I read a very interesting post today. It tries to prove that science – as opposed to religion – is not faith-based. The story also got on Digg frontpage and was heavily discussed there. One comment thread in particular caught my attention, and it gave me some ideas that I feel I should write down before I forget them.
As I understand it, in the original post Phil Plait (the blog’s owner) argues that religion always relies on one or more untestable assumptions (e.g. that there is a god) that are supported solely by faith. According to Phil, science makes only one assumption – that the Universe follows a set of rules (not any specific rules, just that there are rules). This is a testable assumption – you can check if a rule holds up based on what happens “in the real world”.
Here’s the conundrum – one Digg’er suggested that “testability is truth” is a faith-based assumption. In other words, it is only an untestable assupmtion that testable assumptions are not faith-based. Think about it.
I’m not an expert of logic, but I think I see a flaw here. “Testability” is not an assumption, it is instead a part of the framework in which assumptions, theories, proof, etc are assessed – logic (in the wide sense). It’s meta.
But there’s more. One of the definitions of logic is “a system of reasoning”. Logic is rules for reasoning. So is the assumption that “the Universe follows a set of rules” really testable? If it does follow rules, there’s no problem. On the other hand, will logic – a set of rules – apply if there are no rules? Probably not, but how would we be able to find out? No rules -> unclear if logic works -> can’t use logic to determine if logic works. So it appears that we can’t really determine if the Universe follows a set of rules, because there is no way to detect a case where it doesn’t.
What’s worse is that I (hopefully) used logic to ask the above questions and come to those weird conclusions.
Am I wrong?
Addendum. This comment on Digg puts a part the issue a little more eloquently -
Related posts :
Logic (which is the heart of science) relies on axioms; and axioms, by definition, are unprovable assertions. For example: “If A implies B and B implies C, Then A implies C” I have faith that this is true; but any “proof” would necessarily be a circular argument, for how would one link together the steps of the proof without using the result itself?