I think they are saying that descriptive scientific observations have limited utility as prescriptive devices if we are to make progress because descriptions are based on how things have been rather than how things could be (or even could have been).
Especially in soft sciences, the /cause/ of what is described can be hard to reliably tease out. To assume the same cause is an unchangeable aspect of the human condition and that cause, even if true, will only ever lead to one end-state is often a mistake.
