[Karunamon]

(oops!)

I don't think a true null hypothesis exists as long as humans are involved in concocting them, the person could want, consciously or subconsciously, any output from any experiment.

Maybe one outcome leads to more research that's a major paint to secure funding for and one is much easier? Maybe the outfit funding the study clearly wants one particular result?

Which goes right back to what I'm saying: humans are not purely logical, true null hypotheses don't exist, and the only difference between a "hypothesis" and a "belief" by what you just described is the degree to which the person with the idea wants a specific outcome - a variable which is completely unrelated to the eventual truthiness or falsity of the output.

[Nursie]

The fundamental difference is that a hypothesis is not (or should not be, different argument) accepted as fact by the scientific establishment before it is tested and shown true or false with empirical evidence.

As such it doesn't matter where it comes from so long as it is tested before acceptance. Which is why I struggle to understand the bootstrapping problem that the other poster talks about.

--edit-- To put it as simply as I can muster - why does the origin of the idea to test matter, when it is the testing that's the important part?

[crististm]

Hypothesis origin does not matter.

But not all things can be tested - think about dual slit experiment. You can't reproduce the experiment in the sense that you'll know which slit will be taken.

[Nursie]

But you can reproduce the probability distribution aspect of it quite reliably. If you think the dual slit experiment presents a problem to the scientific method or empiricism you're quite wrong.