2 Tempo 61 (239) 2–17 © 2007 Cambridge University Press

DOI: 10.1017/Sxxxxxxxxxxxxxxxx Printed in the United Kingdom

  ’ ‘ ’

Claude Vivier (photo: Jean Antonin Billard, courtesy

Bob Gilmore
of Boosey & Hawkes Music Publishers Ltd)

1. Introduction
There is an uncanny symmetry in the life of the French-Canadian com-
poser Claude Vivier: we do not know the exact time or circumstances
of his birth, and we do not know the exact time or circumstances of
his death. The first of these two facts haunted Vivier all his life. Born
to unknown parents in Montreal in April 1948 and placed in an orphan-
age, he became obsessed with the identity of his birth mother, whom
he never knew. Several of his compositions can be heard as a poignant
attempt to communicate with her. The second fact – his murder in
March 1983 by a young Parisian criminal in circumstances that remain
not fully investigated – has, you might say, haunted the posthumous
reputation of his music. It seems impossible to discuss Vivier’s work
without mentioning the cruel and sordid circumstances of his death.
For some, his murder is the key to an understanding of his life and – even
more controversially – of his work.
Vivier’s music inhabits a shadowy realm between reality and the
imagination. His is a world where human beings express themselves
in invented languages more often than ‘real’ ones; they are seduced
by the allure of distant cities and embark on journeys, often symbolic
ones, in search of love or companionship; and they are haunted by the
omnipresence of death, which in Vivier’s output holds dominion over
everything else. All these themes characterize his largest works, the
operas Kopernikus (1978–79) and the unfinished Rêves d’un Marco Polo
(1977–83). Yet in the deepest sense the most frequently recurring subject
in Vivier’s compositions is himself: almost all his works are essentially
autobiographical. The expressive intensity of his music, together with
its compositional skill and innovation, makes his oeuvre among the
most compelling of the late 20th century.
Nowhere do the various facets of Vivier’s output come together
more powerfully than in Lonely Child for soprano and orchestra. It was
written on commission from the Chamber Orchestra of Radio-Canada
Vancouver, to a text by Vivier himself, and completed in Montreal on
March 5 1980.1 Lonely Child has always held a special place in Vivier’s out-
put: partly because it is the first work in which, by widespread consensus,
he found the musical language unique to him; and partly because of all
his works it is the one that seems most fully to encapsulate the essence
of his artistic personality. It is also the work in which Vivier first intro-
duced the main technical innovation that underlies the composition of
most of the large-scale works in the remainder of his short life. This
technique, derived from the spectral techniques of the Paris-based new
music of the 1970s, involves the use of ‘combination tones’ to produce
a distinctive kind of spectral harmony (which Vivier called, more poeti-
cally, ‘les couleurs’). The technique and its application in Lonely Child is
the main focus of this article, which is the first extended analysis of the
1
Lonely Child is published by Les Éditions Dobermann-Yppan in Quebec. However, in April
2005 Boosey & Hawkes in New York announced the signing of an agreement to publish all
of Vivier’s works, so the score is now available through them.
   ’ ‘ ’ 3

work in English. Drawing on the sketches housed in the Vivier Archives

at the University of Montreal, I attempt to illuminate aspects of the fin-
ished composition and of the compositional techniques that underlie it.

2. Vivier and spectral music

While it would be misleading to categorize Vivier’s music (even his
later works) as ‘spectral’ in the fullest sense of the term, his music was
nonetheless transformed by the use of spectral techniques from 1980
onwards; and he himself was a friend of those composers in whose
work the spectral aesthetic first came to prominence. The term ‘spectral
music’, misleadingly reductionist though it is, has remained in use since
it was coined by the French composer Hugues Dufourt in an article
published in 1979.2 It has come to be associated primarily with the work
of two of Dufourt’s colleagues, Gérard Grisey (1946–98) and Tristan
Murail (b.1947). Although the music that characterizes this approach
first made itself heard in the mid-1970s, it is really only in the past dec-
ade that it has caught the imagination of the listening public, as well as
that of younger composers, performers and musicologists. There seems
now to be an increasing and widespread recognition that spectral music
is one of the defining areas of late 20th-century musical thought. The
pre-eminence of Grisey and Murail provokes an interest in the work of
other composers who established their spectral credentials respectably
early. If Dufourt, Michaël Lévinas and Roger Tessier (co-founders with
Grisey and Murail of the ensemble L’Itinéraire in Paris in 1973) are not
considered spectralists in the true sense, we can however make legiti-
mate claims for the Romanian-French composer Horatiu Radulescu
(b.1942), whose earliest spectral composition dates from 1969, and for
Vivier, whose later works form a distinctive and highly personal contri-
bution to the repertory of spectral music.3
To some extent, spectral music was a reaction against the prevailing
norms of the day: against the post-serial aesthetic that was rampant in
Paris in the 1970s, and against aleatoric music, improvisation and collage
techniques. Murail has often cited the music of figures like Ligeti and
Xenakis as more important in the development of his own thought than
the music of Boulez or Barraqué, Cage or Stockhausen (even though the
latter’s Stimmung (1968) may be considered essentially a spectral work).4
The essence of ‘la musique spectrale’ is its search for compositional
models based on sound, rather than on mathematics, or language, or
visual imagery. In 1982 Grisey famously declared: ‘We are musicians and
our model is sound, not literature; sound, not mathematics; sound, not

2
Dufourt, ‘Musique spectrale’, Paris: Société Nationale de Radiodiffusion, Radio-France,
March 1979. Reprinted in Conséquences nos.7–8 (1986), 111–115. The best history of spec-
tral music to have appeared so far in English is Julian Anderson, ‘A Provisional History of
Spectral Music’, in Contemporary Music Review 19 no.2 (2000), 7–22.
3
There are interesting cases of composers close to, but not quite at the heart of, spectral
music as we have come to understand it – Kaija Saariaho, for example, on whom Grisey’s
music made a considerable impression, or Jonathan Harvey; and, back-projecting, some crit-
ics even speak of Scelsi, Lucier, Tenney and a few others as ‘proto-spectralists’.
4
Vivier had heard Stimmung at the latest by September 1971, as he compares Stockhausen’s
Sternklang unfavourably to Stimmung in an undated letter to his friend Pierre Rochon writ-
ten from Paris in the early autumn of that year (Archives Claude Vivier, Special Collections
Department, University of Montreal). He encountered the work again at its performance in
Darmstadt on August 3 1972. The Canadian composer and conductor Walter Boudreau, who
also attended the Darmstadt Ferienkursen that summer, recalls himself, Vivier and Grisey
incessantly imitating the vocal sonorities of Stimmung as they walked around the town or on
trams, much to the annoyance of one particular driver who threatened to throw them off if
they didn’t stop. Walter Boudreau interviewed by the author, Montreal, November 14 2002.
4 

theatre, or the plastic arts, or quantum physics, or geology, astrology

or acupuncture!’5 The permutation of notes, whether strictly serial or
not, was replaced by a scrutiny of the inner life of sounds, which then
became models for compositional forms. As early as the mid-70s Grisey
and Murail undertook computer-aided analyses of the waveforms of
complex sounds – studies that were sophisticated given the technology
of the time but rather primitive in view of the means available today.
The natural harmonic series was one such object of study, and an
important one; it is the basis of the opening of Grisey’s Périodes and his
Partiels, in which the low E of a trombone creates a spectrum that is
simulated by the instrumental ensemble.6 But the harmonic series was
not the only model explored. In spectral music, harmonic spectra have
no special priority as musical material over inharmonic spectra, which
have just as much musical potential. Early examples of such ‘distorted’
spectra used as the basis of whole pieces are Grisey’s Jour, contre-jour
(1979) and Murail’s Gondwana (1980). Models based on formant regions
of the spectrum of a sound, rather than the whole spectrum, have also
been used as compositional starting-points (for example in Murail’s
Désintégrations (1982)). Even smaller units, such as a single interval with
its sum and difference tones, can yield material of great richness and
complexity, as is the case in the music of Vivier.
We can trace Vivier’s adoption of spectral techniques to a particu-
lar time: his month-long visit to Europe in November–December 1979.
But the ground had already been prepared in his studies with his two
main composition teachers, Gilles Tremblay (with whom he studied at
the Conservatoire de Musique in Montreal in 1967–70) and Karlheinz
Stockhausen (at the Hochschule in Köln in 1972–74). Gilles Tremblay,
still little known in Europe, was the most distinguished of Messiaen’s
Canadian pupils; thanks to him Vivier would have absorbed many of
Messiaen’s ideas, including his ‘chord of resonance’, an eight-pitch
aggregate spanning two octaves and comprising harmonics 4–15 of the
harmonic spectrum (albeit distorted considerably by the limitations of
12-note equal temperament). In his treatise Technique de mon Langage
Musical (1942), Messiaen also describes his use of ‘added resonance’
– dissonant upper pitches added to a triad in the middle register – and
‘inferior resonance’, dissonant clusters of notes in a low register with
triadic configurations above them, citing the complex sounds of percus-
sion instruments such as bells, gongs and tam-tams as an analogy for
this type of harmony. This way of thinking about the nature of sound is
highly prophetic of the spectral aesthetic.7
Studying with Stockhausen, Vivier became familiar with an even
closer precedent to the type of approach he would adopt in Lonely Child:
the use of ring modulation. More so than Stimmung, it is Stockhausen’s
Mantra (1970) for two pianists and live electronics that relates most
interestingly to Vivier’s later work. Two aspects of Mantra are notable
in this regard: its use of a single melodic ‘formula’ to determine both
large-scale form and small-scale detail; and its use of ring modulation to

5
Grisey, ‘La musique: le devenir des sons’, originally written in 1982 and reprinted in Danielle
Cohen-Levinas, ed., Vingt-cinq ans de création musicale contemporaine: L’Itinéraire en temps réel
(Paris: L’Itinéraire/L’Harmattan, 1998), 298.
6
Here we are talking not of actual resonance (of the type Grisey would explore a little later in
his Prologue (1976) for solo viola and optional resonators) but of simulated resonance; a com-
plex sound – in this case, a low trombone note – is treated as though broken down into its
component partials and reconstituted by the ensemble. This technique has become known
as ‘instrumental synthesis’.
7
Messiaen, Technique de mon Langage Musical (Paris: Alphonse Leduc, 1944); cf. Chapter XIV.
   ’ ‘ ’ 5

colour the piano timbre. The first of these two features suggests a com-
parison with Vivier’s obsession with melody, increasingly the focus of
his music from Learning of 1976 onwards, and essential to Lonely Child;
but this is arguably fortuitous and, besides, Vivier’s melodies are actually
more conventional than the formula-type melody used by Stockhausen.
The other aspect of Mantra is of greater consequence: the pianists
modify the piano sound by operating a sine-tone generator and a ring
modulator. This latter, a common feature on early modular synthesiz-
ers, is a device that enjoyed a degree of vogue in the late 1960s. A ring
modulator takes two signals as input and produces a signal containing
the sum and difference of their respective frequencies. If, for example,
the inputs are two sine waves of 200Hz and 300Hz, the output from the
ring modulator will be two frequencies of 500Hz (300Hz+200Hz) and
100Hz (300Hz–200Hz) – the sum and difference tones respectively. If
richer waveforms than sine tones are used as input (in Mantra, the sound
of the grand pianos) then the output signals from the ring modulator
may be very complex; not infrequently the resulting sounds are like
bells or various metallic sounds. Ring modulation was seized upon in
the early studios as a way of creating inharmonic timbres; as we shall
see in a moment, the analogous technique used by Vivier generates a
sort of harmony that is likewise ‘inharmonic’, and not based on the
natural harmonic series. (Shortly before his death Vivier remarked to
Grisey: ‘I’m also composing with spectra now. You’ve influenced me…
only I twist mine a little!’8)

3. ‘l’addition des fréquences’

The new technique in Vivier’s later music, the crucial innovation that
underlies pieces like Lonely Child, Prologue pour un Marco Polo, Wo Bist
du Licht!, Bouchara and several others, came not from Stockhausen but
from Paris, and specifically from the work of Grisey and Murail. (Vivier
had come to know Grisey fairly well in 1972, but had seen little of him
between 1974, when he returned to Montreal upon completion of his
studies in Köln, and his European trip late in 1979; it is not clear when
exactly he first met Murail, whom he knew less well.) When he went
to Paris in 1979 Vivier would have been relatively out of touch with
their recent music, and his renewed encounter with it had the spirit of
a new discovery. In a lecture he gave in Paris in November 1982, Vivier
acknowledged that the music he was then composing made use of
‘a very classical system that Murail uses: a system of addition des
fréquences’.9
In English, this technique goes by different names. The literal trans-
lation, ‘addition of frequencies’, has not entered general usage. The
technique is sometimes referred to as the ‘sum and difference’ principle
of chord generation. François Rose, in a 1996 article on spectral music,
calls the technique one of generating ‘combination tones’.10 Vivier
himself, as we have noted, referred to the materials so produced as ‘les
couleurs’.

8
Gérard Grisey, ‘Autoportrait avec l’Itinéraire’ (1991), reprinted in Danielle Cohen-
Levinas, ed., Vingt-cinq ans de création musicale contemporaine: L’Itinéraire en temps réel (Paris:
L’Itinéraire/L’Harmattan, 1998), 49.
9
Vivier, lecture at the Centre Culturel Canadien in Paris, 23 November 1982, entitled ‘La cou-
leur spectrale développé dans trois oeuvres: à la recherche d’une nouvelle mélodie’: tape
recording courtesy of the Fondation Vivier, Montreal. It seems odd that he should describe
as ‘très classique’ a system that had been in use for less than a decade; perhaps he wanted to
make the technique seem more of an established practice than it really was.
10
François Rose, ‘Introduction to the Pitch Organization of French Spectral Music’, in
Perspectives of New Music 34 no.2 (Summer 1996), 6–39; cf. especially pp. 20–29.
6 

Let us examine this principle in detail, using first an extract from an

early work of Murail, Ethers (1978) for flute and five instruments. Unlike
Grisey’s work with sonograms of trombone spectra, Murail did not
begin from any spectrum analysis of the flute timbre, but rather with the
possibilities of ‘extended techniques’ on the flute. Near the beginning of
the piece the flautist (playing bass flute) enters quietly on a single pitch
(C#4, 277 Hz), then sings the same pitch while still playing it; then keeps

nation of the sung C# and the played B  gives two resulting sounds: a
on singing the C# while playing a new pitch (Bb5, 932 Hz). The combi-

sum tone of 1209 Hz (932+277), which is D6 a quartertone higher; and a

difference tone, 655 Hz (932–277), which is E5. The beginning of Ethers
is entirely based on these pitches – the played notes, the sung note, and
their sum and difference tones, which are then taken up by the other
instruments. Because Murail is working with acoustic instruments rath-
er than sine waves, he takes into account also the harmonics of the flute
timbre and the vocal timbre; thus, as the music develops, the combina-
tion tones used are not only of the form a+b and a–b, but also 2a+b,
2a–b, a+2b, a–2b, 3a+b, 3a–b, and so on. These ‘second-order’ modu-
lation sounds (all actually present, acoustically, as partials) are played
– simulated – by the strings, thereby becoming ‘real’ pitches, and greatly
enriching the harmony.
This is, in essence, the principle used by Vivier in his music of 1980
onward to create the complex sonorities that he called ‘les couleurs’.
His technique does not require the use of electronics at any stage,
merely the calculation of the frequency values (in cycles per second) of
musical notes. Just as in electronic music, where one pitch is ring-modu-
lated against another, so too do Vivier’s calculations involve two pitches.
Typically, this is a melody note and the bass note below it: these two

Claude Vivier
Latest publications
from the Vivier
J. A. Billard

catalogue
Hymnen an die Nacht
soprano and piano £5.50
Pianoforte
solo piano £5.50
Pièce pour violon et clarinette
violin and clarinet £5.99
Pièce pour violoncelle et piano
cello and piano £5.99
Lettura di Dante
soprano and ensemble £10.99
Paramirabo
ensemble £13.99

To purchase music by Claude Vivier visit www.boosey.com/shop

   ’ ‘ ’ 7

pitches are then subjected to an analogy of ring-modulation to produce

new sum and difference tones, although in practice Vivier only makes
use of sum tones. (The dyadic basis of Vivier’s later music, essential to
his use of this technique, will be discussed more fully later.)
As an example, let us look at bars 24–28 of Lonely Child, the first
appearance of ‘les couleurs’ in Vivier’s output (Figure 1). The orches-
tral introduction to the piece, the first 23 measures, does not use them:
the texture is entirely single melodic lines or dyads, always in octaves,
punctuated by the occasional chiming of the rin, a Japanese percussion
instrument that is a crucial ingredient in the overall sonority of Lonely
Child. But when the soprano enters, ‘les couleurs’ appear with her in the
first violins, divisi in six parts. The first note she sings is an A, against a G
in the bass (in the orchestral texture both notes are doubled in octaves,
A by the two clarinets, first horn, second violins and violas, and G by the
second horn, cellos and double basses).

Figure 1:
Vivier, Lonely Child bars 24–28, showing the spectrally derived harmonies
(in Vlns I) accompanying the soprano.
8 

The A of the soprano, 440 Hz, and the G of the second horn and cel-
los, 196 Hz, when ‘ring-modulated’, produce the combination tone of
636 Hz (a+b), a pitch somewhere between the equal-tempered D# and
E (622 Hz and 659 Hz respectively); Vivier notates it as an E with a down-
ward-pointing arrow before it, indicating a quartertone lower. (The
exact frequency of the quartertone between D# and E is 640 Hz, very
slightly higher than the pitch in question, 636 Hz; but this slight degree
of inaccuracy is not considered to be of much consequence.) The result-
ing pitch is played by the fifth of the six first violins. Then the process
continues: the new pitch, E a quartertone lower, is itself ring-modulated
against the original G: 636 Hz plus 196 Hz gives the combination tone
of 832 Hz (a+2b), almost exactly a tempered G# (which is 830.6 Hz);
this pitch is taken by player four of the first violins. This new note is in
turn ring-modulated against the G: 832 Hz plus 196 Hz gives 1028 Hz
(a+3b), a bit lower than the tempered C (1047 Hz) although not as low
as the quartertone below it (1017 Hz); Vivier thus simply writes the note
C without any inflection. This C is played by the third first violin. And so
the process continues, with two more, still higher, combination tones.
The total of five combination tones, played together, make up the ‘cou-
leur’– the spectral harmony – of the generative interval G–A (196 Hz
– 440 Hz).
When the soprano sings her next pitch, a Bb, still against the same

array of combination tones. The B  is 466 Hz, which when ring-modu-

G in the bass, a new set of calculations must begin, yielding a different

lated against the G of 192 Hz gives a combination tone of 662 Hz, which
falls between E 659 Hz and the quartertone above it (679 Hz); Vivier
nonetheless opts for the quartertone, which he notates as F with a down-
ward-pointing arrow (in this case artistic taste would seem to have won
out over acoustical accuracy, as the combination tone generated is much
closer to the E). And so begins the building up of another ‘colour’.
This process, with the sheaves of calculations necessary to it, may
seem exhausting, involving massive amounts of precompositional
working-out; but in actual fact the process is somewhat less extensive
than we might imagine. Because of the nature of the melodic writing
in Lonely Child, which restricts itself to a relatively modest palette of
pitches, there is a comparatively small number of different melody-note-
and-bass combinations used in the work. The first extended soprano
passage, which Vivier calls ‘mélodie 1’ (bars 24–66), uses only 15 differ-
ent such combinations; the second, ‘mélodie 2’ (bars 78–111) uses 38,
of which Vivier builds ‘couleurs’ on 36 (two combinations are used only
in passing); the third, ‘mélodie 3’ (bars 113–140), by far the most chro-
matically elaborate, uses 89; but ‘mélodie 4’ (bars 150–180) uses only 11.
Figure 2 shows a page of Vivier’s sketches for Lonely Child, showing 16
of the combinations from ‘mélodie 3’. The melody note is here notated
as the first of the pitches on the upper of the pairs of staves, followed by
the combination tones they generate when ring-modulated against the
bass note: sounded together, all five of these combination tones make
up the ‘couleur’ of that particular generative interval.
There is, in this manner of composing, an intriguing paradox: namely,
that a process of pitch generation that is so precisely calculated (by work-
ing out Hertz values of the first-, second-, third- and even higher-order
combination tones) should yield a considerable amount of approxima-
tion when the results are written down in musical notation. Clearly, the
Hertz values must be ‘translated’ into values meaningful to performing
musicians. In doing so, Vivier restricts himself to quartertones as a way
of augmenting the limited pitch palette of 12-tone equal temperament
– had he used eighth-tones also (as Grisey and Murail had already done
   ’ ‘ ’ 9

Figure 2: in various of their works of the 1970s), the approximations would be

From Vivier’s sketches for Lonely much less but the work would be correspondingly harder to play, the
Child, showing part of the derivation players having to negotiate 48 divisions of the octave rather than 24.
of ‘les couleurs’ of mélodie 3.
But perhaps this paradox is more imaginary than real. In an approach
such as Vivier’s – and this point has been made by the composer Joshua
Fineberg as true of spectral music in general – microtones are simply
approximations of a set of frequencies to the nearest available musical
pitches; ‘the ear is able to hear past these approximations and hear the
underlying frequency structure whenever the approximation is within
tolerable limits.’11 (This then raises the fascinating question of how play-
ers should approach the intonation of the pitch materials used by Vivier,
or indeed in spectral music generally; in a piece like Lonely Child, where
the notated pitches are approximations of the ideal pitch relationships,
the intervals need not, and perhaps ideally should not, be played as exact
quartertones.)12

11
Joshua Fineberg, ‘Guide to the Basic Concepts and Techniques of Spectral Music’, in
Contemporary Music Review 19 no.2 (2000), 81–113; cf. p.84.
12
It should be noted that Vivier’s calculations of the combination tones used in Lonely Child,
as seen in his sketches, contain a good many errors. These would seem to be the result
of doing the calculations by hand (rather than with a calculator or computer), probably
at Vivier’s characteristic high speed. The errors remained for the most part undetected by
him and were incorporated into the score. In working on this article I found what seem to
me more than ten inexplicable pitch choices out of the 75 combination tones in ‘mélodie
1’ alone, the only plausible explanation for which is mathematical error. And yet, possible
though it would be, it somehow seems utterly repugnant to imagine a ‘corrected’ version of
the score. (W. B. Yeats, in his poem ‘The Scholars’, was surely right to castigate those ‘old,
learned, respectable bald heads’ who ‘edit and annotate the lines / That young men, tossing
on their beds, / Rhymed out in love’s despair’.)
10 

4. Formal design
Lonely Child is, of course, ‘about’ much more than its innovatory use of
this new timbral/harmonic principle, and in this section I will look at
the overall form of the work and the nature of its musical materials in
general before examining Vivier’s use of ‘les couleurs’ in context.
Although Vivier himself did not leave us a written analysis of Lonely
Child he spoke of the work several times in interviews and lectures. The
comments below were made in the context of an interview for Radio-
Canada, and give an accurate although extremely condensed statement
of the compositional techniques used in the piece:
Lonely Child is a long song of solitude. In constructing it musically, I wanted to
have total control in the expression and musical development of the work I was
composing without using chords, harmony, or counterpoint. I wanted to arrive
at a music that was very homophonic, that could transform itself into a single
melody, which would then be ‘intervalised’. I had already composed the first
melody, heard at the beginning of the piece, for dancers. Afterwards I developed
this melody into five melodic fragments that are ‘intervalised’, that is, by adding
one note below another, giving intervals: thirds, fifths, minor seconds, major
seconds, etc. If you make a sort of frequency addition of each of these inter-
vals, you arrive at a timbre. So there are no longer any chords, and the whole
orchestral body is transformed in this way into timbre. The roughness and the
intensity of this timbre depends on the generative interval. Musically, I had one
single thing to master which, in a way automatically, would give rise to the rest
of the music: great beams of colour!13
It is possible to trace the development of these elements by exam-
ining the sketches for Lonely Child. Vivier grouped his sketches for the
work into four small bundles which bear the titles ‘1o mélodie’, ‘2o
mélodie’, ‘3o mélodie’, and ‘4o mélodie’. This shows that, in his mind,
the primary sections of the work were the four long passages for the
soprano which form the main substance of the piece. These are framed
by a prologue and epilogue, the whole punctuated by four interludes
(the first in the middle of mélodie 1, the second between mélodies 1 and
2, the third between mélodies 3 and 4, and the fourth between mélodie
4 and the epilogue). Each of the four bundles contains sketches towards
the vocal lines themselves (and their bass lines), and page after page of
calculations of the combination tones resulting from the various melo-
dy-note-and-bass dyads.
The overall form of Lonely Child is easy enough to describe, although
the details of the description will vary slightly depending on which
aspects of the work one hears as ‘core’. Building from Vivier’s concep-
tion of the four ‘mélodies’, a plausible division would be as follows:
Bars 1–23 prologue: opening melody in the orchestra, first a
single line (doubled in octaves) then, as it proceeds, in
two parts;
Bars 24–77 first mélodie for the soprano (in two halves, punctuated
by an interlude for woodwinds with brake drums
and violins at mm.42–53, and with a brief transition
passage beginning at m.67);
Bars 78–112 second mélodie (with first use of invented language
and the hand tremolo technique – ‘ululation’ – for the
soprano);
Bars 113–140 third mélodie (which follows the second after a gap of
only one measure);

13
Vivier, ‘Lonely Child’, in Véronique Robert, ed., ‘Les écrits de Claude Vivier’, in Circuit:
Revue Nord-Américaine de Musique du Xxe siècle vo.2. nos.1–2 (1991), 108.
   ’ ‘ ’ 11

Bars 140–149 fff passage for bass drum (with rin at the end);
Bars 150–180 fourth mélodie;
Bars 181–194 interlude (ending with harmonic series on the double
bass);
Bars 195–214 epilogue: return to the opening melody in the
orchestra, at a slower tempo and with addition of
‘aleatoric colours’ in the strings
However self-evident this may seem when one gets to know the work,
there are some possible variations. Figure 3 shows Vivier’s own idi-
osyncratic description of the form, as found in his lecture notes for a
presentation he gave on his music at Feedback Studios in Köln on
November 1, 1982.

Figure 3:
A page from Vivier’s lecture notes
for a presentation at Feedback
Studios, Köln, in November 1982,
showing the outline of the form of
Lonely Child.

A transcription of Figure 3 is given below:

Lonely Child
1 mélodie
2 Farben ________ + [ bläsers]
3 _____ huluement rythmen. / ___
4 Haupt farbe < > dauern Kurz lang. bläser + violo.
Neben farbe > 2 Stimmig komplementäre dauern.
Vcl. + Kontrabasse halten die Haupt töne
5 Grosse trommel
6 Farben Akkorde reine interval
7 Interlude
8 Mélodie + Aleatorische farben.
Here, unlike in his sketches, Vivier somewhat confusingly uses the
term ‘mélodie’ to correspond to the opening orchestral section, bars 1–
23, before the voice enters. His second section, ‘Farben’, refers to the first
appearance of ‘les couleurs’ with what in the sketches he calls ‘mélodie 1’,
the first long passage for the soprano. His third section corresponds to
the new ‘couleurs’ with mélodie 2, which features his characteristic voice
hand tremolo (in French, ‘huluement’), a tremolo produced by the hand
12 

in front of the mouth; ‘rythmen’ signifies the increasing accumulation

of written-out rhythms (as opposed to tremolos) on single pitches in
the orchestra parts which begins at bar 84 and goes all the way to bar
105. In section 4, mélodie 3, he distinguishes between the ‘main colour’
(‘Haupt Farbe’) and the ‘neighbouring colour’ (‘Neben Farbe’), a dis-
tinction we shall see later. Then comes the sequence of ten explosive fff
bass drum stokes, spaced ever-closer together, that falls (symbolically, if
not quite literally) at the Golden Section of the piece. It is followed by
the fourth set of ‘Farben’ with mélodie 4, and, after an interlude with
wordless soprano, by the return of the opening melody at a slower
tempo, with ‘aleatoric colours’ (glissandi of harmonics) in the strings.
Graphically, the overall dynamic contour of the piece, based on the 1995
recording by Susan Narucki with the Schoenberg Ensemble conducted
by Reinbert de Leeuw, looks as shown in Figure 4: the bass drum pas-
sage stands out clearly.14

Figure 4: The most striking and, from a compositional point of view, coura-
The waveform of the Narucki/De geous structural feature of the whole work is its wholesale reliance
Leeuw recording of Lonely Child,
showing the dynamic shape of the
on an essentially two-part texture of melody and bass enhanced, ‘in a
work. way automatically’, by ‘les couleurs’. There is, as Vivier himself was
well aware, no real counterpoint to speak of, although even with two
voices there is a kind of functional harmony. The dyad is a character-
istic Vivier sonority; it is to his music what the triad is to tonal music.
Earlier works like Learning (1976) or Pulau Dewata (1977) had been large-
ly homophonic, their single lines sometimes becoming two-part but
rarely contrapuntal in the strict sense. This type of texture reappears in
subsequent works, such as the odd but wonderful Greeting Music (1978);
the piano solo that opens that work juxtaposes gamelan-style melodic
fragments with dyads treated with 19th-century-style octave doublings.
Dyadic textures characterize parts of the opera Kopernikus (1978–79) and
the last section of the orchestral work Orion (1979). However in Lonely
Child this concentration on melody and bass becomes the predominant
texture. In the same Paris lecture mentioned above, Vivier recounts an
experience that made him realize the importance of two-part textures:
One day I heard a rehearsal of a chorale by Bach and I only heard two voices; I
heard the alto and soprano voices together. It was an extraordinary experience
because I heard that the music was still fantastic… two sounds superimposed, as
an interval, are musically as important as a chord.15
The overall ‘sound’ of Lonely Child – the sonic image that lingers in the
mind once the music has stopped – has, arguably, less to do with the
spectral elements than it does with Vivier’s particular sense of melody
and harmony. This is already evident from the opening section, which
is a single line that occasionally becomes two-part; it is melodic, yet
somehow unsure of its direction. Vivier uses almost entirely consonant
14
Philips 454 231–2.
15
Vivier, lecture at the Centre Culturel Canadien, Paris, 23 November 1982.
   ’ ‘ ’ 13

intervals both harmonically (minor and major thirds, perfect fourth,

perfect fifth) and melodically (minor and major thirds, perfect fifth,
major sixth). One simple but delicious harmonic effect is the recontex-
tualization of a pitch in the melody that results from changing what was
a major third to a minor third, or vice versa (Figure 5).

Figure 5: At the beginning of the work the other sonority that creates a spe-
Lonely Child, opening. cial atmosphere is the rin, the Japanese percussion instrument, Buddhist
in origin, shaped like a rice bowl and traditionally made of copper or
bronze. Vivier had fallen in love with the instrument on his travels in
Asia a few years previously (although he had probably encountered
it before; Stockhausen had used it in the realization of several of the
pieces in Aus den sieben Tagen). Struck with a padded mallet, the rin has
a very long ring-time. (The instrument comes in different sizes with cor-
respondingly different pitches, but Vivier specifies a ‘very low’ one.) The
sound, not specifically identifiable to most listeners, unmistakably lends
the piece a ‘non-western’ and quasi-ritualistic aura.

5. The four ‘mélodies’

Having looked at the overall form of Lonely Child and at the constitution
of its spectral harmonies, it is fascinating to see the various ways Vivier
uses these ‘couleurs’ in the four main vocal sections of the piece. In mélo-
14 

die 1, as we have seen, the use of ‘les couleurs’ could hardly be simpler
from a textural point of view: they proceed in rhythmic unison with the
changes of melody note, serving almost literally to colour the melody
as it proceeds.
For mélodie 2, bars 78–111, there is a change. Here the generative
intervals are not strictly speaking the soprano note and bass note, but
the dyads in the cellos (divisi, with cellos 1 and 2 playing the upper note
and cello 3 playing the lower note). ‘Les couleurs’ change in rhythmic
unison with the cellos, rather than with the voice part which, although
having the same pitches as the upper cello line (one octave higher), stag-
gers them rhythmically, anticipating or echoing each new cello note.

Figure 6:
‘Les couleurs’ from the beginning of
mélodie 3, bars 113–114, voice and
strings only.
   ’ ‘ ’ 15

The complexities go further, however, as Vivier mixes some of the

violins play the combination tones arising from the cello dyad E –B ;
spectra arising from the generative intervals. In bar 78, for example, the

but against them the woodwinds play a chord using combination tones
of the interval F–G, the next dyad along, but not yet heard. (This mixing
of sets of combination tones recalls Murail’s technique of premonitions
and echoes, as found for example in his Treize Couleurs du Soleil Couchant
(1978), although the connexion is possibly fortuitous.)
For mélodie 3, bars. 113–140, Vivier introduces yet another new use
of ‘les couleurs’. Because the soprano melody (to the words ‘The stars
make amazing leaps’) is relatively fast-moving, and has more agile pitch
movement than any of the other mélodies, Vivier elects not to have the
first violins play their chords note-for-note against the melody line as
previously. Instead, the chords are hocketed between first and second
violins, so that the combination tones generated by the first interval
are played by Violins I (divisi, as before), those of the second interval by
Violins II, those of the third by Violins I, and so on. To this principle he
introduces yet a further modification: the violins do not play discrete
pitches but glissando constantly from one note to the next. However,
their glissando is not a ‘straight line’; they move first in the direction of
the intermediary note played by the other violin group, as though aim-
ing for that pitch (whether higher or lower), before changing direction
to move to their own next note (Figure 6). The resulting violin texture,
in which every note is marked with a diminuendo (as though its reso-
nance was continually dying away, as with bell- or piano-tones), persists
throughout this third mélodie, creating what Vivier calls ‘aleatory col-
ours’ (a similar texture appears towards the end of the work).
For mélodie 4, bars 150–180, Vivier employs at first a selective prin-
ciple by which only certain melody-note-and-bass combinations
(chosen, it would seem, freely) generate a ‘couleur’ which is then sus-
tained through changes of pitch in the soprano and bass. This creates a
sort of resonance effect, the sonority lasting longer than its generative
dyad. Then, as the melody gains in emotional intensity and rhythmic
momentum, the ‘couleurs’ change more frequently until, by the end,
they are moving in rhythmic unison with it. There follows an inter-
lude, culminating (at mm. 192–4) with a solo double bass sounding in
sequence the first eleven partials of its E string. Curiously, in a work so
rich in spectral thinking, this is the only audible presence of the natural
harmonic series in the whole piece. Its appearance seems symbolic of
something, perhaps the attainment of a certain calm – a mood that is,
however, destabilized by the work’s very last dyad, G–D doubled across
five octaves, and by the two final strokes on the rin, the second of them
suddenly, even cruelly, damped before it has lived out its natural sonic
lifetime.

6. Conclusion
What, then, is this beautiful, innovatory, haunting masterpiece ultimate-
ly ‘about’? Almost everyone seems to hear the work as autobiographical,
with Vivier himself as the ‘lonely child’ of the title. The text, a curious
amalgam of the beautiful and the maudlin, the vivid and the incom-
prehensible, offers us some possible meanings, but only some. ‘Oh
beautiful child of the light’, it begins, ‘sleep, sleep, sleep on’… ‘dreams
will come / sweet fairies will come to dance with you … fairies and elves
will celebrate with you / the joyful farandole will intoxicate you’. A sim-
ple benediction from mother to child? Is Vivier himself, as author of the
text, then speaking as his own mother, the mother he never knew, hear-
16 

ing words he so desperately wanted to hear? The second mélodie, with

its sudden introduction of invented language (‘Ka rè nou ya zo na-ou dè
wa ki’) pulls us away from the world of possible meanings into linguistic
terra incognita. Who is saying these words – is the child now ‘speaking’
to the mother? (It may or may not be relevant to note that Vivier did not
know his birth mother’s own native language, and therefore in which
language he should address her.) The text ends in the same benedictory
spirit in which it began – ‘And the hope of time / of time outside time /
appears, my child / the stars in the sky shine for you, Tazio, and love you
forever’ – although the loved child has now, at the beginning of mélodie
4, acquired a name: Tazio, with its resonances of the beautiful youth,
Tadzio, of Thomas Mann’s Death in Venice. This act of naming suggests
other possible readings: what had seemed at the outset like the love of a
mother for her child can now be understood as love of a sexual, perhaps
homoerotic, kind.
Any great work of art, of course, has multiple meanings, which
enlarge and enrich themselves as the work itself moves further into his-
tory. The autobiographical, self-revelatory nature of Lonely Child, so rich
a part of its meaning, has an added poignancy given the fact that Vivier
himself never heard the work in concert. Its first performance was for
a studio recording by the Canadian radio in Vancouver by the soprano
Marie-Danielle Parent, for whom it was written, and the Chamber
Orchestra of Radio Canada, Vancouver conducted by Serge Garant (7
January 1981). Vivier himself was not present: recently returned from
another trip to Europe, he was unable to attend the session for reasons
that are not certain but are probably to do with the lack of money for
a plane ticket. (Marie-Danielle Parent has pointed out that although
Vivier had a commission fee he didn’t necessarily have a travel grant to
attend the recording: ‘I remember that Claude had told me he couldn’t
be there but that he had every confidence in Serge Garant and myself
to give a good performance of the work.’16) Although he never heard
Lonely Child live, he at least had a copy of the tape made in Vancouver,
which he presented several times in lectures and public talks.17 The first
public performance of Lonely Child was not until nearly two years later,
18 November 1982, in the Salle Claude Champagne in Montreal, with
the same soloist and conductor and the Orchestre des Jeunes du Québec.
This, however, was during the period when Vivier was in Paris for what
would turn out to be the last months of his life. The next public per-
formance was given as part of the commemorative concert for Vivier
in Montreal on 2 June 1983, three months after his death. By then, the
request that ends mélodie 3 of Lonely Child – ‘please give me eternity’
– had already been granted.

Note
I would like to thank Rozalie Hirs and Donnacha Dennehy for their
encouragement as I worked on this article. I also wish to acknowledge
the stimulus of the work of two Canadian scholars who have written on
Lonely Child: Jacques Tremblay’s ‘L’écriture à haute voix: Lonely Child de
Claude Vivier’, in Circuit 11, no.1 (2000), 45–67; and Patrick Lévesque’s

16
Marie-Danielle Parent, e-mail to the author, 18 September 2004.
17
‘Claude must have heard this version’, notes Marie-Danielle Parent, ‘because he gave me
some excellent comments’ (email to the author, 18 September 2004). The recording by
Marie-Danielle Parent and the Orchestre métropolitain de Montréal conducted by Serge
Garant, presently available on the CD Montréal Postmoderne (Centredisques CMC-CD 5194)
is not the Vancouver performance but a slightly later one, recorded in the Salle Claude
Champagne in Montreal some months after Vivier’s death.
   ’ ‘ ’ 17

‘LES VOIX DE VIVIER: Langage harmonique, langage mélodique et lan-

gage imaginaire dans les dernières oeuvres vocales de Claude Vivier’
(Masters thesis, McGill University, Montreal, 2004). My thanks also to
the President of the Fondation Vivier in Montreal, Madame Thérèse
Desjardins, for permission to quote from unpublished materials of
Claude Vivier; and to Marie-Danielle Parent for talking with me about
Vivier and Lonely Child. All translations from the French in this article
are my own.