What is the definition of a physical proof ??

There isn't one. There are mathematical proofs and experimental evidence. You can accumulate evidence to an arbitrarily high level of confidence but you can't physically prove something 100%.

What is the definition of a physical proof ??

We tend to think about falsification of a theory rather than a proof.

 

Basically a theory would be considered good if it agrees with our observations to some agreed upon accuracy taking into account experimental errors and the expected domain of applicability of the theory.

 

If it does not agree well, say it predicts some phenomena that just is not observed, then the theory is bad and is usually considered falsified.

Edited May 11, 201411 yr by ajb

But then, one has to choose whether the theory is the cause, or its interpretation, or a model within this theory, or if the experiment was badly set up...

 

That's daily life. When the 3K background radiation was discovered, the antenna builders had no theory in mind, they just observed excessive noise - and because their antenna was the most silent ever built, they had no guide to pinpoint the cause. More, their noise temperature was maybe 30K (for sure a fantastic value for that time, still excellent today) so the excess was tiny, for a measure difficult to conduct.

 

Only years later, someone read about the excess noise and did remember the little known theory about cosmological background, and suggested a relationship. You imagine that the fellow scientists didn't answer "of course" immediately and enthusiastically. It took more observations, both of the background and of related predictions, to convince people slowly.

 

Same for Michelson-Morley: the first reactions were about the result's reliability (the experiment was extremely sensitive), other reasons - and then people tried to adjust existing models with ether dragging and so on.

 

Or more recently, about supraluminal neutrinos: was the observation real? Completely new physics, or small adaptations, or just a mistake?

 

In the future, maybe the string theory will be abandoned. For no hard evidence, but just because some predicted particles are not found. That won't be a proof against - just a bad feeling. And imagine if detectors persist not to see gravitational waves: what next?

 

So "proofs" are weak in physics, which is a stack of theory, models, experiments and tinkering. Real proofs exist in maths only. That's an excellent reason to want several experiments by independent teams give the same result.