The sciences: unshakable knowledge?

Descartes, who instituted a quite reasonable and modest search for new standards of certainty in science, also had the vision of something much grander that might lie beyond them -- of an absolute, invulnerable certainty, an ideal form of knowledge which nothing could ever shake. This unreal vision has given endless trouble. It is not only excessive; it gives priority to the wrong sort of ambition. The kind of perfection at which the Cartesian project aims is the perfection of epistemic safety. It concentrates on knowledge, not because of any special view (such as Plato had) about what we need to know and why we need to know it, but because knowledge means security from error. By assuming that the possibility of error is the evil that must above all be avoided, it distracts (among other things) from asking about the various reasons why knowledge can be important. And . . . it systematically directs us away from interesting and important questions that possibly cannot be completely settled, and towards less interesting ones where the risk of error seems easier to control.
All this might not matter so much if the demand for unshakable certainty could be satisfied and then put aside. But it cannot. Its ideal is unattainable, going far beyond anything the sciences have ever delivered or could deliver. In all of them since Descartes's time, beliefs which were thought to be unshakable have been first shaken and then abandoned. Mary Midgley, Wisdom, Information, and Wonder: What Is Knowledge for? New York: Routledge, 1995, p.36.