Abstract :
I offer an argument regarding chances that appears to yield a dilemma: either the
chances at time t must be determined by the natural laws and the history through t
of instantiations of categorical properties, or the function ch( ) assigning chances
need not satisfy the axioms of probability. The dilemma’s first horn might seem like
a remnant of determinism. On the other hand, this horn might be inspired by our best
scientific theories. In addition, it is entailed by the familiar view that facts about chances
at t are ontologically reducible to facts about the laws and the categorical history
through t. However, that laws are ontologically prior to chances stands in some tension
with the view that chances are governed by laws just as categorical-property instantiations
are. The dilemma’s second horn entails that if chances are in fact probabilities,
then this is a matter of natural law rather than logical or conceptual necessity. I conclude
with a suggestion for going between the horns of the dilemma. This suggestion involves a
generalization of the notion that chances evolve by conditionalization.
