There’s a simple and valid argument for fatalism based on a proposition which most analytic philosophers would accept. The assumption is this
- There is an actual world, alpha, which contains a complete history. Unlike other merely possible worlds, alpha is the possible world that “obtains”.
For the sake of argument I take this assumption for granted.
Now, fatalism can be understood as the doctrine that no events have an intermediate chance of occurring. By a ‘chance’ I mean a single-case objective probability. By an ‘intermediate’ chance I mean a value between zero and one. If the chance of an event is one, then it is unpreventable–it’s guaranteed to happen. If the chance of an event is zero, then its non-occurrence is unpreventable–it’s guaranteed not to happen. Fatalism simply says that, for any event, its chance of occurring is either zero or one.
Now, consider the actual world, alpha. This includes a complete history. Hence, for every possible event E, either alpha entails that E occurs, or alpha entails that E doesn’t occur. If alpha entails that E occurs, then the chance of E’s occurring given alpha equals one. If alpha entails that E does not occur, then the chance of E’s occurring given alpha equals zero. Using CH() to represent the chance function, this means that for arbitrary E, either
- CH(E | alpha) = 0, or
- CH(E | alpha) = 1.
Now, what is the chance of alpha? That is, what is the chance that alpha obtains or that alpha is the case? Obviously, it’s got to be one. In general, the chance that anything is the case has got to be either zero or one. Consider my sitting at time T. If it is the case that I sit at t, then the chance that I sit at T is one. It’s too late to prevent it. Likewise, if it is the case that I stand at T, then the chance that I sit at T is zero. It’s too late to bring it about that I sit at T. Thus, since alpha (in its entirety) obtains, and thus is the case, we get
- CH(alpha) = 1.
But from 2 and 3 there follows
- CH(E) = 1.
And from 1 and 3 there follows
- CH(E) = 0.
Hence, E is not a future contingent. Since the argument holds for arbitrary E, fatalism is thereby established.
To avoid fatalism the initial assumption must be rejected. We must either deny that there is an actual world (i.e., we must deny that any possible world which includes a complete history obtains), or we must deny that possible worlds must include a complete history, in particular, a complete future history. Call that denial the modal openness of the future thesis. I maintain that the future is modally open. As contingencies are resolved, the modal changes. Things that were possible may not now be possible. Things that are necessary may not always have been necessary.