Errata: The paragraph on determining protein structure with X-ray crystallography has been updated to fix some errors in the description of the technique.
Disclaimer: I am decidedly not a scientific realist, but if I may
I'd like to think that this isn't a question of instrumentation vs realism, but simply the lack of understanding of the principles behind instruments
Knowing what axioms apply or not to a given situation, understanding the rules of inference, being sensible, all these aren't taught when scientists get training
The problem isn't methodological or epistemological per se - it's decidedly institutional
I took pure Math exactly to avoid falling into these pitfalls, so that I'm always above what methods are used and what frameworks are discussed
Edit: was browsing Eastman's StatMech intro, and what he writes sounds fits my thoughts well enough
"In other subjects, you learn about “natural laws”: Newton’s second law, Maxwell’s equations,
Schrödinger’s equation, etc. These laws are not derived from anything else. They were discov-
ered experimentally and then assumed to reflect fundamental aspects of reality.
But statistical mechanics does not involve any natural laws of this sort. Instead, it is a set of techniques that can be applied to nearly any physical system, no matter what laws that system obeys. That is why I
call it the most fundamental field of physics. New theories may replace old ones, and natural laws
may turn out to be merely approximations to deeper laws. But statistical mechanics remains valid
through it all, and whatever new laws are discovered, it will almost certainly work just as well with
Errata: The paragraph on determining protein structure with X-ray crystallography has been updated to fix some errors in the description of the technique.
Disclaimer: I am decidedly not a scientific realist, but if I may
I'd like to think that this isn't a question of instrumentation vs realism, but simply the lack of understanding of the principles behind instruments
Knowing what axioms apply or not to a given situation, understanding the rules of inference, being sensible, all these aren't taught when scientists get training
The problem isn't methodological or epistemological per se - it's decidedly institutional
I took pure Math exactly to avoid falling into these pitfalls, so that I'm always above what methods are used and what frameworks are discussed
Edit: was browsing Eastman's StatMech intro, and what he writes sounds fits my thoughts well enough
"In other subjects, you learn about “natural laws”: Newton’s second law, Maxwell’s equations,
Schrödinger’s equation, etc. These laws are not derived from anything else. They were discov-
ered experimentally and then assumed to reflect fundamental aspects of reality.
But statistical mechanics does not involve any natural laws of this sort. Instead, it is a set of techniques that can be applied to nearly any physical system, no matter what laws that system obeys. That is why I
call it the most fundamental field of physics. New theories may replace old ones, and natural laws
may turn out to be merely approximations to deeper laws. But statistical mechanics remains valid
through it all, and whatever new laws are discovered, it will almost certainly work just as well with
them as it did with the old ones."