Professional physicists quickly learn the power and value of mathematical representations, not only as calculational tools, but as ways to organize conceptual knowledge and reason about physical situations. Often, this is because we started out with enjoyment and success in math and were enthralled by the idea that this beautiful stuff could actually be used to describe the world. Many of our students (especially those in service courses) don't come to physics with this orientation about math. An analysis of epistemological resources and stances chosen by physics faculty and students suggests that including math in our classes in the way most comfortable and natural to us as physicists might not help our students learn to use math in science. A more productive approach might be to run the math "upside down" by beginning with first building a strong physical intuition and then helping students translate to rigorous math.