A Means-End Account of Explainable Artificial Intelligence