252 Contemporary Incompatibilism: Libertarianism
then for a large enough number of instances the possible actions in our class
would have to be freely chosen close to 32 percent of the time. Then, for a suffi-
ciently large number of instances, the possible actions whose physical compon-
ents have an antecedent probability of 0.32 would almost certainly be freely
chosen close to 32 percent of the time. However, if the occurrence of these phys-
ical components were settled by the decisions of libertarian agent- causes, then
their actually being chosen very close to 32 percent of the time would result in a
coincidence no less wild than the coincidence of possible actions whose physical
components have an antecedent probability of about 0.99 being chosen, over a
sufficiently large number of instances, close to 99 percent of the time. The sug-
gestion that agent- caused free choices would not diverge from what the statisti-
cal laws predict for the physical components of our actions runs so sharply
counter to what we should expect as to render it incredible (Pereboom, 1995,
2001, 2014).
If the libertarian agreed, she might propose that there are indeed departures
from the probabilities that we would expect given the physical laws, likely to be
found in the brain. Roderick Chisholm (1964) suggests a position of this sort.
Steven Horst (2011) points out that physics, at least as we presently find it, fea-
tures no departure- free laws when they are construed as describing actual
motions. The law of gravity, for example, will only result in exactly accurate
predictions of motions if there are no other forces, such as electromagnetism, at
play. The better view, according to Horst, is to interpret the laws as governing
causal powers, which in the case of fundamental physics, are plausibly forces.
On this conception, laws do not primarily describe motion, but rather character-
ize causal powers. Horst argues that this conception of the laws is friendly to a
libertarian conception of free will, since any law should be understood as in prin-
ciple open to the existence of powers not described by that law, and the agent-
causal power would be a candidate.
An objection to this proposal is that we would seem to have no evidence that
departures from the known physical laws occur, say in the brain. The libertarian
could reply that we have phenomenological evidence that we are indeterministi-
cally free (Deery et al., 2013), and this yields evidence that the divergences in
question in fact occur. Nothing we know rules out the claim that we are undeter-
mined agent- causes and there exist such divergences, and it may be wise for the
libertarian to develop this approach.
10.11. Is Libertarian Agent Causation Required for Agency?
According to Martine Nida- Rümelin (2007) and Helen Steward (2012), libertar-
ian agent causation is required not just for moral responsibility, but for agency
itself. Here is how we understand the general concern about agency that gives
rise to these views. On the Davidsonian model of agency, action is caused by
mental states or events, and not fundamentally by agents as substances (David-
son, 1963). As Velleman casts this view: