If you were trying to right a capsized business -- let's say a book company -- you wouldn't ask, in an ideal world, how much of the gross national product should be spent on books. Or how profitable the book business should be. Or whether, in the abstract, you should print only books on public safety, or only books that people read in 1789.
Instead, you would ask: Is there a market for such books? Even if people bought them in 1789, are they still relevant to anyone today? Have new technologies, like e-books, fundamentally changed the nature of what I can and should provide? Do competing offerings like movies or the Internet affect my business strategy and how much I should invest in it? Do these changes mean that I should give up on the book business entirely? Or do they open up new opportunities if I simply think differently about what it means to be in the "book business"?
We tend not to think about government in this way. After all, we all know what "government" is, has always been, and should be. But, in fact, what government does and has done has varied widely over time and space. We think of the government providing public safety, or at least meting out punishment for criminal offenses -- but it hasn't always done so. The modern welfare state only emerged in the late 19th century, and, in fact, the state itself is a relatively recent invention. And it simply isn't true that governments only increase the size and extent of their activities over time: Many governments used to develop and impose religious beliefs, determine the disposition of the labor force, or decree what clothes people could wear.