There's something that's really been bugging me about discussions of causal overdetermination.

What is causal overdetermination. Well, its when two events are each sufficient to produce some subsequent event and they both occur.

Its fairly easy to come up with examples that appear to be clear examples of causal overdetermination. For instance, if someone is shot by a firing squad and two bullets enter the same organ at the same moment. Either bullet could have done the job.

I would also consider an example such a commuter who drives to work on two different routes. Suppose there's construction on one route and an overturned truck on the other. Why was the commuter late for work? Well, either of these two events could have been responsible for the lateness. (Though one could argue that the single cause of an inability to find an alternate route when necessary was the sole cause for the employee's tardiness.)

Most people who discuss causal overdetermination would probably not accept my second example. This gets to the point that's been irritating me. Both skeptics and proponents of causal overdetermination would accept the same real world cases as having had occurred. The dispute is about how to divy up the responsibility for any events between its various putative causes. This division of responsibility can have profound philosophical importance. This exact issue often spells the difference between a robust theory of mental causation and epiphenomenalism.

On the other hand, the proponents of theories on either side will made exactly the same predictions about the future and will consider reports of exactly the same sorts of cases having occured in the past. Both proponents and opponents will accept, for instance, that the criminal in question above is actually dead and that the actions of the various members of the firing squad are responsible.
Blame my pronounced empiricist intuitions, if you will, but this makes the debate a little less pressing. There doesn't seem to be a compelling reason to force a conclusion to the debate. Some theorists will accept a "no over-determination" thesis and another will accept a "over determination" thesis. This will have ramifications throughout their philosophical arguments. Maybe discussions will get heated, as they often tend to do between those who accept some two level picture of human and action and more purely physicalist theorists. However, as soon as the point of disagreement is discovered then the more gracious (or weak willed depending on your point of view) will bend and choose to use the other's conventions until such time as they have settled whatever really important question was at the heart of their discussion.

This brings me back to the case of the late commuter. A DAG (see previous posts) could be constructed with the employee's tardiness as the final effect. The nodes feeding into that final effect, however, may be chosen differently for different models. For the employee, various road conditions are important. For the employer, resourcefulness, enthusiasm and alarm clock settings might constitute the relevant variables. (Though in this case, the models in question are not seperated from the variables they include and choice of models will have measurable real world effects.) Either model, or both, could be valid models. The choice between models is based on case by case considerations. I may return to this point, but for the moment, it should be clear that the position I'm hinting at is not committed to either relativism (since there are real world that regulate which theories are acceptable) or incommensurability (since the identification of the point of disagreement in developing theories of causality is often enough to develop appropriate translation tools.)

If anyone's wondering, this note was a from the hip response while reading Jonathan Schaffer's essay "Overdetermining Causes" in this months philosophical studies (volume 114, pp. 23-45).  ¶ #

0 comments Post a Comment