Currently physical effects are classified by actions they help to perform and not by contradictions they help to resolve. But to facilitate the procedure of finding a physical effect that implements a separation in a parameter, they have to be re-formulated to clearly state separation in which parameter is implemented and how.
In the case of such effects as Currie point it is easy. The latter does not even need to be re-formulated because it is already phrased in the from of a separation: below some point on the temperature scale magnet remains magnet, and above it does not. But for other physical effects it is not so easy to phrase them as separations. It is even difficult to discern any separations in most of them.
So how in absence of such a rephrasing could physical effects be selected to resolve a contradiction ? It could be done purely statistically. If statistics is accumulated about what effect may help resolve what contradiction then effects could be selected based on this statistical correlation. But it is a poor approach. The right way is to rephrase effects so that separations they help implement would transpire.