Three decades of research in cognitive development and science education has revealed that students enter the science classroom with rich, though inaccurate, theories of everyday phenomena that interfere with learning. I will present research suggesting that these “intuitive theories" are never truly replaced by scientific theories but rather coexist with them, shaping the kinds of inferences we make, the kinds of explanations we endorse, and the kinds of information we accept as true. While adults with extensive science education are typically able to discriminate between scientific and non-scientific claims, they are slower to make those discriminations for claims that are inconsistent with their intuitive theories, and they justify the endorsement of scientific claims by appealing to intuition and authority rather than theory and evidence. Our understanding of science may thus be constrained by patterns of reasoning that emerge in childhood but persist long thereafter.