James Woodward (University of Pittsburgh)
This paper describes a strategy for dealing with causal complexity involving what I call conditional irrelevance. This describes conditions under which we can legitimately replace a large number of lower-level variables mi that are causally relevant to some effect or explanandum E, with a much smaller number of upper-level variables Mj. The key idea is that fixing the values of the Mj via interventions, further interventions on the mi consistent with the values of the Mj should l not change the value of E. Biological illustrations will be provided.
