According to the CoHE (sorry, reg. req.), Princeton University has suggested imposing limits on the number of "A"s a professor can give in any given undergraduate course. This strikes me as not, perhaps, the best way to go about addressing the problem of grade inflation. The result would be like a grading curve, but without any actual curving--just a cutoff point. As John Bruce reminds us (permalink bloggered--scroll down), grading curves can provide a real disincentive to academic achievement, since "[g]rading on a curve [. . . ] is a sucker's game. There's no way you can put in a 'reasonable' effort and get a predictable result, because you're being set up against the wannabe burnout cases. Smarter students are going to have this figured out, and aren't going to become burnout cases themselves. Thus they'll be farther down the curve." It's not clear how an imposed cutoff point would make things any better. If in a class of 25 students you have 10 who do "A" work--this shouldn't happen, by the way, but let's say that they're all geniuses--then one "A" student will have to be arbitrarily bumped down to a "B" in order to make quotas. How do you tell which student to bump? And in what sense can the instructor defend his or her decision on the basis of simple fairness? Indeed, I can easily see this proposed cutoff generating more administrative headaches, since some grading decisions could not be justified on any grounds other than "well, I hit my quota."
Now, that being said, it's not clear why so many Princeton students are getting "A"s in the first place--a question I've asked before.