I was writing a simple teaching post, on ideas and increasing returns to scale, in micro and macro. I wrote down "Ideas are non-rival". Then I thought I had better explain what I meant by that. Then I thought about professors, who do research (thinking up new ideas), and teaching (communicating existing ideas to other people). Then I thought about how some professors like research but don't like teaching. Then I thought about this post.

Sure, two people can use the same idea (ideas are non-rival), but can't eat the same apple (apples are rival). But the second person can't use that idea unless the first person communicates that idea to the second person. The first has to teach it, and the second has to learn it, and teaching and learning are (sometimes) costly. The cost of communicating the idea to the second person might even be greater than the cost of the first person coming up with the new idea in the first place. Sometimes it might be cheaper to reinvent the wheel than walk to the library.

And the marginal cost of communicating the idea to the n'th person very probably will be greater than the cost to the first person of figuring out the idea, if n is large enough. Teaching and learning can be very hard and very costly at the margin, if you push the margin out far enough, and try to teach a difficult idea to everyone. If you put "number of people who use the idea" on the horizontal axis, starting with those who are easiest to teach and who find it easiest to learn, that marginal cost curve is going to slope up, eventually. Will it always eventually cross the average total cost curve, at some point? And does this matter, for how we teach the economics of ideas? But is it really the same good, as we move along that MC curve? And can you separate the activities of research and teaching (communicating research)? Professors love to argue that you can't. But then they would say that, wouldn't they?

Maybe Frances is right.

Dammit, now I've confused myself. I'm off to the cottage. Have fun.