Advancements in science have often been the direct result of progress in metrology. Galileo’s telescope and the Mars rover each expanded our understanding of space because they allowed us to observe phenomena that were once beyond reach. In that sense, the progress of science is bounded by the rate at which metrology itself advances.
If, for example, one had known in 2010 that future Mars rovers would precisely determine the planet’s sediment composition by 2020, it would seem almost pointless to publish a study inferring the same from coarser satellite data. Even more absurd is trying to recreate Minecraft in Python just because I don’t know Java. The result is the same but worse, since I spend more effort and get inferior outcomes. This highlights a dilemma: the scientific questions that preoccupy us may exist only because we have not yet built the instruments capable of answering them.
However, there are inspiring counterexamples to this. Le Verrier’s mathematical prediction of Neptune’s existence and position, based on irregularities in Uranus’s orbit, was confirmed by observation in 1846, long before direct detection instruments existed. Similarly, Mendeleev predicted undiscovered elements by identifying gaps in the periodic table. The difference is that neither of these scientists could have known whether their hypotheses would ever be confirmed, which show that science has the remarkable ability to predict phenomena before they can be directly observed.
Perhaps there is a potential guide to framing a good scientific question. It must remain meaningful regardless of foreseeable improvements in measurement. The scientific question should therefore feel almost impossible to answer, because there is no clear way to verify or falsify the hypothesis. Yet it should still be in some way falsifiable; otherwise, it is not worth discussion. I think there therefore exists a range in which the question is worth pursuing, although it is not clear in my mind how to quantify such a rule.
Brayden Noh 