Contemporary Incompatibilism: Libertarianism 245Meghan Griffith (2005) employs an epistemic analogue of this line of reason-
ing in her response to Peter van Inwagen’s challenge to agent- causal libertarian-
ism. Van Inwagen (2000) argues that on the supposition that I am an agent- cause,
if in some situation I know the objective probability of my remaining silent is
0.57, and of my speaking and not remaining silent is 0.43, I fail to be in a posi-
tion to promise to you that I will remain silent. My failing to be in a position to
make this promise indicates that on the agent- causal view I lack the sort of
control required for responsibility for remaining silent. Griffith argues that while
on the event-causal conception this kind of reasoning is plausible, on the agent-
causal position it is not. For even if the causally relevant agent- involving events
issue in these probabilities, I as agent- as-substance have the power to definit-
ively determine whether the decision to remain silent will occur by causing it,
and knowledge of how I will exercise my agent- causal power would put me in a
position to make a promise. In these circumstances I could come to know which
decision I will make, even if full mastery of the causally relevant agent- involving
events alone will not yield this knowledge.
Another sort of luck argument might be pressed against the agent- causal lib-
ertarian. When an agent- as-substance causes a decision, an event of the follow-
ing type occurs:
G: A’s causing D at t.But such an agent could not cause events of type G, for it’s absurd to claim that
the agent- as-substance causes herself to cause a decision. Since this agent could
not cause events of type G, she cannot be responsible for them. And because the
agent cannot be responsible for such events, she cannot be responsible for the
choices embedded in them. If an agent cannot be morally responsible for her
causing of her choices at times, neither can she be responsible for her choices.
Ginet (1997: 91) presents a version of this objection against O’Connor’s
agent- causal libertarianism. In O’Connor’s view, in a case of agent causation,
the agent’s causing at t an event e is not itself an event which has a sufficient
causal condition. Consequently Ginet sees him as vulnerable to the objection that
this more complex event will be an undetermined event over which, according to
the agent- causal theorist’s own objection to event- causal libertarianism, an agent
cannot have sufficient control for moral responsibility.
Roderick Chisholm’s (1971) reply to this type of concern is that when an
agent causes a choice by an exercise of her agent- causal power she does indeed
cause an event of type G, for which she can be morally responsible. When an
agent acts freely, what she does most fundamentally is to cause a decision to act.
She can be morally responsible for this decision because she causes it by her
agent- causal power, and such causing supplies the control required for moral
responsibility. Less boldly, note that it is a logical consequence of the agent’s
causing a decision that an event of type G occurs. It would thus follow logically
from the fact that at tn Ralph agent- causes his decision to move to New York
that the more complex event Ralph’s causing at tn a decision to move to New