This can be read as a response to Einstein's Arrogance. There, Eliezer Yudkowsky has argued that usually, when you have enough evidence to even conceive of a theory, you have enough evidence to be pretty sure it's correct. Here, I argue that this state of affairs held only in theoretical...
Maciej Cegłowski has written an article named Scott and Scurvy, which has already been discussed on LW as an example of the "messiness" of science in practice. Cegłowski follows the story of how a working cure for scurvy was found and then lost to an incorrect theory in the face...
This can be read as a response to Einstein's Arrogance. There, Eliezer Yudkowsky has argued that usually, when you have enough evidence to even conceive of a theory, you have enough evidence to be pretty sure it's correct. Here, I argue that this state of affairs held only in theoretical physics roughly between the years 1915–1975, mostly because symmetry and renormalisability created strong mathematical constraints for possible theories, and hasn't been seen before or since.
It's also my own contribution to the mythos around Einstein -- I try to clarify what exactly he accomplished by surveying the attempts at solving the problems with the aether that came before him.
I think we can learn a lot from the dead ends of history that aren't much talked about, because a) the shape of success varies, but failure modes often repeat and b) they are closer to the real way science progresses than the presentation in textbooks, which can skip approaches known not to work.
Part of it is that more complicated interactions tend to create instabilities: https://physics.stackexchange.com/questions/384910/can-einstein-hilbert-action-be-derived-from-symmetry-considerations/385024#385024
Maciej Cegłowski has written an article named Scott and Scurvy, which has already been discussed on LW as an example of the "messiness" of science in practice. Cegłowski follows the story of how a working cure for scurvy was found and then lost to an incorrect theory in the face of new data, which is quite the case study for theories of how science works.
I was fascinated by the story, dug into the primary sources, found that there is a second, more optimistic half to it, and wrote it up. The tale of Scott and Scurvy culminates with the scurvy-accelerated demise of Robert Falcon Scott in 1912, which makes for a pessimistic outlook, but look around: scurvy is not a problem anymore. Why?
I think that people of LW might find this interesting.
For linkposts, is self-promotion frowned upon? I have written a post that I believe would be interesting to readers here, but it's on Substack (free and with no intention of paywalling!). Can I create a linkpost for it?
Why doesn't this apply to Stokes total aether drag theory? George Gabriel Stokes knew about a number of experiments that tried to detect the aether wind and failed, and he made a model where the aether behaved like a non-Newtonian fluid: fluid without viscosity at low speeds (such as the Earth’s), but solid at high frequencies. That way, it could host high-frequency light waves while posing no resistance to ordinary moving objects, and moving objects would drag light along with their movement.
The theory came out in 1845, in 1851 Fizeau measured the speed of light in moving water and found that the water didn't drag the light with its full speed.
What happened... (read more)