Friedman came to Yale once and gave a talk called "Yale versus Chicago in Monetary Theory" before a house of 500 people. [...] It was quite interesting. I didn't get much involved at all in public, but we had a small private session afterwards. The thing I remember most about the occasion was that there was a very earnest, well-meaning graduate student who stood up at the big meeting and asked Friedman politely: "In your mode, money is the basic concept, and yet, you haven't ever told us exactly what money is conceptually. Could you help us understand it now?" Friedman cut the guy down in the withering way he can do by telling him that he didn't understand scientific methods. He said Newton didn't have to tell what gravity was; he only had to tell what it does. The same applied to money. That illustrates Friedman's methodology of positive economics which I think has done great damage. [...] You see that in Lucas, too. Their idea is the as-if methodology in which it is not a question whether the assumptions are realistic, but whether the results derived from the assumptions are consonant with the facts of observation. My reaction is that we are not so good at testing hypotheses so that we can give up any information we have at whatever stage of the argument. The realism of assumptions does matter. Any evidence you have on that, either casual or empirical, is relevant.
Milton Friedman