This is mostly to make it easier to write linear function types without having to invent names for everything, which might be noisy. Also it improves the display of linear function types when the name isn't used in the scope.
