An interesting post over at Renewable Music, addressing just how much repertoire a performer can really know well. In Daniel Wolf's experience as an ethnomusicologist, the upper limit that a performer or ensemble can know really well is about fifty pieces. If you want to have more pieces than that on the go, you find that some of your older repertoire starts to slip.
The whole discussion, of course, was sparked by a post by Kenneth Woods about his own repertoire for the year as a cellist and orchestral conductor. As an organist, of course, I live in a completely different world entirely - where an orchestral conductor's repertoire is measured in fairly large units (overtures, symphonies, concerti), my organ repertoire is measured in Preludes and Postludes. In other words, the music I play is largely assembled to fill approximate five-minute chunks before and after worship services. I've posted below on my dissatisfaction with this custom - almost no-one pays attention to what you're doing, it becomes difficult to differentiate service playing and recital playing, and eventually you find that your entire repertoire consists of spastic, fortissimo toccatas and soupy French pieces for the Voix Celeste. You see the effects of this custom at recitals - people program individual movements of sonatas which suspiciously always turn out to be five minutes long and either unvaryingly loud or soft. Even when you hear a complete work, say Durufle's op. 5 Suite, you can be pretty sure that the performer diced it into voluntary-sized pieces before performing it complete. Hey, look - the Prelude and Sicilienne could work before the service, and that Toccata would be a really impressive postlude!
This is why I like to learn enormous twentieth-century works for recital purposes - it's hard to use a piece like Malcolm Williamson's Symphony in a liturgical situation, so playing it in a recital becomes a special event rather than a rehash of my service voluntaries for the past three weeks. (This argument is also circular - I play huge twentieth-century works, which are impractical for service playing, so no-one else takes them up, so playing them makes me feel unique, so I perform them more often.) On the other side of the coin, however, I love to mess with people's expectations for service music - every Palm Sunday, I play the congregation in with something fairly raucous (Langlais's Les Rameaux, last year), and send them out with Bach's O Mensch, bewein. It's perfect for this somewhat schizophrenic service, which begins with a great procession and ends by leading us into Holy Week. Yet every year people think this is a mistake ("Why were you playing so loudly before the service? Are you practicing for Easter?"), and I try to explain to them what the service music was intended to communicate. Unfortunately, the Christmas services that I play at this time of year are the epitome of "playing to the gallery", and so to satisfy their expectations you end up playing things like Mulet's Carillon-Sortie, which is great fun, but no-one would ever accuse it of being profound.
My issues with the prelude/postlude complex aside, though, I can't help feeling that an estimate of 50 pieces is probably right in terms of how much repertoire I have ready to play at any given time. Certainly I'd play more than that in the course of a year (fifty-two Sundays makes an average of one hundred and four Sunday service voluntaries, for example, and that doesn't include recital repertoire, or mid-week services for the major festivals), but most of the repertoire I've scheduled for the next month needs a fair bit of practice before I can get it up. Which has nothing to do with difficulty (I've been playing Dieu parmi nous a lot recently, so playing it again on Christmas Eve won't be a problem); rather, it seems to be linked to what sort of style I've been playing recently. If I do a lot of Romantic music for a few Sundays, it's difficult to get back into Bach playing at the standard that I'd like, or if I've been taking a break from Messiaen for a while, I wouldn't want to perform one of his works without a week or so to get back into his idiom.
To sum up: we organists have very different performing circumstances than the average musician, but much the same limits apply to us as to everyone else. Even the amazing repertoire feats you occasionally hear about - organists playing the complete works of Messiaen, or Buxtehude - don't change the basic repertoire limit. After all, Messiaen and Buxtehude both have certain formulae in their writing which make it much easier to learn their music once you have a few major pieces under your belt.