Monday, February 23, 2015

Essential properties and self-dependence

My being seated now is causally explained by my having sat down. Suppose that my being seated now is an essential property of me. Then, had I not sat down, I wouldn't have existed. So my very existence would have counterfactually depended on my own causal activity. But that would absurdly make me be something too much like a causa sui.

This seems to generalize into an argument for a general principle:

  1. No property that counterfactually depends on an entity's own non-essential causal activity can be essential to that entity.
I needed to specify that the relevant causal activity is non-essential, because there is nothing deeply absurd about an entity that has some essential properties follow from causal activity that's essential to the entity.

But the form of argument that I used to get to (1) gives other results. For instance, a leading theory of personal identity holds that in cases where symmetric fission occurs—apparently, a person splitting into two—there were already two co-located people there before the fission. But why are there two people there? Presumably precisely because fissioning occurred—otherwise, there would have been only one.[note 1] Thus the existence of the two people is explained by the fission. But surely if fission is possible, it's possible that it be triggered by the non-essential action of the fissioning individual or individuals. In that case, then, the existence of the two people is explained in part by their very own activity. So this account of fission leads to something like self-dependence and should be rejected.

There may, however, be an objection to the argument. Suppose that the causal self-dependence is not fundamental. Instead, more fundamentally, we have a case where the whole depends on the contingent causal activity of a part. For instance, we may think that the length of an event is an essential property of an event. But the length of an event frequently depends on the contingent causal activity of a part of the event. (Thus, the length of World War II depends on the effects of D-Day, even though D-Day is a part of World War II.) In some such cases we might say that the whole depends on its own causal activity, since the causal activity of the part can sometimes be attributed to the whole.

I am not convinced. I think that in cases like this, it is incorrect to say that the whole depends on the whole's causal activity, but that it depends on the part's causal activity, and in this case the part's causal activity is not in fact correctly attributed to the whole.

In any case, the part-based objection will only apply in cases where the whole's causal activity is derivative from the part's causal activity. I think free choices are not derivative from the causal activity of anything other than the person as a whole, so in the fission case it's still impossible that whether there are two persons should depend on their choices. But this is rather controversial.

If there are cases where the whole depends on its own causal activity because its causal activity is derivative from a part's causal activity, then (1) needs to be qualified to apply to fundamental entities.

1 comment:

Heath White said...

This form of argument seems very sound to me. For the idea of counterfactual dependendence of B on A seems to be this: if it were not the case that A, it would not be the case that B. Now if A is contingent, then Poss(~A). But then by a form of MP, Poss(~B). If we stipulate that B is an essential property, however, we have a contradiction.