In economics, public goods are ones that are (a) non-excludable and (b) non-rival. Non-excludable means that it is impossible to stop someone from enjoying the public good. For example, radio broadcasts can be picked up by anyone with a receiver. Non-rival means that one person's consumption of the good does not limit the amount available for others. For example, your consumption of this blog post does not in any way reduce the amount of blog post available for others to enjoy.
As I have argued elsewhere, there are few examples of pure public goods. Radio signals, for example, can be encrypted, and if a million people all tried to read WCI simultaneously the website would crash. So authors of public economics textbooks tie themselves in knots trying to think of examples of public goods other than lighthouses, national defence, and fireworks.
Is firefighting a public good? Surely it is, because even though, in theory, the firefighters could choose not to put out your fire, they never would, would they?
Actually, they might. In Obion County, Tennessee, firefighters stood by as Gene Granick's house burnt to the ground, because he had not paid his $75 contribution to the cost of firefighting.
There are two lessons from this story. One is about the nature of excludability. The ability to exclude is not absolute. It depends upon technology in the conventional sense - for example, the availability of encryption technology for radio signals.
Social technology matters too.
"I thought they'd come out and put it out, even if you hadn't paid your $75, but I was wrong," said Gene Cranick.
Mr Cranick thought that social norms made exclusion impossible. But social norms change.
The second lesson is a point I have made before. The public good concept is not useful. There are few goods that are both non-rival and non-excludable. Confounding the two concepts (non-rivalry and non-excludability) prevents cogent analysis of each individual concept. Moreover the "public" in public goods confusingly suggests that public goods are, in some sense, publicly provided - yet there is nothing whatsoever in the conventional economists definition of public good to suggest that this should be the case.
(Thanks to Bruce Foster for finding this gem.)