'Work in Progress' - Analysing Nono's 'Il Canto Sospeso'

'Work in Progress': Analysing Nono's 'Il Canto Sospeso'
Author(s): Kathryn Bailey

Source: Music Analysis, Vol. 11, No. 2/3, Alexander Goehr 60th-Birthday Issue (Jul. - Oct.,
1992), pp. 279-334
Published by: Wiley
Stable URL: http://www.jstor.org/stable/854029
Accessed: 22-05-2016 17:38 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/854029?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact [email protected].
Wiley is collaborating with JSTOR to digitize, preserve and extend access to Music Analysis
This content downloaded from 193.175.22.68 on Sun, 22 May 2016 17:38:40 UTC
All use subject to http://about.jstor.org/terms
KATHRYN BAILEY
'WORK IN PROGRESS':
ANALYSING NONO'S IL CANTO SOSPESO
In 1960, Alexander Goehr wrote that integral serialism

amounts to a complete neutralization of musical character. The
creative idea is replaced by mere procedure. There is no material in the
traditional sense, but only a pre-compositional abstraction of the
intended course of events.'
In the same year Gy6rgy Ligeti published his analysis of Pierre Boulez's
Structure la for two pianos,2 perhaps the most famous work in this tradition. I was disturbed when I first read this, in my twenties, and realised
that without the composer's explanation (at second hand, in this case) no
amount of labour on my part would have been sufficient for me to discover
the methods used in writing the work and thus to understand why it was as
it was. In those days I was analysing for the first time the twelve-note
music of Schoenberg and Webern, and while I found Schoenberg's
penchant for taking liberties with his system unsettling - because it seemed
to leave the music so susceptible to error (how could one judge the
accuracy of a printing, for example, when the only known control might be
overridden at any time?) - at least one knew in the end how the music as a
whole had been put together and which bits did not conform to the system.
I found much more unsettling the idea that although all aspects of a work
might be closely organised and no liberties taken, any hope of my
discovering the particulars of this organisation through listening to the

piece and studying the score would be futile. Thirty years later I still find
this unsettling.
Luigi Nono's II canto sospeso (1955-6) is a work that stands directly in

the centre of the tradition of post-Webern serialism as it was practised in
the 1950s and 60s; it is also noteworthy as an example of post-war socially
committed music (engagierte Musik). Its text - fragments of letters written
by condemned Resistance fighters - has a life of its own: presumably it was
this, not the integral serialism, that occasioned a bomb attack at the
MUSIC ANALYSIS 11:2-3, 1992 279
KATHYRN BAILEY
Munich Oktoberfest in 1980, killing thirteen people and injuring more

than 200 others, and preventing a scheduled performance. I first became
interested in the work in its role as political protest; I later became
fascinated with the intricacies of its construction. Now, years later, and
after a further six months' close study of the score, I have come to the
conclusion that while I can explain the organisation of several of the
parameters of this work there are still many important decisions for which I
can offer no explanation. I arrived some time ago at the point of diminishing returns. In the ensuing pages I offer what illumination I can
concerning the composition of II canto sospeso; unfortunately this addition
to what has already been written3 still leaves knowledge of the compositional process incomplete. I have deduced the systems used in
determining several aspects of the piece; explanations of most of these have
not, to my knowledge, been put forward before. They have to do mostly
with durations and dynamic indications, which are intricately and variously
organised throughout. I have also solved a few riddles connected with
pitch. The elusiveness of some of Nono's techniques has meant that the
tenor of my analysis is not consistent: where I have been able to apprehend
the system, this is described objectively; where I have not succeeded in
cracking the code, the analysis is of a more impressionistic nature.
Such a gapped analysis is all I can honestly (and safely) offer. In a work
of this sort it is important not to extrapolate and generalise; an early review
of II canto sospeso by the British composer and critic Reginald Smith
Brindle is a case in point.4 Smith Brindle, whose specific knowledge of the

piece was, it is quite clear, limited to No. 2 (one movement of nine) and,
even so, incomplete, wrote in an assured way nevertheless about the entire
piece, on the assumption that the same rules apply throughout. In doing so
he provided several generations of music students and enthusiasts with
considerable misinformation, perhaps the most surprising being his blithe
assertion of the work's simplicity.
In examining Nono's composition principles in the Canto sospeso ... it

will be relatively easy to show how his music is an admirable unity of
pre-determined elements ...
It will be easy to reveal the mechanics of his work ...
The serial aspect of the Canto sospeso is extremely simple, as is the form
of the tone-row itself ... The use of the tone-row is as simple as its
design. It is never transposed, or used in inversion or retrograde. [See
the analysis of Nos 5 and 9 below.] Furthermore, it is not used horizontally in individual voice and instrumental parts, but is applied to
each note (whatever voice this belongs to) as it occurs in time. [See my
analysis of Nos 1, 3, 5 and 7.] The same succession of notes is therefore maintained during the twenty-eight minutes of the whole work.5
280 MUSIC ANALYSIS 11:2-3, 1992
NONO'S IL CANTO SOSPESO
In conjunction with the lack of knowledge of the work betrayed in these

observations, such deceptions as 'It is impossible to analyze here the whole
of the structures in this work. It will be sufficient for our purpose to examine
the second movement'6 and 'Each of these eleven dynamic values is

distributed according to a permutational plan, which hardly concerns us
here'" seem particularly disturbing. Concerning the 'extreme simplicity' of
Il canto's construction, let the following incomplete analysis speak for itself.
THE SERIES
II canto sospeso is built on an expanding all-interval series (referred to

henceforth as the a-is) - A B% A, B G C F# C# F D E Eb. The content of
each hexachord of this series represents half of the chromatic octave (the
row is built from Babbitt's first all-combinatorial source set, with hexachordal content 111117/111117)." It is a symmetrical row whose R0 and P6
are identical. Nono is interested in this series not primarily because of its
combinatorial or symmetrical possibilities, I think, or even because of its
capability of producing eleven interval classes in a single statement. His
fascination seems, instead, to be with the way in which it is arrived at. The
series is produced by the application of either the permutation 7 6 8 5 9 4
10 3 11 2 12 1 or its retrograde to a descending chromatic scale (EL to E or
A to BL respectively), or of its obverse, 12 1 11 2 10 3 9 4 8 5 7 6, or its
retrograde to an ascending one (BL to A, E to Ek). (The two forms I have
listed as retrograde, 1 12 2 11 ... and 6 7 5 8 ..., are less elegant than the
other two because in both cases one of the outer elements of the series
remains stationary throughout all the permutations.) The pitch-class rows

that result from any of these permutations possess the characteristics
shown in Fig. 1. The permutations representing the reconstruction of the
chromatic scale from the series are 1 3 5 7 9 11 12 10 8 6 4 2 and 12 10 8
6 4 2 1 3 5 7 9 11 and their retrogrades. Nos 3, 7, 8 and 9 of Il canto
sospeso are based on these permutations. Nos 2 and 4 use cyclic rotation, a
closely related operation. In fact, the work seems almost to represent an
anthology of the possibilities suggested by the ai-s permutations. The
opening movement, in which I can find no strict serial organisation, and

No. 6a are the only two that would appear not to be based on any kind of
permutation or rotation; they explore instead another aspect of the series,
its symmetry, and in so doing exhibit a structural similarity not shared by
the other movements.
NO. 1
The opening piece, without voices, is divided by double barlines and
changes of metre and tempo into thirteen sections, but two of these (the
KATHYRN BAILEY
Fig. 1 Permutation tables relating to the expanding all-interval series

(a) result of 7 6 8 5 9 4 10 3 11 2 12 1 permutation:
[1 2 3 4 5 6 7 8 9 10 11 12] (descending chromatic scale)
(1) 7 6 8 5 9 4 10 3 11 2 12 1 (expanding wedge, Nono's a-is)
(2) 10 4 3 9 11 5 2 8 12 6 1 7 (alternate intervals tritones)
(3) 2 5 8 11 12 9 6 3 1 4 7 10 (three diminished triads)
(4) 6 9 3 12 1 11 4 8 7 5 10 2
(5) 4 11 8 1 7 12 5 3 10 9 2 6 (six perfect 5ths/4ths)

(6) 5 12 3 7 10 1 9 8 2 11 6 4
(7) 9 1 8 10 2 7 11 3 6 12 4 5
(8) 11 7 3 2 6 10 12 8 4 1 5 9 (four augmented triads)

(9) 12 10 8 6 4 2 1 3 5 7 9 11 (both whole-tone scales)
(10) 1 2 3 4 5 6 7 8 9 10 11 12 (descending chromatic scale)
(b) result of 12 1 11 2 10 3 9 4 8 5 7 6 permutation:
[1 2 3 4 5 6 7 8 9 10 11 12] (ascending chromatic scale)
(1) 12 1 11 2 10 3 9 4 8 5 7 6 (expanding wedge, Nono's a-is)
(2) 6 12 7 1 5 11 8 2 4 10 9 3 (alternate intervals tritones)
(3) 3 6 9 12 10 7 4 1 2 5 8 11 (three diminished triads)
(4) 11 3 8 6 5 9 2 12 1 10 4 7
(5) 7 11 4 3 10 8 1 6 12 5 2 9 (six perfect 5ths/4ths)

(6) 9 7 2 11 5 4 12 3 6 10 1 8
(7) 8 9 1 7 10 2 6 11 3 5 12 4
(8) 4 8 12 9 5 1 3 7 11 10 6 2 (four augmented triads)

(9) 2 4 6 8 10 12 11 9 7 5 3 1 (both whole-tone scales)
(10) 1 2 3 4 5 6 7 8 9 10 11 12 (ascending chromatic scale)
second and eleventh) are so brief that they cannot be considered as

anything more than metric interruptions. The movement is essentially a
binary structure and can be represented as in Fig. 2.
The A and B sections are themselves aba ternary structures. In both
cases two similar sections are separated by a third in contrasting style:
section iv - the central portion of A - resembles VI and VIII, the outer
sections of B, while the music separating these latter two - section vii - is
like A's outer sections, III and V. Sections I, III, V, vii and IX are written
for wind instruments and timpani in 8 metre (and 6 in b.40), playing a
succession of clusters. This style is introduced in section I and developed
in III, V and vii, IX is a varied return of I and marks the end of the first
large unit of the piece. The even-numbered sections within this unit - (ii),
iv, VI and VIII - are for strings only in 3 metre and in a linear contrapuntal
style. Again, the first of these sections serves as a brief introduction of the
style, which is then developed in sections iv, VI and VIII. The two styles of
presentation are combined in sections X and XII, where strings and winds
are pitted against each other; the final section, XIII, for a reduced
orchestra of both strings and winds playing together, seems to represent a
synthesis.
Fig. 2 No. 1: Formal structure

1
I (ii) III iv V VI vii VIII IX X (xi) XII XIII
(intro)
(link)
(A+B)
coda
One of the pervasive characteristics of II canto is the use simultaneously

of a number of different tempos. This is a feature of Nos 1, 2, 5, 6a, 6b, 8
and 9. In No. 4 several tempos, notated as various subdivisions of the
quaver, as in the movements just mentioned, occur in succession. Only

No. 3 progresses at the same tempo throughout. The idea of several
tempos proceeding at once - an idea reminiscent of composers as diverse

as Stravinsky (much earlier) and Carter (a little later) - is given a thorough
exposition in the opening movement. Altogether, ten tempos are used here:
the basic pulses are the demisemiquaver septuplet, the demisemiquaver
quintuplet, the demisemiquaver, the semiquaver septuplet, the semiquaver
quintuplet, the semiquaver triplet, the semiquaver, the quaver triplet, the
quaver and the crotchet triplet. The exact nature of some of these tempos
is obscured by the notation in the sections in 4 metre, where, in order to
facilitate entries on minute subdivisions of the beat, all parts are written as
if the basic unit were twice as fast as it is (written demisemiquaver

quintuplets and septuplets produce, in fact, notes whose pulse must be
understood as divisions of the semiquaver into five and seven).

Five tempos are introduced in the first four bars (section I); two notes
are played in each tempo, the first five units long and the second ten. (The
consistency with which like durations are clustered throughout this
movement supports the suggestion that the nine-unit roll played by the
timpani in bs 3 and 4 should be one semiquaver longer.) The clusters

continue in section III, where the values 6, 4, 9, 2, 7, 12 and 10 are
presented in four tempos (see Ex. 1).

After the interruption by section iv the clusters continue, now in only
three tempos: the values are 7, 8, 4, 12, 3, 11, 7, 10, 6, 2 and 5. (The third
and fourth of these occur in two voices only.) In the next section of this
type, vii, the number of tempos has been further reduced to two: the values
are 5, 8, 11, 2, 6, 12, 1, 7, 8 and 2. Following the B section, section IX is a
KATHYRN BAILEY
Ex. 1
III
_9
9(b)
- 9 ral. 6(d) accel. 4(d) rail.

FI. I
6 _ c) 4 e
Ott.
6(
b)
2bA0
____ __ ____
4. - - - _-~~~_--___ _ _ - _ i ... ..a in
i~I-ziji---7
Clar. I
a4(a) 9 )
____ --- -____ ___________- - rz I' zzz- z:
4(b)mp
2(b)
___;t __- --- __ __ - X4X?-_

Tee - ''j __ - ____ --- -----____{ ____
1 t a= ui pe s units~fL~zzrn~ _61_a)
TJg
--
-9
_ _ -_ - 4- - ---------
4r _ --92 ralL. _ accelcc) :z z ______ -- -~ ------
b---
scisqtcvcr
uit
trpetqivrunt
4_4__euns
qaaser
tail
~~~~- ~ ~ Cprih Ms~- Viv Ve-~~-- ~~~-~ -- --- -- ---- - rla GmbH--- Mainz.------
RepodceIdyeris
by- perisio
ofSht o t. odn L . n
-eprduc
i---Scott&-C.
d = quaver units
Copyright Ars Viva Verlag GmbH, Mainz.

Reproduced by permission of Schott & Co. Ltd., London.

Ex. 1 cont.
iv
=
P
92
al.
1=
60
12(d)
Fl. 1
S-
--
in ---
Clar. bso
12(a) 10(d__
mm--m-m-m - -- i -. ... I_ F.
212(b)
mf
_10(c)
_-Cor-.f--
C . - r i _ _ _ _ _ _ _ _ " "__" L
Trbne.
-6
...
S= 92 rail. 4_ _ ....d
ViT, I
Vc
- -
pppp
...
".
7obl ,ord r - r 3
o
d. .
At
c b . , {-----------b
con sor- d.
..b
KATHYRN BAILEY
return of the opening: again in four bars of ', utilizing the same instruments as bs 1-4, two durations are presented in each of the five tempos
used at the opening, this time in notes of six and seven units' duration.
While the number of tempos is diminishing steadily in the vertically
oriented wind sections of the piece, just the opposite is happening in the
intervening string sections, which are in a linear style. The introduction of
this style occurs in section (ii), in which a single voice plays three notes in
succession (the last two doubled at the same pitch), with values of 3, 8 and
1 quavers. In the first section to develop this idea, section iv, two
rhythmically continuous but pointillistically orchestrated voices both move
in triplets - one in quavers and one in crotchets: the values in the slower
voice are 6 and 9, in the quicker one 4, 2, 5, 3, 1 and 11. The next contrapuntal section, VI, is in three voices, moving in semiquavers, quaver
triplets and quavers. The series of durations presented by these voices are
respectively 7 5 8 10 2 1 3 9 1 4 7, 9 9 11 6 10 and 12 12. In section VIII
the number of voices and tempos has increased to four and the series have
become much longer: 3 10 7 2 4 11 5 7 3 12 8 1 2 6 7 3 8 1 11 9, 9 12 5 3
9 1 5 2 4 10 10 12 12, 4 11 1 8 11 6 5 11 10 5 and 5 9 9 9 4 2 8 (see
Ex. 2). A steady increase in length reinforces this growth in complexity:
section (ii) is two bars long; section iv, although notated as three bars,
overlaps with the following section and thus in fact lasts for four bars;
section VI is six bars long, and section VIII, eight.
In the third large section of the movement, comprising barred sections
X, (xi) and XII, vertical and horizontal styles occur simultaneously, each
played by the instruments associated with it up to this point in section X
(see Ex. 3), with the roles switched in XII. Section (xi) consists of a single
minim, which is the value needed to complete section X. The clusters
presented first by the winds and later by the strings are less consistently
treated than was the case in earlier sections of the movement: although
there are five tempos present, the majority of the durations occur in only
two, three or four of these. Five voices move in continuous polyphony
meanwhile, stating ever longer series of notes, also in five tempos; there is
one break in the continuity of all voices in section XII (the length of this
break varies greatly from one voice to another according to the tempo; all
the breaks occur between bs 74 and 83).
In the final section, XIII, continuous polyphony is abandoned as a
reduced number of instruments representing both winds and strings
presents a series of durations, each occurring in only two parts/tempos, and

not separated into clusters as in earlier sections. Because the several
tempos in this section are neither treated as single continuous voices, nor
presented in exclusive clusters, this final section may perhaps be seen to
represent a consolidation, but at the same time a dissolution, of the two
styles used in opposition until this point.
Although I have identified all the tempos and durations used in this
movement (not as simple-minded an occupation as might be supposed), I
Ex. 2 VIII
Sca .82
FI. 3
Fg.
..-n
Tr. 5 o
(Cdi)
42(a)
(c)
---?? --I- -' - T 7 , <-:i-:.- --?........

7
--
----
, ,7 -
id.
Pj ..--=sord.
(c). __4(a
F 8
Con
sord.
-en:a
'".
40
"
con
12(b)
(div. -9f
enza sor
sord.
3f )
VCb
.
9(b)
,.___ ____ r a ' "-Vlot
44 9(d) 4(d) f 801)

Rdip
dbIyi)
of
4 9_-_ 3:- - - .k "n ,12) ._, -d
" Copyrigh
--' U"- -- - Ars
------z
Viv
__.
Vrg
t< A
b (,
M--_ViC
yg
ms
Mirc 2ro.d .
F A. Ini quaver tiiis I 1 1
d=(bl i r tnl1t2(b
Copyright Ars Viva Verlag GmbH, Mainz.

Reproduced by permission of Schott & Co. Ltd., London.
KATHYRN BAILEY
Ex. 3
9(b)
(a)
Timp
1 (e)
1213 { f a aa-p 9 2
I2(d)
Sr----
! _ - - --- - _--_- _- diAI
"I ;O aI I
7Idi
1040)
_ 24(d
L_. ... t :, V i --- -9--- ------ -- ---_----_ -----------. -f

- -I quaqer uni
I&b4 5-(
24 _(c)
c- = senlq ver units
can offer no explanation for the choice or the order of either durations or
pitches. Judging from the tenor of the other eight movements, all of which
show a high degree of control and integration, I can only suppose that the
opening one is also tightly organised, but I have not been able to find the
key. The a-is is presented haltingly in the opening bars of the movement:
the first cluster (duration 5, bs 1-2) presents notes 1-5 of the series, but in
the order G G# A B Bb; the second cluster (duration 10, bs 2-5) uses notes
6-9 (in the order F F# C# C) plus the first note of the series again; the
remaining three notes, again in the wrong order (E Eb D), appear in the
first two clusters of section III (durations 6 and 4, bs 7-10), along with
other pitches occupying various positions in the series. Pitches 1-9 of the
series are presented by consecutive notes in bs 44-6 with 10-12 in
consecutive notes after the interpolation of G and B. Beyond these two
very unassuming references to the row, I can offer nothing.
NO. 2
'... I die for a world that will shine with light of such strength and
beauty that my sacrifice is nothing. For that world millions of men
have died on the barricades and in war. I die for justice. Our ideas will
triumph... '"
In two movements of II canto sospeso - Nos 2 and 6b - cyclic rotation is
used to determine the order of durations in a very simple pitch-class
structure consisting of repeated block statements of the untransposed a-is.

These two movements are similar in many other, more apparent, ways as
well: both are in ' metre and are linear, with four voices proceeding in
several different tempos, in the style introduced in the even-numbered
sections of the opening movement, and both are fifty bars long.
The second movement is written for unaccompanied eight-part chorus.

The eight voices never sound together but are used as available material in
the pointillistic composition of four continuous voices which proceed
polyphonically, each moving at its own tempo. The four tempos share the
crotchet, dividing it into two, three, four and five units. The four voices
begin the piece together; the first gap in the four-voice texture occurs at the
beginning of the coda (b. 142).
The ordering of both pitches and durations follows what I have

elsewhere called 'block topography','0 cutting across the four-voice linear
structure to present a succession of twelve-note chunks. In each of these
the pitches are arranged so that they articulate the untransposed a-is. Thus
the pitch content of the movement is audibly repetitious: the untransposed
a-is is heard constantly repeated, nineteen times all told.
Each of the nineteen pitch-class rows functions also as a series of
durations. Both the values used and the way in which they are ordered in
KATHYRN BAILEY
successive series are determined by the regular cyclic rotation of a

symmetrical series based on the first six Fibonacci numbers. The durations
of the notes making up the first a-is, in bs 108-10, establish the series 1 2 3
5 8 13 13 8 5 3 2 1. The following eleven statements go through a
complete cycle of rotation in which each series begins with the second
element of the one preceding, the first having moved to the end. This
produces a 'magic square' in which corresponding rows and columns are
identical (see rows 1-12 in Fig. 3). Six of these rotations produce the
inversion of the original series (13 8 5 3 2 1 1 2 3 5 8 13); twelve would
reproduce the original itself. This latter step is avoided. The cycle seems to
begin a second time - series 13-15 are the same as 2-4 - but omitting the
repetition of the original series. Rows 13-15 also represent the first three
steps backwards through the permutation table. This new cycle/reversal is
cut short after three rows, as if all is accomplished once the relationship has
been demonstrated."
Here I should like to quote Smith Brindle one last time. Not only has he
made assertions about the whole work on the basis of his analysis of a
Fig. 3 No. 2: Table of durations
A Bb AX B G C F C# F D E E_
(1)
(2)
13
13
(3)
13
(4)
13
(5)
13
13
13
13
8
13
13
(6)
13
13
(7)
13
(8)
(9)
13
(10)
(11)
13
(12)
13
13
13
13
13
13
13
13
13
[1 2 3 5 8 13 13 8 5 3 2 1 : missing]
(13)
(14)
13
13
(15)
13
13
13
13
(16) 13 13 8 13 13 8 5 8 8 5 3 5
(17)
CODA
(18) 1 2 1 1 2 2 1 3 2 3 8 5
(19) 5 3 8 5 8 5 13 8 13 8 13 13
single movement; it now becomes clear that his remarks about that
movement are based on his analysis of the first two bars. He correctly
identifies the initial durations series and goes part of the way to recognising
how it works when he says
... the proportional series and its retrograde (1.2.3.5.8.13//13.8.5.3.2.1)
is applied to four basic duration values... .Whenever a note ends it is
followed by another formed by multiplying the same basic duration
value by the next number available in the proportional series.'2
(The latter is of course not true at the ends of rows, where the 'next
number available in the series' is always skipped; this first occurs in the
third bar.) The fact is that Smith Brindle's investigation of the movement
must have stopped with the first row; this is the only explanation for his
subsequent statement that
The first complete structure (34 measures) continues this procedure
until the proportional series and its retrograde have been used ten
times.
Not only has he failed to notice that the rows are not all the same, but even
his sums are wrong.
The coda (bs 142-57) works in a different way altogether. Up to this

point the rotation of the durations series has cut across the polyphonic
structure of the music. While the serial operation determines the number
of units in each note it does not prescribe which of the four tempos is to be
used; therefore the same series of numbers results in a variety of combinations. In the coda it is as if this approach were given a 90-degree turn:
now, although the a-is continues to repeat as before, the significant
number series are not produced by successive statements of the pitch-class
row but by the four individual tempo strands/voices, each of which plays
completely the inversion of the durations series (previously heard as row 7,
halfway through the rotational cycle: 13 8 5 3 2 1 1 2 3 5 8 13). The
relative chaos this produces within the four statements of the a-is can be
seen in the row table (rows 16-19 in Fig. 3).
The order of the dynamic indications is also determined through
rotation, by the application of a system which, though similar to, is quite

independent of that governing the durations. Eleven discrete dynamic
indicators are used in the movement: five basic levels plus six indications of
change. The lowest level, ppp, is used twice, in order to obtain the series of
twelve elements shown in Fig. 4. When the dynamic contents of successive
pitch-class rows are displayed on a table like the one used for durations in
Fig. 3 it will be seen that in the first section of the movement (rows 1-12)
each column (representing the successive appearances of each pitch class)
contains all twelve indicators in the correct order, as was the case with
KATHYRN BAILEY
Fig. 4 No. 2: Dynamics series

ppp p mp mf f ppp ppp<f f>ppp ppp<mf mf>ppp p<f f>p
1
10
11
12
durations. Also as there, each column represents a rotation from the

previous one so that neighbouring columns begin at different points in the
cycle (see Fig. 5).13 Unlike the rotations governing the durations, however,
these are not regular; since the rotations of the series (in the columns) are
not determined by the series, the first row is not the same as the first
column, and the product is not a symmetrical square. The pattern of

rotation has been determined by the application of a permutation closely
related to the a-is. This is shown in Fig. 6.
There are eight discrepancies in the score. These can be rationalised in
various ways. The first, the G in row 1, is possibly a printing omission simply, there is no dynamic indication for this note (though this is not
Fig. 5 No. 2: Table of dynamics
A B% A, B G C F# C# F D E E6
(1)
(2)
(3)
(4)
(5)
(3)
12
2 4 3 7 4 10 5 1 6 4 (7) (7)
(3) 5 4 8 5 11 6 2 7 5 (8) 8
4 6 5 9 6 12 7 3 (8) 6 9 9
5 7 6 10 7 1 8 4 (9) 7 10 10
(6) 6 8 7 11 8 2 9 5 10 8 11 11
(7) 7 9 8 12 9 3 10 6 11 9 12 12
(8) 8 10 9 1 10 4 11 7 12 10 1 1
(9) 9 11 10 2 11 5 12 8 1 11 2 2
(10) 10 12 11 3 12 6 1 9 2 12 3 3
(11)
11
(12)
12
(13)
10
(14)
11
12
10
11
(15) 4 6 5 9 6 12 7 3 8 6 9 (9)
(16)
(17)
(18)
10
(19)
10
12
11
12
CODA
NB: Numbers in brackets indicate discrepancies between what is required by the system
and what appears in the score; I have altered these numbers to fit the system.
Fig. 6 No. 2: Pattern of rotation used in table of dynamics

I
-1
I
-3
-5
-7
I
-9
first rows, bs 108-10: 1 3 2 6 3 9 4 12 5 3 6 6

+2
I
+4
I
+6
I
+8
I
+10
I
+0
I
necessarily a sign of omission, as it happens more than once elsewhere). It

is easy to suppose a similar explanation for both the E in row 3 and the F
in row 4, notes marked only f instead of the required f>ppp. The
discrepancies in rows 2 and 5 (2: E and E?, 5: F) might be explained in the
same way (all three are marked ppp instead of ppp<f (in 2) or ppp<mf (in
5), but a second explanation seems more likely, and this would cover the F
in row 4 as well: these four notes are all semiquaver quintuplets, and
nowhere in the movement does Nono assign a changing dynamic to this
short value. If this is the explanation, it is an example of one of the hazards
of writing music in this way: the production of an impossibility through the
inauspicious intersection of two one-dimensional systems. The E? in row

15, ppp<f instead of ppp<mf, may be an error of omission; it is possible, on
the other hand, that its crescendo is exaggerated to emphasise the end of
the main body of the piece and of the four-voice texture that has been
consistent up to this point (but then, why was the dynamic of the E not
altered in the same way?). Only this E, and the A in row 3 (ppp<mf instead
of mp) do not offer easy explanations.
The recurrence of the numbers three and six in the first row results in
identical columns in second, fifth and tenth, and in fourth, eleventh and
twelfth positions. No two rows are the same: rows 2-12 represent the
complete table of transpositions of the initial row, listed chromatically.
Like the durations, the dynamics of rows 13-15 are identical to those of
rows 2-4, and, again, the order within the coda appears chaotic. I cannot in
this case determine the way in which these four rows are organised.
NO. 3
'... they are taking me to Kessariani for execution together with seven
others. I die for liberty and country ...'
'... today they will shoot us. We die as men for our country. Be
worthy of us...'
'... they will hang me in the square because I am a patriot. Your son
goes without hearing the bells of freedom ...'
KATHYRN BAILEY
No. 3 is written for three solo voices and large orchestra (24 wind parts
and at least 19 strings, although at no time are all instruments present).
The parts are grouped into three rhythmic voices which move at the same
tempo, the semiquaver acting as basic durational unit for all three.
Durations are serialised, and, as in the majority of the other movements,
the operator is one of the a-is permutations. The treatment differs from
that in the second movement in several ways: most significantly, both
pitch-class and durations series are given a linear exposition here.
All the durations series in No. 3 draw from the odd numbers between
one and nine; each voice uses a distinctive symmetrical combination of

these numbers. In each series used by the voice that enters first (hereafter
referred to as A) the numbers 1, 3, 5 and 9 all appear twice, and 7 occurs
four times. All the series in voice B contain two each of the numbers 1, 3, 7
and 9, and four 5s; in those of voice C there are four appearances of the
numbers 1 and 3, two of 5 and 7, and no 9. In all voices the series are
subjected to a succession of 2 4 6 8 10 12 1 3 5 7 9 11 - or 11 9 7 5 3 1 12
10 8 6 4 2 - permutations (the symmetry of the original series ensures the
identity of the results of these two operations). Because the starting rows
were symmetrical, all the rows resulting from this permutation are
symmetrical as well. The permutation tables for the rows used by all three
voices are given as Fig. 7.
Fig. 7 No. 3: Permutation tables for durations series
voice
voice
voice
(1) 371579975173 531957759135 171353353171

(2) 759713317957 397515515793 733511115337
(3) 573197791375 955173371559 351137731153

(4) 717935539717 513759957315 517313313715
(5) 195377773591 179535535971 133175571331

(6) 937751157739 755391193557 315731137513
(1) 371579975173 531957759135 171353353171
Each voice is presented by a distinctive instrumental group: A by oboes,

bassoon and trombones; B by clarinets, including bass clarinet, and horns;
C by flutes and trumpets. Strings and singers join all three groups at
various times. The activity of the three voices is shown in Fig. 8.
Fig. 8 No. 3: Three-voice structure

158
176
181
182
201
207
213
219
226
231
239
Each voice plays its original durations series and its first permutation in
the opening section of the movement (bs 158-76). In the middle section
(bs 177-219) voice A goes through three more rows, continuing its
permutations to the penultimate one of its cycle, and voice B completes
one cycle except for the repetition of the starting row (six rows). Like A,
voice C presents rows 3, 4 and 5 of its cycle, then the first four notes only
of its sixth and final row; omitting altogether the repetition of the row with
which it started, it goes on to rows 2 and 3 of a second cycle, which is

continued in the final section (bs 219-39), where the last row (6) is once
again represented by its first four notes only. In this section voice B states
its starting row entirely, thus finishing the only complete permutational
cycle in the movement. Voice A has only fragments in this section: ultimately all but note 10 of the sixth row of its cycle, plus notes 1-6 of its
original series. This is shown in Fig. 9.
The pitch content of this movement is less transparent. Its pattern is
obscured in two ways. Firstly, in the course of the movement eight of the
durations are expressed as rests (in bs 168, 170, 181, 191, 194, 201, 216
and 236-7, shown in bold in Fig. 9); this means that from b.168 onwards a
direct correspondence between durations series and pitch-class rows is
impossible. Secondly, when two or more instruments play the same duration they frequently have different pitches (though there is considerable
pitch-class reinforcement as well). Pitch-class diversity of this sort occurs
sporadically; at its most dense five different pitches express together the
not group
themselves
into
twelve-note
rows,These
nor is
their
vertical
same
duration
(see, ? in bs
169-70,
for example).
'extra'
pitches
do
relationship to each other consistent, although they are in most cases

clusters of semitones or of tones and semitones. Each of the three voices
opens the movement with the untransposed a-is. Voice C plays this row
twice more in the course of the movement (in the first instance, in bs 18996, the row falls on notes 5-12 of a durations series and a four-note
fragment that follows; the second a-is follows immediately, this time
coinciding with a durations series, in bs 196-201). Voice B ends with the

a-is.
NO. 4
The fourth movement, for instruments alone, uses two orchestras: one of
strings only, the other of wind and percussion. These two bodies play at
the same time, treating the same ordered group of seventy-two pitches in
quite different ways. Perhaps the most striking feature of the movement is
its narrow range: all the pitches in both orchestras fall within the octave e'
and eb2. The strings present the basic pitch material in the form of six
twelve-note rows played in very long notes. All these statements take the
form of textural lozenges, beginning and ending with a single voice and
KATHYRN BAILEY
Fig. 9 No. 3: Durations series

voice A [2x1; 2x3; 2x5; 4x7; 2x9]
(1) 371579975173 (b.158): trbnl/2, ob, bssnl, bcl
(2) 759713317957 (b.167): trbnl/2, ob, bssnl
(3) 573197791375 (b.181): trbnl/2, ob, bssnl, vces, str
(4) 717935539717 (b.189): trbnl, ob, bssnl, ten/alt, vln, vlc
(5) 195377773591 (b.199): trbn1/2/3/4, obl, vces, via, vlc
(6) f 937751157739 (bs219; 231): vces (1-5); trbnl, obl, vces (6-9, 11-12)
(1)f 371579975173 (b.235): ob, cll, trbnl/2 (1-6)

voice B [2x1; 2x3; 4x5; 2x7; 2x9]
(1) 531957759135 (b.158): hnl, cl, bcl
(2) 397515515793 (b.166): hnl/2, cl, bcl
(3) 955173371559 (b.185): hns, cl, bcl, vces, vla, vln

(4) 513759957315 (b.193): hnl/2/3, cls, bcl, vln, vla, vlc
(5) 179535535971 (b.202): vces, vla
(6) 755391193557 (b.210): trpl/2/3/4/5, vces
(1) 531957759135 (b.226): hnl/2, cll, vces

voice C [4x1; 4x3; 2x5; 2x7; no 9]
(1) 171353353171 (b.159): trpl/2/3/5, fl

(2) 733511115337 (b.165): trp2/3/4/5, fl
(3) 351137731153 (b.177): trp2, fl, vces, vln, vlc
(4) 517313313715 (b.182): trpl/2/3/5, fl, str
(5) 133175571331 (b.188): trpl/2/3, fl, vces, vln
(6) f 315731137513 (b.194): sop (1-4)
(1) 171353353171 (missing)]
(2) 733511115337 (b.196): trpl/2/3/5, trbl, fl, vln, vlc, c12
(3) 351137731153 (b.207): trpl/2/3/5, fl, vla
(4) 517313313715 (b.224): trpl/2/3/5, fl, vces
(5) 133175571331 (b.229): trp2/3/4/5, fl
(6)f 315731137513 (b.235): trpl/3/4/5, fl (1-4)
f indicates fragment(s) only.
Those series numbered in bold face present a-is pitch-class series.
Instruments in italics are playing outside their usual ensemble.
Duration numbers in bold face are expressed as rests.
becoming most dense at the centre, and they overlap greatly, occurring in
bs 240-51, 246-57, 253-62, 261-70, 264-78 and 270-84. The instrumentation forms an arch: the first three rows are played by, respectively,
twelve solo cellos and basses playing harmonics, twelve solo violins and
violas, and tutti strings (in eight parts), and the subsequent three by the
same groups in reverse order. This arch is reinforced by the dynamics,
which are the same for the length of each row: ppp<p on each note of the
first row, mp<mf throughout the second and f<fff throughout the third,
then fff>f, mf>mp and p>ppp in the fourth, fifth and sixth rows.
The seventy-two notes played by the strings supply the material for the
wind and percussion orchestra as well, but here they are expressed
differently. Each note played by the strings generates two at the same pitch
level in the wind/percussion orchestra, one reinforcing its appearance and
the other emphasising its release. These notes are much shorter than the
corresponding ones in the string orchestra, and all together they produce a
single continuous rhythmic voice, similar to those already encountered in
movements two and three and the linear sections of movement one.
Because several different notes of the same duration are frequently struck
and released simultaneously (corresponding with as many attacks and
releases in the strings), sixty of the 144 notes echoing the string music are
rhythmically redundant. Conversely, fifteen notes played at various times
throughout the movement by unpitched percussion instruments and one
twelve-note row played by the winds in the central section (bs 258-63) are
independent of the strings but contribute to the continuous rhythm of the
winds and percussion (although most notes in the winds' row are also
rhythmically redundant). The result is a succession of 102 durations, or
eight and one-half twelve-unit rows.
These begin in bs 240, 246, 252, 257, 260, 264, 269, 275 and 281,
thereby (except for the final half row, which acts as a coda extending the
final section) corresponding with the internal double bars and changes of
tempo. Four tempos - based on triplet quavers, semiquavers, quintuplet
semiquavers and septuplet semiquavers - are used to produce an arch form
that accelerates towards the centre. A comparison with the arch formed by
the strings' activities shows that the junctures in these two structures do
not in all cases coincide.
The contents of the durations series are given in Fig. 10a. The initial
series in b.240 is a regular, though obscure, rearrangement of the numbers

from one to twelve (see the first row of Fig. 10a and Fig. 10b); subsequent
rows represent rotations of this series. The pattern of rotation is also
regular: it is such that the initial durations of successive rows read in the
order 11 5 4 10 9 3 2 8 7 (see the first column of durations in Fig. 10a and
Fig. O10c).
I can offer less in the way of explanation where pitch structure is
concerned. Two readings of the six rows played by the string orchestra are
suggested by the emphasis provided by the winds and percussion: they may
KATHYRN BAILEY
Fig. 10 No. 4: Durations series

a) b.240 11 2 1 9 7 12 6 3 4 8 10 5
b.246 5 11 2 1 9 7 12 6 3 4 8 10
b.252
10
11
12
b.257 10 5 11 2 1 9 7 12 6 3 4 8
b.260
12
10
11
b.264 3 4 8 10 5 11 2 1 9 7 12 6
b.269
b.275
10
b.281
12
b)
7
5
6
12
11
3
6
2
3
1
4
9
8
7
10
12
5
6
11
3
+3 -4 +5 (6) -3 +4 -5 (6) (semitone pattern)

11 2 1 9 7 12 / 6 3 4 8 10 5 /
-1
c)
-2
(6)
+1
+2
(6)
11 5 4 10 9 3 2 8 7 [1 12 6] (semitone pattern: 6 -1 6 -1 ...)

I
be arranged in order of either entry or release. Fig. 11 shows the first three
rows ordered in the first way and the last three in the second; the central
row in this example (in bs 258-63) is the one played by winds and
percussion at the centre of the movement, ranging across the strings' third
and fourth rows, in bs 258-63.'~ Notes that are bracketed together are
struck (or, in bs 261-84, released) at the same time, thereby leaving the
order undefined.
The order in the outer rows of this list, i.e. the attacks in the first of the
string rows and the releases in the last - conform to the a-is. The central
wind row has in common with this row only its first and (probably) its last
Fig. 11 No. 4: Pitch-class series
bs 240-51 - A B% At B G C F# C# F D E E,
bs 246-57 ~ B D G Ab A F B6 (C# C E) FO E6
bs 253-62 * B F A (At Bb) C D G C# (E F#) E6
bs 258-63 A B (C F) (C# B[) G (At E) F# (D EL)
bs 261-70 -- A B, C G (B C#) (F# E) F (E, AP) D
bs 264-78 -- A B6 (C C#) (G F#) Ab E6 F B E D

bs 270-84 -- A B6, At B G C F# C# F D E E E
note. Although the winds play sufficient notes in this movement to

accommodate thirteen rows (six times twelve notes reinforcing the attacks
in the strings, six times twelve pointing out their releases and one row not
played by the strings, in bs 258-63), and these are in fact arranged so as to
produce a durations structure of eight and one-half rows; because of
duplications these notes do not sort themselves into pitch-class rows
independent of those of the strings.
NO. 5
'... if the heavens were paper and all the seas of the world ink, I could
not describe my sufferings and all that I see around me. I say goodbye
to all and weep ...'
No. 5 proceeds in three continuous voices, each moving at a different

tempo. The basic durational units are the semiquaver, the semiquaver
quintuplet and the semiquaver sextuplet (written as subdivided quaver
triplets). Only three durations are used in each tempo: the semiquaver, and
notes worth two and seven semiquavers (the fact that the three values add
up to ten is probably significant; the numbers five and ten are important in
this movement, as they were in the opening section of No. 1). These are
combined in four symmetrical twelve-unit series, all of which use each
number four times: 7722 1111 2277, 1717 2222 7171, 1271 7227 1721
and 2127 7117 7212. Each voice has ten durations rows in the course of
the movement; the series played by the voices that enter first and third (the
slowest and fastest moving, respectively) are identical. The durations of all
three voices are shown in Fig. 12.
Three versions of the a-is are used to generate the pitch material of this
movement: voice B is based on the prime form, voice A on the retrograde
and voice C on the retrograde inversion. The latter voice also makes use of
transposition. The ten rows in each voice form two retrograde patterns,
one comprising the row on which the voice is based and its retrograde, the
second a series of four rows obtained through successive permutational

rearrangements, and their retrogrades. The first row in voice A is the
retrograde and the second is the retrograde of this (the a-is itself). The next
four rows proceed from this one, the first through a symmetrical
rearrangement of its notes - 1 4 9 12 / 2 6 7 11 / 3 5 8 10 - and subsequent

ones through the application of similar rearrangements in which the central
tetrachord of each row becomes the first tetrachord of the next one and the
notes of the remaining tetrachords are rearranged symmetrically (the four

outer notes in one tetrachord and the four inner ones in the other) - 1 9 12
4 / 2 10 11 3; 102 3 11/ 1 9 12 4; 2 10 11 3/ 1 9 12 4. These four rows

are then heard again, in reverse order and in retrograde. The second voice
has the entire pitch-class content of the first in retrograde, thus ending with
KATHYRN BAILEY

voices Aand C: 7722 1 1 1 1 2277
1717
voice
B:
2222
7171
7171
1717
2222
7722
1111
2277
1271
7227
1721
2127
7117
7212
7171
1717
2222
2127
7117
7212
1271
7227
1721
7722
1111
2277
127
7227
2127
172
7117
7212
1717
2222
7171
7722
1111
2277
1721
1271
7227
7722
1111
2277
1717
2222
7171
2127
7117
7212
1271
7227
1721
2127
7117
7212
NB If 7722 1111 2277=m, 1717 2222 7171=n, 1271 7227 1721=o and 2127 7117
7212=p, then voices A & C contain 3xm, 3xn, 2xo and 2xp, while voice B contains 2xm,
2xn, 3xo and 3xp.
the prime form of the a-is. The remaining voice begins, like voice A, with a
retrograde pair, in this case the retrograde inversion and its retrograde, the
inversion. The next four rows are produced in the same way as the parallel
rows in voice A and are, as there, followed by their retrogrades in reverse
order. This time, however, everything from the eighth note of the first of
these rows in retrograde to the last seven notes of the fourth is transposed
down a tone; the last seven notes are a tone higher than the original. The
pitch-class content of the three voices is shown in Fig. 13. The obvious
potential for canonic imitation between these three series (rows 1-8 of
voice B are identical to rows 3-10 of voice A; voices A and C are related by
inversion) is not exploited.
Five dynamic indications are used in No. 5: ppp p mp mf and f. These
Fig. 13 No. 5: Table of pitch-class rows

voice A:
b.285 F E D F C F C G B A? B% A (R)
b.288 A B A B G C FG CO F D E E~ (P)
b.291 A B F E B% C FO E A? G CO D
b.295 B% C F# E A A D E~ B G CO F
b.298 G
AC
A,FO
D E6
CA
B6BE
BF
b.301
CO G
A?C BFO
F D
E E
F
b.305
b.308
b.311
b.315
(ret.)
E~ E B6 A D F B A C# Fj C G
E F B B6 C# F# C G E D A
F C G B El D A? A E FO C B6
D C G A? E F# CB6 E~ F B A
voice B:
b.288 A B F E6 B6 C FO E A? G CO D
b.290 B6 C F E A A? D E, B G C# F
b.293 A Ab D E6 G C FO C B6 B F E
b.296 G C FO CO A? B F D A B6 E E6e
b.298 EF E B6 A D F B A? C~ FO C G
(ret.)
b.304 E
F F
CO
E DFj
D ACAGEEE
FODCAB6
b.301
B GB
B6 C
A
b.306 D CO G A E FO C B6 E~ F B A
b.309 A
E B6
E DAb
F CG
C G
BF
A?D B6
A (P)
(R)
b.311
B GFO
C FO
C#
E E6
voice C:
b.293
(I)
b.290 A
E A?
D EB6
CGG FB CFOF CB FGC#
B6 EADAE6(RI)b.295 A G CO E6 A? F C D B6 B F E
b.297 A?F C D A B E E6 G B F CO
b.299 A B7E E6 BFO C F A? G CO D
b.302 B FO C F B6G CO E A A FD E~
b.304 E D A? A E CO G A E B6 E A
b.306 C B F FO E6 B6 E A CO D A? G
b.308 B E6 A F C D A? G C B6 E FO
b.310 D E6A A A& C D A? B6 F E6 A B

NB The notes within the single box are at TIo, those within the double box at T2.
KATHYRN BAILEY
are arranged in the same order in voices A and C, thereby reinforcing the
structural similarity of these voices that has already been observed in the
organisation of both pitch classes and durations; as in the case of durations, however, the content of voice B does not represent the retrograde of
that of the other voices, as would be suggested by the pitch-class relationship. The dynamics suggest for the most part trichordal or hexachordal
divisions rather than the tetrachordal ones implicit in the durations series.
The dynamic content of the three voices is given in Fig. 14.

voices A and C:
(1) ppp p ppp p ppp p mp p mp p mp p

(2) p p mp mp p mp mf mp mf mf mp mp
(3) p p mp mp p mp mf mp mf mp mp mf
(4) mf f mp mf f f f mp f f f mf
(5) mp mf mp f mp mp mp mf f f mf mf
(6) f mp f f mf mp mf mp mp mf f mp
(7) mf mf mp mp mf mp p mp p p mp mp
(8) mf mf mp mf mp mp p p mp p mp mp
(9) mf p mf mp mp mp p p p ppp ppp ppp
(10) p mp p mp p mp p ppp p ppp p ppp
voice B:
(1) ppp ppp ppp p p p mp mp mp p p p

(2) p p p mp mp mp mf mf mf mp mp mp
(3) p p p mp mp mf mf mf mf mp mp mp
(4) mf mf mf f f f f f f mf mf mf
(5) mp mp f mp mp mp mf mf mf f f f
(6) f f mf mf mf mf mp mp mp mp mp mp
(7) p p p mp mf mp mf mf mf mp mp mp
(8) p p p mp mp mp mf mf mf mp mp mp
(9) p p p ppp ppp ppp p p p mp mp mp
(10) p p p ppp ppp ppp ppp ppp ppp ppp ppp ppp
I cannot suggest a serial basis for this list; however, several observations
should be made. The last row of voices A and C is the retrograde of the
first; and there is a close relationship between rows 2 and 8. In voice B,
row 2 is the same as row 1 at the equivalent of Ti; row 8 is the same as row
2; row 4 is a palindrome; the first hexachord of row 9 is the retrograde of
the second hexachord of row 1, and vice versa. In six of the ten rows of
voice B a single dynamic indication is in force through entire trichords; the
four rows that do not conform (3, 5, 6 and 7) differ in only one marking.
As in No. 3, also for orchestra and singers (there is only one singer in
the present movement) and also in three voices, each voice is represented
by a unique combination of forces. In this way as in most others, voices A
and C are similar (both use oboe and bassoon, bass clarinet, trumpet and
trombone, harp and strings - A is assigned the second harp and uses also
the clarinet, while C uses consistently the first harp as well as flute and
horn) and B is strikingly different, using none of the instruments of the
other two groups (the tenor shares this voice with the vibraphone and
marimba, all three very distinctive timbres in the total combination).
Again, I cannot find a serial basis for the organisation of timbres but will
make a few observations. In voices A and C the durational value 1 is
always played by harp. In voice C rows 4-6 use only harp and strings, and
strings are used very sparingly in the rows on either side of this central
group; all duration 2s in these outer sections are played by winds. In voice
A, conversely, rows 4-6 are played entirely by wind and harp, with the
strings prominent in the outer sections: the first and last row of this voice
are given entirely to strings and harp, and all duration 2s in rows 1-3 and
7-10 are played by strings.
Inevitably this movement is reminiscent of No. 3, since that is the only
movement to combine instruments and voices prior to this one. The
surface kinship is reinforced in many ways. These are the only movements
thus far that have proceeded in a thoroughly linear contrapuntal manner
(although four voices were maintained consistently throughout No. 2, all
the series - of pitch classes, durations and dynamics - were presented in
block fashion), and both are in three voices. An important difference is the
complication of additional tempos in No. 5: this makes a chart of the
progress of the three voices look very different from the one given earlier
for No. 3 (compare Fig. 15 with Fig. 8 above).
Fig. 15 No. 5: Three-voice structure
285
288
290
313
314
318
A
B
C
NO. 6a
... the doors open. There are our murderers. Dressed in black. They
chase us from the synagogue.'
No. 6a is essentially a cancrizans structure in which chorus and orchestra
perform the same music, going in opposite directions, probably as a
reflection of the text. As the assassins enter, the people exit. The idea of
KATHYRN BAILEY
two groups progressing towards each other and crossing at midpoint was
introduced in No. 1, where the complexity of odd-numbered sections
(played by winds) decreased while that of even-numbered sections (played
by strings) increased as the piece progressed.
In No. 6a the two performing bodies, an eight-part orchestra using only
low-pitched instruments with copious doublings (first bassoon with first
and second horns; second bassoon with third and fourth horns; first and
second trombones; third and fourth trombones) and an eight-part choir,
present essentially the same piece, one starting at either end. The piece, in
2 metre throughout, is in four sections, nine, seventeen, seven and a half,
and eleven and a half bars in length, and comprising respectively six, nine,
six and ten durations in each voice, corresponding to the text, which falls
into four sections of six, nine, six and ten syllables (... le porte s'aprono. /
Eccoli i nostri assassini. / Vestiti di nero. / Ci cacciano dalla sinagoya). Three
tempos are used, based on semiquaver quintuplets, semiquavers and
quaver triplets; each voice plays in one tempo throughout each section, but
the distribution and even the number of tempos change between sections.
In terms of pitch, the piece is static: each part (voice, or instrument) simply
reiterates the same pitch throughout each section, and thus no part has
more than four pitches in the course of the piece. In both chorus and
orchestra all eight parts are present only in the section containing ten
durations (the first section for the orchestra, the last for the chorus); the
heaviest density occurs in the middle of the outer sections, when thirteen
parts are present for just over three bars (bs 322-5 and 357-9). Of the eight
voices in the orchestra's first section (and the chorus's fourth), five move in
semiquaver quintuplets and three in semiquavers. In the 'next' section (the
second for the orchestra, the third for the chorus), only six voices are
present, three moving at each of these tempos. The third tempo 'first'
appears in the following section (the third for the orchestra, but the second
for the chorus), where there are again six voices, two moving at each speed.
In the orchestra's fourth section (the chorus's first), one of six voices
moves in semiquavers, two in quintuplets and three in quaver triplets.
Six durations are used. They correspond to the first six numbers of an
arithmetic progression in which successive numbers represent an expanding increment (n+1, 2, 3, 4, 5): 2, 3, 5, 8, 12, 17. I can discern no
regularity in the initial arrangement of these values, but all the tempos and
durations of the orchestral piece occur in exact retrograde in the choral
one. Since the four sections are of unequal length, chorus and orchestra do
not move from one section to the next at the same time. The largest
discrepancy comes at the centre of the movement, where all parts are
playing the same music, half in retrograde, from b.338 to b.345. This is
illustrated in Fig. 16.
A twelve-element series of dynamic indications is used (see Fig. 17).
Although the dynamic markings are independent of the durations, they
also contribute to the retrograde relationship between orchestra and

Fig. 16 No. 6a: Coincidence of parts

section
II
III
IV
bs 319 20 1 2 3 4 5 6 7 8 9 30 1 2 3 4 5 6 7 8 9 40 1 2 3 4 5 6 7 8 9 50 1 2 3 4 5 6 7 8 9 60 1 2 3
B6b
BEB?
C~
AA
E6-I
rn
IAU
[above: bssnl, hns 1/2; bssn2, hns3/4; trbs1/2; trbs3/4; timp1; timp2; vlc; cb]
[below: sl, s2, al, a2, tl, t2, bl, b2]
chorus, though in this case the relationship is complicated by transposition.
Fig. 18 shows the pitch-class, durational and dynamic content of
movement 6a (dynamic indications are represented by the upper set of
numbers, durations by the lower set). The apparently non-systematic series

of numbers representing dynamics and durations are reminiscent of those
seen in No. 1.
Fig. 17 No. 6a: Dynamics series
ppp p mp mf f fff ppp<fff fff>ppp p<f f>p mp<mf mf>mp15

1
10
11
12
The pitch content of this movement is limited to two incomplete

statements of the a-is, one at To and one at T,. This row's To is the same as
its untransposed retrograde, and while the latter interpretation provides a
more accurate expression of the text as well as better reinforcing the
cancrizans structure of the movement, both rows follow the same contour
in No. 6a - fanning out from the centre in a manner corresponding to the
way in which the a-is was derived from the chromatic scale, and therefore
to the prime form of the row. Only eighteen pitches are used, the activities
KATHYRN BAILEY
Fig. 18 No. 6a: Durations and dynamics

section 1, instruments (bs 319-30) - section 4, voices (bs 352-63) -
B (bssnl/hns): 9 2 4 5 5 4 4 7 2 6 AI (altol): 9 5 10 6 6 8 8 7 4 12 [(R3)]

3 5 17 17 17 12 12 8 8 12 12 8 8 12 12 17 17 17 5 3 (R)
Bb (bssn2/hns): 12 9 6 1 2 12 11 3 9 7 G (alto2): 12 2 8 4 5 7 6 11 2 5 (R5)

5 2 8 8 12 5 3 17 2 5 5 2 17 3 5 12 8 8 2 5 (R)
A (trbl/2): 10 3 5 1 4 4 6 6 6 6 C (tenorl): 2 4 1 5 3 3 12 4 2 9 [(R11)]

8 8 3 17 5 5 2 12 3 8 8 3 12 2 5 5 17 3 8 8 (R)
Ak (trb3/4): 10 7 11 8 3 9 2 8 1 8 B (bassl): 11 4 11 5 12 6 11 2 10 1 (R3)

17 5 17 8 3 12 5 8 8 8 8 8 8 5 12 3 8 17 5 17 (R)
A (timpl): 4 5 5 3 6 8 9 10 8 6 C (bass2): 3 5 7 6 5 3 12 2 2 1 (R9)

8 5 5 12 3 2 3 5 2 5 5 2 5 3 2 3 12 5 5 8 (R)
At (timp2): 6 2 4 1 8 5 6 10 1 12 B (tenor2): 1 1 11 7 6 9 2 5 3 7 (R1)

2 8 3 12 5 2 2 12 17 17 17 17 12 2 2 5 12 3 8 2 (R)
B (vlc): 1 5 12 1 7 5 12 1 2 5 At (sop2): 6 3 2 1 6 8 2 1 6 2 (R1)

3 17 2 3 12 17 2 3 5 17 17 5 3 2 17 12 3 2 17 3 (R)
B6 (cb): 8 7 8 4 2 10 1 3 11 8 G(sopl): 3 6 10 8 5 9 11 3 2 3 (R7)

2 3 2 12 17 3 12 17 5 2 2 5 17 12 3 17 12 2 3 2 (R)
C (bssnl/hns): 2 4 8 6 4 6 F# (sopl): 9 7 9 11 3 5 [(R3)]

8
12
17
12
12
17
12
(R)
G (bssn2/hns): 10 6 11 10 9 11 F (sop2): 4 2 3 4 11 3 (R5)

2 8 3 22 3 3 2 2 3 8 2 (R)
G (trbl/2): 7 2 11 10 12 12 F (tenor2): 7 7 5 6 9 2 (R7)

3
17
17
(R)
F# (trb3/4): 2 4 12 3 6 5 C# (altol): 2 3 12 9 1 5 [(R9)]

5
12
12
12
12
(R)
F# (timpl): 6 3 1 1 9 3 C# (bassl): 2 8 12 12 2 5 (R11)

2
17
17
17
17
(R)
C (timp2): 7 3 4 9 2 10 F# (tenorl): 11 3 10 5 4 8 (R1)

12
12
12
12
(R)
cont. on p.307
of both chorus and orchestra being restricted to the range of a sounding

major sixth: all the chorus notes fall between b and ab', all those of the
orchestra between F# and eb. The chorus sings only one note, e?1, in its
section 1, continuing with two in section 2 (e' and d', notes 2 and 3 of
R/T6),
three
(f',inc#'
and f'1,
notes
4-6) in
section
and four
g', row
b and
ab', notes
7-10)
section
4. The
pitches
needed
to 3finish
the (c1,
choral
(b? and a') do not occur anywhere in the movement.
The instrumental row, PO, is less straightforward. The orchestra plays a
cluster of four pitches in its section 1 (notes 1-4 of the row: A, Bd, A6 and
Fig. 18 cont.
A (bssnl/hns): 7 3 9 2 7 10 1 5 3 D (bassl): 4 6 2 11 8 3 10 4 8 (R1)

12 8 5 5 12 3 3 17 8 8 17 3 3 12 5 5 8 12 (R)
E, (trbl/2): 1 11 11 7 4 9 5 7 11 D (tenorl): 10 6 4 8 3 6 10 10 12 (R11)

17 12 12 2 5 5 3 2 12 12 2 3 5 5 2 12 12 17 (R)
E (timpl): 3 8 2 11 4 12 6 12 10 E (sopl): 5 7 1 7 11 6 9 3 10 (R7)

17 2 17 5 12 8 5 8 3 3 8 5 8 12 5 17 2 17 (R)
Ed (timp2): 9 4 5 7 1 3 10 7 6 D (bass2): 3 4 7 12 10 4 2 1 6 (R9)

3 8 5 2 17 12 5 2 3 3 2 5 12 17 2 5 8 3 (R)
A (vlc): 9 2 10 1 10 11 6 9 4 E (altol): 9 2 11 4 3 6 3 7 2 (R5)

2 12 2 8 2 3 8 2 17 17 2 8 3 2 8 2 12 2 (R)
A (cb): 3 [11] 5 4 9 4 7 5 7 E (sop2): 10 8 10 7 12 7 8 2 6 (R3)

12 2 17 17 3 17 8 17 8 8 17 8 17 3 17 17 2 12 (R)
B6 (bssnl/hns): 9 5 10 8 2 5 EF (bass2): 2 11 5 7 2 6 (R9)

3
17
17
(R)
B6 (trbl/2): 3 8 5 11 5 6 E, (tenor2): 5 4 10 4 7 2 (R11)

8
12
12
(R)
B6 (trb3/4): 8 11 7 2 5 4 Eb (tenorl): 11 12 9 2 6 3 (R7)

2
17
12
12
17
(R)
Bb (timpl): 4 6 12 7 8 1 Eb (altol): 2 9 8 1 [7] 5 (R1)

12
17
12
12
17
12
(R)
Bb (vlc): 6 3 10 11 1 [2] E (alto2): 5 4 2 1 6 9 (R3)

12
12
12
12
(R)
Bb (cb): 3 11 7 4 1 [3] Eb (sopl): 8 6 9 12 4 8 (R5)

17
17
17
17
17
17
(R)
B), and three (notes 5-7: G, c and F#) in its section 2; the two notes it
plays in section 3, however, represent the two extremes of the row (A and
el) rather than the continuation that might be expected, and the single note
played in section 4 is B%, a note already heard in section 2. The two
redundant notes, A in section 3 and the final B% (notes 1 and 2 of the
orchestral row), are the pitch classes needed to complete the choral row,
though for this purpose, in addition to being played rather than sung, they
are both in the wrong order and out of range.
If this structure is intended to reflect the dramatic content of the text
there seem to be slight discrepancies. It is logical that the choir should

represent the people, and the fact that its music begins with a single note
and subsequently fans out seems to parallel their flight from the synagogue,
especially as its progress falls into disarray at the end of the movement (as
KATHYRN BAILEY
represented by the numerous inconsistencies to be seen in the presentation

of the final two notes of its series). The advance of the assassins might, on
the other hand, be expected to become more focused as it progresses. This
is true of the orchestra inasmuch as it begins with its largest cluster of notes
and plays successively fewer as the movement progresses, coming to a close

on the two central notes of its opening section, A and B%.16 At the same
time, however, it is difficult to account for the fact that the instruments
fall, in sections 3 and 4, into greater disarray than the chorus, abandoning
the progress of their row five notes short of completion to sound instead
the final note of their series (which is spatially removed from the notes
played up to this point) and two others which can be seen either as
representing a circularity in which the end is like the beginning - hardly an
appropriate idea here - or as displaced elements of the choral series, or
both. Although it is fitting that the advance is carried out by the original
row and the flight by its reverse, it seems anomalous that the former is
more chaotic than the latter.
NO. 6b
'How hard it is to say goodbye for ever to such a good life!'
Like No. 2, 6b moves in four voices throughout, but here there are only
two tempos, one based on semiquaver quintuplets, the other on quaver
triplets, and each is used consistently by two of the voices. (This is the way
in which tempos were used in No. 5 but not in the otherwise more closely
related No. 2.) Each of the four voices consists of a section of the chorus
and a section of the string orchestra contributing alternately to an
essentially continuous line but never sounding together: the sopranos and
first violins, and the basses and cellos, play at the quicker tempo, while the
tenors and violas, and the altos and second violins - and the trumpet proceed in triplets. Each of the participating sections exhibits slightly
different characteristics: the first violins play high notes exclusively, mostly
from the middle or upper range extending from b2 to c4, and in all but two
of twelve appearances they play only one note, interjected into a wideranging line (c' to bV2) in which the sopranos ordinarily sing three or even
four notes in succession; the relationship of the cellos (range: e?'-e2) and
basses (F-eL) is similar, although the basses' contributions are not as long
as the sopranos' (characteristically two notes only); the second violins play
single long notes exclusively, and only four of them, two relatively early
and two relatively late in the piece, while their partner the second trumpet
plays only two notes altogether - long high ones (d3 and cy3), in Succession,
near the centre - and the altos sing primarily in groups of notes (three in
succession on three occasions, a pair twice), also of long values; the violas'
notes are arranged mostly in pairs, and fall in a fairly high range (b'-b2),
alternating with the tenors, who sing more single notes in isolation than
any other section of the chorus. Thus all the strings play in a high range,
while the voice ranges are fairly wide.
The number series on which the durations in this movement are based
is similar to that used in No. 2. Here, as there, the original version is a

symmetrical arrangement of the first six numbers of an expanding
progression, in this case the one used to determine the durations in the
companion piece, 6a: 2 3 5 8 12 17. The arrangement of these numbers is
the inverse of the arrangement in No. 2: the initial row is 17 12 8 5 3 2 2 3
5 8 12 17. This series has been submitted to the rotational operation used
in the earlier piece, but here the cycle stops one row short of completion.
In No. 2 the thirteenth row, which would have reproduced the original
series, was avoided, as if having demonstrated the complete cycle were
sufficient without hammering home its circularity. Here the twelfth row is
avoided as well, being represented by its first note alone, which, as it
happens, is the first note of the piece and therefore seems to satisfy the
desire for closure. Although the two movements contain the same number
of bars, there are many fewer notes in 6b (133, compared with the 118 of
No. 2), since this structure not only falls short of a complete rotational
cycle but also lacks the retrograde group and coda that followed the
completion of the cycle in No. 2 (compare Fig. 19 with Fig. 3).
Fig. 19 No. 6b: Table of durations
1) 17 12 8 5 3 2 2 3 5 8 12 17
2)
12
3)
12
17
17
12
17
17
12
4) 5 3 2 2 3 5 8 12 17 17 12 8
5) 3 2 2 3 5 8 12 17 17 12 8 5
6) 2 2 3 5 8 12 17 17 12 8 5 3
7) 2 3 5 8 12 17 17 12 8 5 3 2
8) 3 5 8 12 17 17 12 8 5 3 2 2
9) 5 8 12 17 17 12 8 5 3 2 2 3
10) 8 12 17 17 12 8 5 3 2 2 3 5
11) 12 17 17 12 8 5 3 2 2 3 5 8
12) 17 [17 12 8 5 3 2 2 3 5 8 12] bracketed notes

[13) 17 12 8 5 3 2 2 3 5 8 12 17] do not occur
The dynamic indications, also, seem to have been handled in much the
same way as those in No. 2, though a large number of anomalies exist.
After puzzling over these for a considerable time, I have decided that they
must represent alterations made for contextual reasons.'7 Six dynamic

levels are used; these have been arranged symmetrically to produce the
series in Fig. 20. The table of dynamic markings in this movement (see
Fig. 21) resembles in several particulars the one in Fig. 5 above. In the
KATHYRN BAILEY
Fig. 20 No. 6b: Dynamics series

ppp pp p mp mf f f mf mp p pp ppp
1
10
11
12
table for No. 2 the numbers in all the columns were ordered one to twelve,
although the columns started at various points in the series. Here, as can
be seen in Fig. 21, the same thing nearly happens: while there are too many
'wrong' numbers (36 of 133: these are in bold face in Fig. 21, and the
numbers they replace are given in brackets at the end of each line) to
suppose all of them to be errors, much too much of the table conforms to
the pattern to allow the dismissal of the correspondence as simply
fortuitous.'" I can offer no explanation for the irregularity in the order of
the first row, which begins like that of No. 2 but does not continue in the
same way (although seven numbers occupy the same position as in the
earlier table - compare the first row in Fig. 21 with that given in Fig. 6).
Fig. 21 No. 6b: Table of dynamics
1)
A Bb A, B G C FO CO F D E E6
1
2)
3)
4)
5)
6)
5
6
5
6
7
8
10
11
10
9
8
(2,
8
10
11
(4,
(1,
3)
4,
2,
4)
6,
2)
(1,1,3,3)
10
(6,
(2,
3)
5,
2)
7) 7 9 8 7 12 8 10 9 11 9 12 11 (4, 1, 1, 3)
8) 8 10 9 8 1 9 11 10 12 10 1 12 (6, 6, 4, 3)
9) 9 11 10 9 2 10 (12) 11 (1) (11) (2) (1) (4, 2, 2)
10) 10 12 11 10 3 11 1 12 2 12 3 2 (4, 4, 1)
11) 11 1 12 11 4 12 2 1 3 1 4 3 (3, 3, 6, 2)
12) 12
NB. Numbers in bold face do not fit the scheme. They can be seen as errors with varying
degrees of success: a misprint involving two numbers that are adjacent on the scale is more
likely than one involving two non-adjacent numbers. Brackets indicate notes that have no
dynamic indication in the score. It is assumed that each of these is to be played at the same
level as the note directly preceding in the same voice, but this creates some ambiguity, since
it is also conceivable that omission might be involved.'"
The pure vowel sounds used in this movement are also arranged
symmetrically (and serially). Each of the five vowels appears four times in
the body of the piece, and in each instance all five are heard before any is
repeated. The order u i e o a is established in bs 364-72. The next
appearance, in retrograde order, is scattered over ten bars, 375-85. The
series appears twice more, first forward and then in retrograde, in bs 386310 MUSIC ANALYSIS 11:2-3, 1992
93 and 396-407. A coda (bs 407-9) repeats u, the last vowel heard, two
more times.
NO. 7
'... goodbye, Mother, your daughter Liubka goes into the damp
earth.'
In No. 7 up to three voices progress at a time, all in the same tempo, using
the semiquaver as the basic unit. Statements of the row are linear once
again in this movement; this linearity, as well as the number of voices,
immediately connects this movement with Nos 3 and 5, both of which are
written, as is No. 7, for solo voice(s) and orchestra; the consistency of
tempo relates particularly to No. 3. Twenty-four twelve-note rows are
presented one, two or three at a time in twelve clearly distinguished
sections. Both pitch-class and durations rows are submitted to a-is
permutation.
The durations series on which the movement is based consists of
identical hexachords derived from a 1 6 2 5 3 4 permutation of the first six

numbers of the Fibonacci series (a shorter version of the permutation that
produces the a-is from the descending chromatic scale A to Bb: 1 12 2 11 3
10 4 9 5 8 6 7). The resulting row, which reads 1 12 2 8 3 5 1 12 2 8 3 5,
appears alone in bs 414-20. In the following section (bs 420-6) three voices
proceed in counterpoint, each presenting a single row. These three rows,
which represent a reading of, respectively, alternate numbers, every third
and every sixth number of the original row, are characterised by repeating
trichords, dyads and single units (see Fig. 22).
Fig. 22 No. 7: One possible derivation of the three durations series in bs
420-6
Permutation 1 12 2 11 3 10 4 9 5 8 6 7 (retrograde a-is) of

series 1 2 3 5 8 12 * 1 2 3 5 8 12 (Fibonacci nos) gives
series 1 12 2 8 3 5 1 12 2 8 3 5 (bs 414-20).
Then...
permutation:
2468 10 12* 1 3 5 7 9 11 gives 12 8 5 * 12 8 5 * 1 2 3 * 1 2 3 (s.sop, vlns,...)

36912 * 25811 * 14710 gives 25 * 25 *123 123 * 1 8 * 1 8 (hp, sop, mar,...)
612*511*410*39*28*17 gives 55*33*88*22*1212*11 (alt,fl,vibr,...).
KATHYRN BAILEY
Fig. 23 No. 7: Derivation of durations series

(1)
(2)
12
12
12
2
5
(3)
(4)
12
12
row
12
12
row
12
row
(5)
12
12
row
(6)
12
row
1
3
12
row
120
Two further durations series occur in the third section (bs 426-32). The
first of these represents the reading of every fourth number of the original
series (4 8 12 3 7 11 2 6 10 1 5 9), so completing the set of derivatives

begun in the previous section. The result of this permutation is not a
repeating series like those in Fig. 22; it is a less symmetrical and therefore
in a way a more complex rearrangement of the original series than those
that have preceded it: 8 12 5 2 1 3 12 5 8 1 3 2. The second row of this
section is the least regular of all: 12 2 3 5 1 2 8 5 1 12 8 3. This row cannot
be related directly to the original row but represents a 2 4 6 8 10 12 1 3 5 7
9 11 permutation of the row that directly precedes it, and this relationship,
finally, supplies the key to the real etymology of the lot. In fact, of course,
the rows described above as representing various permutations of the same
row (2 4 6..., 3 6 9..., 4 8 12..., 6 12 5 11...) are all produced by
repeated applications of the single permutation 2 4 6 8 10 12 1 3 5 7 9 11.
Their true relationship is not immediately apparent because they have been
reordered. In Fig. 23 they are placed in the order of their derivation. It can
be seen that these six rows first appear in the order 1 6 2 5 3 421- in other
words, they have been ordered according to the same permutation that was
used to determine the hexachordal content of the row initially. They
comprise the complete durations content of the movement; the entire set
occurs three more times: each row again (but in a different order) in bs
430-51, and each in retrograde form in 451-69 and again, reordered, in
463-88 (rows from both the third and fourth sets are heard together in bs
463-9). In these subsequent appearances the rows occur in the order 3 4 2
5 1 6, 6 1 5 2 4 3 and 3 2 1 4 5 6, all short forms of a-is permutations. It
will be noted that the order of the (retrograde) rows in the third section is
the direct retrograde of that in the preceding section, while the order in the
final section represents a more complex reversal - 5 3 1 6 4 2 - of the first.
The derivation of the pitch content is more difficult to apprehend. The

twelve durations rows and their retrogrades are conveyed by twelve pitchclass rows and their retrogrades. The first of these is the result of a 1 6 2 5
3 4 reordering of the first hexachord of the a-is and a corresponding 12 7
11 8 10 9 reordering of the second: A C BI G Ab B EL, F E C) D F. Two

rows are derived from this - one directly, the second from the first -
through the rearrangement both of notes within each trichord and of the
trichords themselves. In the first case the trichords of the original row are
reordered as an expanding series - 2 3 1 4 (a short form of the a-is

permutation 6 7 5 8...) - and the notes within the outer ones turned
around to produce two expanding wedges: 6 5 4 7 8 9 1 2 3 12 11 10. In
the third row the trichords of the second are rearranged as just such a
wedge - 2 1 3 4 - while the internal order of each is changed - differently
in every case, so that those in the first hexachord represent expanding, and
those in the second contracting series: 5 6 4 2 1 3 7 9 8 12 10 11.
A second row, C# G D B A F E6 A6 B6 C E F, also appearing within the
first three sections (bs 414-32), spawns two offspring in a manner similar
to but less straightforward than that just described. In this case the three
tetrachords of each new row consist of notes that occupied symmetrical
positions in the previous one. The first is a 2 4 11 9 6 3 7 10 1 5 12 8
reordering, the second 4 9 3 10 5 8 2 11 6 7 (1 12 - the last two notes of
this row are missing in the score).
Thus are the first six pitch-class rows arrived at. Like the durations
rows, however, they first appear in the order 1 6 2 5 3 4. The remaining six
come about in an exactly parallel way and occur in the order 9 10 8 11 7
12, thus corresponding with the 3 4 2 5 1 6 recurrence of the durations
rows to which they are necessarily attached. The derivation of all twelve
pitch-class rows is shown in Fig. 24. (I cannot suggest a source for the
Fig. 24 No. 7: Derivation of pitch-class series22
[a-is: A Bk A6 B G C F# C# F D E E]
(1) A C Bk G A6 B Et F# E C D F
1
12
11
10
(2) B A6 G E FF E A C B F D C#
6 5 4/ 7 8 9 / 1 2 3 / 12 11 10
(3) F# E EE A B G A Bk C C# F D
5 6 4/ 2 1 3 / 7 9 8 / 12 10 11
(4) Cg G D B A Ft Et A6 Bk C E F
(5) G B E B F# D Et C C A F A6
2 4 11 9 / 6 3 7 10 / 1 5 12 8
(6) Bk C# E A F# C B F D Et (G A6)
4 9 3 10 / 5 8 2 11 / 6 7 (1 12)
(7) C A6 E Ct B E, D A G F B F"
(8) E, B C# D A G C AB E F B F
6 5 4 / 7 8 9 / 1 2 3 / 12 11 10
(9) A G D B Et C# C E A6 F FF Bb
5 6 4 / 2 1 3 / 7 9 8 / 12 10 11
(10) G E Et A6 F F B C D A Bk C
(11) E A6 Bk D F Et B A G F# C C
2 4 11 9 / 6 3 7 10 / 1 5 12 8
(12) D (G) Bb F F A A6 C# EB B E C
4 9 3 10/ 5 8 2 11 / 6 7 1 12
KATHYRN BAILEY
generating rows 4, 7 and 10.) In the second half of the movement, bs 45188, the association of pitch-class and durations series remains the same;
the twelve rows heard in bs 414-51 are heard in retrograde, in the order
121110987 531642.23
Only three dynamic indications are used in this movement: ppp, p and
mf. While this is the smallest of the pools of material represented compared with six durations and twelve pitch classes - the way in which
they are employed seems somehow to embody the patterns used in the
other two row structures.
Twelve-element rows are constructed by grouping the three markings in

three ways - 3xppp + 5xp + 4xmJf 5xppp + 3xp + 4xmf; and 4xppp + 4xp +
4xmf - and forming four distinct rows from each group. Referring once
again to the numbers used to represent the order of derivation of the pitchclass rows (see the bracketed numbers in Fig. 23), the dynamics of the
rows numbered (1), (4), (7) and (10) there use the combination 3xppp +
5xp + 4xmf, those numbered (2), (5), (8) and (11) contain 5xppp + 3xp +
4xmf, and each indication occurs four times in those numbered (3), (6),
(9) and (12). In this way twelve rows24 have been built which can be
reordered in conjunction with the pitch-class rows (in the second half of
the piece the retrograde of each of the twelve pc rows expresses the
retrograde of the appropriate dynamics row). Thus the dynamic structure
seems to encompass both the other structures: just as the durations rows
exhibit a tighter organisation than the pc rows (six durations combined
into six discrete series compared with twelve pitch classes in twelve series),
so the dynamics are in one sense even more tightly organised (only three
indications grouped in three ways); yet in the end these three elements lead
to twelve discrete series, thus matching in number those of the pitch
classes. All three structures undergo the same permutational reordering.
The orchestration of this movement seems to be in direct response to
the text, which is from the letter of a young girl and the briefest of the lot.
The small orchestra consists of strings, flutes and harp, together with a
large group of high-pitched colour instruments - glockenspiel, celesta,
vibraphone and marimba. The tessitura is very high: the lowest pitches are
f, f? and g, each played once by the harp; aside from this only the altos
descend below c'. The cellos and basses play in harmonics exclusively. The
interest in unusual colours extends beyond the choice of instruments: the
flutes articulate the piece with harmonics in the first and final pairs of rows
and just before the centre (these are approximately, but not precisely,
symmetrically placed), and, at irregular intervals throughout, by the
fluttertongued doubling of harmonics played by the cellos or violins (in bs
426-7, 432-3, 437-8, 453-4, 456-7 and 472). Most of the singing is done
with closed or half-closed mouth. The instrumental parts are very
scattered: only five times in the course of the piece does an instrument play
two notes in direct succession - the violas in bs 429-30 (continuity is in
this case offset by the fact that the first note is bowed and the second
pizzicato), the second flute in bs 456-7, the violas in bs 457-9, the second
violins in bs 461-3 (the only instance of three notes) and the marimba in
b.467.
NO. 8
No. 8 is the third and last purely instrumental movement, written for
flutes, clarinets, horns, trumpets, trombones and timpani. Although its
surface structure is reminiscent of the first movement, this is as far as the
resemblance goes: this movement contains none of the linear polyphony
that characterised the string sections of No. 1, and, while a texture rather
like that of the wind sections in the earlier piece is maintained throughout,
there are no metre changes, and groups of similar durations are not
gathered together here. Double bars and changes of tempo and instrumentation define eleven sections of varying lengths; in Fig. 25 this
structure is compared with that of No. 1.
The music proceeds at four tempos. In the quickest of these the quaver
divides into seven units; the other three represent divisions into five, four
and three. The texture is not linear as in most of the other multi-tempo
movements (Nos 2, 5 and 6b), where all notes moving at the same speed
connect to form a continuous strand. Neither is there any strict
correspondence between tempos and instruments (as there is, say, in No.
6b) or tempos and structural divisions (as in No. 4), with the exception of
sections 2 and 10, which are very short and entirely in triplet and septuplet
subdivisions respectively. In most cases each instrument plays in only one
tempo for the duration of a section.25
As in Nos 2, 6b and 9, the series of No. 8 unfold in block fashion.

Thirty-six twelve-note series appear as successive clumps of music, each
comprising twelve discrete pitch classes and durations, and the serial
manipulations of these are unquestionably based on the a-is permutations.
With one exception, however, I have not been able to discover the exact
nature of the operations used. The system I have been able to identify is
that used to determine the way in which pitch classes and durations are
paired. As was the case with Nos 3, 4 and 6a, I cannot explain the order of
notes in each row.
Durations are expressed as notes worth one to twelve semiquavers (in

the triplet tempo) or demisemiquavers (in all the others), or as groups of
two to six repeated notes of the same duration. Within the block of time
given to each row, notes appear and disappear in apparently free fashion.
Both the order of notes within the rows and the sequence of the rows
themselves give the appearance of having been somewhat randomly
determined. The untransposed a-is appears on seven very unevenly spaced
occasions - as the third, sixth, eleventh, eighteenth, twenty-fourth, twenty-
fifth and twenty-sixth rows, in bs 493-5, 498-500, 506-8, 516-17 and

KATHYRN BAILEY
Fig. 25 Nos 1 and 8: Structures compared

No. 1
section
no.
I I I I i
(ii)
bars
III
iv
VI
vii
VIII
IX
(xi)
15
XII
XIII
20
19
speed (mm) 92 60 92 60 92 72 72/92 82 92 92 60 92 72/56/72/60

no.
tempos
(1)
(1)
metre 4/8 3/4 4/8 3/4 4/8 3/4 4(&6)/8 3/4 4/8 4/8 2/4 4/8 4/8
instr. fl, 2vns wind, str wind, str ww, str brass, all timp all ww, str
brass,
timp
timp
timp
hn,
timp
trp
No. 8
section
no.
ii
bars
III
iv
vi
VII
viii
IX
XI
15
speed (mm) 92 rall. 112 92 112 66/60 72 rall. 56 46 92

no.
tempos
metre 4/8 -- -throughout ---
instr. fl, 3 hns, ww, (cl), fl, fl, all 2 tbns, fl, fl, fl,
brass, timp tbn, trp, trp, timp timp hn, trp hns,
timp
timp
tbn
tbn,
timp
trp,
timp
trp,
tbn,
timp
524-31 - and the remaining twenty-nine rows show no tendency to

regularity: they are not the product of symmetrical permutation, nor are
they the conventional variants (inversion, retrograde, transposition) of the
a-is. Successive rows tend to retain the same pitch classes in many
positions (it is particularly noticeable that the first nine rows begin on A
and end on E6, with B6 occupying the second position in eight of these and
E the penultimate position in seven; subsequently thirteen more begin A B6
and fourteen end E EB); this pitch-class stickiness results in a large number
of rows that are nearly but not quite the a-is (see the table of pitch-class
rows in Fig. 26).
An examination of the durations discloses a similar situation. A
comparison of the thirty-six durations rows with the permutation tables

associated with the a-is (see Fig. 27) shows too many similarities for this
relationship to be coincidental, but too many anomalies to be treated as
Fig. 26 No. 8: Pitch-class series
(1)
(2)
(3)
(4)
(5)
bs
bs
bs
bs
bs
489-91 A B% B Ab G Fo C Co D F E Ek
491-3 A Bk B Ab C G Fo Co F E D E6
493-5 A B6 A6 B G C Fo Co F D E E6 a-is
495-7 A B A6 B6 G C Co Fo F D E E6
497-9 A B, Ab B G C F# Co D F E E6
(6) bs498-500 A B, A6 B G C Fo Co F D E E, a-is

(7) bs 500-3 A B6 A6 G B C Fo Co F E D Et
(8) bs 502-4 A B, A6 B G C Fo Co D F E E6
(9) bs 504-5 A B A6 B Co C Fo F G D E E6
(10) bs 505-7 B6 A A6 B G C Fo Co F D E E6
(11) bs 506-8 A B, A6 B G C Fo Co F D E E6 a-is
(12) bs 507-9 B6 G A6 B A C F# C# F D Et E
(13) bs 509-10 A B B6 At G C Fo Co F D E E6
(14) bs 510-11 At B A B6 G C Fo Co F D E Et
(15) bs 511-12 B6 A At G F# B C C# D F E E6
(16) bs 512-14 A B6 At B C Fo G Co F D E EF
(17) bs 514-16 A B6 At B C Fo G Co F D E E6
(18)
(19)
(20)
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
bs 516-17 A B6 At B G C Fo C# F D E E6 a-is
bs 517-18 A B6 A, B G C Co F D E E6 Fo
bs 517-20 A B, At B G C Fo Co F D Et E
bs 519-20 B%, At A B G C Fo Co F D Et E
bs 520-1 [A] B6 At G B C F Co F D E E6
bs 521-4 A B~ A, G B C Fo Co F D E EtF
bs 524-6 A [B,] A, B G C F C# F D E FE a-is
bs 526-9 A B, A, B G C F C# F D E E, a-is
bs 528-31 A B, At, B G C F C# F D E E, a-is
bs 531-2 A B, Ab B F# C# G D F C E E,
bs 532-4 A A, Fo G B Co B, C E6 F D E
(29)
bs 535-6
534-5 C
A A,
A, AB,Fo
F#
Co B
GG
C E6
F E_
(30) bs
CoBB6,
F DD EE
(31) bs 536-8 C A Bk, Fo Co A, B G F E D E,
(32) bs 537-9 A A6 B B, C G Co Fo F E6 E D
(33) bs 539-41 A Ab B, B Fo Co G C F E, D E
(34) bs 540-2 C A, A Fo Co B, B G F E, D E
(35) bs 541-3 C Fo Co A B, Ab B G F D Et E
(36) bs 543-4 A B C A, G D E, B, F# E F C#
simple mistakes or misprints. The eighteen rows in the first and second
sections closely follow the 7 6 8 5 ... table, rows 1 to 9 presenting the rows
of that table in retrograde, and 10 to 18 in forward-going fashion. The
eighteen rows of sections three and four bear a similar relationship to the
rows of the 12 10 8 6 ... table. The seven rows that present the pitch
KATHYRN BAILEY
classes in the a-is order also contain no durational aberrations: in these
rows the series of durations read the same as the corresponding rows on
the permutation tables (these rows are printed in bold face in Fig. 27). In
most instances anomalies take the form of two, or perhaps three or four,
numbers exchanging positions directly - but not systematically. In terms of
both pitch and durations, the final section (rows 28-36) is more chaotic
than any of the previous ones.
Although the close adherence of the durations rows to these
permutation tables makes it plain that the order within the rows has been
in some way determined by them, I can offer no explanation for the
numerous deviations from the table or, therefore, for the sequence of notes
in the thirty-six rows.
What I can explain is the rather circuitous route by which the

correspondence of pitch classes and durations has been determined. For
this purpose the thirty-six rows do represent four complete cycles of the a-
Fig. 27 No. 8: Durations content compared with permutation tables

durations rows permutation tables
(1) 1 12 11 2 3 4 10 9 8 5 6 7 - 7 6 8 5 9 4 10 3 11 2 12 1
(2) 7 1 12 6 2 8 5 11 9 4 3 10 10 4 3 9 11 5 2 8 12 6 1 7
(3) 10 7 4 1 3 6 9 12 11 8 5 2 2 5 8 11 12 9 6 3 1 4 7 10
(4) 2 7 5 10 8 4 1 11 12 3 9 6 6 9 3 12 1 11 4 8 7 5 10 2
(5) 6 2 9 10 3 5 12 7 8 1 11 4 4 11 8 1 7 12 5 3 10 9 2 6
(6) 4 6 11 2 8 9 1 10 7 3 12 5 5 12 3 7 10 1 9 8 2 11 6 4
(7) 5 4 12 3 6 11 7 2 10 1 8 9 9 1 8 10 2 7 11 3 6 12 4 5
(8) 9 5 1 4 8 12 10 6 3 2 7 11 11 7 3 2 6 10 12 8 4 1 5 9
(9) 11 9 7 5 4 1 2 6 3 8 10 12 12 10 8 6 4 2 1 3 5 7 9 11
(10) 6 7 8 5 9 4 10 3 11 2 12 1 - 7 6 8 5 9 4 10 3 11 2 12 1
(11) 10 4 3 9 11 5 2 8 12 6 1 7 10 4 3 9 11 5 2 8 12 6 1 7
(12) 5 12 8 11 2 9 6 3 1 4 10 7 2 5 8 11 12 9 6 3 1 4 7 10
(13) 6 12 9 3 1 11 4 8 7 5 10 2 6 9 3 12 1 11 4 8 7 5 10 2
(14) 8 1 4 11 7 12 5 3 10 9 2 6 4 11 8 1 7 12 5 3 10 9 2 6
(15) 12 5 3 10 9 7 1 8 11 2 6 4 5 12 3 7 10 1 9 8 2 11 6 4
(16) 9 1 8 10 7 11 2 3 6 12 4 5 9 1 8 10 2 7 11 3 6 12 4 5
(17) 11 7 3 2 10 12 6 8 4 1 5 9 11 7 3 2 6 10 12 8 4 1 5 9
(18) 12 10 8 6 4 2 1 3 5 7 9 11 12 10 8 6 4 2 1 3 5 7 9 11
(19) 11 9 7 5 3 1 4 6 8 10 12 2 p 12 10 8 6 4 2 1 3 5 7 9 11
(20) 9 5 1 4 8 12 10 6 2 3 11 7 11 7 3 2 6 10 12 8 4 1 5 9
(21) 4 12 5 6 3 11 7 2 10 8 9 1 9 1 8 10 2 7 11 3 6 12 4 5
(22) 4 6 11 8 2 9 1 10 7 3 12 5 5 12 3 7 10 1 9 8 2 11 6 4
(23) 6 2 9 3 10 5 12 7 1 8 11 4 4 11 8 1 7 12 5 3 10 9 2 6
(24) 2 10 5 7 8 4 11 1 12 3 9 6 6 9 3 12 1 11 4 8 7 5 10 2
(25) 10 7 4 1 3 6 9 12 11 8 5 2 2 5 8 11 12 9 6 3 1 4 7 10
(26) 7 1 6 12 8 2 5 11 9 3 4 10 10 4 3 9 11 5 2 8 12 6 1 7
(27) 1 12 2 11 4 9 3 8 5 10 6 7 7 6 8 5 9 4 10 3 11 2 12 1
(28) 12 8 1 4 6 3 10 2 11 5 7 9 - 12 10 8 6 4 2 1 3 5 7 9 11
(29) 11 3 7 12 2 8 6 10 4 9 1 5 11 7 3 2 6 10 12 8 4 1 5 9
(30) 7 8 9 11 3 1 10 2 5 6 12 4 9 1 8 10 2 7 11 3 6 12 4 5
(31) 1 5 12 9 8 3 7 10 2 6 11 4 5 12 3 7 10 1 9 8 2 11 6 4
(32) 4 8 1 11 12 7 3 5 10 6 2 9 4 11 8 1 7 12 5 3 10 9 2 6
(33) 6 3 9 12 4 8 1 11 7 2 5 10 6 9 3 12 1 11 4 8 7 5 10 2
(34) 9 8 2 6 3 5 11 12 1 10 4 7 2 5 8 11 12 9 6 3 1 4 7 10
(35) 5 2 8 10 4 3 9 11 12 6 7 1 10 4 3 9 11 5 2 8 12 6 1 7
(36) 7 5 4 8 9 2 1 6 10 12 11 3 7 6 8 5 9 4 10 3 11 2 12 1
is permutation, and with no anomalies, but this fact is obscured by

subsequent reordering. Rearranged so that the durations express a
graduated scale from 1 to 12, the pitches of the first row form the
descending chromatic scale A-Bb. This can be seen to represent a 12 10 8 6
... permutation of the retrograde of the a-is (or of the a-is itself at T6).
When the second row is reordered in the same way the pitch classes
produce a 12 10 8 6 ... permutation of the first. Pitch classes and
durations are coupled in the following thirty-four rows in such a way that
when the durations of all rows are reordered to read 1 to 12 the pitch
classes undergo four complete permutational cycles, two using the pattern
12 10 8 6 ... and two its obverse, 7 6 8 5 .... Rows 1-9 (bs 489-504)
represent a 12 10 8 6 ... permutation of the retrograde of the a-is; rows
10-18 (bs 505-17) the same permutation of the forward-going version. In
rows 19-27 (bs 517-31) the retrograde of the a-is is put through a complete
7 6 8 5 ... cycle (the retrograde of the 12 10 8 6 ... cycle used on the
previous nine rows), and the same operation is applied to the original a-is
in the nine rows of the final section (bs 531-45; again a retrograde of the
permutation used in the corresponding section 1, bs 489-504). In all four
cycles the first and last rows (the initiating a-is and the return to it) are
omitted, so no rows are repeated. Large-scale repetition does occur,

however: the rows of sections 3 and 4 are the same as those of,
respectively, sections 1 and 2, in reverse order. In Fig. 28 all the rows of
section 1 have been reordered so that the durations read from one to
twelve; a comparison of this figure with Figs 26 and 27, and with the first
table in Fig. 1 above, will show how the pitch/durations couplings have
Fig. 28 No. 8: Determination of pitch class/duration couplings in section
1; rows 1-9 reordered to show the permutation of pitch classes
pitch-class rows rearranged so durations read 1-12:

1 2 3 4 5 6 7 8 9 10 11 12
[EI E D F C F# C G B Ab Bb A] [Nono's a-is in retrograde]
(1) A Ab G F# F E E D C# C B B% (descending chromatic scale)

(2) Bb C D E F Ab A G F E C# B (whole-tone scales)
(3) B E6 G Ab E C Bb D F# A F C# (augmented triads)
(4) C# A D C Ab E6 B G E Bb F F
(5) F Bb G Eb C A C D Ab B E F#
(6) F# B D A Eb Bb F G C C Ab E (six p5s/p4s)
(7) E C# G B6 A B F# D E6 F C Ab
(8) Ab F D B B6 C# E G A F E, C (diminished triads)
(9) C F# G C# B F Ab D Bb E A Eb (alternate intervals tritones)
[Eb E D F C F# C G B Ab Bb A] [Nono's a-is in retrograde]
KATHYRN BAILEY
been arrived at. This relationship can of course be observed from the
opposite direction, by reordering all the rows in each section so that the
sequence of pitches rather than the sequence of durations is constant. If all
the rows in section 1 are arranged to follow the order of the retrograde a-is
(or the a-isT6) the durations will produce a complete cycle of the
permutation 7 6 8 5 . . . (see Fig. 29). Section 2 produces the same cycle
when the notes are arranged in the order of the forward-going a-is at its
original tonal level. The durations of sections 3 and 4 represent complete
cycles of the 12 10 8 6 ... permutation, section 3 around the a-isT6 and
section 4 around the a-isTo.
Fig. 29 No. 8: Determination of pitch class/durations couplings in section
1; rows 1-9 reordered to show the permutation of durations
durations rows rearranged so pitches read as in a-is:
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
A B6 Ab B G C F C# F D E E6
7 6 8 5 9 4 10 3 11 2 12 1
10 4 3 9 11 5 2 8 12 6 1 7
2 5 8 11 12 9 6 3 1 11 7 10
6 9 3 12 1 11 4 8 7 12 10 2
4 11 8 1 7 12 5 3 10 1 2 6
5 12 3 7 10 1 9 8 2 11 6 4
9 1 8 10 2 7 11 3 6 10 4 5
11 7 3 2 6 10 12 8 4 1 5 9
12 10 8 6 4 2 1 3 5 7 9 11
I am unable to suggest anything concerning the organisation of

dynamics in this movement. Fifteen indications are used, seven simple
ones (ppp, pp, p, mp, mff, , fff) and seven compound ones (ppp<fff, fff>ppp,
p<mf, mf>p, p<f, f>p, mp<mf and mf>mp), thus making the thought of any
twelve-element serial operations impossible. There is also considerable
repetition of the same dynamic indicator within individual row statements.
All in all, despite one very interesting discovery, No. 8 has proven to be
resistant to my efforts.
NO. 9
'... I have no fear of death.. .'
'... I shall be calm and collected in front of the firing squad. Are they
as tranquil as those whom they have condemned? ...'
'... I go with faith in a better life for you ...'
The similarity of Nos 2 and 6b was noted earlier. A much closer

relationship exists between Nos 2 and 9, the latter seeming, indeed, to

represent some sort of recomposition of the former. Like both 2 and 6b,
No. 9 is in i metre and, although it lasts considerably longer than either
(the greater extent of its sixty-one bars compared with their fifty is further
enhanced by its slower tempo: [Q]=54 compared to [Q]=60-66 and 66
respectively), its structure closely replicates that of No. 2. It is written for
divisi chorus,26 a cappella as in No. 2, except for the presence here of

timpani. No. 9 proceeds in three tempos, based on the quintuplet
semiquaver, the ordinary semiquaver and the triplet quaver; in this respect
it occupies a middle ground between No. 2, with four tempos, and 6b,
with two. In most ways, however, it is considerably more complex than
either. Rows of pitch classes, durations and dynamics are expressed in
block fashion throughout, as in the two earlier pieces, but here the tempos
do not form continuous strands; instead of a four-voice structure we are
presented with a series of moments in which density increases and
subsides, each beginning with one or two voices and swelling to as many as
nine at once (as on the second beat of b.573). Thus the microtexture
reflects the macrotexture: the constant fluctuation in the number of voices
heard at once is similar to the continually changing number of voices
participating. There are eighteen rows, only one less than the nineteen of
No. 2.27
Probably the most significant difference between this movement and the
others is that it is organised almost entirely back to front. If we are to
understand the logic controlling both the order of notes within the rows
and the order of the rows themselves, all but the last three - the coda must be read from end to beginning, since the order in which notes are
released is the defining feature (read from right to left). (Because of the
variation in the length of notes, this succession of releases, when read from
start to finish, does not produce retrograde statements of the a-is.)
As in the earlier pieces, the durations are derived from the first six
numbers of a series, in this case the series of prime numbers, and here
again two identical groups have been arranged symmetrically to produce a
row of twelve units - 13 11 7 5 3 1 1 3 5 7 11 13 - but now the generating
series is not used as the first row in the cycle. The rotation used in Nos 2
and 6b is replaced by more complex permutation. Two permutations, in
fact- 12 1 11 2 10 3 9 4 8 5 7 6 and7 6 8 5 9 4 10 3 11 2 12 1 - both
operating from back to front, beginning with rows 15 and 5. Each of these
rows has been obtained by the application, to both hexachords of the
generating series in turn, of a scaled-down version of the permutation to be

used in the ensuing cycle (thus, for example, the permutation 6 1 5 2 4 3 a short version of 12 1 11 2 10 3 9 4 8 5 7 6 - applied to both hexachords
of the series 13 11 753 11 357 1113 gives 1122 11 57 12 111 357,
the row (15) which is the structural beginning of the piece). Because of the
relationship of the two permutations used, rows 5-1 (sic: see Fig. 30) are
the retrogrades of 15-11, just as the parallel rows 13-15 were the reverse of
KATHYRN BAILEY
10-12 in No. 2 (see Fig. 3). The permutation table in Fig. 30 represents a
reading of the piece from end to beginning, with notes listed according to
the order in which they are released (but see the qualifying note).
A comparison of Fig. 30 with Fig. 3 shows the close structural similarity
(in retrograde) of movements 2 and 9. No. 9 falls into three sections
defined by the treatment of the durations series; these sections are exactly
parallel to those in No. 2, with the important exception of the coda, which
by definition comes at the end of both movements, and is thus misplaced
structurally in No. 9 (shown by the fact that while the coda appears at the
end of Fig. 3, it comes at the beginning of Fig. 30). Thus the 'starting' row
- the one from which subsequent ones are derived - appears, in Fig. 30,
not as the first, but as the fourth - the first row in the main body of the
movement, the section following the coda. (In performance, of course, it is
the last heard in this section, and directly precedes the coda.) In both Nos 2
and 9 the main section (the second in Fig. 30, the first in Fig. 3) contains
one complete cycle (the rotation used in No. 2 requires twelve rows for
completion, the permutation in No. 9, only ten). In the tables this is
followed in both cases by a shorter section (three rows in No. 2, five in No.
9 to compensate for the shorter first cycle), which begins to go through the
process in retrograde (in performance this is the second section of No. 2,
the first of No. 9). In both pieces the coda behaves differently from the
preceding sections. The coda of No. 2 was discussed on page 291; that of
No. 9 was conceived in the same way. It is of particular interest, in the
context of this movement, that in this section alone the forward order of
the notes (the attacks, rather than the releases) is definitive: the music in bs
595-605 must be considered from start to finish, in the usual way. In these
bars the durations progression 13 11 7 5 3 1 1 3 5 7 11 13 occurs in each
of the three tempos used in the movement. Here, finally, is the generating
series of the piece.
As the permutation used in No. 9 is more complex than the rotation in

No. 2, so is the process of reversal between the two sections of the piece
that follow permutational procedures. When the 12 1 11 2 10 3 9 4 8 5 7 6
cycle has been completed (as is the custom in II canto sospeso, just short of
repeating the original row), the reverse of this permutation - 6 7 5 8 4 9 3
10 2 11 1 12 - is applied to the last row. This results in the series 5 7 3 11
1 13 7 5 11 3 13 1, which is then subjected to the second permutation
used in the movement, 7 6 8 5 9 4 10 3 11 2 12 1. (This row also
represents the application of a short form of this permutation to both
hexachords of the generating row.) The new permutation, itself essentially
a reversal of the first one, is applied four more times.
Similarly, the dynamic indications of No. 9 have been determined by a
more obscure version of the method that was used in No. 2 and
complicated considerably in 6b. Concerning the derivation of dynamics in

No. 9, I can in fact offer nothing beyond a progress report. Twelve
indications occur: the six fixed markings used in both the earlier

A B6 A B G C F? Ct F D E EB
(18) 7 1 1 3 5 11 11 7 11 13 13 13
(17) 3 1 1 1 5 3 3 5 3 1 7 5 CODA
(16) 13 13 13 11 11 11 7 7 7 5 5 3
(15) 1 13 3 11 5 7 13 1 11 3 7 5 permutation 1:
(14) 5 1 7 13 3 3 11 11 1 5 13 7 I I I I I1
(13) 7 5 13 1 5 7 1 13 11 3 11 3 12 1 11 2 10 3 9 4 8 5 7 6
(12) 3 7 11 5 3 13 11 1 13 5 1 7 I I I I I I
(11) 7 3 1 7 5 11 13 5 [1] 3 11 13
(10) 13 7 11 3 3 1 1 7 5 5 11 11
(9) 11 13 13 7 5 11 5 3 7 3 1 1
(8) 1 11 1 13 [3] 13 7 7 3 5 5 11
(7) 11 [1] 5 11 5 1 3 13 7 3 7 13
(6) 13 11 7 1 3 5 7 11 13 5 3 1
[1 13 3 11 5 7 13 1 11 3 7 5: missing]
(5) 5 7 3 11 1 13 7 5 11 3 13 1 permutation 2:
(4) 7 13 5 1 11 11 3 3 13 7 1 5 1 1 1 I I I
(3) 3 11 3 11 13 1 7 5 1 13 5 7 6 7 5 8 4 9 310 2 11 1 12
(2) 7 1 5 13 1 11 13 3 5 11 7 3
(1) 13 11 3 1 5 13 11 5 7 1 3 7
NB In rows 16-18, which are listed in the same way as the others here for the sake of consistency, the
order in which the notes appear is definitive; therefore the content of these three rows is correct when
read from left to right, though such a reading of the rest of the table represents the movement in
retrograde.
movements and six representing change - ppp pp p mp mf f ppp<f pp<mf

p<mp mf>p mf>pp f>ppp. There is some reason to suppose that the order
of the generating series is ppp ppp<f pp pp<mf p p<mp mp mp>p mf mf>pp f
f>ppp, but, as in the case of the durations rows in this movement, this
series is not used directly, and those that are used do not make their origin
clear. The series that actually occurs is shown in Fig. 31. I cannot explain
(as I could in the case of the durations series) the way in which this
ordering has been derived from what I believe to be the generating series;
nor can I explain the table of durations, which appears in Fig. 32. I can
only make some observations.

It is obvious that this is not a permutation table and that the dynamics
in this movement do not represent the regular permutation or rotation of a
series of twelve elements: there is a great deal of internal repetition. In fact,
no row and only one column - column 8, representing Ci - contains twelve
discrete elements (compare this with the situation in Figs 5 and 19). What
KATHYRN BAILEY
Fig. 31 No. 9: Dynamics series

mp mf>pp pp<mf mf pp ppp<f f ppp mp>p p<mp f>ppp p
1
10
11
12
does seem fairly clear is that the dynamics are arranged in such a way as to
impose the proportions of No. 2 on the present movement in addition to the
proportions natural to this movement - those resulting from the
permutations applied to the durations series. (Since it contains twelve
elements, the one whole series to appear as a column in this movement
begins in the 'first' row, row 15, and continues to row 4, which is two rows
into the 'next' section on the table in Fig. 30, but is parallel to the last row
of the first section in No. 2 (see Figs 3 and 5).) We will see presently that
this division of rows 15-1 into two sections of twelve and three to replicate
the structure of No. 2 is reinforced in other ways.
Only column 8 contains twelve discrete elements in succession; thus
this table, while hauntingly like those of Nos 2 and 6b, is essentially
different. In it are also resonances of various other constructions we have
observed while working with the a-is permutations. Columns 5 and 10 of

the section comprising rows 15-6 and columns 3 and 8 of the section
consisting of rows 5-1 represent the continual alternation of two elements;
such columns are a feature of the permutations used in this work, and even
occur in these positions (see, for example, columns 3 and 8 in table (a) or
5 and 10 in table (b) of Fig. 1), although in a regular permutation the same
two numbers alternate in both columns. This can be no accident; at the
same time, it is difficult to see how all four of these alternating columns
could have been produced by the same operation. What is very apparent,
and more to the point, is the structural resemblance of the table in Fig. 32
(the dynamics table) to that in Fig. 33 (the durations table, expressed in
order numbers rather than actual durations as in Fig. 30).
Two of the rows on the table in Fig. 32 are strikingly different from the
rest, and closely related (they are constructed in the same way, but in
format they are the retrograde of each other): these are rows 15 and 5,
which are parallel to the structurally important 'first' rows of the two
sections on the table in Fig. 33 - those rows from which the regular
permutations proceed, the second of which is the retrograde of the first.
The alternating columns are 3 and 8 in the larger section, 5 and 10 in the
smaller on the table in Fig. 33 - a mirror image of their positions in the
Fig. 32 table. In addition row 14 and most likely row 4 of the dynamics
table - the 'second' row in each section - is a palindrome. Thus the table
representing the organisation of dynamics in this movement seems to be a
cosmetic one, drawing its features from various permutational tables and
combining them in a somewhat eclectic fashion so that the result resembles
a permutational table without being one.
An interesting and characteristic feature of this movement is the
A B6 Ab B G C F# C# F D E EF
(18) 3 5 1 8 11 10 10 8 1 10 5 8
(17) 8 5 1 1 3 8 9 7 7 1 1 1 CODA
(16) 7 6 4 6 12 11 8 11 9 11 2 11
(15) 8 11 8 11 8 11 1 1 1 1 1 1
(14) 3 5 6 8 2 4 4 2 8 6 5 3
(13) 1 9 7 12 8 10 4 3 8 1 12 6
(12) 8 4 11 10 2 3 5 4 5 6 8 12
(11) 2 12 12 7 8 9 12 5 - 1 10 8
(10) 11 3 1 1 2 12 4 6 7 6 8 5
(9) 6 6 4 9 8 6 11 7 10 1 7 7
(8) 1 2 1 2 - 2 8 8 3 6 4 9
(7) 4 - 2 3 8 1 1 9 4 1 9 10
(6) 8 7 12 5 2 8 7 10 12 6 4 8
(5) 3 3 3 3 (3) 3 8 11 8 11 8 (11)
(4) 1 6 5 5 12 9 9 12 5 5 (6) 1
(3) 5 2 3 11 1 12 7 11 5 10 4 3
(2) 2 8 5 6 12 6 1 12 7 8 9 6
(1) 11 7 3 12 6 2 4 11 10 4 (2) 12
NB In rows 5, 10 and 4 I have made alterations in order to preserve a pattern that seemed
to me very likely correct. The 3 and 11 in row 5 should read 4 and 8, following the score;
the 6 (f) in row 4 is really 7 (ppp>f), and the 2 (mf>pp) in row 10 is 4 (mf), both easy
misprints. There is a mistake in row 1: the dynamic indication on E is mf>ppp, an
indication that does not exist in this movement.
frequent occurrence of the same duration in two parts simultaneously.
This happens somewhat less often than it would seem at first glance, if we
discount those instances where the same pitch occurs on two staves for
technical reasons as the result of forced divisi notation before or following a
real division of function (the basses and tenors in bs 553-60 and the altos
in 577-81). This is not the case, however, with the middle C sung by
sopranos and tenors together in bs 552-3. This C is interesting because it is

the only example of the doubling of the same pitch by two parts, but especially - because it corresponds to the only example of doubling of any
sort that occurs in No. 2: in both movements the pitch concerned is a
middle C in row 2 (see soprano 2 and alto 1 in bs 110-12). Different
pitches occur in rhythmic unison in the following rows: 4 (F# and C# in
KATHYRN BAILEY
Fig. 33 No. 9: Permutation of durations series
A B% Ab B G C F# C# F D E Eb
(18)
(17)
(16)
(15)
(14)
10
(13)
(10) 2 12
(9) 6 7 5
12
10
10
10
(6)
(5)
12
(4)
(3)
10
(2)
10
(1)
1
8
11
12
10
12
7
10
10
6
10
1
2
11
11
12
9 4 10 7 6 5 8 11 3
10 8 4 11 9 3 12 1
12
7
4
(8)
(7)
12
12
12
11
11
11
(12)
10
11
(11)
11
11
12
11
12
11
11
12
10
12
11
10
8
bs 560-1), 7 (E and F in 567-8, A? and G in 568), 8 (E and D in 569-70,
C and B in 580-2), 9 (E and D?, F? and G, both in 573), 10 (F and D in

574-5, C and F?, G and B, both in 576), 12 (C? and E in 581), 14 (C? and
F? in 587-8, C and G in 588-9), 17(A? and B in 599) and 18 (B6 and G? in
601).
More striking than these, because of the repetition involved, are
rhythmic unisons involving the A at the end of one row and the E6 at the
beginning of the next. Nearly all the rows in the main body of the piece (615) are elided in this way (see bs 569-70, 573-5 and - following
immediately without a break - 575-7, 579-80, 586-7, 589-90). Only
between rows 12 and 13 is there no hint of elision. (Could this be in
reference to the structural break that occurs between rows 12 and 13 of
No. 2?) Such a connection is closely approximated between the first two
rows of the section, rows 6 and 7: in bs 85-6 the second sopranos' D? and
the second basses' A both have a value of 13; their sounding and release
are as nearly simultaneous as possible, given that they are not in the same
tempo. This also is an important juncture in No. 2: there row 6 represents
the completion of the first half of the rotational cycle; 7 is the exact
inversion of the generating row. The next join that is not rhythmically
elided is that between rows 8 and 9; here such an elision would occur if the
than
the beginning
of 573, and,
in the light
of at
thethe
other
timpani's
D. were written
one quaver
earlier,
endcarefully
of b.572 planned
rather
connections, one is tempted to see this as an error. The rows of the first
section of the movement (1-5) and the last (16-18) are not so joined,
except for the final two: the E6 at the end of row 17 and the initial A of 18
(both these rows are forward-going) are sounded exactly together.
And so in some way the circle is closed: Nos 8 and 9 represent the
return of many aspects of Nos 1 and 2, although the arch form that this
suggests is not realized. The design of the whole is asymmetrical, with little
in the way of reference beyond the similarities just noted: Nos 1 and 8, and
Nos 2, 6b and 9. The disposition of several features is shown in Fig. 34 for
reference; those movements exhibiting a structural similarity are bracketed
together.
Fig. 34 Il canto sospeso: overall design

1 ORCHESTRA alone, winds vs strings
4/8, 3/4: .=56, 60, 72, 92; .=60, 72, 82

2 unaccompanied eight-part CHORUS; fragmented text
2/4: =60-66
3 soprano, alto and tenor SOLOISTS; ORCHESTRA

4/8/3/8: = 152
4 sixteen solo strings and ORCHESTRA

4/4, 3/4: .=92
5 tenor SOLOIST, solo strings and ORCHESTRA

3/4: ==46, 56, 66
6a eight-part CHORUS and low INSTRUMENTS

2/4: =66
6b four-part CHORUS and strings; vowel sounds extracted

2/4: .=66
7 soprano SOLOIST, women's CHORUS and high INSTRUMENTS;

singing with closed mouth
5/8: .=80, 96, 120, 132, 152
8 wind INSTRUMENTS and timpani

4/8: ,>=46, 56, 60, 66, 72, 92, 112
- 9 eight-part CHORUS and timpani; fragmented text
2/4: i=54, 44
KATHYRN BAILEY
CODA
At this point I should like to discuss musical aspects of Il canto sospeso; but
the truth is that I find it very difficult to write about the work as music,
because it doesn't seem to me either to consist of musical fabric or to
behave in a musical way. I return to my original quotation from Alexander
Goehr: 'There is no material in the traditional sense, but only a pre-
compositional abstraction of the intended course of events.' Material

sounds a fairly unspecific term, but when a musician speaks of material, he
is in fact speaking of the very essentials of music-building. He means
melody as distinct from the arbitrary succession of pitches, rhythms that are
contrived either to reinforce or to suspend a metrical orientation,

simultaneities brought together to some purpose, motives that recur or
develop and in some way provide unity. These things are the essence of
music, and they have no place in the music of integral serialism, in which
decisions concerning the combination of ingredients are made on nonmusical grounds before composition begins.
Perhaps ironically, the downfall of this music finally is its sameness
throughout. Carefully contrived contrasts in instrumentation notwithstanding, it sounds monotonous. Aurally, four movements of Il canto
sospeso provide some degree of contrast to the rest: No. 5, because of its
slow tempos; No. 6a, because of its restricted, low range and the large
number of voices playing together (the effect is of an extended tone
cluster); No. 7, because of its high range and thinness of texture (never
more than three voices present, and often only one); and No. 9, again
primarily because of its slow tempos. The overall effect, however, is of
continual strident dissonance in, generally speaking, a very high range, in
which shrillness is exaggerated. Though consonant intervals and octaves
frequently do occur, they do nothing to allay the overriding harshness,
since their presence is the accidental result of precompositional decisions
of a numerical rather than a musical nature and their placement therefore
without (musical) purpose. Again I quote Alexander Goehr:
... this music is dull in its lack of formal complexity, of dramatic
gesture. No amount of technical ingenuity can break the monotony of
regularity - even though this regularity can be shown to contain,
within itself, the widest possible degree of isolation and differentiation.
The dull impression is simply due to the fact that all serial possibilities
are continually present in the work. Musical interest is always
produced by the restriction of possibilities - a restriction which is
determined, in its turn, by the character of the material, by individual
invention.28
Goehr of course was writing about integral serial music in general, not
about Nono's piece. Yet perhaps it is time to submit some of the 'classics'
of that tradition to the detailed analysis they deserve (and which we

routinely accord the classics of the past). If they really are classics, they will
repay such attention richly. If they are not, then it is time they lost some of
their mystique. I am well aware of the ambivalence of my own feelings in
this regard. While fascinated by the intellectual operations that lay behind
the composition of Nono's work, I am much less attracted to it as 'music'.
This leads me to ask - at the risk of sounding philistine - whether 'music'
is in fact an accurate description of the piece at all. When sense and sound
- constructive achievement and aesthetic effect - are out of kilter to the
extent that they appear to be here, something must be wrong. This is not
to denigrate the intellectual accomplishment of the work: many of its
problems are challenging, many of its solutions elegant. Perhaps, as writers

like Nicolaus Huber and Ivanka Stoianova argue, it is expressive of its text
(though it is surely not trivial that these authors have turned to theories of
film making and of language to make their point). But if the problem I
have highlighted is a real one, it is one that affects many modem 'classics'
apart from II canto sospeso.
APPENDIX
One of the inevitable results of an analysis of this sort is a list of probable misprints
and omissions. I offer the following.
No. 2: b.110 figure accompanying bracket in bass 2 (over D) should be 5,

not 3, and note should be a semiquaver
b. 135 durations of 2 notes in altol&2 should be reversed: 1st, not
2nd, should have a dot
No. 3: bs 169-70 2nd tied note in bcl should be a quaver

bs 169-70 2nd tied note in trp4 should be a semiquaver
No. 4: b.249 xyl should have B, not G
b.278 fl should have B, not D
No. 5: b.290 vibr should have C instead of B
b.292 [omission: trp3 should be mp]

No. 6a: b.319 2nd and 3rd vlc notes should be tied
b.332 decrescendo missing from bssnl

b.334 sop l's 1st and 2nd notes should be tied
b.339 there should be no tie in the cb
b.340 mp<f in cb should be mp<mf

b.359 [omission: p missing from vlc]
b.360 [omission: mp missing from cb]
KATHYRN BAILEY
No. 6b: b.387 vnl should have Fi , not Ag

b.392
should
be17,
E,, not
not 26
E
b.394 vlc
notenote
should
equal
b.410 vnl's D? should be worth 8

No. 7: b.428 A should be a semiquaver
b.432 alto's E should end with dotted quaver
b.445 soprano's F should be G
b.459 E in marimba should be G
b.464 [omission: there should be an A on last half of 4th beat]

No. 8: b.497 2nd tied note in bcl should be a semiquaver
b.498 dot belongs with 1st tied note in trbn rather than with
preceding rest
b.504 1st note in trp2 should be a quaver
b.510 G in trbn2 should be 7 units long
b.516 last 2 notes in hn2 should be 4+3 units in length
b.517 a treble clef is needed in hn5 part
b.520 F in hnl should be A (bass clef?)
b.522 trp4 note should be a semiquaver
b.524 B in trbnl should be Bb
b.533 1st note in trpl should be a quaver
b.535 tied note in fll should be a quaver

b.535 hns5&6 should have F? , not F
b.542 1st note in trp3 should be a quaver

b.542 trbn4 should have dotted rest, demisemiquaver
No. 9: bs 545-6 mf>ppp should be mf>pp
bs 562-4 C? should be C?
b.564 A should be 5 units long, not 4
bs 566-8 C? should be 13 units long, not 12

b.584 D should be 3 units long, not 4
b.592 C should be 7 units long, not 4
b.568(9?) [omission: Bb (=1 unit) missing from b.568 or 569]

bs 571-2 [omission: G (-3 units) missing from bs 571-2]
b.577 [omission: F (=1 unit) missing from b.577]
NOTES
1. 'Is There Only One Way?', The Score (January 1960), p.64.
2. 'Pierre Boulez', Die Reihe, Vol. 4 (1960), p.36.
3. In English there is, to my knowledge, only Udo Unger, 'Luigi Nono:
"Polifonica-Monodia-Ritmica - Il canto sospeso"', Die Reihe, Vol. 4, pp.5-13
(II canto sospeso is mentioned only on pp.10-13); Karlheinz Stockhausen,
'Texte und Sprache', Die Reihe, Vol. 6, pp.40-67 (Il canto sospeso discussed on
pp.47-57); and Reginald Smith Brindle, 'Current Chronicle: Italy', Musical
Quarterly, Vol. 47 (1961), pp.247-55. Both Stockhausen's and Smith
Brindle's discussions are limited to certain aspects of No. 2 (although this is
not clear in the case of Smith Brindle, whose pretence of an intimate
knowledge with the remainder of the work leads him to make egregious
blunders), and both fail to notice essential features of the organisation of that
movement. There is, in addition to these: Massimo Mila, 'La Linea Nono.
Aproposito de "Il canto sospeso"', Rassegna Musicale Italiana, Vol. 30 (1960),
pp.297-31 1; T. Ekbom, 'Der om6iliga uppdraget. Om Luigi Nonos k6rverk Il
canto sospeso', in Nutida Musik, Vol. 6, No. 1 (1962/3); Wolfgang Motz,
Analytische Betrachtungen iiber 'Il canto sospeso' von Luigi Nono (Freiburg i.Br.,
1978-80); Nicolaus A. Huber, 'Luigi Nono: Il canto sospeso VIa, b', MusikKonzepte 20: Luigi Nono, ed. Heinz-Klaus Metzger and Rainer Riehn (1981),
pp.58-79; and Istvin Balizs, "'Il giovane Prometeo". I "peccati" di Nono
contro il serialismo ortodosso nel periodo darmstadtiano', and Ivanka
Stoianova, 'Testo-musico-senso. "Il canto sospeso"', both in Nono, ed. Enzo
Restagno (Torino, 1987), pp.102-15 (Il canto sospeso is discussed on pp.10815) and pp.126-42. I have not seen the first two of these; of the remaining
three, only Nicolaus Huber discusses the serial and permutational basis of the
music, and, even so, this is not his primary concern. And again, of course, his
article deals with only one (two?) movement(s). I feel a kinship with him
when he writes that his analysis is bound to be 'eine subjektive Interpretation
..., die durch Daten abgesichert, somit einsehbar ist und Anregung zu
4.
5.
6.
7.
sch6pferischer Weiterarbeit sein kann' (p.59).

'Current Chronicle: Italy', Musical Quarterly, Vol. 47 (1961), pp.247-55.
Ibid., pp.248-9.
Ibid., p.249 (my emphasis).
Ibid., p.254 (my emphasis).
8. Milton Babbitt, 'Some Aspects of Twelve-Tone Composition', The Score
(June 1955), pp.57ff.

9. Translations of the texts are my own.
10. I use this term for serial music in which the series is expressed through the
succession of attacks throughout the entire range of a texture which may be
linear in nature but most often is not. This is the opposite of linear topography, in which several rows unfold simultaneously in separate voices that
sound together. See Bailey, The Twelve-Note Music of Anton Webern

(Cambridge: CUP, 1991), pp.31ff.
11. Stockhausen identifies 'a series of six factors which can be multiplied with the
four unit values: 1 2 3 5 8 13' in this movement, but curiously fails to see that
this series was doubled and the result submitted to a regular pattern of
rotation. Thus his rather complicated explanation of the organisation of
durations misses the point entirely (Stockhausen, p.52). Smith Brindle, on the
other hand, recognises the twelve-element series 1 2 3 5 8 13 13 8 5 3 2 1 and
KATHYRN BAILEY
its application to the repeating pc rows, but (even more curiously) seems not
to have noticed that the series underwent rotation (Smith Brindle, p.250).
12. Smith Brindle, p.250.
13. Concerning dynamic markings, Stockhausen writes: 'The structure of the
dynamic values is ... linked with the pitches, each of the 12 chromatic pitch
values receiving a different intensity whenever it recurs' (p.52). While this is
nearly true (it holds only within the two main sections - rows 1-12 and 13-15
- and not in the coda, rows 16-19, and does not take into account the
appearance of the ppp indication twice (as elements one and six) in each of the
columns representing rows 1-12), Stockhausen has here again failed to notice
the essential fact of rotation.
14. The figure does not show all the possibilities (the arrangement of the first
three rows in order of releases and of the last three in order of attack have
been omitted, for example), because the symmetrical organisation of other

aspects of the movement make it fairly clear that the series shown must be the
significant ones.
15. Although I cannot find any serial manipulation of the dynamic indications in
this movement, the order in which I have placed them here is not arbitrary.
Besides representing a logical progression as in other movements where
dynamics are serialised, this order is clearly Nono's: the transpositional
relationship between the dynamics of the chorus and those of the orchestra is
apparent only when the dynamics are ordered in this way. This relationship is
missed by Huber in his otherwise thorough analysis of 6a. He says in this
connection that 'die Lautstirken im Orchester sind selbstandig, gehorchen

nicht dem Rticklauf, sind normaler Musikh6r-Zeit unterworfen ...' (p.66).
16. Huber agrees with this interpretation, but for different reasons, which are
difficult to discern. He discusses the matter on pp.66-7: 'Das Orchester ist die
Stimme der Mbrder. Sie wiederholen den Text auf der Bedeutungsebene des
Berichts, anders als der Chor jedoch von der Basis der Handelnden,
Eingreifenden, Schicksalbringenden. Die M6rder sind gleichsam die direkte
Realitdt, von der der Chor berichtet, auf die der Chor der Opfer reagiert. Die
Orchester-Anschlige schlagen gleichzeitig um in neue Ausdrucksqualitiit. Sie
sind die Schlige, die fremde, feindliche Emotionalitit besitzen (vgl. die eigene
neue Lautstdrkekomposition [my emphasis]' , etc....
17. I must thank Anthony Gritten for suggesting this as a possibility after studying
the score.
18. Huber, who analyses 6b, does not notice this (but then, it is unlikely that it
would come to mind unless the scheme used in No. 2 had been found first,
and to my knowledge no one has seen this until now). On pp.77-8 he
suggests an explanation rather like Stockhausen's explanation of the dynamics
in No. 2 (Stockhausen, pp.52-4).
19. In many cases it is easy to see the reason for the alteration. I give the following
as examples:
b.368, alto has p instead of f on the third note of a melodic group in
which first and second notes are sung p and mp;
b.368, cello has mp instead of p on note five of a six-note crescendo:

ppp-[f, pizz]-pp-p-mp-mf;
b.372, soprano has mp instead of f on second note of 'Com' &', after
first note was p;
b.376, soprano has f instead of mf on second of three notes, the other

two of which are f;
bs 378-9, soprano and bass have ppp-p for 'addio' instead of the mf-f
(soprano) and f-f (bass) prescribed;
b.381, soprano has f instead of p on third of three notes, the first and
second of which are f;
b.407, cello has ppp instead of p to match preceding two notes of a

three-note group.
At other times the reason is not immediately apparent: for example, why the
second violin has been written ppp rather than pp in bs 370-2, why the viola
is marked p in bs 385-6 and then pp-pp in 387-9 rather than mp then f-f, or
why the bass has p and f instead of mp and p in bs 395-6.
20. In the following paragraphs and, particularly, in the tables I will make a
distinction whenever it seems necessary between the actual order of the rows'
appearance and the order in which they were derived by putting the numbers
identifying their real position in the score in bold face and placing the
numbers representing the order of derivation in brackets.
21. These numbers and those later in this paragraph are derivation numbers;
brackets have been left out in the interests of simplicity.
22. The rows are listed and numbered here in the order of derivation. The order
in which they actually appear in the score is that described earlier:
row 1 bs 414-20 (1)

rows 2, 3 and 4 bs 420-26 (6), (2) and (5)
rows 5 and 6 bs 426-32 (3) and (4)
rows 7 and 8 bs 432-8 (9) and (10)

rows 9, 10 and 11 bs 438-44 (8), (11) and (7)
row 12 bs 445-51 (12)
23. To continue the table begun in the previous note:
rows 13
rows 15,
rows 18,
rows 21
and 14 bs 451-7 (12)R and (7)R

16 and 17 bs 457-62 (11)R, (8)R and (10)R
19 and 20 bs 463-9 (9)R, (3)R and (2)R
and 22 bs 469-75 (1)R and (4)R
row 23 bs 476-82 (5)R

row 24 bs 482-8 (6)R
24. It should be mentioned that although there are twelve discrete rows, two of
them, (1) and (7) (or 1 and 11, as they actually occur), are the same with the
KATHYRN BAILEY
hexachords exchanged: (1) = ppp mf mfp p ppp p ppp p mf mf p, (7) = p ppp p

mf mf p ppp mf mf p p ppp. The other pairs of rows that share the same
contents are not related in this way.
25. Exceptions are trumpet 5 in section 1; trumpet 4 and trombones 1, 3 and 4 in

section 5; trumpets 3, 4 and 5 and trombone 3 in section 7; and horns 1 and
2, trumpets 2, 3 and 5, trombone 1 and both timpani in the final section, by
far the longest and rhythmically most diverse of all.
26. Whereas No. 2 divided the four sections of the chorus regularly into eight, the
division here grows from only four parts at the outset to ten by b.567. This is
accomplished through a constant fluctuation: the sopranos are divided into
three groups in bs 564, 567-8 and 586-7; the altos are similarly divided in
b.568; the tenors in bs 564-5, and the basses in 570-7 and 588-90. The
number of parts diminishes towards the end: all the altos join together in
bs 577-98 and again in 602-4, as do the sopranos in 595-8, the tenors in
598-9 and the basses in 603-5.
27. It will be seen presently that this difference is the direct result of there being
one less tempo in No. 9 (see the discussion of the coda below).
28. 'Is There Only One Way?', p.64.

'Work in Progress' - Analysing Nono's 'Il Canto Sospeso'

Uploaded by

Copyright:

Available Formats

'Work in Progress' - Analysing Nono's 'Il Canto Sospeso'

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

'Work in Progress' - Analysing Nono's 'Il Canto Sospeso'

Uploaded by

Copyright:

Available Formats

'Work in Progress': Analysing Nono's 'Il Canto Sospeso'

Author(s): Kathryn Bailey

ANALYSING NONO'S IL CANTO SOSPESO

In 1960, Alexander Goehr wrote that integral serialism

intended course of events.'

music of Schoenberg and Webern, and while I found Schoenberg's

might be closely organised and no liberties taken, any hope of my

discovering the particulars of this organisation through listening to the

Luigi Nono's II canto sospeso (1955-6) is a work that stands directly in

Munich Oktoberfest in 1980, killing thirteen people and injuring more

of II canto sospeso by the British composer and critic Reginald Smith

Brindle is a case in point.4 Smith Brindle, whose specific knowledge of the

assertion of the work's simplicity.

In examining Nono's composition principles in the Canto sospeso ... it

280 MUSIC ANALYSIS 11:2-3, 1992

NONO'S IL CANTO SOSPESO

In conjunction with the lack of knowledge of the work betrayed in these

the second movement'6 and 'Each of these eleven dynamic values is

II canto sospeso is built on an expanding all-interval series (referred to

remains stationary throughout all the permutations.) The pitch-class rows

6 4 2 1 3 5 7 9 11 and their retrogrades. Nos 3, 7, 8 and 9 of Il canto

anthology of the possibilities suggested by the ai-s permutations. The

opening movement, in which I can find no strict serial organisation, and

the other movements.

The opening piece, without voices, is divided by double barlines and

Fig. 1 Permutation tables relating to the expanding all-interval series

(5) 4 11 8 1 7 12 5 3 10 9 2 6 (six perfect 5ths/4ths)

(8) 11 7 3 2 6 10 12 8 4 1 5 9 (four augmented triads)

(5) 7 11 4 3 10 8 1 6 12 5 2 9 (six perfect 5ths/4ths)

(8) 4 8 12 9 5 1 3 7 11 10 6 2 (four augmented triads)

second and eleventh) are so brief that they cannot be considered as

282 MUSIC ANALYSIS 11:2-3, 1992

NONO'S IL CANTO SOSPESO

Fig. 2 No. 1: Formal structure

I (ii) III iv V VI vii VIII IX X (xi) XII XIII

One of the pervasive characteristics of II canto is the use simultaneously

quaver, as in the movements just mentioned, occur in succession. Only

tempos proceeding at once - an idea reminiscent of composers as diverse

if the basic unit were twice as fast as it is (written demisemiquaver

understood as divisions of the semiquaver into five and seven).

consistency with which like durations are clustered throughout this

timpani in bs 3 and 4 should be one semiquaver longer.) The clusters

presented in four tempos (see Ex. 1).

- 9 ral. 6(d) accel. 4(d) rail.

____ --- -____ ___________- - rz I' zzz- z:

___;t __- --- __ __ - X4X?-_

4r _ --92 ralL. _ accelcc) :z z ______ -- -~ ------

Copyright Ars Viva Verlag GmbH, Mainz.

284 MUSIC ANALYSIS 11:2-3, 1992

NONO'S IL CANTO SOSPESO

MUSIC ANALYSIS 11:2-3, 1992 285

reduced number of instruments representing both winds and strings

presents a series of durations, each occurring in only two parts/tempos, and

286 MUSIC ANALYSIS 11:2-3, 1992

NONO'S IL CANTO SOSPESO

---?? --I- -' - T 7 , <-:i-:.- --?........

,.___ ____ r a ' "-Vlot

44 9(d) 4(d) f 801)

4 9_-_ 3:- - - .k "n ,12) ._, -d

--- - _______- - rz I' zzz- z:

_;t - --- - X4X?-_

,._ __ r a ' "-Vlot