The expectation that science is a stable body of relatively objective knowledge on which the law can draw to settle legal controversies may seem benign. However, this expectation often corresponds to a romantic notion of the scientific enterprise and thereby eclipses not only the instabilities and controversies within science itself, but also the social and rhetorical aspects of even the best science. We see the idealization of science in law whenever there is a presumption that if two scientific experts disagree, one of them must be a "junk scientist." This presumption ignores the theoretical presuppositions and limitations of data that lead to genuine scientific disputes. We also see the idealization of science in law whenever we associate "bias, interest, and motivation" with unreliable expertise. This association misses the practical advances made by scientists who have strong theoretical biases, institutional interests, and financial motivations. Finally, we see the idealization of science in law whenever a legislator, administrator, or judge demands certainty from science, not recognizing its probabilistic nature and dynamic history. It is neither a critique of scientific progress nor an exaggeration to acknowledge scientific debates, the conventional aspects of scientific methodology, the importance of networking and "social capital" with respect to publications and grants, and the persuasive elements in scientific discourse. To think that these features are somehow markers of bad science is to idealize science.