Kedaman
What's the use of having only false and potentially true statements? For logic to be applied you need axiomatic approaches. Besides, isn't it possible that a falsified statement can be reverted to potentially true, since observational evidence doesn't justify anything?
OK, according to Popper, the purpose of science is to weed out the false theories (from an array of conjectured hypothesis) and from those left, select the best one. Eventually, you end up with the best available theory and go with that. The benefit of not having actually true theories is that you keep things open to revision. You acknowledge the fact that one day a new and better theory may come along and usurp the previously held one.

The reason why observational evidence doesn't justify anything is, not because it might be erroneous (although that is a valid point in itself), but because any finite collection of observations may be consistant with an infinite number of potential theories.

Assuming that an observation is valid, if it does not falsify your particular theory, it doesn't mean that your theory is in anyway justified. There are an infinite number of other theories that would also not be falsified by this observation.

That is why an observation can falsify but not confirm a theory.
Isn't that a bit how inductionists think? Why does the event of falsification become less probable if you have more correct applications of your statement?
I was refering to non-generalisation statements (such as "there is a chair in the den") which can be confirmed by observation.