Journal article: In an environment with privately informed agents who can produce evidence, we study implementation of a social choice function by reading mechanisms: mechanisms that simply apply the social choice function to a consistent interpretation of the evidence. We provide sufficient conditions on the social choice function and the evidence structure for ex post implementability by such mechanisms. If the first-best policy of a mechanism designer satisfies this condition, then its implementation by a reading mechanism does not require commitment. We show that with rich evidence structures, (1) a function that is implementable with transfers is also implementable with evidence but no transfer, (2) under private value, the efficient allocation is implementable with budget balanced and individually rational transfers, and (3) in single-object auction and bilateral trade environments with interdependent values, the efficient allocation is implementable with budget balanced and individually rational transfers.