The issue of whether the assumptions made in economics or any other science are "reasonable" depends upon one's philosophy of science. If "reasonable" is taken to mean realistic or plausible then the answer is no. Almost all models in Econ are set up as constrained maximization problems, and it's obvious that people don't figure out lagrangeans in their heads when they're deciding how many apples to buy at the market. Most economists and physicists fall into the instrumentalist/operationalist camp in the philosophy of science. This is the belief that theories are neither true nor false, but simply tools (i.e., instruments) used to make predictions about reality. Under this view, any assumption is reasonable as long as it allows one to make correct predictions. e.g., concave indifference curves predict that people will spend all of their income on one good, and we generally don't see that in real life and so concavity is an unreasonable assumption. The issue of whether an assumption is ad hoc under the instrumentalist worldview is mostly an aesthetic criterion. Since one can assume whatever one likes in order to generate predictions, there is no way to judge whether the assumption is ad hoc. Every discipline develops its own conventions about things like what constitutes "good" modeling techniques and what level of statistical significance is sufficient for counting something as a correct prediction. I think where economists and other instrumentalists get into trouble is when they decide that because their model makes correct predictions then the assumptions that they used to generate those predictions must be "reasonable" in the sense of representing reality. This is called the pragmatic conception of truth-- if it works it must be true. This is when you get economists arguing absurd things like that people really, truly are rational maximizers or can actually form rational expectations.
Daljit Dhadwal > I guess what I was origially wondering was not "Is math appropriate?" > But rather, given that math is appropriate, are the > assumptions about rationality, preferences, convexity, > continuity, inter alia, as well as some of the more ad > hoc assumptions of many macro models as reasonable as > the assumtions made in the king of all sciences?
