Agent Causation, Event Causation, and State Causation

By | March 21, 2007

It is commonly held among philosophers that events are the kinds of things that can most properly be said to serve as causes and effects. On this view, the event of the rock’s hitting the window causes the event of the window’s breaking.

Some, however, prefer to analyze causation in terms of states (or states-of-affairs). On this view, the intrinsic character (or state) of the rock-window system (including the momentum, density, etc., of the rock, the fragility of the window, and the position of the rock in relation to the window) results in a new state of the system, with an altered position and momentum of the rock and a shattered window.

Some draw no clear distinction between state-causation and event-causation, holding them to be virtually interchangeable. Thus, we can conceive of the rock’s breaking the window either as a series of events, each causing the one after it, or as a series of instantaneous states, each causing the one after it.

In a famous essay entitled “Human Freedom and the Self” Roderick Chisholm distinguishes between what he calls transeunt causation and immanent causation. Transeunt causation includes both event causation and state causation, which Chisholm treats as equivalent. Immanent causation, on the other hand, is supposed to be in a distinct category, one that he also refers to as agent causation. Here it is not a state or an event that does the causing but a person or agent. Chisholm tries to illustrate this distinction with a quote from Aristotle:

A staff moves a stone, and is moved by a hand, which is moved by a man. (Aristotle, Physics, 256a)

Here we are to imagine a man pushing a stone along the ground with a stick held in his hand. The man, says Chisholm, is an agent cause of the motion of his hand. The stick serves merely to transmit motion from the hand to the stone. It has no intrinsic efficacy to move the rock. Thus, the motion of the stick, says Chisholm, is merely a transeunt cause of the motion of the rock.

The main problem I have with this is that Chisholm seems to be confusing two different distinctions:

  • An agent’s causing an event (state) versus an event (state) causing another event (state).
  • Something’s (e.g., a man’s) having the intrinsic efficacy to move a rock versus something’s (e.g., a stick’s) having only an extrinsic efficacy to accomplish the same result. (In other words, the stick can only move the rock only if something else is moving the stick.)

The quotation from Aristotle suggests that nothing could happen unless some things were intrinsically efficacious. Such things could provide the starting points of new causal chains. The alternative is to suppose that everything is extrinsically efficacious and thus merely a transmitter of causal efficacy. But if that were so, then where does causal efficacy come from in the first place?

2 thoughts on “Agent Causation, Event Causation, and State Causation

  1. Ian

    To me, the essential difference between state-causation and event-causation is this:

    With state-causation, there are at least two successive points in time at which the cause persists unchanged. This gives us state-state causation

    With event-causation, there are no two successive points in time at which the cause persists unchanged. This gives us event-event causation.

    In the case of event causation, if it is accepted that in a sequence of events a particular cause is an effect of the prior cause, we have two possibilities – an initial case of state-causation or an infinite regress of event-causation. If state-causation of an event seems to be incoherent, then we would seem to be left with the infinite regress.

    This situation is not resolved with the introduction of agent-causation. The infinite regress is simply shifted to the agent – the agent has its own (mental) event-causation continuum, and the infinite regress matter.

  2. Ian

    Perhaps we could refer to these as Static Causation and Dynamic Causation? Static is long-term unchanging whereas dynamic is transient.

    If we draw a representation on a time-line, with static the cause is identical at any two points on the line whereas with dynamic, the cause is always different at any two points on the line.


Leave a Reply

Your email address will not be published. Required fields are marked *