Inspecting sacred cows
By Thoreau
In a continuing effort to highlight the way that even the most fundamental assumptions are subject to experimental testing in science, today I point you to Physical Review Letters, where a paper was just published on the masses of the electron and proton. Some cosmologists have speculated that some of the fundamental constants of the universe, (e.g. electron charge, speed of light, electron and proton masses) might be changing over time. (Strictly speaking you can generally only detect changes in the relative values of constants, but we’ll leave that aside for the purposes of this blog post.)
By analyzing the spectrum of light emitted and absorbed by ammonia molecules far away in space (and hence detecting light emitted 10 billion years ago), they can check to see whether the frequencies of certain transitions have changed over time. Since these transitions are related to molecular vibrations, they depend on the masses of the electrons and protons. So if you measure the energy of a transition in an ancient molecule, you can measure the masses of electrons and protons 10 billion years ago. They find that the masses (or, more accurately, the ratio of the masses) have changed by less tha one part in a million over 10 billion years.
Even though this isn’t my field, I like reading these sorts of papers. The analyses are generally clever and understandable, and the significance of what they’re doing goes way beyond the field, and to the very heart of science: Constantly checking assumptions against observational data. I won’t claim all science is done this way, but it should be done this way, which makes it all the more important that such articles appear in prestigious journals, as reminders to us all.
Comment by Gene Callahan —
June 16, 2007 @ 3:27 am
“In a continuing effort to highlight the way that even the most fundamental assumptions are subject to experimental testing in science…”
This, of course, logically cannot be the case. How, for instance, could the fundamental assumption that experimental testing is a good way to get at the truth be tested experimentally without logical circularity? Or the idea that their is some sort of quantitatively stable relationship between a sequence of measurements? (You have to simply assume this, or all of your measurements are merely random pokes in the dark that prove nothing.)
Comment by Thoreau —
June 16, 2007 @ 1:19 pm
Gene-
Fair point. But if experimental testing of nature never yielded any reproducible patterns then that would be a refutation of the value of experimental testing. So in some sense our experiments are tests of the whole experimental principle.
Comment by Gene Callahan —
June 16, 2007 @ 7:44 pm
Yeah, but…
1) If you don’t assume measuring “3 meters” today and measuring “3 meters” tommorow means the same thing, then what is the meaning of “reproducibloe results”?
2) Even given we buy into 1), why should the reproducibility means anything without assuming some sort of inductive principle? We could instead assume the principle “It’s time for a change!”, which is every bit as backed by experience as ordinary induction. Or, see Nelson Goodman’s “New Problem of Induction“.
Note: I think the assumptions necessary to generate science are justified, but I just want to note that even science must make assumptions it can’t test with its own methods.
Comment by Gene Callahan —
June 16, 2007 @ 9:28 pm
Ooh, sorry, that page I linked to is a very half-assed, amateur page on Nelson — try here.