Discussion Questions on Salmon's "Statistical Explanation"

1.  What moral does Salmon draw from counterexamples such as the one involving John Jones' consumption of birth-control pills or of vitamin C?

2.  When Salmon talks of "relevance", what exactly does he have in mind?

3.  What is the "reference class" and why are there ambiguities concerning what reference class an event belongs to?  What determines the reference class to which the explanandum event belongs?

4.  What is it for a reference class to be homogeneous?

5.  Define what it means for some property or event type C to "screen-off" A from B.

6.  What is the connection between causation and screening off?

7.  Why is it a mistake to partition with respect to a statistically irrelevant factor?

8.  Why does Salmon reject the view that explaining requires citing factors that increase the probability of the explanandum?

9.  To what extent is Salmon a defender of the symmetry thesis?  How does his symmetry thesis differ from Hempel's?