Sunday, 4 January 2015

Problems with Agent Causation

What is causation? As I understand it, if X causes Y then:
  1. X and Y are temporally indexed propositions (or some metaphysical analog to propositions; i.e. facts)
  2. X is temporally prior (or maybe simultaneous) to Y
  3. X counter-factually entails Y
These three conditions seem to be built into our very concept of causation, and while maybe they are not alone sufficient, they surely seem necessary.

But this conflicts with the agent causation brand of libertarian freewill, which requires that causation is the sort of relationship that can involve—as operands—particular objects. Say the effect in question is "Sam's choosing to eat breakfast", then if it's a free choice it must be caused by Sam himself. It cannot be caused by any events containing Sam, or any sort of fact about Sam (e.g. his being hungry), but must be the object that is Sam. In this case, how can we understand causation? We must give up not only (1), but also (3) since particular objects cannot stand in counter-factual entailment relationships, and (2) since it makes no sense to talk about objects themselves being temporally indexed. Events containing objects might happen in some temporal order, and propositions about objects might be true at particular times, but it doesn't make much sense to attribute temporal attributes to persons. And even if it did, still Sam existed prior to the event, he exists while the event takes place, and hopefully he will continue to exist afterwards. So regardless, (2) must be given up.

In other words, to accept this brand of libertarian freewill, we have to reject everything we intuitively understand to be causation. Can the libertarian say; fine, maybe Sam's choosing to eat breakfast wasn't caused? Of course not: how could Sam exercise control in any meaningful way over an uncaused event? Such an event would be random, and randomness precludes freewill.

The libertarian might offer non-personal examples of 'object causation' in his defense, like a ball breaking a window. But to me this only sounds like a case of sloppy language. Everyone implicitly recognizes that when we say "the ball broke the window", what we really mean is "the ball hitting the window broke the window". Obviously if the ball had not hit the window, the window would not have broke. But then "the ball hitting the window" is an event, and not a particular thing: the cause and effect are both events. This therefore isn't actually a good example of object causation, because once we clarify what needs to be clarified, it's pretty obvious that it's really an event and not an object that's causing the window to break.

With all this said, it's very unclear how those who accept this brand of libertarian freedom can make sense of 'causation' without it ruining their view.