Introduction
My project is to defend a toy argument for metaphysical fatalism. This paper has two sections. In the first section, I explain a couple of assumptions that I make about the nature of time. Presumably, these assumptions are plausible and widely held – or, at the very least, they are reasonable positions to hold about the nature of time. In the section after that, I outline and explain a thought experiment designed to bring our fatalist-friendly intuitions to the forefront.
However, before I get too far into the philosophical weeds, I want to briefly explain the claim at the center of my project. This claim is called Metaphysical Fatalism – or just fatalism, for short. Crudely sketched, fatalism is simply the claim that everything that happens must have happened just the way it did. In other words, there is some kind of necessity to the way the world is configured across time and space, such that it could not have been arranged, in either dimension, in any other way. Put differently, to say things about how the world might have been arranged otherwise (e.g. that claim that I could have worn a green shirt today, rather than a red one) is to say something false. In the following sections, my project is to construct and defend an argument for fatalism.
§1.
There are a few highly plausible, common sense claims about the nature of time that most of us accept, and I think rightly so, without thinking twice. Consider the following:
The events, individuals, and so forth, which reside in the past, are unchangeable. This seems to be about as secure a claim as one could hope to find. Call this Past.
We have all sorts of strongly held intuitions that reinforce this claim. We take it, for instance, that there is little reason to be upset over mistakes made in the past in the following sense: there is nothing that can be done to change the events that reside in the past. Even though we might be able to learn from them, there is little reason to dwell on them for anything other than lessons to be learned; simply put, events cannot be altered, once they recede into the past. Consider that people often terribly regret their poor choices in the past, in part, because of their consequences in the present, and future, but also because they are unchangeable, once it has past. They intuitively know that once in the past, their choices cannot be undone.
It might be tempting to invoke time machines, as a way, in principle, to change the past. So, it might be argued that a time machine, if they could be had, would afford us the opportunity to go back in time and change moments in the past. I could, for instance, decide to go back and do a better job in my relevant classes, so I could alter my present in a way that would allow me admission to be a better graduate school. This solution is suggested by all sorts of movies, television shows, and so forth, which allow characters to alter the present (or the future), in virtue of their changing something in the past. This might be dubbed the Time Machine Worry – or just Machine, for short.
However, Machine is not much of a worry. The reason for this is simple. We don't really think that we can go back and change the past. Indeed, our intuitions, even though a bit foggy, provide good evidence that we are committed, even if we allow the possibility of past-oriented time travel, to the unchangeable nature of the past. Consider, as a first stab at an explanation, you try to go back in time to kill your grandfather. Presumably, in light of his genetic contributions to your mother, he was a necessary condition of your existence; if you exist, then it must be that your grandfather does also. Surely, you do exist; if you didn't, it would be difficult to explain your ability to travel back in time in the first place. After all, people who do not exist cannot go back in time to change the past. This means that you cannot alter the past, at least where your grandfather and his reproductive activities are concerned.
It might seem that these killing-your-grandfather type cases are the exception, rather than the rule, when it comes to changing the past via time travel. If so, these sorts of cases would do little to show that changing the past cannot be done. There are general problems with proposed changing-the-past cases; I am going to argue, against Machine, that either the past already includes a visitor from the future, or the change to the past moment makes the moment different from its temporal counterpart before it was changed.
To illicit the relevant intuitions, consider the following example:
Suppose that Stewie wants to travel back in time to warn his younger self about doing poorly in school. Perhaps this is because Stewie is tired of having to work menial jobs because he failed out of school. Stewie travels back in time to have a long, serious conversation, with his younger self, about the importance of finishing school. Younger-Stewie seems to understand the point; he promises his older self that he will do better in school. Satisfied, Stewie uses the time machine to return to the present. But, upon returning, he notices that everything is the same.
Call this example Grades. It may, or may not, be immediately obvious why Stewie failed to change his fate, by going back into the past and convincing his younger self to do whatever he failed to do the first time around. The reason that Stewie failed is simply this: if Stewie succeeded in convincing younger-Stewie to do a better job in school, and secure a degree, then when the future moment rolled around, and he had to convince his younger self to do better, because of how poor he was in the future, Stewie wouldn't have a reason to make that temporal trip. However, it is supposed to be that the explanation for why younger-Stewie did better in school, thereby improving the prospects of his older self, is simply that older-Stewie went back in time to convince younger-Stewie to do just that. If older-Stewie was doing better, he thought, because he did better in school – all the while unaware of the trip his other self had made – then he would lack the motivation to go back in time and convince younger-Stewie to finish school; but if this trip into the past was the explanation for younger-Stewie's improved grades (i.e. younger-Stewie would not have improved grades without the temporal trip), then younger-Stewie would not do well scholastically without older-Stewie going back in time to lecture him. Then again, if younger-Stewie did poorly in school, without his lecture from older-Stewie, then older-Stewie would have failed to secure a better financial situation. This time-traveling produced paradox is easily resolved by the following:
If fact A, in the past, is the explanation for your traveling back in time, to change the past, then you cannot go back in time to change A.
This is a simple rule that helps us avoid time traveling paradoxes like the one we just saw. However, the savvy reader might think there are still instances where we could travel back in time and change the past, which does not violate the rule above. We might think of a simple example: you get into the time machine, squish a butterfly, and then return to the present; there is nothing about squishing a butterfly in the past, that would prevent me from wanting to travel back in time, while in the future, to squish the butterfly. Simply put, unlike Grades, this example of time traveling to change the past does not violate the rule I just laid out.
There is a serious worry with the butterfly example. The worry is this: either the past already includes you squishing a butterfly, after arriving by time machine – in which case, it seems that you didn't change the past at all – or, you transformed the moment in the past into an entirely different moment. Call the former option, Alpha, and the latter, Beta. I want to explain why Alpha and Beta do not help those who think the past could be changed, if only they could travel to it in a time machine. Suppose we start with Alpha: it turns out that the past already contains an older version of you, who traveled from the future, to squish a butterfly. The intuition behind this way of looking at the example can be stated as follows. If the past is something that has already happened, and you went to the past to squish a butterfly, it must be that you squishing a butterfly already happened. Thus, Alpha does not show that you could travel to the past, with a time machine, to change it.
Consider Beta. There is a serious problem for Beta; let's call it the locating-the-change problem. Roughly, the problem might be explained as follows. Suppose we consider the timeline where you travel back in time, to squish a butterfly that was previously un-squished. After going back in time, and squishing the butterfly, then it seems that your timeline always has you in the past, squishing a butterfly. So, if your traveling-to-the-past-squishing-a-butterfly action doesn't create a new timeline, then we have to ask what happened to the old timeline where you didn't squish the butterfly. Did it just vanish? It seems like we would have to say so; but then there is the question of where it went. Think about it like so: suppose you are in the way distant future. From your perspective, it would seem that the past always contained a squished butterfly. The sequence of events would look like it was arranged as follows: first, there is someone, from the future, squishing a butterfly, then someone enters a time machine and vanishes, then, finally I arrive, in the distant future, to observe the past. (A quick and dirty way to think about this problem is this: the past there is how things already happened. So, if you changed the past, it must have already happened. Thus, however the past is configured it has already been changed. This is so because the past simply is how things already happened. For instance, if I travel back to the past to change my grades, then I am traveling to that part of time that has already happened; so, whatever I do has already happened. It might be argued that we could say the same thing about the future, i.e. that the past and future are similar in this respect. The future is just that part of the timeline that is going to happen in such-and-such a way. So, for instance, if I go to the future to change how things are arranged, go back to the present, then, after enough time has passed, I could come upon myself changing the future). Put differently: there doesn't seem to be a moment where someone wasn't already in the past, having arrived by time machine, squishing a butterfly. This is a strange view, especially since Beta was supposed to provide an account of changing the past via time travel.
This provides good reason in support of Past. If something lies in the past, it cannot be changed, even if we have a time machine.
Before moving on to the next section, I want to explain another common sense aspect to the nature of time. This is the claim that time itself is not efficacious in the following sense: time itself neither increases, nor diminishes the causal powers of anything. The passage of time is only a conduit that allows other processes (e.g. entropy), to either supplement or diminish the causal powers of those individuals, and such, which reside in that that section of the timeline. For example, the passage of time, by itself, does not transform me into a frail, feeble old man. After all, if my aging processes were frozen, such that time could pass, even though I failed to age, this illustrates that temporal passage, on the one hand, and the mechanics of aging, on the other, are distinct things. Even though it might be that time must pass in order for my gaining processes to play out, this does not mean that they are the same thing. Put differently, time might pass without any change in my age. My losing or gaining the power to do something must be the result of something other than the mere passage of time. If I lose the ability to do X, over time, then there must be an explanation for that loss other than the mere passage of time. Call this claim Toothless.
Comparing the following situations that motivate Toothless. The first situation, I mistakenly leave the accounting computer in cleaning mode, and go home for the night. During the night, the computer program I left on, wipes and reformats the hard drive, destroying the only copy of the company's financial records. It seems that although the passage of time enabled the computer-wiping program to do its job, it was the program itself that is efficacious in destroying the company's financial records. Suppose that instead of leaving the computer-wiping program on, and going home, I remember to turn the computer off before leaving the office. Clearly, we would not expect that computer-wiping program to wipe the company's financial records, without interference, on the basis of time passing. But it seems that the computer program either running or not is the only difference between these two situations. This strongly suggests that the computer program itself is efficacious, rather than having anything to do with the mere passage of time.
If I have elicited the appropriate intuitions, then we have some reason to accept Past and Toothless. In the next section, I want to bring to the fore an argument for metaphysical fatalism.
§2.
Suppose that Bob the physicist built a time shuffling machine. The machine works similarly to a card shuffler in that it takes time segments, much like a card shuffler takes cards, and changes their sequence, randomly. Call this machine Shuffle. We might suppose that there are a few restraints on Shuffle. The first is that Shuffle is not allowed to reside in any of the moments in which it is shuffling. So, if Shuffle partially resides in time slice Alpha, then Shuffle is prohibited from moving Alpha around. The second is that Shuffle is not allowed to change anything except the sequence in which time slices are arranged. Put differently, Shuffle cannot alter anything within the time slices themselves, rather it is only allowed to alter the sequence, on the timeline, in which those time slices occur; although, I will explain in a bit why this is not actually a problem. In light of what I've said about the past (i.e. if x exists in the past, then it must have always been that x existed in the past) it must be that in whatever time slices Shuffle resides, it cannot move from those slices; it must always exist, where it exists. This is because Shuffle is a machine capable of shuffling the sequence in which moments occur, not a time machine.
For our purposes, I want to ignore the implications of changing the sequence of moments. There might be some who worry that we would be changing the past (or the future), if we moved around moments such that causes were preceded by their effects – not to mention the worry that changing the sequence where certain events happened is merely an instance of changing the past. For instance, suppose in our timeline there is one event followed by another: the first is my flipping a light switch, and the second is the light coming on. Presumably, the right story to tell about the relationship between the first and second events is that something about the first moment caused something about the second moment; specifically that my hitting the light switch caused the light to come on. Suppose that someone used Shuffle to change the order of moments I've set up. The light would come on first (the effect), and in the next moment, I would be flipping the light switch (the cause); if you think that my flipping the light switch did a good deal of explaining why the light came on. But if you think that the flipping of the switch caused the light to come on, but the light came on before someone flipped the switch, after someone shuffled the moments around, then it looks like something screwy is going on. But I don't care about that.
My worry amounts to this: it seems like we alter the past, if we have a machine that could shuffle the order in which moments happen. After all, if Shuffle is able to change the sequence of moments in a timeline, then it seems that Shuffle is able to change the relational facts about the past. For those who are unsure, a relational fact is a fact about how one thing relates to another; for instance, Mary being shorter than Bob is a relational fact because it's a fact about how Mary relates to Bob with respect to their respective heights. Likewise, if Shuffle can change the sequence of moments in a timeline, then it seems that Shuffle can change the facts as to how which past moment occurred when, in relation to other past moments. But, as we have already seen, there is good reason to suppose that the past cannot be altered. Thus, it seems like there cannot be a time shuffling machine. Call this the No-Shuffle Worry.
There are several ways to deal with the No-Shuffle Worry. The first of these is to deny that Shuffle changes the past. If you recall, Shuffle does not change anything about the moments themselves, but only their relationship to one another. We might have the intuition that changing the past involves changing something about the moments in the past, not merely changing the order in which they happen. Although someone might take this particular move, I do not find it all that compelling; it concedes too much ground to the critic, and thus undermines much of the appeal of using an idea like Shuffle to defend fatalism in the first place. Call this approach Denial.
The next two approaches have greater promise.
The first of these is simply to pretend that Shuffle is a machine that could in principle exist. Call this approach Pretend. I only need for us to imagine a machine such as Shuffle, even if there really could be no such machine; the point of imagining this machine, then, is just to bring features had in common by all moments, whether they exist in the past, present, or future – it doesn't matter, for the sake of my argument, whether Shuffle could exist, even in principle, as long as imagining that it could exist helps to clarify my argument. The purpose of Pretend is to clarify, not justify, such that once my point has been understood, the Shuffle thought experiment might be kicked away.
Rather than rely on Pretend, or Denial, I want to introduce the approach, to the No-Shuffle Worry, on which I'll rely, for the remainder of this paper. The approach is simply to stipulate that we are discussing a universe where there is a single moment, located in the future, which Shuffle can then move around on a particular timeline. (For those who are sceptical that there are moments in the future, consider the following argument: if you can travel to x, then there must be an x to be traveled to; after all, you can't travel to nowhere. It is possible for you to travel into the future. Thus, it must be that there are already future moments.) This is while Shuffle resides in a distinct moment causally unconnected to the single, future-located moment. Call this the approach Single. The reason that Single is a superior approach to Denial and Pretend is simply that Single tells a story where Shuffle could, in principle, exist, even if there are plenty of details that would have to be worked out. Simply put, it is a more permissive story to tell, then either Denial or Pretend.
Finally, we are ready to consider an argument for fatalism. The argument amounts to this:
Consider a single moment that exists in the future. In this moment, I am putting on a red polo shirt, before going out for the evening. Call this moment Temporal. Suppose that Shuffle moves Temporal from the future, into the past. Surely, on the basis of our accepting Past, we already accept that just in virtue of residing in the past, Temporal is unable to be altered. So, once Shuffle moves Temporal into the past, there is nothing about that moment that could have been different. But, there is something to notice about Shuffle moving Temporal from the future, to the past: the mere passing of time is not efficacious. This was established earlier as the claim we dubbed Toothless. Unless there is something extra-temporal happening to Temporal, then the passage of time itself, or moving the position of Temporal, on a timeline, should not do anything to alter the properties of Temporal. However, we accept that in the past Temporal has the property of being-unalterable. So, if Shuffle didn't change anything about Temporal, except its position on the timeline – and the position on a timeline, had by a moment, is not efficacious (from Toothless) – then Temporal never gained, or lost, any properties. So, if in the past Temporal has the property of being-unalterable, it has always had that property. This story about Temporal having the property of being-unalterable generalizes to all moments. Thus, there is never a time, which anything could be other than it is. This is fatalism. Call this the Time Shuffle Argument – or just Time Shuffle for short.
Before concluding, I want to emphasize an aside. It should be apparent, upon reflection, that Time Shuffle need not make references to any particular laws of physics, in order to serve as an argument for metaphysical fatalism. We only need accept a few basic principles, all of which seem plausible on their own; they seemingly lend support to fatalism. My project has not been to show that we must accept fatalism, but rather than there are good reasons supporting the view.