What sort of taxonomy of causation do we need for language understanding?

Open in new window