Recent trends in cognitive science have not made scientific literacy easier to attain, but they have made the practices through which educators meet its challenges more interpretable. Traditionally, cognitive scientists viewed knowledge as a set of propositions comprised of classical concepts, thought as logical inference and language as a literal representation of the world. They attributed the same denotative characteristics to cognition as to science text. In contrast, many contemporary cognitive scientists view knowledge as comprised of fuzzy and contextual concepts, thought as perceptually rather than formally grounded, and language as largely metaphorical and narrative. In this view the expressive characteristics of cognitive representations differ from the relatively denotative characteristics of science texts. Science literacy education bridges this difference through practices that help students: build classical concepts on a fuzzy cognitive architecture, achieve formally valid reasoning using perceptually driven operations, and construct written explanations and arguments using speech-like and narrative language.