minutes. An adroit theologian might even be tempted to try to exempt sinful functions
from God's support. To be sure, this version will invoke questions analogous to those
listed earlier in the paragraph. But worse yet, it rests on a faulty assumption. A function
may outlast some of its ancestral hosts, but no part of it can occur without being
embedded in some host or other. And those hosts must be sustained in their existence.
The last photons emitted from the sun immediately prior to its annihilation must
themselves be sustained in existence if they are to irradiate my garden eight minutes
hence: after-effects do not earn an exemption just in virtue of being after-effects. More
generally, a function must be a function of some ensemble or sequence of things. If the
function is spread over a period of time, the things on which the function depends must
be kept in existence long enough to host the function. Sins are no exception; they must
have perpetrators. Even if we suppose that a sinful act is freely committed, in some
strong, indeterministic sense of freedom, that supposition does not gainsay the fact that
the sinner must be kept in existence long enough to commit the sin.
It is not obvious, then, that intermediate positions are philosophically better off than
divine conservation. But how bad is the case against divine conservation? Recall that two
considerations were raised against it. One rested on a comparison to the plight of Atlas.
Theists are entitled to regard the comparison as invidious. Atlas's chore is burdensome
because it is imposed as a punishment and his strength is limited. But God is supposed by
most theists to be a being of unlimited power and a being against whom no other being
can prevail. Thus, it is hard to see how, for such a being, the conservation of creation
could be exhausting drudgery. Conservation would be a problem if it took all of God's
unlimited power to create and conserve something ex nihilo, or if God inflicted the
burden of creation on himself as some sort of act of supreme self-flagellation. Neither
hypothesis seems remotely plausible.
The second worry about divine conservation was that it appears to confer on theism a
particularly nasty version of the problem of evil. Theists typically concede that God
permits evil to occur while denying that God commits evil. It is possible to see too much
moral difference in the distinction between doing and allowing to happen. But in this
case, the strategy of downplaying the difference is a dangerous one for the theist to
pursue. It might have the unhappy result of assimilating divine doing to a type of mere
passive allowing. Alternatively, it might promote divine allowing up to the level of active
doing, which would validate the second worry. I suggest a different strategy, one more
narrowly tailored to divine conservation. The strategy is to argue that divine conservation
does not increase the problem of evil for a theist who is willing to grant that God permits
evil to occur.
Let us begin by considering this principle:
(1) If x keeps y in existence while y does φ, then x is also responsible for doing φ.
(1) is surely false. An oxygen tank may enable an arsonist to continue breathing while
setting fire to a building, but the arsonist's crime cannot be imputed to the tank. If some
modification of (1) is going to be plausible, it must incorporate appropriate restrictions
into x's knowledge, x's power, even the sort of responsibility ascribable to x. Skipping a
few intermediary iterations, we can examine this descendant of (1):
(1′) If x knows (a) that she is keeping y in existence while y does φ, (b) that y's doing φ is
morally impermissible, and (c) that she could have terminated y's existence but chose not
to, then x has done something that is morally impermissible.
nextflipdebug5
(nextflipdebug5)
#1