Announcing Canon #88, “Carminite.”
The essence of a canon is imitation: the follower copies the leader. In some canons, imitation is exact: the follower plays the same notes that the leader played moments before. In other canons, imitation involves some kind of transformation. There might be a chromatic or diatonic transposition. There might be an inversion, where an upward movement is imitated as a downward movement and vice versa. There might be a stretching or shrinking of note durations. There might even be a reversal or retrograde statement of the material. In all of these cases, imitation is deterministic: once the leader plays a note, it’s clear what note the follower must play – there’s no choice. Of course, during the composition process, liberties may be taken. The composer might decide that even though the follower is supposed to play a certain note to stay faithful to the leader’s example, the required note doesn’t sound good and a different note will be chosen. These exceptions to the imitation scheme are precisely that: exceptions and not the norm. When I write canons I often agonize over these exceptions: is it better to keep the canon strict or to make alterations that diminish the formal purity of the piece while making it more musically engaging?
For a long time I’ve wanted to write a canon with a different concept of imitation: one that would be approximate or flexible instead of prescriptive or fully determined. There are lots of ways you could flesh out what “approximate imitation” means. The idea I had in mind was that interval direction would be preserved while interval size could be changed arbitrarily; meanwhile, rhythm would be preserved exactly. That’s to say, if the leader plays two quarter notes that move up by a major third, the follower would have to play two quarter notes as well, and they would have to move up as well, but they could move up by a minor second or a perfect fifth or any other interval.
Writing a piece like this, I hoped, would help me address a question I’ve been thinking about for years: what makes a canon sound like a canon? What formal elements are essential to creating the experience that we seek when we listen to a canon? Is the strictness of the imitation important, or is it dispensable? Is rhythmic imitation more important, or is the replication of melodic contour more important? To shed any light on this, a piece employing approximate imitation would have to be more than a canon that takes liberties here and there – many canons already do this. It would have to be a canon where the leader’s intervals are very frequently changed by the follower, not in a predictable way, like always cutting them in half (see Interval Compression), but in a free and unpredictable way.
One might expect that the idea of approximate imitation would make a canon much easier to write: more flexibility, fewer constraints. So, after months of attempting it, I was perplexed at my inability to get such a canon working. The project, in fact, seemed much harder than writing a strict canon. Why should this be so? The reason is what I alluded to above: approximate imitation introduces a new burden. To really showcase this concept, you have make the follower regularly – not just occasionally – vary the leader’s intervals. So it’s like you’re writing two melodies at once. You do all the work to make the leader’s line sound good, and then, instead of having the follower copy the leader’s already-good-sounding line, you have to make the follower do something different, but still similar, and this different-but-similar thing also has to sound good. That’s hard.
In Canon 88, I’ve finally been able to create an example of approximate imitation. What made it possible this time around? I’m not sure there was any particular trick, although I did make the early decision that this would be a dissonant canon without a clear tonal center. I spent more than a week working on possible outlines, and then another week trying to develop them: two false starts before something finally came together. The false starts were interesting in their own way. I think what went wrong with both is that my melodic developments on top of the non-tonal outline started to become too tonally suggestive. I was beginning to turn a non-tonal skeleton into a tonal piece, but this was a lost cause. Each possible tonal focus was hinted but not maintained long enough to give any satisfaction. As one possible tonal center was abandoned and another suggested, it created the disappointing, sloppy variety of chaos, rather than that very different variety I seek: the one that is intriguing, potent, and structured. In the final attempt, the one that worked, I managed to create lines that were less tonally binding, so they played better with the non-tonal outline, not constantly working up an expectation and then disappointing it.
This canon also explores a V-shaped registral contour I had been wanting to execute in a canon for some time. Both voices start very high and gradually descend until they’re both very low. At the midpoint, they turn around and begin ascending till they return to their high starting range.
In the first section, the lower voice leads. At the midpoint, the upper voice takes over as leader.
The two sections of the piece have different characters: the first is more whimsical and skittish; the second is more patterned and goal-oriented.
The second part of the piece is built from a retrograde or reversed version of the outline I used for the first part. So, the bones of the piece exhibit a kind of mirror structure around the midpoint, but the details in each half are completely different.
As the canon uses approximate imitation, of course, there’s no specific “interval of imitation” used here.
Subjectively, the piece brings to my mind a kind of cartoon dance or chase: fast, erratic, and maybe a bit comical.
The choice of name, “Carminite,” is not particularly significant in this case, though I liked its sound and I noted that Carminite is a red mineral. Sometimes pieces bring to mind certain colors. Red seemed like a good match in this case but to try to explain why, I’d have to be more inventive than I’m feeling right now.
So, does this piece answer any of my questions about the nature of canons? It leads me closer to the conclusion that strictness is non-essential. Approximate imitation sounds, or can be made to sound, very much like classic, strict imitation. Perhaps proving this was my unstated intention when writing the piece. While I set out to showcase approximate imitation as something distinct from strict imitation, my impulse was to reign it in and apply it in a very controlled way. The ear can be guided into ignoring countless discrepancies between leader and follower when more important similarities call for its attention. It’s possible that as I wrote the piece, I tended to put the discrepancies in places that didn’t draw focus.
What also may have happened is that the more glaring discrepancies in my outline got smoothed over by later melodic elaboration. For example, the outline might have had a minor third leap in one voice imitated as, say, a tritone leap in the other – a difference that would be easily noticed. But these leaps might have been filled in as a run of three semitones in one voice versus a run of three whole steps in the other voice. Of course, a semitone run is different from a whole-tone run, but if the ear is more interested in the destination, these paths may sound quite similar: they are both sequences of “small” steps. So while there are many discrepancies between leader and follower situated all throughout the piece, they’re not as glaring as they could have been made to be.
For me, the experience of listening to Canon 88 is much the same as what I get when I listen to my previous canons, barring aesthetic differences from piece to piece. It doesn’t sound like an outlier to me. But this experiment yields no final or general conclusion about approximate imitation because the idea can be realized in so many different ways; only one of them is explored here. It’s perhaps a tame application of the concept, and yet it gave rise to what I feel is one of my freer-sounding pieces. At the moment, it’s one of my favorites. My goal now is to try using approximate imitation again, in a way that makes the discrepancies even more apparent. ■