of a work using Kopfton ^5 but then encountered a coda that
emphasizes ^3 (at 243, 263, 283, and 293, all omitted from Burkhart’s
graphs). I wonder if this bothered him at all – if he at any point
considered how an analysis with Kopfton ^3 might transpire. Nor am I
comfortable with the path of descent from ^5 that he proposes. In my
reading of the work, there is no direct voice-leading connection between
F♯ and E, as his ex. 8a proposes. The score conveys quite clearly that F♯
instead ascends to G♯ (in two registers). My (D♯)<F♯ and G♯>E thirds in
13.8 assert that two distinct strands are at play, with the lower strand
((D♯)<E) serving as the principal line. (That is, E is the upper neighbor of
D♯, not a passing note from F♯.) I also question his prolongation of E in
the soprano against E♯ in the bass at 224. In my graph the line descends
from E through D♯ to C♯ (note the voice exchange with the bass) during
IV’s diatonic 6 phase, prior to the emergence of that E♯. Consequently
the local C♯>B>A♯ line, at whose conclusion C♯ is reinstated at 233,
resembles the upper-staff structure of 63 through 72, displayed in 13.6a.
Beyond that, I struggle with consistencies in Burkhart’s interpretation
of measures 21 and 22. In my view measure 21 projects IV, targeted by the
preceding surging tonic, while measure 22 projects IV’s 6-phase chord,
here asserted as II and evolving into a surging II➔ (C♯-E-G♯ into E♯-G♯-
B-D♮). At some points Burkhart seems to concur with this reading: the
half-note bass C♯ in his ex. 8b suggests IV5–6, even if IV is displayed
without further annotation. Jumping to ex. 8a, I discover the “5 –(6)” that
might have been included beside IV in ex. 8b. (The parentheses around
the 6 probably are intended to acknowledge that the 6-phase chord has
been unfurled (my term), thus preventing a literal sounding of sixth C♯
above bass E.) Yet at other points it seems Burkhart is proposing a
contrasting reading: just after the dotted bar line in his ex. 9, one
encounters a provocative slur covering a stepwise fourth descending
from E (with upward stem) to B (with downward stem). Likewise
162 Harmony in Mendelssohn and Schumann

Burkhart’s ex. 8c displays the annotation 5 – 6 – ♮7 during these measures,

thereby not allowing the 6-phase chord to be established as the 5-phase
chord’s successor, but instead showing a local chromatic passing note E♯
emerging in the bass against what might be interpreted as IV8–♮7 (with the
filled-in E-to-B fourth, G♯, and minor seventh D♮). If 5–6 in fact is what
Burkhart intends, then the E>D♯>C♯>B slur of ex. 9 is untenable, since by
that point C♯ would outrank B hierarchically (that is, B would be a passing
note connecting C♯ and the dominant’s A♯ that follows).

“Zwielicht”: introduction and stanzas 1 and 2 (measures 1–23)

Twilight (Zwielicht) is a period of indistinctness, when the paucity of

atmospheric light makes visual perception challenging. One may need to
take a second or third look before discerning what visual field lies before
one’s eyes. Schumann makes a sonic correlation with that visual phenom-
enon by only gradually allowing the song’s foundational structure to reach
its full luster. Though what transpires during measures 1 and 3 from the
introduction could reasonably be interpreted as conveying chords rooted
on tonic E, that tentative perception is not solidified until later: by bass E
during measure 8 from the first stanza, and by bass E during measures 16
and 18 from the second stanza. Given that foundation, the binary structure
displayed in 13.9 ultimately is fleshed out over the course of the song. Its
similarities to 13.1 and 13.5 are obvious.

Example 13.9 Analysis of Schumann: “Zwielicht” from Liederkreis (op. 39, no. 10),
mm. 8–41.
 Schumann: Three songs from Liederkreis (op. 39) 163

The focused E>B>E bass of measures 16 through 18, against the

melody’s G<G♯, gives definitive shape to an initial tonic prolongation
culminating in a surge. Though surges might correlate with a wide
range of poetic images, here the slow tempo, piano dynamic, offbeat
arrival of chord members, and peak dissonant intensity of G♯-B-D-F♮
conspire to evoke key words of the text: schaurig (eerily), schwere
(heavy), Grau’n (horror). In retrospect we come to understand that
components of the first stanza’s musical structure were stated already
during measures 1 through 3. A failure to fully comprehend what was
transpiring at that early point (integrated with the potential emergence
of a competing hypothesis regarding the structure, realized at the onset
of the fourth stanza) appears to have been Schumann’s purposeful
intent, in order to evoke the mood of an eerie twilight, when dangers
lurk in a threatening world.
Though more fully realized during the second stanza than during the
first, measure numbers for both annotate the structure on display in 13.10.
Whereas an initial tonic expansion typically would support a local G>F♯>E
third-progression, here G>F♯>F♮ substitutes, so that minor ninth F♮
sounds within the surging tonic chord (where concurrently G♯ displaces
G en route to upper neighbor A). The D (rather than D♯) deployed during
the interior embellishing chord anticipates the D of that I➔. The surge’s
dissonant focus helps the listener to perceive G<G♯<A as the principal
melodic line of the first four measures in both stanzas, even if a B sounds
after G♯ and a C sounds before A.

Example 13.10 Analysis of Schumann: “Zwielicht” from Liederkreis (op. 39, no. 10),
mm. 8–23.
164 Harmony in Mendelssohn and Schumann

The arrival of surge-targeted IV helps to clarify the broad outlines of the

tonal trajectory, where the composer generally chooses from a range of
options for the principal intermediary between I and V♯. By initiating that
trajectory with I ➔ IV, a continuation as IV5–6 V♯ becomes probable. Given
that the song is in a minor key, IV’s 6 phase is especially susceptible to an
evolution into a supersurge (II⇨), another chord especially evocative of the
text’s content. Thus IV5–6 (A-C-E into A-C-(E)-F♯) might develop further,
into A♯-C-E-F♯ (often referred to as the “French” augmented sixth) or
A♯-C-E-G (the “German” augmented sixth). Because Kopfton G’s incomplete
upper neighbor, A, is featured melodically during IV, the “German” option’s
G is tapped for use in the melodic descent from that A to the dominant’s ^2
(F♯). Three unfoldings transpire in the vocal melody: A>F♯ (into which local
passing note G is inserted, during measures 11–12 and 19–20), G>E (mea-
sures 13 and 21), and D♯<F♯ at the cadence (measures 14–15 and 22–23). The
latter third’s D♯ disconcertingly emerges in the vocal line over the piano’s
prolonged E, which holds out until the second beat of the following measure.
Likewise G descends to F♯ in the vocal line while G is prolonged by the piano.
Eerie!5 As is typical of supersurging chords, the pitch C ultimately takes up a
position in the bass (at 143), so that C>B and A♯<B half steps may coordinate,
resolving the augmented sixth into the dominant octave.
The introduction’s first measure is audacious because listeners cannot be
sure – even after repeated hearings of the work – what it represents.
Through comparison with measures 3 and 4 (which proceed as E➔ A),
one might interpret C♯-E-G as an incomplete presentation of A♯-C♯-E-G
proceeding to B-D-F♯ (thus F♯➔ B, inaugurating a segment of the circle of
fifths). Schumann reinforces that conception at the onset of the song’s
fourth stanza (measures 33–34, where A♯ is explicitly presented, as well as
root F♯). Yet through comparison with the closer-at-hand measure 8, it
would seem that an E minor tonic triad may be intended at the onset,
perhaps with two distinct chords transpiring during the measure: I corre-
sponding to the initial sounding of G and E, followed by II➔ with the
emergence of C♯ (which in this scenario would imply A♯ as well).
Those alternative readings both flow into the subdominant during mea-
sure 4. By the time the local downward cascade of the lower line (E>C>A
filled in by passing notes) reaches IV’s root at 51, 6-phase F♯ has emerged
above it. From this point a fresh problem takes hold (problematic in compar-
ison with the well-behaved structure projected during the first two stanzas
that follow). We noted above how the melody’s unfolded AF♯ third is followed
by two additional thirds – GE and FD♯♯ – resulting in the sounding of ^2 exactly
 Schumann: Three songs from Liederkreis (op. 39) 165

where it “should” sound: in the vocal melody above dominant root B at the
HC (measures 15 and 23). Yet in the introduction the G and F♯ instead take
up positions in the texture’s interior (during measure 6). Though the V♯ at
63 may be accepted as resolving the preceding II⇨, one needs to forgive
both the absence of soprano F♯ and the tardiness of bass B. It appears that
Schumann has used these inherent defects to motivate a quick reiteration.
That dominant is followed immediately by a renewed thrust from the tonic
to the subdominant (I➔ IV), and upon the arrival of IV’s 6 phase, upper
neighbor A is restored in the soprano (now an octave lower), after which
the G and F♯ resound within the melody. Schumann adds one further
creative wrinkle as this reiteration proceeds: the subdominant 6-phase
chord substitutes chromatic F♮ for diatonic F♯ during 72.
A further nuance of Schumann’s writing is documented in 13.11. The
left model shows the evolution of IV’s 6-phase chord into a supersurge. The
D♯ from the end of measure 12 is not displayed, since that pitch serves as a
local neighbor to chord member E (which sounds after but not before the
D♯, acknowledged by the parentheses around E in the model). Four of
II⇨’s chord members sound during the first beat of measure 13. In the
right model, D♯ is projected more prominently, though its chord is dis-
played using filled-in noteheads to indicate a subordinate layer of struc-
ture. The sonority at that point has long been recognized as what is called a
common-tone diminished seventh chord.6 Locally it resolves into the C-E-
G triad that follows. Yet that vignette occurs as such only because the A♯
that normally would follow directly after A (as on the lower staff of the left
model) is withheld temporarily. In such a context, the C major triad plays
no independent harmonic role, but instead projects the fifth, seventh, and
ninth of the emerging II⇨ harmony. The measure numbers indicate that
this idiosyncratic realization of a conventional supertonic evolution takes
place both during the introduction and during the fourth stanza.

Example 13.11 Schumann: “Zwielicht” from Liederkreis (op. 39, no. 10), relationship
of mm. 4–6 and of mm. 36–40 to mm. 11–15.
166 Harmony in Mendelssohn and Schumann

In broad terms Ferris and I share the same conception regarding the
initial stages of the song’s projection: “The song is based on a musical
phrase that is not clearly defined when we first hear it, gradually comes
into focus as it is varied in each succeeding stanza, and ultimately turns
out to be something quite different than what we had first imagined”
(p. 140). Though I will offer a contrasting interpretation of the “ulti-
mately turns out” phase (regarding events of the third and fourth
stanzas as aberrational rather than as culminating), we agree on the
path toward increasing clarity through the second stanza.
Ferris suggests that the introduction lacks a “coherent harmonic
progression until the cadential dominant at the very end” (p. 143). His
block-chord reduction of the musical content (ex. 2a on p. 145) is
annotated by question marks in place of Roman numerals until the V
at 73. Though I agree that how the A chord of measure 4 is attained is
open to question, my ear more quickly establishes E Minor as a tonal
center, through the A<A♯<B bass of measures 5 and 6 (including an
augmented-sixth approach to the dominant) and the A<B bass of
measure 7 (including a distinctive ♮II approach to the dominant).
Ferris’s assessment is hampered by a misrepresentation of measure 6
in his ex. 2a. The outer-voice EA♯ diminished fifth resolves normatively to
D♯
B before DG♯ arrives (all contorted from their conventional formulations
by metrical displacements). The display of the bass there as A♯>G♯
omits the introduction’s goal root, B, which is reiterated at 73–4.
Likewise the reduction displayed for measure 1 clumps all the important
pitches into one vertical entity (C♯-E-G). I suggest that one ought to
leave open the possibility that the E-G dyad operates on a deeper plane
than the C♯ that emerges at beat 3. (That interpretation is supported by
the three Es and two Gs that sound during the first two beats of measure
8.) So, instead of projecting a vague ? as my interpretation of the open-
ing measures, I propose that two potential trajectories (both developed
further as the work unfolds) are at play: E ➔ A (reinforced by the onset
of stanzas 1 and 2) and F♯➔ B E➔ A (reinforced by the onset of stanza
4). Both confirm the A chord of measure 4, after which its 6-phase
transformation into F♯⇨ is a normative turn of events in the context of
IV5–6 in E Minor, despite Ferris’s timid postponement of that key’s
assertion until 73.
Though I would want to fill in some of the question marks annotating
Ferris’s reading of the first stanza (his ex. 2b) with the Roman numerals
 Schumann: Three songs from Liederkreis (op. 39) 167

that finally appear in his reading of the second (ex. 2c), his conception
now begins to coincide with what I display in 13.10. As described above
with reference to measure 1, I question the inclusion of C♯ in the chord
displayed for measures 8 and 16. For the first two beats of those
measures (in the context of a slow tempo), the sounding chord presents
only Es and Gs, fully compatible with a tonic interpretation. (Note also
that measure 15 concludes with a surging B➔ chord, which makes tonic
E the expected successor.) I also find his [V] symbol at measure 18
problematic. For me, the chord is a surging tonic (I➔). Thus the tools
Ferris uses for his harmonic analysis are impeding a realization that the
tonic is conveyed at the onset of these stanzas. Though the three symbols
iv, ii, and Aug. 6 make measures 19 through 22 seem less integrated than
do the hierarchically differentiated symbols I deploy in 13.10, they
conform to the conventional practice of harmonic analysis, which I
propose is inadequate for sophisticated musical thinking.

“Zwielicht”: stanzas 3 and 4 (measures 24–41)

At first, stanza 3 exhibits even greater confidence in projecting its chordal

content than did its predecessors: note how tonic E’s B-D-F♯ embellishing
chord during measure 25 sports a B<C♯<D bass against the D>C♯>B
melody (thus forming a voice exchange). Yet that confidence appears to
have been misplaced, since what follows in measures 26 and 27 amounts to
a major mishap. Whereas E➔ proceeding to A is expected (in accordance
with the precedent of the first two stanzas), instead G♯➔ C♯ transpires.
(Observe how the vocal melody’s G♯<B third from measures 10 and 18 is
stretched an extra third to become G♯<D♯ in response to the startling G♯ –
instead of E – that emerges in the bass during measure 26.) I interpret this
event as a seismic shift – an external force that powerfully jolts the
progression, moving its I➔ IV trajectory to a higher tonal plane. In this
context the C♯ chord should not be interpreted as ♯VI♯5 in E Minor, but
instead as a diatonic IV that has been abruptly hoisted a major third higher,
as if the composition might proceed to a cadence in G♯ Minor. (It is
characteristic for the chords buffeted by a seismic shift to retain their
original functions: here IV is hurled onto a different tonal plane.) Yet
that alternative outcome’s viability is short-lived: a reciprocal lowering by
a major third eventually restores the initial tonal plane. For the remainder
168 Harmony in Mendelssohn and Schumann

of the third stanza the harmonic progression essentially is put on hold, with
inverted and root-position C♯ chords sounding repeatedly (interspersed
with embellishing chords). Schumann instead forges a link between this
hoisted IV from the third stanza and the restored E Minor IV from the
fourth stanza (measures 27 and 36), thereby circumventing a tonic restora-
tion at the onset of the fourth stanza. By this means the two stanzas
together constitute one structural entity, as shown in 13.9.
The unexpected emergence of a seismic shift has allowed Schumann to
offset one of the disadvantages inherent in the way the text is set. Over the
course of measures 8 through 25 the listener develops the expectation that
each stanza will project the same local harmonic progression – that of
13.10. Consequently the musical content likely would relate to the poem in
a general way rather than responding to specific textual material from any
one stanza. Yet the mishap of measure 26 (the G♯➔ chord displacing an
E➔ chord) emerges against the verb trau’ (trust). The third stanza’s
admonition that one’s friends may be tückisch (deceitful) coordinates
with the musical deceit of sabotaging the E Minor tonal center!
Whereas two times (at the end of the introduction and the end of the first
stanza) a

the third stanza’s concluding

C♯ chord is followed by a new stanza’s F♯➔ B E➔ A

whose E➔ chord is not invested with an assertive tonic role. Instead, a

circular progression (emphasizing C♯, B, and A in turn) connects the
hoisted (C♯) and original (A) subdominant chords. (Compare this progres-
sion with how Schumann initiates “Schöne Fremde”: D♯ G♯➔ C♯ F♯➔ B.)
What makes this unexpected turn in the song’s trajectory so stupendous is
that it is accomplished without altering the vocal melody’s pitches.
(Compare the vocal lines of measures 16–19 and 33–36.) That in turn
helps explain why the song’s enigmatic opening is so effective: measure 1
could correlate equally well with measure 8 (root E) or with measure 33
(root F♯). The conceptual dilemma lies in the fact that those later contexts
provide a strong bass support (E or F♯), whereas measure 1 transpires with
no such defining component. Consequently it inherently adumbrates both
of the principal stanza-initiating trajectories within the song.
It was noted above how the foundational descent from Kopfton G’s
incomplete upper neighbor to ^2 (A>G>F♯) initially was neglected during
 Schumann: Three songs from Liederkreis (op. 39) 169

the introduction, resulting in a second approach to the dominant (during

measure 7) during which that trajectory belatedly sounds in the melody.
The melody’s A during measure 36 equally deserves a stepwise continua-
tion through G to F♯, now proceeding onward to E for a PAC. Yet after the
G at 391 the vocal melody again descends (now more deeply) into the
chordal interior: the A♯<B shown in the tenor register of 13.10 is appro-
priated for the voice. (This departure from the structure that was projected
during the first two stanzas may be understood as Schumann’s response to
the poetic text: “Much gets lost in the night.”) This time no second chance
is offered. Analysts (and performers) are left with two options. One might
propose that the song remains enigmatically unfinished.7 Or, the line
continuing from G to F♯ and then E could be understood as taking place
in the tenor register (as was the case also during measure 7), where all the
requisite pitches in fact reside in the piano part. Though I favor the latter
interpretation (having succeeded in projecting the line in that register,
below cover tone B, on my fortepiano), this is not a matter about which a
compelling case could be made one way or the other.

Though Ferris and I agree that Schumann “departs from his underlying
model” during the third stanza, his interpretation of what happens as
“essentially turn[ing] the stanza into a contrasting B section” (p. 147) is
a bolder reading than mine. I instead hear a correlation established
between internal points within stanzas 3 and 4 (the hoisted and restored
IV chords), making one indivisible entity out of the combined content of
both stanzas (thus incompatible with his notion of ternary form). An
unfortunate consequence of Ferris’s failure to hear the first three mea-
sures of stanzas 1 and 2 as a projection of I ➔ (= E–B–E➔) in E Minor is
his willingness to interpret the internal B embellishing chord as a local
tonic during stanza 3. Though I would not expect an analyst working
with conventional harmonic tools to replicate my seismic shift inter-
pretation, I might have hoped not to see the Roman numeral i appear
below the B-D-F♯ chords of measures 25 and 34. Consequently my
hoisted IV (now suddenly in G♯ Minor) at measure 27 is interpreted
as ii in B Minor. At this point Ferris’s and my hierarchies are totally at
odds: the B root of measure 34 is for me a passing note between the C♯ of
measures 27–31 and the A of 36, within a circular progression traversing
a descending major third, whereas for him it is the endpoint of a ten-
measure B Minor tonicization. Ferris calls measure 34 the “climax” of
the composition, featuring “the first tonic resolution of the song”
170 Harmony in Mendelssohn and Schumann

(p. 148). By placing his reductions of the introduction and the four
stanzas in vertical alignment in his ex. 2, he intends to correlate the C♯-
E-G sonorities of measures 1, 8, 16, and 24 and the F♯-A♯-C♯-E-G of
measure 33, proposing that in the fourth stanza “the opening harmony
is finally revealed as the dominant seventh chord that has been obliquely
implied in each of the preceding stanzas” (p. 149). I instead regard
measure 33’s structural purpose as strongly contrasting that of its pre-
decessors during the first three stanzas, since the aberrational event of
measures 26 and 27 has not yet been offset. From this perspective the
content preceding subdominant A’s restoration at 361 is fundamentally
different from what transpired earlier. Schumann succeeds in making
the passages seem similar, since he deftly maintains the same melodic
content at the fourth stanza’s onset. But that is a deception (the gist of
stanza 3’s text). The subdominant will be “newly born” at the word
neugeboren!
Ferris’s treatment of the closing measures likewise seems to me off the
mark. I do not endorse his converting beat 2 of measure 40 into a
downbeat (as displayed in his ex. 5 on page 152). The harmony shifts
at 401 even if the vocal and piano entries are staggered. Nor do I
subscribe to the notion that “the phrase appears to break off . . . on the
second beat of measure 39” (p. 151). The pacing of the chordal entries
corresponds closely to what happened during stanzas 1 and 2 (taking
into account the elaborate projection of II⇨ as proposed in 13.11). The
dominant (with delayed bass B and with resolution of the 64 embellishing
chord timed similarly to how it occurred in the piano part of measures
15 and 23, rather than their peppier vocal equivalents) again occurs
during the phrase’s eighth measure, necessitating a ninth measure for
the concluding tonic. As I see it, the only matter of contention regards
the F♯ and E successors of the vocal melody’s G at 391. Ought one to
regard those pitches as neglected, or instead to transfer the G to the
interior piano G at 393 for descent through F♯ to E in that register? Ferris
boldly asserts that “the cadence does not resolve the final phrase, and
thus does not really provide closure for the song” (p. 151). Thus he
regards Schumann’s work as a “romantic fragment” (p. 153). I am more
forgiving of the oddities of the close.
14 Schumann: “Im wunderschönen Monat Mai”
from Dichterliebe (op. 48, no. 1)
in response to deborah stein

Whereas Allen Cadwallader (to whom I respond in chapter 10) and I

attended classes together at the Eastman School of Music, Deborah Stein
and I both enrolled in seminars at Yale University between 1977 and
1979. Her research since then has focused on German lieder from the
nineteenth century, including a book devoted to extensions of tonality in
Hugo Wolf’s songs. She also spearheaded an especially creative and useful
pedagogical initiative: the volume Engaging Music: Essays in Music
Analysis, with contributions from twenty-one leading North American
analysts. Stein’s own contribution to her volume is a study of ambiguity,
featuring an analysis of Schumann’s “Im wunderschönen Monat Mai”
from Dichterliebe.1
An important proviso of Stein’s undertaking is that the analyses
should be accessible to “upper-level [undergraduate] music students.”
Consequently my own analysis of the song below (the unboxed por-
tions of this chapter) is pitched at a level that I would expect my own
upper-level undergraduates to be able to follow. (Lower-level under-
graduate students receive a thorough grounding in my harmonic
theory during their third semester of study, which I teach.)
Of course, Dichterliebe has been addressed often in the analytical litera-
ture, with some authors assessing individual songs and others attempting
to come to terms with how Schumann organized the entire cycle. Several of
these authors make cameo appearances in the endnotes, inserted where
some extra reinforcement or a contrasting viewpoint might enrich the
central discussion.

The introduction (measures 0|1–4)

When purchasing an automobile, it is useful to undertake a careful inspec-

tion to make sure that everything is in proper working order. If a problem
surfaces during a test drive, one would insist that it be resolved before
bringing the car home. In encountering musical compositions, one shifts
from a consumerist attitude to an artistic stance, allowing composers great 171
172 Harmony in Mendelssohn and Schumann

leeway in deploying the devices that were stylistically viable in their time to
create works that convey sentiments in memorable and moving ways.
When first encountering the initial three pitches of “Im wunderschönen
C ♯





Monat Mai” – A ♯ – one might misjudge the date of composition by
D



a century or more. What is such a dissonant concoction doing in a
composition by Schumann? (As the initial sonority of a nearly thirty-
minute song cycle, no less!)
If we open the door a bit wider (through the end of measure 1), the sense
of Schumann’s scintillating sonority begins to reveal itself. Against stable
bass D, A♯ resolves upward to B and C♯ resolves downward to B, with a
doubling of D and an F♯ joining in as well. A first-inversion B minor chord
has emerged out of the cacophony.
From the traditional grounding in tonal composition that Schumann
learned in his youth, we may understand tensions like those generated by
A♯ (augmented fifth above bass D) and C♯ (major seventh above bass D) as
the second phase of a three-phase initiative:

Preparation – Suspension – Resolution

We thus may be drawn to the upbeat preceding measure 1 (prior to D’s

arrival in the bass) to get a sense of what might have served as preparation
for the upcoming suspensions. Taking into account that dissonant A♯
resolves upward in the manner of a chordal third (leading tone), one
might propose that an F♯-A♯-C♯ sonority is implied by the lone C♯ that
inaugurates the composition.
Analysts are sharply divided regarding the appropriateness of making
such additions to the actual sounding pitches of a composition, falling into
camps that may be referred to as literalist and imaginative. This chapter
espouses an imaginative perspective. We may be guided by some words of
wisdom from one of Schumann’s most illustrious predecessors, C. P.
E. Bach, who advises: “There are many things in music which, not fully
heard, must be imagined . . . Intelligent listeners replace such losses men-
tally, and it is primarily such listeners whom we should seek to please.”2
Because music involves spreading out individual harmonies over time
and may also feature backtracking for repetitions, it often is useful to
complete a harmonic analysis below block chords notated on staff paper
rather than in the score. Though most of these chords will be extracted
from the composition, this practice also provides an opportunity to display
imagined pitches or chords as the peers of the composition’s sounding
Schumann: “Im wunderschönen Monat Mai” from Dichterliebe (op. 48, no. 1) 173

elements. In 14.1 the forty-six noteheads that appear before the vocal
entrance in Schumann’s score are represented by fifteen noteheads, of
which two are filled in to indicate their status as embellishing notes;
while seven additional noteheads have emerged through imaginative
capacities grounded upon a sense of how tonal music of Schumann’s
era tends to go. An x in the row of measure numbers marks the
chords that are imaginatively conceived. (Though the practice will not
be continued later, in this first example a box encloses the imagined
pitches.)

Example 14.1 Analysis of Schumann: “Im wunderschönen Monat Mai” from

Dichterliebe (op. 48, no. 1), mm. 0|1–4.

The imaginative procedure introduced above to generate F♯-A♯-C♯ has

been expanded in two ways in 14.1. First, since the B minor chord of
measure 1 sounds in its first inversion, the preceding F♯ chord is displayed
in 42 position, the typical arrangement when a surging F♯-to-B root succes-
sion occurs with bass D sounding in place of B in the latter chord. Second,
since the progression of roots is shaping up to be F♯–B–C♯, we likely will be
inclined to interpret the harmonic trajectory as arising from the tonic to
the dominant by way of the subdominant. Both Schumann’s deployment of
a three-sharp key signature and the fact that the B chord is of minor quality
lead one to understand the progression as transpiring in F♯ Minor, thus
warranting the imaginative projection of a foundational F♯-A-C♯ tonic
chord preceding the tonic’s surging state (thus I ➔ IV V♯37 ). The omission
of a progression’s initiating tonic chord is so common in tonal music that
Heinrich Schenker coined a term for the phenomenon: auxiliary cadence.3
Though it may seem a stretch for an analyst to so actively engage in
composing “missing” material from a composition, Schumann himself in
part confirms this reading later: as this chapter will clarify below, the
imagined F♯➔ chord proceeding to B proposed for the introduction is
realized concretely during the first stanza, at 92 to 101!
174 Harmony in Mendelssohn and Schumann

The introduction’s melodic focus hinges on two descending seconds:

C♯>B and F♯>E♯. In 14.1 the C♯-to-B second is emphasized through the
annotation ^5 >^4 above the system, whereas an arrow is deployed to convey
that the high F♯ is actually an “alto” pitch temporarily transferred to the top
of the texture. The dominant arrival point in measures 2 and 4 comes
across as a dissonant chord whose resolution is stymied. (Augmented
fourth EB ♯ persists in a suspenseful state but does not resolve.) Over the
short term, Schumann gets around this dilemma by traversing a contrast-
ing path once the voice enters (as will be explored below). Yet he eventually
reaches the same V♯ 37 stymie point – in measures 13 and 15 and again in
measures 24 and 26. After the former, the same solution as occurred at the
juncture between the introduction and the first stanza is applied again.
After the latter, a contrasting path (in the cycle’s next song) is substituted in
order to get beyond the impasse.

Stein’s essay on musical ambiguity begins with an unintended analytical

ambiguity: the very first Roman numeral annotating the score provided
in her ex. 7.1 reads as ii6 (measure 1, beat 2), whereas the textual
commentary addressing that passage instead indicates iv6 (page 78).
Looking for how recurrences of the chord in the same context are
analyzed later in the work is of no help in resolving the discrepancy,
since those chords are not annotated by a Roman numeral in her
presentation.4 Despite this confusion, I suggest that Stein intends for
the D-F♯-B chord of measure 1 to be understood as iv6 in F♯ Minor, not
ii6 in A Major (a key that arrives later in her analysis).5 If that is the case,
her reading is in accord with mine in 14.1, taking into account the
inherent differences in our styles of analytical notation.
My imaginative interactions with the score contrast Stein’s literalist
approach. For me, measure 1’s first beat is an exceedingly important
venue for analytical endeavor. By definition a suspension is a pitch that
once was meaningful in a way that is thwarted by its prolongation into a
new context. For me the concurrent sounding of A♯ and C♯ where they
clearly do not fit as chord members triggers an analytical backtracking
to reconstruct a context (prior to the first downbeat) in which they
might have thrived.6 In contrast, Stein’s first Roman numeral (intended
as iv) does not appear until the suspensions have about run their course,
on beat 2 of measure 1, rather than at the point where the subdominant
harmony’s assertion is initiated, at beat 1.
Schumann: “Im wunderschönen Monat Mai” from Dichterliebe (op. 48, no. 1) 175

The first stanza through the arrival of the subdominant

(measures 4|5–10)

An appealing way for a composer to correlate a song’s introduction and its

first stanza is to shape the former as a miniature version of the latter,
deploying the same foundational harmonic progression in both contexts.
Note how the IV of the introduction’s measures 1 and 3 holds forth in the
first stanza’s measure 10, and how the V♯37 of measures 2 and 4 recurs in
measures 13 and 15. Unless something emerges later to challenge the
assertion, let us proceed under the hypothesis that a I–IV–V♯37 harmonic
progression in F♯ Minor fulfills such a double duty.
Perhaps because the initial tonic is only hinted at rather than robustly
asserted, Schumann does not backtrack to that tonic at the dividing point
between the introduction and the first stanza. The root progression works
as follows, with the ellipses conveying crucially that each B chord should be
understood as following after an initial F♯ tonic:

Introduction
F♯ ➔ B C♯ (before the composition
begins through
measure 2)
... B C♯ (measures 3–4)
Stanza 1
... B E➔ A (measures 5–6)
... B E➔ A (measures 7–8)

After twice proceeding directly to dominant C♯ during the introduction,

Schumann supports the opening vocal line by proceeding instead to the
mediant (F♯ Minor’s diatonic A-C♯-E).7 That path is explored in 14.2.

Example 14.2 Analysis of Schumann: “Im wunderschönen Monat Mai” from

Dichterliebe (op. 48, no. 1), mm. 4|5–8.
176 Harmony in Mendelssohn and Schumann

A composer often will deploy several individual chords to prolong

one foundational harmony. For example, in the tonicization of, say, a
VI harmony, a local progression from the root of VI to its dominant
(perhaps incorporating some other chords along the way) and back
will serve to dynamically convey the sense of VI within the broader
harmonic trajectory. Third-relationships sometimes take on a role
similar to fifth-relationships in such initiatives, especially during the
nineteenth century. To prolong an F♯ minor tonic harmony, one may
proceed from an F♯ minor chord to an A major chord (perhaps
incorporating some other chords along the way) and then back to
F♯. The progression in 14.2 shows the tonic-to-mediant phase of
such a prolongational strategy, wherein the circle of fifths serves as
the means of locomotion. The surging F♯ chord is a local event
within the circle’s initiating F♯-to-B fifth. It does not convey the
key of F♯ Major. (All of the chords in measures 5 through 8 are
diatonic in F♯ Minor.)
The broader context in which that F♯-to-A trajectory transpires is shown
in 14.3. Contrasting the introduction’s bass D at 11, the subdominant of
measure 10 emerges in its root position, with bass B, and so Schumann is
able to deploy the strongest possible bass motion during the I-to-IV
succession: root F♯ to root B. This is accomplished in the context of a
surging tonic: after mediant A-C♯-E, the restored tonic of 92 surges from its
onset, replacing diatonic A with A♯ and retaining the E introduced by the
mediant. This is a vigorous and potent harmonic succession. No longer
must the analyst excuse tonic chord members for their truancy, since F♯,
A♯, C♯, and E all have reported for duty.8

Example 14.3 Analysis of Schumann: “Im wunderschönen Monat Mai” from

Dichterliebe (op. 48, no. 1), mm. 4|5–10.
Schumann: “Im wunderschönen Monat Mai” from Dichterliebe (op. 48, no. 1) 177

In my interpretation of the song’s harmonic content, the introduction’s

I ➔ IV V♯37 progression expands to encompass the entire first stanza.
Given the luxuriant framework afforded by that time span, Schumann
undertakes a creative double deployment of a B chord. Clearly B-D-F♯
in measures 5 and 7 sounds in a context that contrasts that of the B-D-F♯
in measure 10. Which of these chords fulfills the same role as the B
chord of measures 1 and 3, as the principal internal point between the
tonic and the dominant? The only viable answer, I think, is the latter B
chord, in measure 10.
If my proposal of a foundational F♯-A-C♯ tonic chord in F♯ Minor as
the predecessor of the first stanza’s content is rejected, then my analysis
falls apart completely. But if one posits such a primordial chord, then its
means of prolongation and evolution through measure 9 may be con-
veyed concisely as follows:

F♯ A C♯ (prior to the stanza)

A C♯ E (measure 6, repeated in measure 8)
F♯ A♯ C♯ E (measure 9)

From this perspective the A-C♯-E chord does not warrant a III label
indicating an asserted scale-step along the path to V♯37 , but instead falls
entirely within the orbit of the F♯ tonic chord.9 Stein assigns a more
prominent role to the A chord, not only shifting her analysis into the key
of A Major, but also proposing that the crucial B-D-F♯ chord of measure
10 functions as ii in A Major. My analysis interprets the chordal
hierarchy in a much different way, with the F♯ chord prevailing through
measure 9, where, in a typical harmonic evolution, it surges just prior to
the progression’s conquest of IV. Though Stein shows iv as one of two
alternative readings for the chord of measure 10, her analysis goes blank
in measure 9. To me that measure’s F♯➔ chord is the culmination of the
stanza’s broad tonic prolongation.
I interpret the foreground scurrying up to the mediant chord that
transpires during measures 5 and 6 as a circular – not a harmonic –
progression, warranting letters indicating roots (as in 14.2) rather than
Roman numerals (as in Stein’s ex. 7.1). In this delightful moment a witty
Schumann allows a B chord like those encountered during the intro-
duction to veer off in a different direction. Though Stein binds small
clusters of chords via Roman numerals in her analysis, how the goal A
chord correlates with the prior F♯ Minor focus is left unexplored.
178 Harmony in Mendelssohn and Schumann

The broad progression of the first stanza (measures 5–15)

In the spirit of exploring how the first stanza expands upon the struc-
ture that occurs during the introduction, it is reasonable to expect that
the IV harmony attained at 101 will undergo some sort of prolongation
before the dominant’s arrival, which in fact does not occur until 131.
Whereas there are many ways to prolong a IV harmony, in the context
of this particular song one option is especially attractive and appropri-
ate: to again proceed to the upper-third chord and back. Schumann
does everything possible to emphasize the relationship between those
two chords in his song. A connective chord at 91 (shown using filled-in
noteheads in 14.3) leads into an F♯➔ tonic chord that targets subdo-
minant B at 101. Then the procedure is repeated a third higher: another
connective chord at 111 (again shown using filled-in noteheads, now in
14.4) leads into an A➔ chord that targets the subdominant’s upper-
third chord on D at 121.

Example 14.4 Analysis of Schumann: “Im wunderschönen Monat Mai” from

Dichterliebe (op. 48, no. 1), mm. 10–12.

By this point in the song we have grown accustomed to hearing exact

repetitions of content (for example, with measures 3 and 4 repeating measures
1 and 2). Schumann departs from that practice in the connection between the
subdominant and the dominant. Whereas the return from upper-third D back
to B is entrusted to a mere sixteenth-note B sounded in the bass at the end of
measures 12, measure 14 offers a potent confirmation that this return actually
has occurred. That upper-third D chord is not reinstated during measure 14.
(Both the tonic’s and the subdominant’s upper-third chords are displayed
within large parentheses in 14.5.) By this point Schumann has begun to
concern himself with another important initiative: making measures 14 and
15 sound like measures 3 and 4, so that the second stanza may commence in a
context that exactly matches that of the first.
Schumann: “Im wunderschönen Monat Mai” from Dichterliebe (op. 48, no. 1) 179

Example 14.5 Analysis of Schumann: “Im wunderschönen Monat Mai” from

Dichterliebe (op. 48, no. 1), mm. 6–15.

Another notable feature of the expansion concerns the vocal melody.

In 14.1 observe how the foundational melodic line ^5 >^4 is complemented
by the emergence of a high F♯ (embellished by G♯) in the soprano
register at the end of measure 1 (and measure 3). The first stanza’s
expansion of IV likewise proceeds melodically upward to F♯ (now with
a G♮ embellishment in measure 12, “corrected” to G♯ by the piano later
in the measure to correlate exactly with measures 1 and 3). Given the
premise that the harmonic progressions of the introduction and the
first stanza are foundationally the same, encountering this distinctive
trait at both locations instills a confidence that the analysis is attuned to
Schumann’s conception.
In that all the song’s harmonic content has sounded by measure 15, we
may now speculate regarding how Schumann’s chordal progression con-
veys the sentiments of Heine’s poem. Three factors – two unequivocally
positive and one still to be resolved – seem to me central:

Winter gives way to Spring

A loveless state gives way to Love

Reciprocation is as yet Uncertain

Though many commentators have noted that the song contains no F♯

minor chord, if Schumann assigns winter and a loveless state to that founda-
tional harmony, then its absence is to be expected: winter has yielded to
spring; the former loveless state has given way to a burgeoning love. The
imagery of blooming buds and singing birds (at some height above the
earth’s surface) both convey a rise above the dormant state. In harmony
there are two principal ways to bring a minor triad to life: its parallel major
180 Harmony in Mendelssohn and Schumann

(as in F♯-A-C♯ to F♯-A♯-C♯, perhaps with added seventh E), and its relative
major (as in F♯-A-C♯ to A-C♯-E). I propose that the song’s three founda-
tional triads each undergo one or both of these transformations:

Wintry, loveless chords Springtime, love-impregnated chords

F♯ minor F♯ major (with minor seventh) A major

B minor D major
C♯ minor C♯ major (with minor seventh)

The song captures a moment when the love has been revealed though
not yet reciprocated. What sort of harmonic progression might best
convey this blend of yearning, hope, suspense, and uncertainty? I suggest
there is no more suitable choice than to proceed from the tonic through
the subdominant to a dominant seventh (V ♯37 ). The lack of resolution in
no way indicates a negative outcome. It is simply that our snapshot of the
scene occurs at that delicate moment when reciprocation is as yet
uncertain.

Schumann’s deployment of IV’s upper-third chord matches what tran-

spired in the context of I (now lacking the surge, of course, since IV
cannot surge to V♯). The pitches may be displayed as follows:

B D F♯ (measure 10)
D F♯ A (measure 12)
B D F♯ (end of measure 12, confirmed in measure 14)

Sixteenth-note B at the end of measure 12 (repeated at the end of

measure 14 in a more overtly subdominant context) reinstates the
subdominant root.10 Though Stein has placed a iv numeral below
measure 10 (as one of two options) in her ex. 7.1, the reinstatement of
iv after the excursion to its upper-third chord is noted neither in
measure 12 nor in measure 14. Thus neither of the two foundational
harmonic successions that I display in 14.5 (I to IV, deploying a surge;
and IV to V♯37 ) has a counterpart in Stein’s analysis of the first stanza.
Whereas her summary of the stanza emphasizes a “strong arrival on A
major followed by the return to V7 of F♯ minor” (p. 78), I instead would
emphasize how an A major chord (not key) helps to prolong the F♯ tonic
and how the dominant’s arrival proceeds from IV, as it did during the
introduction.11
Schumann: “Im wunderschönen Monat Mai” from Dichterliebe (op. 48, no. 1) 181

Given my reading of how the harmonic progression conveys the

meaning of the poem, I do not sense the “pain of lost love” or “the
poet’s ultimate lack of fulfillment” (p. 78) that Stein proposes. Granted,
during the early stages of any love relationship one might harbor self-
doubts and worries about rejection. Yet the transformation of all three
of the progression’s principal chords through parallel and/or relative
relationships into chords of major quality encourages interpretation in
an optimistic light. As the song ends, the season is spring and the poet is
in love, awaiting the loved one’s response – bracing for the possibility of
12
a No, but full of hope for a Yes.

The second stanza (measures 15|16–26)

There is not much to say analytically about the second (final) stanza. The
open-ended content of the first stanza is left intact during its repetition.
Consequently there is no conventional cadence. The melodic C♯>B second
(indicated by ^5 followed by ^4 in 14.5) leaves a dangling dissonance
sounding at the movement’s close. This is, of course, an exact parallel to
the impasse at the end of the introduction. Schumann responds in a similar
way: the mediant, now established as the key of A Major, again comes to
the rescue in the next song (“Aus meinen Tränen sprießen”), where a
C♯>B>A third like that of 14.2 is traversed multiple times during that
song’s opening eight measures (supported by a harmonization that both
begins and ends on an A major tonic chord), thereby resolving the impasse.
In retrospect, we come to realize that the cycle’s first song is a fragment – a
beginning whose ending (if one occurs at all) does not fall within the song’s
boundaries.

I have resisted expanding this chapter into a broader study of

Dichterliebe as a whole.13 What I state above regarding the C♯>B>A
third from 14.2 migrating to the second song is tantalizing, as is a look at
how that song’s middle section relates to the tonic chord:
A C♯ E (measures 0|1 through 8)
C♯ E♯ G♯ (measure 12)
A C♯ E G♮ (end of measures 12 and measure 13)
182 Harmony in Mendelssohn and Schumann

Though this reading contradicts one of the most studied analyses in

Schenker’s Free Composition (fig. 22b),14 it so clearly relates to how I
propose both I and IV are expanded in “Im wunderschönen Monat
Mai” that my analytical claims are bolstered significantly. I close leaving
these loose ends dangling, awaiting further analytical contextualization,
just as Schumann closed his first song with loose ends.15 In retrospect,
perhaps Stein’s selection of one song from a cycle was not an ideal way
to generate a demonstration analysis intended for undergraduate
students.
15 Schumann: Sonata in A Minor for Violin
and Piano (op. 105)
in response to peter h. smith

Peter H. Smith has been one of the most productive of all tonal analysts
during the past few decades. His numerous publications and lectures reveal a
special devotion to music by Schubert, Schumann, and Brahms. His route to
music theory was via study of the viola at Juilliard. This has resulted in a
special emphasis on music for stringed instruments in his analytical writings.
Smith explores Schumann’s opus 105 in three publications: an article from
2009, a book chapter from 2011, and another article from 2013.1 Because I
dart back and forth among these sources during my commentary, I indicate
the publication year before each page or example number cited.
Smith proceeds from technical analysis to a consideration of broader
issues, especially tonal pairing and the TMS complex (where TMS refers to
the tonic, mediant, and submediant). Because my technical analysis con-
trasts his in many ways – with disagreements not only on some of the finer
points but also regarding basic parameters such as form, tonicization, and
Kopfton – in many cases my perspective does not affirm the premises upon
which his broader assertions are made. My Schumann composes more
within the mainstream channels of early- to mid-nineteenth-century tonal
practice than does Smith’s Schumann. This chapter invites readers to take a
stand on a range of the work’s compositional details. Just how wide a gap
separates Schumann’s practice from the music he knew? Do we need fresh
analytical notions to deal with his practice? Or can his innovations be
accommodated within the analytical framework that has been developed
for his immediate predecessors?

Movement 1 exposition: P (measures 1–27)

A sonata exposition’s P is charged with introducing the tonic key, generally

by traversing the harmonic path from tonic root to dominant root and then
back to tonic root in the bass. That arpeggiation usually will support a
middleground melodic third- or fifth-progression descending by step from
the Kopfton (almost always ^3 or ^5 ) to the tonic pitch, resulting in a PAC.
183
184 Harmony in Mendelssohn and Schumann

The opus 105 first-movement exposition’s P is built from four inter-

connected, varied traversals of that path, creating points of articulation at
measures 5, 9, 16, and 27. (Locate these goals in 15.1, noting that each
corresponds to the endpoint of a C>B>A third-progression.) Only the final
iteration deploys both root-position dominant and root-position closing
tonic chords, as befits its comparative structural depth. Schumann prevents
a sense of full closure at the first three articulation points: a C covers the
melodic A of measure 5, and the preceding dominant is inverted; and the
melodic A at both measures 9 and 16 is supported by an inverted tonic.
(The final traversal’s majestic dominant expansion, whose truncation
within 15.1 is indicated by a hairpin symbol, will be addressed separately
below.)

Example 15.1 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 1, mm. 1–27.

Though the violin’s first pitch – a Middle C – fulfills the Kopfton role for the
time being (note the B and A that follow in that register over the course of
measures 1 through 5), the higher pitches that soon emerge arpeggiate the
tonic chord in an upward trajectory, ultimately resulting in a registral shift: E
(preceded by an appoggiatura F) in measure 1 begins the process, while the
piano’s A inaugurating the theme’s second iteration at 61 continues it. Though
a higher C is not attained prior to measure 6’s upper-neighbor D, clearly by
that point the octave above Middle C has taken over as the principal register
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 185

for the main melodic initiatives. (The Cs in measures 11–12 and 16 confirm
Kopfton C’s ensconcement in that register as P continues.)
The violin’s juxtaposition of the tonic triad’s C and E during measure 1
amounts to an inauguration of two distinct melodic strands, which des-
cend together during measures 1 through 5: C>B>A and E>D♯>D>C.2 As
is fitting for this early point in the work, the bass does not jump dynami-
cally among harmonic roots but instead shifts by half steps: A>G♯<A.
Consequently the progression’s supertonic and dominant harmonies are
inverted, muting the thrust of the surging II➔.3 (As we shall see, the
theme’s later deployment in the bass generates a more conspicuous projec-
tion of II➔ to a root-position dominant, during measures 17 and 18.)
The theme’s transposed redeployment, starting at IV during measure 6,
coordinates with a subtle reinterpretation of its initiating gesture. Whereas
C<F>E during measure 1 introduces the tonic triad’s third and fifth (with F
serving as an upper neighbor to E), A<D>C during measure 6 emphasizes
the subdominant’s fifth and octave (with C serving as a passing note
from D). The dominant that follows at first sports root E as its bass, but
seventh D emerges there (at 82) before the chord resolves. Consequently
the tonic of measure 9 sounds in its 63 position. Though on the one hand
this inverted tonic may seem disappointing as the goal of the local harmo-
nic progression, on the other hand it is an ideal starting point for the
particular trajectory Schumann has in mind for the next few measures. In
15.1 the chordal activity during measures 9 through 12 is conveyed as a
linear progression connecting the tonic’s 5- and 6-phase chords. As often is
the case when deploying an ascending 5–6 sequence, the full path between
those points is curtailed. The complete rendering would proceed as
A5––6 B5––6 C5––6 D5––6 E5––6 F5
A Minor: I5–– ( ) ––6

Given that the A5 and C6 chords are built from the same pitch classes, it is
possible to abbreviate the sequential ascent by starting on C6, which sounds
during measure 9.4 Schumann’s progression includes all the components
between C6 and F5 except for an elided D6 (which is conveyed parenthe-
tically in 15.1). Concurrently the lowering of B to B♭ during measure 11
enhances the focus on the goal F chord (so that E6 surges toward F5). In
that this broad I5–6 connection generally would proceed to II, it is impor-
tant to allow the pitches of measure 14 some time to fully fall into place.
With the arrival of B♮ within the melody’s descending arpeggiation, we
understand that the chord being projected indeed is II (not IV). Because
186 Harmony in Mendelssohn and Schumann

the seventh of the V♯ harmony that follows resides in the bass (D following
root E, as was the case also during measure 8), the resulting ♯ 42 chord again
leads to a tonic in 63 position in measure 16. P is not over yet!
In the next approach to the dominant, the D♯ of measures 17 and 18
dynamically targets dominant root E, contrasting diatonic bass D in
measures 6 and 14. During the theme’s initial presentation (measures 1
through 5), D♯ resides within a descending filled-in third (E>D♯>D>C♯).
Now, with that theme deployed in the bass, a AD ♯ diminished fifth resolves
to the dominant’s GE ♯ third (measure 19), fulfilling D♯’s ascending ten-
dency. Over the course of an eight-measure expansion, the dominant’s
dissonant seventh (D) emerges in the chordal interior and resolves to the
tonic chord’s C at 271, as shown in 15.1.
Before that D takes hold the dominant undergoes a dynamic expansion that
may be broken down analytically into three parts. In 15.2a a descent from B to
A in the melody sounds in a context that prevents the A from coming across
as the broad fifth-progression’s goal ^1. Instead it connects the preceding B and
an interior G♯ that emerges in conjunction with a reinstatement of the
dominant harmony during measure 26. (Thus the C>B>A third within each
of the earlier tonic prolongations is complemented by a B>A>G♯ third in the
context of the dominant.) The intervening F-A-C and D-F-A chords (at 251
and 261, respectively) embellish the dominant. Another phase of the domi-
nant’s prolongation takes an altogether different approach. Whereas 15.2b
shows the transfer of an inner-strand G♯ to the top of the texture (above the
B>A that appeared on the beam in 15.2a), 15.2c clarifies how that transfer is
carried out (the open noteheads) and how those dominant-chord pitches are
embellished by upper neighbor embellishments (the filled-in noteheads). A
third distinct component of the dominant prolongation is the local D-F-B♭
chord (shown within parentheses in 15.2c) that intervenes before all the
dominant’s pitches have fallen back into place. The D that sounds during
this stunning and unexpected chord is retained once the dominant is restored,
serving as its seventh (the D of measure 24 in 15.2a, not that of measure 26 in
15.1). Though this interpretation projects the D-F-B♭ chord of measure 23
and the D-F-A chord of measure 26 as embellishments of an already estab-
lished dominant, these conspicuous events do to some extent have the effect of
backtracking within the harmonic progression – of a fresh approach to the
dominant. Thus by the time this dominant resolves to the tonic (measure 27)
it has been paired with three different predecessors: II➔ (measures 17 and 18),
♭II (measure 23), and IV (measure 26).
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 187

Example 15.2 Schumann: Sonata in A Minor for Violin and Piano (op. 105), mvmt. 1
(a) analysis of mm. 19–26; (b) foundational voice leading of measures 19–25;
(c) detailed voice leading of measures 19–25.

In my presentation above, I segment P into four regions based on the

tonic arrival points at measures 5, 9, 16, and 27. My comments on
Smith’s analysis (as conveyed principally in his ex. 2009/4) will be
organized within those boundaries as well.
Measures 1–5: Smith regards the sforzando F of measure 1 as the
onset of a foundational neighbor-note motive, with a resolution to E at
42. I instead interpret it as a local neighbor of E within measure 1 (with
F>E matched immediately by E>D♯ on the following downbeat and later
by D>C at 51). Though that F does not appear in my graph, I would
grant that its introduction at the onset readies the sensitive ear for
188 Harmony in Mendelssohn and Schumann

similar deployments – for example, in measures 12 (bass) and 25

(violin) – later.5 These contrasting interpretations have resulted in
7
conflicting symbols in our graphs. Whereas I deploy ♯ 5 (which would
♯3
♯6
translate into ♯4 were I to acknowledge the chordal inversion) beside my
2

II numeral at measure 3, Smith instead proposes 4þ 2 (where his + here is

the equivalent of my ♯), implying diatonic 6. Though indeed Schumann

remains noncommittal for a while, during the second half of measure 3
an F♯ (♯6 above bass A) sounds thrice. I cannot endorse Smith’s reading
of F>E in light of such an obvious conflict.6
The first upper-staff notehead in Smith’s graph is a Middle C,
bearing a downward stem. What transpires in that register during
the chords that he labels as 4þ2 and as 6, neither of which accommo-

dates a C? Are we to hear an upward leap of an augmented second to

the downward-stemmed D♯? My graph shows a descent from C to B
(delayed until 32 in the score due to the suspension of C) in conjunc-
tion with II➔’s arrival. That B persists during V➔, with the emergence
of A coordinating with the resolution to I. (Both B and A sound at the
ends of measure-long descending violin trajectories, following their
less pronounced introduction in the piano part.) Though Smith dis-
plays an unstemmed B an octave higher at measure 4 and an
unstemmed A in the register of the initial C at the tonic resolution
(measure 5), he does not connect the dots. That A’s upward transfer
for the piano presentation of the opening theme (see the arrow in 15.1)
secures the register in which C>B>A will occur repeatedly during the
remainder of P. Though the initial C>B>A is set below other content, it
is important to acknowledge it here, since it will play a significant role
in the higher register as P continues.
Measures 5–9: I am in wholehearted agreement with Smith’s beamed
line C>B>A during this second iteration of the local I–V♯–I harmonic
progression (which is acknowledged via Roman numerals only in its
third and fourth traversals in his graph). Yet our treatments of the line
differ in two respects. First, I regard it as a reiteration of a C>B>A line
already introduced during measures 1 through 5, rather than as the
second half of a line emanating from E. Second, I follow Schumann’s
explicit projection of the line an octave higher than where Smith dis-
plays it. Regarding the latter, Smith’s note 2009/14 clarifies that he has
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 189

placed “the piano’s ‘covering’ line . . . down an octave” (p. 2009/82), an

intervention that I suggest should be rescinded. Given that the pitch B is
articulated four times in that upper register (above a persistent E)
during measures 7 and 8, I feel justified in interpreting the melody’s
E>C>A (above another E) during measure 9 as the introduction of goal
A via an arpeggiation from its upper fifth. (Compare with B<E>A at the
cadence of measures 26 and 27.) In Smith’s view that A instead is an
internal point within an E>C>A>E octave arpeggiation.
Because IV – rather than II➔ – now serves as the preparation for V♯
within the harmonic progression, F>E is no longer problematic as a
neighboring embellishment. The question thus becomes whether
E<F>E (projected by the violin) or C>B>A (projected by the piano) is
the defining melodic line. Over the broader course of P, my graph
displays focused C>B>A third-progressions consistently, whereas
Smith displays static neighbor-embellished E prolongations with equal
consistency. It seems to me that if one perceives a PAC (which I regard
as normative for P) transpiring at measure 27, then a descent to A is the
more compelling reading here.
Measures 9–16: Though Smith and I both acknowledge an F major
chord at measure 12, for me it serves as a conventional 6-phase expan-
sion of the tonic (A-C-E to A-C-F, here unfurled into F-A-C), whereas
for him it is the principal harmony between i6 and V. (Keep in mind that
Arabic 6 is being used in contrasting ways in his and my graphs.) I hear
bass D (measure 14) as the principal intermediary between the tonic’s C
(measure 9) and the dominant’s E (measure 15). Whereas a similar D
supported IV during measure 6, now it supports II7, whose seventh, A, is
delayed in the melody until the downbeat of measure 15. An unfolding
symbol connects bass D and F in both of our graphs. Yet whereas he
adds a flagged stem descending from the F (and refrains from labeling
the D chord as ii°), I have deployed such a stem from D (and incorporate
II7 within my harmonic analysis).7
Smith’s omission of a broad C>B>A melodic line stems in part from
his reading of the linear trajectory during measures 9 through 12. Note
that a B♭ sounds in the piano part before C emerges (measure 11). This
has led me to display a double application of the reaching-over techni-
que in my graph, so that from the pitch A at measure 9 first B♭ and then
C emerge, resulting in what is (for me) a reinstatement of the Kopfton.8
The tonic’s CA third (which may be interpreted as soprano and alto
190 Harmony in Mendelssohn and Schumann

strands) then proceeds to B

A during II7 (with the melody’s F>D>B
arpeggiation echoing the E>C>A of measure 9), BG ♯ during V♯, and
finally a unison A (surmounted by a fresh C for the next iteration of
the cycle) at I. Smith instead explicitly shows the B arriving from below
(via a slur connecting A and B in his graph), and he leaves the domi-
nant’s leading tone (upward-stemmed G♯) dangling, despite the piano’s
A at 161.
Measures 16–27: The three pitches of my C>B>A third-
progression are all present in Smith’s graph: C and B both sport
upward stems (measures 16 and 17), though B’s stem is reversed to
downward at the onset of V in measure 19; and A is present above
tonic root A in measure 27, though presented unstemmed as an
appendage of the E above. Smith again does not connect the dots,
proposing “a precipitous plunge to ^1, rather than a generic stepwise
descent to closure” (p. 2009/54, emphasis added). In that the visual
projection of such a third-progression certainly would trigger ques-
tions regarding whether perhaps ^3 (instead of ^5 ) serves as the
Kopfton, its omission in Smith’s presentation makes for a more
consistent (though in my view faulty) interpretation. Yet if F>E is
again to play the principal role, then the problem noted above
recurs: his 65 annotation at measure 17 corresponds to a spot that
6
in the score instead would be figured as ♯53 . My sensibility again rebels
against the proposal to prolong a neighbor F against a sounding F♯. I
display this B-D♯-F♯-A chord as one of the four principal harmonic
events of P (as presented in the second row of Roman numerals in 15.1).
♭
Not so for Smith, where the II and VI that precede reiterations of this
dominant are not complemented by some sort of harmonic label
between the long-prolonged initial tonic and the onset of the dominant.
I likewise am curious why VI rather than IV is displayed during mea-
sures 25 and 26. Though in my reading F-A-C appears in the same
places as in Smith’s (measures 12 and 25), I do not deploy a VI numeral
at either location.
My proposed B>A>G♯ third-progression over the course of the
dominant prolongation (as shown in 15.2a) likewise has no counterpart
in Smith’s reading. A downwards-stemmed B is displayed in the alto
register in his graph at measure 19 and an upward-stemmed A with flag
appears an octave higher at measure 25. But the G♯ that completes the
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 191

third-progression during measure 26 is missing, despite the sharp in the

figured bass indicating its presence in the chord.

Movement 1 exposition: TR ⇨ FS (measures 27–64)

Once A Minor’s root-position tonic chord is secured cadentially at 271,

thereby concluding P, Schumann shifts from a tonic-establishing to a
tonic-departing agenda. In an exposition that juxtaposes A Minor (the
key of the movement as a whole) and a C Major tonicization (the venue for
the upcoming Fortspinnung – “spinning-out” – that concludes the exposi-
tion) in the context of Kopfton ^3, the latter key often will be prolonged by
means of a fifth-progression descending from an inner-strand G trans-
ferred to the top of the texture, as is conveyed in the foundational model of
15.3. Schumann deploys a segment of the descending circle of fifths to link
those two more stable regions. The A chord of P’s cadence is expanded by
means of an A–E–A bass arpeggiation (during measures 27 through 34)
that supports two instances of reaching-over in the melody, so that E and
then G sound at the top of the texture, as shown in 15.4. The latter pitch
emerges against the raising of the A tonic’s third to C♯, resulting in a surge:
A➔ targeting D. Though D’s seventh, C, is introduced during measure 35,
that chord’s minor quality is retained during the succession to a G-rooted
chord, which surges from its onset. The expected resolution (following the
7–10 precedent of A➔ D, as marked in the graph) would be G➔ to a root-
position C chord for the onset of the Fortspinnung coinciding with the
mediant tonicization. The first sign that what lies ahead will involve many
disparities between expectation and compositional realization emerges
with that initial C chord: though we expect F>E in the soprano (below a
G) and G>C in the bass, that resolution’s E and C trade places in
Schumann’s rendering. (Compare the score and 15.4, which projects the
normative resolution pitches within parentheses.) Repetitions of the cir-
cular progression’s chords (now with a surging D➔) continue through
measure 43, finally attaining a C-rooted chord with the expected soprano E
and bass C at 432. Yet while resolving one issue, Schumann introduces
another: a B♭ sounds along with C, E, and G in this tonic chord.
Consequently there is no point in the vicinity of the C Major tonicization’s
onset where a root-position, consonant C major chord sounds.
192 Harmony in Mendelssohn and Schumann

Example 15.3 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 1, foundational model for how an exposition in A Minor might be structured.

Example 15.4 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 1, mm. 27–43.

Given the circular progression’s reliance upon surges during its reitera-
tions (with D➔ followed by G➔ leading toward C), adding B♭ to C-E-G at
432 might come across initially as a C➔ surge targeting F, putting into
question C’s role as the transition’s goal. Schumann here deftly navigates a
shift of context: C’s seventh B♭ ultimately functions instead as 6-phase A’s
ninth (with A➔ represented by C♯-E-G-B♭). We may excuse C♯’s tardiness
(coinciding with the arrival of bass F at 441 rather than sounding before the
bar line, as occurs during 452), since it facilitates Schumann’s ruse.
Colliding with the arrival of II’s bass F, this surging VI➔ definitively
terminates the transition’s circular progression and ultimately helps con-
firm C Major as the prevailing key for the remainder of the exposition, as
shown in 15.5.
Example 15.5 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105), mvmt. 1, mm. 38–63.
194 Harmony in Mendelssohn and Schumann

Schumann’s writing during the exposition’s Fortspinnung (FS) is

extraordinary because it seems intent upon preventing a full realization
of the expected outcome, embracing numerous partial or contorted
iterations of a compositional initiative that in a more routine deploy-
ment would offer a straightforward means of tonicization. This initia-
tive’s principal component is displayed unadorned in 15.3: a stepwise
descent from G down to C. The harmonization of that line might be
expected to incorporate C Major’s II, IV, or V7 as support for F; and
certainly the harmonization of the concluding D>C would be V(7)–I, both
in root position. Consider how Schumann’s fertile imagination and secure
mastery of tonality conspire to thwart this scenario repeatedly, as conveyed
below. Once all these iterations have transpired, the listener confronts the
discomforting fact that the exposition has run its course without projecting
an EEC.
Measures 42–47: By taking into account the transition’s C chord during 421
(which itself hearkens back to that of 381), one may perceive how a high G sets
the stage for a stepwise descent, so that the violin’s E>D during measure 44
comes across as part of an interior strand. What has happened to the G?
Schumann goes so far as to prevent the piano G an octave lower from
descending to F during measure 44, though this prohibition is lifted in the
reiteration two measures later. The graph offered in 15.5 shows an E>D
second appearing below a broad set of open parentheses that convey how
the G starting point for a potential descending fifth-progression has been
neglected.
Measures 47–49: Taking a somewhat broader view, one may regard the
E>D second discussed above as an event of the chordal interior, preceding a
soprano descent, with the G reinstated at 472 serving as the starting point
for a fifth-progression occurring in the piano. This rendering is inconclu-
sive because the goal tonic chord is presented in 64 position and because a G
is retained at the top of the texture.
Measures 49–51: The preceding descent is here reiterated, with a differ-
ent irregularity at its close: a tonic 6-phase chord (C-E-A, unfurled), rather
than 5-phase C-E-G in second inversion.
Measures 51–55: Instead of starting again on the tonic, the A minor
chord (the tonic’s 6-phase chord) reached at 511 proceeds to II. Schumann
accomplishes this through two descending-third successions in both outer
voices, resulting in the introduction of the fifth-progression’s F an octave
lower than expected. As 15.5 shows, that F is transferred up an octave in
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 195

coordination with the arrival of V7, after which the descent to C proceeds
without a hitch. Once again, though, the goal C is supported not by the
C-E-G tonic harmony, but by its 6-phase surrogate.
Measures 55–57: At 552 Schumann backtracks all the way to the C-E-G-
B♭ chord introduced at 432. Whereas there and at 452 the melody focused
on the interior strand E>D, now G>F is projected, though the thread is lost
in measure 57. Thus we have experienced yet another failure in the quest
for closure.
Measures 57–59: Schumann’s backtracking at this point is indicated in
15.5 by the two asterisks placed between the graph’s staves. (We now
proceed from the chord marked by the second asterisk, with a surging
A➔ chord sporting root A in the bass.) Both the harmonic progression
(I(5)–6 II V I) and the melodic descent (G>F>E>D>C) are exactly what we
would want in order to achieve the long-awaited closure. Yet success
remains elusive once more, since the concluding tonic chord sounds in
its 63 position.9
Measures 59–63: As a last-ditch effort to get all the factors to work
together in achieving a PAC, Schumann revises the harmonic progression,
substituting I IV5–6 for I5–6 II. Alas, the dominant that follows after IV once
again misses its tonic goal, with the tonic’s 6-phase chord sounding yet
again at 611 and at 631. Though Schumann of course could have con-
structed the desired cadence with no problem, he now bows out – as if done
in by his unruly progressions – leaving behind a panorama of failures. The
A minor chord lingers precariously for two measures and then is fully
embraced for the exposition’s repeat and, the second time, for moving
onward to the development.

Though its analytical notation contrasts mine, Smith’s Example 2009/9

conveys a conception of the transition that is very similar to that on display
in 15.4. By consulting the score, one might become convinced that the
transfer of the piano melody from 272 through 291 down a fourth for the
violin line from 292 through 311 would warrant placing the onset of
Smith’s local V at 292 rather than at 311. (That is, the C-E-A chord at
301 serves as the subdominant within a tonicization of dominant E, rather
than as a reiteration of the preceding A tonic.) We both are coming to
terms with the broad progression of four roots, A–D–G–C, which I
interpret as an indivisible circular progression connecting A Minor’s I
196 Harmony in Mendelssohn and Schumann

and III Stufen (with III tonicized over the course of FS).10 Smith instead
interprets A–D in one key and D–G–C in the other, with D serving as a
pivot chord. In his ex. 2009/6, the completion of the motion to C is
qualified by the bold placement of the symbol ♮VII (in A Minor) under
the G chord, which is then prolonged for most of the section; whereas in
ex. 2009/10B (= ex. 2011/9.9), the C chord does not even appear after the G
chord.
Our disagreement regarding the C chord at 381 is of great conse-
quence. Though Smith grants the chord a I label in ex. 2009/9, this I is
conveyed as subordinate to a prolonged V, a view that is carried over
into his FS graph (ex. 2009/6), where the G chord’s relative prominence
is conveyed through the impressive length of its stem (over twice as long
as that of the C tonic’s bass E). I instead hear G➔ fully resolving to C –
despite the C chord’s inversion at 381 and 421, and despite the added B♭
at 432. The A-C-E chords that transpire during FS all relate convincingly
to a C tonic, just as the F-A-C chords during P relate at some level to an
A tonic. My conviction wavers only at measure 63.
Given my contrasting perspective, I would recommend a few altera-
tions in the details of Smith’s ex. 2009/6. Most urgently, I would want to
see a wholehearted embracing of C as a tonic Stufe in C Major at 432,
452, and 552.11 I also would propose C as a melodic goal (descending
from D) at 491, at 511, and at 551. Roman numeral IV should be removed
as a harmonic label for measure 51 (in that that chord is internal to a
descending-thirds trajectory between A and D) but added for 601 and
621 (where a G held over from the preceding tonic harmony prevents
chord member F, and with it A in the bass, from sounding on the
downbeat). Finally, I would refrain from using Schenker’s interruption
symbol at measure 56, instead interpreting the backtracking and reitera-
tion as merely a local rhetorical flourish.
Whereas Smith rejects his ex. 2009/10A, which (except for its projec-
tion of E as the Kopfton) corresponds closely to my 15.3, and instead
endorses his ex. 2009/10B, which proposes that G rather than C is the
principal bass pitch between the As at the endpoints of the exposition, I
propose that Schumann in fact pursues the “conventional tonal struc-
ture” of 2009/10A. For me, the local quirks in his writing can be (and
ought to be) fully accommodated within that framework, as my analysis
demonstrates.
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 197

Movement 1 development (measures 65–ca. 115)

The exposition’s failure to attain a PAC in C Major has seriously compro-

mised the viability of the common minor-key tonal plan of proceeding from I
to III during the exposition, followed by a continuing upward course to V♯
during the development. The A minor chord of measure 63 not only reintro-
duces the tonic for the repeat of the exposition, but also serves as the departure
point for the chordal trajectory that will occur within the development. We are
back on square one, with no inkling of what will happen next.
Because so many chords appear between the end of the exposition and
the first hints of P’s return for the recapitulation, it is challenging for
listeners (and analysts) to get a clear sense of what motivates Schumann’s
writing during this movement’s development section. My approach, whose
outcome is presented in 15.6, has been to apply hierarchical thinking
sensitively yet adamantly, so that a more manageable number of chords
may be assayed. Before entering into the details of how these foundational
moments emerge out of their rich chordal context, one might attempt to
discern what factors could have motivated their incorporation within the
progression. If logical relationships – in keeping with compositional pro-
cedures of the time – are noted, a confidence that one has hit upon a viable
means of coming to terms with this development section may emerge.
Analytical endeavors then might focus on assaying how well this repre-
sentation corresponds to what Schumann actually wrote.

Example 15.6 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 1, mm. 65–119.
198 Harmony in Mendelssohn and Schumann

The trajectory displayed in 15.6 may be interpreted as a broad and

roundabout traversal of a descending circle of fifths, incorporating all of A
Minor’s seven diatonic pitch classes as roots, arranged in the following order:

A D G C F B E A
This scenario may have developed in Schumann’s mind as an outgrowth of the
thematic content from measures 1 through 5, where a revised interpretation of
the pitch F in measure 1 (corresponding to what was discussed above regarding
the pitch D in measure 6) would generate the F B E A segment of this circle.
Each pair of adjacent chords that may be enhanced by a surge in fact sports such
an evolution at least once during the development section (thus A➔ D, D➔ G,
and so on, with only F and B not so connected). The circle proceeds through its
fourth chord during measures 65 through 75; a backtracking to the second of
the circle’s chords ensues through measure 81; a temporary period of stalling
involving a hovering around that second chord (incorporating its upper- and
lower-third chords) persists through measure 99; and finally the circular
progression begins to move forward again, reaching the tonic in measure
119. Our exploration of how the model of 15.6 corresponds to the composi-
tion’s surface details will consider each of these four regions in turn.
Measures 65–75: Just as the work proceeds from the exposition’s closing
measures by looping back to P in the context of the repeat sign, the initial
measures of P likewise prevail as one moves beyond that sign into measure
68. At a local level, a harmonic progression like the one that opens the
exposition occurs:
m. 65 66 68 69
A Minor: I II➔ V➔ I

From a somewhat broader perspective the circular progression that will

guide the entire development section is inaugurated by the surge that
emerges between the tonic harmonies at that progression’s endpoints:
m. 65 69
A ( ) ➔

The circle’s next component – a surging D chord – sounds for only one
measure (71), a terseness that will be compensated for later in the development.
Then G arrives. This G chord’s B♭ is a wobbly lowering of diatonic B♮, a factor
that Schumann offsets with some effort. Reprising an initiative also deployed
during measures 9 through 12, a local sequential progression now connects the
broader circular progression’s two versions of the G chord, as follows:
m. 72 73 74
♭6 ♭5–——–6
G5———– A B♮5–——–6
G ( )➔
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 199

Whereas normally an ascending 5–6 sequence’s sixth chord is built from the
same pitch classes as its first chord, here that relationship is tempered by a B♭-
to-B♮ inflection that allows the G chord to target the C chord of measure 75 as a
surge. In that this C chord likewise is introduced with minor quality, an
eventual shift to E♮ might be expected. In this case, however, Schumann instead
rescinds most of the progress made thus far, as we shall see below.
Measures 75–81: Given the setbacks encountered during the exposition
attempts to arrive at a PAC in C Major, it hardly seems appropriate that a
circle of fifths would proceed through all of its stations without resistance.
Though Schumann might have relished the superlative drive that a circular
progression offers, he likewise could enjoy setting to work on it and making
it yield to his wishes. Though the C minor chord of 751 is transformed into
a surge (at bass B♭ during 761, where E-G-B♭-D♭ is an evolved state of C➔),
the chords that transpire over the next few measures are of local impact,
not components of the broad circular progression we have been mapping
out. Schumann is drawing upon a property particular to circular progres-
sions. Since a closed circle contains seven distinct descending fifths (one of
them diminished), one may proceed forward six fifths to backtrack by one
fifth. In this case Schumann incorporates some of the flats characteristic of
C Minor and G Minor, as follows:

C ➔F B♭ ➔ E♭ A D➔ G
Once G is reached in this manner, another backtracking by a fifth occurs
during measures 78 through 81, as:

G ➔C F ➔ B♭ E A➔ D
Though at first it may seem that a third backtracking (which would lead to
the initiating A chord) gets underway soon after this D chord arrives, it
turns out that that trajectory goes elsewhere, as will be explored below.
Thus the backtracking may be regarded as having ceased with the arrival of
the D chord at 811.
Measures 81–99: In the midst of a circular progression in which every
internal component is related to adjacent chords by both ascending and
descending fifths, Schumann pauses to replicate such relationships on a
smaller scale, with ascending and descending thirds. Whereas the earlier
local deployments of circles of fifths that began after measures 75 and 78
led to a chord rooted a fifth higher, such a circle leads instead up a third in
its deployment after measure 81, as follows:
D ➔G C➔ F
Then borrowing the local harmonic progression that landed on its sixth
scale degree near the end of the exposition (starting at 592), descents by a
200 Harmony in Mendelssohn and Schumann

third occur twice: from F to D, and then from D to B♭. Consequently both
upper- and lower-third chords reinforce the D chord attained at measure
81. From lower-third B♭ the restoration of D is accomplished by again
deploying a segment of an ascending 5–6 sequence (with the bass now
chromatically enlivened), as follows:

m. 88 91 92 95 96
B♭5 B♮ 6
C 5
C♯ 6
D5
Measures 99–119: The D minor chord of measure 96 returns with surge
characteristics during measure 99, the endpoint of a local tonicizing I II➔
V I progression with minor dominant.12 The G minor chord of measure
100 is prolonged in a similar manner, with its own tonicizing I II➔ V➔ I
progression (with major dominant) concluding with a surging G➔. Given
this exemplary commitment to moving forward along the circular path
that was inaugurated at the start of the development section, the momen-
tum continues with C in measure 104, which surges at its onset.13 In that
Schumann will inaugurate a broadly paced version of the P theme above
this C root in measure 110 to launch the recapitulation, it is fitting that the
C chord is treated to an expansion, as displayed in 15.7. That example’s
open noteheads convey the foundational succession from C-E-G-B♭ to
F-A-C. Observe how Schumann heightens the passage’s intensity by per-
sistently preventing a recurrence of a pure C-E-G-B♭ sonority: at least one
ascending diatonic or chromatic passing note (the graph’s filled-in note-
heads) infiltrates every sonority between the C➔ of measure 104 and the F
of measure 111. The F chord is intensified through its presentation in 53
position during measure 115, thereby confirming that the melody’s F
introduced at measure 111 is transpiring in a chordal context that contrasts
its deployment during measure 1, where the introduction of the A-C-E
tonic chord was of paramount importance.

Example 15.7 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 1, mm. 104–111.
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 201

Among these internal chords, that of measure 110 is the most

enigmatic. Schumann bridges the divide between the C and F chords
with a C-E-A sonority that might be interpreted as the C-E third of the
prolonged C➔ with an anticipation A (the third of goal F-A-C), or
instead as the A-C third of the upcoming F-A-C with an E suspended
from C-E-G-B♭. (These alternative perspectives are juxtaposed in 15.7,
where open oval noteheads correspond to C➔ and open diamond
noteheads correspond to F.) Yet that is not all! Given the thematic
content initiated by the violin in measure 110, there is also a resonance
with the A-C-E tonic chord of measure 1. My reading proposes that an
A tonic is not asserted at this point, but instead that the context
(dominated by the extended C➔ surge) molds the theme to its own
agenda, which the development’s overriding circular trajectory suggests
would inevitably proceed as C➔ F. The circular progression then
concludes with the B, E, and A sonorities that were deployed in the
theme’s harmonization during measures 2 through 5. Consequently
there is no tidy dividing line between the development and the reca-
pitulation. The circle-of-fifths initiative concludes on the recapitulation
equivalent of measure 5’s tonic during measure 119. There is no tonic
equivalent of measure 1.

Smith’s presentation passes over a large chunk of the development

section. After the A Minor tonic harmony at the end of the exposition
(measure 65, displayed at the right edge of his ex. 2009/6 with an open-
notehead A in the bass), we regain contact with the progression at the C-
E-G-B♭ chord of measure 104 (labeled as III at the left edge of his
Example 2009/11). Despite the fact that III possesses a minor seventh
(acknowledged by a ♭7 symbol annotating his graph), his middleground
harmonic analysis shows the chord proceeding (via a 5–6 shift) to II♯.
This all seems dubious to me.
Whereas Smith would reject my reading of the exposition (prior to
the reinstatement of tonic A during its final measures) as I ( ) III, I reject
his reading of the development as I ( ) III. The fifth-relationships
projected in my 15.6 correspond to a procedure that Schumann would
have known and experimented with at the keyboard from an early age.
In that context the C➔ chord has a fifth-related predecessor (G➔) and a
fifth-related successor (F). Schumann here integrates the P theme’s II
V➔ I within the circle, modifying the chordal trajectory to
202 Harmony in Mendelssohn and Schumann

accommodate root F (thus making the neighboring note F of measure 1

a stable chord member), so that a seamless continuation of the circular
progression may be achieved as P’s thematic material emerges. There is
no rationale for picking and choosing among the fifth-related chords
(resulting in Smith’s III-to-II♯ succession, reinforced by a C–B slur in
the bass of his ex. 2009/11), nor is an E [Minor] tonicization (between
measures 110 and 118) warranted. Measure 119 should be displayed as
the definitive return of the A Minor tonic (corresponding to measure 5),
and not as a subordinate i (subservient to a prolonged V), as Smith
conveys in his ex. 2009/11.14
My 15.7 organizes the pitch content between measures 104 and 111
into three categories: members of the circle’s surging C-E-G-B♭ chord,
members of C➔’s F-A-C successor, and connective pitches. At various
moments along that trajectory, chords of no structural significance are
formed as members of the C➔ chord and one or more connective
pitches coincide. Smith interprets some of these incidental chords as
“motivic sonorities” (p. 2009/63), labeled using the letters F and A
(which he proposes bring to mind the other players of his TMS
complex), including one A chord that he interprets as iv in E
[Minor] (ex. 2009/11).15 In my view, he has at this point lost sight of
the essence.

Movement 1 recapitulation and coda (measures ca. 115–209)

Because most of what is presented during a typical recapitulation will

have been explored analytically in the context of the exposition,
recapitulations generally receive limited attention in analyses. Here
we focus on three specific questions: one pertinent to all recapitula-
tions, one to all recapitulations within minor-key sonata movements,
and one to this particular recapitulation. A graph of the entire move-
ment, presented in 15.8, should be consulted as required. (This graph
accommodates the unusual development section – which essentially
prolongs the tonic – by displaying it as a parenthetical passage
between the tonic harmonies at the end of the exposition and begin-
ning of the recapitulation. Kopfton ^3 is maintained through to the
recapitulation’s TR.)
Example 15.8 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105), mvmt. 1.
204 Harmony in Mendelssohn and Schumann

How does the recapitulation’s TR differ from the exposition’s, in order

to accommodate the fact that what follows will take place in A Major rather
than in C Major? The exposition TR’s A ➔ D G➔ C circular progression
(15.4) is ignited by a surging A➔ chord during measure 34. At the
equivalent spot during the recapitulation TR (measure 148) what begins
to take shape as another A➔ surge is transformed into an F♯➔ surge. As a
result, the former continuation is transposed down a minor third, to F♯➔ B
E➔ A.16 The E chord here is asserted as a structural dominant between the
A Minor and A Major portions of the recapitulation. The former supports
the background ^3 prolonged since the movement’s onset; the latter sup-
ports middleground fifth-progressions whose goal is background ^1.
Consequently the ^2 of TR’s dominant harmony is displayed in 15.8 as a
component of the Ursatz.
Given that the exposition’s juxtaposition of A Minor and C Major
regions is converted into a juxtaposition of A Minor and A Major
regions during the recapitulation, does the recapitulation fully embrace
the new key? Or is A Minor still a force to be reckoned with even after
A Major has been established? Because the material presented during
the recapitulation’s A Major FS is a transposition of the exposition’s C
Major material, A’s major modal characteristics are firmly supported –
until measure 177. From that point onward, the piece seems torn
between these contrasting pitch collections. Though the ESC sounds
as an A major chord (measure 195, replicated at least once every two
measures through measure 205), the F♮ and C♮ of A Minor infiltrate the
local chordal progressions as the coda unfolds. Schumann’s last word
on the subject is devastating: the tonic shifts to minor quality in
measure 206, the supertonic that follows is of diminished quality
(indicative of A Minor), and the final tonic is a chord of minor quality.
The shift to A Major has been decisively revoked.
The A minor chord at the end of the C Major tonicization during the
exposition is easily accommodated, since the movement is in the key of A
Minor. How is this unusual detail handled in the context of the recapitula-
tion’s A Major region? The F♯-A-C♯ chord of measure 175 indeed is
worrisome. How will Schumann get beyond the impasse? The unexpected
shift to F-A-C at measures 177 through 179 calls listeners to attention:
certainly something unusual is brewing. This chord hearkens back to the
F-A-C chord of measure 25. Though they occur in different parts of their
respective sections (near the end of P versus near the end of FS), these
chords share the characteristic of preceding a strong cadence on an A-
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 205

rooted tonic: minor during the exposition, major during the recapitulation.
(This discussion is predicated upon the fact that the recapitulation does
eventually achieve a cadence, the ESC, whereas the exposition lacks an
EEC.) It was noted earlier how the exposition P (and of course the
recapitulation P as well) juxtaposes three distinct predecessors of the
dominant leading up to P’s cadence on tonic A: II➔, ♭II, and IV. By adding
D♯ to F-A-C during measure 180, Schumann’s already impressive collec-
tion grows to include II⇨ (in the formulation referred to in conventional
harmonic analysis as a German augmented sixth). The distinctive
augmented-sixth interval DF ♯ alternates with its FD ♯ inversion over the
course of the measures preceding the dominant restoration in measure
189, followed by the tonic resolution (the ESC) at measure 195.

In our considerations of the recapitulation and coda, Smith and I both

focus on the reintroduction of A Minor beginning with the F-A-C chord
of measure 177 in a context that has projected A Major since measure
157. The main difference in our perspectives is that I regard this F to
reside within the recapitulation (whose close resembles the exposition
and recapitulation P’s cadence because the exposition FS does not offer a
suitable cadential model), whereas for Smith the F of measure 177
instead inaugurates the coda (p. 2009/63).

Movement 2 refrain and transition (measures 0|1–15)

Local upper neighbor F in an accented metrical position embellishes the

fifth of the A-C-E tonic triad that opens the sonata’s first movement. Local
lower neighbors E and G in an accented metrical position embellish the
root and third of the F-A-C tonic triad that opens the second movement,
which is in rondo form. (See 15.9.) In both movements the violin melody
proceeds swiftly from Kopfton ^3 to its upper third (C<F>E and A<C).
Likewise in both the ^2 that follows is harmonized first by II➔ and then
by V(♯).17 In the first movement, this applies both to measures 3 and 4 and
to the broader trajectory leading to P’s principal dominant (measures 17
through 19). Whereas the dominant proceeds directly back to the tonic
during the first movement’s measures 4 and 5, the dominant of measure 19
prevails until measure 27. Like this latter prolongation, the dominant
attained in measure 2 of the second movement prevails until measure 12.
206 Harmony in Mendelssohn and Schumann

Example 15.9 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 2, mm. 0|1–12.

Already the duration of the second movement’s opening I–II➔ succession

is lopsidedly short compared with that of the prolonged V that follows. (The
first movement P’s components are more evenly balanced.) A reading of the
E-C-G-C chord at 11 as the onset of V (rather than as an embellishment of I)
would contort that relationship even further, to a degree that I doubt
Schumann would have sanctioned. (C Major would be so ensconced that
listeners would be thrown into confusion with the resolution to F at measure
12.) He in fact offers guidance against that interpretation through a
later reworking of the opening progression, in measure 57, where a diatonic
II (G-B♭-D) surges (as G-B♮-D-F) preceding the onset of V. That turn of
events would suggest a similar conception – in which a surging supertonic
harmony likewise precedes the dominant – during measures 1 and 2.
The first movement P’s dominant expansion features two chords – in
measures 23 and 26 – that have the impact of backtrackings, allowing the
dominant to be attained a second and third time (following the initial II➔
V♯) via the alternative paths ♭II V♯ and IV V♯. Schumann offers a similar
reinforcement of the dominant during the second movement’s refrain,
replacing measure 2’s G-B♮-D-F predecessor of V with B♮-D-F-A♭ and
B♮-D-F-A before restoring G-B♮-D-F prior to the long-awaited V7–I
cadence. In the foundational model shown in 15.10a, a C dominant
harmony is expanded through two concurrent melodic initiatives above a
stable bass C: the pitch F serves as a neighbor to the dominant’s E, and B♭
and A pass between its C and G. Although dissonant EB ♭ is followed by FA ,
the broader context prevents the internal C-F-A chord from coming across
as a resolution of V7 to I.
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 207

Example 15.10 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 2, (a) foundational model; (b) mm. 2–3; (c) mm. 3–5; (d) mm. 5–7; (e) mm. 7–11.
208 Harmony in Mendelssohn and Schumann

A composer with Schumann’s creative gifts would regard such a struc-

ture as an enticing opportunity to inject a subordinate tonic assertion into
the midst of a broad dominant prolongation. Thus the I II➔ V progression
that served so well in attaining the dominant during measures 0|1 through
2 may be redeployed as a means of prolonging that dominant prior to its
eventual resolution. In this refrain that progression is repeated almost
obsessively in a range of varied elaborations, with the initial dominant
arrival at 21 reinforced at 31, 41, 51, 71, 92, and 112 prior to the tonic
resolution at 121. Schumann here puts the forward trajectory from V to I
on hold while exploring a range of alternatives to the straightforward I II➔
V of his original conception. The ritardando and fermata in measures 2
and 3 signal the onset of a passage devoted to reverie. They alert listeners
not to expect the dominant to resolve punctually. What transpires between
the dominant of measures 3 and its restoration at the end of measure 11
may be interpreted aptly as a parenthetical passage.
The onset of Schumann’s elaboration hardly extends beyond the bare
bones of the foundational neighboring/passing model proposed in 15.10a.
The A>G second of the descending fourth is filled in chromatically at the
end of measure 2, and a D serves as a neighbor to C (with a concurrent shift
up an octave) in 15.10b. (In the models, the foundational pitches of 15.10a
are displayed using open noteheads, while the various accretions appear as
filled-in noteheads.) Though one might regard this new sonority as mere
connective voice leading, what happens later encourages one to interpret it
as an incipient formulation of a surging II➔ – that is, as (B♮)-D-F-A♭ above
dominant pedal C. Though under normal circumstances positing the
assertion of a G➔ chord when neither a G nor a B♮ sounds would be
regarded as excessively imaginative, in this context it might be justified, in
that the imagined B♮ emerges in sound at the equivalent location just one
measure later (as the alignment of 15.10b and 15.10c helps clarify). Two
other novelties enhance model 10c: the cessation of the C pedal point, and
the introduction of the pitch D, projecting the tonic’s 6-phase chord,
between I and II➔. A different mix of chords emerges in 15.10d.
Whereas the C pedal point is reinstated and the tonic 6-phase chord does
not intervene, B♮ sounds against pedal C for the II➔ chord, in which A
(rather than A♭) now serves as the ninth.
The next variant (shown in 15.10e) begins in measure 7. Reprising the
ritardando and fermata to remind us that, yes, the broader progression is
still on hold, this variant is much more elaborate than any of its predeces-
sors. Not only is the tonic’s 5–6 shift restored, but also it is doted on by the
rhetorically emphasized A➔ embellishing chord that targets the 6-phase D
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 209

chord.18 Both the 6-phase chord and the supertonic that follows are
presented in two stages: first diatonic and then surging. Though this
surge-rich region is accommodated within the harmonic progression (as
conveyed by the Roman numerals annotating the model), listeners also
might note a correlation between this trajectory and the circular progres-
sion of the first movement’s development (15.6, measures 65 through 104).
The expected cadence at 101 is delayed by a second pass through this
variant’s trajectory, now with the diatonic onsets of VI and II suppressed
and with II➔ presented in root position at 112. The PAC finally is realized
at 121.19 All that has transpired constitutes one phrase with multiple
internal backtrackings (offering opportunities for creative modifications
that Schumann has exploited admirably) rather than multiple phrases.
What may seem at first to initiate a repeat of the refrain’s contents
(starting at the upbeat to measure 13) ends up serving as a transition.
The dominant reached at 141 resolves without much ado to an F Minor
tonic chord to initiate the movement’s first episode, to be discussed below.

Both the violin melody and the piano bass descend a second over the
movement’s first bar line: A>G above F>E. Schumann’s slurring calls
attention to the 10–10 relationship. These lines immediately snap back:
a third 10 ( AF ) completes the gesture, with leading tone E resolving to F
in the bass. Whereas the violin’s G is a local neighbor between two As,
the piano melody’s A>G in the middle of measure 1 is a more substantial
structural element. G’s arrival coincides with the onset of II➔, and
though G is transferred to the tenor register at 31 it eventually is restored
to prominence (most notably just prior to the arrival of the third-
progression’s goal F at 121). The dominant’s extended prolongation
(which I describe as reverie above) makes the normalcy of this basic
structure (as projected in 15.9) somewhat challenging to appreciate. As
an orientation, one might perform from the beginning of the movement
through the fermata in measure 3, and then pick up with the dominant
at the end of measure 11 (noting how the violin’s G<A<B♭>A>G there
complements the initial A<C>A), so that the PAC at 121 arrives norma-
tively on the phrase’s fourth downbeat.20
Smith’s Example 2009/12 displays the opening F major chord only as
IV in C Major. (Earlier, in ex. 2009/5 – reprinted as ex. 2011/9.3 – the
option of a tonic interpretation was acknowledged for the F chord at the
onset, but not for that during 11.) The possibility of the most normative
210 Harmony in Mendelssohn and Schumann

of all structural openings – tonic harmony with Kopfton ^3 – is thereby

removed from consideration. (Smith describes the refrain, which he
concludes at measure 8, as “tonally unstable and harmonically open”
(p. 2009/64). My 15.9, which conveys the refrain as a twelve-measure
entity, is tonally stable and harmonically closed.) I suggest that readers
endeavor to hear the opening three chords as a projection of I support-
ing ^3, pondering how performers might convey that interpretation
effectively, as well as how the bass G that follows later in measure 1
might be projected as the fulcrum between roots F and C.
Though ending ex. 2009/12 with the III♯ chord of measure 8 might
seem justified by Schumann’s placing a fermata on that chord in his
score, I suggest that this chord actually is in the middle of things – that
the A<D in the bass of both measures 3 and 4 is about to happen again,
with D sounding at 91, just after the graph terminates. Smith’s display of
the III♯ numeral in the same row as the preceding V numerals (with a I
in parentheses, corresponding to 72, placed in a subordinate position a
bit higher on the page) gives the misleading impression that V and III♯
are directly related. (He goes so far as to interpret the A chord as “the
cadence of the refrain’s consequent phrase” (p. 2009/64).) My 15.10e
proposes instead that this A-C♯-E chord embellishes a conventional 5–6
shift of the tonic harmony. (That is, C♯ serves as a chromatic passing
note between structurally deeper pitches – diatonic C and D, the 5 and 6
of the F tonic chord’s 5–6 shift.) The following accounting of hierarchies
(fully fleshed out in 15.10) is totally at odds with Smith’s conception:

F C (15.10a)
F G C (15.10b and d)
F D G C (15.10c)
A➔
F D G C (15.10e)
F Major: I5–––––––––––––––––––––––––6 II V

Note how Schumann progressively builds up the structure. The late

incorporation of A➔ (beginning in measure 8) is indicative of its fore-
ground role. To me it seems whimsically presented, with the fermata
placing the indivisible C<C♯<D chromatic motion (hopping between
registers and instruments) within an uncommon framework.21 There is
no cadence (in a harmonic sense) in the vicinity. Instead the will to
move forward from the surging A➔ chord is temporarily held in check,
making the arrival of the D chord all the more satisfying when it finally
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 211

does occur.22 I hear the D, G, and C chords of measures 9 and 11

(15.10e) as creative variants of their sounding during measures 3–4
and 4–5 (15.10c). Smith instead proposes a significant formal break
after measure 8’s fermata, with the onset of a “scherzando idea” that
“contrasts . . . the movement’s more leisurely main theme” (p. 2011/
257). In my interpretation, that main theme terminates instead at 121
and in no sense requires an A’ continuation (within a ternary ABA’, as
Smith proposes on p. 2009/64).23 By setting the stage for the arrival of F
Minor at the onset of the first episode, what transpires during measures
12|13 through 15 performs the duties of a transition, in the process
deploying the melodic and harmonic content of the opening measures
(just as material from P frequently recurs at the onset of a sonata
exposition’s TR).

Movement 2 first episode (measures 16–25)

The shift to F Minor for the rondo’s first episode is accomplished through
the temporary lowering of Kopfton A to A ♭, coinciding with the shift to a
four-flat key signature. In that the section ends on the dominant, the
principal melodic trajectory is A♭>G, with upper-neighbor B♭ intervening
during the IV harmony of measure 22. (That line is displayed prominently
in 15.11.)

Example 15.11 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 2, mm. 16–25.
212 Harmony in Mendelssohn and Schumann

The harmonic progression from I to V♮ in F Minor features an internal

tonicization of A♭ Major. A comparison of 15.3 and 15.4 (which docu-
ment the relationship between the minor tonic and its mediant during the
first movement) and 15.11 is instructive. Note that in both cases a circle
of fifths serves as the mechanism for shifting from one key into the other
(A Minor to C Major in the first movement, F Minor to A♭ Major in the
second movement). Given that both structures deploy Kopfton ^3, the shift
to the mediant induced Schumann to create some space above ^3 for a
descending linear progression. During the first movement a G emerges
above Kopfton C, creating space to accommodate a descending fifth-
progression; during the second movement a C emerges above the lowered
Kopfton A♭, creating space to accommodate a descending third-
progression. Though what transpires during measures 19|20 through 22
is somewhat more intricate than what happens during measures 16 and
17, both the tonic and the mediant regions within the second movement’s
episode essentially pursue the same broad harmonic progression (in their
respective keys).
Given the restoration of tonic A Minor near the end of the first
movement’s exposition, the mediant’s potential to instigate a broad
A<C<D<E trajectory in the bass is there left unfulfilled. The second
movement episode’s broad F<A♭<B♭<C bass thus in its own way
belatedly fulfills that earlier expectation. The return of the A♭ chord
in its first inversion during measure 22 (at the end of the mediant
tonicization) becomes a motivic feature of the bass: A♭<C during III
followed by B♭<D♭ during IV, with C>A♭ and then D♭>B♭ reciprocating
in the melody (thus creating voice exchanges, as marked in 15.11).
Though the melody’s goal pitch G does not sound in any register at the
onset of the dominant (during 231), the reiterations of the cadential
succession over the following two measures makes up for that lacuna,
now instead leaving neighbor D♭’s resolution pitch C unsounded in the
upper register.24

Smith’s ex. 2009/13 and my 15.11 have much in common. Though he

reads the movement’s Kopfton as ^5 (contrasting my ^3 interpretation),
during the episode his ^3>^2 exactly matches my ♭^3>^2. Since that is the
case, I suggest that he might have flagged B♭ rather than D♭ in the
melody of measure 22. Readers might compare our respective means
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 213

of conveying how Schumann moves from tonic F to mediant A♭

(where I more overtly project the circular course of the chordal
progression) and my emphasis upon the C-E♭-A♭ chord during 221
as the terminus of III’s prolongation and goal of a voice exchange.
(Smith emphasizes the melody’s A♭ during 221 via both a stem and a
slur from an earlier A♭, while the bass C curiously appears as a passing
note between D♭ and B♭.)

Movement 2 second episode (measures 41–56)

In a conventional account of rondo form, one might encounter a list of the

keys in which the episodes are presented. The first episode’s F Minor (the
parallel minor key) is a common choice to contrast an F Major refrain,
while the second episode likely would be listed as transpiring in the key of
D Minor (the relative minor key), as is conveyed in 15.12. Yet I propose
that Schumann is attempting something subtler here. My ear is not quite
willing to accept D as a local tonic. The G chord of measure 44 still asserts a
II function in F Major, with the A chord that follows reverting to an earlier
point in the progression. In this conception a harmonic progression in F
Major – I5–6 II ➔ V8–7 I – appears to get mired hopelessly, with a way out
offered only at the return of the refrain, starting at the upbeat to measure
57. (This new context for the refrain motivates a reharmonization of its
opening.) The harmonic progression is in fact lifted from measures 7
through 9, where an A-C♯-E embellishing chord comes between the tonic’s
5- and 6-phase chords, as shown in 15.10e. That model’s A➔ D ➔ G ➔ C ➔
letters recur in 15.13, where a comma indicates that the A-C♯-E chord of
measure 44 is not the supertonic’s expected successor. The progression has
backtracked, and it will do so again before the episode is over. (These
recurrences are displayed within large parentheses in the graph to distin-
guish repeats of what precedes G from what follows G beginning in measure
57.) I hear the G chord at 441 relating also to the B♭ chord from the first
episode, at 222. (Both chords support lowered Kopfton A♭’s incomplete
upper neighbor, B♭.) These multiple references to earlier moments instill a
sense that a C dominant with soprano pitch G will sound next. We are
rewarded eventually for not giving up hope: that long-awaited C chord
finally arrives at 572.
214 Harmony in Mendelssohn and Schumann

Example 15.12 Faulty analysis of Schumann: Sonata in A Minor for Violin and Piano
(op. 105), mvmt. 2, mm. 37|38–44.

Example 15.13 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 2, mm. 37–69.

In this scenario the supertonic is given multiple opportunities to proceed

to the dominant. That of measure 44 (replicated in measure 54) fails, with a
reprise of D’s A-C♯-E embellishing chord sounding where V is expected.
Eventually the supertonic is inverted in the context of a BG♭ GB♭ voice
exchange, after which it evolves into a surge, with bass B♮ dynamically
targeting dominant root C at 572. This path forward emerges only as the
second episode gives way to the thematic material that opens the refrain,
now presented in an altered harmonic context. Note that an F-A-C chord
sounds in the middle of the succession from A➔ to D➔ during measure
56.25 That event helps explain the otherwise curious F-A-C chord of
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 215

measure 47: instead of continuing the transformation with a descent to D

there, the A➔ chord is reinstated.

Though Smith does not provide a full account of the second episode, his ex.
2009/15 shows some highlights. It is clear that he regards D Minor as the
local tonal center, where the A-C♯-E chord serves as V. (Compare with the
V♯ in my 15.12, which I do not fully endorse.26) Though the sense I make
of the F major chord during measure 47 comes through association with
what occurs during measure 56, any chord may be embellished by such a
third-related chord: see “Chromatic lower-third and upper-third chords”
in my Harmony in Schubert, pages 56–60. According to 2.17 in that
volume, an A5–6–5 embellishment that proceeds from A-C♯-E through
A-C-F (unfurled as F-A-C) rather than through A-C♯-F♯ (diatonic in the
context of an A major chord’s prolongation) would be deploying the 6-
phase chord’s Chromatic Variant ♯2. If one grants a III label to the F-A-C
chord of measure 47, then I suspect one would find it reasonable also to call
the A-D-F♯ and A-D-F chords of measures 49 and 50 I♯ and i♮, respectively.
In 15.13 I propose instead that D and F(♯) serve as passing notes leading up
to the A chord’s fifth (E) and seventh (G), without asserting a tonic
function. The tension of the A➔ surge is not released during measure 49
or 50, but instead during the second half of measure 51. Smith may have
been trying to convey something along those lines by placing the I♯ and i♮
numerals in a separate row, though the fact that his example stops abruptly
after measure 50 leaves one without a full picture of his conception. I
suspect he would be as troubled by the paucity of Roman numerals in my
15.13 (seven numerals over thirty-three measures) as I am by their abun-
dance in his ex. 2009/15 (eight numerals over six measures).

Movement 2 final refrain and coda (measures 56|57–79)

Due to the complications during the second episode, described above, the
third presentation of the reprise does not begin with a tonic harmony.
Instead, the episode and reprise together form an integrated structure
(conveyed in 15.13) that cadences only at the end of the reprise (measure
69), prior to the onset of the coda. The hairpin symbol before the final tonic
chord corresponds to that deployed in 15.9 (the initial statement of the
reprise). We saw how the range of related structures juxtaposed in 15.10
provided an abundance of content for that hairpin location during the
216 Harmony in Mendelssohn and Schumann

opening refrain. Though much of this material recurs now during the
closing refrain, Schumann has added a fresh variant during measures 63
through 68, as shown in 15.14. Here the internal F-A-C chord undergoes a
brief tonicization (displayed within large parentheses) before its 6-phase
chord emerges in measure 66. That turn of events replaces the A-C♯-E
chord (measure 8) that later was so thoroughly ingrained within the second
episode that Schumann likely regarded its potential as used up by this
point, motivating his alternative trajectory. Once the tonic root F arrives in
the bass at 691, a coda commences, deploying local harmonic progressions
(which get underway even as the two suspensions from the preceding
dominant harmony are just resolving) to further secure the tonic.

Example 15.14 Analysis of Schumann: Sonata in A Minor for Violin and Piano
(op. 105), mvmt. 2, mm. 63–68

The end of the movement is not addressed in Smith’s commentary.

Movement 3 exposition: P (measures 0|1–24)

The twenty-four measures that Schumann earmarks for the third movement’s
P are divided into three eight-measure units. The second unit is structurally
related to the first, in that both fulfill the responsibilities of an antecedent
phrase (proceeding to the dominant with a melodic descent to ^2). The third
serves as a consequent phrase, with measures 16|17 providing the same
initiating material as was deployed in measures 0|1, proceeding this time to
a cadence on the tonic (supporting ^1). In 15.15 the second unit is interpreted
as integral (rather than as a mere reiteration or as a contrasting region), since
the consequent draws upon content from both of the antecedent units.
Example 15.15 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105), mvmt. 3, mm. 0|1–24.
218 Harmony in Mendelssohn and Schumann

We noted above how the neighboring-note embellishment of one or two

tonic triad members is featured at the beginning of the first two move-
ments. That initiative is expanded for the third movement: both lower and
upper neighbors embellish in turn the A minor triad’s root, fifth, and third.
Once that chord is secured, an initial ascent to Kopfton C (= ^3) transpires.
The remainder of the phrase proceeds through II➔ to V♯, incorporating a
6 6
4 embellishment of II➔ that initially is unfurled into 3 position. (Observe

how bass G<A<B coordinates with the melody’s B>A>G above.)

Though the pitches that shape the second phrase’s chordal trajectory
migrate both between instruments and between registers, the depiction
in 15.15 proposes an underlying voice leading featuring exemplary surge
(and supersurge) deployments. The restored tonic of measure 9 already is
surging, setting off a circular progression that proceeds through A
Minor’s diatonic D, G, C, and F chords. Roots F and B form the one
imperfect fifth within the circle, thus not susceptible to a surge. Though
the abundance of embellishment complicates matters, I read the line near
Middle C in measure 13 as (C)<D<D♯, where passing note D comes
between the imagined C of an F chord and the D♯ of a B chord.27
Though the latter might be fashioned either as D♯-F-A-C or as D♯-F-A-B
(both conveying a B⇨ supersurge), I have incorporated the latter into my
graph (corresponding to a location in the score where neither B nor C
sounds, thus requiring some imaginative chord-completing operation)
because B would be the expected soprano pitch given the voice leading of
the previous chords and because a B is featured (within a contrasting
harmonization) at the parallel location during the consequent phrase
(D-F-B at 232).28
The consequent phrase repeats the first phrase’s initial ascent, though
the C that would be expected as its goal is elided, making way for the C♯
corresponding to the onset of the second phrase (at 91), now with C♯
sounding at the top of the texture (justifying its parenthetical placement
there at measure 9 in the graph). Whereas that earlier surge set a circular
progression in motion, now a more harmonically oriented chordal pro-
gression is pursued, still supporting the same melodic trajectory (from C’s
upper neighbor D through passing note C down to B). Strongly at odds
with the meter, Schumann fills the first three eighth notes of measure 22
with a projection of IV (incorporating a FDDF voice exchange), then devotes
three eighth notes (straddling measures 22 and 23) to the projection of a
passing chord, and finally one eight note to IV’s 6-phase D-F-B (into which
one might imaginatively prolong the alto A from the passing chord,
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 219

preceding the G♯ of the dominant that follows). This leaves one eighth note
at the end of measure 23 for V ♯7 , so that I may arrive on the phrase’s final
downbeat.

Again Smith and I are at odds regarding a movement’s Kopfton. In my

view an initial ascent in the opening measures (targeting the sforzando C
at 51) is a very strong marker of ^3, as is the descending fifth-progression
from ^2 that follows immediately thereafter. The C♯ of measure 21 leads
chromatically upward to D (as a neighbor), whereas Smith instead dis-
plays a downward trajectory from E to D in his ex. 2009/18, with no
indication of how or when Kopfton ^5 was established.
The unusual array of chords during measures 22 and 23 likewise has
generated contrasting interpretations. If ^3 does serve as the Kopfton, then
it seems to me that the IV harmony of measure 22 would have a broad
impact, as in my IV5–6 leading to the phrase’s V♯. Smith instead shows an
internal tonic restoration, with a stemmed D to a stemmed C in both
outer voices, followed by a flagged bass D that, though unlabeled, suggests
an inverted II as the principal intermediary between I and V♯.
Though we agree that a dominant arrives at the end of measure 23, in
the broader scheme of things the dominant of measures 14 and 16 (not
shown in Smith’s ex. 2009/18, though labeled as V in the score excerpt
provided in his ex. 2009/19) is hierarchically deeper (as beamed in 15.15).
Since the issue will come up again later in the exposition, I should com-
ment on his VI label for the F-A-C chord of measures 13 and 15 (in his ex.
2009/19). His reading highlights the affinity between tonic A-C-E and
submediant F-A-C. I instead regard the principal chord of those two
measures as F-A-(B)-D♯. We both acknowledge that a circle of fifths is at
work, though I proceed along its path further than he does. The awkward
diminished fifth within the A Minor circle resides between F and
B. Schumann smoothes out that blemish by retaining F as the bass, with
the B-rooted chord conveyed without an actual sounding of the pitch
B. Because two of the circle’s chords share the same measure, the traversal
through seven chords fills the phrase’s first six measures, allowing for a
reprise of the final three chords during the next two measures. The
dominant goal is prepared potently by II⇨, not meekly by VI, as the
deployment of Roman numerals in his ex. 2009/19 seems to imply.29 As we
shall see presently, a supertonic chord figures prominently between what
Smith might label as VI and V also during the remainder of the exposition
(beginning in measures 35 and 36).
220 Harmony in Mendelssohn and Schumann

Movement 3 exposition: TR ⇨ FS (measures 25–58)

The exposition of a minor-key sonata movement usually will proceed to the

mediant (a chord of major quality) or to the dominant (of either minor or
major quality). In this exposition Schumann samples both in turn – twice! This
reading emanates from the conspicuous fact that the material
over the measures 41|42 bar line corresponds to that of measures 29|30.
What happens during measures 42 through 57 should come across as a
varied repetition (a strategy similar to that which occurs during measures 1
through 16): the cadential E-G♯-B chord of measures 56 and 57 (the EEC)
was attained already in measure 41. In fact, it was attained earlier as well, in
measure 27. The seemingly inconsequential I-to-V♯ trajectory of measures
25 through 27 turns out (abetted by multiple reiterations) to evolve into the
main harmonic trajectory during this latter portion of the exposition, with
the B and D that sound above pedal point A during measure 26 eventually
developing into a full-fledged II harmony, coming between I and V♯ during
measures 36, 38, 40, and 56.
Three foundational trajectories (of which the second and third are closely
related) are presented in 15.16. In the first a local motion to tonic A’s 6-phase F-
A-C (measures 29 through 31) is followed by the arrival of A’s mediant, C-E-
G.30 The A-to-F connection is facilitated by what I interpret as a segment of the
descending circle of fifths, with one elided chord: A (D➔) G C➔ F. The
imaginative insertion of D➔ here is confirmed by Schumann’s inclusion of
that chord during the next A-to-F trajectory, during measures 33 through 35. If
this C mediant chord were to be accepted as the exposition’s tonal goal, then
what follows might initiate a tonicization of C Major, with I5–6 II➔ transpiring
during measure 33. Yet once that progression proceeds a bit further we
recognize it as a reiteration of the circular progression from A to F (now with
no elided chords). Schumann has backtracked to the A starting point, where
this time a trajectory through II to V♯ follows after the F chord. (This, in my
view, is the most normative continuation from I5–6.) Schumann pursues that
course three times in close succession (as is conveyed by the multiple rows of
measure numbers annotating 15.16b, with the first two dominants resolving
back to A), concluding with the E dominant at 411. Though not strongly
marked, this V♯ is essentially the endpoint of the exposition’s structure. All
the chordal energy (wherein “TR becomes FS” (TR ⇨ FS) in the context of a
continuous exposition) has been devoted to attaining the E major dominant
(after first auditioning mediant C as an alternative), and none saved for its
tonicization thereafter. Instead this robust trajectory is repeated, with
modifications.
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 221

Example 15.16 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 3 (a) model for trajectories from I to III during measures 29–45; (b) model for
trajectories from I through II to V♯ during measures 33 through 57; (c) model for a
trajectory from I through IV to V♯ during measures 45 through 56.

It indeed is unusual that A Minor’s tonic chord is still in play as late as

measures 41 and 45 within a 57-measure exposition (see 15.16a and 15.16c).
Yet due to the repetition initiative, that is indeed the case. Schumann modifies
the progression in several ways during this second pass. First, given how the F
and C chords are juxtaposed in 15.16a, a composer might consider how C
could be made to serve as F’s dominant. Between measures 47 and 52 (where
only the F-A-C chord appears in 15.16c), Schumann undertakes a brief
tonicization of the submediant, pursuing a I II7 ➔ V ➔ I harmonic progres-
sion in F Major. Then the supertonic of the earlier broad progression is
replaced by the subdominant during measures 52 through 55, with goal V♯
attained at 561. Just as the arrival on E at measure 14 is repeated at measure 16,
now the arrival on E at measure 56 is repeated at measure 57, creating a
context for the reinstatement of II (instead of IV) as the dominant’s prede-
cessor. (Thus a part of 15.16b is integrated into the structure, as its lowest row
of measure numbers indicates.) Note especially how the violin’s B<D<F
arpeggiation from 401 is presented in retrograde (as F>D>B) over both the
measures 55|56 and 56|57 bar lines, now with added embellishment. Though
the dominant serves as the goal harmony for the exposition, the D among the
violin pitches in both measures 56 and 57 (contrasting the situation in
measure 41) ignites a surge, targeting the tonic first for the repeat of the
exposition or later for the onset of the development.

After a cursory glance at the score, most musicians would describe what
transpires at measure 58 as a “first ending” and a “second ending.” I propose
222 Harmony in Mendelssohn and Schumann

that these measures are not endings at all, but instead beginnings. Observe
that Schumann has placed the reciprocal repeat sign between measures 1
and 2. Consequently the doubled E before the downbeat of measure 1
corresponds to the doubled E before the downbeat of measure 58. The
exposition ends on the preceding dominant (E) – not on tonic A, as Smith
proposes in his ex. 2009/18 (labeled as “graph of exposition,” reprinted as
ex. 2013/16). And though the dominant root E of measures 56 and 57 is
attached to his principal bass beam, the E of measure 41, which I regard as
its equivalent, is not. (That chord appears just to the left of measure number
41 in Smith’s ex. 2009/18, where it is interpreted as a local dominant within
a i6–V–i progression spanning measures 39 through 41.31)
Smith’s reading lacks an acknowledgement of tonic A’s pervasiveness as
a starting point. At the onset of TR, A proceeds to E effortlessly (measures
25–27). Then a conflict arises: a slide from A down to F, followed by a C
chord (measures 29–33). This is a serious challenge to A’s hegemony. Yet A
ultimately regains control by incorporating the F chord within its broader
A-to-E trajectory (as shown in my 15.16b). Smith’s graph omits the A
restoration at 332, and so the A>G>F>E>D bass that corresponds to A
Minor’s I5–6 II is not projected. Concurrently the melodic B (at 362) that
serves as C’s successor in a descending linear trajectory is displayed as an
ascending passing note in his graph. Schumann persists in this initiative
through 411: a dominant harmony supporting ^2 in the soprano, the goal of
all that has transpired since the onset of TR. The following beat’s A chord
(with a restored ^3 in the soprano) is a backtracking to 292.32
The two threats to A’s preeminence are reprised. Just as an F chord
emerged at 311, so also at 431, and with the same F-to-C continuation.
Likewise the F chord of 351 recurs during measure 47, though now
inverted. (I thus regard Smith’s reading in ex. 2009/18 as far from the
mark. The inversion of the F chord has to do with the fact that the
impending F Major tonicization proceeds to a supertonic chord with
bass B♭. Schumann merely takes the shortcut of allowing A to precede
B♭, rather than descending all the way to F before the B♭.) Temporarily
departing from the earlier presentation, this F chord mounts an even
greater appeal for hegemony, though ultimately the F>D<E bass of mea-
sures 35–36 (and 39|40–41) is reprised during measures 52–57. (Though
Smith displays bass F in the vicinity of measures 53–54 with a flagged stem,
no harmonic role is indicated for the D-F-A chord in his analysis. Since the
stemmed soprano D at that point migrates to the bass, where it ascends to
E, D<E should not be displayed also in the soprano. I interpret that soprano
D as an incomplete upper neighbor to Kopfton ^3, followed by B (= ^2).)
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 223

Thus the A-to-E trajectory first stated during measures 25–27 ultimately is
confirmed. Though Smith reads a “closure in F” and “the return of A minor
as a structural harmony” (p. 2009/69) at the end of the exposition, the
exposition’s goal instead is dominant E, a far more conventional choice
than either A or F and not a member of his TMS complex.33

Movement 3 development (measures 57|58–ca. 116)

In a useful essay on development sections, Edward Laufer offers thirteen

models of “typical development-section paradigms” (his figs. i–viii), most
of which commence on the dominant, though III to V and IV to V are
represented as well. Turning the page, one finds seven more models,
representing “further development-section paradigms” (figs. ix–xii), the
first of which shows the tonic leading to the dominant during both the
exposition and the development.34 In fact, the deep structures of these two
sections are identical: in both cases CA proceeds to BE (transposing from the
published model’s C Major into A Minor). Laufer comments as follows:
“fig. ix presents a rather different situation, of relatively infrequent occur-
rence . . . The sonata’s characteristic dichotomy – that of three formal
divisions at odds with two harmonic-contrapuntal divisions articulated
by an interruption . . . – is here undone . . . In a sense, fig. ix is rather closer
to a rondo design.”
Schumann’s thought process in producing this outcome may have run
something like the following: “If I am going to proceed again from I to V♯
(supporting ^3 to ^2) during the development, why not use exactly the sort of
trajectory that generates that structure so admirable during an exposition?
In this movement I have pursued what those American analysts of the next
millennium, Hepokoski and Darcy, will label as a continuous exposition
(with P followed by TR ⇨ FS). Why don’t I create a trajectory resembling
what they call the two-part exposition (P TR ’ S / C), which I have not
deployed during either of this sonata’s expositions, for the development
section! Damschroder certainly will find that to be an exciting turn of
events to discuss when writing his book on my music.”35
Exuding that prophesied excitement, I suggest that the restored tonic of
measures 57|58 through 61 (which deploys motivic content from P) leads into
a TR that concludes with a medial caesura (MC) on B in measure 76.
Schumann’s breaking out into song in E Major starting in measure 78 thus
resembles an S. The cadence at 1011 serves as the equivalent of an EEC. What a
novel and interesting way to proceed through a development section!
224 Harmony in Mendelssohn and Schumann

Given that A Minor’s dominant (E-G♯-B) will be tonicized during this

pseudo-S, we would expect TR to lead to a chord that aptly precedes that
dominant. No trajectory is more suitable than I5–6 II ♯♯ 53 , which is pursued
over the course of 15.17. To convey the essence of a medial caesura, the B
chord at 761 initially is consonant, before its seventh (A) emerges, first in
the piano part and then in the violin line during the caesura-fill.36 In a
minor key the diatonic I5–6 II trajectory is complicated by the fact that the
tonic’s 6-phase pitch, F, is not related by perfect fifth to B, the root of II.
Schumann responds to this situation in a most interesting way. At first the
6-phase chord is transformed completely for a dominant-emulating surge
targeting the supertonic. This requires three chromatic pitches: F♯, A♯, and
C♯ (asserting ♯VI➔ during measure 69). Concurrently the supertonic’s
fifth shifts from diatonic F to F♯, making it more amenable to a surge. (That
is, if a surge by definition emulates a dominant, then the chord that it
targets should emulate a tonic, which is never a chord of diminished
quality.) Yet after the supertonic is attained (as a chord of minor quality)
during measure 70, its prolongation backtracks to an alternative of the
tonic 6-phase chord: C-E-(G)-A♯ during measure 75, thus more faithful to
the diatonic context of the original A Minor key. This is, of course, a
supersurging chord – ♯VI⇨ – with root F♯ omitted. The supertonic to
which it resolves now is of major quality, and so the ♯VI⇨ II➔ V♯
trajectory (labels that I suggest need to be retained, since ultimately the E
goal chord is A Minor’s major dominant) has the impact of II⇨ V➔ I in E
Major, by far the most common context for a supersurging chord.

Example 15.17 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 3, mm. 58–80.
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 225

The B and E chords that play such foundational roles in 15.17

(displayed as open-notehead arrival points at measures 70/76 and
78–80) take on subordinate roles earlier in TR (with arrival points
during measures 62 and 66). Though some analysts might be inclined
to project the decisive dominant arrival as early as measure 66, I
interpret that chord as a connector between the tonic’s 5- and 6-
phase chords. The whole purpose of TR is to prepare for the arrival
of V♯ – not to have V♯ creep in early, fourteen measures before S gets
underway.
The violin’s dissonant A during 772 sets a descending fifth-
progression from B (= ^2 in the context of the entire movement) in
motion. As occasionally happens, the B and A of that fifth-progression
occur before II➔ resolves to V♯, thereby taking advantage of the AD ♯
dissonance, which targets the GE ♯ of V♯. (See measures 70 through 78 in
15.18.) The harmonic trajectory between measures 78 and 84 may,
with some modest chord-completion in the bass, be interpreted as I5–6
II➔ V in E Major, with ascending lines (G♯<A♯<B above B<C♯<D♯)
preventing progress toward closure. The next phrase, which commences at
842, builds from the same starting point but soon makes distinct progress,
with the broad fifth-progression’s ^2 arriving during 891. Yet the potential of
the following B and E chords to serve as asserted V and I harmonies
evaporates as the momentum pushes onward to the A chord of measure
92. In 15.18 these chords are interpreted as members of a surge-rich circular
progression leading from II to its upper-third chord, from which one might
still succeed in reaching a cadence (if followed by V and then I). However,
this A chord instead undergoes a disconcerting tonicization during mea-
sures 92 through 96.37 As if nonplussed by that turn of events, Schumann
elects not to proceed forward from that point. Instead he backtracks a bit –
to the onset of II, from which a surge (with EA ♯ at 982 matching that of 892)
targeting an asserted V that, unlike the B chord of measure 90, does in fact
succeed in resolving to the cadential I, which unfolds during measures 101
through 103.
226 Harmony in Mendelssohn and Schumann

Example 15.18 Analysis of Schumann: Sonata in A Minor for Violin and Piano (op. 105),
mvmt. 3, mm. 70–103.

Once the harmonic progression tonicizing E Major has run its

course, the remainder of the development (measures 101 through
116) takes on a more typical developmental stance (borrowing melodic
content from the exposition P), awaiting the A Minor tonic restoration
at the onset of the recapitulation. The C♮ deployed during 64 embellish-
ments of the E-G♯-B dominant during measures 104–107 and 115 is a
reminder that the tonic that will emerge is that of A Minor – not A
Major. The distinctive bass ascent to F♮ in measures 110 and 112
resonates with A Minor as well, but also (with its successor D) relates
potently to two earlier points within the sonata: the latter part of the
third movement’s exposition (as displayed in 15.16b) and the domi-
nant prolongation during the first movement’s P (as displayed in
15.2a). The recapitulation’s tonic arrival in measure 117 brings the
extended dominant prolongation to an end.

In its broad conception Smith’s ex. 2009/16 is quite similar to my 15.17.

We both interpret root B as a fulcrum between the tonic A at the onset of
the development section and the dominant E that is tonicized beginning
in measure 78. And though Kopfton ^5 appears at the onset of his graph,
soon ^3 takes over, descending to ^2 (which I interpret as a background
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 227

event). I applaud the use of II♯ (rather than V in the upcoming tonicized
key of E Major) as the Roman numeral at measure 78.
One feature of Smith’s graph puzzles me. Where he shows an DF♯ sixth
in the space between measure numbers 68 and measure 71, I instead
hear an F♯➔ surge. In that I regard the chord of 752 (which he labels as
Itl. 63 ) as F♯⇨, my reading of the harmonic progression contains a
reiteration: F♯➔ B, then (subordinately) F♯⇨ B. His melodic C>B>A♯
spans measures 61 through 75, whereas my similar third (with an
imagined A♯ at the top of the texture) spans measures 58–61 through
69. Thus my supertonic arrives sooner than does his and is at first a
chord of minor quality. The voice leading I propose between the initial
A tonic and the F♯ chord involves parallel tenths, whereas his broader
trajectory involves a voice exchange.

Movement 3 recapitulation (measures 117–213)

The recapitulation’s principal harmonic task is to get beyond the domi-

nant endpoint of the exposition’s TR ⇨ FS (measures 41 and 56–57) to
achieve a PAC in A Minor – the ESC. The graph showing this outcome is
straightforward. (See 15.19, awaiting commentary below that will explain
the discrepancy between the score and the chords displayed within the
box.) The novelty of Schumann’s writing during the recapitulation cen-
ters on an elective undertaking during this typically lower-interest por-
tion of a sonata movement.38 To enliven what otherwise might be a
predictable pasting in of content from the exposition for reuse during
the recapitulation, Schumann allows the material to be jolted by a seismic
shift, wherein everything moves unexpectedly to a different tonal plane (in
this case up a third).39 Sometimes such a shift will be maintained for the
remainder of a movement. Yet in this case, after almost thirty measures in
the elevated tonal region (which shifts further during its prolongation), a
breathtaking collapse back to the original tonal plane occurs, allowing
Schumann to proceed to the ESC (after additional flourish) as if nothing
out of the ordinary had happened. (My graph displays the foundational
pitches as they would transpire without the seismic shift. Thus the G♯
“dominant” chord of measure 157 corresponds to the post-shift E chord
of measure 176.)
228 Harmony in Mendelssohn and Schumann

Example 15.19 Analysis of Schumann: Sonata in A Minor for Violin and Piano
(op. 105), mvmt. 3.

To get our bearings in the unstable tonal terrain of the seismic shift, it
will be useful to talk through Schumann’s strategy using root letters that
also indicate chordal quality (for example, a = minor triad A-C-E; F =
major triad F-A-C, b° = diminished triad B-D-F, etc.), rather than the
Roman numerals of 15.16. The exposition’s TR ⇨ FS material projects two
trajectories in alternation: from a to E, and from a to C. Between measures
25 and 41 these initiatives play out as follows:

m. 25 27 29 31 33 33 35 36 36/41
a E | a~~~~~~F C | a~~~~~~F b° E
tonic to dominant tonic to mediant tonic to dominant

(A special symbol – ~ – is used to designate a 5–6 shift.) As proposed

earlier, these trajectories together constitute the essential structural content
of TR ⇨ FS. After the E chord of measure 41, the second and third
segments are repeated, with an internal expansion of the latter, as follows:

a~~~~F b° E becomes

m. 45 47 48 50 52 52 56
56 57
a~~~~F(= F g C F)
D or b° E
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 229

Making sense of Schumann’s writing during the recapitulation can come

about only by carefully assessing how each of these five units fares in its
new context.
Measures 25–28 into measures 141–144: The exposition maintains a
tonal continuity between the end of P and the onset of TR: an a chord both
concludes the former and inaugurates the latter. The bold new thematic
idea that launches TR deploys modest harmonic support, traversing the
path from a to E without much fuss. That continuity is annulled during the
recapitulation presentation. After P’s cadence on a, C to G transpires at
TR’s onset.40 Granted, in many recapitulation transitions the tonal trajec-
tory shifts because the exposition S and the recapitulation S are in different
keys. But this TR does not proceed to an S. The exposition’s goal E will be
attained later during the recapitulation as well. Thus the shift from a to C is
essentially an elective overhaul that eventually will be rescinded. While
adding interest for the listener, who now should be on high alert regarding
what Schumann has in store, ultimately it is superfluous to the unfolding of
the broad structure. We simply traverse a part of the recapitulation in a
higher tonal plane, attained via a seismic shift. That said, the juxtaposition
of a and C is a signature feature of the composition as a whole. Recall that
the first movement exposition’s TR ⇨ FS was “supposed” to be in C but
slipped down to a. Now the third movement recapitulation’s TR ⇨ FS is
“supposed” to be in a but has risen up to C!
Measures 29–33 into measures 145–149: The audacious jump up to C at
the onset of TR remains tonally allied to the A Minor key of the movement as
a whole, where C serves as the diatonic mediant. (This situation will change
later, as will be explained below.) The exposition’s a~F C emerges during
measures 145 through 149 not in an exact transposition (for example, as
c~A♭ E♭), but instead as C~a E, thus closely aligned with A Minor. Whereas a
circle of fifths works ideally in traversing the span of the descending major
third a~F, C~a is less amenable to that device (since one of the fifths would
need to be of diminished quality). Schumann does not pursue a circular
trajectory in the local connection of measures 145 through 147.
Measures 33–41 into measures 149–157: The C~a connection of mea-
sures 145 through 147 “should” be replicated during measures 149 through
151 (based on the traversal of a~F in both measures 29–31 and 33–35).
Schumann defies that expectation in two ways: first by omitting the C
chord entirely, and second by substituting A for a. Though A lies outside
the movement’s diatonic realm, its major quality resonates with the earlier
F (within the exposition’s a~F). But if that is the case, then A’s predecessor
would have been c♯. As this model is repeated, a c♯ chord in fact sounds
230 Harmony in Mendelssohn and Schumann

during measures 153 and 155. The exposition’s a~F b° E is conveyed as

c♯~A d♯° G♯ – an exact transposition! – during measures 155 through 157.
Measures 41–45 into measures 157–161: Given the intervening shift to
C♯ Minor, measures 157 through 161 do not replicate measures 145
through 149 (C~a E), but now project c♯~A E (thus exactly transposing
the exposition’s a~F C). The major third c♯~A now is filled in by chords
from the descending circle of fifths, as was its a~F predecessor during the
exposition.
Measures 45–57 into measures 161–176ff.: Though the shift from A
Minor to C♯ Minor was undertaken in stages, by measure 161 the latter
key’s stability is secure. Attentive listeners might wonder whether the
broad progression will continue upwards to dominant E. Schumann
instead pursues a wonderful and unexpected course at this delicate
moment. Recall that a’s 6-phase chord, F, is tonicized during the exposition
(beginning in measure 47). It was conveyed above as

F(= F g C F)

The equivalent passage within C♯ Minor takes place beginning in measure

163 as
. . .)
A(= A b E

Though the E chord (the A Major tonicization’s dominant) surges, its

successor is not the expected A, but instead F! Suddenly the bubble has
burst. Our long-extended pursuit of the exposition’s TR ⇨ FS content in
an upper-third context (C, then c♯) comes tumbling back down to the
foundational tonal plane. The F chord of measure 168 corresponds to the F
chord of measure 52. Whereas during the exposition first d and E and then,
in a modest backtracking, b° and E follow the F chord, b⇨ and E (remi-
niscent of measures 180 through 189 from the first movement) occur
during measures 171ff.
The movement’s harmonic resolution (getting beyond V♯ to a PAC on I)
is gradual. Though the dominant’s root E takes hold in measure 176, it is
prolonged as a pedal point below a new approach to V♯, incorporating IV
beginning in measure 182 and its 6-phase chord, in an assertion as II➔,
beginning in measure 184. Though V♯ gives way to I at measure 192, the
pitch A sounds neither as that tonic chord’s soprano nor as its bass. Further
harmonic activity features IV (starting at 1961) and its chromatic 6-phase
D-F-B♭ (unfurled at 1991). The tonic’s bass A is attained finally at 2001,
while soprano A emerges at 2011 and again at 2031, corresponding to the
ESC.
 Schumann: Sonata in A Minor for Violin and Piano (op. 105) 231

Smith does not offer a substantial account of the recapitulation. He

focuses only on the F chord that emerges at measures 168 (which he
interprets as residing within a coda). While he aptly correlates this F
chord and that of measures 177ff. of the first movement, his interpreta-
tion of what it means to add D♯ to F-A-C contrasts mine. Whereas he
proposes a “transformation of VI into an augmented-sixth chord, as a
stepping-stone back to V” (p. 2009/75), I view the VI as an extension of
the initial I (as its 6-phase chord), with F-A-C-D♯ serving as II⇨, the
principal intermediary between I5–6 and V♯ (just as diatonic II did
multiple times earlier in the recapitulation). In my view the practice of
referring to such chords via their ubiquitous “Italian,” “French,” and
“German” nicknames unfortunately encourages interpreting them as
belonging to a subordinate hierarchical level. One benefit of my II⇨
label is that it puts such chords on an equal footing with other promi-
nent supertonic deployments.
 Notes

1 Foundational diatonic processes

1. Mendelssohn’s strategy here is to segment a pervasively upward arpeggiation

engaging the tonic triad’s scale degrees ^1, ^3, and ^5 into discrete subgroups.
Foundationally the melody’s arpeggiation would proceed as follows:

C♯ E A C♯ E A C♯ E A
Because such a lengthy ascending arpeggiation soon would extend beyond
music’s normative range, each downbeat note after the first is understood to be
replicated an octave lower (actually sounded by an inner voice in the composi-
tion), so that the arpeggiation takes on a zigzag shape, as follows:

The five R pitches are enclosed within boxes, with scale degrees indicated in a
separate row. The underlining within the diagram indicates that the first
interval within each of the resulting three-note groups is filled in by passing
notes (the topic of question 3).
2. Though the ^1 at R5 provides a suitable ending for the initiative explored in
question 1, Mendelssohn launches another trajectory that hoists that A
upwards by two octaves. A confirming descent that restores A in the register
of R5 follows, achieved during the passage from Z1 through Z3 by means of the
downward arpeggiation ^8>^5 >^1 in which the initial ^8>^5 is projected as a
descending eleventh so that the concluding ^1 (= Z3) corresponds in range to
that of R5.
3. The melody’s intervals at the spots marked W result from the concurrent
pursuit of two distinct initiatives: (a) the filling-in of spans formed by adjacent
pitches within an ascending arpeggiation; and (b) persistent downward
registral shifts. Only W1 transpires without registral shift: the diatonic
interval formed by ^3<^4 (a minor second) occurs. At W4 the pursuit of the
232 same initiative, concurrently incorporating a downward registral shift (thus
 Notes to pages 4–8 233

^3>^4), results in the melodic traversal of that minor second’s inversion, a

major seventh. W2 and W3 correspond to the ascents ^1<^2 (within ^1<^3) and
^5 <^6 (within ^5 <^8), respectively. Given the major scale’s construction, these
intervals both are major seconds, projected here as minor sevenths through
inversion.
4. The underlines in the second diagram accompanying the answer to question 1
designate filled-in segments of the arpeggiations. Because two pitches are
added between tonic chord members, the span from ^5 to ^8 is presented as a
diatonic ascent: E<F♯<G♯<A (incorporating a descending registral shift at the
onset). Thus neither X3 nor Y3 is chromatic. Adding two notes between third-
related pitches (^1<^3 and ^3<^5 , also incorporating registral shifts) requires a
chromatic pitch: A<B<B♯<C♯ (where chromatic B♯ occurs at Y2) and
C♯<D<D♯<E (where chromatic D♯ occurs at Y1 and at Y4).
5. R1, R4, R5, R8, R10, R12, and R14.
6. The pitch A sounds at all fourteen locations.
7. D and F sound above A at R2, R6, R9, R11, and R13. B and D sound above A at R3
and R7. At first the upper pitches are arranged as double-neighbor figures
embellishing the tonic triad’s third and fifth (C<D>B<C and E<F>D<E). From
R9 onwards, only the upper neighbors are deployed.
8. The chords at R9, R11, and R13 result from the contrary motion described.
Above an unwavering bass A, the pitches G♯, B, D, and F sound, forming an
embellishing diminished seventh chord that resolves back to the tonic at R10,
R12, and R14.
9. E<G♯<B<E is an arpeggiation of the E Major tonic harmony.
10. Given that there is no deviation from the tonic harmony during the span from
R1 through R4, bass B at R3 corresponds to the tonic in 64 position.
11. R2.
12. Bass B at S1 is reminiscent of what occurs at R3. If that context were to prevail
again, the tonic prolongation would persist. However, B shifts from its former
role as fifth of I to an assertion as the root of V, so that instead of members of
the tonic triad sounding (at S3), V’s third and fifth (D♯ and F♯) emerge above
the prolonged bass B. In the first interpretation the 64 chord corresponds to the
second inversion of I, whereas in the second the chord serves as a “cadential 64 ”
embellishment of the dominant 53 .
13. The pitches D♯ and F♯ both sound (along with B retained from the tonic), and
both embellish members of the tonic harmony as neighbors: E>D♯<E and
G♯>F♯<G♯ (both traversed twice). Given that to this point a slow harmonic
progression prevails, conveyed using only three Roman numerals (I––––––
V––––I), it would be more logical to interpret what happens at T as a tonic
prolongation (displayed as I–––––) than as the sudden onset of frenetic
harmonic activity, as I V I V I. The supposed V chords instead are linear
embellishments, lacking assertion as individual harmonies.
234 Notes to pages 10–12

14. C♯ sounds both at the initial upbeat and during measure 1’s second eighth
note. An AF♯ dyad sounds during that measure’s third eighth note.
15. In the accompaniment an GE ♯♯ dyad sounds, while in the melody an E♯<G♯ third
is traversed.
16. During X1 the tonic root F♯ is maintained below the pitches of the dominant
harmony, serving as a pedal point. During X2, dominant root C♯ sounds in the
bass.
17. The pitches are B (minor seventh) and D (minor ninth).
18. As conveyed in the answer to question 1, a C♯ leads upwards to an AF♯ dyad
above the tonic root F♯ during R. That trajectory is spread out to become
C♯<A>F♯ during W1, conveyed as C♯ for over two measures, followed by A
(the second eighth note of W1’s third measure) and, after a passing note, F♯
(the fourth eighth note of that measure). Even R’s initial C♯>F♯ is replicated
during the last measure of W1.
19. Only the inversion changes. The tonic harmony – I – initially is presented in 53
position and shifts at S to 63 position.
20. During X2 Mendelssohn stacks the two measures of X1 (now with a swifter
succession from E♯ up to G♯) on top of the two measures that follow it during
the introduction. Thus the B and D earlier marked by the Y1 and Z labels
sound concurrently with the traversal from E♯ to G♯ during X2.
21. W2’s fourth measure does not correspond to content from W1. That measure’s
highest pitch is an F♯, and chromatic B♯ is deployed.
22. The suspension is from F♯ to E♯ over bass C♯, thus a 4–3 suspension (here with
3 raised to ♯3 to function as the leading tone within the dominant harmony).
E♯ is the second of X3’s three pitches.
23. During the antecedent phrase an A sounds as the second pitch in W1’s third
measure, followed by G♯, the second pitch of X2. During the consequent phrase
an A likewise sounds as the second pitch in W2’s third measure, followed by
G♯, the first and third pitches of X3. A half cadence and a perfect authentic
cadence, respectively, conclude the two phrases.
24. E and G.
25. The major dominant harmony (V♯) is being arpeggiated as B<D♯<F♯. The G of
the opening I and the F♯ of this V♯ project ^3>^2.
26. The AB dyad that sounds at S is an abbreviated presentation of the dominant
with seventh (B-D♯-F♯-A). The introduction proceeds only so far as consonant
V♯. The addition of dissonant A at S motivates the arrival of an E tonic chord
on the following downbeat.
27. The E<G third of Q is reprised by T1 and T2. Though with some internal
backtracking in the melody’s ascending contour, the line eventually reaches
the G an octave higher at T7.
28. In the former case, the melody proceeds from one of the tonic chord’s pitches
to a higher one, so a passing note is deployed: G<A<B. In the latter case, a
 Notes to pages 12–15 235

single pitch is being prolonged, so a double-neighbor embellishment is

deployed: B>A<B<C>B.
29. As explored in question 2, the foundational melodic initiative of the introduction
is ^3>^2. Exactly those scale degrees occur from T7 to U1. (The melody’s broad E<G
followed by F♯>D♯ correlates with most of the introduction’s melodic content,
while the harmonic support is enriched.) The bass during that beat is A, resulting
in the projection of II in 63 position. (Though ^2>^7 ♯ may be supported by a
prolonged V♯, it is very common for II V♯ to occur instead.)
30. Had F♯>E occurred before the bar line, II8–7 (in first inversion) would have
6
transpired. Here E instead participates in a cadential 4 chord on the phrase’s
final downbeat. The phrase ends in a half cadence.
31. A C♯ sounds in both the soprano and the bass. The chord is of minor quality,
and it is presented in 53 position.
32. Though both S1 and S2 present a descending arpeggiation – C♯>G♯>C♯ – some
of their sixteenth and eighth notes do not match in terms of pitch. During S1
the second and third of these faster notes initiate the next pitch in the
arpeggiation

while during S2 the equivalent notes repeat what just preceded them

33. C♯-E-G♯ sounds at the perimeters of S2, whereas G♯-B♯-D♯-F♯ against bass C♯
sounds during the middle.
34. S2 projects I in 53 position at its perimeters, whereas T projects I in 63 position.
35. For the melody’s F♯ at U, either II or IV would be a common choice for
harmonization. Here Schumann deploys IV.
36. A PAC results from the succession from V♯37 to I at the end of the phrase.
37. A dissonant seventh (B above bass C♯) occurs at S. The melody begins at R with
a consonant B preparation (supported by the tonic harmony) that persists as a
common tone over the emergence of supertonic root C♯ at S. The conventional
downward resolution of B to A♯ at T1 coordinates with the harmonic
succession from II to V♯ (both chords with seventh).
38. A root-position tonic harmony at U shifts into first inversion at W. The bass
and vocal melody coordinate to form a voice exchange: DB  DB .
39. B<C♯<D occurs in the vocal melody at R, T2, and U; D>C♯>B occurs in the
keyboard bass at R, S, and U.
40. A single harmony prevails during this half-measure: V♯ 37 . The melody traverses
one of the intervals of the dominant chord: from A♯ at T1 to C♯ at T2, filled in
by passing note B.
236 Notes to pages 17–20

2 Surges

1. The tonic triad is arpeggiated upwards in the bass during the first three beats of
measure 1 and of measure 2: E♭<G<B♭. That arpeggiation is both reversed and
filled in by accented passing notes at the melody’s onset: B♭>A♭>G>F>E♭.
Taking into account that A♭ and F are embellishing pitches, an GE♭  EG♭ voice
exchange transpires during the beats labeled R2 and R3.
2. Whereas only E♭, G, and B♭ sound at R1, the R2 through R4 beats witness the
emergence of the tonic’s dissonant minor seventh, D♭. This may be effectively
♭
conveyed in an analysis either as I8– 7 or as I ➔. (The space between the Roman
numeral and the arrow indicates that the D♭ emerges only gradually. If it were
present already at the tonic’s onset, the symbol would appear as I➔.) This
surging tonic targets the IV of measure 3.
3. In proceeding from I to IV, the peak melodic note ascends from B♭ to its
neighbor, C. In both cases a local descent via an accented passing note occurs:
B♭>A♭>G during the R1 and R2 beats; C>B♭>A♭ during the S1 and S2 beats. As
C descends to A♭, the bass ascends from A♭ to C and back, thereby reinforcing
the subdominant harmony and relating to the voice exchange that took place
during measure 2 (mentioned in the answer to question 1). When neighbor C
returns to B♭ at T2, the “obvious” choice for the melody’s continuation would
be another descending third: B♭>A♭>G. Here Mendelssohn conveys that third
in a varied form, with an F substituting for passing note A♭. (A♭, the
dominant’s seventh, sounds instead in the chordal interior.)
4. During measure 4 the pitches B♭, D, F, and A♭ project a V7 (or V➔) harmony,
whose resolution to I at U results in an IAC.
5. The tonic is presented in 53 position at S1 and in 63 position at S2, with the
melody and bass participating in a DB  BD voice exchange.
6 6
6. The pitches F♯, A♯, C♯, and E are presented in 5 position at R1 and in ♯ 4
3 2
position at R2.
7. The minor-key tonic surges toward IV when its third is raised (D to D♯) and its
minor seventh is added (A). These shifts may be conveyed via analytical
3
♯ 3 or as I ➔.
symbols either as I ♮ 8
7

8. A perfect authentic cadence transpires, resulting from the harmonic trajectory

8
7
IV V 6
5 I.
4
♯ 3
9. IV arrives at S1, and I arrives at U1 (with melodic embellishment resolving at
U2). One would expect V to occur between those points. A root-position
dominant (with seventh) occurs at T3.
10. One may describe the relationship from R1 through R3 as the tonic’s evolution
into a surge: from stable A-C♯-E to dissonant A-C♯-E-G♮. That surge targets
the subdominant, which arrives at S1. The broad harmonic analysis might
appear as follows: A Major: I8–♮7 IV, or as I ➔ IV.
 Notes to pages 20–23 237

11. The chord at R2 is spelled as E-G♯-B-D. (One might regard it as a local V7.) The
pitches G♮, F♯, and F♮ come between the initial tonic bass A and this chord’s
bass E.
12. The major second between subdominant D-F♯-A’s F♯ and T1’s E is
chromatically filled in by F♮, resulting in a shift of the subdominant
harmony’s quality to minor.
13. The D>E span is filled in as D>C♯>C♮>B>E. (Most of these notes have lower-
octave doublings.) The C♯ is joined above by E and A, the very notes that form
a cadential 64 when E resides in the bass. A viable harmonic reading of the
entire excerpt would be a four-measure prolongation of I (eventually surging)
leading to a one-measure presentation of IV, followed by two measures during
which V reigns (at first incorporating embellishing pitches C♯ and A), and
finally a measure of I. Here the conventional cadential 6
4 is presented in a 6
3

unfurling, with C♯ rather than E in the bass.

14. As mentioned above, the bass proceeds as C♯>C♮>B (corresponding to 6–5
within a cadential 6
4 ’s resolution into 5
3 ). The chromatic line E>D♯>D♮
(corresponding to the dominant’s 8–7) sounds concurrently near the top of
the texture. The third moving line, the 4–3 of a cadential 64 ’s resolution, cannot
be chromaticized, since D>C♯ is a minor second.
15. At the top of the texture, the tonic’s fifth, E, is embellished by upper neighbor
F♯. In the chordal interior, the tonic’s fifth, E, is embellished more locally by
chromatic lower neighbor D♯.
16. The tonic’s root E♭ and third G are embellished by neighboring notes: D in the
bass, and F and A♭ in the piano accompaniment and/or voice. The resulting
6
chord is in 5 position.
3
17. First B♭>G is traversed; then B♭<E♭ is traversed.
18. The chord in question is both dissonant and chromatic, a surging supertonic
targeting the dominant. It is constructed using the pitches F, A♮, and E♭, and
might be labeled as II♮37 or as II➔. Because it is in root position, the figured bass
symbols are ♮37 .
19. A B♭-D-F dominant harmony (with F omitted) occurs at U, with B♭ in the bass
and D in the vocal melody.
20. The excerpt divides into five segments as follows: R1 through R4; S2 and T; U1;
U2 through U5; and W1 through X. (Though an F♯ does not sound in the vocal
melody until W2, it is introduced by the piano at W1.)
21. The BE perfect fifth is projected melodically in its inverted state, as an ascending
perfect fourth, between R1 and R2 and is filled in melodically: B<C♯<D♯<E.
Then a descending fourth from E to B transpires by leap between R2 and R3,
resulting in the reinstatement of the initial B.
22. It is very common to concurrently embellish a triad’s third and fifth by upper
neighbors, which here may be displayed as I 5
6
5
3
4
3 . The vocal melody’s C♯ at S1
238 Notes to pages 23–26

is a neighbor to the tonic’s fifth, B, while concurrently the piano’s A is a

neighbor to the tonic’s third, G♯. Though often the embellishing chord retains
6
the prolonged harmony’s root (here E) in the bass, such a 4 chord may be
6 5
unfurled into either or position. The emergence of A in the bass during the
3 3

S1 beat results in a 53 unfurling of the embellishing chord.

23. I II7 V I. Because of the prevailing downward trajectory of the melody (from
B to G♯), some listeners likely will mentally retain the vocal pitch A during the
T beat, resulting in a V(7) harmony whose dissonant seventh A resolves to G♯
at U1.
24. The pitch classes E, G♯, and B are arranged in 53 , 63 , 64 , and 53 positions,
respectively, during the four locations indicated.
25. Within a tonicization of the dominant key, B Major, the passage from W1
8
7
through X would be interpreted as I IV V 6
5 I. The excerpt’s foundational
4
3

progression would be I followed by V (emerging at W1). Maintaining an E

Major perspective and instead interpreting the chords W1 and W2 as repre-
senting the same embellishing function, the progression from U5 through X
8
7
would be analyzed as I II 6
5 V (with the surging supertonic fully in place at
4
♯ 3

W3). Only the latter interpretation corresponds to this chapter’s surge topic,
hinting at which of the two hypotheses the author endorses.

3 IV5–6 V

1. The tonic pitches are arranged in 63 position at R1 and in 64 position at R2. The
outer-voice melodic lines work in parallel motion (compound parallel sixths):
B up to D in the vocal melody against D up to F♯ in the piano bass. In contrast,
those outer voices work in contrary motion during the projection of IV that
6 5
follows, with chords in 3 and then 3 position participating in a sixth-to-tenth
voice exchange (  GE ).
E
G

2. The passing chord deploys the pitch classes E, G, B, and C♯, for which the
appropriate figured bass is 65 .
3. The parallel motion between outer voices would have persisted if bass G>E at
T1 and T2 had been matched by E>C♯ in the vocal melody. Though the vocal
E<G third contrasts that trajectory, the leap down to C♯ after G belatedly
fulfills it. C♯ anticipates the dominant’s fifth, in the context of IV5–6.
4. Whereas in the context of R2 the pitches B, D, and F♯ represent the tonic
harmony in its second inversion, the pitch F♯ is asserted as the dominant’s root
at U1, at a point nearing the phrase’s final cadence. Evidence of F♯’s assertion
comes from the fact that the chord’s B resolves to A♯ and its D resolves to C♯,
 Notes to pages 27–29 239

resulting in the full sounding of the dominant harmony at U2. (Concurrently

the dominant’s seventh – E – emerges, resulting in V ♯ 37 .) The resolution to I
produces a PAC.
5. Initially the G Minor tonic chord is prolonged through melodic traversals of
chordal intervals: D>B♭ during R1 versus D<G during of R2. Though the
following measure begins as if an exact repetition were being pursued, the
tonic chord surges (targeting IV) at R4 through the shift of chordal third B♭ to
B♮ and the addition of chordal seventh F.
6. Tonic root G’s lower neighbor F♯ sounds, its third B♭’s upper neighbor C
sounds, and its fifth D’s upper neighbor E♭ sounds.
7. The surging tonic chord at R4 harbors a diminished fifth ( BF ♮ ) that yearns for
resolution to the subdominant’s root and third ( CE ♭ ). That resolution indeed
transpires between R4 and T1 (covered by pitch G). This may be conveyed
♮ 3 IV or as I ➔ IV.
8––– 7
analytically either as I♭ 3

8. Just as the initial G-B♭-D chord surges as G-B♮-D-F, this C-E♭-G chord could
surge as C-E♮-G-B♭, leading to F-A♭-C, which could surge as F-A♮-C-E♭,
leading onwards to B♭-D-F, tonic G’s mediant. That is a common chordal
trajectory in a minor key (a topic explored in chapter 8). Here Mendelssohn
instead seeks to prevent such a trajectory. He deploys a common device that
impedes further surge-formation: the 5–6 shift. Asserted as IV, C-E♭-G
is prolonged via the melody’s descending arpeggiation during the span
from T1 through T2: downwards in thirds from G through E♭ and C to A,
the 6 of IV5–6. This A serves as an intermediary between IV’s C-E♭-G and
V♯’s D-F♯-A.
9. The perfect fourth DG at U1 generally would resolve as 4–♯3, projecting a
dominant harmony. Yet here a downward cascade of thirds transpires from
IV at T1 through its 6 phase at T2 to V♯ 37 at U2: G E♭>D<E♭ C>B♭<C A>G<A F♯,
leading to some uncertainty regarding whether the G in question should be
interpreted as a lower neighbor to A or as a passing note to F♯. Either way, V♯ 73
is attained at U2, leading to I at W for a PAC.
10. F, A, C, and E♭.
11. No.
12. A 5–6 shift transpires during the prolongation of IV. (A passing chord, to be
explored in question 4, comes between 5-phase E♭-G-B♭ at S1 and 6-phase E♭-
G-C at S2.) The 6-phase chord is unfurled from its foundational 63 state into 5
3

position, with the pitch C sounding in the bass at S2.

13. Bass E♭>C is filled in diatonically (D) and the soprano B♭<C chromatically
(B♮). The pitches F and G sound along with D and B♮ at T, resulting in a
surging embellishing chord that targets the subdominant’s unfurled 6-phase
chord, C-E♭-G.
14. After I IV5–6, one would expect to proceed either to V, or through V to I. Here
V7 followed by I fulfills our expectations for the phrase’s conclusion.
240 Notes to pages 30–34

15. Before the tonic root C arrives in the bass, only the three pitches G, C, and E (in
ascending order) sound at Q. The major third formed by C and E is traversed
in the vocal melody’s opening E>D>C melodic gesture.
16. The initial tonic proceeds to IV at S1. Just as I is projected by means of the
melodic third E>C (= chordal third to chordal root) in the vocal melody at R,
this IV is projected by means of the melodic third A>F (= chordal third to
chordal root) in the piano bass between S1 and S2. Consequently the
subdominant shifts from 63 to 53 position over the course of its presentation.
17. In the succession from IV (F-A-C) to V (G-B-D), it is common for IV’s
6-phase pitch D to emerge, either replacing 5-phase C or as an addition to
the chord. Here Schumann pursues the latter option (F-A-C-D). Concurrently
he allows A to mutate to A♭ (diatonic in C Minor), prominently projected by
the vocal melody’s leap from F up to A♭ between S2 and S3. (That F<A♭ third is
a contorted retrograde of the A>F third heard in the piano bass between S1
and S2.) Then, when the harmony shifts to V7, this A♭ persists for some time as
a suspension, with the ♭9–8 resolution transpiring only at the end of the
measure (T2).

4 I5–6 II

1. The span from R1 to R2 projects the progression’s initial tonic harmony.

Whereas root B serves as the bass for the chords at both R1 and R2, the
melodic trajectory ascends from B to D♯. The bass prolongation of B is
facilitated by neighboring note A♯ at Q2, whereas the melodic ascent from B
to D♯ is facilitated by passing note C♯ at Q2. (These trajectories reach their
goals only on the second try: measures 15|16–173 repeat in slightly varied form
the material from 13|14–153.) The chord at Q2 emerges via a voice exchange
from Q1 ( AC ♯♯  CA ♯♯ ). (The pitches of the embellishing chord sound first in 43
position and then in 65 position.)
2. A descending major third it traversed over the course of R2 through R3:
D♯>C♯>B. This span is the retrograde of the B<C♯<D♯ ascent discussed in
question 1.
3. I5–6 transpires during R2 through R3. The F♯<G♯ step corresponding to the 5–6
shift (below Middle C in the piano part) is filled in chromatically, as F♯<FÜ<G♯.
4. Following the robust arpeggiation employing the pitch classes G♯, B, and D♯,
an F♯ sounds in the melody just before the 6-phase tonic yields to its successor.
This F♯ forms a dissonant minor seventh against the earlier G♯.
5. The previous 6-phase tonic yields to II7 at S. The phrase concludes with
V♮ 79 at T.
6. The passage projects A Major’s tonic harmony, which sounds unadorned at R1,
R2, and R3, and whose impact persists through the EA pedal fifth maintained
 Notes to pages 34–36 241

during the three internal Q locations. Both lower and upper neighbors to the
tonic chord’s root and third participate in the formation of embellishing
chords. At Q1 upper neighbors B and D sound, and at Q2 first lower
neighbors G♯ and B and then upper neighbors B and D sound. (The lower
and upper neighbors are connected by passing notes.) Though the upper
neighbors B and D extend from Q2 to Q3, the latter context makes those
pitches come across as suspensions, which resolve (as 9–8 and 4–3) at R3.
7. Tonic A-C♯-E shifts to A-C♯-F♯ during R4. This shift may be conveyed
analytically as I5–6.
8. The phrase follows the familiar harmonic trajectory from I (at R1) to V (at T).
Because I undergoes a 5–6 shift, the root succession A<B<E is likely. In this
case the supertonic chord’s diatonic state is omitted: it is already surging (II➔)
when introduced at S. This surge results from the concurrent sounding of the
diatonic chordal seventh (A) and the chromatic raised third (D♯), with
dissonant AD ♯ resolving to the dominant’s GE ♯ (at T).
9. The tonic 6-phase pitch F♯ that emerges during R4 is a crucial voice-leading
event. Its impact extends to the F♯ that sounds near the end of the measure
labeled S, before its voice-leading descent to E at location T. The minor third
from F♯ down to D♯ is filled in, so that the E at R5 comes across as a passing
note – and not as a reinstatement of the E from the earlier tonic 5
3 chord.
Concurrently C♯ shifts to C♮, so that C♮>B rather than C♯>B transpires during
the succession from the tonic’s 6 phase to the supertonic (at location S).
Consequently the chord at R5 is an incomplete statement of A-C♮-E-(F♯),
both chromatic and dissonant in quest of its successor, B-D♯-F♯-A.
10. A voice exchange occurs between R3 and R4. The intervening connective chord
6
at Q2 is built from the pitch classes A, C♯, E, and G, arranged in 5 position.
3

11. The three unfoldings are A>F♯, G>C♯, and D<F♯. The dissonant diminished
fifth of the second unfolding is resolved by the major third of the third.
12. From F♯ at R4 an A and a D lead to an F♯ an octave higher. The D sounding at R5
coincides with the emergence of the tonic’s unfurled 6-phase chord (B-D-F♯).
Consequently the C♯-E-A chord at Q3 should not be interpreted as an asserted
dominant harmony (in 63 position). It serves instead as a connective chord
between I5 and I6. Though the melody’s continuing arpeggiation up to F♯
might well have arisen during I6, Mendelssohn elected to introduce II just as
that goal pitch is attained, thereby denying it the anticipated consonant chordal
support. Consequently, F♯ descends immediately to E.
13. At S: F♯>E>E>B. (The F♯ embellishment of II’s E might be interpreted either as
a suspension with an imagined preparation or as an incomplete upper
neighbor. E and B are chord members of II.) At T: D>C♯>B–B>A. (The
initial D actually “belongs” in the preceding measure, where it would have
served as the II harmony’s seventh. Its delay until the following downbeat,
where it clashes with the onset of V7, converts it from an unaccented into an
242 Notes to pages 38–39

accented passing note. The following C♯ and A are chord members of V7, while
B connects those pitches as a passing note.)
14. Four thirds are involved in projecting these two strands: A♭<C, D♭>B♭, C>A♭,
and B♭>G.
15. While the melody’s initial A♭<C third is being traversed from R1 to U2, the
tonic evolves from consonant A♭-C-E♭ to dissonant A♭-C-E♭-G♭ (thus I ➔).
That surge leads to IV at W1. The unfolding of the melody’s D♭>B♭ third
results in the immediate shift to IV’s 6 phase (thus IV5–6). The dominant
follows at X. Though only pitch classes E♭ and G participate in its projection,
no successors to the recently stated D♭>B♭ third have sounded, and so listeners
should mentally merge them with the E♭ and G (which might be conveyed
analytically using parentheses as V(7)). Finally, the tonic is restored during R2
(thus I). Note that at that goal point the melody’s A♭<C and D♭>B♭ thirds have
been followed only by the upper-strand restoration of C. The full flowering of
the C>A♭ third transpires during the passage from R2 to R3, coordinating with
the onset of further chordal activity that will lead the broader harmonic
progression from I to the phrase’s goal at T2. (See question 4.)
16. Because the chords at R1 and U1 are almost identical, the connective chord at
Q1 features neighboring notes: G and D♭ embellish the tonic’s A♭ and C,
respectively, while E♭ is maintained as a common tone. In contrast, the
chord at U2 is situated higher than that at U1, and so the connective chord at
Q2 features passing notes. Three concurrent ascending stepwise motions
occur: that between root and third (A♭<B♭<C), that between third and fifth
(C<D♭<E♭), and that between fifth and seventh (E♭<F<G♭).
17. Once the initial tonic prolongation has concluded, Mendelssohn undertakes a
5–6 shift of the tonic between R2 and R3. (The 6-phase chord of I5–6 is here
unfurled, with pitch F in the bass. That F is in fact a sixth above the low bass A♭
that has persisted for nearly three measures.) The harmony that gradually
emerges at location S is II, followed by V at T2.
18. The connective chord at location Q3 engaged pitch classes E♭, G, B♭, and D♭.
At T1 a cadential 64 chord sounds. We expect the A♭ and C above dominant
root E♭ to resolve downwards by step to dominant chord members G and B♭,
respectively (as in fact occurs at T2). In this particular instance the dominant
harmony is further embellished during the T1 beat, still targeting a 53
resolution at T2. The C shifts to C♭, and the pitches D♮ and F are added, so
that ultimately a D♮-F-A♭-C♭ embellishing chord emerges. Its dissonances
target all three members of the upcoming dominant harmony, all above
dominant root E♭.
19. Between R1 and R2, the bass ascends from A to C♯, so a tonic chord in 53
position is followed by a tonic chord in 63 position. Between R1 and R3, the
soprano ascends from E to A. Because the bass is not affected, both chords are
5
in 3 position.
 Notes to pages 39–41 243

20. A, G♯, F♯, E, D♯.

21. The tonic is expanded via a 5–6 shift: A-C♯-E to A-C♯-F♯. The latter chord is
unfurled into 53 position at S1.
22. The tonic’s 5–6 shift is here traversed as E<E♯<F♯ in the tenor register. The E♯
at location X is joined by C♯ and G♯ (clashing with tonic pedal point A) as a
connective chord (C♯➔) targeting the tonic’s 6-phase F♯-A-C♯ chord.
23. 42 .
24. I6 typically would proceed to II (B-D-F♯). The chord at T, rooted on B, has
evolved into a surge, with a shift of its third from D to D♯ and the addition of
minor seventh A. The chord’s analytical symbol would be II 73 ♯ or II➔.
25. 65 .
26. During T the pitch D♯ serves as II➔’s raised third, contributing to the surge
effect because D♯ and the A above it form a dissonant diminished fifth that
should resolve to V’s E and G♯, respectively. Though A descends cleanly to G♯
at the point marked U1, the D♯’s ascent to E is less decisive. Though the
expected E sounds twice during the beat labeled U1, the preceding D♯
persists into that space, embellishing the E. This E serves as V’s root. In
preparation for the tonic chord that will begin the next phrase, the
dominant’s seventh, D♮, emerges in the bass at U2. The analytical symbol for
the measure would be V8–7, with the 7 phase conveyed by a chord in 42 position.
(The B that follows after D in the bass prepares for the root-position tonic that
will sound at the following downbeat.)
27. The harmonic progression I5–6 II V transpires. The foundational melodic
succession G♯>F♯ (= ^3>^2) here expands into G♯>G♯<A>F♯. (At location W,
incomplete upper neighbor A is embellished by its own upper neighbor, B.)
28. The introduction’s pitches form an interlocking augmented fourth and
F♯ G♯
diminished fifth – C♮ (which would resolve to B ) and A
D♯ (which would
G♯
resolve to E ). The B, E, and G♯ that resolve those dissonances all sound, in
the appropriate register, during the melody of R1 through R3, in the ascending
order B<E<G♯.
29. The chord at R4 matches that at R1. The chord at S serves as a 64 embellishment
of the tonic, resolving to 53 at R4. (That is, E persists as the tonic root during S,
while embellishing pitches A and C♯ sound. The A is not asserted as a
subdominant root.) The outer voices of the chords at R3 and R5 relate as a
voice exchange ( GE ♯  EG ♯ ), thus with the tonic chord sounding in its first
inversion at R5.
30. The two chords are spelled as B♯-D♯-F♯-G♯ and C♯-E-G♯. Observe how the
former chord contains a dissonant diminished fifth ( FB ♯♯ ) that resolves into (as if
“targeting”) the latter chord’s minor third ( EC ♯ ).
31. A parallel progression of 63 chords is traversed over the span from R5 through
T5, as follows:
244 Notes to pages 41–46

E D♯ C♯ B A G♯
B A G♯ F♯ E D♯
G♯ F♯ E D♯ C♯ B♯
During the final chord, Mendelssohn adds an F♯, which joins with the chro-
matic bass B♯ (substituting for diatonic B) to create a diminished fifth whose
E
inward resolutional tendency (to C♯ ) prevents a further continuation of the
parallel progression.
32. The dominant B-D♯-F♯ is secured at X1, resulting in a half cadence. Later
minor seventh A (at X2) and major ninth C♯ (at X3) emerge, helping to
destabilize the dominant prior to the onset of the next phrase.

5 Surging 6-phase chords

1. The bass notes B♭, C, F, and B♭ suggest a foundational I II V I harmonic

progression.
2. The II harmony, which is introduced in its root position at S1, is followed by its
first inversion (with bass E♭) at S2. Whereas the melody’s F♯ serves as a lower
neighbor to II’s fifth (G) during S1, F♮ (in the upper register) serves as an upper
neighbor to its third (E♭) during S2.
3. The dominant is embellished by a cadential 64 chord, with fourth B♭ proceeding
to third A sooner than sixth D descends to fifth C. In addition, the dominant’s
seventh E♭ emerges.
4. Had only the soprano F<G motion transpired between R1 and R2, the passage
would demonstrate how a 5–6 shift may facilitate the harmonic succession
from I to II, with 6-phase G anticipating the supertonic’s fifth. Yet
Mendelssohn provides a more assertive thrust toward that supertonic in two
ways. First, the 6-phase chord is unfurled, with G doubled in the bass (thereby
inducing a potent G<C succession from R2 to S1). Second, not only is G added
to the B♭-D-F triadic pitches, but the B♭ also shifts chromatically to B♮. The
resulting G-B♮-D-F chord emulates a dominant seventh and thus may be
analyzed aptly as VI➔. The G is transformed from a mere anticipation of
the supertonic’s fifth into an asserted chordal root.
5. With the harmonic succession V7 I allied to melodic ^2> ^1, a PAC occurs.
6. The tonic’s initial presentation involves an upward trajectory that proceeds
from root position at R1 to first inversion at R2. The outer voices ascend in
parallel tenths (soprano C<D♭<E♭ against bass A♭<B♭<C). (An upper
neighbor, F, embellishes the apex E♭.) Consequently the chord at Q plays a
connective role. The leap from C to G♭ in the bass at R3 adds the chordal
seventh that converts a diatonic major triad into a surge. Consequently a
suitable harmonic analysis for the passage from R1 through R3 would be I ➔.
 Notes to pages 46–47 245

7. Any surging chord should fulfill the same voice-leading conventions as does the
V7 that it emulates. Consequently listeners should expect that the surging I➔ in
4
2 position at R3 will be followed by IV in 63 position. For that to be the case, it is
necessary to interpret the first melodic pitch at S1 as a suspension (E♭ resolving
to D♭), so that the harmony projected is F-A♭-D♭ (= inverted IV) rather than
F-A♭-E♭ (= VI7). The melodic E♭>D♭ corresponds to the F>E♭ from R2.
8. The melody proceeds as E♭>D♭>C–C–C>B♭. The underlined D♭ (discussed in
question 2) and B♭ are the principal pitches, which in this context correspond
to IV5–6. In this instance the 6 is attained not as a step upwards from 5 but
instead as a filled-in third downwards from 8. (Because the IV chord is
inverted, this line proceeds as a sixth to a fourth above bass F.)
9. In a diatonic context, IV’s 6-phase chord may proceed directly to V: D♭-F-
(A♭)-B♭ to E♭-G-B♭. Thus the chord at S3 might have been prolonged for two
beats. Mendelssohn elected to modify the 6-phase chord to generate a surge,
accomplished through the replacement of D♭ with chromatic D♮ (thus B♭-D♮-
F-A♭, in which B♭, though unsounded, is understood to persist from the
preceding beat). Consequently IV’s 6-phase chord has evolved into II➔ at S4.
10. V7.
11. Pitch classes F♯, A♯, C♯, and E sound against the tonic root and fifth pedal
points during Q. These pitches form a connective chord between two
presentations of the tonic. Over the course of the piano melody, the
ascending major third from B to D♯ is traversed. Thus Q’s melody pitch C♯
serves broadly as a passing note.
12. Melodic pitch C♯ at S1 is supported by II, which evolves into II➔ at S2. The
C♯>A♯ third traversed from S1 through T complements the B<D♯ third that
transpires from R1 through R2.
13. Whereas only the pitch classes B, D♯, and F♯ sound at R2, an A is added to the
mix at R3, resulting in a I➔ surge. That surge’s tendency is fulfilled by the
arrival of IV at U1.
14. When IV serves as the principal intermediary between I and V, it often
undergoes a 5–6 shift. The sounding of 6-phase C♯ in the melody at U2 helps
guide IV along the path to V. Concurrently chord member E shifts to E♯, so a
C♯<F♯ root succession becomes pronounced. Thus IV5–6 V is helped along in
its trajectory as IV’s 6-phase chord evolves into II➔.
15. It is very common, in the vicinity of a cadence, for ^2 (supported by II or by IV6)
to serve as the starting point for a descending third to ^7 (supported by V) –
here filled in as C♯>B>A♯. The intervening passing note (the pitch B) often
6
sounds against the dominant’s root (F♯), participating in a cadential 4 chord.
The text Schumann is setting at this point conveys a rushing downwards to a
loved one’s feet. Rather amusingly, he stretches the C♯>B>A♯ third into a tenth
by lowering the A♯ by an octave (sounding at the end of the beat labeled as
W2), thereby integrating the low point of the human body with a low-sounding
246 Notes to pages 49–54

pitch. Before the melody comes to rest, the C♯ of U2 is restated in the lower
register as well (at X), before goal B sounds.
16. The phrase’s initial tonic surges from its onset, as I♯ 37 or I➔. Since the
subdominant that it targets sounds as a major chord (E-G♯-B), it should be
labeled as IV♯.
17. The E<E♯<F♯ bass from S1 through Q2 is exactly what one would encounter in
5
6
a IV♯ 3


V♯ harmonic succession in which the subdominant’s 6-phase chord is

asserted as II➔. The B that sounds above bass F♯ at Q2 would be expected to

behave as a suspension (4–♯3), and the G that sounds in the bass at S2 would
serve as an upper neighbor between the F♯s at Q2 and at T1.
18. Alto G♯>F♯>E coordinates with bass E<E♯<F♯<G♮, resulting in the projection
of a chromaticized voice exchange from S1 through S2: GE ♯  EG ♮ . As noted in
the answer to question 1, G♯ is a chromatic pitch in B Minor, and so its
eventual lowering to G♮ comes across as a restoration of the diatonic state.
The IV♯ with raised third (E-G♯-B) at S1 undergoes both a diatonic restoration
and a 5–6 shift with the arrival of 6-phase E-G♮-(B)-(C♯) at S2. In this reading
the chord at Q1 is not an asserted II➔, but instead serves as a connective chord
wherein the goal S2 chord’s B and C♯ are in place while G♯ has yet to descend to
G♮, all above local chromatic passing note E♯.
19. Above bass F♯, the D that sounds in the melody at T1 serves as the 6 of a
cadential 64 chord, in which context we might expect a descent to C♯ at T2. The
high G that sounds instead (with the expected C♯ sounding only in the tenor
register) serves as the dominant’s ninth. Unlike the dominant’s seventh, a
ninth may resolve within the chord. After G is transferred to the alto register,
G>F♯ transpires at the end of the T2 measure, just prior to the resolution to I
and in conjunction with an anticipation of the leading tone’s resolution to the
tonic pitch (A♯<B) in the soprano.

6 Surges with ninths

1. I IV V 6
5
4
♯ 3 I.

2. The G that sounds at R2 shatters the cadential stability of the initiating A chord
by forming a compound minor seventh against the bass. The B♭ that sounds at
R3 likewise dissonates against the bass, as a compound minor ninth. It forms a
minor third with the melody’s earlier G. The C♯ at R4 is consonant against the
bass (a compound major third) but forms a dissonant melodic interval with the
earlier G (augmented fourth) and B♭ (augmented second).
3. I ➔ IV V♯ 37 I.
4. The tonic arrival occurs at R1. The preceding dominant is presented in 65
position.
 Notes to pages 54–56 247

5. The dominant is presented as G♯-B-D-E at Q1 and Q3. During the intervening

Q2, the outer voices ascend to the adjacent diatonic pitches and then revert to
their starting points: G♯<A>G♯ in the bass, and E<F♯>E in the soprano. The
resulting A-B-D-F♯ chord plays no independent harmonic role. Instead, it
results from the persistence of the dominant’s fifth and seventh (B and D)
against upper neighbors of the root and third (A and F♯).
6. The chord’s arrival is embellished by a 4–3 suspension (A to G♯), whereas its
tendency to resolve back to A is strengthened by the addition of the chordal
seventh (D) at S2.
8
7
7. The progression concludes as II V 6
5 I. A shift from F♯ to F♮ transpires during
4
3
the prolongation of II at T1-2. In part the F♮ gives the passage a somber hue. Yet
note also that this shift is part of a broad alto-register descending trajectory
that is preceded by A>G♮ and continues with the dominant’s E>D and the
tonic’s C♯.
8. The tonic at R2 (A-C♯-E) often would absorb a sixth (F♯) in a trajectory that
proceeds to the supertonic (B-D-F♯ at T1). The F♯-A-C♯-E chord may be
asserted as VI➔ through the shift of A to A♯ (resulting in F♯-A♯-C♯-E). One
may increase the surge’s potency by adding a chordal ninth, G♮ or G♯. Often
the root will be omitted when the ninth is added. At R3 this chord sounds as
A♯-C♯-E-G♮ (against an A pedal point in the bass). Though this chord is a
highly evolved representative of the tonic’s 6 phase, note that the 6-phase pitch
F♯ does not actually sound within the chord, but instead is understood as its
unsounded root.
9. The melody’s initiating A<D fourth ultimately is complemented by the filled-in
descending fourth D>C♯>B>A, whose realization is twice prevented by the
chord at Q. This descending fourth is supported by chords that correspond to
a very common figured-bass pattern, 8
7
6
53


4
3 , whose second chord is repre-
sented here by F♯-A-C♯ (with bass D absent) and third chord by G-B (likewise
with bass D absent).
10. The prolonged tonic root D is followed by bass pitches G<G♯<A. The diatonic
pitch G supports II (in first inversion), and the diatonic pitch A supports V.
11. The chord at S1 is derived from the supertonic triad, E-G-B. To make that
supertonic surge, chromatic G♯ must replace G. The seventh also typically will
emerge – here the pitch D. In this case another dissonance sounds as well at S2:
in addition to 8–7, a 10–9 descending motion transpires, so that the supertonic
chord is represented by its raised third (G♯), seventh (D), and ninth (F♯).
Because the 8 is not restored elsewhere in the texture, the resulting chord at S2
lacks its root, a very common occurrence when the chordal ninth is present.
The simplest representation of the harmonic function for the chords at S1
and S2 would be II ➔ (with some space between II and the arrow to indicate
that the surge only gradually emerges). This notation lacks precision, however.
One always may append Arabic numerals and accidentals to the right of a
248 Notes to pages 58–59

Roman numeral to indicate the complete chordal content. In this case the
10

9
analysis would appear as II ♮ 8

7
3
♯ 3
(with • designating an absent root).
•

12. The R chords project the E♭-G-B♭ tonic in tonicized E♭ Major (the mediant of
the movement’s C Minor tonic). The Q chords (A♭-C-E♭) serve as neighboring
embellishment of the E♭ tonic (G<A♭>G and B♭<C>B♭ against prolonged E♭).
Though these neighboring chords sound with A♭ in the bass, it is as if the E♭
root were prolonged through them, in the manner of I 5
6
5
3
4
3 .

13. Whereas Q1 and Q2 embellish the E♭ tonic chord (R1 and R2), the S chord
embellishes the Q embellishing chord. Against the held pitch E♭, Q’s A♭ is
embellished by both lower and upper neighbors (A♭>G<A♭ and A♭<B♭>A♭)
while its C is embellished by an upper neighbor (C<D♭>C).
14. The four pitches that sound during R2 and R3 are B♭>G>F>E♮. Though the B♭
and G sound in the context of the E♭ tonic harmony, during R3’s F and E♮ the
chord shifts (to be explored in question 4), with the initially expected traversal
of B♭>E♭ replaced by B♭>E♮, a diminished fifth. That dissonant interval, here
unfolded in the downward direction, is resolved by the upward unfolding of a
minor third (F<G<A♭) during T1.
15. The passage proceeds from I (E♭-G-B♭) to II (F-A♭-C). Often a 5–6 shift (I5–6)
helps direct the trajectory toward the supertonic. In this case the 6-phase tonic
chord (E♭-G-C) has evolved into an asserted VI➔ chord: E♭-G-C not just to
surging C-E♮-G-B♭ but also to a more dissonant state, E♮-G-B♭-D♭. The 6-
phase pitch (C) is absent.
16. The II harmony’s seventh (E♭) emerges at T2. Though in this case II does not
surge toward V, the seventh is a dissonance that calls for a downward
resolution, as will occur during the dominant harmony that follows. The
analysis might appear as II8–7.
17. Though the chord at R3 restores the tonic harmony after intervening chords,
that tonic now is surging (I➔), targeting the IV harmony that emerges at S2.
18. The chord at S1 is II♯ (or II➔): F♯-A♯-C♯. Though an E concurrently sounds in
the bass, its retention throughout the excerpt argues against incorporating it
into the II♯ harmony (as its seventh), since there is no downward resolution to
D♯. Instead, a consonant II♯ clashes against an E pedal point. Whether or not
the supertonic surges, it is commonly attained via a 5–6 shift of the tonic. In its
diatonic state the tonic 6-phase chord would be spelled as E-G♯-C♯. With C♯
asserted as a root and surging (thus VI➔), that chord might be transformed
into C♯-E♯-G♯-B. Mendelssohn incorporates an even more potent variant, E♯-
G♯-B-D♮ (all above a tonic E pedal point) at R2.
19. Both II➔ and IV generally will proceed to V within a harmonic progression.
Here the chords at T1 and T2 project that dominant with its minor seventh: B-
D♯-F♯-A (against a tonic E pedal). Though we would expect a resolution to the
tonic in both cases, the I at R3 is surging (as I➔), and thus is not suitable for a
cadence. Only at U does V7 I result in a PAC.
 Notes to pages 62–66 249

7 Colorful variants of II

1. G♯, B♮, D, and E.

2. Soprano A at T1 conflicts with the dominant’s pitch classes. It resolves to G♯
at T2.
3. Whereas three of the dominant seventh’s pitch classes sound at T1 (E, B, and
D), with G♯ delayed by the sounding of A, at T3 only the dominant’s root
sounds (E, with doubling), while A and C delay the arrival (at T4) of the
dominant’s G♯ and B, respectively. The chord at T3 often is referred to as a
cadential 64 chord.
4. The tonic chord’s fifth, E, is embellished by upper and lower half-step
neighbors in the melody during R2: F>D♯<E. The asserted chord remains A-
C-E (= I). At U the embellishing pitches are re-deployed within a different
harmonic context, so that now F and D♯ are elevated to the rank of chord
members. That chord, spelled as D♯-F-A-C, is a supersurging supertonic (II⇨)
whose root, B, is absent.
5. Though the direct succession from I to II⇨ (as occurs earlier in the song) is
always a possibility within harmonic syntax, often that succession is expanded
to become I5–6 II⇨. Bass G at S1 represents an unfurling of the tonic 6-phase
chord within that trajectory. In this case the chord is transformed from
diatonic G-B♭-D to G-B♮-D, a magical moment because such a juxtaposition
of major chords is not characteristic of the key. (The B♮ is a wobbly note, a
temporary diversion from the diatonic norm: B♭ wobbles to B♮, and eventually
B♭ will be restored.)
6. The chord embellishing G-B♮-D at T is spelled using the pitch classes D, F♯, A,
and C (against a G pedal point in the bass). What follows at S2 not only restores
the G chord of S1 but also transforms it to its diatonic state (within B♭ Major):
G-B♭-D. (Thus the wobbly note B♮ has reverted to B♭.)
7. The phrase ends on V. Given the I5–6 opening (through S2), we might expect
some sort of supertonic chord at U1. The sounding pitch classes are E♮, G♭, B♭,
and D♭, which form a supersurging supertonic (II⇨).
8. The sounding pitches at U1 may be interpreted as the supertonic’s raised third
(E♮), lowered fifth (G♭), seventh (B♭), and lowered ninth (D♭) in B♭ Major
(noting that two of the pitches, G♭ and D♭, are borrowed from the parallel
minor key). At that point, root C is omitted. Yet the D♭, which might persist
until the chord resolves to its successor (V), here instead resolves within the
chord, so that ultimately – at U2 – the chordal root C is present in a
supersurging supertonic in 43 position.
9. The tonic harmony in C Minor is projected via an CE ♭  EC ♭ voice exchange
between R1 and R2. At R3 that tonic surges (targeting IV) through the shift of
E♭ to E♮ and the addition of B♭. (This chord is presented in 42 position, so that
the IV that follows is introduced in 63 position, as was the initial tonic.)
250 Notes to pages 66–68

10. Though an FA ♭  AF ♭ voice exchange transpires in the outer voices, thereby

providing a context for the prolongation of IV matching that of the earlier I,
concurrently an interior strand descends from F to D♭. Consequently, whereas
the initial tonic is transformed via a surge (appropriate when IV is imminent),
that IV is transformed via a 5–6 shift (appropriate when V♮ is imminent).
Chromatically lowered from diatonic D♮, the pitch D♭ serves as the 6 phase of
♭
the 5–6 shift. In this context the symbol IV5– 6 aptly conveys the situation.
(The pitch D♭ is a wobbly note, a temporary shift from diatonic D♮, which will
be restored during V♮.)
11. Both T chords embellish the related chords that surround them. Because
both R1 and R2 project the C-minor tonic harmony, at T1 C’s leading tone,
B♮, sounds (along with G and D). If the subdominant were to be similarly
projected by S1 and S2, then the chord at T2 might have been T1’s exact
transposition: G-C-E♮. But that is not the case. The subdominant instead
shifts to its 6 phase (with chromatic D♭ replacing D) at S2, as explored in
question 2. Consequently E♭ sounds instead of E♮ within the embellishing
chord. Melodically, this E♭ is the interior component of a descending
F>E♭>D♭ third that coordinates with the FA ♭  AF ♭ voice exchange that sur-
rounds it.
12. The cadential 64 chord at U1 generally will be followed by a ♮ 35 resolution (often
with the dominant’s seventh emerging also at that point). That resolution here
transpires at U4. The chord at U2 serves as an embellishing chord, incorporat-
ing the upper neighbors of all three of the cadential 64 ’s members (F, A♭, and D
above dominant root G).
13. A PAC appears to be imminent, with U4’s V♮ 37 poised to resolve to I with a C
residing at both edges of the texture. (Though Mendelssohn does provide a
tonic chord with a C in the soprano, his bass note on the following downbeat is
G, not C.)
14. The B♭>A>G third during R1 projects a third from the tonic harmony, which
transpires at that point. The A♭>G>F♯ third from S2 through U2 harbors a
wobbly note – A♭ – that reverts to A♮ at U2 for the projection of V♯.
15. In a diatonic context, A-C-E♭ (= II) often is deployed between I and V♯ in a
minor-key context. In a common deployment of chromaticism, that chord’s A
may be lowered to A♭ (as a wobbly note), thereby changing the supertonic’s
quality from diminished to major. This chord often will be represented by the
Roman numeral ♭II in analyses.
16. The tonic harmony’s G and B♭ are embellished by neighbors F♯, A, and C,
which sound along with a prolonged D to create the embellishing chord at
F♯
T. At the conclusion of T, the outer voices project a C compound augmented
fourth. Whereas the bass descends as expected to B♭ at R2, the F♯’s resolution
to G is accomplished within the following chord’s interior, since the melody
instead leaps up to D.
 Notes to pages 68–72 251

17. Though the bass descent from root D at U2 to C at U3 is a conventional means

♯4
of dominant intensification (added minor seventh, in 2 position), concur-
rently the soprano F♯ of U2 is embellished by a lower neighbor, E♮.
♯4
Consequently the full sounding of the dominant 2 occurs only during the
beat’s final eighth note.
18. The B♭>A>G third of R1 is presented in retrograde, as G<A<B♭, during R3–S3,
maintaining the same rhythm. Whereas the former is supported by the tonic
harmony throughout, the latter – despite the GB♭  BG♭ voice exchange – pro-
jects both the tonic harmony and its syntactic successor, since inner voices C♯
and E♮ emerge against the concluding G and B♭, resulting in II➔. (Due to the
impact of the voice exchange and the maintenance of G and B♭ in the outer
voices, what occurs at the end of the S3 beat may be interpreted as a collision –
the simultaneous sounding of I and II➔.)
19. A half cadence transpires at location W. Though both of the outer voices sound
a momentum-adverse dominant root during the second beat of W, C emerges
in an inner voice, instilling some forward propulsion that targets the tonic
harmony that will inaugurate the next phrase.

8 III along the path from I to V

1. The circle proceeds from E through A and D to G (at location U1). This
corresponds to I ( ) III in E Minor. Though that circle could lead onward to C,
the C chord at location X comes across instead as an internal harmony within
the tonicization of the preceding G arrival point.
2. A straightforward surge process transpires at R1–2. The tonic’s diatonic E-G-B
chord at R1 is adjusted in two ways at R2: its third, G, is raised to G♯, and its
seventh, D, is added. Though the situation at S1–2 is similar, the intensity of the
surge is greater: instead of A-C-E to A-C♯-E-G, Mendelssohn deploys C♯-E-G-
B♭. The third surge, at T, is equivalent to that at R2: D-F♯-A-C. Its context
differs from the earlier instances in that the surge lacks a consonant, diatonic
predecessor.
3. The chord is introduced as F♯-A-D at Y1. By the measure’s end (at Y2), root D
has lodged in the bass and seventh C has been added. Thus the chord shifts
7
6
from 3 to 5 position.
3
8–7
4. I IV V I.
5. The upper lines at R1 and R2 begin with a CE third, followed by an AC third, thus
introducing all three pitches of A Minor’s tonic triad. That initiative is
repeated, up a step, as neighboring embellishment during S1. That
F
embellishment’s foundational idea should be understood as D followed by DB ,
252 Notes to pages 72–78

G
though other local embellishing pitches sound as well: E upper neighbors
F E D
(appoggiaturas) to , and
D C passing notes leading down to . B

6. The chromatic pitches D♯ and F♯ join with A in forming a version of II➔

(wherein root B is absent).
7. R1 and R3 are equivalent locations within their respective phrases, and indeed
the melodic onset – E>C – is the same in both. The harmonic support is
contrasting, however. Whereas I prevails during R1, that phrase’s V♯ goal
(attained at U1) persists through R3, leading to the unusual situation of a
dominant E-G♯-B-D harmony colliding with the C of an E>C>A tonic
arpeggiation in the melody.
8. The progression proceeds from the tonic to the mediant (which shares two
pitch classes with the tonic) between R4 and R5. The progression is led toward
that mediant through a surging embellishing chord – G-B-D-F – at S2. Its
diminished fifth, BF , resolves to the mediant triad’s lower third, CE .
9. The D is supported by II7 during T2, followed by a cadential 64 inaugurating the
7
dominant at U3 and then V 5 at U4. Because the embellishing chord preceding
♯3

the mediant is built above bass B and the II7 chord is presented in 6
5 position,
the bass ascends A<B<C<D<E against the melody’s E>D>C>B.
10. Location R1 projects a tonic harmony that is diatonic and minor in quality.
Location S1 projects a lowered supertonic harmony that is chromatic (B♭
wobble) and major in quality. This I–♭II succession would be expected to
proceed to V♯. The chord at T1 (with sounding members G♯, B♮, D, and F) is an
evolved (added seventh and ninth; absent root) and inverted representative of
that dominant.
11. The dominant chord at T1 is robust, built from four pitch classes (G♯, B, D, and
F). Over the course of the measure, three linear trajectories of an ascending or
descending third transpire, in a manner that results in a connective chord at
X1. Interestingly, all three of those trajectories land on a B or a D at T2. Though
not robust, as was the chord at T1, this B-D chord should be understood to
represent a continuation of the dominant harmony.
12. The dominant harmony of T1–2 resolves to the tonic at R3. Over the course of
the measure, the tonic shifts to mediant C-E-G, presented in its first inversion.
13. II7 (presented in 65 position), V♯ 73 , I.
14. A circle of fifths connecting the tonic and the mediant (E♭–A♭–D♭–G♭)
transpires between R2 and R3. Whereas the chord at T1 would be of minor
quality in its diatonic formulation, Schumann transforms it into an A♭-C♮-E♭-
G♭ surge. The D♭ chord at T2 is of major quality in its diatonic state. Schumann
does not add a minor seventh to create a pronounced surge effect.
15. There is a superficial similarity between these chords, in that each projects a
low A♭ at the bottom of the texture. Yet their roles are contrasting. The A♭-C♭-
E♭ chord at S is an unfurling of tonic E♭’s 64 embellishment. (The tonic third G♭
 Notes to pages 78–87 253

is embellished by A♭; the tonic fifth B♭ is embellished by C♭.) In this context the
A♭ does not serve as tonic root E♭’s successor. Instead E♭ is maintained
throughout the passage from R1 to R2. In contrast, the A♭ at T1 represents
the initiation of a root progression (the circle of fifths explored in question 1)
that leads beyond the initial tonic.
16. A pitch that Schumann spells as Eº for notational reasons (though a
juxtaposition of D♭ and D♮ would more appropriately convey its role as a
chromatic passing note between D♭ and E♭) sounds along with members of the
mediant chord at R3. The mediant’s inversion arrives at R4.
17. The bass pitches are G♮, A♭, and B♭. The G supports an inverted I➔ (G♮-(B♭)-
D♭-E♭), the A♭ supports IV, and the B♭ supports V♮, first with a cadential 6
4
7
embellishment at X1, then with 5 spread between X2 and X3 (connected by a
♮3

passing chord).

9 Mendelssohn: Octet in E♭ Major (op. 20), movement 1

1. Philadelphia: Gordon and Breach, 1992. The first movement of the Octet is
explored on pages 72–96.
2. Cambridge University Press, 2011. The first movement of the Octet is explored
on pages 60–72.
3. Oxford University Press, 2006.
4. The manuscript, housed at the Library of Congress, was printed in facsimile in
1976, issued by the U.S. Government Printing Office in Washington, D.C.
5. The C>C♭>B♭ line from neighbor C helps differentiate Kopfton B♭ from the
tonic chord’s other arpeggiated pitches, as does the sforzando C>B♭ stated
multiple times in the bass, beginning in measure 10.
6. See my Thinking About Harmony, pp. 155–161.
7. This particular evolved state of the supertonic will be deployed again later in P,
at 274 and at 314.
8. Without the “harmonic reduction” caption, I might be able to convince
myself that the example shows merely a series of initiation points – paths
from E♭ to Fm, as emphasized by the first three brackets below the staff and
described as “a series of overlapping harmonic expansions up a step of ever
increasing scale, which become progressively more firmly established”
(p. 61) – with no intention to convey a broader continuity. Taylor’s
analysis transpires in the context of an attempt to demonstrate the
following proposition: “Both thematically and harmonically, this opening
movement reveals a continual process of growth out from its opening phrase
that indeed justifies the analogy with the aesthetic ideal of organic growth
and unity claimed by several commentators” (p. 60). As my chapter unfolds,
254 Notes to pages 88–94

readers will encounter evidence of “organic growth and unity” not so much
in disjointed juxtapositions of E♭ and Fm as in diverse manifestations of I5–6
II V I.
9. None of these accents are included in the score excerpts as printed in Taylor’s
exx. 2.2 and 2.3a–b.
10. The vi of fig. 1 certainly is a typographical error. It should appear as VI, as in
figures 3 and 6, confirmed also by the words “major submediant” in his
commentary (p. 73).
11. Vitercik is not consistent in his application of this protocol. The E♮-G-B♭-D♭
chord of measures 23 and 24 is labeled as VI in figures 3 and 6. Yet C-E♮-G-B♭-
D♭ in measure 41 is labeled as V/ii in fig. 4 – and then as VI in fig. 6!
12. Readers concerned that my style of Roman numeral analysis is more
cumbersome than Vitercik’s are welcome to use my shorthand symbol VI→
instead of the four-tiered entity displayed in 9.2a. Though I appreciate my
system’s capacity for precision (keep in mind that several different evolutions
of the submediant could be labeled as VI→ but have contrasting appearances
when my Arabic numerals, accidentals, and bullet symbol are deployed),
others may find my arrow notation a convenient and intuitive alternative.
13. My conception of harmony as a constantly churning, vibrant process involving
frequent shifts in a chord’s configuration (generally resulting in a more
dynamic thrust toward its successor, as in II’s eventual surge targeting V)
makes Vitercik’s word “sitting” seem especially inappropriate.
14. Whereas my harmonic analysis features a conventional E♭–B♭–E♭ bass
arpeggiation (beamed in 9.2a), Taylor’s slur from C to E♭ in his ex. 2.1
(where I read the C as corresponding to that of both measures 25 and 31) is
matched by a slur from C to G (for the E♭ chord’s inversion) during measures
31 through 33 of ex. 2.5 Thus my description above of his slurring in the
vicinity of measures 28 and 29 as “adequate” may be wishful thinking on my
part. If pressed, he might clarify that his three small slurs break up what
foundationally would be a broad slur from the C of measure 25 to the E♭ of
measure 29.
15. Though many analysts espouse the ubiquitous symbols Fr+6 and Ger+6, one
may reasonably ask whether the chords they represent have roots. The vital
presence of A♮⇨ and C⇨ within the circular progression of 9.3c offers a
strong incentive to answer that question in the affirmative. In both cases the
augmented sixth interval is formed by the pitches a major third and a
diminished fifth above the root (which is omitted in the “German” version
of the chord).
16. I might hope that, if pressed, Vitercik would name C as the root of the chord in
measure 58. In his chapter he follows most other analysts in merely applying
the “French sixth” nickname (p. 76). (See also note 29, below.)
17. In his commentary Vitercik alludes to “the abrupt arrival in G minor in m. 52”
(p. 75) – the key of G Minor, whose dominant root D emerges at 521. Even if I
 Notes to pages 94–99 255

were to concur that the passage is “in” G Minor, I would not interpret either
the G-B♮-D or G-B♭-D chord of measure 55 as an asserted tonic. G Minor’s V♯
holds sway for the entire duration of measures 52 through 56.
18. My 9.4c also omits several foreground chords that systematically traverse an
upward path between the accounted-for F and C: F E♭→ A♭ G→ C.
19. Another infelicity in Vitercik’s Roman-numeral deployment should be noted
as well: his use of the symbol VI/ii for the D-F♯-A chord of measure 52.
Appropriately, the VI numeral is capital because the D chord is of major
quality. But VI in the key of ii (F Minor) is D♭-F-A♭, not D-F♯-A. In that
either D♭ or D♮ may be deployed as a root in this context, the analytical
notation must be able to differentiate between them. (Compare with my
9.11, measures 179–180.)
20. As much as possible, conventional harmonic analysis neutralizes the impact of
chromatic chords. For example, chromatic D-F♯-A-C in C Major is made to
appear as diatonic V7 in G Major through the symbol V7/V. Alas, that tactic
will not work for D-F♯-A♭-C, since that chord’s pitches are not all diatonic in
any one key. What can be done to make such intense chromaticism palatable?
Nicknames! But in doing so, information about D’s role within the chord’s
construction and the chord’s location within tonal space are sacrificed.
Though this lamentable situation is not Taylor’s fault, of course, it helps to
explain how he could fail to come to terms with the chords at the ends of
measures 51 and 58 within his conception.
21. Mendelssohn’s Instrumental Music by Erez Rapoport (Hillsdale, NY:
Pendragon Press, 2012) is a book that all Mendelssohn enthusiasts should
know. I have elected not to interact with it in a substantial way in this volume
because, as editor of Pendragon’s Harmonologia series (in which it appears
as volume 18), I was deeply involved in its production. I will honor his
achievement by simply mentioning our contrasting, equally viable
treatments of one specific chord from the Octet’s first movement: D♭-E♮-
G-B♭ at 1161. I take an overtly harmonic approach in 9.6, interpreting the
chord as a manifestation of II→. He instead leaves an open area between I6
(his symbol for the tonic in first inversion) in measure 115 and V at 1163 in
his ex. 1.9. His approach thus is more linear than mine: a formation
consisting of multiple appoggiaturas delaying the dominant’s arrival by
two beats. Janet Schmalfeldt offers a middle path in her book, In the
Process of Becoming: Analytic and Philosophical Perspectives on Form in
Early Nineteenth-Century Music (Oxford University Press, 2011). While
her ex. 7.13 shows I6 and V 4
7
3 numerals akin to Rapoport’s, she fills in the
space he left between them with the annotation (°7).
22. Vitercik’s brief comment regarding the G major submediant – “strongly biased
toward its own minor submediant, C minor” (p. 82) – does not make sense to
me. I think the word subdominant was intended, instead of submediant, in
256 Notes to pages 100–108

reference to the G-C-E♭ chord at 781 and 821. In my view, Mendelssohn is

merely upholding the B♭ Major tonicization through the deployment of its
diatonic E♭, rather than G Major’s E♮. In any event, the chord is a local 64
embellishment of G-B♮-D. I suggest that it does not warrant being described as
a subdominant, which would imply C’s assertion as a root.
23. Though the placement of parentheses around V/vi appears intended to imply a
temporary diversion from a broad V extending from measure 90 to the
purported I in measure 96 – “a momentary shift from V of B♭ to V of G
minor” (p. 82) – it is in fact the V/vi that resolves instead of the F dominant.
While rejecting Vitercik’s proposal that the B♭ tonic emerges at measure 96, I
would advise against merging the dominant of measures 96–101 with that
which sounded in measure 90. That earlier dominant’s tendencies are used up
with soprano E♭’s descent to D (supported by the B♭ tonic’s 6-phase rather
than its 5-phase chord).
24. The latter chord is referred to as “V 65 of IV” in Vitercik’s commentary (p. 85).
25. My thought that perhaps V/c was intended to convey “V of c” followed by “c”
was shot down when I noted the juxtaposition of V/f and f, V/g and g, etc.,
earlier in the same diagram.
26. The precise wording of Taylor’s remark at this point exemplifies the wide gap
between our harmonic conceptions: “This G major then moves, via C minor,
to a temporary A♭” (p. 61). Whereas my G→ C conception proposes that the
former chord is dependent upon the latter, Taylor appears to hear a G<A♭
second with an intervening C chord. Likewise my interpretation of A♭ as an
intermediary between C and F is foiled not only by Taylor’s omission of F from
his purview but also by his subordination of C to A♭.
27. The D♭-to-G succession is efficiently realized by adding B♮ to the D♭-F-A♭
triad. (Compare with 9.4a.) The G chord’s supersurge is especially appealing
here because F Minor will be tonicized. The most common location for a
supersurge is a key’s supertonic, and so G⇨ C→ F has the effect of II⇨ V→ I.
28. For further evidence that Vitercik’s measure numbers may not always be
reliable, see page 94, lines 7 and 10, where the number 235 certainly should
read as 253.
29. In note 16, above, I express a hope regarding what pitch Vitercik might regard
as the root of an augmented sixth chord. The fact that what he describes here as
a “circle of fifths” (fig. 9) includes a D♭-F-A♭-B♮ augmented sixth chord that
would fit within the circle only if G were regarded as its root gives support to
my hope.
30. Observe that a prolonged G→ chord (beginning in measure 77 and similar to
what precedes the C minor chord of measure 137) targets the C minor chord of
measure 86. Thus I can find no basis for Taylor’s assertion that C minor is
“comparatively absent from the exposition.” His emphasis on C Minor stems
instead, I believe, from his reading of a broad connection between the C minor
chord of measure 137 and the C major chord of measure 148. (He describes the
 Notes to pages 109–120 257

arrival of the latter as “back to C” on page 71.) Though his full conception of
the hierarchical relationships among chords cannot be fully reconstructed
from his example and commentary, it appears that whereas I hear a
connection between the B♭ chords of measures 127b and 141, leading up a
fifth (via D♭) to F at measure 149, he instead is proposing a B♭ C → F trajectory
(measures 127b, 137–148, and 149).
31. I raise this concern also in Harmony in Beethoven, chapter 13, in coming to
terms with a movement (String Quartet in A Minor (op. 132), movement 1)
that contains departures from the norm far more radical than Mendelssohn’s
here.
32. Continuing my assessment of Schmalfeldt’s example cited in note 21, above,
observe that she and I offer contrasting readings of the melodic content in the
vicinity of measures 269 through 271. Whereas she places an unequivocal ^1
atop the G, F, and E♭ that sound during 2701–2, in my view the melodic line has
descended only as far as ^3 at that point (with ^3 more overtly “on top” in
measure 271 than in measure 270). Consequently the G>F>E♭ of measures
277–278 (repeated in measures 279–280) concludes the linear initiative, not
acknowledged in her reading. From my perspective, her “CODA” and PAC
markings at 2761 should be moved to 2801.

10 Mendelssohn: Song without Words in F Major

(op. 85, no. 1)

1. “Form and Tonal Process: The Design of Different Structural Levels,” in Trends
in Schenkerian Research, ed. A. Cadwallader, New York: Schirmer Books, 1990,
pp. 1–21.
2. The initial impetus for the A chord certainly has something to do with
symmetry: during stanza 1 the trajectory proceeds down a third from F to D
(measures 5 and 6), whereas during stanza 2 the equivalent measures (13 and
14) instead lead up a third from F to A. Once there, however, Mendelssohn was
free to use the A sonority however he pleased. My reading ultimately connects
the F chord of measure 13 and the D chord of measure 19 (thus again F5–6),
more akin to what happened during the first stanza than one might have
expected at first, given the potency of the intervening A chord’s presentation.
3. Given the persistence of intervallic unfoldings in the melody (C>F in measures 3
and 11, E<B♭ in measures 4 and 12), it is mildly unsettling that only the A of the
A
tonic’s resolving F third is presented by the melody during measures 5 and 13.
Though perhaps a stretch, one might regard the A<B♭<C<D<E<F sixth-
progression of measures 13 through 23 as belatedly fulfilling the expectation
of an unfolding, with an ascending sixth replacing the registrally normative
descending third.
258 Notes to pages 127–134

11 Schumann: “Warum?” from Phantasiestücke (op. 12)

1. Music Theory Spectrum 27 (2005), pp. 159–185. The discussion of “Warum?”

appears on pages 180 through 182.
2. Burstein is unequivocal on this point: “The work’s conclusion . . . lacks both an
Urlinie descent and perfect authentic cadence” (p. 181).
3. After completing this chapter I discovered that a variant of Burstein’s essay was
published in Essays from the Third International Schenker Symposium, ed.
A. Cadwallader, Hildesheim: Georg Olms, 2006, pp. 1–36. Its ex. 18, which
corresponds to the Music Theory Spectrum article’s ex. 17, in fact positions the
V7 label below melody pitch E♭ in all three instances.
4. Compare Schumann’s mediant usage here with Mendelssohn’s in his Song
without Words in F Major, explored in chapter 10. (See measures 13–19 in
10.2.)
5. Many Schenkerian analysts from New York deploy parentheses around a
Roman numeral to indicate that it does not correspond to an actual Stufe (a
functional harmony), but instead plays a linear role. I instead favor reserving
Roman numerals for asserted Stufen.
6. Burstein’s care in showing the descent of the D♭ from a DG ♭♮ diminished fifth in
G♭
measures 1 and 2 is not matched by showing the ascent of the C from a C

diminished fifth in measures 9 and 10.

7. In the microfilm copy of the collection, the score is labeled as item 91 and is
found on reel 42.
8. Of course, one should guard against granting Schenker an aura of infallibility.
For example, in the same score he analyzes 122 through 13 as ♮IV V below the
system (with a II placed between the staves above the latter numeral). In my
analysis above, I discount the G♮ that motivated Schenker’s ♮IV numeral
(which corresponds to II➔ in my analytical notation), proposing instead a
full measure of 6-phase tonic, a full measure of II➔, and a full measure of V.
(This reading is confirmed during the next complete cycle, where only measure
31 – corresponding to measure 13, not measure 12 – may serve viably as II➔.)
9. These chords often go by the names French and German augmented sixth.
Schumann here camouflages them by spelling their augmented-sixth interval
º
( CA ♭♮ ) as a minor seventh ( BC ♭ ), justified by the neighboring-note context (Bº
descending to chordal seventh A♭ rather than the typical ascending resolution
of A♮ to chordal root B♭). Though in conventional theory the augmented sixth
chords generally precede the dominant harmony, I propose that a supersurge
relationship may prevail between any pair of chords whose roots relate by
descending perfect fifth.
10. The application of an augmented fifth also may imbue a tonic chord with a
surge character. Thus the A♭<A♮<B♭ line during measures 34–35 and 38–39
contributes to the surge effect along with the addition of minor seventh C♭.
 Notes to pages 136–143 259

12 Mendelssohn: Song without Words in A♭ Major

(op. 53, no. 1)

1. Lewiston, NY: Mellen, 2008. The Mendelssohn graph with which I interact is
part of his ex. 7.65. That example’s caption reads as “Other harmonizations of
V7-8-7 motion.” In his reading V7 (E♭-G-B♭-D♭) is expanded via the ascending
step from seventh D♭ to octave E♭ and back. The E♭ is supported by a downward
shift from dominant root E♭ to C, with a return to E♭ coinciding with the
restored D♭ above.
2. Though the introduction II➔’s D♮ is an upward-tending pitch, in its context the
broader downward trajectory mandates a D♮>D♭ half step. The situation is not
so clear with A1’s II➔ chord: does the D♮ at the end of measure 6 proceed to E♭,
followed by D♭; or does the F of 63 descend to E♭ while D♮ descends to D♭? (In
12.1a, a D♮<E♭ resolution is proposed.)
3. H. Schenker, Der freie Satz: Neue musikalische Theorien und Phantasien III,
Vienna: Universal, 1935; rev. edn, ed. O. Jonas, Vienna: Universal, 1956; as Free
Composition, trans. and ed. E. Oster, New York: Longman, 1979; reprint trans.
Stuyvesant, NY: Pendragon Press, 2001.
4. Other graphs in Free Composition that resemble my 12.1a include fig. 67, ex. 4b
(second model); fig. 76, ex. 2 (where ^4 is the upper neighbor of ^3 rather than a
descending step from ^5 ); and fig. 152, ex. 4.
5. In my conception of how harmony works, the mediant typically would play one
of two roles. It might be the principal intermediary between I and V; or, it might
be a means of extending the scope of the initial tonic, followed by the chord
(generally a form of II or IV) that serves as the principal intermediary between I
and V. The Roman numeral analysis in 12.2 suggests that Mendelssohn’s
writing in this composition exemplifies the latter of those two possibilities.
6. Two chapters of the Harmonielehre that serves as Part One of my Harmony in
Beethoven address the mediant. Of the thirteen excerpts analyzed, three deploy a
circle-of-fifths connection between the tonic and the mediant. (See 4.2, 4.7, and
5.4a in that volume.) Certainly the insight gained from previous efforts should
help analysts as they probe other compositions. I am unabashedly “on the
lookout” for a circle-of-fifths progression whenever I become aware that the
mediant is in play.
7. That E♭ is embellished by appoggiatura F during measure 10, appoggiatura D♮
during measure 17, and appoggiatura F♭ during measure 19.
8. The first beat of measure 29 is packed with content that warrants some further
elucidation. Because chord member G is absent at the downbeat and the
melody’s embellishing F displaces chord member E♭, only one of that beat’s
three chords is fully and “correctly” presented. Those chords should be
understood as C-E♭-G, A♭-C-E♭, and A♮-C-E♭-G♭. Given the circle-of-fifths
approach and its sounding on a downbeat, the C mediant chord overpowers the
260 Notes to pages 144–169

fleeting A♭ chord that follows in my perception of the passage. I hear the

mediant as a connector between the two chords shown in the first model of
12.3b, with an A♭>G>G♭ chromatic line (with a migration up an octave and
back) in the alto register of the second model despite the A♭ that sounds on the
measure’s second eighth note in the score. Other analysts might reasonably
propose that the mediant C chord comes between two tonic A♭ chords.
9. Recall my observation earlier that the introduction’s E♭>D♭>C third sounds in
retrograde at the onset of A1. Now an A♭>G>F>E♭ fourth and its retrograde are
juxtaposed during the coda.

13 Schumann: Three songs from Liederkreis (op. 39)

1. Schenker Studies, ed. H. Siegel, Cambridge University Press, 1990, pp. 146–164.
2. The Journal of Musicology 22 (2005), pp. 131–153.
3. The first two stanzas break down into four four-bar hypermeasures, for example
measures 6–9, 10–13, 14–17, and 18–21.
4. Compare with Schenker’s analysis of another work by Schumann, displayed in
C♯
Free Composition, fig. 22b. Whereas the upper graph shows a consonant A

tenth at the onset of the A2 section of an A1 B A2 ternary form, the middle and
lower graphs reveal that in its surface realization this tonic (bearing the Roman
numeral I in Schenker’s analysis) is surging from its onset. In her In the Process
of Becoming: Analytic and Philosophical Perspectives on Form in Early
Nineteenth-Century Music (Oxford University Press, 2011), pp. 230–236,
Janet Schmalfeldt borrows from Burkhart’s analysis, with his permission repri-
nting and amplifying some of his voice-leading graphs. Below measures 47
through 50 of his ex. 4 (which she presents as her fig. 9.2a), she adds the
distinctly un-Schenkerian label [V 7




65 ]➔, targeting the IV of measure 51.
Though her annotation of the score (in her fig. 9.1) follows a similar course at
measure 59, her amplification of Burkhart’s graph displays a I numeral at that
point.
5. “The resulting clash is profoundly moving, even anguished” (Charles Rosen,
The Romantic Generation, Cambridge, MA: Harvard University Press, 1995,
p. 688).
6. For example, August Swoboda describes this usage in his Harmonielehre (1828).
See my Thinking About Harmony, p. 313, note 14.
7. Sometimes in tonal music an expected event will be delayed, rather than
omitted. So one should investigate what happens at the onset of the cycle’s
next song. Alas, turning the page for “Im Walde” reveals no such outcome. In
response to Patrick McCreless’s “Song Order in the Song Cycle: Schumann’s
Liederkreis, op. 39” (Music Analysis 5 (1986), pp. 5–28), one might also
investigate “Auf einer Burg” (which follows “Zwielicht” in one of “The Four
 Notes to pages 171–174 261

Orderings of the Songs of Liederkreis” listed in his fig. 2). Though that song
retains E as tonic, it lacks such a descent. (It instead projects an ascending third –
E<F♯<G – in its second measure!)

14 Schumann: “Im wunderschönen Monat Mai” from

Dichterliebe (op. 48, no. 1)

1. “Introduction to Musical Ambiguity,” in Engaging Music: Essays in Music

Analysis, ed. D. Stein, Oxford University Press, 2004, pp. 77–88.
2. Essay on the True Art of Playing Keyboard Instruments, trans. and ed.
W. J. Mitchell, New York: W. W. Norton, 1949, p. 106.
3. We encountered the notion of auxiliary cadence in the context of another work by
Schumann in chapter 11. It should be noted that Schenker and I apply the concept
in entirely different ways in the context of this song. I hear the work in F♯ Minor
throughout, and thus an unrealized initial F♯-A-C♯ tonic precedes subdominant
B and dominant C♯. Schenker offers a graph in which the C♯ chord of measures 2
and 4 is interpreted instead as III♯ in A Major, followed by V (the end of measure
5) and I (measure 6). Thus for him the unrealized chord is an initial A-C♯-E tonic.
His bass D at 51 is a move forward from C♯, toward dominant root E; for me (as
will be explained as this essay unfolds) it instead is a backtracking to the D of
measures 1 and 3, now proceeding along a contrasting trajectory. See Schenker’s
Der freie Satz: Neue musikalische Theorien und Phantasien III, Vienna: Universal,
1935; rev. edn., ed. O. Jonas, Vienna: Universal, 1956; as Free Composition, trans.
and ed. E. Oster, New York: Longman, 1979; reprint trans. Stuyvesant, NY:
Pendragon Press, 2001, fig. 110, ex. c2.
4. In attempting to convey where those chords are located, I encountered another
obstacle: the measure numbers supplied in Stein’s score are all off by one. A
chord equivalent to that of measure 1 is found in measures 3, 14, and 25 – or
Stein’s 4, 15, and 26. Henceforth in this chapter I shall use the correct numbers,
as Stein herself does in her textual commentary.
5. In his “Schumann’s Fragment” (Indiana Theory Review 28 (2010), pp. 85–109),
Nathan John Martin states unequivocally that “Schumann’s song is in A”
(p. 106). He provides a chord-by-chord harmonic analysis in that key (his
ex. 10 on pp. 100–101) that predictably begins with ii6. (That analysis later
incorporates tonicizations of ii and of IV.) His ii6 label followed by III 7♯ for the
C♯ chord of measure 2 is consistent within one key, whereas Stein’s ii6 followed
by V7 (“of f♯”) is not, leading me to suspect a printing error rather than an
intentional switch from one key into another so early in the work. It might
have emerged, though, through initially pursuing a conception resembling
Martin’s but not erasing it entirely upon switching to a reading of the opening
measures in F♯ Minor.
262 Notes to pages 174–176

6. My analytical endeavor to “reconstruct” a non-sounding tonic opening using

music notation is simply a diligent pursuit of a perception conveyed by
Charles Rosen as follows: “The first song of Schumann’s Dichterliebe begins
in the middle . . . It starts as if continuing a process already in motion . . . There
is, in fact, no F sharp minor chord anywhere in the song – and some doubt
whether F sharp minor is really the key of the song. With an introduction that
does not fix the tonality but seems to take a previously settled harmony and
rhythmic motion for granted . . . “Im wunderschönen Monat Mai” is a brilliant
and famous example of the open form which was one of the ideals of the
period.” See his The Romantic Generation, Cambridge, MA: Harvard
University Press, 1995, pp. 41–48. The quoted passages (here with emphasis
added) appear on page 41.
7. Both historically and in contemporary thinking, music theorists have debated
what pitches should be regarded as forming the minor mode’s foundational
state. In my analytical practice the “natural minor” pitches – those projected by
the conventional minor key signatures now almost universally employed –
form that foundational state. Thus F♯ Minor’s diatonic mediant chord is A-C♯-
E, in my view.
8. In his “Schumann, Heine, and Romantic Irony: Music and Poems in the First
Five Songs of Dichterliebe,” Intégral 10 (1996), pp. 93–123, Lauri Suurpää falls
squarely in the A Major (rather than F♯ Minor) camp regarding this song’s
tonal center. (For example, the Roman numeral III♯ appears under bass C♯ in
both measure 2 and measure 13.) Though a comparison of his Example 1 and
all my examples would be a rewarding undertaking for the reader after
finishing my chapter, I limit my comments here to the region around the
F♯➔ chord at 92. For me, this moment is the culmination of tonic F♯’s exended
prolongation, where (as very often occurs at the juncture between I and IV) it
now surges. Suurpää’s tenor G♮>F♯>B beam, alto B stem, alto B<C♯<D slur,
and soprano B<F♯ unfolding symbol all conspire to diminish the hierarchical
value of the F♯-C♯ fifth that represents 92 in his graph. I suggest that he has
misconstrued the voice leading in several ways: tenor G♮ is a passing note
between A and F♯, alto B is a passing note between A and C♯, and soprano C♯
reinstates the open-notehead C♯ prolonged since the introduction (the upbeat
to measure 1 for me, measure 2 for Suurpää). That C♯ descends to B, below an E
(attained via reaching-over) that descends to D. The B later becomes the
seventh of measure 13’s C♯ dominant.
Beate Julia Perrey’s discussion of the song in her Schumann’s Dichterliebe
and Early Romantic Poetics: Fragmentation of Desire (Cambridge University
Press, 2002) features a display of the “succession of harmonies” as ex. 9
(p. 166). While I concur with “F♯7” for 92, I would argue that the preceding
chord is not a “harmony” at all. Her reading of measures 8 through 10 – A e6
F♯7 B – obfuscates what is most essential: the third-relationship between A and
F♯7, connected by passing notes (displayed as filled-in noteheads in my 14.3).
 Notes to pages 177–181 263

On the other hand, the display of the relationship between the F♯7 and B
chords in her ex. 11c (p. 173) is artistically sensible in a way that Suurpää’s
graphs are not.
9. Consulting chapter 5 of my Harmony in Beethoven (“The mediant within the
orbit of the tonic”) is recommended to gain a deeper sense of how the mediant
may be deployed in such contexts. In that chapter’s music examples, the
mediant sometimes is labeled as I7• (wherein the bullet indicates an absent
root), sometimes is given no label, and sometimes (particularly when it is
tonicized) bears the label III.
10. Schumann’s bass B at the end of measure 14 appears incorrectly as D (a third
higher) in Stein’s ex. 7.1.
11. I here refrain from engaging with the principal word in the title of Stein’s
essay: ambiguity. Instead I have proceeded simply to interpret the work as I
hear it – in F♯ Minor throughout. Consistently my analyses are less prone to
invoke the notion of ambiguity than do those of other analysts. (I address this
issue head-on in chapter 7 of my Harmony in Chopin, where I respond to
Alison Hood’s article “Ambiguity of Tonal Meaning in Chopin’s Prelude
Opus 28, No. 22.”)
David Kopp offers a sensitive study of the notion of key among prominent
nineteenth-century theorists, which he folds into a discussion of “Im
wunderschönen Monat Mai” in his “Intermediate States of Key in
Schumann,” published in Rethinking Schumann, ed. R.-M. Kok and
L. Tunbridge, Oxford University Press, 2011, pp. 300–313. He assesses how
F♯ Minor and A Major may be understood as alternative tonics, supplying
nuances of interpretation from moment to moment within the song. Though
I do not endorse his outcome, I recommend his essay as an example of
mainstream thinking on such matters among contemporary analysts.
12. My reading of the song contrasts that of Jon W. Finson in his Robert
Schumann: The Book of Songs (Cambridge, MA: Harvard University Press,
2007), pp. 62–63. He contends that Heine’s poems should be read ironically as
“shameless lies.” I propose instead that Schumann sets up his cycle by
conveying a sincere hope of love in the first of the cycle’s songs (which
Finson refers to as “psychological snapshots”), only later to be devastatingly
undermined. Whereas I interpret the tonal trajectory of “Im wunderschönen
Monat Mai” as consistent with the (perhaps naïve) expectation of a positive
outcome, Finson reads Schumann’s music as a denial of the text’s affirmative
message: “Schumann employs tonal instability in the first number to suggest
that all is not what it seems. The accompaniment begins disoriented, as if in F-
sharp minor, but the voice part always enters in the relative A major. Just when
the tonality seems secure at the end of each stanza, Schumann retreats into the
vacillation between the subdominant and dominant seventh in F-sharp minor,
and he finally ends inconclusively on a half-cadence . . . The tonal basis of the
264 Notes to pages 181–183

first song, and by implication the foundation of the whole cycle, is a sham: the
home key, like love itself, does not exist.”
13. David Ferris assesses the Schenker-oriented studies of the entire work by
Arthur Komar and by David Neumeyer in chapter 2 of his Schumann’s
Eichendorff Liederkreis and the Genre of the Romantic Cycle, Oxford
University Press, 2000. He advises: “And so the song cycle sets up a kind of
vicious circle for the analyst: we try to solve the problem of organic unity in the
song by seeking it in the cycle, and the reason we believe that we can solve the
problem of organic unity in the cycle is that it is lacking in the song” (p. 26).
Komar’s study is found in “The Music of Dichterliebe: The Whole and Its
Parts,” in Dichterliebe, by R. Schumann, ed. A. Komar, New York: W.
W. Norton, 1971, pp. 63–94; Neumeyer’s can be found in “Organic Structure
and the Song Cycle: Another Look at Schumann’s Dichterliebe,” Music Theory
Spectrum 4 (1982), pp. 92–105.
14. In a seminal article on Schenkerian analysis, Allen Forte chose Schenker’s
graphs of “Aus meinen Tränen sprießen” as the principal focus of his
explication. See his “Schenker’s Conception of Musical Structure,” Journal of
Music Theory 3 (1959), pp. 1–30; reprinted in Readings in Schenker Analysis
and Other Approaches, ed. M. Yeston, New Haven: Yale University Press, 1977,
pp. 3–37.
15. David Neumeyer labels the C♯ chord that concludes the introduction in
measure 4 (and, by extension, those of measures 15 and 26 as well) as an
applied chord of vi in A Major, with that vi transpiring during measure 1 of
Dichterliebe’s second song. Analyzing only the first song leaves one
unequipped to respond to such an assertion. See his “Organic Structure and
the Song Cycle: Another Look at Schumann’s Dichterliebe,” Music Theory
Spectrum 4 (1982), pp. 92–105. His analytical graphs integrating content
from the first two songs appear as ex. 8 on p. 104.

15 Schumann: Sonata in A Minor for Violin and Piano

(op. 105)

1. “Harmonies Heard from Afar: Tonal Pairing, Formal Design, and Cyclical
Integration in Schumann’s A-minor Violin Sonata, op. 105,” Theory and
Practice 34 (2009), pp. 47–86; “The Drama of Tonal Pairing in Chamber
Music of Schumann and Brahms,” in Expressive Intersections in Brahms:
Essays in Analysis and Meaning, ed. H. Platt and P. H. Smith, Bloomington:
Indiana University Press, 2011, pp. 252–290; and “Tonal Pairing and
Monotonality in Instrumental Forms of Beethoven, Schubert, Schumann, and
Brahms,” Music Theory Spectrum 35 (2013), pp. 77–102.
 Notes to pages 185–189 265

2. The D that continues after D♯ in the violin line “belongs” in measure 4 but is
delayed until the downbeat of measure 5. A more immediate D♯>D connection
transpires in the piano part.
3. For the exposition’s repeat Schumann introduces the supertonic’s root in the
bass (measures 66 and 67), though by the time the C suspension above has
resolved to B, the bass has descended to chordal seventh A, so that the
connection at the juncture between II➔ and V➔ is again A>G♯.
4. Compare with the similar deployment displayed in Harmony in Beethoven’s
2.6.
5. Smith juxtaposes these passages (and others) in his ex. 2011/9.5 (reprinted as
ex. 2013/15(A)), with some hesitancy conveyed in measure 1 through the
placement of the VI numeral within quotation marks and surrounded by
parentheses: i (“VI”). (He explains: “initially the potential for the emphasized
F to express itself in the harmonic dimension remains latent. The 63 chords
above A in mm. 1 and 2 are byproducts of an implicit 5–6 contrapuntal motion
within the tonic” (p. 2009/52).) Yet when that passage sounds again (with the
melody in the bass) at measure 16, those markers have been removed: i VI. (My
graph displays a VI numeral neither in measure 1 nor in measure 16.) One
should consider also whether or not the F descends immediately to E. In some
of the passages assembled for Smith’s example, the annotation ^5 follows
directly after the featured ^6. In others, it does not. Through comparison with
Example 2009/4 we learn that Smith interprets the violin E at the end of the
first movement’s first measure as a passing note between F and D♯. Yet if that is
the case, his comment at another location – that “the finale’s main theme takes
as its motivic kernel a double neighbor figure that recalls the haunting opening
gesture of the first movement” (p. 2013/89) – seems curious, since the two
passages are analyzed in diametrically opposing ways (F>E>D♯ in movement
1, F>E in movement 3). As will become apparent later, I read both Fs in the
same way: as neighbors resolving immediately (F>E).
6. A context for the slurred F>E>D♯ that Smith displays twice during measures 1
through 3 will emerge near the end of the movement – see measures 179 and
180. (That deployment begins against root F rather than against root A, thus
reversing the hierarchical relationship between F and E.) As will become clear
as my analysis unfolds, Schumann deploys a most potent chord (II⇨) at that
decisive moment as a means of exceeding the impact of three other approaches
to the dominant in earlier measures. Smith’s graph suggests that we encounter
this highly charged “supersurging” chord already in measures 2 and 3!
7. Though I suspect that Smith intends for the Arabic 6 he has placed below bass
D in measure 14 to indicate that the D-F-B chord is in 63 position, it is possible
(given the conspicuous connection from the F-A-C chord of measure 12) that
he instead wants to convey that the D is the 6-phase pitch within VI5–6, prior to
B’s arrival.
266 Notes to pages 189–200

8. Though Schumann’s projection of the linear trajectory is not quite complete –

D’s 6-phase pitch is displayed within parentheses in 15.1 – the sense that a
linear trajectory is being traversed should nevertheless be apparent. As such, I
question Smith’s breaking up of the C<D<E<F bass fourth into a third plus a
second and the V/VI harmonic label. This “tonicization” of VI apparently
justifies the denial of the following D-F-B chord’s role as ii° within the broader
progression. (Hierarchically, the supertonic seems to me more significant than
the alleged V/VI.)
9. Whereas I propose a resolution from a dominant 42 chord to a tonic 63 chord
during measure 59 based on normative harmonic tendencies, Schumann’s
slurring opens the possibility of an alternative reading: an abandoned
dominant at 591, with a fresh tonic emerging at 592.
10. My analytical perspective was inspired by FC, fig. 76, ex. 7.
11. While acknowledging that Schumann does “touch on C several times . . . but . . .
only in 63 position,” Smith proposes that “the harmonic progressions center on
the local G dominant” and that the “tonicization of C fails to articulate a
middleground mediant Stufe” (p. 2009/55). His label V/IV (introduced in his
textual commentary on page 2009/55) for what I regard as a tonic harmony
conflicts with Schenker’s mature practice: see FC, fig. 49, ex. 2, where a I numeral
extends through B-D♯-F♯-A♮ in B Major prior to the onset of IV; or fig. 63, ex. 1,
♭
where D♭-E♭-B♭ in E♭ Major is analyzed as I 7. I reject the assertion that
Schumann’s addition of B♭ to C-E-G at (rather than after) the chord’s onset
disqualifies it from serving as a means “to resolve the overriding G dominant”
(p. 2009/55). Given his stance here, I am surprised to read later in the essay that
“the exposition’s promise of C articulation reaches fulfillment in the
development, where the mediant finally arrives as a tonal goal at m. 104”
(p. 2009/60). So far as I can tell, the chords at 432 and 552 during the
exposition and at 1041 during the development are all spelled as C-E-G-B♭.
Why are the earlier ones denied the status granted to the latter one? Smith’s
answer apparently would be that the exposition chords are presented as
simultaneities, whereas in the development the chord’s pitches are
arpeggiated, giving a tiny sliver of time for a consonant C-E-G to assert itself,
though it “immediately acquires a V/F function” (p. 2009/61). (In my reading of
the development, to be explored in this chapter’s next section, the motivation for
measure 104’s C-rooted chord is at odds with its role within the exposition,
further distancing my view from what Smith proposes here.)
12. Compare with the minor dominant deployed in Harmony in Beethoven, 12.11,
m. 44.
13. My analytical notion of surge acknowledges voice leading like that which
transpires between V7 and I without concurrently stipulating that the
surging chord actually be rooted on scale degree 5 in some key. Especially in
the context of a circle of fifths, the various surging chords do not imply that
their successors are tonics. Consequently I find Joel Lester’s use of the word
 Notes to pages 202–209 267

dominant in his “Robert Schumann and Sonata Forms” (19th-Century Music

18 (1994–1995), pp. 189–210) to be excessively prescriptive: “The development
ends on a C dominant pedal, as if preparing a return in F (mm. 104ff.)”
(p. 205).
14. It appears that Smith was unaware of Edward Laufer’s analysis of this
movement in his “Voice-Leading Procedures in Development Sections,”
Studies in Music from The University of Western Ontario 13 (1991),
pp. 115–116 and ex. 30. Though Laufer’s presentation is less detailed than
Smith’s, their readings of measures 104 through 121 are essentially the same.
Neither of us concurs with Laufer’s interpretation of the mediant’s onset: his
graph shows a connection between the C major chord of 592 and the C minor
chord of 751, with bass A at 611 labeled as Nn (neighboring note) and slurred to
the G of 721.
15. I am not sure whether the omission of a stem for the alto-register E during
measure 106 in Smith’s ex. 2009/11 is intended to establish a hierarchy (in
which E would serve as a leading tone of the F in the following chord, which he
labels as “F”), or instead results from an oversight. Though I use open and
filled-in noteheads rather than stems to denote relative hierarchical weight in
my 15.7, my reading proposes that this E is more deserving of a stem than
either the preceding D or the following F.
16. The B chord’s minor quality hints at A Major, as does the following E chord’s
major quality. Though these chords (as well as the surging F♯➔ chord that
targets B) might transpire in the context of A Minor, they help prepare for the
emergence of A Major during the recapitulation’s FS.
17. A II➔ chord (with 64 embellishment) also is featured during the first phrase of
the third movement. (Readers may want to look ahead to 15.15, mm. 1–8.)
18. I interpret the violin’s accented sixteenth-note D at the end of measure 7 as a
local neighbor between C and C♯, resulting in a C<C♯<D trajectory for the
emergence of the tonic’s 6-phase chord. Some listeners instead might hear that
D as the initiation point for the 6-phase chord, in which case the following C♯
would serve as a lower neighbor.
19. One should emphasize that the F major chord at 121 concludes the harmonic
trajectory that began at the movement’s onset and thus brings about a PAC in F
Major. Michael Struck proposes that the F chord instead concludes the internal
b region within a local a b a’ form (measures 0|1–8, 8|9–12, and 12|13–15),
listing the Tonalität for those regions as F (A)d (F). Though a harmonic analysis
is not fully elaborated, it would appear that he regards measure 12’s cadential
F chord as III in D Minor. See his “Beziehungs-Probleme: Zum Verhältnis der
Komponisten Schumann und Brahms, dargestellt am Beispiel von
Violinsonaten,” in “Neue Bahnen”: Robert Schumann und seine musikalischen
Zeitgenossen: Bericht über das 6. Internationale Schumann-Symposion am 5. und
6. Juni 1997 im Rahmen des 6. Schumann-Festes, Düsseldorf, ed. B. R. Appel,
Mainz: Schott, 2002, pp. 294–347.
268 Notes to pages 209–218

20. In that both A and G are embellished by distinctive upper-third excursions, it

is fitting that goal F is likewise so embellished near the end of the movement:
F<G<A>G>F in both the piano and the violin parts during measure 75. It is
rare for all of the Urlinie’s pitches to be embellished in the same manner.
21. Whereas I regard Schumann’s treatment of the A➔ chord as a key component
of his whimsical initiative, Smith regards it as “a chromatic intruder within an
otherwise whimsical refrain” (p. 2011/259, emphasis added). I am baffled by his
suggestion that the A chord’s “harmonic urge” “remains repressed during the
two refrain statements” (p. 2011/259). What fault is there in my 15.10e’s
proposed A➔D succession? The repression perhaps resides instead in
Smith’s analytical imposition of a formal barrier between the A➔ chord of
81 and the D chord of 91.
22. I regard the emergence of an A➔ surge between tonic F’s 5- and 6-phase
chords to be a common occurrence in harmonic practice, so conventional that
it emerges already in chapter 1 of my Tonal Analysis: A Schenkerian
Perspective. (See the commentary to that volume’s ex. 1.7, where I discuss
the progression from the example’s Chord 1 (C) to Chords 5 (E➔) and 6 (A).)
Consequently I am less inclined than Smith to dwell on the A chord as “an
associative connection with the first movement” (p. 2009/68). Also, in my view
a surge may target any diatonic chord, not just a tonic. (See my remark
concerning a quotation from Joel Lester in note 13, above.) Thus Smith’s
references to “D tonicity” and “D-minor gravity” and his suggestion that
Schumann is “pitting the two dominants [C and A] against each other”
(p. 2011/259) seem overstated.
23. Smith acknowledges the possibility of a transition interpretation for what he
labels as A’ in his note 27 on p. 2009/85.
24. Though Cs in two other registers within the chord compensate for the absent
upper-register C, one certainly may appreciate the fact that the return of the
refrain at the upbeat to measure 26 belatedly supplies that in-register C.
25. Though typically an F major chord would sound at the beginning of the
refrain, here the sounding of an F-A-C chord does not come across as the
assertion of the tonic. This situation exactly parallels what occurs at the first
movement recapitulation’s onset: though normally an A-C-E tonic chord
would be asserted, the A-C-E chord during measure 110 (shown in 15.7)
does not project a tonic function.
26. The lack of a graph for the second episode in Smith’s 2009 essay leaves readers
uncertain regarding the full scope of his conception. Apparently he has
something different in mind from what I display in 15.12, since he proposes
that “a D Stufe never materializes” (p. 2009/68).
27. The D>C♯<E>D♯ of 132 is here interpreted as a double-neighbor
embellishment of D coinciding with a chromatic shift from D to D♯.
Compare with the addition of D♯ to an F-A-C chord during measures 179
 Notes to pages 218–222 269

and 180 of the first movement, and also later in this movement: measures 168
through 175.
28. The chord F-A-D♯ at the end of measure 13 corresponds to what many analysts
call the “Italian augmented sixth” chord. Yet, as 15.15 shows, the melody has
proceeded from upper neighbor D through C and will land on B at the cadence.
So for analysts who are attentive to voice leading, the question becomes whether
the C still prevails against F-A-D♯ (thus the “German” variety of the chord), or
whether C has by that point descended to B (the “French” variety). Given that
this question ultimately is unanswerable (due to the omission of a structurally
significant pitch in the score), I have pursued the analytical course of favoring
one of the two options in my graph (for stated reasons) but including
commentary here to alert readers regarding the delicacy of the situation.
29. How the “augmented sixth” chords should be labeled in the context of a
Roman numeral analysis is of course a controversial matter. The use of the
familiar nicknames provides no clue regarding what scale degree serves as the
foundation for such chords. Since Smith employs the label Itl. 63 (indicating the
“Italian” version of the augmented sixth chord) for a later passage in his ex.
2009/16, one might have expected something similar between VI and V in his
ex. 2009/19. I can easily convert another analyst’s Itl. 63 or It+6 into my II⇨.
That issue is merely a matter of notation. In contrast, the juxtaposition of VI
and V implies that the intervening F-A-D♯ chord does not play a significant
role in the progression, an interpretation that I reject. In my 15.15 it is one of
only four chords within P that does warrant a harmonic label in my deepest
level of Roman numeral analysis.
30. Because there is no forward trajectory from the C chord of 331, it remains
somewhat indeterminate harmonically. On the one hand, the earlier A–E fifth
might make F–C seem like a local progression from I to V in F Major; on the
other hand, the A and C endpoints of the progression would correspond to I
and III if followed by a motion up to dominant E (which ultimately does
prevail, though via a different harmonic routing). I display the latter reading in
15.16a, here acknowledging the uncommonly speculative nature of that
assertion.
31. Given my reading of the A chord at 412 as the onset of a very notable
backtracking, Smith’s reference to “the consequent phrase of mm. 35–42”
(p. 2009/71) seems wayward. My confusion is compounded by the number
42. His ex. 2009/18 displays the content of that measure as passing between the
structurally deeper chords of measures 41 and 43. Was 41 or 43 intended
(instead of 42) as his consequent phrase’s endpoint?
32. I regard the piano slur from E to A in the bass of measure 41 as an indication of
a legato connection between the downbeat E goal of the section to the A that
sets us back to where we were at 292. It does not imply the same sort of
relationship as a Schenkerian analytical slur. My interpretation contrasts
Smith’s assertion that “Schumann steers the theme back onto an A minor
270 Notes to pages 223–229

path in mm. 35–41, with the result that the first tonal motion of the secondary
area leads to none other than the home tonic!” (p. 2013/89).
33. Failing to acknowledge the E major dominant goals at both measures 41 and
56–57, Smith makes the exposition seem exceptional: “Schumann delays, until
the development, the kind of deep-middleground progression that typically
occurs within the exposition as part of a large-scale dissonance relationship”
(p. 2009/69). In my view what is exceptional is the tonic restoration that follows
the arrival of the dominant goal, with the development section proceeding
from I to V♯ instead of prolonging an already attained V♯. Smith is in
fundamental agreement with this latter point, facilitated by his focus on F
rather than E, since “closure in F corrals this submediant onto a relatively
shallow level of the tonal hierarchy . . . The submediant thus fails to achieve the
relative independence that would be needed for it to form a dissonant pole in
relation to the opening tonic” (p. 2009/69).
34. Laufer, E., “Voice-Leading Procedures in Development Sections,” Studies in
Music from The University of Western Ontario 13 (1991), pp. 69–120. The
figures cited appear on pages 72 and 74, while the quotation is from page 73.
35. Giving voice to Schumann’s thoughts was undertaken also (and admirably) by
Harald Krebs in his Fantasy Pieces: Metrical Dissonance in the Music of Robert
Schumann, Oxford University Press, 1999.
36. Though it may appear that what I am proposing is the equivalent of Smith’s
interpretation of measure 104 from the first movement, discussed in note 11,
above, observe that the B chord was introduced during measure 70, with the
addition of minor seventh A not emerging until measures 76 (piano) and 77
(violin).
37. Schumann’s writing here may be interpreted as a reverie triggered by the
uncanny sounding of the movement’s tonic root, A, in a distinctly non-tonic
context: the upper third of the supertonic of the dominant!
38. I have proposed elsewhere that a composer may be inspired to formulate a
creative reworking of conventional recapitulation procedures to offset the
predictability and long-enduring tonic focus inherent in the standard
practice. See, for example, Harmony in Schubert, chapter 6, and Harmony in
Beethoven, chapter 13.
39. References to seismic shifts can be found throughout my Harmony in . . .
volumes, beginning with Harmony in Schubert, p. 173.
40. The transposition of the exposition’s melodic motive A–A>E to C–C>G puts it
into alignment with the bell motive in Berlioz’s Symphonie fantastique
(movement 5, Songe d’une nuit du sabbat), which proceeds also as C–C>G.
(Schumann had reviewed the symphony in his Neue Zeitschrift für Musik during
1835.) Consequently the motive may have a diabolical implication in
Schumann’s work, though on the other hand the similarity may be coincidental.
Select bibliography

Agawu, V. K., “Structural ‘Highpoints’ in Schumann’s Dichterliebe,” Music

Analysis 3 (1984), pp. 159–180
“Theory and Practice in the Analysis of the Nineteenth-Century Lied,” Music
Analysis 11 (1992), pp. 3–36
“Ambiguity in Tonal Music: A Preliminary Study,” in Theory, Analysis and
Meaning in Music, ed. A. Pople, Cambridge University Press, 1994, pp. 86–107
Music as Discourse: Semiotic Adventures in Romantic Music, Oxford University
Press, 2009
Appel, B. R., Robert Schumanns “Album für die Jugend”: Einführung und
Kommentar, Atlantis Musikbuch-Verlag, 1998; Mainz: Schott, 2010
Vom Einfall zum Werk: Robert Schumanns Schaffensweise, Mainz: Schott, 2010
Bahr, R., “‘Dreiklang = Zeiten. Terz vermittelt Vergangenheit und Zukunft als
Gegenwart’: Kadenzharmonik bei Robert Schumann und Moritz
Hauptmanns harmonische Dialektik,” in Musiktheorie im Kontext: 5.
Kongress der Gesellschaft für Musiktheorie, Hamburg 2005, ed. J. P. Sprick,
R. Bahr, and M. von Troschke, Berlin: Wiedler Buchverlag, 2008, pp. 275–288
Bak, C.-Y., “An Eclectic Analysis and Performance Guide of Robert Schumann’s
First Two Violin Sonatas, Opus 105 in A and Opus 121 in D,” Ph.D. diss., New
York University, 2000
Bartoli, J.-P., L’harmonie classique et romantique (1750–1900): Éléments et
évolution, Paris: Minerve, 2001
Bischoff, B., “Robert Schumanns Klaviertrio g-Moll op. 110: Einstehung –
Rezeption – Analyse,” in Schumann in Düsseldorf: Werke-Texte-
Interpretationen: Bericht über das 3. Internationale Schumann-Symposion
am 15. und 16. Juni 1988 im Rahmen des 3. Schumann-Festes, Düsseldorf,
ed. B. R. Appel, Mainz: Schott, 1993, pp. 221–248
Burkhart, C., “Departures from the Norm in Two Songs from Schumann’s
Liederkreis,” in Schenker Studies, ed. H. Siegel, Cambridge University Press,
1990, pp. 146–164
Burstein, L. P., “Unraveling Schenker’s Concept of the Auxiliary Cadence,” Music
Theory Spectrum 27 (2005), pp. 159–185
Cadwallader, A., “Form and Tonal Process: The Design of Different Structural
Levels,” in Trends in Schenkerian Research, ed. A. Cadwallader, New York:
Schirmer Books, 1990, pp. 1–21

271
272 Select bibliography

Carey, N., “An Improbable Intertwining: An Analysis of Schumann’s Kreisleriana I

and II, with Recommendations for Piano Practice,” Theory and Practice 32
(2007), pp. 19–50
Dommel-Diény, A., L’analyse harmonique en exemples de J.-S. Bach à Debussy:
Contribution à une recherche de l’interprétation: Schumann, Paris: Éditions
Delachaux & Niestlé, 1967
Elias, A., “Zwei Klavierstücke aus Schumanns ‘Album für die Jugend’,” Der
Dreiklang: Monatsschrift für Musik 7 (1937), pp. 161–164
Ferris, D., Schumann’s Eichendorff Liederkreis and the Genre of the Romantic Cycle,
Oxford University Press, 2000
“‘Was will dieses Grau’n bedeuten?’: Schumann’s ‘Zwielicht’ and Daverio’s
‘Incomprehensibility Topos’,” The Journal of Musicology 22 (2005), pp. 131–153
Finson, J. W., Robert Schumann: The Book of Songs, Cambridge, MA: Harvard
University Press, 2007
Forte, A., “Schenker’s Conception of Musical Structure,” Journal of Music Theory 3
(1959), pp. 1–30; reprinted in Readings in Schenker Analysis and Other
Approaches, ed. M. Yeston, New Haven: Yale University Press, 1977, pp. 3–37
Goldenberg, Y., “‘Negative Texture’ and the Prolongation of Seventh Chords,”
Theory and Practice 29 (2004), pp. 95–123
Prolongation of Seventh Chords in Tonal Music, 2 vols., Lewiston, NY: Mellen,
2008
Hatten, R. S., “Performing Expressive Closure in Structurally Open Contexts:
Chopin’s Prelude in A minor and the Last Two Dances of Schumann’s
Davidsbündlertänze,” Music Theory Online 20/4 (2014)
Hoeckner, B., “Paths through Dichterliebe,” 19th-Century Music 30 (2006–2007),
pp. 65–80
Hunt, G., “‘How Much is Enough?’: Structural and Formal Ramifications of the
Abbreviated Second A Section in Rondo Finales from Haydn to Brahms,”
Journal of Schenkerian Studies 8 (2014), pp. 1–48
Kaminsky, P. M., “Aspects of Harmony, Rhythm and Form in Schumann’s
Papillons, Carnaval and Davidsbündlertänze,” Ph.D. diss., University of
Rochester, 1989
“Principles of Formal Structure in Schumann’s Early Piano Cycles,” Music
Theory Spectrum 11 (1989), pp. 207–225
Komar, A., “The Music of Dichterliebe: The Whole and Its Parts,” in Dichterliebe,
by R. Schumann, ed. A. Komar, New York: W. W. Norton, 1971, pp. 63–94
Kopp, D., Chromatic Transformations in Nineteenth-Century Music, Cambridge
University Press, 2002
“Intermediate States of Key in Schumann,” in Rethinking Schumann, ed. R.-M.
Kok and L. Tunbridge, Oxford University Press, 2011, pp. 300–325
Krebs, H., Fantasy Pieces: Metrical Dissonance in the Music of Robert Schumann,
Oxford University Press, 1999
 Select bibliography 273

“Sentences in the Lieder of Robert Schumann: The Relation to the Text,” in

Formal Functions in Perspective: Essays on Musical Form from Haydn to
Adorno, ed. S. Vande Moortele, J. Pedneault-Deslauriers, and N. J. Martin,
University of Rochester Press, 2015, pp. 225–251
Krummacher, F., “Zur Kompositionsart Mendelssohns: Thesen am Beispiel der
Streichquartette,” in Das Problem Mendelssohn, ed. C. Dahlhaus, Regensburg:
Bosse, 1974, pp. 169–184; as “On Mendelssohn’s Compositional Style:
Propositions Based on the Example of the String Quartets,” trans. D.
Seaton, in The Mendelssohn Companion, ed. D. Seaton, Westport, CT:
Greenwood Press, 2001, pp. 551–568
Mendelssohn der Komponist: Studien zur Kammermusik für Streicher, Munich:
Wilhelm Fink, 1978
Laufer, E., “Voice-Leading Procedures in Development Sections,” Studies in Music
from The University of Western Ontario 13 (1991), pp. 69–120
Lerdahl, F., Tonal Pitch Space, Oxford University Press, 2001
Lester, J., “Substance and Illusion in Schumann’s ‘Erinnerung,’ Op. 68: A
Structural Analysis and Pictorial (geistliche) Description,” In Theory Only
4 (1978), pp. 9–17
“Robert Schumann and Sonata Forms,” 19th-Century Music 18 (1994–1995),
pp. 189–210
Lewin, D., Studies in Music with Text, Oxford University Press, 2006
Malin, Y., “Review of Schumann’s Dichterliebe and Early Romantic Poetics:
Fragmentation of Desire by Beate Julia Perrey,” Music Theory Spectrum 28
(2006), pp. 299–310
Marston, N., Schumann, Fantasie, Op. 17, Cambridge University Press, 1992
“Schumann’s ‘Gesänge der Frühe’, op. 133: A ‘False Dawn’?,” in “Le ultime opera
di Robert Schumann: Analisi dei Gesänge der Frühe per pianoforte op. 133
(1853),” ed. A. Rostagno, Rivista di analisi e teoria musicale 16/2 (2010),
pp. 29–49
Martin, N. J., “Schumann’s Fragment,” Indiana Theory Review 28 (2010),
pp. 85–109
McCreless, P., “Song Order in the Song Cycle: Schumann’s Liederkreis, op. 39,”
Music Analysis 5 (1986), pp. 5–28
Meeùs, N., “Robert Schumann: Quatre Lieder du Dichterliebe op. 48,” Analyse
musicale 21 (1990), pp. 23–35
Michaelsen, R., Der komponierte Zweifel: Robert Schumann und die Selbstreflexion
in der Musik, Paderborn: Wilhelm Fink, 2015
Morgan, R. P., “Dissonant Prolongation Again: Nontonic Extensions in
Nineteenth-Century Music,” Journal of Music Theory 60 (2016), pp. 1–21
Moßburger, H., Poetische Harmonik in der Musik Robert Schumanns, Sinzig:
Studio Verlag, 2005
“Poetische Harmonik,” in Schumann Handbuch, ed. U. Tadday, Stuttgart:
Metzler, 2006, pp. 194–211
274 Select bibliography

“Robert Schumanns frühe und späte Fantasien,” in Der späte Schumann, ed. U.
Tadday, Munich: Boorberg, 2006, pp. 51–67.
Ästhetische Harmonielehre: Quellen, Analysen, Aufgaben, 2 vols.,
Wilhelmshaven: Florian Noetzel Verlag, 2012
Neumeyer, D., “Organic Structure and the Song Cycle: Another Look at
Schumann’s Dichterliebe,” Music Theory Spectrum 4 (1982), pp. 92–105
“Synthesis and Association, Structure, and Design in Multi-Movement
Compositions,” in Music Theory in Concept and Practice, ed., J. M. Baker,
D. W. Beach, and J. W. Bernard, University of Rochester Press, 1997,
pp. 197–216
Oinas, C., “Toward a Performer-Oriented Analysis: Communication between
Analysis and Performance in Schumann’s D-minor Piano Trio,” Res musica
3 (2011), pp. 134–145
Perrey, B. J., Schumann’s Dichterliebe and Early Romantic Poetics: Fragmentation
of Desire, Cambridge University Press, 2002
Ponce, A., “Form, Diversity, and Lack of Fulfillment in the First Movement of
Schumann’s Fantasie op. 17,” Music Theory Online 20/3 (2014)
Pousseur, H., Schumann, le poète: Vingt-cinq moments d’une lecture de
Dichterliebe, Paris: Méridiens Klincksieck, 1993
“Schumann ist der Dichter,” Intersections: Canadian Journal of Music 25 (2005),
pp. 7–22
Rapoport, E., Mendelssohn’s Instrumental Music: Structure and Style, Hillsdale,
NY: Pendragon Press, 2012
Reiman, E., Schumann’s Piano Cycles and the Novels of Jean Paul, University of
Rochester Press, 2004
Reti, R., The Thematic Process in Music, New York: Macmillan, 1951
Reynolds, C. A., Motives for Allusion: Context and Content in Nineteenth-Century
Music, Cambridge, MA: Harvard University Press, 2003
Roesner, L. C., “Schumanns ‘Parallel’ Forms,” 19th-Century Music 14 (1990–1991),
pp. 265–278
Rosen, C., The Romantic Generation, Cambridge, MA: Harvard University Press,
1995
Rothgeb, J., “A Structural Function of Dynamics in Schumann’s ‘Ich grolle nicht’,”
In Theory Only 5/2 (1979), pp. 15–17
Rothstein, W., Phrase Rhythm in Tonal Music, New York: Schirmer Books, 1989
Salzer, F., Structural Hearing: Tonal Coherence in Music, 2 vols., New York: Boni,
1952; reprint edn, New York: Dover, 1962
Schachter, C., “Rhythm and Linear Analysis: Aspects of Meter,” Music Forum
6/1 (1987), pp. 1–59; reprinted in Unfoldings: Essays in Schenkerian Theory
and Analysis, ed. J. N. Straus, Oxford University Press, 1999, pp. 79–117
“The Triad as Place and Action,” Music Theory Spectrum 17 (1995), pp. 149–169;
reprinted in Unfoldings: Essays in Schenkerian Theory and Analysis, ed. J. N.
Straus, Oxford University Press, 1999, pp. 161–183
 Select bibliography 275

“Che Inganno! The Analysis of Deceptive Cadences,” in Essays from the Third
International Schenker Symposium, ed. A. Cadwallader, Hildesheim: Georg
Olms, 2006, pp. 279–298
Schenker, H., Der Tonwille: Flugblätter zum Zeugnis unwandelbarer Gesetze der
Tonkunst, einer neuen Jugend dargebracht, 2 vols., Vienna: A. J. Gutmann,
Leipzig: F. Hofmeister, 1921–1924; as Der Tonwille: Pamphlets in Witness of
the Immutable Laws of Music, Offered to a New Generation of Youth, ed. W.
Drabkin, trans. I. Bent et al., Oxford University Press, 2004–2005
Der freie Satz: Neue musikalische Theorien und Phantasien III, Vienna:
Universal, 1935; rev. edn, ed. O. Jonas, Vienna: Universal, 1956; as Free
Composition, trans. and ed. E. Oster, New York: Longman, 1979; reprint
trans. Stuyvesant, NY: Pendragon Press, 2001
The Oster Collection: Papers of Heinrich Schenker, New York Public Library,
Lincoln Center, New York
Schmalfeldt, J., In the Process of Becoming: Analytic and Philosophical Perspectives
on Form in Early Nineteenth-Century Music, Oxford University Press, 2011
Smith, P. H., “Harmonies Heard from Afar: Tonal Pairing, Formal Design, and
Cyclical Integration in Schumann’s A-minor Violin Sonata, op. 105,” Theory
and Practice 34 (2009), pp. 47–86
“Associative Harmony, Tonal Pairing, and Middleground Structure in
Schumann’s Sonata Expositions: The Role of the Mediant in the First
Movements of the Piano Quintet, Piano Quartet, and Rhenish Symphony,”
in Rethinking Schumann, ed. R.-M. Kok and L. Tunbridge, Oxford University
Press, 2011, pp. 235–264
“The Drama of Tonal Pairing in Chamber Music of Schumann and Brahms,” in
Expressive Intersections in Brahms: Essays in Analysis and Meaning, ed. H.
Platt and P. H. Smith, Bloomington: Indiana University Press, 2011,
pp. 252–290
“Tonal Pairing and Monotonality in Instrumental Forms of Beethoven,
Schubert, Schumann, and Brahms,” Music Theory Spectrum 35 (2013),
pp. 77–102
“Schumann’s Continuous Expositions and the Classical Tradition,” Journal of
Music Theory 58 (2014), pp. 25–56
“Hausmusik for Cognoscenti: Some Formal Characteristics of Schumann’s Late-
Period Character Pieces for Instrumental Ensemble,” Music Theory Spectrum
37 (2015), pp. 51–70
“Schumann’s A-minor Mood: Late-Style Dialectics in the First Movement of the
Cello Concerto,” Journal of Music Theory 60 (2016), 51–88
Stein, D., “Introduction to Musical Ambiguity,” in Engaging Music: Essays in Music
Analysis, ed. D. Stein, Oxford University Press, 2004, pp. 77–88
Struck, M., “Beziehungs-Probleme: Zum Verhältnis der Komponisten Schumann
und Brahms, dargestellt am Beispiel von Violinsonaten,” in “Neue Bahnen”:
Robert Schumann und seine musikalischen Zeitgenossen: Bericht über das 6.
276 Select bibliography

Internationale Schumann-Symposion am 5. und 6. Juni 1997 im Rahmen des 6.

Schumann-Festes, Düsseldorf, ed. B. R. Appel, Mainz: Schott, 2002,
pp. 294–347
Suurpää, L., “Schumann, Heine, and Romantic Irony: Music and Poems in the First
Five Songs of Dichterliebe,” Intégral 10 (1996), pp. 93–123
“Cover Tone in Brahms’s Immer leiser wird mein Schlummer op. 105, no. 2 and
Mendelssohn’s Lied ohne Worte op. 67, no 1,” in A Composition as a Problem
III: Proceedings of the 3rd International Conference on Music Theory, Tallinn,
March 9–10, 2001, Tallinn: Estonian Academy of Music, 2003, pp. 109–121
“The Undivided Ursatz and the Omission of the Tonic Stufe at the Beginning of
the Recapitulation,” Journal of Schenkerian Studies 1 (2005), pp. 66–91
“Schumann’s ‘Das ist ein Flöten und Geigen’: Conflicts between Local and
Global Perspectives,” in Structure and Meaning in Tonal Music: Festschrift
in Honor of Carl Schachter, ed. L. P. Burstein and D. Gagne, Hillsdale, NY:
Pendragon Press, 2006, pp. 173–178
“The Fourth Piece of Schumann’s Kreisleriana, Op. 16, As a Musical Fragment:
Discontinuity and Unity Intertwined,” Rivista di analisi e teoria musicale 19/1
(2013), pp. 7–24
Taylor, B., Mendelssohn, Time and Memory, Cambridge University Press, 2011
Thomas, M., Das Instrumentalwerk Felix Mendelssohn-Bartholdys: Eine
systematisch-theoretische Untersuchung unter besonderer Berücksichtigung
der zeitgenössischen Musiktheorie, Göttingen: Hubert, 1972
Tunbridge, L., Schumann’s Late Style, Cambridge University Press, 2007
Vitercik, G., “Mendelssohn the Progressive,” Journal of Musicological Research
8 (1988–1989), pp. 333–374
The Early Works of Felix Mendelssohn: A Study in the Romantic Sonata Style,
Philadelphia: Gordon and Breach, 1992
“Mendelssohn as Progressive,” in The Cambridge Companion to Mendelssohn,
ed. P. Mercer-Taylor, Cambridge University Press, 2004, pp. 71–88
Wadsworth, B. K., “Directional Tonality in Schumann’s Early Works,” Music
Theory Online 18/4 (2012)
Waldura, M., Monomotivik, Sequenz und Sonatenform im Werk Robert
Schumanns, Saarbrücken: Saarbrücker Druckerei & Verlag, 1990
Webster, J., “Ambivalenzen um Mendelssohn: Zwischen Werk und Rezeption,” in
Felix Mendelssohn Bartholdy: Kongreß-Bericht Berlin 1994, ed. C. M. Schmidt,
Wiesbaden: Breitkopf & Härtel, 1997, pp. 257–278
Index of Mendelssohn’s works

Incidental Music to A Midsummer Night’s op. 53, no. 1 (A♭ Major), 44–46,
Dream 136–145
“Wedding March”, 143 op. 53, no. 5 (A Minor), 5–7
Octet in E♭ Major (op. 20), 81–115 op. 53, no. 6 (A Major), 19–20, 52–54
Songs without Words op. 62, no. 3 (E Minor), 11–12
op. 19, no. 6 (G Minor), 27 op. 62, no. 5 (A Minor), 71–72
op. 30, no. 4 (B Minor), 48–49 op. 62, no. 6 (A Major), 3–4
op. 30, no. 6 (F♯ Minor), 9–10 op. 67, no. 6 (E Major), 58–59
op. 38, no. 1 (E♭ Major), 17, 66–68 op. 85, no. 1 (F Major), 116–126, 135
op. 38, no. 2 (C Minor), 56–58, 64–66 op. 85, no. 4 (D Major), 35–36
op. 38, no. 3 (A Major), 40–41 op. 85, no. 6 (B♭ Major), 43–44
op. 38, no. 5 (A Minor), 61–62 op. 102, no. 1 (E Minor), 69–70
op. 38, no. 6 (A♭ Major), 36–38 op. 102, no. 5 (A Major), 38–39

277
 Index of Schumann’s works

Dichterliebe (op. 48) no. 15 (“Aus alten Märchen”), 22–23

no. 1 (“Im wunderschönen Monat Mai”), no. 16 (“Die alten bösen Lieder”), 7–8, 16
171–182 Liederkreis (op. 24)
no. 2 (“Aus meinen Tränen sprießen”), 181 no. 2 (“Es treibt mich hin”), 18, 24
no. 3 (“Die Rose, die Lilie, die Taube, die no. 3 (“Ich wandelte unter den Bäumen”),
Sonne”), 55–56 31–33
no. 5 (“Ich will meine Seele tauchen”), no. 7 (“Berg’ und Burgen schau’n herunter”),
14–15 33–34
no. 7 (“Ich grolle nicht”), 29–30 Liederkreis (op. 39)
no. 8 (“Und wüssten’s die Blumen”), 51–52, no. 5 (“Mondnacht”), 146–154
73–75 no. 6 (“Schöne Fremde”), 154–162
no. 11 (“Ein Jüngling liebt ein Mädchen”), no. 10 (“Zwielicht”), 162–170
21–22 Phantasiestücke (op. 12)
no. 12 (“Am leuchtenden Sommermorgen”), “Aufschwung”, 129
28–29, 62–64 “Grillen”, 129
no. 13 (“Ich hab’ im Traum geweinet”), “Warum?”, 127–135
75–78 Sonata in A Minor for Violin and Piano
no. 14 (“Allnächtlich im Traume”), 46–47 (op. 105), 183–231

278
Index of names and concepts

applied dominant, 16 hypermeter, 149–150, 152, 153, 260

assertion, 8, 148
augmented sixth chords, 164, 205, 231, 254, idiosyncratic progression, 120
255, 256, 258, 269; see also supersurge interruption, 111, 125, 146, 154, 196
auxiliary cadence, 127, 129, 141,
148, 173, 261 Komar, A., 264
Kopp, D., 263
Bach, C. P. E., 172
back-relating dominant, 140 Laufer, E., 223, 267
Beethoven, L. v., 104, 129 Lester, J., 266
Berlioz, H., 270 lower-third chord, 198, 200, 215
Burkhart, C., 146–170 lowered supertonic see bII
Burstein, L. P., 127–135
Martin, N. J., 261
cadential 64 , 8 McCreless, P., 260
Cadwallader, A., 116–126 Mehrdeutigkeit, 53, 83, 106
chordal evolution, 16, 88, 177 medial caesura, 91
chordal ninth, 50–59 mediant, 24, 69–78, 121–122
chromatic variant, 215 multiple meaning see Mehrdeutigkeit
circle of fifths, 69–70, 75, 92, 114, 142,
198–202, 212, 220, 229, 254 Neapolitan sixth see ♭II
collision, 114, 123, 125, 128, 131, 147, 159, 192 Neumeyer, D., 264
common-tone diminished seventh
chord, 106, 165 Oster Collection, 133

Darcy, W., 81 parallel fifths, 141

diminished seventh chord, 104–106 parenthetical passage, 208
pedal point, 53
elision, 97, 115, 218, 220 Perrey, B. J., 262–263
embellishing chord, 5–6, 156
enharmonic equivalence, 92, 114, 115 Rapoport, E., 255
reaching-over, 47, 189
Ferris, D., 146–170, 264 registral shift, 184
Finson, J. W., 263–264 Rosen, C., 260, 262
Forte, A., 264
Schenker, H., 118, 133, 139, 173, 182, 261
Goldenberg, Y., 136–145 Schmalfeldt, J., 255, 257, 260
seismic shift, 167–168, 227–230
Heine, H., 179 sequence, 198
Hepokoski, J., 81 Smith, P. H., 183–231

279
280 Index of names and concepts

Stein, D., 171–182 unfurling, 19, 90

Struck, M., 267 upper-third chord, 69, 113,
supersurge, 60, 61–64, 91–92, 93, 95, 134, 224, 178, 180, 198, 200, 225,
256; see also augmented sixth chords 230, 270
surge, 16–23, 42–59
Suurpää, L., 262–263 Vitercik, G., 81–115
Swoboda, A., 260
wobbly note, 63, 67, 249
Taylor, B., 81–115 5–6 shift (5- and 6-phase chords), 24–41,
TMS complex, 183, 223 42–49, 96, 118
tonicization, 115, 119–120 ♭II, 60, 64–68, 73–75