Stuart Buck points out that the traditional view of teaching v. scholarship (that good scholars make for good teachers) is likely not true (or likely not often true) but that teaching might make for better scholarship:
No matter how intensely you study a particular subject, if time goes by without regular review, it’s easy for the details to slip from your memory. But teaching a course inherently requires regular review — not just of your own scholarship on a given subject, but of everything else that is relevant to that subject. If you’re going to stand in front of a group of people and explain a particular legal subject, you have to know the ins and outs of all the important cases/statutes/commentary. It’s not enough to know this stuff “on paper” — you have to know it stone cold, so that you can answer practically any question that students might throw your way.
What’s more, you have to know the subject well enough to explain it to beginners. I think that this requires more in-depth knowledge than merely being able to converse with other “experts.” When you’re talking to beginners, you have to understand the topic well enough to boil it down to the basics. You can’t get away with casually genuflecting in the direction of some abstraction on the assumption that everyone else will know what you’re talking about.
I think there’s something to this. At least for political theory (my field), the best scholars are those who can work their way through a problem and bring to bear a wide range of analytical tools and concepts. There’s a place, of course, for the detailed study of what Locke thought about parental relationships, but if you’re trying to think about how we ought to understand (and capture in law) such relationships, you’re better off if you can draw from a wider range of thinkers and histories. Scholars who teach widely seem to me more likely to do that.
That said, it’s still worth noting that when research universities have great teachers it’s because they want to be great teachers, not because there’s any particular incentive in that direction. For most, teaching is something they “get through” in order to do their “work” (i.e. research and writing). Most try to find ways to minimize the amount of work they have to do for the classes (Ever had a course where the students were doing presentations the last four weeks?) and the better scholars get rewarded by having to teach fewer classes. Until the incentives change, teaching will always be a sidelight, not the main event, at research schools.