Agent (Immanent) Causation
17-Dec-2013: Original version

Agent causation is a hypothesis related to the existence and nature of free will. The concept that can goes back at least as far as George Berkeley (Berkeley 1710). More recently it is particularly associated with Roderick Chisholm who used the term immanent causation (Chisholm 1964).

The existence of free will is often attacked by a well known argument that goes roughly as follows:

  1. Everything that happens is either caused or not caused
  2. If events are caused then they are predetermined and occur without regard to free will
  3. If events are not caused then they are random and occur without regard to free will
  4. Therefore free will does not exist
There have been numerous attempts to escape this conclusion, including the attempts by some compatibilists to redefine freedom and free will so as to escape point 2 above. An alternative approach is to question the assumptions that arise from the dichotomy in which the argument is phrased.

The term "caused" is normally taken in this context to refer to strict causal determinism: if we know the exact status of the universe at any one moment then we can in theory predict the status of the universe at any future moment. This classical concept of strict determinism has been seriously undermined if not completely discredited by quantum mechanics. However if strict determinism does hold then it is indeed true that free will does not exist.

Point 3 of the argument above is more contentious. "Not caused" is taken as meaning "random". This can be extended to include stochastic determinism of quantum physics. Again, if events are randomly determined then they are clearly not the result of free will.

That seems reasonable. The problem comes from equating 'not caused' with 'random'. This is an unjustified assumption.

"Caused" and "Not caused" are two different levels of categorisation. It's like saying that all animals are either "Cats" or "Not cats". That's a valid statement that would appeal to all cats and many cat owners, however it is inherently misleading. "Not cats" is a higher level of category than "Cats" can include numerous further subdivisions such as "Dogs", "Birds" and "Mice". These alternatives are at an equal level to "Cats" (much to the annoyance of the feline population).

Similarly "Not caused" might encompass more options than just "Random". One of these options might preserve the concept of free will.

One such option is agent causation. This provides a "third way": an alternative to "Random" for events which are "Not Caused" in the strict deterministic sense. It is causation through the free will of a conscious entity rather than through the interaction of unconscious elements of the universe.

How does immanent causation differ from determinism? Is it not just determinism with the "Agent" as the determining factor? No. Agent causation posits a third class of determinism beyond "determined" and "random": "chosen". Events initiated by immanent causation could not even theoretically be predicted in advance. We might make reasonable guesses on the basis of our knowledge of the agent, however we cannot be sure what the agent will decide.

Chisholm describes the difference between "ordinary" determinism and agent (immanent) causation as follows:

"I shall borrow a pair of medieval terms, using them, perhaps, in a way that is slightly different from that for which they were originally intended. I shall say that when one event or state of affairs (or set of events or states of affairs) causes some other event or state of affairs, then we have an instance of transient causation. And I shall say that when an agent, as distinguished from an event, causes an event or state of affairs, then we have an instance of immanent causation."
(Chisholm, 1964)
So what is the "agent"? It is, I contend, consciousness. The (currently unknown) mechanism through which it acts is free will.

This might sound like spooky dualism, especially given Berkeley's use of the word "spirit", but it need not be so. Agent causation could exist alongside and interact with the familiar physical universe through some currently unknown scientific mechanism.

The idea of agent causation is logically sound. The question is whether or not it is correct. Does agent causation actually exist? If it does then it is certainly outside of our current scientific understanding. But without agent causation we are left with no answer to the original argument; we have no free will and hence no moral responsibility. Derk Pereboom summed it up thus:

"Agent causation is a coherent possibility, but it is not credible given our best physical theories. Consequently, no position which affirms the sort of free will required for strong accountability is left standing."
(Pereboom 2002)
Pereboom believes that even without strong accountability we can still have some form of moral responsibility, a position with which I disagree. Rather than accept his (to me) depressing conclusion I prefer to think that agent causation will one day be made credible by an advance in our physical theories.

References:
Berkeley, George. 1710. A Treatise Concerning the Principles of Human Knowledge
 Chisholm, Roderick M. 1964. Human Freedom and the Self
 Pereboom, Derk. 2002. Meaning in Life Without Free Will. Philosophic Exchange 33