The issue with arguments of the form "Mutation is..." is that all these things are models of mutation: merely representations. Different models help us elucidate different properties of the same underlying phenomena. This does not mean that either one is "correct" or the one true representation of mutation.