Hmm... but we can only ever inductively infer the causes of the things we observe. We observe observables, not beables. That's the nature of science. Our inferences can never be proven, only falsified. That's why we look for as many ways of confirming our inferences as possible. I just don't understand: do you mean to take issue with the scientific method in general? Because if you do, you've got grounds to do so: as you said, A==>B does not mean observing B proves A. What if A=/=>C and we also observe C? There's got to be some A' that's compatible with both B and C.
But instead of hypothesis confirmation, what would you do? I really don't mean to sound negative. Maybe I've got this wrong somewhere...