Summary
Report
See the full content of this document
Extract
Abstracting away from preemption.
In cases of deterministic preemption, there is a backup, c', to the actual cause, c, of an effect such that had c failed to occur, c' would have caused that same effect. Had Assassin 1 failed to fire his weapon and kill Jones, Assassin 2 would have killed him. In cases of indeterministic preemption, the backup cause would have raised the probability of the effect to the same degree as did the actual cause had the latter failed to occur. In either form, preemption is a key test case for any theory of causation. Passing this test requires classifying the preempting cause as genuine and the preempted cause as non-genuine in any case of preemption. There are also the related challenges of the phenomenon of overdetermination, which an adequate theory of causation must handle. I have argued elsewhere that one of the comparative advantages of what I will call a physical connection theory (which includes a certain kind of singularist component involving persistence) is that such a theory is better equipped to deal with preemption than are other standard reductionist theories (Ehring 1990, 1997, 2004).
In this paper, I will criticize a general strategy for making the latter theories compatible with the preemption and overdetermination. This strategy maintains that causation is reducible to some non-causal relation R between events that does not involve a physical connection, but only when those events are screened off from the distorting circumstance of accompanying preempted causes or other overdetermining causes. Abstracting away from such circumstances, the preempting/overdetermining cause is R-related to its effect. (Or, in another version, intrinsic analogues of the cause/effect pair, connected by an intrinsically indistinguishable process, are R-related). My goal is to show that that strategy--which I will call the "abstracting away" strategy--does not provide an adequate basis for dealing with preemption and overdetermination. I will conclude that we are better off focusing on physical connection theory. I. PRELIMINARY DISTINCTIONS AND TRADITIONAL THEORIES I begin with a number of distinctions among theories of causation in order to indicate more precisely the kind of "physical connection" theory of causation I defend. Reductivist/Non-reductivist. A non-reductionist about causation holds that causal facts are primitive and that they cannot be reduced to non-causal facts. A reductionist holds that causal facts are reducible to non-causal facts, with a wide range of disagreement over the non-causal reduction base for causation. Singularist/Non-singularist. I will here take non-singularist theories to make causation depend in some way on laws of nature. This dependence may be direct as in "nomological accounts" (the causal pair, c and e, must fall under an appropriate covering law) or indirect as in "quasi-nomological" accounts (for example, in Lewis's counterfactual theory, laws help determine which counterfactuals are true, but it is not required that cause and effect fall directly under a covering law). Singularists, as understood here, reject any dependence of causation on laws, even the indirect dependence of a Lewis-style counterfactual theory. The fact that c causes e entails neither any particular law under which c and e fall, nor even that there exists some covering law applicable to c and e, nor even that there are any laws. Causation is, thus, possible in a lawless world. This "singularist/non-singularist" distinction can also be extended to the components of a theory of causation. An example of such a singularist component, arguably, is the requirement that causes be spatially and temporally contiguous with their direct effects. Finally, a theory of causation may fail to be singularist as a whole, although it includes singularist components. Intrinsic/Nonintrinsic. Intrinsic features of a situation are, very roughly, those qualitative features of a situation that do not depend on what conditions hold elsewhe...See the full content of this document
Sponsored links
