"If you can't explain it simply, you don't understand it well." - Albert Einstein
I've heard some people say before that they consider undergraduate teaching to be a necessary evil--as something one has to do in order to get paid to do research. Now, I don't know how many philosophers actually think something like this. It could be few, it could be many. What I do know is that, in my experience, nothing could be further from the truth. My experience has been that undergraduate teaching--in addition to being valuable in its own right--can be a vital part of doing research. Now, of course, this is just my experience. Still, for all that, I'd like to suggest that there are some general reasons to think that, if you really put a lot of time and energy into undergraduate teaching, it is likely to improve your research.
I always wanted to be a good teacher--the kind that inspires students the way my mentors inspired me. However, before coming to the University of Tampa, career-concerns focused my attention primarily on research. My main concern coming out of graduate school was to get a job, and because I was of the belief that research promise is a primary determinant of one's job prospects, I spent my first year out, at the University of British Columbia, focusing primarily on research. Alas, despite spending the vast majority of my time on research, I was spectacularly unproductive. I mucked around on a couple of unpromising papers for most of the year, and only published a short reply.
For family reasons, I transferred jobs the next year to the University of Tampa, where the primary expectations are teaching. Because it was made clear to me that I was being hired to be a teacher, I threw myself into my teaching. I spent most of my days developing lectures, trying out new teaching methods, grading daily (yes, daily) assignments for three courses--all of which left little time for research. At first, this concerned me a great deal. Was I selling myself short as a researcher by putting so much time and energy into teaching? Yet, this worry quickly disappeared. Much to my surprise, I found myself getting far more research done than I ever had before, and I felt like I was developing better philosophical ideas than I had before. The secret, it increasingly seemed to me, was undergraduate teaching. Why?
Two thing that undergraduate teacher challenges one to do are to set aside philosophical jargon and focus on making things simple and clear. For instance, when teaching Kant's moral/practical philosophy, it's really easy to get sidetracked by Kant's technical terminology, etc. But, when teaching a first-year undergraduate course, getting mired in that stuff is a recipe for disaster. Students tend to tune out. In order to get them to tune in, the challenge is to explain, in the simplest and most intuitive way possible, what Kant is up to, and how his theory is philosophically motivated. This, of course, isn't so easy to do--but it's great practice for writing papers. In order to get through to journal reviewers, one has to make things clear and well-motivated to them--and, as we all know, reviewers can often interpret things incorrectly, uncharitably, etc. Teaching undergraduates can help greatly in this, as they will express confusion, give uncharitable interpretations, etc., unless you can make things crystal clear to them! Now, of course, grad students can do the same thing--but there's a difference. Grad students tend to already be "in the know"--they know what you're lecturing about, and are already "in the game" (or dialectic)--in a way that undergraduates and journal reviewers may not. In other words, in a strange way, my experience has sort of been that undergraduates are in some ways more like journal reviewers than grad students.
On a related note, another thing that undergraduates tend to have are good "B.S. meters", or the ability to detect when we, professional philosophers (for all of our other virtues), may have gone of the rails. Sometimes I'll find myself lecturing on an argument that professional philosophers take super-seriously, and my undergrads will look at me like philosophers are from another planet. When it happens, I often step back and ask myself, "Why do we take that argument seriously?" Then, I tend to find, when I work through this very question with my students, we tend to come to one of two results, both of which are well-won victories! Either (A) we find a clear reason for thinking that argument really is worth taking seriously, or (B) we find that there's something really seriously amiss with the argument (one of my favorites here are standard intro-textbook "refutations" of moral relativism--none of which, thanks in no small part to my undergrad students, I now find convincing). Both are philosophical "wins." And, in my experience, oftentimes the same sort of thing doesn't happen with grad students. Grad students, in my experience, sometimes have a certain tendency to accept a certain theory or dialectic simply because it's taken seriously in the discipline. They'll say, as it were, "Everyone knows X is serious problem/theory/argument. Great philosophers A, B, C have all published on it." Undergrads, on the other hand, haven't been so indoctrinated. They're more like, "Philosophers take X seriously? Really? Why?" (Note: a long time ago, when I asked my undergraduate advisor, Dan Dennett, why he stayed at Tufts rather than work in a place with a PhD program, that was the answer he gave--that he likes the way undergrads "keep us honest" as philosophers. I feel safe sharing this because he's written similar things publicly in the introduction to his recent book Intuition Pumps and Other Tools for Thinking).
A third thing I've found incredibly valuable in undergraduate teaching is that it encourages one to repeatedly go "back to the beginning", revisiting and rethinking basics that one thought one knew really well. The book manuscripts on ethics I just finished, for instance, came from a rather unexpected place. I wasn't even doing research in Kantian ethics at that point, but I had an introduction to ethics class in which I had to teach the Groundwork for the umpteenth time. Each time I teach it, I try to look for simpler, more intuitive ways to explain what I think Kant is up to and why--and this time it led me to what I thought (and think) are (1) a couple of false assumptions at the bottom of his project that Kantians have generally accepted, which (2) when corrected, justify fundamental revisions to the theory that should push Kantian moral philosophy in a new and better direction. In other words, returning time and time again to philosophical basics in undergraduate teaching can lead one to question things that one might not have questioned if one didn't have to revisit the basics.
Now again, these are just my experiences and impressions. Maybe my experiences are unrepresentative, and perhaps my impressions are inaccurate (maybe I just want to believe that teaching has improved my research!). Who knows. What I know is this. I love teaching for its own sake, and find it incredibly gratifying to see my students grow and improve. But, in addition to that, my experience has been that undergrad teaching can inform one's research in unique, exciting and unexpected ways, and so I thought these experiences might be worth conveying! :)