DOI: 10.1111/j.1468-2249.2011.00294.

robert hasegawa
Ge rard Grisey and the Nature of Harmony

musa_294

349..372

Grard Grisey (19461998) was a founding member of the spectral movement

a group of French composers born in the 1940s whose best-known members
are Grisey, Tristan Murail, Michal Levinas and Hugues Dufourt. Spectral
music emerged in the 1970s, in part as a reaction against the abstraction of serial
music. Instead of basing their music on the manipulation of rows or motives,
spectral composers take inspiration from the physical properties of sound itself.
Each of these composers defines spectral music differently (some even rejecting
the label altogether), but as a generalisation we could say that the essential
characteristic of spectralism is the dissection of sounds into collections of partials
or overtones as a major compositional and conceptual device. Spectral composers use the acoustical fingerprints of sounds their spectra as basic musical
material.
In their writings, spectral composers have often emphasised the natural
origins of this material, even while acknowledging the artificiality of some of the
procedures used to transform and develop spectral pitch sets. In this article I
explore how Griseys music invokes the idea of nature and what this idea might
mean for listeners and analysts. For Grisey, the mimicry of features of natural
sounds is an essential compositional technique; such procedures are amply
documented in analytical studies based on sketch material.1 While such studies
can explain how nature is harnessed in Griseys music, they tend to overlook the
equally important role played by nature in a quite different sense: the way our
innate mechanisms of aural perception make sense of musical sound. For the
analyst of Griseys music, these contrasting concepts of nature one based on
the objective, physical nature of external reality, the other on the subjective,
internal nature of aural perception lead to very different ways of thinking about
musical structure. There is often a significant gap between the theoretical structures produced by spectral compositional procedures and the perception of these
structures by the listener; due to this gap, an account of a work based solely on
a reconstruction of compositional procedure often fails to reflect a listeners
experience of the work. By developing an analytical method which reflects the
natural biases of our aural perception, we can arrive at an analysis more sensitive
to the actual experience of listening to Griseys music.
Instrumental Synthesis and Inharmonicity
Among the most characteristic procedures of spectral composition is instrumental synthesis: this technique mimics the electronic music technique of additive
Music Analysis, 28/ii-iii (2009)
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK
and 350 Main Street, Malden, MA 02148, USA

349

350

robert hasegawa

Ex. 1a Spectral analysis of a trombone sound

synthesis, but replaces pure sine tones with the complex sounds of real instruments. An iconic example is the opening of Griseys Partiels for chamber orchestra (1975), based on a fortissimo trombone E2.2 The trombone sound can be
analysed into a set of partials of varying frequencies and amplitudes; this can be
expressed either as a numerical table or graphically as a spectrogram (Ex. 1a). In
a spectrogram of a sound, the sounds evolution in time is represented on the x
axis from left to right, and frequency is shown on the y axis, with low frequencies
at the bottom and high ones at the top. The intensity of vibrational energy at any
frequency is indicated by shades of grey from light (weak) to dark (strong).
Ex. 1b reproduces the opening page of the score. We first hear the trombone
itself, accompanied by sforzandi in the double bass an octave below; as the
trombone fades out, instruments from the ensemble enter gradually from low to
high, playing pitches which match selected partials of the analysed trombone
sound. For example, the third partial (played by the clarinet) is a perfect twelfth
above the trombones fundamental, with a frequency three times that of the
fundamental; the cellos G approximates the trombones fifth partial, and so on.
Grisey uses the strength of each partial in the trombone analysis to assign
dynamics to the instruments participating in the synthesis, and also to shape the
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

351

Ex. 1b Instrumental synthesis at the opening of Grisey, Partiels (1975)

order of entries. In most brass sounds, the upper partials emerge slightly later
than the lower ones, a phenomenon which Grisey imitates (on a greatly
expanded time scale) with the staggered entries in his synthesised replica of the
trombone.The goal of instrumental synthesis is not a precise reproduction of the
trombone sound which would in any case be impossible given the complex
spectra of acoustic instruments but rather a hybrid sonority permitting us to
hear both the individual instruments and their fusion into a unified timbre.
The physical properties of sound are brought into focus by these techniques
of analysis and re-synthesis; this is an appeal to nature in the objective sense of
the term. For music theorists, Griseys technique will have strong echoes of
Rameaus corps sonore. The essential difference, however, is that Grisey is dealing
with real sounds, not with an idealised source of overtones. Recall that Rameaus
corps sonore, as formulated in the Gnration harmonique, conveniently stopped
vibrating after the sixth partial to avoid the out-of-tune natural seventh.3 In
contrast, Grisey carries into his music the complexities of real sounds, including
their often distorted and imperfect spectra.
Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

352

robert hasegawa

Though we tend to think of the frequencies of a sounds partials as corresponding precisely to the harmonic series (x, 2x, 3x ... ), some of the most common
musical sounds have inharmonic spectra. A piano string, for example, produces a
stretched spectrum: that is, the first overtone is not exactly twice the frequency of
the fundamental (a perfect octave), but slightly higher. (Only an idealised string
with no mass or resistance would produce a pure harmonic spectrum.) The
stretching continues into the higher partials.We might not realise it (although our
piano tuners do), but by the fourth octave the partials of a low piano note are
approximately a third of a whole tone (65 cents) higher than their equivalents in
a pure harmonic series. Other spectra, such as those of certain brass instruments,
are compressed: each partial is lower than its harmonic counterpart.

Inharmonic Spectra in Vortex Temporum

Grisey exploits these real-world departures from ideal harmonicity in the design
of his 1996 chamber ensemble piece Vortex Temporum, for flute, clarinet, string
trio and retuned piano (four pitches are lowered by a quarter tone). His compositional procedures are extensively documented in sketch-study monographs
by Jean-Luc Herv and Jrme Baillet, and sketches in the Paul Sacher Foundation confirm their findings. The section which follows retraces some of
Griseys techniques to illustrate how he brings inharmonic spectra into his
music; after exploring these compositional derivations, we shall examine how the
harmonies might actually be perceived.
Throughout Vortex Temporum, Grisey uses only three types of spectra harmonic, stretched and compressed which are transposed to start on different
fundamentals. The first stave of Ex. 2 shows the pitches of a harmonic spectrum
on B0; each partial is represented by the nearest equal-temperament pitch, with
the deviation in cents (rounded to the nearest cent) indicated below the note. For
example, the seventh partial of the spectrum is A3, flattened by approximately 31
cents. The spectrum of a piano tone on the same fundamental, computed by
Fourier analysis, is shown on the second stave of Ex. 2. In the case of low partials,
the stretching of the spectrum is barely perceptible; the second partial is just 4
cents higher than its equivalent in the harmonic spectrum.The higher the partial,
the more pronounced and obvious the stretching becomes, to the point that the
twelfth partial is 41 cents (almost a quarter tone) above its harmonic equivalent.
Grisey is not content with the relatively slight stretching of the natural piano
spectrum. He constructs a more exaggeratedly stretched version for use in Vortex
Temporum (see the third stave of Ex. 2, which shows Griseys stretched spectrum
in his preferred quarter-tone approximation). Although the natural twelfth
partial of the piano tone is 41 cents sharp relative to its harmonic partial, in
Griseys stretched spectrum it is 198 cents sharp almost an equal-tempered
whole tone above the equivalent harmonic partial.To calculate this exaggeratedly
stretched spectrum, Grisey uses the equation fn = f0(n1.046); that is, the frequency
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

+2

-14 +2 -31

+4

11

12

13

14

15

-14 -49 +2 +41 -31 -12

10

17

0 +5

16

19

+4 -2

18

21

22

23

24

25

26

-14 -29 -49 +28 +2 -27 +41

20

+6

27

Music Analysis, 28/ii-iii (2009)

+9

+4

10

11

12

13

14

15

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

10

11

12

13

14

15

16

-5

26

17

17

19

20

21

22

23

24

25

26

18

19

20

21

22

23

24

25

26

(approximated to the nearest quarter tone)

18

(approximated to the nearest quarter tone)

0.954

1.046

4. Grisey's compressed spectrum on B 0, calculated by the formula y = Fx

16

+12 -2 +15 -14 +21 +30 +16 -14 +43 -13 +22 +47 -35 -20 -15 -13 -17 -21 -34 -47 +36 +17

3. Grisey's stretched spectrum on B 0, calculated by the formula y = Fx

28

-31

28

30

27

27

28

28

31

29

30

29

29

30

30

32

32

31

31

32

32

-45 +22

31

+30 -12 +45

29

-29 +45 +17 -14

27

2. Spectrum of a piano tone on B 0 (data from Fourier analysis, deviation in cents from equal temperament below each note)

1. Harmonic spectrum on B 0 (deviation in cents from equal temperament below each note)

Ex. 2 Comparison of a harmonic spectrum, the spectrum of a low piano pitch and Griseys stretched and compressed spectra
from Vortex Temporum

Ge rard Grisey and the Nature of Harmony

353

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

354

robert hasegawa

of partial number n equals the fundamental frequency f0 multiplied by the partial

number raised to the power of the constant 1.046.4 This constant produces
stretching of approximately a quarter tone per octave. Grisey constructs a
compressed spectrum with the equation fn = f0(n0.954), which produces comparable distortions in the opposite direction: the spectrum is compressed by a
quarter tone per octave (see the fourth stave of Ex. 2). Throughout Vortex
Temporum, the degrees of distortion of the stretched and compressed spectra are
fixed.The spectra typically are heard only in filtered form: only selected partials
are played, and the rest (often including the fundamental itself) are omitted.
Griseys distorted spectra are based on the stretching and compression found
in some natural sounds but exaggerate these features to an unnatural degree.
Why does Grisey exaggerate the stretching so drastically? In part, it may be to
make the spectrums inharmonicity apparent even when the partials are rounded
off in a quarter-tone grid: with a smaller degree of stretching, the approximation
to quarter tones could erase the difference between stretched and harmonic
spectra. Also, the exaggerated distortion produces a sonic result reminiscent of
the broken octaves characteristic of much atonal and serial music. (Despite
spectralists professed antipathy to serialism, the characteristic sound of serial
music continued to exert a strong influence on spectral composers.) Most
important, however, we must consider Griseys self-described interest in borderline cases and thresholds in this case, the perceptual threshold, where, as the
degree of stretching increases, a spectrum is no longer heard as a fused timbre
but instead breaks up into a collection of independent pitches.5
This is one way in which nature enters Griseys music: as the reproduction
through instrumental synthesis of the acoustical spectra of real-world sounds,
with their characteristic distortions maintained or even exaggerated. Straightforward as this evocation of nature might seem in an early work such as Partiels,
when we listen to later works such as Vortex Temporum it is often impossible to
hear Griseys harmonies as versions of the natural spectra from which they were
derived. After the harmonies have been subjected to extensive compositional
manipulation exaggerated stretching or compression, approximation to a
quarter-tone grid and omission of many partials their natural acoustical source
is no longer recognisable. What, then, does talking about Griseys compositional
techniques really tell us about how we hear his music? When the natural
derivation of his material collapses and we hear the distorted spectra as complex
chords rather than fused timbres, the appeal to nature in the objective, external
sense fails, and the working of our internal nature the nature of our auditory
perception becomes more relevant to our musical understanding.6
Towards a Theory of Tone Representation
Although Griseys compositional techniques often strain the audible connection
between real-world physical models and their scored adaptations, the basic
tenets of spectral harmony do reflect some of the intuitive ways in which listeners
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

355

make sense of pitch combinations. Modern psychological research confirms the

important role that harmonic spectra (with partial frequencies in the series x, 2x,
3x ... ) play in our parsing and organisation of aural information. Psychoacousticians suggest that we have developed a mental template of the relationships
between partials of a complex harmonic tone from our frequent encounters with
such sounds, and that this template is used to make sense of incoming auditory
data. Such templates are crucial in auditory scene analysis, which is the mental
separation of jumbled aural input into sounds from distinct sources. For
instance, when we hear two violins playing different notes at the same time, we
separate the two sound sources from one another by subconsciously matching
their partials to different templates. When we find a template which matches the
partials which we hear, we perceive the fundamental pitch of the sound. As the
acoustician William Hartmann explains, Modern theories of pitch perception... are foremost pattern matching theories.They assume that the brain has stored
a template for the spectrum of a harmonic tone, and that it attempts to fit the
template to the neurally resolved harmonics of a tone.7
One of the essential concepts of spectralism is the transfer of theories about
the auditory organisation of heard partials to musical contexts, where the basic
building block is not a simple, pure-wave partial, but rather an instrumental
tone (with its own complex spectrum consisting of many partials). Many music
theories can be understood in terms of this analogy: for example, Rameaus
acoustical justification of the major triad is based on its match with partials
16 of an idealised vibrating body, the corps sonore. The Pythagorean tradition
of defining musical intervals by ratios (between either vibrational frequencies
or string lengths) can also be seen as an example of the analogy between
overtones and complex pitches: the whole-number ratios which define just
intervals are also found between the individual partials of a harmonic tone.
The tendency of the ear to group partials which can be understood as overtones of the same fundamental suggests that we have a built-in bias towards
such just intervals: the composer and theorist James Tenney has called the
overtone series and the just intervals it contains the only perceptual givens in
our understanding of pitch relationships.8 Just intervals are the historical basis
of Western music theory: octaves, fifths, fourths, thirds and sixths are all based
on simple just intervals whose frequency ratios can be expressed as two, three,
five and multiples thereof. A number of theorists have argued that the just
intervals are referential sonorities, in the sense that we understand them as the
ideal versions of intervals, even when the intervals we actually hear are out of
tune. As Tenney puts it:
I propose as a general hypothesis in this regard that the auditory system would
tend to interpret any given interval as thus representing or being a variant of
the simplest interval within the tolerance range around the interval actually heard
(where simplest interval means the interval defined by a frequency ratio requiring the smallest integers). The simpler just ratios thus become referential for the
auditory system ... .

Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

356

robert hasegawa
Another hypothesis might be added here, which seems to follow from the first
one, and may help to clarify it; within the tolerance range, a mistuned interval will
still carry the same harmonic sense as the accurately-tuned interval does, although
its timbral quality will be different less clear, or transparent, for example, or
more harsh, tense, or unstable, etc. (Tenney 2001, p. 110; emphasis in original)

Our tolerance for mistuned just intervals is evident in the historical development of temperaments: the essential harmonic meaning of the just
interval remains, even when it is heard only in an approximate, tempered version.
If, following the spectralists, we apply our knowledge of the perception of
partials to the analysis of chords made up of many complex tones, we can make
some musical observations which are impossible in theories not based in psychoacoustics.When we match a heard interval to a referential just interval, we produce
two essential pieces of data: the ratio relating the two pitches and an implied root
or fundamental. Given the pitches E4 and G4, for example, we identify both a just
interval between the two (5:6) and the implied fundamental, C2. The number
assigned to a pitch imparts a harmonic meaning in this example, the 5 means
that we hear the E as the fifth partial of C, not as an independent fundamental.The
process of matching a given collection of pitches to a just-intonation interpretation
is similar to Hugo Riemanns concept of Tonvorstellung, or tone representation.
Riemann proposes that the harmonic meaning of a pitch is determined by how we
imagine it as one of the factors of a major or minor triad: According to whether
a note is imagined as 1, 3, or 5 of a major chord or as I, III, or V of a minor chord,
it is something essentially different and has an entirely different expressive value,
character and content (Riemann 1992, p. 86).
Riemanns triadic model of tone representation allowed only the ratios of
Renaissance just intonation, based on two, three and five but we can expand the
theory of tone representation to allow more complex interval ratios with higher
prime factors. This brings us into the harmonic world of extended just intonation, developed by the American experimental composers Harry Partch, Lou
Harrison and Ben Johnston. Extended just intonation includes many microtonal
intervals which fall between the keys of twelve-note equal temperament, such as
the flat minor seventh (4:7, or 969 cents) or the undecimal tritone (8:11, or 551
cents). If we accept that approximations of these extended just intervals still
convey the same harmonic meaning as the true ratios, many atonal sonorities of
music of the twentieth century can be understood as equal-temperament
approximations of pitch collections in extended just intonation.9
In translating a collection of heard pitches to a referential just-intonation set,
we are guided by what Riemann calls the Principle of the Greatest Possible Economy
for the Musical Imagination (Riemann 1992, p. 88). We choose the simplest
just-intonation pitch set which matches the heard pitches while minimising the
level of mistuning between the heard pitches and their just-intonation counterparts. The tone representation of a given pitch set can be expressed in a simple
notation: to describe the pitches D, E and G as the ninth, tenth and twelfth
partials of a C fundamental, we can write C(9:10:12). Depending on the context,
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

357

we may wish to provide a specific register for the root (for example, C1) or
indicate a microtonal deviation from equal temperament (which can be
expressed in cents from the nearest tempered pitch, for example C+20). Because
many factors combine to determine the simplest tone representation, it is difficult (and not necessarily desirable) to completely formalise the theory; what I
propose instead is a simple model based on preference rules, which gives intuitively satisfying results.10 The flexibility of this model is not a weakness, but
rather one of its greatest strengths, for the way we understand pitches and their
relations needs to be context sensitive to allow for the interaction of other
musical parameters with our harmonic perception.The preference rules outlined
here suggest the most likely ways to interpret any given harmony while allowing
the analyst to weigh the impact of contextual factors.
Preference Rule 1: Prefer interpretations in which the referential just intervals
correspond as closely as possible to the actual intonation of the music that is,
tone representations which require the least retuning from the heard intervals to
the referential just intervals.

The first preference rule is based on the commonsense principle that our tone
representation of a heard pitch set the just-intonation proportion which lends
each pitch a harmonic meaning in relation to a root should match the heard set
as closely as possible. Although it is simple to determine the closest just-interval
representation for a dyad by referring to a chart of just-interval sizes, larger groups
of pitches can be more difficult to match to a just-intonation interpretation.With
the mathematical tools outlined by Clifton Callender,11 it is possible to quantify
how much retuning is required to map a given pitch-class set onto a target
just-intonation pitch-class set by calculating the Cartesian distance between the
two sets. Using basic calculus, we can find the transposition of the just-intonation
set which minimises this distance that is, the transposition which results in the
least total retuning. In this study, Ive used a computer program to find the
just-intonation sets which best match any given input set; the number of justintonation sets is theoretically infinite, but Ive limited my tone representations to
sets which do not invoke integers above 33. (The complex intervals created by
higher integers are difficult to comprehend in most musical contexts.) The
application of this preference rule can provide a list of many possible tone
representations of a heard set, each associated with a specific fundamental and a
measurement of the amount of retuning between the set and its representation.
Preference Rule 2: Use the simplest possible interpretation of a pitch collection:
the tone representation with the simplest just intervals between its members.
(Simple intervals have low integers in their frequency ratios when reduced
to lowest terms.) The presence of the fundamental (or one of its octave transpositions) tends to considerably strengthen the plausibility of a tone representation.

After we have determined several just-intonation sets which closely fit the
input set, we can choose among them by applying the second preference rule,
Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

358

robert hasegawa

selecting the simplest possible tone representation. There are several ways to
quantify the relative simplicity of a given tone representation of a pitch collection. One is to choose the representation with lower partial numbers: given two
tone representations of the set C5:D5:E5:F5 as F0(24:27:30:34), or as
D2(7:8:9:10) we can easily recognise the greater simplicity of the second
representation by its lower partial numbers. In comparing tone representations
for the same set, this is equivalent to choosing the tone representation with a
higher virtual pitch: D2 is higher than F0. (The use of virtual pitch as a guide to
the relative consonance of a pitch set is common among spectral composers.12)
As a general rule, this criterion is useful, but it ignores the question of factorability: in our comparison of sets, we should also seek the representation with the
simplest just intervals between its members. This fits with our intuition that the
tone representation with partial classes 8:10:12:17 should be simpler than
7:9:11:17 despite the higher virtual pitch of the second list.
Clarence Barlow has proposed a measure of harmonicity, determined not
only by the absolute size of the numbers in an intervals ratio when reduced to
simplest terms, but also the divisibility of those numbers in other words, their
prime limit.13 For example, although the intervals 25:27 and 23:29 are quite
similar in the size of their constituent integers, 25:27 is easier to comprehend
because it can be broken down into simpler intervallic steps. Both 25 and 27 are
products of simpler primes, 5 and 3, while 23 and 29 are prime and cannot be
simplified. For sets larger than dyads, Barlow sums the harmonicities of all the
intervals between set members.
Another metric for simplicity based on factorability is harmonic distance, as
explicated by James Tenney. Tenneys theory is based on a theory of harmonic
space, a multidimensional extension of the Riemannian Tonnetz, with each axis
representing a different prime factor.The distance between any two points on the
lattice is calculated by the sum of all the steps in between the points; however,
steps along the low prime-number axes are considered shorter than those along
the axes of the higher primes. The axes are weighted by their logarithms base 2:
thus, a step on the 2 axis is a harmonic distance of 1, a step on the 3 axis is a
harmonic distance of log23, or 1.58, and so on. Steps along each axis can be
summed for composite intervals: thus the perfect fifth, 2:3, can be seen as a
combination of one step on the 3 axis and one on the 2 axis, a distance of 2.58.
Like Barlow,Tenney calculates the simplicity of larger pitch sets by summing the
intervals between each member:
[Y]ou could go through a piece and say, Alright, weve heard in the beginning of
the piece two pitches.You take the simplest ratio representation of that interval
tempered. Now we hear the third pitch. What specific, rational intonation for that
approximate pitch will give us the simplest configuration in harmonic space, the
most compact configuration in harmonic space? Lets call it that.14

Barlows and Tenneys metrics have slightly different biases with respect to the
means by which they weigh interval simplicity but tend nevertheless to give
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

359

roughly similar results. In general, Barlows method gives a greater weight to

interval simplicity: the presence of an interval made up only of multiples of three
and two contributes significantly (perhaps too significantly) to his overall harmonicity score.These simple ratios also contribute to shorter harmonic distances
in Tenneys calculations of harmonic distance; but the curve drops less sharply in
his case, meaning that the inclusion of higher primes has a less drastic effect on
the overall simplicity. Tenneys metric is more liberal about including higher
prime numbers. A choice between the two metrics might be determined by the
repertoire being examined. By considering the factorability of interval, both
Barlows and Tenneys metrics yield more intuitive results than a measurement
of simplicity by virtual pitch alone.
Preference Rule 3: Use the smallest possible number of fundamentals; invoke
multiple fundamentals only if they yield a significantly simpler interpretation than
is possible with a single fundamental.

If no just-intonation set fits the input set reasonably well, we can turn to
preference rule 3 and describe the input set as the combination of just-intonation
sets on two or more fundamentals; this division of problematic pitch sets into
simpler entities has precursors in Rameaus dual generator derivation of the
minor triad in the Dmonstration du principe de lharmonie and Hermann Erpfs
idea of Mehrklnge.15 This preference rule reflects how perceptual templates are
used in auditory scene analysis to sort partials into smaller, overtone-based sets.
As the psychoacoustician Albert Bregman notes, we seem to apply a sceneanalysis mechanism that is trying to group the partials into families of harmonics
that are each based on a common fundamental. If the right relations hold
between an ensemble of partials, they will be grouped into a single higher-order
organization (Bregman 1990, p. 507).
A tone representation analysis must continually balance the conflicting
demands of the three preference rules while also taking into account other
contextual aspects of the musical surface. The flexibility of the preference rules
allows musically sensitive readings which can be tested aurally. In choosing one
tone representation over another, we are not dealing with abstractions; these
choices have something concrete to say about our musical understanding of each
pitch in the collection and its relationship to all of the others. When we perceive
a diminished fifth as representing the ratio 8:11 instead of 5:7, it has different
tonal implications and, as Riemann notes, an entirely different expressive value,
character and content (Riemann 1992, p. 86). A change in the understood root
changes the meaning of each of the chord members: a pitch that is relatively
stable in one reading can become exotic and harmonically distant in another.
Even if we do not entirely agree on the precise tone representation for a given
pitch set, the terminology introduced here offers a way of discussing what we
hear asserting one tone representation over another is a meaningful, and, above
all, a musical activity. We can continually test our analyses by playing potential
Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

360

robert hasegawa

roots under a harmony or by experimentally adding pitches to see how they

strengthen or weaken our hypotheses.
An Example of Tone Representation: Schoenberg, Op. 11 No. 2
Tone representation offers a useful alternative to other analytical methods currently used for works from the atonal repertoire. Unlike pitch-class set analysis,
which focuses on motivic relationships between chords, tone representation allows
us to closely examine the tensions within a single harmony in a way which is
sensitive to vertical spacing and to the delicate balance of different tonal implications. This sensitivity is particularly valuable for the music of the twentieth
century, in which striking individual sonorities are such an important feature.
Rather than attempting to illustrate an organic coherence through the repetition
of identical harmonic motives, we can discuss changing colour and degrees of
harmonic rootedness.We do not need to compare these chords to one another to
get at their internal tensions and qualities, since we can refer to a consistent
interpretative strategy based on the overtone series instead.This approach allows
us to discuss postcommon practice harmony from a phenomenological rather
than an organicist standpoint.We can observe the application of tone representation in the analysis of a well-known and often-discussed passage by Schoenberg.
Ex. 3a reproduces the Chorale from Schoenbergs Piano Piece, Op. 11 No. 2.
Because the tetrachords of the Chorale fall into different set classes, this passage
poses a challenge to standard pitch-class set analysis. David Lewin has discussed
this passage at length, using Klumpenhouwer networks to answer the question:
[I]s there some way in which we can sense the harmonic field of the phrase as
unified, rather than diverse? (Lewin 1994, p. 79). He sets out an agenda for
analysis: to relate the tetrachords of different pc-set classes into a unified overall
view which includes the five- and six-note sets that appear at the end of the
phrase.16
Unlike pitch-class set analysis, which tends to emphasise motivic relationships
between sonorities, tone representation makes it possible to discuss the competing
root implications and inner harmonic tensions within a single chord.17 Ex. 3b
lists several plausible tone representations of each chord of the Chorale, with the
most convincing representation appearing in boldface type. (Occasionally two
representations seem equally convincing; in such cases, both are printed in
boldface.) In this table, I have listed only the best matches based on closeness
of fit and simplicity of intervals from the list of possibilities produced by
computer calculation. A typical situation can be seen in my analysis of the first
chord, BEFA. The tone representation which entails the least retuning is
A-3(17:24:27:32); according to this representation, B is heard as the seventeenth partial of a notional low A fundamental (lowered by 3 cents from equal
temperament), E is heard as the 24th partial, and so on. While the intonational
fit is very precise with a distance of only 5.41 between the heard chord and its
just-intonation tone representation the intervallic relationships between the
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

361

Ex. 3a Schoenberg, Op. 11 No. 2, Chorale, set classes labelled

1

(0256) (0267) (0157)

4Z15 416 416

(0256) (0267) (0347)

4Z15 416 417

10

11

10

11

(0356)
413

(0346)
412

(0148)
419

(01458)
521

(023578)
6Z25

12

13

12
8
cresc.
poco string.

12
8

Ex. 3b Table of plausible tone representations for each chord in Ex. 4a (with the
most convincing tone representations shown in boldface)

Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

362

robert hasegawa

pitches are complex and obscure. For example, were asked to hear the interval
from B to F as the exotic interval 17:27, though we would intuitively prefer a
simpler interpretation such as 5:8, a just minor sixth. We can find a simpler
interpretation of the whole tetrachord by accepting a slightly greater mistuning
between the heard set and its just-intonation representation. The tone representation F+12(10:14:16:19) provides the most convincing compromise between
intonational accuracy and simplicity of interval ratio, and the inclusion of the
fourth octave of the fundamental (16) further strengthens its appeal.
An extended discussion of this excerpt is impractical here, but a few general
observations will illustrate how the theory of tone representation might contribute
to an analytical reading.When we look at the most likely roots for each chord in the
passage, we see the frequent repetition of just a few pitch classes (allowing for some
variability of tuning): F, F/G and G. These three pitches account for nine of the
eleven chords of the Chorale. G is the root of chord 3, at the end of the first gesture,
as well as of chord 8 and the cadential chord 11, although the chords differ in
cardinality and set class. An upwards progression by semitone from one fundamental bass pitch class to the next recurs frequently first as F to G from chords
2 to 3, then as F to F to G in chords 6, 7 and 8.This fundamental bass progression
is repeated in chords 911 as F to G to G, even though the pitch content of the
chords is different. The fundamental bass progression cuts across the phrase
structure in an interesting way, inviting the listener to group chords 6, 7 and 8
across the notated phrase boundary between chords 6 and 7. This is one way we
could make sense of the crescendo beginning below chord 6: the increase in
intensity accompanies the beginning of the ascent in the fundamental bass.
In a different hearing of the passage, we can hear the boundaries between the
phrases as revoicings of the harmony over a repeated fundamental bass; thus,
chords 3 and 4 can be heard as rooted on E, while chords 6 and 7 share a root
of A. Note that this reading interprets the roots of these chords differently than
the previous analysis the divergent interpretations reflect two possible ways of
hearing the structure of the passage, which is rich and complex enough to
support a range of competing analyses.18
By invoking tone representation, we are no longer treating this music as atonal,
but rather as exhibiting a kind of extended tonality. As weve seen, the ratio model
of interval focuses our attention on very different aspects of pitch structure than
those illuminated by the distance-based models of pitch-class set analysis. I do not
deny the utility of this and other such atonal theories for this repertoire, but they
are designed to describe different kinds of relationships to those Im interested in
exploring here. In a sense, no music is truly atonal; there is music for which atonal
relationships are the basis of convincing analytical interpretations, but this does
not rule out the possibility of other tonal or quasi-tonal readings. If we do not insist
on forcing musical works into the framework of just one theory at a time, the two
methods could be usefully combined atonal theorys emphasis on motivic
transformation could be complemented by tone representations attention to
vertical spacing, sonic colour and implied roots.
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

363

Tone Representation in Vortex Temporum II

As noted above, the techniques and plans which Grisey used to construct Vortex
Temporum have been described in detail in studies based on the composers
sketches for the work.The description of a compositional process, however, is not
necessarily a good description of a pieces aural and musical effect. Even though
many spectral techniques take acoustic and psychoacoustic facts as their starting
point, there is often no clear, unambiguous relationship between such compositional techniques and their audible musical results. Instead of analysing the
music by reconstructing Griseys derivation of the harmonies, we can use the
theory of tone representation to approach the music from our own harmonic
intuitions.19 This is a turn from one sense of the natural to the other: from an
external idea of the natural, based on how Griseys harmonies draw on natural
models, to an internal one, based on how we intuitively that is, naturally make
sense of complex sonorities. We can describe Griseys harmonies with reference
not to their source, but rather to our own aural experiences. Examining the work
through the lens of tone representation can offer new insights into its harmonic
relationships as actually heard: tone representation can function as a listening
grammar for complex microtonal sonorities.20
We will concentrate here on the harmonic analysis of several excerpts from the
second movement of Vortex Temporum. Ex. 4 is an outline of Griseys deployment
of spectra throughout the movement.21 Each section is based on a selection of
notes from a single spectrum, either harmonic, stretched or compressed; in the
figure, the partial number of each pitch in its respective spectrum appears above
each note. Over the nine sections of the movement, the pitch class of the nominal
fundamental descends chromatically from B to E, with the exception of the
central spectrum on C although, owing to the changing spectrum types, this
does not create a clearly audible sense of downward transposition. The texture
remains consistent throughout the movement. The piano plays on every beat,
gradually cycling downwards through the available pitches; stems indicate the
pitches repeated by the piano on every beat.The descending pitches of the piano
imitate the well-known aural illusion of the endlessly descending Shepard tone
the descent seems continuous because as the entire complex of partials drifts
downwards, new high partials gradually fade in from silence and the lowest
partials drop out.22 Because the piano cannot play microtones (with the exception of its four retuned strings), it often rounds off partials to the nearest available
semitone; these approximations are shown as letter names below the pitch that
they replace.
Ex. 5a illustrates how the stretched spectrum on B in Ex. 2 is presented in
section II of the movement (rehearsal numbers 47). All the pitches of the spectrum are rounded off to the nearest quarter tone (or, for the piano, the nearest
semitone). The pianist cycles continuously downwards through the boxed notes
on the bottom stave, rearticulating the stemmed notes on every beat. The other
instruments of the ensemble flute, clarinet, violin, viola and cello play sustained
Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

robert hasegawa

364

Ex. 4 Deployment of spectra in Vortex Temporum, ii

2

13

18

12

13

14

15 16

17

17

18

19

20

I: B compressed
D

10

10

11

12

13

11

18

22

27

21

22

23

II: B stretched
F
7

A
14

15

37

41

26

28

29

32

27

28

32

35

24

25

26

28

32

21

22

23

25

21

23

25

32

33

37

III: A harmonic
F
5

10

11

D
12

13

14

15

16

17

18

19

20

35

38

41

IV: G compressed
D

10

10

C
11

12

13

14

15

16

17

18

19

10

11

12

13

14

15

16

11

12

13

14

15

16

17

18

19

20

21

22

17

18

19

20

21

22

27

V: C harmonic
F
20

VI: G stretched
E
9

VII: F harmonic
C
8

11

13

15

17

19

20

22

25

26

10

11

14

28

30

32

C
33

38

VIII: F compressed
A
3

D
5

IX: E stretched
C

pitches, shown in the upper two staves of Ex. 5a with bars indicating their
duration.
Since no one fundamental offers a convincing tone representation for all of the
pitches in the excerpt, Ex. 5b identifies four plausible tone representations, each
describing different subsets of the complete harmony. Our choice of one tone
representation over another depends on the musical context and changes over the
course of the passage. Often, not all pitches are equally important to our decision
we give more weight to repeated and held notes.The sonority rooted on B seems
especially convincing at the beginning of the excerpt, when the B is present as a
held note in the flute and violin. However, the B representation cannot account for
some salient pitches in the texture: the held G1/4 and the repeated C1/4 and D1/4.
As the B fades out, our attention turns to the tone representations which give these
salient pitches more weight, and we experience the sonority as a combination of
harmonies with roots on E and C1/4 (the low repeated notes in the piano). At the
end of the excerpt, the upper-register held pitches G, D1/4, F and B strongly
imply the tone representation D1/4 (11:16:21:26); the pianos B, B and D1/4 also
fit into this interpretation, as partials 13, 14 and 17.
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Music Analysis, 28/ii-iii (2009)

5

6 7 15 16 17 18 24 30)

distance: 35.90

B +6(5
distance: 45.20

E+7 (8 13 15 17 18 20 24 25 27 28 29 31 38 48)

distance: 32.00

7 8 11 13 18 19 23 29) D+47(11 13 14 16 17 21 26)

distance: 37.80

C-45(6

Ex. 5b Several plausible tone representations for the excerpt reproduced in Ex. 5a

The piano cycles continuously through the boxed notes; stemmed notes are repeated on every beat.

glissando...

Sustained notes are played by the winds and strings.

Ex. 5a Harmonic transcription of Vortex Temporum, ii, section II (rehearsal numbers 47)

Ge rard Grisey and the Nature of Harmony

365

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

366

robert hasegawa

The complexity and multivalence of this analysis reflects the aural richness of
the harmony. I would argue that the competing pull of different tone representations is one of the things which keep our attention engaged throughout this
minute-long, largely static passage. When we view the chord through the lens of
tone representation, its derivation from an exaggeratedly stretched spectrum is
irrelevant we focus instead on the complex ways in which the harmony plays on
our aural intuitions. The choice between the two types of nature Ive discussed
here, one external and one internal, illustrates a broader decision between two
models of analysis for Griseys music. One is essentially formalist, based on
sketch study and the re-creation of the composers material and ideas, while the
other is essentially phenomenological and pragmatic the analysts subjective
experience of the piece is taken as the essential explicandum. The ambiguity of
the pragmatic approach is not a weakness, but rather a strength: by allowing such
ambiguity, we recognise the richness of musical listening and avoid flattening our
experience, forcing it into a simplistic formal mould.
Tone representation is a valuable tool of this pragmatic approach to analysis:
an approach based on taking aural experience seriously, instead of formalistic
abstractions or speculations on the composers intent. By focusing our attention
on sonic quality and rootedness, tone representation suggests new readings of
music in a variety of styles, from Schoenbergs atonal music to Griseys spectral
works. The theory makes it possible to put into words some of the most elusive
aspects of our experience of complex harmonies and by emphasising listening
instead of mathematical or formal abstraction, it offers a promising alternative to
existing analytical techniques.
NOTES
Copyright clearance for musical examples was obtained from the following sources:
Grisey, Partiels (1975), copyright 1976 by G. Ricordi & C.; Schoenberg, Op. 11 No. 2,
Copyright 1910 by Universal Edition.
1. Baillet (2000) is the most complete sketch-based study of Griseys work; see also
Herv (2001), a monograph on Griseys Vortex Temporum.
2. This discussion of Partiels is indebted to Fineberg (2000), pp. 11518; see also Rose
(1996), pp. 811. It would be an oversimplification to associate the spectralists
solely with the technique of instrumental synthesis, but particularly in the early days
of the movement this was an essential and frequently used technique. Later developments in the spectralists technique added a variety of effects and transformations, including many (such as frequency modulation) based on the tools of the
electronic music studio.
3. See Christensen (1993), pp. 13368.
4. See Baillet (2000), p. 217. Griseys equation results in a curve shaped differently
from the natural stretching of the piano spectrum, which is described by the
equation fn = nf0(1 + Bn2)1/2 in Fletcher, Blackham and Stratton (1962), p. 756.
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

367

5. Grisey (1984) includes the composers discussion of the concept of the liminal.
See also Stahnke (1999).
6. The composer Roger Reynolds (1993, pp. 2823) has questioned the relation
between such spectral models and their musical realisation, stating that it is clearly
absurd that instrumental synthesis could possibly result in an orchestrated product
that bears anything other than an incoherent and metaphoric relationship to the
supposed model. Reynolds is at pains to emphasise that his characterisation of the
relationship between spectral structures and their acoustical models as incoherent
and metaphoric is not an aesthetic judgment so much as a theoretical one; the music
may well succeed artistically despite the incoherence of the compositional technique. Reynoldss observations raise an important question: if the models underlying the music do not have a perceptible relationship to the musical surface, then
can an analysis based on a reconstruction of compositional procedure tell us
anything about how a work of music is heard and understood? If the compositional
model is not clearly reflected in the musical surface, an analysis which proceeds
instead from a perceptual standpoint is likely to tell us more about the experience
of hearing a work.
7. See Hartmann (1998), p. 135.The distorted spectra Grisey uses in Vortex Temporum
lack the unique organisational potential of harmonic spectra. One of the unique
qualities of sounds with harmonic (or near-harmonic) spectra is that they are easily
resolved into separate streams when presented at the same time. But, as Albert
Bregman notes (1990, p. 238), when two stretched series of partials are sounded at
the same time, you do not hear only two distinct sounds as you do when listening to
two harmonic sounds. The composite of two inharmonic spectra is not easily
resolved into two distinct sources and often results in a vague or ambiguous sense of
pitch, or sometimes the chimerical perception of more than two illusory sound
sources.
8. See Tenney and Dennehy (2008), p. 87.
9. The use of the overtone series to explain complex harmonies was a frequent trope
in twentieth-century theoretical writing. Arnold Schoenberg, for example, suggested that the future of musical evolution would rest on the growing ability of
the analyzing ear to familiarize itself with the remote overtones (1978, p. 21).
Similar ideas are found in the work of authors from Paul Hindemith to Henry
Cowell.
10. Preference rules make their first appearance in music theory in Lerdahl and Jackendoff (1983). The rules proposed here appear in a somewhat different form in
Hasegawa (2006).
11. See Callender (2004), pp. 2631.The Cartesian distance between sets is the square
root of the sum of the squares of the differences between each pitch and its mapping
(measured here in cents). Following Callender, this result is multiplied by a scaling
factor of n n 1 (where n is the number of pitches in the set). This brings the
distance into conformity with our intuitions about pitch distance, so that, for
example, the minimum distance between sets {0, 4, 7, 10} and {0, 4, 7, 11} is equal
to 1 and not 3 2 ; see Callender (2004), p. 29.
12. Fineberg (2000), pp. 1248. Spectral composers have adopted Ernst Terhardts
virtual pitch algorithm, designed to predict the most likely assignment of overall

Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

robert hasegawa

368

pitch to a collection of partials; see Terhardt (1979). Terhardts algorithm seeks

matches (or near matches) between the subharmonics of a given set of frequencies;
a match of subharmonics means that both frequencies can be heard as overtones of
a fundamental pitch at the frequency of the match. Like the theory of pitch
representation advanced here, Terhardts algorithm for finding virtual pitch frequently locates several possibilities for the virtual pitch of a given set of components. Terhardt invokes criteria similar to my three preference rules to choose
between competing interpretations (Terhardt 1979, p. 169).
13. See Barlow (1987), pp. 4455.
14. See Belet (1987), p. 462. Tenneys theory that we prefer simple explanations is
closely related to gestalt psychologys principle of Prgnanz (conciseness).
15. See Rameau (1750) and, for example, Erpf (1969).
16. See Lewin (1994), p. 86.
17. Visl (2002) explores overtone-based harmonies in works by Webern, Berg,
Scriabin and Debussy.Visl draws extensively on Richard Parncutts research into
psychoacoustics and harmony (Parncutt 1988 and 1989). Another application of
the overtone series to the analysis of atonal harmonies is Delige (2005).
18. The analysis offered here might seem to imply that Schoenbergs music is out of
tune and needs to be corrected to just intonation. This is not my intent, and
Schoenberg made his preference for equal temperament very clear. Rather, Id
argue that equal temperament allows the careful balancing of ambiguities between
several tone representations, an ambiguity which seems integral to the aesthetic of
early atonality. A similar conclusion is drawn by Gary Don in his research on
overtone series chords in the music of Debussy: he concludes that Debussy was
content to incorporate the overtone series into his music through the lens of equal
temperament, thus suggesting a particular sonority, without requiring a literal (i.e.,
just intonation) realization of those sonorities (Don 2001, p. 69).
19. Stahnke (1999) and (2000) take a similar approach to the analysis of chords from
the first part of Vortex Temporum.
20. The term listening grammar was coined in Lerdahl (1988). Lerdahl makes a distinction between listening grammars (which a listener uses to make sense of a musical
work) and compositional grammars (which a composer uses to create a work).
21. For different viewpoints on the construction of this movement, see Baillet (2000),
p. 224, and Herv (2001), pp. 567.
22. See Shepard (1964). An example of a continuously descending Shepard tone can be
found on the website of the Acoustical Society of America: http://asa.aip.org/
demo27.html (accessed 19 July 2008).

REFERENCES
Baillet, Jrme, 2000: Grard Grisey: Fondements dune criture (Paris:
LItinraire).
Barlow, Clarence, 1987: Two Essays on Theory, Computer Music Journal, 11/i,
pp. 4460.
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

369

Belet, Brian, 1987: An Interview with James Tenney, Perspectives of New Music,
25/iii, pp. 45966.
Bregman, Albert S., 1990: Auditory Scene Analysis: the Perceptual Organization of
Sound (Cambridge, MA: MIT Press).
Callender, Clifton, 2004: Continuous Transformations, Music Theory Online,
10/iii.
Christensen, Thomas, 1993: Rameau and Musical Thought in the Enlightenment
(Cambridge: Cambridge University Press).
Delige, Clestin, 2005: Lharmonie atonale: de lensemble lchelle, in
Sources et ressources danalyses musicales: Journal dune dmarche (Sprimont:
Mardaga), pp. 387411.
Don, Gary, 2001: Brilliant Colors Provocatively Mixed: Overtone Structures in
the Music of Debussy, Music Theory Spectrum, 23/i, pp. 6173.
Erpf, Hermann, 1969: Studien zur Harmonie- und Klangtechnik der neueren Musik,
2nd edn (Leipzig: Breitkopf und Hrtel).
Fineberg, Joshua, 2000: Musical Examples, Contemporary Music Review, 19/ii,
pp. 11534.
Fletcher, Harvey, Blackham, E. Donnell and Stratton, Richard, 1962: Quality of
Piano Tones, Journal of the Acoustical Society of America, 34/vi, 74961.
Grisey, Grard, 1984: La musique: le devenir des sons, Darmstdter Beitrge zur
Neuen Musik, 19, pp. 1623.
Hartmann, William, 1998: Signals, Sound, and Sensation (New York: Springer
Science+Business Media).
Hasegawa, Robert, 2006: Tone Representation and Just Intervals in Contemporary Music, Contemporary Music Review, 25/iii, pp. 26381.
Herv, Jean-Luc, 2001: Dans le vertige de la dure: Vortex Temporum de Grard
Grisey (Paris: LItinraire).
Lerdahl, Fred, 1988: Cognitive Constraints on Compositional Systems, in John
Sloboda (ed.), Generative Processes in Music: the Psychology of Performance,
Improvisation, and Composition (Oxford: Oxford University Press), pp. 231
59.
Lerdahl, Fred and Jackendoff, Ray, 1983: A Generative Theory of Tonal Music
(Cambridge, MA: MIT Press).
Lewin, David, 1994: A Tutorial on Klumpenhouwer Networks, Using the
Chorale in Schoenbergs Opus 11, No. 2, Journal of Music Theory, 38/i, pp.
79101.
Parncutt, Richard, 1988: Revision of Terhardts Psychoacoustical Model of the
Root(s) of a Musical Chord, Music Perception, 6, pp. 6594.
______, 1989: Harmony: a Psychoacoustical Approach (Berlin and New York:
Springer-Verlag).
Rameau, Jean-Philippe, 1750: Dmonstration du principe de lharmonie (Paris:
Chez Durand, Pissot).
Reynolds, Roger, 1993: Seeking Centers, Perspectives of New Music, 32/ii, pp.
27291.
Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd

370

robert hasegawa

Riemann, Hugo, 1992: Ideas for a Study On the Imagination of Tone , trans.
Robert Wason and Elizabeth West Marvin, Journal of Music Theory, 36/i, pp.
81117.
Rose, Franois, 1996: Introduction to the Pitch Organization of French Spectral
Music, Perspectives of New Music, 34/ii, pp. 639.
Schoenberg, Arnold, 1978: Theory of Harmony, trans. Roy E. Carter (Berkeley,
CA: University of California Press).
Shepard, Roger, 1964: Circularity in Judgments of Relative Pitch, Journal of the
Acoustical Society of America, 36, pp. 234653.
Stahnke, Manfred, 1999: Die Schwelle des Hrens: Liminales Denken in
Vortex Temporum von Gerard Grisey, Osterreichische Musikzeitschrift, 54/vi,
pp. 2130.
______, 2000. Zwei Blumen der reinen Stimmung im 20. Jahrhundert: Harry
Partch und Grard Grisey, Hamburger Jahrbuch fr Musikwissenschaft, 17,
pp. 36989.
Tenney, James, 2001: The Several Dimensions of Pitch, in Clarence Barlow
(ed.), The Ratio Book: a Documentation of the Ratio Symposium, Royal Conservatory, The Hague, 1416 December 1992 (Cologne: Feedback Studio
Verlag), pp. 10215.
Tenney, James and Dennehy, Donnacha, 2008: Interview with James Tenney,
Contemporary Music Review, 27/i, pp. 7989.
Terhardt, Ernst, 1979: Calculating Virtual Pitch, Hearing Research, 1, pp. 155
82.
Visl, Olli, 2002: Prolongation of Harmonies Related to the Overtone Series
in Early-Post-Tonal Music, Journal of Music Theory, 46/iii, pp. 20783.

ABSTRACT
Grard Grisey (19461998) was a founder of the influential spectral movement. Reacting against the abstractions of serialism, spectral composers derived
their musical material from the physics of sound and the mechanisms of aural
perception. The present study explores the tensions between Griseys natural
sonic models and their alterations and distortions in his music. One common
spectral technique is instrumental synthesis the scoring for instrumental
ensemble of the partials of a complex natural sound. Instrumental synthesis
creates a musical effect which is neither atonal nor tonal in the traditional sense
rather, we can best understand this music as exhibiting an extended tonality
based on the upper overtones of the harmonic series. The analysis of this
extended tonality calls for new theoretical tools which can account for the
complex harmonic relationships between high overtones.
I propose a modification of Hugo Riemanns theory of Tonvorstellung (tone
representation): the idea that, given a collection of pitches, we will understand
them as connected by the simplest possible just intervals. As a model of harmonic
2011 The Author.
Music Analysis 2011 Blackwell Publishing Ltd

Music Analysis, 28/ii-iii (2009)

Ge rard Grisey and the Nature of Harmony

371

meaning based on our processes of auditory cognition, tone representation can

illuminate the way we hear and understand harmony in a wide variety of works.
The theory is demonstrated in a discussion of Sehoenbergs Piano Piece, Op. 11
No. 2, then applied in an analysis of Griseys Vortex Temporum (19946). Sketches
for the piece indicate Griseys use of distorted stretched and compressed
spectra in addition to the familiar harmonic series. Applying my theory of tone
representation makes possible a sensitive description of the aural effect of such
distorted spectra, rather than the interpretations we frequently find of these
sonorities which contradict their natural origins in Griseys sketches.

Music Analysis, 28/ii-iii (2009)

2011 The Author.

Music Analysis 2011 Blackwell Publishing Ltd