Claude Vivier’s Zipangu

1. Mistaken Memory

When I first heard Claude Vivier’s Zipangu, I was struck most of all by the melody that opens and closes the work. As I recalled the piece in the days that followed–or rather, as the piece repeatedly worked its way into my thoughts–the middle part of the piece, the material in between the two statements of the melody, gradually disappeared, leaving only the melody. This was not like taking a catchy tune home from the symphony. Instead, like a computer virus, the melody took over the rest of the piece, replicating itself and altering structures where it went. As my memory of its details faded, the piece became a single melody. Had I thought about it at the time, it might have struck me as odd that Vivier could have stretched out that single simple, albeit striking, melody for nineteen minutes.although some recent music by younger American composers like Aaron Jay Kearnis, David Lang and Joan Tower has featured a single melody spanning an entire piece.

My next encounter with the piece, and my first glimpse at the score, took place about a year later under the guidance of my composition teacher at the time, who mostly drew attention to the work’s textural aspects, pointing out the details in various passages of Vivier’s characteristic string writing. This time I was surprised at how little of the work was taken up by the melody that I had remembered as being all-important. Although I appreciated the work’s textural accomplishments, hearing the work as a textural essay left me somewhat dissatisfied.

When I recently began to listen to the piece again and study the score in earnest, the relationship between melody and texture became clearer. It also became clear why my memory of the piece in those days after first hearing it was mistaken, if mistaken in a way that jibed nicely with the work’s broader purposes. In addition to the work’s melodic attractions, I also became increasingly interested in proportional structures–metrical and durational schemes–that leapt at me from the score. Once again, this work was surprising me: passages that had seemed rhapsodic turned out to be structured and controlled.

2. Melody and Colour

Zipangu–the title refers to the name of China during the time of Marco Polo–was composed in 1980 with a commission from New Music Concerts. It is scored for thirteen strings, divided into two groups. Group I, seated to the left, consists of six violins. In group II are one violin, three violas, two cellos and one double bass, all seated to the right. Vivier calls for moderate amplification (if possible), with each group mixed into one channel and the left-to-right output of the speakers matching the placement of the players. The score’s list of notational signs contains a few that are less common, including signs for irregular, accelerating and decelerating tremolos, and indeterminate artificial harmonics. This last sign is used in conjunction with glissando figures. Vivier also includes a sign for an effect that he refers to as “granular sound,” obtained by using exaggerated bow pressure. In Vivier’s usage granular sound has a greatly increased noise content, but a sense of pitch is not completely eliminated, as is sometimes the intention when some other composers call for this effect.

In his program note for the work, Vivier sets out his main concerns: “Building around a melody,” he writes, “I explore different aspects of ‘colour’ in this piece. I have tried to veil my harmonic structures by using different bow techniques. A colourful sound is obtained by applying exaggerated bow pressure on the strings as opposed to pure harmonics when returning to normal technique. In this way, melody becomes ‘colour’ (chords), grows lighter and slowly returns as though purified and solitary.” The score includes an additional comment by its editor, Serge Garant, which explains the workings of the piece in more straightforward terms. “The melody the composer is speaking of is always present in this work,” writes Garant. “It is clearly expressed both at the beginning and the end of the work, but undergoes all sorts of transformations throughout the rest of the piece. In one of the most beautiful passages, we hear a solo violin playing a very fanciful air against a texture made up entirely of harmonics and in which we recognize the basic harmony and its harmonization. It is a work which towards its end achieves a deeply moving lyricism in a grave and sombre passage.” [2]

Vivier’s treatment of melody in this work, or rather his conception of melody, undoubtedly owes much to his studies of Eastern music during travels in the mid 1970s. (It should also be noted, however, that many such Eastern ideas concerning melody have become an established part of Western working methods in the second half of the twentieth century; such ideas were certainly brought up during the course of Vivier’s studies with Gilles Tremblay, whose own teacher had been Messiaen.) In a classic Western melody, notes are related to each other in ‘precedes’ and ‘succeeds’ predicates. One can refer to, for example, the third note of “Happy Birthday” and know its precise relationships to the notes immediately surrounding it. In Zipangu, the notes of the melody are ordered in looser terms, such as ‘comes earlier than,’ and ‘comes later than.’ For example, in a melody of this type containing ten different pitches, the first three pitches will tend to occur earlier than the last three, but might be re-iterated many times, in many different orders before those last three occur. This is similar to a cellular melodic approach, sometimes adopted in serial music, in which more or less independent cells follow one another in a rigid order, while the ordering within the cells themselves remains loose.

Specifically, the melody in Zipangu derives from a source set of fifteen pitches (ex. 1). Four of the pitches are represented twice, giving a total of eleven discrete pitches. Each pitch, although not absolutely fixed registrally, has a preferred register in which it usually occurs (in passages with octave doubling, the upper register is taken to be the operative one).

Ex. 1, Main Pitch Material

The work as a whole is divided into sections, with clear textural changes marking the divisions. Each section, as Garant notes, has some form of the melody through it. There is one notable exception, Garant’s “grave and sombre passage.” The melody is always in the highest register, usually played by violin 1. But the melody is not always given the preferential treatment it receives at the beginning and end of the work, at which points it dominates the texture. At other times, the melody is woven into the texture, or covered with more active material, as it is in Garant’s “fanciful air.” Furthermore, the entire source set is not exhausted in each section; sometimes only a few of the pitches are used, in one case, only a single source pitch is used. Given that each section is rather single-minded and closed, and since the work opens with a clear, thematic statement of a melody that is subsequently transformed throughout the piece, it is possible to view the work as a kind of theme and variations. I would not want to push this characterization of the piece too hard, especially since Vivier himself included none of the usual trappings of variation sets (e.g., numbering the variations) but it does help to convey an overall sense of the work’s form. Ex. 2 (see appendix for longer examples) provides a summary reading of the piece as a set of variations. Ex. 3 outlines the melodic use of the source set throughout the work.

The initial presentation of the melody, which occurs immediately at the opening of the work, is striking and memorable on several counts. The texture here is conceptually simple, but sonically complex. The texture consists of two parts: the melody, presented in group I, with violins 4, 5, 6 doubling violins 1, 2, 3 an octave below; and a drone bass in group II. The drone, four octave Es, supplies the complexity: each instrument varies the bow pressure from normal to granular to normal, with each player maintaining an individual tempo, or rate of change. Even though the E pitch centre comes through forcefully, the overtone and noise content of the sound is constantly changing. This process is roughly similar to klangfarbenmelodie, which Vivier also uses later in the piece.

Following the unambiguous presentation of the melody in the opening of the work, Vivier proceeds to more subtle, ‘veiled’ versions of it. At mm. 40.54 the melody is presented in a homophonic setting, with block chords supporting each note of the melody. The harmony, consisting of strident tetrachords and hexachords–this after the simple intervallic music of the opening–together with the use of granular sound, tremolo and rapid repeated notes, draws attention away from the melody and towards the texture. However, the melody–this time without any internal repetitions–is still plainly present in the top voice. Next Vivier all but forces us to abandon the melody for the texture, withholding everything but texture. Mm. 55.68 consist of a single chord, the first tetrachord of the previous section, and a constantly changing texture made up of elements of granular sound, tremolo, dynamic fluctuation and repeated notes. In other words, this passage is total Klangfarbenmelodie. It is not totally devoid of melodic import, however, since the top voice plays the first melody note throughout, creating an explicit, literal prolongation of that note. In effect, Vivier is playing with the notion of scale here, and has zoomed us in to an extreme close-up on this first member of the source set. Like taking a magnifying glass to the type on a page, we find that smooth edges transform into jagged patterns of dots and smears.

3. Structures and Schemes

During an interview in 1981, Vivier spoke on the relative importance of melody, structure, and intuition in his work:

In my opinion, Kopernicus and Orion were disturbing works. I wrote them as a reaction against some contemporary music which reflects that composing today is the equivalent of inventing structures. It is crazy, structural frenzy, to think that this is the only way to create real works of art and to forbid all inspiration resulting from what I call musical emotion. In my case, a melody is often the base for a whole work. I compose the melody and sing it to myself all day until it develops itself and takes its own form. Sometimes it can suggest the larger form of the piece as the organization of its smallest parts. Some expressions of emotion are too subtle to be manipulated or interpreted by ‘structuring’ instruments like mathematical models. I have to feel close to my materials, I have to live it. Some of the parameters of my music–such as length and proportion–are very organized, but my approach is intuitive as well as organized.

It would be all too easy for a casual reader of the above to gather that Vivier worked more from melody and inspiration than from ‘mathematical’ structure. It seems almost as though Vivier is trying to convince himself that this is the case. But in fact, ‘structural frenzy’ cannot readily be separated from melody or intuition in Vivier’s music; what may appear to be rhapsodic turns out to be highly controlled. In Zipangu, proportional and durational schemes are manipulated with the same sense of play that another composer might show when improvising a melody.

The structural play in Zipangu takes place principally on the two levels of metre and individual durations. Organizationally, these two levels are distinct, but Vivier often links them in interesting ways. An earlier example of Vivier’s manipulation of metres is offered by the solo piano piece Shiraz of 1977. Typically for a Vivier piece, Shiraz’s program notes point the listener to the work’s exotic inspiration, involving, among other things, two blind singers in a marketplace in Iran. The score itself, however, draws the observer’s attention to interlaced ascending and descending metrical patterns in the opening measures. The descending pattern consists of successive decreasing members of the fibonacci sequence (which is heralded by the initial metre signature of 55/16). Interlaced with this is an ascending pattern that begins as a doubling series, but quickly becomes unpredictable. In other words, although there are predictable, algorithmic elements within the section, these elements are combined in an unpredictable, or intuitive way to form a whole that could not have been produced by an algorithm.

Throughout Zipangu, Vivier plays with similar metrical patterns of interlaced ascending and descending series. Often an entire phrase will fall within the scope of such a pattern. The very first phrase of the piece is an example. A descending series, beginning with eight quarter notes is alternated with an ascending series that begins with one quarter note. When both series meet at their first common member, the pattern is complete and the phrase is closed (ex. 4). Vivier marks this closure in no uncertain terms, with a crescendo to ff in the last measure of the phrase leading to p subito on the downbeat of the next phrase (the first change in dynamics of the piece).

Ex. 4 mm. 1.11

(The numbers in ex. 4.8 refer to the number of beats in the metre signatures; all are measured in quarter notes.)

The same procedure is used to generate the next phrase-length as well (ex. 5).

Ex. 5 mm. 12.18

Later on, Vivier uses this scheme to produce a metrical palindrome, which governs an entire section (ex. 6). This is the klangfarbenmelodie section. In this case, a palindrome is particularly apt: the palindrome adds, if perhaps in a conceptual way only, a further element of stasis to the passage, since there is no difference between forward and retrograde motion through it.

Ex. 6 mm. 55.64

Another feature of this device is that successive pairs will form an invariant sum. This allows Vivier to create a metrical pattern that oscillates on one level, but which is constant on the next higher level. Vivier exploits this in the section mm. 69.81, where the higher level metre is brought down to the lower level in the final measure (ex. 7).

Ex. 7. mm. 69.81

Vivier also seems to have been working with a higher level metre in the following example creating a metrical acceleration to the final cadence (ex. 8). It should be emphasized that each of these examples constitutes the pattern for an entire music unit, a phrase or a section. In the examples that are least likely to produce an audible impact on the music, these patterns are still important indicators of Vivier’s working methods (or what kinds of constraints he placed on his intuition). In any event, the first two examples certainly do have an audible impact.

Ex. 8 mm. 239.244

In addition to these metrical patterns, Vivier also creates small-scale structures using durational series. These series also occupy prominent positions, especially at the opening of the piece, where several of them occur. The first rhythmic gesture of the piece is built on a series of durations in which each successive duration is shorter than its predecessor by one sixteenth note. Immediately following this is an ascending series in which each successive duration is larger by an eighth note triplet (with some fudging, necessary to make the durations fit the metre). These ascending and descending rhythmic gestures occur within the context of the interlaced metrical structure mentioned above (ex. 4). It was noted that in this case the metrical structure was audible. This is made clear in ex. 9, which gives the melody for the first six measures of this section. In each of the two subphrases, the goal of the subphrase occurs in the measure corresponding to the ascending metrical series. This is both the melodic goal of the subphrase, and the goal of the durational series, i.e., the duration required to complete the descending series in the first subphrase (four sixteenth notes), and the maximal duration of the ascending series of the second subphrase (six eighth note triplets).

Ex. 9 mm.1.6

Throughout the piece Vivier creates similar gestures using ascending, descending, palindrome and interlaced (e.g., 4, 1, 3, 2) series. Again, these constructs are not employed in a systematic or algorithmic way, they are simply a part of Vivier’s rhythmic language, used freely. These are summarized in ex. 10. These patterns do, however, have a significance on the interpretation of rhythm in Vivier’s music (as well as many other twentieth century composers). Figures that appear to be ‘syncopated’ or ‘complex’ often turn out to be quite simple in conception. The idea of syncopation is based on a system of normative sub-accents within the measure. Given Vivier’s manner of constructing rhythms, it would be wrong to try to infer a system of sub-accents. The downbeat, however, retains its normal accentual role. Many cases of suppressed downbeats are actually instances in which the measure in question is a subdivision of a larger metrical unit. For an example of this we need look no further than m. 2. In this case, the operative metre is 8/4, which Vivier has divided in two for simplicity’s sake.

4. Conclusions and Comparisons

Vivier was not a composer who felt the need to invent himself anew with each composition; in many of his pieces we find similar techniques, similar textures and similar contrasts. But beyond this, there are some links in several of his later works that verge on the spooky. Probably the most notable example involves a recurring solo double bass gesture. In m. 208 of Zipangu, a double bass plays the series of natural harmonics based on its low E string, up to the tenth partial, G#. This gesture is not totally mysterious: it serves to link the two boundary registers present at the opening of the work, and thus prepares for the recapitulation of the main melody in m. 229. In Lonely Child (also composed in 1980), the same gesture appears, also near the end of the piece, on p. 27 of the manuscript. This time the series goes one note beyond the one in Zipangu, to A#. Prologue Pour Un Marco Polo (1981), also scored for thirteen strings (this time within a full orchestra) ends with the same gesture, this time extending another note up the series, to B. In the same piece, at rehearsal 3 there is a dramatic sixfold repetition of the word ‘Zipangu’ in the vocal parts, supported by an E-G# sonority in the orchestra. Immediately following this, the full string section enters with the boundary pitches E-G# in the same registers as at the opening of Zipangu. Finally, Wo Bist Du Licht? (1981) opens with an interlaced rhythmic gesture, featuring granular sounds (which are used in Marco Polo, as well). Zipangu, Marco Polo and Wo Bist Du Licht? all open with an identical tempo of quarter note = 70 MM. Whatever these congruencies may have meant to Vivier, they are further signs of a musical mind that was capricious as well as craft-conscious. By all accounts, these contrasts were reflected in his personality. These features add up to a musical style that is engaging and admirably direct, even if somewhat mystical in the end.


1: Taken from the score, published by Doberman-Yppan, 1985.

2: ibid.

3: Sylvaine Martin, “Intuition as well as reason guide Vivier when composing,” The Music Scene, July-August 1981, p. 6.


Ex. 2 Formal Overview

Ex. 3 Zipangu–Progress of a Melody

Ex. 10 Zipangu Durational Series