I signed a lot of study cards last week, not just because I am Director of Undergraduate Studies and have picked up a lot of advisees this term from faculty on leave, but because it seemed like every other student who came in had two or three study cards for me to sign. They listed various combinations of courses, and while a few students just didn't want to decide until the last possible moment, most of them were waiting for the results of course lotteries – the process by which limited enrollment courses are whittled down to size when too many students want to take them.
Limited enrollment courses have been around forever. Seminars obviously have to be capped, because they wouldn't be seminars otherwise. I once decided to lift the cap on my freshman seminar from 12 to 15, because about 60 students wanted to get in. It was a big mistake. Talking in a group of 13 is different from talking in a group of 16.
Courses with unusual equipment or classroom demands sometimes have to be capped. Again, film courses have long been capped, because cameras and whatnot are costly.
As we move away from lecturing to flat-floor classroom teaching, the shortage of appropriate teaching spaces creates capacity problems. These are exaggerated, in the case of Computer Science, by huge enrollment swings in our direction. Our teaching facilities were not designed for the numbers of concentrators we have, and it is awfully costly to renovate space even if the space exists, which it doesn't. I know this is a problem in the lab sciences as well. It is a fair question whether the earnest efforts we have gone to in order to increase enrollments were matched by earnest planning about what would happen if we succeeded.
In some fields, enrollment caps are a response to a shortage of TFs. CS has this problem for graduate courses (yes, some graduate coursed draw multiple dozens of students). Most undergraduate courses use many talented undergraduate course assistants, a resource which happily grows and shrinks in almost direct proportion to the enrollments in the courses. In humanistic fields that doesn't work, and even a small reduction in the number of PhD students makes the TF shortage even more acute.
All that said, I can't ever remember a year when I had such a strong impression that lotteries are now accepted as a part of ordinary undergraduate life at Harvard. The difference is that a lot of the courses being lotteried are big courses, not small ones. They are General Education courses, not boutique courses on specialties that have suddenly become trendy.
The Crimson reports on this a bit today, listing some of the lotteried courses, and also noting, importantly, that not all "lotteries" are run the way the Numbers Game is run. What is politely termed a lottery is actually a second admissions process, as though, having gotten through the eye of the Admissions Committee's needle and made your way to Harvard, you should have to prove to a professor that you are worthy of enrolling in his or her Gen Ed course.
This seems crazy, and wrong. We got rid of limited-enrollment concentrations some years ago, on the theory that anybody should be able to get an education in any concentration [major] at Harvard, if they were willing to work at it. And now we have filters on Gen Ed courses instead? This is what Harvard students should expect for their $60,000 per year? Sounds more like Berkeley.
I wish I understood how this happened. In the old days, Bill Bossert used to repeat his lectures in Nat Sci 110, a classic Gen Ed course on computer science, because there was no lecture hall big enough to hold the whole class. That was extreme, but I wonder if there has not been an erosion of an old feeling that teaching was a mission, not just a job, and if a lot of students wanted to learn from you, it was an honor rather than a nuisance. One hears that some courses are limited just because … the professor doesn't like teaching big courses. That seems wrong.
It seems to me a pretty simple expedient would solve most problems of over enrollment in large courses. Teach them at 9am! That would dis-enroll the gut-seekers and schedule-fitters, and would leave the professor with only the most committed students to teach. I really don't know why my colleagues don't try this -- in fact why not make it a rule? Any course lotteried one year has to be offered at 9am the following year. Bingo. Many fewer lotteries.