If I guess the time, and get it right, do I know the time? No, says common sense, and nearly all theoretical and formal epistemology. If I guess that it will rain tomorrow, am I any better off? Presumably not. Yet we assess predictions almost entirely by whether they are right.
I do think Swedish predictive work was broadly accurate, compared to, for example, the models produced by Imperial College London. But more importantly, I think their stance was rational. They did what was right given the evidence. That isn’t the same as being right in the sense of landing on the truth. But there’s nothing either epistemically or morally significant about the latter. The former, however, is both. Sweden behaved more reasonably than any other country, or perhaps at least as reasonably as the most reasonable, given that there was room for reasonable disagreement.
The stance on Sweden is another version of the intellectual intolerance of the age. And it ignores the evidence. Sweden has done well: not perfectly, but no country has, that I can think of. Whether it comes out tops long-term is up in the air. But there is good reason to think it will – at least as good as the reasons to think it won’t.
Katy Balls talks to Sunetra Gupta, Professor of Theoretical Epidemiology at the University of Oxford. An expert in the fight against infectious diseases, she is the lead scientist behind the Oxford study that disputed Imperial College’s dire coronavirus predictions. She is also a novelist and translator. On the podcast, she talks to Katy about her writing and how it was inspired by her intellectual father; her dispute with the mentor of Imperial College’s Neil Ferguson; and how she has found being in the public eye.
The table below shows 250 days after first infection. Yes yes infection didn’t all start on 1 Jan but this model would only be ballpark correct if the infection started globally 6 weeks ago.
It’s easy to be wise in hindsight – except that quite a few people were saying this sort of thing, at every step of the way. I predict that in future this – the model, the politics, all of it – will become a classic study in how science policy should not be made. In the meantime, as we climb down from the heights of our panic, it’s just so fascinating to witness ideas that start off as dangerous – like “maybe we shouldn’t be locking down” – gradually become more common, and to feel the tug in oneself of trying to decide who to trust.
The usual disclaimer: I’m not saying the virus isn’t dangerous, that lives don’t matter, that we should do nothing… just that we haven’t reacted well.