Contemporary Music Review

Vol. 25, No. 3, June 2006, pp. 263 – 281

Tone Representation and Just Intervals

in Contemporary Music
Robert Hasegawa

Simple integer ratios between frequencies (just intervals) have historically been invoked
as the basis of musical harmony. Hugo Riemann’s theory of Tonvorstellung proposes
that we mentally ‘represent’ pitches as members of justly tuned triads. Broadening the
scope of Riemann’s theory to include more complex interval ratios, James Tenney argues
that we understand heard intervals as approximations, within a tolerance range,
of referential just intervals. This essay develops the idea of tone representation as an
analytical tool for contemporary music. An excerpt from György Ligeti’s Melodien
(1971) is analyzed as a succession of just intervals pivoting above a shifting fundamental.
A close reading of the opening of Gérard Grisey’s Vortex Temporum (1994 – 1996)
explores tone representations of complex microtonal harmonies, and examines the gap
between compositional techniques and aural experience in this ‘spectral’ work.

Keywords: Just Intonation; Hugo Riemann; James Tenney; György Ligeti; Gérard Grisey;
Spectral Music

I
Throughout the history of music theory, just intervals have been defined by simple
integer ratios—measured by ancient Greek theorists as ratios between string lengths
on a monochord, and later (beginning in the sixteenth century) as ratios between
vibration frequencies. Simple ratios such as 2:3 (the fifth) or 5:6 (the minor third)
produce the pure intervals which are the basic building blocks of tonal music.
However, much of the theoretical and analytical writing on twentieth-century music
has focused on a different concept of interval: interval as the distance between two
pitches measured in equal-temperament semitones. The important distinction
between ratio and distance models of interval has been discussed in detail by Ben
Johnston (1964). Reviving the idea of intervals as frequency ratios can help illuminate
our experience of intervals as qualitative sonic phenomena, describing the particular
acoustic qualities of each interval which are key to our understanding of their musical

ISSN 0749-4467 (print)/ISSN 1477-2256 (online) ª 2006 Taylor & Francis

DOI: 10.1080/07494460600726529
264 R. Hasegawa
role. This ratio-based approach to interval can be useful for analyzing harmony in
many contemporary musical works.
All just intervals can be expressed as integer ratios between the frequencies of two
pitches. For example, the interval of the just major third C4/E4 can be expressed as
the frequency ratio 4:5—we could also say that these pitches are in the same
relationship as the fourth and fifth overtones of a low fundamental C2. Just interval
ratios can be converted logarithmically into distance intervals measured in cents
(hundredths of an equal-temperament semitone); the interval 4:5 is equivalent to a
distance of approximately 386 cents, or 3.86 semitones.
Renaissance theorists such as Gioseffo Zarlino defined musical consonances as the
intervals with ratios made only with the integers one through six and eight; these
ratios describe the octave, fifth, fourth, and major and minor thirds and sixths in just
intonation (along with their octave equivalents, such as the twelfth and double
octave). By admitting higher multiples of two, three and five, one can create the just
intervals which have historically been considered dissonances, such as the whole tone
(8:9) or the semitone (15:16). The inclusion of higher prime numbers (such as 7,
11 and 13) and their multiples leads to intervals which fall between the customary
semitone divisions of the octave, such as the flat natural seventh (4:7, or 969 cents) or
the flat tritone (8:11, or 551 cents)—this is what Ben Johnston calls ‘extended just
intonation’.
Some twentieth-century composers have turned to these microtonal just intervals
as the basis for new musical systems: extended just intonation has been explored by
Harry Partch and Ben Johnston, and the spectralist composers Gérard Grisey and
Tristan Murail have used various types of overtone-based harmonies. The interest in
just interval and the closely related phenomenon of overtones is tied to a renewed
interest in harmony and ‘sound as sound’, in reaction to the abstract distance-based
conception of interval in serialism and motivic atonality. Adopting just interval as an
analytical concept can provide a way to discuss this qualitative aspect of musical
experience.
My aim is not to disparage the concept of interval as distance—its enormous
power and practicality of application are undeniable. Rather, I propose that having
recognized that interval can be understood in different ways, we may find
applications where the less familiar concept of intervals as frequency ratios offers
new analytical insights which would not come to light with a distance-based
intervallic approach.
To this end, I will discuss excerpts from two major works from the last half of
the twentieth century—György Ligeti’s Melodien (1971) and Gérard Grisey’s Vortex
Temporum (1994 – 1996). In Ligeti’s work, we shall see how the pitches of a chord
progression can be understood as approximations, to the nearest semitone, of
the high harmonics of a shifting fundamental tone, or ‘root’. In Grisey’s Vortex
Temporum, we will explore the ambiguity which arises from his use of various
distortions of an overtone series, and how Grisey’s compositional techniques are
often at odds with the aural sense we make of the complex resulting harmonies.
 Contemporary Music Review 265

II
If just intervals are significant to our musical understanding, how do they apply to
music written in equal temperaments, such as the Ligeti and Grisey works? We must
accept that equal-temperament approximations of just intervals still carry some of
the same musical qualities as their just intonation counterparts—in other words,
that we can recognize the harmonic implications of just intervals even when they are
not perfectly tuned. The relationship between just intervals and their mistuned
counterparts has been addressed in compatible ways by two very different music
theorists: the early-twentieth-century German theorist Hugo Riemann and the
contemporary American composer-theorist James Tenney.

Riemann
In his idealist approach to music theory, Riemann asserts that we are not merely
passive recipients of musical sounds, but active interpreters of those sounds
into logical structures; the mental representation of musical relationships is
more important than their actual manifestation as sound. Every musical tone is
‘represented’ or ‘imagined’ as part of a justly tuned major or minor triad. (Riemann’s
work focuses on the primarily triadic music of the tonal tradition.) The harmonic
meaning of each tone depends on its context—whether it is, for example, the third of
a minor triad or the fifth of a major triad. A single, isolated tone may pose problems
of ambiguity—but once we are familiar with the piece of music in which it occurs, it
takes on a character depending on its harmonic representation: ‘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’ (Riemann, 1992, p. 86).
Riemann labels the members of a major triad with the arabic numerals 1, 3 and 5,
and the members of a minor triad with the roman numerals I, III and V. In a minor
triad, intervals are labeled from the fifth of the chord downward rather than from the
root upward—this is in keeping with Riemann’s dualist conception of the minor
triad as an upside-down version of the major triad. All of these triadic relationships
are governed by just intonation: thus, the equal-tempered fifths and thirds of chords
played on a piano are ‘imagined’ as pure, just intonation intervals in the listener’s
mind. Heard intervals are transformed into just intervals in the mind through the
process of Tonvorstellung, which I will translate here as ‘tone representation’. Tone
representation also provides a harmonic meaning for each pitch, associating it with
its harmonic root by a just interval.
Riemann proposes a general principle governing the way that our minds
understand tones harmonically—we prefer the simplest possible interpretation
consistent with the music. ‘This Principle of the Greatest Possible Economy for the
Musical Imagination moves directly toward the rejection of more complicated
structures, where other meanings suggest themselves that weigh less heavily on
266 R. Hasegawa
the powers of interpretation’ (ibid., p. 88 [emphasis in original]). Thus, given a
collection of pitches, we will understand them as connected by the simplest possible
just ratios, even when our ears are confronted by the complex and irrational intervals
of equal temperament: ‘our organ of hearing fortunately is so disposed that absolutely
pure intonation is definitely not a matter of necessity for it’ (ibid., p. 99).

Tenney
For Riemann, the possibilities of tone representation end with the tonal relationships
found within a triad: just intonation thirds and sixths, fourths and fifths. James
Tenney has proposed some similar theoretical ideas, but from a vastly different
aesthetic position—while Riemann was aesthetically conservative, Tenney is a major
creative figure in American experimental music. Like Ben Johnston and Harry Partch,
he expands the concept of just interval to allow more complex integer ratios
(Gilmore, 1995). However, unlike these more strict just-intonation advocates,
Tenney allows what he calls tolerance—‘the idea that there is a certain finite region
around a point on the pitch height axis within which some slight mistuning is
possible without altering the harmonic identity of an interval’ (Tenney, 1992, p. 109).
This notion brings Tenney very close to Riemann’s idea of tone representation, but
with a much more liberal attitude towards which ratios might act as ‘referential’ just
relationships.

Now, 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 . . .
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, 1992, p. 110 [emphasis in original])

Tenney separates the timbral quality of a heard interval from its harmonic sense;
thus, we can imagine two different heard intervals representing the same harmonic
sense, or the same heard interval representing two different harmonic senses. As an
example, an equal-temperament major third (400 cents) played on a piano can
represent a just 4:5 major third of 386 cents. The equal-tempered third is audibly
sharper than the smaller just third (it has a different ‘timbral quality’), but we can still
identify it as projecting the ‘harmonic sense’ of the just interval. However, one might
object that there is a just ratio that much more closely approximates the equal-
temperament third: the 19:24 interval of 404 cents. Why then should we tend
to understand the 400 cent equal-temperament third as 4:5, not 19:24? Tenney’s
explanation is in the same spirit as Riemann’s ‘principle of the greatest possible
 Contemporary Music Review 267
economy’: ‘Given a set of pitches, we will interpret them in the simplest way possible’
(Tenney & Belet, 1987, p. 462). The context in which we hear an interval is thus very
important in determining its harmonic sense. If we hear the equal-tempered third
C/E as part of a larger set of pitches implying a fundamental of A, the context could
lead us to understand it as the complex just ratio A(19:24). If we hear the same
third on its own, though, we are more likely to understand it as the simpler ratio
C(4:5).
Like Riemann, Tenney proposes that our mental representation of harmony is
based on simple frequency ratios, and that we tend to experience slight deviations
from a referential just interval as still representing that interval. This is not to say that
we do not hear the specific sonic quality of, for example, an equal-temperament third
as compared to a just third—we can notice the difference in intonation while still
recognizing the harmonic sense of the referential just third.
Tenney does not specifically define what the tolerance range of a just interval might
be, but he notes that the degree of tolerance tends to ‘vary inversely with the ratio
complexity of the interval’ (Tenney, 1992, p. 110); that is, simple intervals such as
octaves and fifths are more likely to be recognized in spite of mistunings, while
complex relationships, such as the 19:24 major third, are likely to lose their identity if
mistuned by a comparable amount.

Tools and Guidelines

Before exploring the applicability of tone representation in specific works, it will be
useful to establish some basic theoretical tools. Tone representation is based on just
intervals, which can be understood as intervals between the partials of an overtone
series. Figure 1 shows the first to thirty-second partials of C1 rounded off to
quartertones, with the distance above the fundamental in cents (mod 1200) of each
pitch. (Quartertone rounding has been chosen here to match the quartertone
notation of Grisey’s Vortex Temporum; in other analytical contexts, it may be more
practical to use coarser or finer approximations. For example, my analysis of Ligeti’s
Melodien rounds each partial to the nearest semitone.)

Figure 1 Overtone series on C, partials 1 to 32.

268 R. Hasegawa
For the current study, I will limit tone representation to the just intonation
intervals which can be formed using the integers 1 to 21 and their products when
multiplied by powers of 2—that is, the partials 1 to 21 and their octave equivalents.
By analogy to the concept of pitch class, the collection of an integer and its products
when multiplied by powers of 2 could be thought of as a partial class. Thus, we will
admit into our possibilities of tone representation the intervals 13:17 and 17:26 (26 is
twice 13, which falls within our partial class limit of 21), but not 17:25. Using partial
classes ensures that the inversions and octave equivalents of every interval within
the partial class limit of 21 will also fall within that limit; so, in addition to the
15:17 major second of 217 cents, we will include its inversion to a 17:30 minor
seventh and its octave equivalent 15:34 major ninth. The idea of partial class is
derived from Riemann’s own formulation of octave equivalence for triadic tone
representation:

Through the equivalence, or at least the more intimate relationship, of tones that
stand in octave relation, the number of tones belonging to a single harmony is
reduced to three altogether: the prime (and its octaves ¼ [Link], etc.), the fifth
(and its octaves ¼ [Link], etc.), and the third (and its octaves ¼ [Link], etc.), above
as well as below. (Riemann, 1992, p. 98 [emphasis in original])

Why limit tone representation to partial classes of 21 and below? In examining music
which uses quartertone approximations, such as the Grisey work analyzed below,
errors in rounding begin to become a serious problem with the inclusion of higher
partials. Furthermore, with the lower degree of tolerance of the complex intervals
involving partial classes above 21, it is difficult to convey a strong sense of their
identity when they are approximated to a quartertone grid.
A basic notational convention will make the description of tone representations
much simpler. To indicate the tone representation ‘the fourth and fifth partials of an
F fundamental’, I will write F(4:5). The same notation is applicable to single pitches,
e.g. B(17), or to larger collections, e.g. A([Link]).
Figure 2 uses this notation to list all of the intervals (within an octave) which use
the partial classes 1 to 21. They are sorted by their quartertone approximations
(cf. Figure 1); the symmetrical layout of the chart puts each interval opposite its
inverse modulo 12 (for example, 1½ is opposite 10½). An instance of each interval is
shown beginning on the pitch class C. The possible representations of that interval
are then listed: for example, the tritone C/F-sharp can be either B(17:24) or F(12:17).
These tone representations are sorted by the size of the just interval, which is shown
(in cents) beneath each ratio. The rounding off of each partial to the nearest
quartertone leads to some anomalies between an interval’s quartertone notation and
its just intonation size: for example, the interval 20:21 (84 cents) is rounded down to
a quartertone (50 cents), even though the smaller interval 21:22 (81 cents) is rounded
up to a semitone (100 cents). If the existence of these rounding anomalies is kept in
mind, they should not present a major obstacle when looking up possible tone
representations for a given interval.
 Contemporary Music Review 269

Figure 2 Chart of just intervals using the partial classes 1 to 21.

270 R. Hasegawa
This chart will be useful in determining possible tone representations for
a collection of pitches. We can define several guidelines for choosing a tone repre-
sentation for a given collection:

1) Use the smallest possible number of fundamentals; invoke multiple funda-

mentals only if they yield a significantly simpler solution than a single
fundamental can. Collections of pitches which can be related to the same
fundamental will tend to imply that fundamental as a kind of ‘root’—especially
when the partial class of the fundamental (1, 2, 4, 8, etc.) is present
(see Parncutt, 1988 and Väisälä, 2002). As Ben Johnston observes:

A group of pitches may be very complexly related to one another, but often all of
them can be simply related to another pitch, which need not even be present. Thus,
the missing pitch is strongly implied by the complex group. The root of a chord, the
tonic of a tonality, the principal tonality of a modulating movement are all examples
of this principle. (Johnston, 1964, p. 61 [emphasis in original])

2) Use the ‘most economical’ possible representation of a pitch collection—that is,

choose the representation which uses the smallest integers in its interval ratios.
A corollary of this is that intervals such as fourths and fifths, which can be
represented by low integer ratios, should use these low integers when possible—
this means that they will have a stronger effect on the determination of a
collection’s fundamental than more complex intervals. Paul Hindemith, who
also argues for a version of Tonvorstellung (‘our ability to accept complex
intervals as versions of their nearest simple equivalents’), makes a similar
point when he states that the root of a chord is determined by its ‘best’
interval—the interval with the simplest just intonation ratio (Hindemith, 1942,
pp. 94 – 98).
3) Prefer interpretations in which the just intervals of the tone representations
correspond closely to the actual intonation of the music—that is, interpretations
which require the least possible mental retuning from heard intervals to the
referential just intervals.
4) Consider the position of pitches in a chord: the lower a pitch sounds, the more
weight it should be given when considering harmonic relationships.

III
Ligeti’s Melodien for orchestra is not a microtonal piece (it’s written entirely in
standard twelve-tone notation), but certain passages strongly imply tone representa-
tions which involve microtonal intervals such as 7:8 (231 cents) and 8:11 (551 cents).
These just microtonal intervals are approximated to the nearest semitone—in a sense,
then, one could argue that these passages are examples of microtonal music
forced into a semitone grid. Ligeti has long been interested in microtones and
overtones; Richard Toop (1999) discusses an abandoned electronic piece from
 Contemporary Music Review 271
the late 1950s based on a ‘synthetically produced overtone series’ of sine tones.
Microtones produced as high string harmonics appear in the first movement of
the Cello Concerto (1966) and the Second String Quartet (1968); the Double
Concerto (1972), Ligeti’s next work after Melodien, incorporates notated microtones
in the first movement. Bob Gilmore (2003, pp. 27 – 30) discusses how Ligeti’s 1972
encounter with Harry Partch and his music influenced the Double Concerto and later
works. Ligeti’s continuing interest in microtones and just intonation is evident in
such recent works as the Violin Concerto (1990 – 1992), the Viola Sonata (1994) and
the Hamburg Concerto (1998 – 1999).
Figure 3 is an abstraction of all the pitches from m. 11 to m. 19 of Melodien
(an excerpt of about 40 seconds). Jonathan Bernard (1999, pp. 3 – 10) has discussed
part of this section in terms of transformations in pitch space (that is, the space of
pitches in register, as opposed to modular pitch class space). In this approach,
interval is conceptualized as distance; Bernard’s analytical diagram of the music as a
‘graph of durations in pitch space’ makes the analogy of interval to spatial distance
explicit.
Turning from distance intervals to just intervals, we can arrive at a very different
reading of the passage as an example of changing tone representations. The excerpt
begins with a unison A6. Given the lack of a context which would imply otherwise, it
seems reasonable to assign this pitch the simplest possible tone representation of
‘fundamental’: A(1).
At m. 13, an F appears below the A (this dyad is doubled an octave above by the
celesta, but this does not significantly affect our harmonic understanding).
The simplest, most ‘economical’ way to hear the F/A dyad is as a just major third,
F(4:5). At this moment, the tone representation of the A changes—it changes from a
fundamental (1) to a just major third (5) above the new fundamental F.
The addition of E-flat in m. 14 does not affect our sense of F as root—rather, it
sounds like the approximated seventh harmonic of F([Link]). Other tone represen-
tations of these three pitches are possible, such as E-flat([Link]), B([Link]),
G([Link]), or D([Link]), but with no reason to favor these more complex
representations, the more economical F([Link]) is a better choice. The F has also been
already established as a fundamental in m. 13, so a sort of ‘inertia’ makes us likely to
keep the same partial classes for the F and A. In music which rounds off microtonal
just intervals to a semitone grid, it is necessary to accept larger degrees of tolerance

Figure 3 Pitch collections in Melodien, mm. 11 – 19.

272 R. Hasegawa
than in music rounded to a quartertone grid—thus, the 200-cent interval E-flat/F is
taken to represent 7:8 (231 cents) here, with a 31-cent discrepancy between the heard
and referential intervals. While F and A are still represented by the same partial
classes as in m. 13, the representation of the sonority must be understood as
occurring an octave higher in the harmonic series (changing the representation of F
from 4 to 8), to allow the inclusion of E-flat as F(7). The move into higher integers
corresponds with the addition of more complex and less traditionally consonant
intervals.
At m. 16, the addition of an F-sharp creates a harmonically ambivalent situation.
We can fit the new collection into an F spectrum, F([Link]), which requires a
move still higher into the overtone series, reinterpreting F from 8 to 16. However, a
competing interpretation emerges which better satisfies the ‘Law of Economy’: we can
hear the collection as an approximation of B([Link]). It’s difficult to choose
between the two; our sense of the F fundamental is weakened, but the sense of B as
fundamental is still not established (particularly as the pitch class B is not part of the
collection). The addition of a B to the sonority in m. 18 resolves this ambiguity, and
decisively shifts our sense of fundamental from F to B. The fifth B/F-sharp (one of the
simplest just intervals) plays a strong role in confirming B as the root. So, now our
Tonvorstellung of the total sonority is B([Link]).
As our sense of fundamental changes from F to B, the harmonic sense we make of
each of the pitches also changes: for example, the E-flat stops sounding like a natural
seventh above the fundamental, or F(7), and takes on the quality of a major third:
B(10). The tone representation of the interval F/A changes from F(8:10), at 386 cents,
to B(11:14), at 418 cents. Our tolerance for mistuning of these just intervals makes it
possible to hear the equal-tempered third of 400 cents as an approximation of either
interval, and as a pivot between the two fundamentals. Thus, by accepting a degree of
tolerance in our tone representations, complex just intonation patterns can be
conveyed in twelve-tone equal temperament.
The added D in m. 19 throws our recognition of B harmony into doubt, much as
the F-sharp weakened our sense of the F fundamental in m. 16. We can persist
with a B tone representation of B([Link]), but the simpler option
G([Link]) is more aurally convincing—and, analogously to the arrival of
B in m. 18, the G eventually appears a fifth below the D to confirm this reading.
Note that the shift in fundamentals from B to G mimics the shift from A to F at the
beginning of this passage.
In the bars following m. 19, Ligeti continues to add pitches more rapidly; our
analysis can keep up with only a few more additions before the density of pitches
overwhelms our capacity to discern a clear harmonic structure. After this point, a
motivic or transformational analysis (such as Bernard’s) could better describe the
music’s progress. The tone representations discussed here can easily coexist with such
atonal, piece-specific approaches to analysis. Motivic and atonal interval structures
may also take part in purely harmonic processes, which tone representation can
describe in detail.
 Contemporary Music Review 273
Unlike motivic or transformational analyses, which focus on unique intraopus
relationships such as the repetition or transformation of characteristic motives or
pitch collections, this analysis has used tone representation to describe how the
unfolding of pitches creates a sense of shifting just intonation sonorities. In place of
unique, piece-specific relationships, the set of just intervals is taken as an interopus
constant governing our harmonic perception. That is, instead of assuming a com-
pletely atonal world, in which the only landmarks are motivic correspondences
within a work, an analysis using tone representation posits a world of extended
tonality governed by a consistent set of just harmonic relationships.

IV
Vortex Temporum, a three-movement work for flute, clarinet, violin, viola, cello and
piano, poses some intriguing analytical difficulties. The work is a ‘spectral’
composition, insofar as it is composed by reference to models based on the acoustic
spectra of instruments. However, Grisey’s compositional techniques significantly alter
these spectra, sometimes to the point of unrecognizability.
The compositional techniques and plans which Grisey used to construct the work
are described in detail in studies by Jérôme Baillet (2000) and Jean-Luc Hervé
(2001)—and substantial corroborating sketches are also held by the Paul Sacher
Stiftung. The description of compositional process, however, is not necessarily a
good description of a piece’s 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.
Manfred Stahnke (2000) has made some brief but tantalizing observations about
how some of the harmonies of Vortex Temporum might be interpreted by ear—his
interpretations address aspects of the harmonies which are not evident from a
consideration of their compositional origin. By examining the beginning of the
work’s first movement through the lens of tone representation, I hope to offer some
new insights into its harmonic relationships as I hear them. Tone representation can
function as a sort of ‘listening grammar’ for complex microtonal sonorities.
The first two minutes of the first movement use a very limited set of harmonic
materials—we hear the alternation of three distinct ‘chords’, arpeggiated by the flute,
clarinet and piano. Each of these three chords was conceived by Grisey as a subset of
a ‘stretched’ harmonic series. The normal harmonic spectrum is systematically
distorted, so that each octave is stretched to approximately an octave plus a quar-
tertone. From the resulting distorted spectra, Grisey selects certain pitches for each
chord. Figure 4, drawing on Baillet and Hervé, illustrates the derivation of the three
chords (labeled, in order of appearance, x, y and z).
The stretched harmonic spectrum is conceived by analogy to the mildly in-
harmonic spectra of many common musical sounds: the upper harmonics of a piano
string, for example, are slightly sharp in comparison to the precise multiples of
 274
R. Hasegawa

Figure 4 Derivation of Vortex Temporum chords x, y and z from stretched spectra.

 Contemporary Music Review 275
the fundamental frequency. If Grisey’s more intensely stretched spectra were played
in their entirety, and with simple sine tone partials, we might recognize their
connection to this acoustic model. However, in Vortex Temporum, only a small subset
of each spectrum is heard, and that is ‘resynthesized’ using complex instrumental
timbres. The selection of a subset of pitches from a spectrum can completely efface
the spectral derivation of that subset—careful selection of pitches can even imply that
their source is a different spectrum altogether. Grisey’s own awareness of this
possibility is clear from a sketch in the Sacher Stiftung—by selecting carefully from
the stretched G spectrum in Figure 4, he derives subsets which imply a stretched
D-sharp spectrum (D-sharp/D- 3=4 sharp/B/E/G-sharp—partials 3, 6, 9, 12 and 15) and
even a compressed F-sharp spectrum (F-sharp/F-¼sharp/C-¼sharp/F-¼sharp—
partials 7, 13, 19 and 25).
Given this gap between spectral compositional procedures and the aural effect of
the derived chords, recounting the compositional process sheds little light on the way
the harmonies are actually heard. It will be more productive to analyze the chords
without reference to their derivation, concentrating instead on how we hear the
chords: their tone representations, internal tensions, and relationships to one
another.
Figure 5 shows how the three chords x, y and z are deployed in rehearsal numbers
1 through 19 (this figure is based loosely on Baillet, 2000, p. 214). At each rehearsal
number, the arpeggiation is punctuated by a cluster of very high piano notes, and
sometimes quick notes in the strings—however, these short-lived punctuations do
not seem substantially to affect the harmonic perception of the three sustained
arpeggio chords, so they will be omitted from this analysis. Sometimes, the
arpeggiation is joined by sustained single tones in one of the strings (shown here as
half notes followed by bars indicating their length). As fixed points opposing the
rapidly moving arpeggios, these sustained tones capture our aural attention very
strongly, and they will be given corresponding weight in our analysis.

Chord x
Figure 6 illustrates some possible tone representations for the chords x, y and z.
(Boxed pitches indicate the notes which are sometimes held as ‘pedal points’ by the
strings.) At the beginning of Vortex Temporum, the arpeggiation of chord x is
sustained for more than thirty seconds; it seems false to musical experience to
assert that we don’t begin making sense of this chord until there’s something to
compare it to. One advantage of tone representation is that we can say things about
collections of pitches without comparing them to other collections, as in pitch-class
set analysis. The technique of pitch-class set analysis is essentially comparative; in
technical terms, this means that we can say very little about a pitch collection until we
find something to compare it to. In many musical situations, however, we will want
to say something right away. The idea of tone representation will allow us to make
observations about the internal composition of this complex chord from the very
 276
R. Hasegawa

Figure 5 Arrangement of chords x, y and z in Vortex Temporum, rehearsal numbers 1 to 19.

 Contemporary Music Review 277

Figure 6 Tone representation analysis of Vortex Temporum chords x, y and z.

beginning of the work. This analytical approach does not view the identification of
motivic repetition within a work as the most significant type of relationship—instead,
we can cite the harmonic tendencies of any given collection, along with any internal
tensions or ambiguities of tone representation. These tensions are essential in making
the chord compelling to listen to for such a long time.
One way of hearing the first chord is as a combination of tones related to two
harmonic centers, D and D-¼sharp (see Figure 6): the A and F-sharp imply a D root,
while the A-¼sharp and D-¼sharp imply D-¼sharp. This tone representation is
essentially what we would expect from the compositional derivation of the chords:
the stretching process moves the C-sharp ‘fundamental’ up to D, then to D-¼sharp.
However, the pitches G- 3=4 sharp and B-flat don’t fit well with either the D or
D-¼sharp fundamentals (the question marks on Figure 6 indicate that the
corresponding pitches are unusually out of tune for the indicated tone representa-
tion). Describing the chord as based on D or D-¼sharp (or, for that matter, as a
stretched C-sharp spectrum) gives only an overall, statistical impression, which
doesn’t account well for details; to me, though, the harmonic details are precisely
what make this chord interesting.
278 R. Hasegawa
Another, more nuanced interpretation recognizes an F root for the lower part of
the sonority, combined with an E-¼sharp root for the upper part. An advantage of
this interpretation over the D/D-¼sharp one is that we can recognize a definite tone
representation for every pitch, accepting less mistuning between the sounding pitches
and their just intonation tone representations. Aurally, I find this tone representation
much more convincing, even though one must accept quite high partial classes,
17 and 21, to account for the F-sharp and A-¼sharp as part of the F spectrum. A low
F played beneath the harmony seems to fit well as a ‘root’ for the lower half of the
chord; this is largely due to the harmonic strength of the fourth C/F (an interval
which was problematic in the D/D-¼sharp reading). This interpretation gives us a
clear explanation for the prominent C-F fourth, and does not require that we distort
it into a more complex interval.
The arpeggio figuration tends to temporally separate the upper and lower parts of
the chord. Playing the upper part of the chord alone (from the G- 3=4sharp up) makes
its orientation towards E-¼sharp clearly audible. By recognizing a clearly delimited
E-¼sharp collection above the F collection, the sonority seems much more like
a compressed spectrum than a stretched one (as the chord’s derivation would
suggest); we also have a good explanation for the quartertone-flat octaves F/E-¼sharp
and C/B-¼sharp, which are responsible for much of the chord’s characteristic
tension.
The ambiguity between the two possible tone representations of chord x—as
a stretched spectrum with D and D-¼sharp roots or a compressed spectrum
on F and E-¼sharp—seems to reflect an essential part of Grisey’s musical
style. As Stahnke (2000) writes, ‘Grisey plays with the shape-finding capability
(Gestaltfindungsfähigkeit) of our ear, with the thresholds of our awareness’ (p. 383).
The play between two plausible tone representations helps to animate the chord
through this long passage.

Chord y
The first major harmonic change in the piece occurs at rehearsal number 6,
with the move to chord y. As we know, this collection shares a common origin
with chord x; both are subsets of the same stretched spectrum. We can easily hear
that the two chords are closely related, since more than half of chord y’s pitches
are common tones with chord x. Does this mean that the chords have the same
harmonic effect? In fact, the changed pitches, and particularly the overall higher
register, make the new chord’s tone representations subtly different from those of
chord x.
We hear a continuation of the F/E-¼sharp tone representation quite strongly—in
fact, the F root of the lower part of the chord is much stronger than in chord x, in the
absence of the F-sharp and A-¼sharp which were fairly weak partial classes of F.
Without these pitches, the C/F fourth can be represented as 6:8 rather than the more
complex 24:32. Three of the four new pitches (F- 3=4sharp, D, and D- 3=4sharp) fit well
 Contemporary Music Review 279
with the E-¼sharp fundamental, while the high G-sharp could be F(19) or a rather
flat E-¼sharp(20).
The sense that the chord might be heard (at least in part) as based on D is
considerably weakened by the absence of the low A and F-sharp. However, we can
still sense the possibility of D-¼sharp as a root for three of the lower notes of
the chord (C, D-¼sharp and F- 3=4sharp, but not F-natural). This is a byproduct of the
stretched spectrum in the compositional process—if we take a sample from further
up in the spectrum, the higher the perceived ‘root’ of that sample will be. The sense of
D-¼sharp as fundamental is strengthened by the absence of the A and F-sharp, which
were ‘out of tune’ with the D-¼sharp root, and by the added F- 3=4sharp, which is
D-¼sharp(10).
It’s also possible to recognize a weaker tone representation based on B-flat.
The strength of the B-flat interpretation is that it can relate nearly all of the pitches to
a single fundamental, but the very high partial numbers that it requires make it seem
less convincing than the F/E-¼sharp interpretation.
When the music returns to chord x from chord y at rehearsal number 7, our
perception of chord x is subtly colored by the experience of chord y. The sustained
D-¼sharp, a sustained note in the viola in chord y, is heard as closely linked to the
sustained cello A in chord x. The strength of this linkage makes an ‘in tune’ version
of the interval A/D-¼sharp desirable—since this interval is clearly emerging as a
harmonically important one, it makes sense to understand it as a representation of a
just interval. Only the F tone representation allows us to hear the interval between
A and D-¼sharp as a just relationship in the context of chord x, F(10:28), further
strengthening the sense of an F root as opposed to D or D-¼sharp.

Chord z
At rehearsal number 10, we hear chord z for the first time. The pitches of this chord
were selected from a spectral source set a tritone lower than that of chords x and y.
Is there any way that we experience this harmonic change as a tritone transposition?
Many of the pitches of chord z are a tritone below pitches in chords x and y
(F, G- 3=4sharp, B-flat, B-¼sharp and C-sharp in chord x map onto B, D-¼sharp, E,
F-¼sharp and G in chord z). However, the overall range remains the same, which
works against the sense of tritone transposition; also, as Stahnke (2000) points out,
the chords are linked by close voice leading, creating a sense of continuity with chord
x and weakening the sense of tritone motion: ‘In comparison to the spectrum on
C-sharp, we experience an almost stationary (gleichbleibendes) tone field, that seems
only to be illuminated by a different light. The quartertone movements are important
for this effect’ (p. 382).
Even acknowledging the many pitches in chord z which are related by a tritone to
pitches in chord x, I don’t hear an overall transposition by tritone here. To me,
the most aurally convincing account of the sonority is as a combination of
pitches implying an A root in the lower part of the chord and an A-¼sharp root in
280 R. Hasegawa
the upper part. (The sense that the sustained cello A of chord x lingers aurally into
chord z strengthens this tone representation.) While in chords x and y, the upper part
of the chord was based on a fundamental a quartertone below the fundamental of
the lower part (E-¼sharp above F, or in spectral lingo, a ‘compressed’ spectrum),
here, the upper part of the chord is based on a fundamental which is a quartertone
above that of the lower part (A-¼sharp above A, or a ‘stretched’ spectrum).
The contrast between these two types of tone representations further weakens the
case for a tritone transposition—we have a sense of a different type of harmony,
not a transposition of the same type. If we do hear a change of overall ‘root’ from
chord x to chord z, it is likely to be from F to A—transposition up a third, not down
a tritone.
It’s also possible to hear tone representations of E and C in this sonority—
particularly before the entrance at rehearsal number 16 of the sustained viola
D-¼sharp, which contradicts both of these tone representations. The D-¼sharp
strongly focuses our tone representations toward the A root for the lower notes of the
chord; the lowest trichord, B/D-¼sharp/E, is only possible with A as fundamental:
A([Link]). The common D-¼sharp sustained tone with chord y also makes chord z
sound similar to y in some ways—both are heard as contrasts to the more ubiquitous
chord x. The shared high G-sharp and nearly identical overall register also help to
cement a relationship between the chords, although they are harmonically quite
different.
The return of the D-¼sharp over chord z as a sustained tone is a striking harmonic
event; this sustained pitch has previously only appeared with chord y, and we hear it
quite differently in its new context. D-¼sharp is now heard as A(11) instead of F(7),
and the dyad of sustained string notes A/D-¼sharp is reinterpreted to A(4:11) from
F(10:28). The reinterpretation of this emphasized dyad, highlighted by the long
duration of its component notes, is heard as an essential part of the harmonic move
from F to A. (One can find a similar harmonic relationship, but in reverse, in
Melodien: the dyad B/E-flat changes from B[16:22] to G[10:14].)
The close reading offered here indicates a possible route towards a deeper
engagement with the harmonies of spectral music. By moving away from a des-
cription of compositional tools towards an analytical method which describes audible
harmonies, we can begin to describe and discuss the fascinating tensions and
ambiguities of this harmonic language in detail.

V
Tone representation is a highly flexible analytical tool—it can be applied to styles
ranging from tonal music (as in Riemann’s original formulation) to ‘atonal’ works in
twelve-tone equal temperament such as Melodien or quartertone works such as
Vortex Temporum. The idea of interval as ratio helps to reflect important aspects of
musical experience which are not dealt with convincingly by distance-based
approaches to interval—particularly the effect of specific intervallic qualities and
 Contemporary Music Review 281
the acoustic relationships of pitches and intervals to their fundamentals or
roots. Tone representation provides an alternative way to explore our harmonic
intuitions, and can shed new light on the ways that we come to understand the music
we hear.

References
Baillet, J. (2000). Gérard Grisey: Fondements d’une écriture. Paris: L’Harmattan.
Bernard, J. (1999). Ligeti’s restoration of interval and its significance for his later works. Music
Theory Spectrum, 21(1), 1 – 31.
Gilmore, B. (1995). Changing the metaphor: Ratio models of musical pitch in the work of Harry
Partch, Ben Johnston, and James Tenney. Perspectives of New Music, 33(1/2), 458 – 503.
Gilmore, B. (2003). The climate since Harry Partch. Contemporary Music Review, 22(1/2), 15 – 34.
Hervé, J.-L. (2001). Dans le vertige de la durée: Vortex Temporum de Gérard Grisey. Paris:
L’Harmattan.
Hindemith, P. (1942). The craft of musical composition. New York: Associated Music Publishers, Inc.
Johnston, B. (1964). Scalar order as a compositional resource. Perspectives of New Music, 2(2),
56 – 76.
Parncutt, R. (1988). Revision of Terhardt’s psychoacoustical model of the root(s) of a musical
chord. Music Perception, 6, 65 – 94.
Riemann, H. (1914/1915). Ideen zu einer ‘Lehre von den Tonvorstellungen’. Jahrbuch der
Musikbibliothek Peters, 21/22, 1 – 26.
Riemann, H. (1992). Ideas for a Study ‘On the Imagination of Tone’ (R. Wason & E. W. Marvin,
Trans.). Journal of Music Theory, 36(1), 81 – 117.
Stahnke, M. (2000). Zwei Blumen der reinen Stimmung im 20. Jahrhundert: Harry Partch und
Gérard Grisey. Hamburger Jahrbuch für Musikwissenschaft, 17, 369 – 389.
Tenney, J. (1992). The several dimensions of pitch. In C. Barlow (Ed.), The ratio book
(pp. 102 – 115). Cologne: Feedback Studio Verlag.
Tenney, J. & Belet, B. (1987). An interview with James Tenney. Perspectives of New Music, 25(1/2),
459 – 466.
Toop, R. (1999). György Ligeti. London: Phaidon.
Väisälä, O. (2002). Prolongation of harmonies related to the harmonic series in early post-tonal
music. Journal of Music Theory, 46(1/2), 207 – 283.
Wason, R. & Marvin, E. W. (1992). Riemann’s ‘Ideen zu einer ‘‘Lehre von den Tonvorstellungen’’’:
An Annotated Translation. Journal of Music Theory, 36(1), 69 – 79.