Entity resolution (ER) identifies database records that refer to the same real-world entity. In practice, ER is not a one-time process, but is constantly improved as the data, schema and application are better understood. We first address the problem of keeping the ER result up-to-date when the ER logic or data "evolve" frequently. A na < ve approach that re-runs ER from scratch may not be tolerable for resolving large datasets. This paper investigates when and how we can instead exploit previous "materialized" ER results to save redundant work with evolved logic and data. We introduce algorithm properties that facilitate evolution, and we propose efficient rule and data evolution techniques for three ER models: match-based clustering (records are clustered based on Boolean matching information), distance-based clustering (records are clustered based on relative distances), and pairs ER (the pairs of matching records are identified). Using real datasets, we illustrate the cost of materializations and the potential gains of evolution over the na < ve approach.