Francis Spufford considers which kinds of questions match which kinds of answers.
I think that the history of 20th-century scepticism is the history of attempts to take scientific knowing – which seems so gloriously solid, and in its own domain is solid – and to extend it beyond its natural reach, into the rest of life. Because it would be lovely if things were verifiable in the scientific, you know, experimental sense. If you could have a moral hypothesis tested by experiment and then know exactly what to do afterwards, life would be great and much simpler. And I suspect that some of the darkness and confusion of 20th-century history has been what has created the appetite for trying to think of ourselves with science all the way down like that.
But actually, science doesn’t work so well as a method when you take it out of the questions it’s designed to ask, and I am radically unconvinced by the idea that there are verifiable answers of that kind to human and historical quandaries. We don’t get to prove things quite like that. Other kinds of assurance are available – and I’m not just talking about religion, I’m talking about culture, and imagination, and art. But we don’t get to write “QED” at the end of every proposition about humans.