Experience suggests to me that different composers "hear" in strikingly different ways. That is, "a musical experience" can rest upon varied bases and have markedly different character depending upon who is reporting. It would seem inescapable that the composer knows a great deal about the music s/he has written before it is actually heard. I am rarely surprised in any way by the objective fact of the sounding of what I have written. But this does not translate into a comparable certainty about how the music will feel, what it will evoke emotionally. Part of the difficulty here is, of course, that a performance emerges gradually as the musicians individually (and as an ensemble collaboratively) find their ways. There are, I think, explicit stages through which any rehearsal process passes. Almost at the same time that the notes are being learned, the question of temporal orientation arises and becomes primary: when do I play? with whom am I to coordinate? So, during the first rehearsal, one is hearing an incomplete, rather approximate evocation of what one has written (and, of course, of what the musicians will eventually achieve). It takes time for a composer to acquire the necessary patience to be able to suppress the feeling that one should jump in each time something does not immediately sound as one believes it should. This process of "successive approximation" is shared by the individual musicians who know, in all probability, nothing at all about the work as an imagined entity. They only begin to hear themselves as participants in a larger, cooperative enterprise as rehearsals continue. Often (too often), it may not even be until the performance itself that they begin to actually experience the substance that the musical work is attempting to address. This is a matter of economics17.
If one is not comfortable with what one is hearing (and also feeling) as rehearsals proceed, there are some adjustments that can be made, but it is impractical to actually recompose in major ways. The time constraints and the conventions of the musical world discourage it. Three main kinds of adaptations can be made without undue disruption: dynamic balances, tempi, and what one might broadly term punctuation or formal articulation. Instances of each of these follow. The beginning of the second Transition section (m 178) involves a "shriek" on the tam-tam that is followed by, and must be balanced with a series of descending violin harmonics (and, later, piccolo) relative to the remainder of the ensemble. The loudness level and tonal purity of such shrieks is relative to the nature of the individual tam-tam. Some produce very assertive sounds, while others are unreliable. I wanted this sound to color the beginning of the transition but not to be so strong as to seem raucous and distracting. A succession of violin harmonics needs to be played with considerable force in order that they sound clearly above the ensemble. Further, there was a tendency for the level of the ensemble, particularly the percussion and brass (and especially in S), to grow too quickly, so that the conductor needed to control the overall evolution. The specific dynamic markings in the score and parts, then, became relativistic.
It was necessary to slow Theme 5 from the indicated MM 90 in the ensemble version particularly. When a line is passed from one instrument to another, the information content of the material is raised; one needs more "processing time", I think, to hear the components of a lyrical line if they are performed by a timbrically varying series of instruments than would be required when the same materials are performed by the timbrically uniform piano.
The issue of punctuation sometimes involves several factors at once. The last section of Theme 4 presented two instances of this sort. Firstly, at measures 305-307, the strong rise from the deepest bass to the highest register is very forceful. I had included in the score a poco ritard indication over it. In rehearsal, it became clear that a very delicate balance was necessary between allowing, on the one hand, the energy of this culminating gesture to be released (before the final phrase begins), and, on the other, breaking the continuity so decisively that the final phrase seemed cut off from the body of the theme, and its overall shape became confused. Although many approaches were tried in the Paris rehearsals, the timing never satisfied me. I also realized that the 9-second-long pause after measure 309 could not be empty of resonance. (This similarly disrupted the clear articulation of formal structure.) It was necessary, therefore, to add the parallel piano version to the ensemble version so that an appropriate resonance carried over into the pause.
Although it is faintly metaphysical to say so, I often wonder about the level of responsibility/credit a composer can appropriately take in relation to works that s/he has composed. Naturally, it is the case that everything that is happening has been to a large degree elicited by the notation that I use to represent my intent. But music appears to me to have an existence that is external to and larger than the prescriptions a composer makes for any work. It is as though one is luring into specific manifestation a phenomenon already in existence. As rehearsals progress, then, two processes are underway at the same time: one listens for the structure of sound that one has created to begin to come into focus. At the same time one listens for the emergence of "music" itself.
In rehearsal, it was particularly interesting to compare the effects of TR2 to 4 (beginning at m. 178) in its two versions. Although the Sectional version is evidently more complex, richer in terms of the variety of its simultaneously evolving musical threads, the solo piano version is more cohesive and strongly directed in its immediacy and processual cohesiveness. The materials of the two are closely related, the position and function of the formal elements are the same, but their emotional charge and character are markedly different. One feels the variegation of the musical fabric in S (the accumulation of the woodwind parts from m. 203 to m. 212, and again from m. 224 to m. 257, for example) strongly, but the impression left seems more formal or "distant" than the parallel passage as performed in the solo piano version in D. In the latter case, one is led and participates more directly in the arrival at Theme 4. It was the force of this arrival that made me decide to allow the ensemble to take control of the last stages of the theme at m. 302. (Although the idea of a shared treatment of Theme 4 was a feature of the early overall plan, the S version did not, in fact, develop in this way.) In both cases, the Transition feels broad and cumulatively forceful, but the nature of that force is strikingly different.
Another rather sensitive point was the shaping of the inner phrases of the Repetitive Strata section (substituted for the beginning of what had originally been planned as COMB1/2/3). The first three subphrases — patterns of 5, 5, and 3 3/4 measures, ending with one of 3/8 — each ended with a comma. But it became clear in rehearsals that the irregularity of temporal structure and the complexity of the cross-accenting undercut the clarity of the iterating accumulations. In order to focus attention on this aspect of the passage, it was necessary for the conductor to add an acceleration to each of the initial subphrases as well as to observe the comma.
One instance of tempo adjustment was just given in relation to Theme 5. At an earlier stage of work on Angel, however, a more crucial instance arose in this regard. I had conceived Other to be performed at MM 150, and had sent the piano versions of the five themes plus Other to pianist Jean-Marie Cottet before traveling to Paris for recording sessions. Although M. Cottet had forwarded several small questions, it was my impression that he would have no undue difficulties performing them as notated. When we began to work on Other, however, he said quite firmly that this material could not be performed at the tempo I had indicated. But the necessity of performing this element at such a markedly different tempo (MM 92) created serious issues in relation to the temporal proportions of the work as planned.
I had no choice but to urge him to play at the maximum possible tempo and to accept the fact that I would have to accommodate this unanticipated reality. Because of Other's ostinatic overlays, there was no possibility of reducing its duration: the planned cyclic structure had to be played out even if it took longer than had been anticipated. I considered the possibility of radically compressing the duration of the recording with phase vocoding. But, in view of the very fine renderings that M. Cottet had done of the five thematic elements, it seemed highly unlikely that he (or other pianists) could be expected to greatly improve upon the speed at which he found it possible to perform this section in the recording session. I made the decision to adjust his recording (less radically and more practically) so as to allow it to conform to the next highest durational value in the logarithmic series controlling the proportional aspect of the piece: Other was expanded from 99.5 seconds to 161.
As indicated, composing nearly identical versions of the thematic elements for solo piano and chamber ensemble was sometimes challenging. The capacity that an individual musician has for subtle temporal coordination and nuance cannot be matched by an ensemble of 16 musicians. The core element of Theme 5 (T5.5 at m. 374) is a telling instance of the difference. For the solo pianist (in D), the task is clear: a series of chords, variously voiced is the basis of the passage, but the primary verticalities are preceded by a leisurely grace note, marked with a tenuto accent. There is a chorale of sorts and each component chord is given harmonic weighting by the fact that one of its members occurs early, as a slightly emphasized anacrusis. In orchestrating the ensemble version, I orchestrated the chords so that each would draw differently on the available timbric diversity. Thus, there was an issue of coherence and balance: a low-register bass clarinet might have to balance with a high register piccolo (as at m. 376 of S). The practical impact of register on each instrument's realization of a dynamic level is a subtle thing. If a passage continues a particular instrumental voicing over some period of time, one expects the players will adapt to the situation, and the chords will gain in cohesiveness and consistency as a result. But when the voicing changes from chord to chord, it is difficult to achieve such balance. Each instrument speaks at a different rate of responsiveness, and this too is register dependent. Added to the problem of consistent chordal articulation, was the element of the anacrusic grace notes, again, assigned to different instruments. (This is a problem that could have been avoided had the grace notes all fallen in a restricted register and I had decided that they were to be performed by the same musician. But the piano version of Theme 5's core had explored register as one of its essential features, and this facilitation was not desirable.)