1. The logic of principles supercedes experiment, because without such principles the experiment would be meaningless. Perhaps the objective state of the world does not depend on those principles, yet the accurate operation of the experiment depends more on the formulation of results than on the state of the world.
For example, how do we determine what's interesting (in an experiment)? There must be certain categories or properties which merit interest. Deriving results depends on comparing two interesting things. So there must be a principle of correspondence.
2. Since certain principles are the foundation for deriving truth from experiment, it follows that such principles are an appropriate ground for non-experimental positivism--a second type in which universal laws of correspondence can be tapped for special cases and therefore special solutions, ranging from new types of bottlecaps to over-unity devices. The greatest inventors are well aware that the correct experiment would be impossible to find without a considerable appreciation of the second type of positivism: positivism as meta-logic, as an internal method for finding solutions.
Since the art of inventing concerns originality and special utility, we can say that it concerns "The Principle of Asymmetry" within the meta-logic of categories.
My unbalanced (although non-perpetual) wheel by the same name can be found under "Principled Asymmetry".