Contemporary Incompatibilism: Libertarianism 247Clarke’s proposed solution is that it is a matter of natural law that the propen-
sities of the reasons to cause actions are the same as the propensities of the
agent- as-substance to cause actions (2003). One issue for this strategy is that a
natural law of this sort would seem to be in a sense brute, since there would
appear to be nothing about the agent- as-substance per se that explains why its
propensities to cause actions match those of the reasons. A second issue is that
the agent- as-substance, in causing an action, would then seem to lack the right
sort of relation to reasons. Intuitively, this agent must be influenced by reasons
in causing an action. However, on Clarke’s integrated view, it would appear that
there could be no such influence, but rather only a correspondence of propensi-
ties of the reasons on the one hand and the agent- as-substance on the other by
virtue of a natural law that’s brute in the sense just specified.
Hence what seems to be needed is an explanation for how the agent- as-
substance could be causally influenced by reasons. O’Connor proposes an
account of this kind, one that employs Fred Dretske’s (1993) distinction between
structural and triggering causes. To illustrate the distinction, the structuring
cause of the bomb’s explosion is the process by which the bomb is made, while
its triggering cause is what detonates it. In O’Connor’s view, reasons are struc-
turing causes of a decision by virtue of structuring the propensities of the agent-
causal power, while the agent- as-substance, in her exercise of this structured
power, is the decision’s triggering cause. The result of the causal structuring by
reasons is the alteration of the propensities of the agent- causal power toward a
range of effects:
While nothing produces an instance of agent- causation, the possible
occurrence of this event has a continuously evolving, objective likelihood.
Expressed differently, agent- causal power is a structured propensity
towards a class of effects, such that at any given time, for each causally
possible, specific agent- causal event- type, there is a definite objective
probability of its occurrence within the range (0,1), and this probability
varies continuously as the agent is impacted by internal and external influ-
ences.... [T]he effect of the influencing events is exhausted by their alter-
ation of the relative likelihood of the outcome, which they accomplish by
affecting the propensities of the agent- causal capacity itself. (O’Connor,
2009: 197–8)The core of O’Connor’s account is that the reasons structure the agent- causal
power by changing the objective probabilities of its propensities toward effects—
toward intention- formations and decisions. Andrei Buckareff (forthcoming)
argues that on this view reasons are causally relevant but are not causally effica-
cious when the action occurs, and this may be a cost of the view.^1
To this account Pereboom (2014) objects that agent- causal libertarianism
cannot accommodate the claim that the propensities of the agent- causal power
are governed by probabilities specified in this way. To answer the disappearing
agent version of the luck objection, the causal power exercised by the agent