IME most of the classes you take in grad school are irrelevant to your research. Maybe 2 will be useful, 3-5 if you're lucky. That's basically 1-2 semesters of

IME most of the classes you take in grad school are irrelevant to your research. Maybe 2 will be useful, 3-5 if you're lucky. That's basically 1-2 semesters of learning compared to 3-6 years doing research for a PhD. My point is that the coupling of teaching and research isn't necessary, teaching could be covered in a 4+1 undergrad/masters. I've personally had professors from each combination of (good at teaching) x (good at research), and I think the (good,bad) is more common than (good,good). Coupling of research and university in order to protect research against financial/political shocks does seem to have some merit. But I think we have to examine the costs. For example, in my department one of the professors has a poster that reads what is expected of their time commitment: 30% of their time devoted to service, 70% to teaching, and 100% to research. If the reaserch is that important to us, shouldn't it be done by someone who can spend all their time thinking about it?