Tristan Murail Lecture (an excerpt)
August 18, 2003
Tristan Murail: What I will do is to try to tell you simply how I came to write the music I write - as much as I know myself. I won’t go all the way back to the beginning, but start from my early “professional” experiments in composition. In fact, you are going to hear three pieces performed here at Ostrava Days, which all belong to this first period of my experiments with composition. The most recent piece performed here will be Gondwana, which I wrote in 1980 - quite a long time ago! Another piece will be Ethers, which I think will be performed at the end of the festival. But perhaps I would like to start with another piece I wrote at more or less the same time.

You may know that I was a student of Olivier Messiaen at the Conservatory in Paris (not that there is anything special about this - everybody wanted to study with him, and so did I). At that time, the musical atmosphere in Paris was very much dominated by serial music, and by Boulez’ influence, even if he was not physically present in France at that time. So serial techniques were hugely predominant. The most famous ensemble for new music was the “Domaine Musical,” which had been founded by Boulez and, consequently, mostly played pieces belonging to those aesthetics. Messiaen himself thought that serial music was very ”advanced,” as he said; something all young composers should try to use as a compositional technique. So, many colleagues of mine composed “serial” music, and I tried it too. I wrote a few pieces that were sort of ”twelve-tone,” without being strictly serial, but I soon realized that it didn’t fit me. I had musical ideas, sound images that I wanted to express, and I could not do that with the serial or twelve-tone technique.

Very soon, I had the feeling that it was very important to find some other way for my compositional development. Of course, you cannot start from scratch; you need to have examples or models, or cues, in order to start a trip like this. A good thing about Messiaen’s classes was that he invited many composers to speak and present their own music. I remember seeing and speaking with [Iannis] Xenakis in Messiaen’s class. Xenakis had brought some of his big orchestra pieces — Metastasis, Pithoprakta — and he explained them. I was quite impressed by his approach, which was very different from what you were taught at the conservatories. You were taught melody, harmony, counterpoint,etc.; while Xenakis thought of sound masses, in which individual lines — i.e., the notes played by the performers — were not necessarily that important. What was important was the structure of the mass of sounds.

I discovered a very similar approach in [the work of Gyorgy] Ligeti when Messiaen played some of his music for us: orchestral pieces like Lontano or AtmosphŹres. These two pieces had just been released and he brought the recording with the scores. It was kind of a shock for most of us, because we saw that there were possibilities other than serialism. You didn’t have to think about music in terms of accompanied melody or counterpoint. Ligeti and Xenakis had in common these other ways of thinking about music, masses of sound, processes, etc.... And the importance given to timbre, as a way of structuring the form. The concept of “process” - where a musical situation changes from one to the next – was also very different from the old notion of musical development.

But all that was not enough. The big problem remained the same - and is the same even today: Which notes should I write in my score? Dodecaphony had tried to address that issue. But if you don’t want to use this technique, and at the same time don’t want to go back to tonality or modality, what, then, do you do? Ligeti, in a way, had an answer to this question. In a piece like Lontano he reintroduced very clear harmonic compounds. Lots of octaves throughout the piece, and also a very striking feature: a series of tritones repeated from octave to octave. Then there are clusters, which oppose these clear harmonic structures. You will also find pentatonic chords. Ligeti uses clear harmonies, which, in a way, are related to tonality; but the music is clearly not tonal.

Ligeti’s example showed us that it was possible to reintroduce very clear harmonies, but also to use harmony as a function, which had not been the case for many decades. In serial music, and even in the music of Messiaen, harmony is not functional. It’s more like a coloration. Messiaen likes to “color” time as it passes, but his beautiful harmony doesn’t have a real formal function, like it has in tonal music. However, I do think that harmony is a very powerful way of structuring form in music.

Having left Messiaen’s class, I started to explore electronics. The electronics of that time, of course, were quite different from what we have now. They were based on analog principles. Having received the Prix de Rome, I spent two years at the French Academy in Rome, where there was a small studio that I could use for experimenting, although the electronic equipment was rather basic. A few of the things I did there were close to Alvin Lucier’s experiments — for example, toying with a feedback loop. Recording sounds on one tape recorder; the loop carried the sound to the second tape recorder; the second tape recorder fed the sound back into the first recorder. What happens in this process is that, as the sounds are copied and copied again, they get transformed; degraded in some way. This way, you get a process, which works by itself, and I was very interested in this.

Perhaps we can start listening to some music now. One of the first pieces I wrote after my studio experiments was Mémoire/Erosion, for ten instruments. There [were] no electronics in the piece, but the instrumental scoring took electronics as a model — a formal and timbral model. In a feedback loop, you need to have someone playing or singing into a microphone. In this case, the horn plays into a “virtual” microphone, and the nine other instruments – five strings and four woodwinds – play the role of the two tape recorders. Everything the horn plays is repeated, after a delay – well, not really repeated, because it is also transformed - by one of the other instruments, until the music drifts into something else. So you have a process going on, but also a certain kind of entropy - entropy meaning going back to “chaos,” to noise, in a way. This resulted in a metaphor for the entire piece, the opposition of “pure” sound versus “noise” - noise in a musical sense (we’ll have to come back to the issue of “noise”). So, when the horn plays a sound or a small phrase, it is repeated and transformed by the other instruments until it becomes something totally different. You wind up with a new element, which the horn then takes over and uses as a new idea, a seed, so to speak, which is fed into the system again. This is the main idea of the piece. Now you can see how the electronics served as a model. This was the beginning of a long search for models, which I’ll try to describe to you. But let’s hear a few excerpts from the piece. The title – Mémoire/Erosion – refers to the two aspects on which the piece is built, memory and erosion.

[Plays recording]

You heard that the piece started with a C. Well, you heard it if you have perfect pitch.... [Laughs] So, the music at the beginning is very simple, centered on this C4. At the end of the first section you hear a Bb, and the music now revolves around the Bb sound; there is, in fact, a continuous transformation from the C to the Bb. This is something I found during my experiments with loops. Quite often you fed some sound into the system and, after a number of repetitions, you got another frequency, different from the one you had played into the system, which in fact depended, I suppose, on the resonant frequencies of the room where you worked. One of the ideas of the piece was to exploit these ”flaws,” in a sense.

You may have noticed that the piece starts with instrumental ”noise.” The instruments are required to play noise sounds, which today is quite standard, but was less so at the time. To produce this noise, all the strings have to bow directly on the bridge, and make a hissing sound without a pitch. But it became a challenge when we recorded the piece for a disk. At that time we still had vinyl disks, and this hissing sound wound up sounding like background noise, instead of musical noise. So we had to reinforce the noise in the recording, to make it clear that it was something desired, and not just a fault in the recording!

[missing part of lecture]

Let’s go further into the piece. The idea of the feedback loop was used by many composers at that time. It was often used in real time, in concerts. Usually, there were a lot of technical problems with it, caused by the tension of the tape. Another problem was the length — the duration of the loop — which didn’t change, so you had always the same interval of time between repetitions. Somebody — I think it was in Freiburg, Germany — tried to build a very complex machine which could change the length of the loop, but it never really worked because of difficulties adjusting the tension.

But when you write this idea on paper, imitating the tape loop with instruments, changing speeds is not a problem. In this piece the speed of the “loop” changes very often. In our next example, you will hear a very clear harmony, which drifts away. Actually, the harmony does not drift, but the timbre of the chord does.

[plays excerpt]

When you use feedback loops,a number of problems can happen; for instance, saturation. The volume of the sound gradually increases and there is nothing you can do about it; at the end of the process, you get a kind of crashing sound. The only thing you can do is to stop the process. This is exactly what happens in the middle of the piece, the saturation being here “transcribed” for the instrumental ensemble. Or, when you work with a real tape loop, you have to use a sticker to create the loop. After a while, the tension of the loop can pull the sides of the sticker apart, and you end up hearing a ”click.” In this piece, the repeated dampened [pizzicato] of the viola is a kind of an imitation of something that really happens with a physical tape loop. This “click” sound is also used as a seed for the next situation, as the horn plays very short sounds together with the “sticker sound,” and a new section starts from that.

The notion of repetition is, of course, present throughout the piece, because of the virtual tape loop. But it isn’t textual repetition; nothing is ever exactly repeated. There is always a small transformation, depending on the process I am using. That’s different from what you find in minimalist music, for instance, where repetitions are exact and where changes proceed by clear steps. Though there is some connection between the two styles of music, in essence they are different.

Later, I kept two or three ideas from this period: the concept of process, especially timbre process; [and] the notion of regular repetition or, rather, the notion of ”periodicity.” I also came back to using technological models in various contexts. Because technology has evolved so dramatically in the last two or three decades, it has been a great source of new ideas for me.

Alvin Lucier: Did you use some kind of process to determine the pitches, or did you do it by ear?

TM: At that time, the pitches were chosen very intuitively.

Lucier: You mean you created the distortion intuitively?

TM: Yes, it was the product of an additive process – in this case, adding more and more inharmonic partials to a simple inverted seventh chord, which you can equate to a portion of a harmonic series. My colleague Gérard Grisey started to write pieces which were consciously using harmonic series, or instrumental spectra as models. But at that time, I was mostly toying with ideas based on technology. The harmony here is very simple. In fact, in this piece, the timbre is more important than the harmony.

Lucier: How about the rhythm? I understand part of it comes from the loops; but beyond that, where did you get the microstructures?

TM: Partly from the loops. But I also tried to bring in microstructures that contradicted the loops in some way, to get a more interesting rhythm.

Lucier: Was this contradicting intuitive also?

TM: Yes, very much so. But once the process had been decided, it was very mechanical. I just notated all the consequences of the process: Sometimes, I suppose, I made changes, but I knew that I was going from A to B in some way, and I tried to deduct everything from this idea.

Student: Did you know in the very beginning what [the ending point] was going to be or did you just figure out a way of getting there while writing it? Or did you know how the process is going to evolve?

TM: It worked both ways. Sometimes I didn’t know where it would go, and sometimes I did, I suppose - to be honest with you, I don’t remember exactly. But both possibilities exist: You can have a process where you know the starting point and the ending point, and then you have to decide how to get from one point to the other. In a way, you can call this a process of interpolation. But you can also just have the starting point, and a system which transforms progressively that starting point: this is ”extrapolation.” In general, I prefer to know where I am going, but sometimes I just try something, and see what happens.

Perhaps I should speak a little bit about the next piece, Ethers, which will be performed here in Ostrava. I wrote Ethers two years after Mémoire/Erosion. You still find the idea of transformation and repetition, but this time there is no looping. The durations of the repetitions are also subject to a process, which means you never have a steady pulse. The tempo is never stable; it always accelerates or slows down (which creates lots of complications for the performance!). For me, the piece is mainly about relativity. Everything here is relative; there is no fixed time reference. There are also a few technological models involved in this piece, like echo chambers, reverberation, and my first attempt to use frequency modulation in order to organize pitches. The piece is written for flute and five other instruments. The flute is the principal instrument, but does not have a real solo part. Actually, its role is to provide acoustic models for the piece. I didn’t try to analyze the spectrum of the flute at that time, but I experimented with some of its “extended techniques,” and used them as models for the instrumental textures.

[draws on blackboard]

At the beginning of the piece, the flutist is singing – that’s the square note - and playing another pitch at the same time. In addition to these two pitches, you will get two resulting sounds. So, how do you get them? Well, let’s say this C#4 (the sung pitch) has a frequency called “A” and the Bb5 (the played pitch) has a frequency called “B”: then the resulting sound [is] A plus B, and B minus A, i.e., a D6 (in fact, a quarter-tone too high) and an E5. So you see that with a very simple calculation, you can predict the result very precisely. This is similar to amplitude modulation, often known as “ring modulation.” Ring modulation was one of the first devices used in studios early on to create inharmonic timbres. The initial section of the piece is based on these modulation sounds proposed by the flute. The strings start playing very high sounds: flageolets, which are, in fact, harmonics of these four pitches that compose the flute multiphonic sound. [draws on board]

Progressively, the frequencies played by the strings come down and introduce the flute, anticipating its multiphonic sound. After the flute plays, we go back progressively to the high ranges – the string harmonics - and start the process again, until the flute proposes a new combination. This acoustic model is also used at the very end of the piece.

[TM plays Ethers]

I said that with ring modulation, if you input two sounds, you get the sum and the difference of their frequencies. This is supposing that the two sounds are very pure, like sine waves. Of course, if you use instruments they never are, and you also have to take into account their harmonics. Let’s make an example with simple sounds: let’s say sound “A” has two harmonics [draws on board]. The second harmonic of “A” is, of course, twice the frequency of the fundamental. Let’s say sound “B” has three harmonics: B, 2Bs and 3Bs. Now you have to take into consideration all of these five frequencies and intermodulate them. The result will be: A+B, A–B, but also, 2A+B, 2A-B, A+2B, A-2B, etc.... It can be quite complex. In the first section of Ethers, the trombone is playing one of these second-order modulation sounds, which enriches the harmony.

Lucier: When you sing the C# and play the Db, can you actually hear those side bands?

...full version in Ostrava Days 2003 Report