Consider the following argument for fatalism (let t* = a time 1000 years ago, and t = tomorrow):
1) It was true at t* that Susan goes to Anstruther at t. (Assumption)
2) Necessarily, if it was true at t* that Susan goes to Anstruther at t, then Susan goes to Anstruther at t. (From 1 and the necessity of logical consequence)
3) Susan never had and never will have a choice about whether 1 is true. (From 1 and the fixity of the past)
4) Susan never had and never will have a choice about whether she goes to Anstruther at t. (From 2 and 3)
5) Therefore Susan is not free to refrain from going to Anstruther at t. (From 4)
The idea behind premise 3 is that the past if fixed, over and done with, and so whatever was true about the past will always be true, regardless of our choices. So we construe the fixity of the past in terms of an agent’s choices, and it is fairly uncontroversial to affirm the thesis that an agent has no choices about matters 1000 years past. It is somewhat more controversial to assume, as the argument clearly does, that there is truth about the future. Rejecting this assumption, however, involves rejecting classical logic for a limited domain of propositions (those about the future), solely in order to avoid fatalism; and this just seems ad hoc. So I will assume that there is truth about the future.
The inference pattern relied on in moving from premises 2 and 3 to 4 looks something like this (where Np = ‘p and the agent never had and never will have a choice about p’):
Necessarily (if p then q)
I assume that this is a valid inference pattern. It is difficult to see how there could be a plausible counter-example to it. Furthermore, an objection to FAT which relied solely upon a questioning of this inference pattern would not be very philosophically satisfying.
So how should we respond to FAT? Consider how the argument might call for a divergence of responses depending upon one’s temporal ontology. The standard anti-fatalist metaphysical explanation that is brought to bear against such arguments is that future truth does not necessitate or fix future or present actions; rather, future and present actions make propositions about those actions true. According to Rea (2006), the anti-fatalist explanation is a valid response to FAT, but only on an eternalist ontology according to which past, present, and future are all equally real. His argument for this claim is that premise 1 is best interpreted as a tenseless truth that is eternally made true by Susan’s going to Anstruther at t. Thus, on the basis of this claim, the eternalist can reject premise 3: Susan does have a choice about whether 1 is true. But, Rea argues, this response is not open to the presentist (i.e. someone who thinks that only present objects and events exist), since, for her, Susan’s going to Anstruther does not exist until it occurs, and so is not available to make premise 1 true. So, regardless of whether 1 is true at t* or eternally true, since Susan’s going to Anstruther neither exists at t* nor exists eternally, she has no choice about its truth, and the fatalist conclusion follows.
I think that Rea is correct in arguing that the eternalist can resist FAT in this manner, but I disagree that it’s curtains for the presentist. What is it about the past that makes us think that it is fixed, or that we (now) have no choice about it? What is it that grounds the intuitive notion of the fixity of the past? Plausibly, for the presentist, fixity is grounded in the occurrence of concrete events, and there is no concrete event that corresponds to ‘Susan goes to Anstruther in 1000 years’. There is at most the abstract state of affairs, it’s being the case that Susan goes to Anstruther in 1000 years; and whatever this state of affairs amounts to, it certainly is not something that occurs at t*. After all, it is not as though when t* is present, it’s being the case that Susan goes to Anstruter in 1000 years happens. Thus, intuitively, this state of affairs is not fixed, and therefore neither is its corresponding proposition. What ultimately grounds the truth of ‘Susan goes to Anstruther in 1000 years’ is the concrete event of Susan’s going to Anstruther at t* + 1000, and this event is not fixed until it occurs. Granted, there is an issue here of whether the presentist can ground the truth of premise 1, when it’s corresponding event does not yet exist; but there are a range of solutions that the presentist might offer to this ‘grounding’ problem. As long as these solutions do not beg the question against the fatalist (and I see no reason why they should), then the grounding problem can legitimately be deemed independent of worries about fatalism.