To me learning cause-and-effect is a non trivial process, when recognizing relationship often comes first and intrigues "formal" reasoning process afterwards (i

To me learning cause-and-effect is a non trivial process, when recognizing relationship often comes first and intrigues "formal" reasoning process afterwards (if any). The more formal in the later process, the closer we reach to the real "knowledge". So your suggestion is quite practical, in the sense that we can start from that and figure out how to push the engine toward the more formal spectrum.