One of the first things I noticed in learning Palais de mari is that I had to count constantly. The music sounds so spacious with all those isolated phrases floating in the air, but underneath that I was going “1 2 3 4 5, 1 2 3 4 5” like a madman. The fact is that rhythm is deceptively hard in Feldman. Things happen slowly, but they vary by such small degrees that you really have to pay attention to play the music accurately.
Here’s a quick inventory of what happens metrically in Palais de mari. The bulk of the piece is in 5/8. The grouping used is sometimes 3+2, sometimes 2+3, so you have to keep track of the distinction between quarter notes and dotted quarters. Thrown into this are a number of 3/4 bars, which are usually either dotted halves or (less often) pairs of dotted quarters, so I think of these as 3+3 whenever they occur. There are a number of 2/2 bars, which are almost always just whole notes, but I think of them as 3+3+2, just to keep the 3-and-2-ness going in my mind. There are a few other meters scattered here and there: bars of 1/2, 3/8, 5/4, and, just to keep you on your toes, a couple of bars of 3/16. The last two pages have their own metrical world that I’ll describe shortly. The point here is that the piece moves mostly in quarters, dotted quarters, dotted half notes, and whole notes, and that these can be thought of in the framework of various combinations of 3s and 2s.
In the beginning, I counted all of this, as I described above. Now that I’m getting more comfortable with the piece, I’m thinking in terms of the quarter/dotted quarter pulse without subdivision as much as possible. It’s much easier on the brain and more relaxed in effect. But my score is heavily annotated with numbers indicating the various counts of eighth notes, partly because it can be difficult to read Feldman’s manuscript notation at a glance (at least for me). But even then, the effect of the rhythms shifting constantly around the basic quarter and dotted quarter units forces you to really pay attention carefully to what you’re doing. The piece is too repetitive over too long a period of time for you to just memorize the rhythms and go on autopilot. You stay more present from measure to measure in the piece because of these subtle shifts that you need to keep on top of.
The ending of the piece does an odd metrical thing that I’m still getting worked out. Unlike the rest of piece so far, here Feldman has his repeated ideas start on the second eighth note of the bar, at least some of the time:
Personally, I see this as similar to those passages in Schumann’s piano music where it sounds like the downbeat, but it’s really the upbeat, or (as here) the second beat (a great example of this is the “Paganini” piece in Carnaval). In these places in Palais de mari, I count the rhythms as if the main notes are on the downbeat, and I treat the rest as the last beat of the prior bar. So in the example above, I would count the first bar as eight (3+3+2), then the second bar as nine (3+3+3), thus adding the eighth rest to that bar. I then treat the C# in the third bar as the downbeat of a 7/8 bar, including the next bar’s opening rest as the end of that measure, and so on. This may be cheating, but I’m not sure that I think there’s an appreciable difference in sound, so it works for me.
This example shows the other rhythmic quirk that appears towards the end. Feldman shifts occasionally between grace notes (as in the first five bars of the example) and full-fledged eighth notes (as in the last two bars). In these cases, I treat these eighth notes as the upbeat. Indeed, these written-out slow grace notes fill the same function as the leading eighth rests. There are no bars in which both features appear: it is always either one or the other. So again, I just add the eighth note to the bar before it when I count it out. So the D in the fifth bar is the downbeat of a 7/8 bar that ends with the F-natural leading to the E-flat on the downbeat of the next 7/8 bar.
How much does a listener actually hear all this rhythmical and metrical complexity? I’m not sure that one hears it directly in a lot of the piece. Certainly one is aware of small changes in rhythms, of patterns appearing slightly more quickly or slightly more slowly. I think the changing around between 2+3 and 3+2, etc. lends to the “fuzzy” effect that the music can have for us, making the patterns recognizable as something repeated, but yet not mechanical or sharp-edged (as they might seem with classic “minimalists”). And I would not discount the energy that is created by the attention of the performer. All that counting keeps me on a kind of edge, and I’d like to think that a listener is aware of that, too. It keeps the piece from going soft and flabby, even as the sound is quiet and pliable.