Pothos suggests dispensing with the distinction between rules and similarity, without defining what is meant by either term. We agree that there are problems with the distinction between rules and similarity, but believe these will be solved only by exploring the representations and processes underlying cases purported to involve rules and similarity.