Arandur comments on Expecting Short Inferential Distances