Two problems on invariance for interventionist or policy purposes

The use of invariant knowledge for achieving interventionist or political purposes is commonly accepted by manipulability approaches. However, such use raises two philosophical issues that will be examined in this paper. The first problem is ontological, and refers to the difficulty of finding genuine stable causal factors in the social sciences. In this regard, it is showed that social phenomena do not respond to the logic of stable factors such as mechanisms or << capacities >>, but is more suited to the logic of << possibility trees >> or << open-ended results >>. The other problem is epistemological, and refers to that even if stable social contributions are founded, nothing guarantees that such stability prevails in future. Different thesis supporting this stance will be examined, such as the so-called << Lucas critique >> or the << black swan theory >>.