From Lewis, Letter’s to Malcolm: Chiefly on Prayer, 38-39.
It is true in one sense that the mark of a genuine science is its power to predict. But does this mean that a perfected science, or a perfected synthesis of all the sciences, would be able to write reliable histories of the future? And would the scientists even want to do so? Doesn’t science predict a future event only in so far as, and only because, that event is the instance of some universal law? Everything that makes the event unique – in other words, everything that makes it a concrete historical event – is deliberately ruled out; not only as something which science can’t, or can’t yet, include, but also as something in which science, as such, has no interest. Take away from the sunrises that in which they differ and what is left will be identical. Such abstracted identicals are what science predicts. But life as we live it is not reducible to such identities. Every real physical event, much more every human experience, has behind it, in the long run, the whole previous history of the real universe – which is not itself an “instance” of anything – and is therefore always festooned with those particularities which science for her own purposes quite rightly discounts. Doesn’t the whole art of contriving a good experiment consist in devising means whereby the irrelevancies – that is, the historical particularities – can be reduced to the minimum?
Later in his essay Burnaby seems to suggest that human wills are the only radical unpredictable factor in history. I’m not happy about this. Partly because I don’t see how the gigantic negative which it involves could be proved; partly because I agree with Bradley that unpredictability is not the essence, nor even a symptom, of freedom. […] But suppose it were true. Even then, it would make such a huge rent in the predictability of events that the whole idea of predictability as somehow necessary to human life would be in ruins. Think of the countless human acts, acts of copulation, spread over millennia, that led to the birth of Plato, Attila, or Napoleon. Yet it is on these unpredictables that human history largely depends. Twenty-five years ago you asked Betty to marry you. And now, as a result, we have young George (I hope he’s got over his gastric flu?). A thousand years hence he might have a good many descendants, and only modesty could conceal from you the possibility that one of these might have as huge a historical effect as Aristotle – or Hitler!