MTS.

Reviews_pp131-174

5/11/04

11:06 AM

Page 149

reviews
Dave Headlam. The Music of Alban Berg. New Haven: Yale
University Press, 1996.

contain a wealth of diagrams, examples, figures, and tables

that, in addition to the authors astute observations, give
readers a leg up in their own studies of this music; his references to the existing analytical literature and Bergs own
theoretical ideas and analytical commentaries are especially
helpful. Unifying threads for understanding Bergs entire
oeuvre are symmetry and cyclic interpretations of musical
materials, for which Headlam freely acknowledges his debt
to George Perle (61). Preceding these chapters is a fine introduction that places his work into the context of Berg scholarship as a whole. Yet, despite the books admirable clarity in
layout, its theoretical and analytical claims suffer from imprecise definitions. As we shall see, Headlam makes assumptions about tonality and cyclic construction in particular that
tend to obfuscate many of his often engaging analytical points.

reviewed by richard hermann

Dave Headlams study of Bergs music is the successor to

Douglas Jarmans identically titled book published in 1983.
At 460 pages Headlams text is more detailed than Jarmans
and incorporates much of the important scholarship on Berg
and post-tonal theory that has appeared in the intervening
thirteen years. Headlam aims for a wider readership than
professional music theorists and defines his terms accordingly (x).1 He claims that Bergs music warrants a broad set
of critical and analytical approaches: Bergs music is . . .
susceptible to both a modernist approach, treating each piece
as a self-contained musical entity, and to a postmodernist
approach, searching for meaning in the works symbols and
references, history and context, both past and present, and
the effect on the listener (910). In the end, Headlams
principal aim is to contribute to the continuing study of
Bergs music (2), and does so through recourse to modernist
theoretical tools.
The main argument of The Music of Alban Berg is laid out
in five chapters. Three analytical chapters are each devoted
to a different musical period: tonal, atonal, and serial.
(Headlam defines these periods on the basis of musical resources rather than mere chronology, though of course to
some extent the two intersect.) These analytical chapters alternate with two chapters that introduce the theoretical ideas
needed for the analytical discussions. The analytical chapters

the tonal period

The first analytical chapter addresses Bergs tonal compositions, defined by Headlam as the early student works
through the first three songs of op. 2. Unfortunately, this is
the only analytical chapter not preceded by an introductory
theoretical chapter. Headlam simply assumes that Bergs
early pieces are tonal, a question very much in dispute among
other Berg scholars.2 On the surface, this music may seem
tonal: Berg uses key signatures after all. But the use of roving harmonies (to borrow Schoenbergs term), non-triadic
simultaneities, and successive unresolved dissonances points
to a more extended tonality or even atonality. Headlam
chooses to pass over this controversial issue in silence. He
also chooses not to consider Bergs own thoughts on tonality
or those of his mentor, but instead employs without rationale
2

Right after the acknowledgments follows A Note on Terminology,

a brief three-page section (ixxi) which tacitly assumes readers have
taken an introductory course in post-tonal theory.

149

Janet Schmalfeldt (1991) claims that op. 1 is tonal. Yet Joseph N. Straus
writes of the second song from the Four Songs, opus 2: Let us put
aside thoughts of E ! minor and see how the music is organized (1990,
84).

MTS.Reviews_pp131-174

150

5/11/04

11:06 AM

Page 150

music theory spectrum 26 (2004)

this ! II. How can such a harmonic retrogression (an outof-order, even reversed, succession of harmonies) be tonally
confirming in such a highly chromatic context? Should
Headlam claim that this phrygian II is an upper fifth that
expands the V, the idea founders on the distance of six semitones in pitch-class space between the two chordal roots.
These two Grundtne do not occupy the same diatonic
(home key) or harmonic-series space and hence can not
define an upper fifth relationship.5 As noted, these chords
contain unresolved dissonances and are out of grammatical
order at the surface. Perhaps these observations underlie my
inability to hear these moments in the manner Headlam
specifies. Example 2 further shows the next structural harmony to be V at m. 54 of the score. Vs root, however, occurs in what from a Schenkerian perspective is an inner voice
and is rhythmically syncopated within a descending sequential pattern, placing its status as controlling harmonic event
into question. Unfortunately, Headlam gives no explanation
for his choices.
Such analytical problems are not isolated. Consider
Headlams discussion of op. 2, no. 3, the score for which is
given as Example 3. In op. 2, no. 3, the principal harmonic
and scale-degree relationships are around notes {A !,B #,F !,E !},
or the conventional nineteenth-century emphasis on scale
degrees 1, ! 2, ! 6, 5; these scale-degree relationships are reflected harmonically in motives that can be described in setclass terms as members of [01], [016], and [0156] (34). On
what theoretical basis may we invoke set-classes in a tonal
analysis? Let us briefly examine these assertions. A in the
bass at m. 5, for instance, can assume a dominant function in
relation to the D-minor triad on the following downbeat;
two voice-leading chords occur in between this D-minor

the theories of Heinrich Schenker, which were not built to

elucidate highly extended tonality much less atonality.3
Since Headlam does not directly concern himself with
considerations of analytical method, the reader surmises that
his actions are at least partly justified by his assertion that
proof of any theoretical approach is, of course, in the analyses that result (65).4 With this comment in mind, let us
consider his analyses of the Piano Sonata in B minor, op. 1
(190708), and the song Nun ich der Riesen Strksten
berwand, op. 2, no. 3 (190910).
Headlam writes of op. 1: The closing theme of the exposition, shown in Example 1, leads toward I with a motion
through V and ! II, that is, F " and C (mm. 4955) (23). ! II,
in other words, occurs after Vwhich is odd, since ! II normally prepares V. This chord, at m. 51, is also unusual in
terms of its (tonal) intervallic construction: {C2,D4,G !4,B !4,
D5}. As ! II, this simultaneity lacks a chordal third, E, and includes an altered (lowered) fifth, G !. Headlams phrygian II
is also joined by a dissonant seventh (B !) and ninth (D). D
resolves to E !, a mixed chordal third, on the second beat of
the measure, but B ! is transferred up an octave and left unresolved. Meanwhile, the bass arpeggiates !C2,F "2,B2", which
the tenor follows canonically with !C "3,G3,C4 ". In short,
notwithstanding notions of mixture and of being out of context, the chord-type could also strongly suggest either a ii7
function in B ! minor or a vii7 function in D ! major. Why,
then, phrygian II? The latter part of Headlams Schenkerian
graph of the exposition is reproduced as Example 2. Here a
structural V supporting 2 in the soprano at 46 progresses to
3

Why not use one of the tonal theories of Bergs contemporaries, such as
Schoenberg, Riemann, Louis and Thuille, or even of subsequent theorists, such as Hindemith, or Lerdahl and Jackendoff, among many? Of
course, these theories and teachings have differing purposes, but nevertheless remain viable to the task at hand. Not to consider them or the
larger theoretical issues concerning the choice of Schenker is curious.
At least since Popper (1962), theories are understood to be only supported or refuted and can not be proven by empirical (here, analytical)
results.

On the upper fifth chord and back-related V, see Cadwallader and

Gagn (1998, 180, 420 n. 47) and Aldwell and Schachter (1989, 145,
2356, 417). In neither source is the upper fifth used to support the
substitute transformation of ! II for II. The back-related V derives from
Schenkers comments on the bertragende Teiler, which Ernst Oster
translates as applied divider in 279 of Free Composition (1979).

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 151

reviews

42

151

dimin.
e
n

m m m m
[[ breiter
n
nn n
n

,
,

n
*
e

accel.

Viel langsamer (Quasi adagio)

ritard.

46

\\
r.H.

*
*
*

*
*
*
l.H. *

*

51

(Tempo I)

( )

(( )) *
l.H. n
\

l l

dimin.

poco accel.

example 1. Berg, Piano Sonate op. 1, score excerpts.

triad and its reiteration on the second beat of m. 8, which

coincides with the end of a textual unit. The arrival on D
minor, in other words, would appear to be a salient, structural event in the song. On what groundstonal or
atonalshould A (as B #), then, be included in Headlams
set of principle harmonic and scale-degree relationships
when D is excluded? Headlam goes on to say about the re-

mainder of the song: Tonally, the guide leads the narrator

back to the home tonic (A !) from the point of furthest removethe tritone or the darkest land (D as IV of ! II). The
ending in sleep is inconclusive, however, signified by the
closing E ! sonority, V of A ! (43). How can the D minor
chord bounding the suddenly denser, soft left-hand syncopated chords of mm. 68 function as IV of ! II, when no A

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 152

152

music theory spectrum 26 (2004)

[29]

[33]

wt

4
2

(D:) V

[II

V]

II

[V]

V]

II

II

B : [VII]

(3)

I6

( )

B: II7

[44]

[52]

[VI III]

wt

[VI

[49]

5-cycle

[41]

[38]

[36]

[55]

wt

(5)

,g

II7

example 2. Headlams Graph of Berg, op. 1, Exposition.

major or minor chord, enharmonically equivalent to B #

(major or minor), appears afterwards, not to mention the
fact that its only appearance is as part of a progression functioning as V leading to that D minor chord?

In Headlams words: The secondary key area is ! II, B #

respelled enharmonically as A, achieved by reinterpreting F !
( ! VI of A !) as V of ! II (B #) (43). This F ! to E7, a V7-type
chord, occurs between mm. 34, but the harmonic progres-

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 153

reviews

153

II.
Erst ziemlich bewegt, dann langsam

[
/ , , ,
0
3

*) Nun ich der Rie-

/
0

-fand

molto ritard.

/ / / // ,
/ / / *

*) an ei- ner wei-en

*
dimin. e molto ritard.

sten

u-

ber-

wand,

,
nn
[
n


*
n

/0

Stark- -

\\ langsam
0
* 0

cresc.

Mar- chen-hand,

n
l.H.
n

* \\ r.H.
,


*l.H.

hal-

7777777777

(riten.)

, ,

* *

*
sen

Alban Berg, op. 2, no. 3

rit.

a tempo

00

len schwer

cresc. (molto)

r.H.
,

*

heim-

n
n

00
*
die

- kel-sten Land

n
,

n *

,
* *

mich aus dem dun-

,

][

r.H.

r.H.

*)

// 0
00 0

Glok-

n
00

, 0 ,
0

*
*

ken;

, /
* 0
3

*)

und ich

*
*
r.H. (poco accel.)
,


n n
* n
[ espress. n n
3

/0
/0

*) Diese Stellen nicht hastig, sondern im Tempo des gesprochenen Wortes.

example 3. Score of Berg, Four Songs, op. 2, no. 3.

sion in question is functionally out of order. It should be IV
(D minor with mixture) to V 7 (above E) to I in A major
(or minor). Instead, we have E7 in m. 4 to A7 in m. 5 to
D minor, also in m. 5. Other factors emphasize D minor,
including changes of density, dynamics, and durational

patterning. The local D minor tonic is approached by its

V 7, which is in turn preceded by its own applied V 7; both
applied dominants are constructed, spelled, and resolved in
traditionally tonal ways. Headlam acknowledges this reading: The circle-of-fifths progression from aligned ascending

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 154

154

music theory spectrum 26 (2004)

/0 *

wan-

\molto dimin.rit.
n
/
/
/ / 0 * ,
0
ke

durch die Gas- sen

n
n n rit.
/
0

l l l
l l l
([ )
(\ )
,
/

0
]\espress.
abdampfen
molto

,
*

schlaf-

be-

dimin.

,

m
\\

n
n

example 3. [continued]

The !E,A,D" pitch-class segment from the ascending 5-cycle would

model the harmonic roots of the progression, and the !G " ,G,(F " ),F"
segment from the descending 1-cycle would simultaneously model
an upper voice in that progression. The F " in parentheses is, of course,
neither present in nor implied by the progression.
For an exposition of " IVs tonal status in Schenkerian thought, see
Brown, Dempster, and Headlam 1997.

\\\

tion of dissonance into changes of chords, and extended

dominants by cyclic voice-leading and collections. In all
three songs, tonic chords are expressed with dominant characteristics and thus have an ambiguous status as . . . in op. 2,
no. 3, A ! as I or IV of E ! and, in the area of ! II, A (B #) as I
or V of IV in D (E #) (34). How can a piece be tonal, no
matter how extended, when its main harmonies are never
secure in their most elemental functions?8
Earlier Headlam was right to invoke set-class entities in
this song, as I will, though for different reasons. The ambiguous functional status of important harmonies and the
problematic tritone tonal relation in this song can be explained post-tonally. My counter analysis examines the bass
line, then the harmonies, and, finally, their interactions.

5- and descending 1-cycles may be interpreted as [V]VI

in D minor, but given the tonal context a preferable alternate
reading is VI arriving as V/IVIV (in ! II) (44).6 How can
we understand this as Headlam suggests, when A itself is
never established as a clear tonal region before or after this
progression? He himself notes that . . . rather than E # or
D as " IV of A !, a tonal situation which is untenable (44).
But Headlam does not tell us why.7
Both Headlams analysis of this song and my discussion
of his analysis shift from a hierarchical perspective to a more
surface-oriented discussion of harmonic grammar. Headlam
seems to recognize the shift as follows: The first three songs
in op. 2 are tonal, in characteristic late nineteenth-century
style, with rhythmic displacements of consonances, resolu6

fan- gen.

,



n

\\
n
,
*

Bergs use of these triads would correspond to the last of Schoenbergs

(1978, 153) four ways of describing the nature of chords and their ways
of relating to a tonic as follows: The harmony is nowhere disposed to
allow a tonic to assert its authority. Structures are created whose laws do
not seem to issue from a central source (Zentrum); at least this central
source is not a single fundamental tone.

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 155

reviews
I interpret the main features of the bass line as a global
projection of an ordered abstract subset !A ! 2 (m. 1),D2 (mm.
68),E !2 (m. 1112)" of the opening !A ! 2,C3,G2,D ! 2" bass
line (octave doublings are omitted). These are the beginning,
middle, and ending pitches of the songs bass line. NB: Here
the beginning bass pitch of this global member of set-class
3-5[016] is also the first of the local opening bass-note
superset. The global middle pitch D2 is preceded locally by
A2 and is followed by A !2 ; that is, D2 occupies the medial
position here, too, as another member of 3-5. The globally
ending E !2 is preceded by a B !2 and an E2, yet another member of 3-5. These observations can be formulated in terms of
Klumpenhouwer networks, henceforth k-nets (see Lewin
1990). Example 4 displays these k-nets. Here, LBLX stands
for local bass-line k-net of ordinal position X, and GBL
stands for global bass-line k-net; these represent the ordered
sets mentioned above. The circled pitch classes of the
LBLXs represent pitch classes in the beginning, middle, and
end positions constituting the GBL as previously described.
The k-net of Example 4(f ) represents transformations among
the transformations between the global and local k-nets. My
interpretative choices in both this k-net and that of Example
4(d) reflect the medial emphasis created by the extended
stylistic quotation of tonal progression (one-third of the
songs duration), represented here by LBL2. Note that the
transformational pattern is the same between that of the emphasized medial LBL2 to the boundary position LBLs as
that of GBL to the boundary LBLs. In addition, the medial
LBL2 shares the same transformations as GBL.
Taking Headlams interpretation of the initial triad as
that of A ! minor along with the medial D-minor triad and
the closing E ! -major triad, Example 5 presents the triadic knets for the song in a similar manner as before. (LTX stands
for local triad of ordinal position X.) Just as with Headlams
tonal interpretation, there appears to be no path to a next hierarchical level yet more distant from the surface. The GBL
is the key to understanding the organization of the triadic
k-nets at that next level. Example 6 displays the transforma-

155
tional paths between the bass line and triadic k-nets. As seen
in Example 6(d), the boundary LTs share the same transformational configuration with GBL, and the remaining transformational paths have been commonplace in the various
k-net structures for the song. The triads that accompany
GBLA ! major (m. 1), D minor (mm. 68), and E ! major
(mm. 1112)can thus be understood as undermining conventional harmony by organizing its prime icon, the triad,
in coherent but non-tonal ways. Headlams symbolic interpretation linking events of the text with the tritone (furthest
tonal distance away) between the opening and medial triads
also fits well with this analysis.
Clearly, op. 2, no. 3 pushes Schenkers monotonal theory
beyond the systems explanatory bounds: witness the unresolved ambiguity of harmonic function, the large-scale tritone root progression from the tonic (resulting in a rather
forced interpretation of the preceding surface progression),
and the resulting collapse of tonal hierarchy. I will not even
touch upon the difficulties involved in explaining dissonance
usage. By comparison, the post-tonal analysis puts no strain
whatsoever on post-tonal theory. Instead of focusing on the
problematic tritone progression and the ambiguity of a
theoretically-strained Schenkerian analysis, a k-net analysis
emphasizes the commonalities of T7 in all triadic and bassline k-nets along with homologous commonalities between
GBL and LTXs transformations (of transformations). T7 relates the framing boundary pitch classes of the triad organizing GBL about the songs weighted center. As we now
have an efficient explanation encompassing this songs redeployment of traditionally tonal materials, we may ask: Is this
music tonal or does it simply refer to that system within a
larger post-tonal context?
the atonal and serial periods
Headlams discussion of Bergs non-tonal music develops
certain new theories and reorients some past approaches,
particularly Perles work on cycles and symmetry (545).

MTS.Reviews_pp131-174

5/11/04

11:06 AM

156

Page 156

music theory spectrum 26 (2004)

8

I3
LBL1
(mm. 12)

T7

(a)

T7

(c)
3

I5

!1, 11"
LBL1

LBL3

LBL2
I10

(d)
(e)
GBL

LBL1

!1, 0"

!1, 9"

LBL3

LBL2
!1, 10"

!1, 9"

(f )
example 4. Bass line K-nets for Berg, op. 2, no. 3.

T7

GBL

!1, 9"

!1, 10"

T7

I7

(b)

!1, 10"

LBL3
(mm. 1012)

I10

I2

LBL2
(mm. 58)

I8

I5

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 157

reviews
A ! m
I2

I2
T7

I7

(a)

I5

LT2
(mm. 68)

I7

E ! M

Dm

LT1
(mm. 14)

157

T7

(b)

LT3
(mm. 1112)

I10

T7
3

(c)

!1, 3"
LT1

LT3
!1, 3"

!1, 0"
LT2
(d)

example 5. Triadic K-nets for Berg, op. 2, no. 3.

This reorientation employs even-interval cycles, odd-interval
cycles, wedges, and motives. The importance of whole-tone
plus collections have long been recognized in the literature
on Berg; these collections contain all even or all odd pitch
classes, save for one member. Naturally, Headlam relates
these to even-interval cycles. He develops three forms of
pitch-class set structures from the odd-interval collections,
which may be defined as those with only odd intervals between adjacent notes, although these collections are much
less distinctive as TnI-classes, excluding only the whole-tone
and whole-tone+ collections (72). Collections from odd-

interval cycles appear in three forms: (1) as cyclic segments

. . . ; (2) as cyclic segments plus an added note . . . ; and (3) as
gapped collections, cyclic segments with one or possibly two
gaps in the cycle (73). He adds that because of their mix of
even and odd intervals, odd-interval-based collections lack
the harmonic distinctiveness of whole-tone collections. Although 3-cycle collections are sufficiently restricted in content to be recognizable even when the cyclic interval does
not appear between all adjacent elements, collections from
1/B and 5/7 cycles rely on the presence of the cyclic interval
for definition in their gapped and larger forms (73).

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 158

158

music theory spectrum 26 (2004)

I2

A ! m

LT1
(mm. 14)

!1, 1"

LBL1
(mm. 12)

T7
8

I7

I3

T7
1

I8
!1, 11"

(a)
Dm
I2

LT2
(mm. 68)

5
I7

!1, 3"

I5
8

T7
2

LBL2
(mm. 58)
I10

T7
2

!1, 9"
(b)
example 6. Bass Line and Triadic K-nets for Berg, op. 2, no. 3.
Moreover, these observations are not always sufficient to
decide whether a segment is a cycle+ collection or a
gapped-cycle collection. For instance, Example 3.1d on
page 69 (not reproduced here) labels the !A2,D3,E3,G3,B ! 3"
whole-tone chord from Act 1 of Wozzeck (m. 330) as a 5cycle+ collection. This pitch set could also be identified,
however, as a 5-cycle gapped collection in pitch-class space,
or !E,A,D,G,,B ! ", with C and F constituting the omitted
gap pitch classes. His labeling is of added importance because Headlam posits a dissonant status for non-cyclic

members of cyclic based collections.9 Perhaps the deciding

factor could be Headlams suggestion that the dissonant status of some pitch classes rely on the presence of the cyclic
9

The questions of consonance, dissonance, and embellishing tones in

Bergs music hinge on the dichotomy between the underlying cyclic
collections and transformations and the surface collections, which contain cyclic and non-cyclic elements. The problem is similar to that in
late tonal music, where the functional basis of the system is often difficult to separate out or decipher from the surface. For the most part, it is
possible to distinguish cyclic-based chord and non-chord tones and

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 159

reviews

I5
7

E ! m

159

!1, 9"

I2

LT3
T7
(mm. 1112)
I10

LBL3
T7
(mm. 1012)

I7

!1, 3"
(c)
GBL
!1, 9"

!1, 0"
LT1

!1, 0"
LT3

LT2
!1, 9"

!1, 0"
(d)

example 6. [continued]
interval for definition in their gapped and larger forms (73).
But where is this presence manifested? In pitch space
measured from low to high, the adjacent ordered intervals of
their dissonance-consonance relationships in Bergs music. The cyclicbased collections present no hierarchy or distinction between harmony
and voice leading, however, except where established contextually. In
my view, the cyclic collections in Bergs atonal music are referential, and
are the basis of the pitch language, but they are not prolonged in a tonal
sense. Cyclic collections are quickly superseded, are not in force in their
absence, and require constant reiteration for their continuing referential
status. Thus, I do not posit large-scale cyclic collections comprised of
largely non-adjacent notes spanning a piece or large sections (634).

this chord are !5,2,3,3", which suggests a 3-cycle at least as

much as a 5-cycle derivation. In pitch-class space, this
chords interval-class vector is !122131", its set class is
5-29[01368], and its trichordal and tetrachordal vectors of
abstract inclusion respectively are !010110212110" and
!00000000000011010000001000100".10 Abstract subsets of
cardinalities 2 through 4 that hold the maximum number of
ic 5 are set-classes 2-5[05], 3-9[027], and 4-23[0257]. Note
that set-class 2-5 is the most numerous dyadic abstract
10

Vectors are taken from Morris 1991. These correspond to results

yielded by employing Lewins embedding function (Lewin 1977, 197).

MTS.Reviews_pp131-174

160

5/11/04

11:06 AM

Page 160

music theory spectrum 26 (2004)

subset; set-class 3-9 is tied for the most numerous with setclass 3-7; and all tetrachordal abstract subsets are represented equally. Thus, this chord presents a mixed case for
pitch-class space presence of the 5-cycle. By implication, the
authors following statement clarifies this chords status: In
Bergs atonal music, passages of an expository nature are
characterized by distinctive cyclic materials in clear presentations; however, long dissonant passages also occur in which
distinctions between materials are not as clear. This dichotomy is reminiscent of late nineteenth-century music,
which although tonal, contains long passages of extended
dissonance and harmonic ambiguity (64). Certainly, this
chord is dissonant with regard to the 5-cycle, but theoretically it is better characterized as ambiguous with regard to
cyclic structure. This explanation undercuts the rhetorical
structure that is metaphorically based on consonance and
dissonance, and on harmonies whose function are ambiguous. It is unfortunate that Headlam chooses this particular
chord to illustrate the foundational principles for the study
of this style period. His example 3.2(b) (not provided), a reduction of mm. 372373 from Act 1 of Wozzeck, would have
been clearer. This is one of many passages that convincingly
demonstrate the analytical relevance of the cycles.
Headlams gap theory also is problematic. It features reflexivity, symmetry, and transitivity; thus, it is a similarity relation. As such, it suffers from the various problems that
similarity relations face: pitch space and other realization
factors can undermine perception and even lead to counterintuitive results. In addition, its lack of development creates
theoretical problems too.11
Also problematic are the numerous analogies to tonal
practice (such as cadence, resolution, and meter) and
the application of standard atonal techniques. I will consider
several instances of each in turn.
11

For instance, a simultaneous pitch-space realization of set-class

6-20[014589] (two gapped but in each case with two missing pitch
classes per gap) as !E3,A3,C5,F5,A ! 6,D ! 7# might reasonably be heard as

On mm. 314 of the Orchesterlied, op. 4, no. 5, Headlam

writes: The upper line arpeggiates !G5A ! 5B ! 5C " 6E6",
accompanied by an inner-voice sum-0 wedge !F4E4D4
(B !3)A !3" and the E-based motive 4/a, and harmonized
with whole-tone+ chords that characteristically alternate
source collections C and C " (93). Example 7 excerpts motive 4/a from Headlams Example 3.7 (87). Several questions
arise on examining the score. Is a sum-0 wedge a pitchspace or pitch-class space construct? Why is B ! 3 in parentheses when it is present in the cellos? If the accompaniment
is based on a sum-0 wedge, then where is pitch-class G, the
counterpart to pitch-class F? Questions also surface when
Headlam invokes symmetry: The opening vocal phrase of
op. 4, no. 2, !B ! BG F " FE ! AE" [is] symmetrical
around A4/B ! 4 (91). But pitch-classes A !, C, C " , and D,
needed to sum-pair with pitches B4, G4, F " 4, and F4 of the
vocal phrase, are not members of the phrase.
Headlam recognizes that analogies with tonality (in particular) are referential and not structural. At one point, for
instance, he writes parenthetically that characteristically, the
interval 7s over the bass allude to a local tonal stability
(85). But at times his language strongly implies that they
may also have a structural role. As in many Berg pieces
[here the String Quartet, op. 3, first movement], the focus
on the bass G2 in mm. 39 sounds vaguely tonal; characteristically, the upper D3 in mm. 69 has a dual role as a dissonance in the C " whole-tone+ context and a consonant upper
interval 7 support for G2 (83). Headlams Example 3.6(a)

more a projection of ic 5 than a simultaneous pitch-space realization of

6Z40[012358] (which possesses two gaps with one pitch class missing
in each gapsee below, however) as !C4,D ! 4,D4,E ! 4,F4,A ! 4". Further,
if this is truly a pitch-class space relation, then spreading 6-Z40 on the
pitch-class clock with either five or seven semitones between the
hours reveals three gaps, two with one pitch class missing and another
with more than one pitch class missing. A further problem arises: even
if we ignore the larger third gap, 13 of 50 hexachords (26%) share this
property with respect to ic 5. This is not a very fine-grained tool.

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 161

reviews

ints

T6

161

example 7. Motive 4/a in op. 4 from Headlams Example 3.7.

Langsam

.0

MT

.0
.
% 0

,

\\

,


,

WGE

,

\\

l l l l l
3

*
\\
CYC

\\

*
\\

l l

l l

l l
l l l

SMT

[
^[_

.0

(Tempo I)

l l

l l

\\\

\\\


\\\

example 8. Headlams Example 3.6a, the opening of String Quartet, op. 3, i.

MT

\\espress. n

/
l/ ,
\\

MTS.Reviews_pp131-174

162

5/11/04

11:06 AM

Page 162

music theory spectrum 26 (2004)

(81), given here as Example 8, is the annotated opening

of this movement. He has previously identified set-class
3-3[014], aligned 1- and 5-cycles, wedge voice leading, and
horizontally unfolding 3-cycles as important in this movement (79). Recall that the + note, here D3, is analogically
dissonant in the local cycle+ collection. As the aligned cycles
are not present and ic 5 is either not present or present in the
remaining structures as an unstable interval, how can an interval 7 provide structural support in this atonal movement?
A similar situation obtains with the term cadence. To
create an atonal analogy to a tonal cadence requires a great
deal: simultaneous and coordinated intervallic patterning of
soprano and bass, and stereotypical harmonic successions.
Cadential points in tonal music, moreover, are typically followed by repetition (with or without variation) or by clearly
contrasting material. Further, cadences are frequently separated from one another or are at least expected after more or
less standard phrase lengths (e.g., 4, 8, or 16 measures). Yet
the only criterion Headlam invokes for hearing atonal cadences is that a succession of cyclic segments is involved:
The phrase [from the first movement of the String Quartet,
op. 3, shown in Example 8] ends with an open cadence
(m. 9): a C " [sic] whole-tone+ chord {G,B !,D,F " } leading to
a less structural neighboring 5-cycle+ chord {A,D,C,E}
(83). While this and other cadences described throughout
chapter 3 occur at segmental points of articulation, he provides us with no segmentation criteria for deciding when
points of articulation are cadential as opposed to serving
merely as markers for semiphrase boundaries.
If Headlams considerations of the tonal and atonal music
are problematic, his elucidation of Bergs highly idiosyncratic
serial music deserves fulsome praise. Of Bergs serial tactics, he writes that the basis of Bergs pitch language in his
later music continues to be the cyclic collections of his earlier
period, developed with regard to order-position relationships
and aggregate completion, and still capable of tonal allusion
by registral spacing and intervallic emphases (195). Headlam makes telling use of the rich legacy of sketches, row-

tables, self-analyses, and musically revealing letters by Berg

on his serial music, and of the various studies of that material
in print, the direct analytical application of which, however,
remains an area of disagreement among scholars. Headlam
neatly summarizes the contending positions and then gives
his own: any relationships that Berg reveals between rows
and material on the surface may have dramatic associations
or even local musical structural bases, but they are not necessary for musical coherence. Since the rows are not central,
their treatment or relationship to the surface need not be
consistent. Thus, Berg can reorder rows and even add or
omit notes without disturbing the language. Although he
often carefully related derived materials to the original row,
the use of row-derived materials in non-row contexts, the
reordering of row segments, and the free addition of nonrow-derived notes suggests that the basis of the language is
not the rows but the smaller derived and non-derived materials, which are mostly, as in his atonal music, cyclic-based
collections (19798).
Each of the serial pieces receives a detailed analysis engaging pitch structure, rhythmic structure, and musical form;
readers will benefit enormously from the plentiful and useful
information on Bergs serial music. While Headlam makes
frequent use of Bergs writings, he does so with a critical eye.
A particularly insightful discussion relates these materials to
problems of compositional realization in mm. 4567 of the
third movement from the Lyric Suite (26470). In effect,
the series of important questions he raises amounts to a call
for a new critical edition of the work.
conclusion
Headlams attempts to view Bergs tonal, atonal, and serial
work as unified by cyclic materials and their derivatives
(386), occasionally ignores or overlooks some theoretical and
analytical issues. His conclusions laudably point out some of
the difficulties in such an enterprise. He notes that an important question about Bergs music has not yet been asked.

MTS.Reviews_pp131-174

5/11/04

11:06 AM

Page 163

reviews
I paraphrase it here: how do passages, strata, or even entire
movements that do not fit the assumptions made here cohere? (388). Still other questions lurk. Are we to read his account of Bergs music as a Bildungsroman, the story of a composer perfecting his craft (389)? This may leave some of us
uncomfortable with the unstated implication that later works
must be better than earlier ones. And are we to believe in a
form-building processes rooted in a rhetoric of order versus
chaos from a composer who has well documented his constructive means (64)? What of the assumption of unity itself
in the body of work by a composer of keen dramatic sensibilities in the very city and time of Freud?12
Headlam has patiently and lovingly woven the fabric of a
story about Bergs great music from the threads of a multitude of scholars and his own work. Although there are some
tears and even gaps in that fabric, Headlams book is an important point of departure in exploring the gorgeous tapestry
of Bergs music.
list of works cited
Aldwell, Edward, and Carl Schachter. 1989. Harmony and
Voice Leading. 2nd ed. San Diego: Harcourt Brace
Jovanovich.
Brown, Matthew, Douglas Dempster, and Dave Headlam.
1997. " IV Hypothesis: Testing the Limits of Schenkers
Theory of Tonality. Music Theory Spectrum 19/2: 155
205.
Cadwallader, Allen, and David Gagn. 1998. Analysis of
Tonal Music: A Schenkerian Approach. Oxford: Oxford
University Press.
Hayes, Malcom. 1995. Anton von Webern. London: Phaidon.

12

Hayes (1995, 108) and Moldenhauer (1979, 17880, 195) both document an occasion on which Bergs friend, Anton Webern, sought help
from Freuds follower and later apostate, Alfred Adler, at the suggestion
of Schoenberg.

163
Jarman, Douglas. 1979. The Music of Alban Berg. Berkeley:
University of California Press.
Lewin, David. 1977. Fortes Interval Vector, My Interval
Function, and Regeners Common-Note Function.
Journal of Music Theory 21/2: 194237.
. 1990. Klumpenhouwer Networks and Some
Isographies that Involve Them. Music Theory Spectrum
12/1: 83120.
Moldenhauer, Hans, with Rosaleen Moldenhauer. 1979.
Anton von Webern: A Chronicle of His Life and Work, New
York: Alfred A. Knopf.
Morris, Robert D. 1991. Class Notes for Atonal Music Theory.
Lebanon, New Hampshire, Frog Peak Music.
Popper, Karl R. 1963. Conjectures and Refutations: The
Growth of Scientific Knowledge. New York: Harper.
Schenker, Heinrich. 1979. Free Composition, trans. Ernst
Oster. New York: Longman.
Schmalfeldt, Janet. 1991. Bergs Path to Atonality: The
Piano Sonata, Op. 1. In Alban Berg: Historical and
Analytical Perspectives, ed. David Gable and Robert P.
Morgan. Oxford: Clarendon Press, 79109.
Schoenberg, Arnold. 1978. Theory of Harmony, trans. Roy E.
Carter. Berkeley: University of California Press.
Straus, Joseph N. 1990. Introduction to Post-tonal Theory.
Englewood Cliffs, New Jersey: Prentice Hall.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.