Ruth Millikan and Alvin Plantinga claim, roughly, that knowledge is true belief produced by processes in circumstances for which they are (successfully) designed to yield truth. Neither offers the account as a conceptual analysis of knowledge. Instead, for Plantinga it represents the core concept of knowledge characterizing central cases, and for Millikan an empirically warranted theoretical definition of knowledge as a natural phenomnenon. Counterexamples are then dismissed as appropriately called "knowledge" only in some analogically extended sense. I argue instead that a definition of knowledge is better thought of, like an account of justice, as an explication, or a decision about what normative constraints to adopt. It reflects trade-offs in intuitive judgements and pragmatic concerns, and perhaps background theories, for the sake of some overall coherence. However the contentious scientific and historical claims that underwrite their views gives us reason not to accept the design account's norms for knowledge. Further, the controversial character of these claims gives us reason not to rest an interpersonally acceptable account of knowledge on them.