Series Title: Science for Use: In praise of the inexact, the inelegant, and the unassuming

The theme of these two talks is science for use: what demands should we expect scientific claims and concepts to fulfil when we want science not just to help us understand the world but to help us change it? As my title suggests, contrary to what is widely assumed, exactness is not the answer. The slogan for the talks comes from my late husband, Stuart Hampshire, who enjoined: Do not bring more exactitude / precision to a subject than it can bear – and, I add, than is useful.

 

Lecture 1:  A laudatio for scientific laws that are inelegant, inexact, and unassuming

 

Abstract: I begin with an injunction to focus on the reliability of scientific laws, not their truth.  That’s because what’s so great about modern scientific laws is that they have proven to be so useful. And when it comes to usefulness, their truth is of little consequence; it’s reliability that matters. We need our law claims to be able to do what we expect of them. And if we're going to be justified in our use of them, we need to be warranted that they can do what we expect. That is, scientific claims need to be reliable and we need warrant that they are reliable. Being true – or even being properly ‘truth apt’ – is neither necessary nor sufficient for reliability. I shall illustrate with an account – and a defence – of the importance of ‘middle-level principles’ and what JS Mill called ‘tendency principles’. This will take us straight  into a discussion of generics and a rejection of what Sharon Crasnow has dubbed ‘the detachability thesis’ – that scientific claims can be regarded as free-standing, detached from the vast tangle of work that supports and interprets them.