I think we should use the definition he refers to when he applies the inductive/deductive dichotomy, rather than the one which is explained at the start of the article, also given that it is machine translated and thus some meaning is lost in the process.