Evidence and meaning
When we’re inferring general patterns from individual events, we put too much weight on events of personal, moral, or political significance.
We focus too much on whether we were harmed by some risky behaviour, relative to what’s happened to other people. Obviously on average there’s nothing special about yourself.
And we put too much evidentiary weight on what happens in large countries, such as the US, relative to small ones. We can learn a lot from events in small countries, even though their practical importance is smaller.
Likewise, we over-generalise from events of great historical significance, like World War II, whereas we neglect less significant events.
*
Conversely, we’re too inclined to explain significant events by general patterns. The greater the significance, the smaller the role of chance.
And not only that - the general patterns should be morality tales, that tell us something meaningful.
On this view, obscure policies on little-known issues may be down to chance, but a country’s Covid-19 strategy must be explained by deep-rooted national characteristics.
And a state leader can’t have been assassinated by some random madman, but must have been the victim of a major conspiracy of wider moral and political significance.
*
Evidence and meaning are orthogonal. More significant events aren’t necessarily better evidence. And they don’t necessarily have more meaningful explanations.
That’s because whereas we ascribe meaning to events, evidence is independent of us. The world gives us clues, and we must do our best to find them. They appear among significant and insignificant events alike.
It can be hard to internalise this. It’s difficult to disenchant the world.
We do well, e.g. in astronomy. We don’t put disproportionate evidentiary weight on large planets, the way we often put disproportionate weight on large countries. Little Mercury provided important evidence for the theory of general relativity.
But in the social world, we often fail to tear evidence and meaning apart. Careful social scientists probably do succeed, at least to a large extent. But most people tend to be oblivious about this bias, and fall for it hook, line, and sinker.