Abstract: In the context of linear causal systems, ‘How to Get Causes from Probabilities’ shows that given correct background information about other causal facts, certain probabilistic relations (like correlations and partial correlations) are both necessary and sufficient for the truth of new causal facts. This is done by showing how simple structural models from econometrics can be read causally if the conditions for identification of the model are met, and a generalized version of Reichenbach's principle of the common cause is assumed.

Keywords: cause, common cause, correlation, econometrics, identification, probabilities, Reichenbach,