Thread by @lymanstoneky

This paper might be one of the greatest NBER papers ever published and also such a perfect example of epistemic paradigms being important. nber.org/papers/w29182?…

Short version: 1644-1912, Chinese counties were MASSIVELY more likely to have rebellions in years where they experienced a total eclipse. In counties that also have memory institutions (like temples), this rebellious history survives today, and there is more unrest TODAY.

The authors say this is because in Confucian beliefs, an eclipse is a sign of heaven's displeasure with a ruler, and so peasants took it as a sign that their existing grievances could be licitly expressed in rebellion. This is a religion-as-social-narrative story.

But maybe they're wrong.

Maybe eclipses REALLY DO signal heaven's displeasure with a Qing Dynasty ruler, and so eclipses "cause" rebellions not because of peasant beliefs but because rebellions are ALSO are a sign of heaven's displeasure!

Obviously to scientific moderns this explanation sounds kind of crazy, but the key takeaway is that the only thing separating these two theories is "Might heaven in fact feel displeasure and act as an agent in history on that basis?"

As causal explanations, these two theories both "work." Nor is one necessarily simpler than another: postulating an complicated ideological narrative widely held among illiterate peasants which motivates lethal action more than e.g. famines or earthquakes...

Is *hardly* more parsimonious than "God was TICKED at the emperor."

Nor is it even more observable: historic peasant ideology is almost completely unobservable and we have to just accept elite attestations about it plus some archeological type data which is also debatable.

The difference then is not in parsimony or empiricism, but in *theoretical* empiricism. The modern explanation says IF we could go back in time and interview peasants, then we could observe their beliefs which would be such-and-such.

Whereas the traditional-religious explanation says *even if you were there* you would not be able to like interview heaven and ask "How ticked were you?"

The difference is not actual empirical evidence nor an Occam's Razor type judgment, but a hypothetical argument about what we could do, in our imaginations.

Confucian ideology recognized the predictability of eclipses and had begun accurately predicting them centuries prior to the period in question!

https://twitter.com/AlexGodofsky/status/1437449606091128837

Thus, the acceptance of the modern scientific account is essentially an *imaginative* act.

Now, it's an imaginative act I firmly believe. And I'm perfectly happy to impose it: you *should believe* this imaginative act.

But that's because I accept virtually all beliefs as volitional. :)