But my biggest problem with his suggestion is just a practical one.  Much of the point of education is to prepare students to be able to learn from and react to "knowledge" they encounter outside of school.  Since the discussion of "public goods" is very prevalent in economics and policy research, students should be taught the meaning of the term.  A better proposal, similar to Caplan’s, would be to teach students that "non-rival" = "zero marginal cost" and that "non-excludable" = "positive externalities".  That way, they will still know what people mean when they say "public good", but they will have a fuller understanding of that class of goods, and be able to understand better how they differ and relate to "regular" private goods.

There’s a word for people who don’t want the government to do anything, not even define and enforce property rights. The word is “anarchist”, not “libertarian”.

For what it’s worth, Caplan is a self-professed anarchist.  See, for instance, this article of his on the private provision of law.  I don’t agree with his conclusions, but I do think he has at least thought through a number of the objections Dean has raised. 

Caplan may also call himself libertarian – I’m not sure.  Often libertarians include both anarchism and "minarchism" as sub-classes of libertarianism.  By the way, another famous libertarian-anarchist is David Friedman, Milton’s son.

I think what you’re thinking of as a public good is any good which is owned by the public (either formally owned by the government; or no formal ownership, leaving usage open to any member of the public), which makes sense given the literal implicaitons of the termand is consistant with colloquial usage of the term. The economic term for a good owned by the public is a "commons" (as in "The Tragedy of the Commons"). It gets more confusing, since there’s considerable overlap between the concepts of commons and public goods.

Anyway, I think I myself am having a little trouble with the meaning of the term "public good" which may be part of why I’m arguing, and may also relate to the contention that we should use different words because "public good" is confusing even to people who’ve read up on economics.

The problem I was examining in my original post was how content winds up being broadcast on the airwaves once broadcast rights are assigned. We can examine that question narrowly, independently of the question of how to allocate broadcast rights (whether by FCC-style licensing, by auctioning or homesteading property rights to broadcast bands, or through another method).

So let’s say that in recognition of my services to mankind, I am created the Baron of 42.1 FM. I have the exclusive right to broadcast on 42.1 FM in the United States, but do I want to? Buying or making content takes time and effort, and broadcasting equipment isn’t free. I need some revenue source in order to make broadcasting economically viable, and that’s a tricky issue because having content broadcasts on 42.1 FM is a public good.

Maybe I can get the government to tax every potential listener to pay my expenses — this is the classic solution to the public goods problem.

Maybe there’s someone whose private benefits are so high that he’s willing to pay the entire costs himself and let everyone else freeload. E.g. Warren Buffett pays me $1 million/year to play his favorite songs 24/7 so he doesn’t have to worry about losing his iPod.

Maybe I’m a philanthropist who is willing to subsidize the effort out of the goodness of my heart.

Or maybe I figure out a clever way to sell a private good associated that piggybacks on the broadcast (such as advertising) or to make the broadcast excludable (through encryption).

My point being that while government provision of public goods is a widely accepted solution to the public goods problem, and it’s often the best solution, it isn’t necessarily the only viable solution when the public goods problem crops up.