The arrow of chance

Some say that, while the future is often chancy, the past cannot be. I disagree: I say that the past can be chancy, in some possible situations. Specifically, I say that past chanciness is possible in some cases of backwards causation; that is, in some cases whereby a later event causes an earlier one. In this dissertation, I defend the possibility of past chanciness, along with a view of chance that permits it. First, I set out a commonly endorsed and intuitively plausible view of chance, according to which chance is intimately related to time, and the past cannot be chancy. Then, using a particular case of backwards causation, I raise two problems for this view, and argue that it ought to be rejected. Finally, I propose and defend an alternative view, according to which chance is intimately related to causation, not time. According to this view, past chanciness is possible: it's not the past that can't be chancy, but rather the causal history. I argue that this causal view of chance delivers many of the benefits of its temporal cousin, while correctly handling cases of backwards causation, including our problem case. I defend the causal view against potential objections, and conclude that we ought to accept it