What Tim Can't Do
Once more on the ineptitude of time travelers
Tim’s Grandfather was an evil man; he stole candy from babies on a staggering scale! Tim, on the other hand, is an adept inventor, morally aghast at his grandfathers actions. Tim will try to stop his grandfather, whatever the cost. In order to achieve his goal Tim invents a time machine and travels to 1930, ten years before his birth. Tim finds his grandfather asleep in his bed. Tim draws his gun (which he brought back in his time machine) and aims it at his grandfather. Nothing, it seems, is stopping Tim from killing his grandfather. A bullet is chambered, Tim’s aim is true, grandfather is not encased in ann inconspicuous bullet proof case. Nonetheless, we might wonder: is Tim able to kill his grandfather?
Well of course! Tim’s pointing a gun at him! Tim is perfectly competent, he’s practiced for years. All Tim has to do is pull the trigger and no more grandfather.
Well of course not! That’s his grandfather! If Tim killed his grandfather, there would be no Tim! No Tim, no grandfather killer, no grandfather killer, no dead grandfather.
So Tim can, but he can’t. This puzzle, the grandfather paradox, seems to constitute a reason for thinking that time travel is impossible. If time travel was possible (so the argument goes) a contradiction would be true: Tim could, and couldn’t, kill his grandfather. This seems to present an apriori argument against time travel. The existence of an apriori argument against time travel is itself a bit of an oddity. The possibility of time travel seems like the sort of thing that could be empirically confirmed, or disconfirmed, by looking to our best scientific theories. If a suggested theory allows for the existence of closed time-like curves, and rather cooperative worm holes, why should a presumptuous philosopher barge into the physics department and shut the whole business down? What to do...
In The Paradoxes of Time Travel, David Lewis articulated what is now the canonical response to the grandfather paradox. Lewis argued that we equivocate when we invoke Tim in our refutation of time travel. We use ‘can’ (or ‘possibly’) in two different senses. When we evaluate counterfactuals (e.g. what would happen if Tim tried to kill grandfather?) we hold some facts fixed, and allow others to vary. According to one set of facts — concerning Tim’s competence with his firearm, and apparent opportunity for grand-parricide — he is able to kill his grandfather, but concerning another (broader) set of facts — that Tim’s grandfather survives past 1930 so that he might continue the lineage — Tim is unable to do so. If we don’t equivocate, the puzzle evaporates. If time travel is possible then it is true, in one sense, that Tim can kill his grandfather, and false in another sense; but it is not true in one and the same sense that Tim can, and cannot, kill his grandfather.
So far, we agree. There is no apriori argument against time travel. The appearance of contradiction comes from equivocation, and once we have our heads on straight, we should realize our presumptuous philosopher is worthy of his name. Yet there is an itch. Which is it? Can he or can’t he? Tim, I mean. Suppose time travel is possible. We use modal words like ‘can’, and ‘possibly’ in everyday contexts. I can open a can of soup, but I can’t open a jar of pickles—I lack the herculean strength necessary. Maybe we are ‘lured into equivocating’ by the fact that Tim is a time traveler, but once we recognize the equivocation I want to ask again: Can Tim kill grandfather, the same way I (can opener in one hand, Campbell’s in the other) can open a can of soup?
You may think this question answered: “On one delineation of the facts, yes, but on another, no.” I present a question in turn: can anyone move faster than the speed of light? God could, maybe, but any human? Obviously, no. It’s impossible to move faster than the speed of light, so no one is able to do it. But I’ve tricked you; faster than light travel is only impossible if we hold fixed the laws of nature. If we allow the laws of nature to vary, it is clearly possible for someone to move faster than the speed of light. Is that really a trick, though? We would never say that any actual human is able to run faster than light in ordinary contexts. If Smith could make it to the recital by moving faster than light, but not otherwise, we wouldn’t say that Smith was able to make the recital. The lesson? There are some delineations of the facts that are too permissive, to comport with our ordinary usage of ‘ability,’ even if we sometimes adopt those delineations in other contexts. Some delineations are too narrow; we would never hold fixed all the facts when evaluating our abilities, lest we have no abilities at all. So there is an upper and lower bound — too hot or too cold — and if we cannot possibly successfully perform an action within those bounds, then we lack the ability to perform the action given what we ordinarily ask when we wonder whether someone has an ability or not.
Here is where I depart from Lewis. Tim can’t kill his grandfather, he lacks the ability. There is, of course, a delineation of the facts on which he can; but this delineation would be far too permissive to comport with ordinary usage, just as permissive as one which would allow the possibility of faster than light travel. Tim can only kill his grandfather if the laws of nature are broken and no one has the ability to do something which can only be done if the laws of nature are broken. So Tim can’t kill grandfather.
The point is easiest to see when analogized to a similar case: Suzy is much like Tim, except she hates the actions of her younger self (just as much as Tim hates his grandfather). Suzy uses a time machine to return to a moment right after her birth, when her baby self is alone in the nursery. Suzy approaches her crib, poised to suffocate her younger self. She gets closer... closer... Suzy is well positioned to commit retro-suicide. Is she able?
No, Suzy cannot commit retro-suicide, she cannot kill baby Suzy. Suzy lives long enough to invent her time machine, and return, so we know Suzy fails, but there is more than that. I had cheerios for breakfast this morning, but I could have had something else. I could have had frosted flakes. If I had frosted flakes, other facts may have differed, there may have been variations in the laws of nature or changes in the near or distant past, but I could have had frosted flakes even if the laws were the same. A difference in my choice of breakfast necessitates no miracle (a miracle is an event inconsistent with the actual laws of nature). Every world where Suzy succeeds must be one where her action is undone. After killing her baby self, a miracle occurs, and baby Suzy is resurrected. Suzy must live on to invent the time machine, so baby Suzy can’t stay dead. If Suzy is to succeed, her success cannot be permanent, and her success can be undone only by miraculous fiat. We can vary the history as much as we like, we can change when Suzy time travels, or when she is born, or facts about the distant past, but in all scenarios that hold the laws fixed, Suzy will fail to kill her baby self. There is, of course, a delineation of the facts according to which Suzy has the ability to kill baby Suzy. Let the laws range, and Suzy will have no issues. As we have seen, however, this is far too permissive. Suzy can’t kill baby Suzy.
Suzy is much like Tim, although her situation is simpler. Tim’s scenario may be rectified by miracles other than resurrection; for instance, Tim’s parents might appear spontaneously, despite the death of Tim’s grandfather, ready to give birth to Tim in 1940. Lewis himself notes similar possibilities as the sorts of stories that must be told in order to render Tim’s successful grand-parricide consistent. So there are more miracles that might render Tim’s scenario consistent than there are that would render Suzy’s consistent. However, the fact that Tim’s success might be set right by a wider array of miraculous events should be of little comfort to those who think (as I do) that we cannot do anything which necessitates a miracle. Tim can succeed at killing his grandfather only if the laws were different; the laws could differ in a variety of ways and allow Tim to succeed, but they cannot be as they actually are. Tim, therefore, cannot commit grand-parricide.
Before I turn to objections, I must insure those who have followed me do not stumble over the precipice. Some readers may see what I have said, and interpret this as an argument for the false conclusion that time travelers lack the ability to do anything other than what they actually do during their temporal romp. This argument does not prove that time travelers are so inept. Suzy cannot kill her younger self, but she could have picked up a different pillow before making her attempt, or taken a different path to the hospital. Suppose Suzy had two pillows before her in the hospital, one red and one green. She chose the red, but she could have chosen the green. “This would require a miracle!” We have no good reason for thinking it would, at least, no reason as good as the one we have for thinking that Suzy successfully killing baby Suzy would require a miracle. If Suzy kills her baby self, we are left with many questions: “How is Suzy there in the first place? Why didn’t she just disappear after her younger self died?” No questions of comparable difficulty arise on the supposition that Suzy picked up the red, rather than the green pillow; certainly no questions arise whose answer necessitates miraculous intervention. So I affirm that Suzy, and Tim, suffer from a limited ineptitude. They lack some abilities, and not others. Suzy cannot kill her younger self, but she was able to pick a different pillow.
There are many possible objections to this argument. Much of what I have said follows Kadri Vihvelin; she defends the same conclusion I do, and I draw largely on her respondents to pre-empt objections.1
My first objector: “You have neglected a story, one perfectly consistent with the laws of nature, on which Suzy succeeds. There is a world where the baby in the nursery, the one which Suzy kills, is baby Sally. The nurses made a mistake, mislabeled the cribs. This world’s laws do not differ from the actual world, so Suzy can kill the baby in a world that shares our laws.” This response, and replies like it, engender confusion. My claim is simple: to be able to perform an action, your success must not necessitate a miracle. This is a necessary condition on abilities. To be able to sink a three pointer, my success must not require divine intervention. If I’m Seattle’s best shooter only when all my prayers are answered, this gives little comfort to my actual pickup team. The scenario described, where Suzy kills baby Sally, is not one where Suzy succeeds. It is a scenario where Suzy kills Sally, not one where Suzy kills Suzy.
There are similar, but more sophisticated, objections that suggest scenarios which invoke the concept of ‘person stages’ and the doctrine of Origin Essentialism. Consider the following theory: today, you are one way, 10 seconds ago, you were another. At each instant, there was a person stage, and those stages were each a part of you. You, anyone, is the collection of these person stages, throughout their history. This is, very roughly, the idea invoked by talk of person stages. When we wonder ‘can Suzy kill baby Suzy?’, so the objection goes, we are wondering whether the Suzy person stage can kill the baby Suzy person stage. Origin Essentialism is a view which entails that people have their person stage parts essentially, so Suzy could never be part of a different person than baby Suzy, the two person stages are necessarily connected. But what if Origin Essentialism is false? Then there is a world where Suzy and baby Suzy are person stages of different people. In such a world, Suzy (the person stage) might kill baby Suzy (the person stage) with no resurrection. These sophisticated replies get off on the wrong foot, because they fail to describe a scenario where Suzy succeeds in killing her younger self. In the possible world where Suzy (the person stage) kills baby Suzy (the person stage) we aren’t looking at a world where Suzy kills her younger self. We are looking at a world where one person kills a different person. To reply to my argument, one must either deny that consistency with the actual laws of nature is necessary for ability, or show that Suzy’s success necessitates no miracles. This objection does neither, and fails.
My second objector: “Your reply to the first objector is too quick. You deny that this problem should be put in terms of person stages, the talk of ‘baby Suzy’ and ‘adult Suzy’ should thus be able to be articulated as talk of ‘Suzy’ simpliciter. This leaves us with the question of whether, in the nursery, ‘Suzy can kill Suzy.’ Obviously she can, by jumping out the window. Thus, your attempted escape of the more complicated articulation of the problem leaves your suggested answer untenable.” Here I am largely adapting an argument from Vranas, 2009. How do I reply? Arm Suzy with a firearm (after she has used her time machine and entered the nursery). Point the gun at Suzy, who is in the crib in front of Suzy. Wrap Suzy in chains (after she has used her time machine and entered the nursery) such that she cannot move. Suzy cannot jump out any windows. Can Suzy kill Suzy? Clearly not, for the reasons we’ve already explored. If Suzy killed Suzy in the only way available to her, she would need to be resurrected, and Suzy can’t do anything that necessitates a miracle. “She could refuse to eat!” Fine. A nurse is there, who wants to ensure Suzy succeeds. She hooks Suzy (after Suzy has used her time machine and entered the nursery) up to a powerful IV drip which sustains all her bodily needs. Can Suzy kill Suzy? With enough ingenuity, it is clearly possible to engineer a scenario where Suzy will live a long and healthy life unless she kills (younger) Suzy by firing her gun. It seems that she can. (Baby) Suzy is right there, at the end of the barrel, defenseless in the crib. But she cannot, as we have just seen. The gun will jam, or explode, or miss, forever and always, because no one can bring about a miracle. The puzzle is restored, with no talk of person stages.
My third objector: “No compatibilist can accept your argument. Deterministic worlds are worlds that are alike, in terms of concrete matters of fact as well as natural law, at all times or none. Thus, if someone were to refrain from performing an action that they actually did, the world in which they refrain must differ from the actual one everywhere in terms of concrete matters of fact, or natural law. The compatibilist must maintain that we still have the abilities involved in free will, to deliberate, choose, and decide. They must think we have these abilities even at deterministic worlds, where the counterfactual exercise of these abilities necessitates a world which differs everywhere. So the compatibilist cannot accept that nomological possibility is necessary for the possession of ability.” I am a compatibilist of the commonsensical sort, and so I accept the objector’s presumption of compatibilism. My suggested constraint is merely a strengthened version of a constraint placed on ability by the classical compatibilist par excellance, with whom my objector is likely to agree. In Are We Free to Break the Laws? David Lewis accepts that no one has the ability to break the laws in the following sense: no one can perform an action which is itself, or causes, a law breaking event. I agree. My principle entails Lewis’. If it is necessary to possess the ability to perform an action that the action in question is can occur at worlds with our laws, then no one is able to perform law breaking actions, or actions which cause law breaking events (because worlds with law breaking events do not share our laws). My principle is stronger only in that it also rules out actions which, although they themselves do not break the laws, nor do they cause the laws to break, can only be performed if the laws were broken. My objector interjects:
“But your principle contradicts Lewis’ weak thesis, that we are able do things such that, if we did them, a law would be broken.” Not so. I claim that actions we are able to perform cannot necessitate miracles, this does not commit me to denying the weak thesis. An action can require no miracle and be such that, if someone were to perform the action, a miracle would occur. Take the standard Lewisian counterfactual analysis, where a counterfactual is true if the consequent is true in all the worlds closest to the actual world where the antecedent is true. There might be some world, very far away from actuality that still shares our laws, where Suzy chooses the red pillow. All the same, in all of the words closest to actuality, those relevant for evaluating the counterfactual, a miracle might occur to bring about Suzy taking the green pillow. Suzy taking the green pillow requires no miracle, yet it is still true that if Suzy had taken the green pillow, a miracle would have occurred. So even a Lewisian ‘local miracles compatibilist’ can, and should, accept my argument without revision. That said, there may seem to be a tension here. I agree there is a tension. This tension, alongside other arguments concerning counterfactual miracles, are why I deny Lewis’ weak thesis. However, my denial of the weak thesis does not come at the expense of my compatibilism. I cannot develop arguments for either of these claims here, but rest assured I have future articles which will address the tension. As a quick teaser, the apparent tension is easily remedied if the nomological possibility of the antecedent and consequent of a candidate counterfactual is necessary for the truth of the relevant counterfactual, and this is true if we hold the laws of nature counterfactually fixed (a practice widely maligned, especially by Lewis).
“You have not sufficiently demonstrated the need to give a definitive answer on the question at issue. Why is it not enough to say that Tim can’t kill grandfather, and that he can on different delineations of the facts?!” I’ve already addressed this throughout the article but it’s worth going over again. First, agreement: Modal terms are shifty in that, in different contexts, we might (correctly) think some statements are true when in other contexts we might think (correctly) that those same statements are false. Consider: I can’t speak Finnish. My duolingo streak never got long enough. But also, I can speak Finnish. I have a larynx, vocal cords, a functional tongue. I can speak Finnish in the same sense an iguana can’t. But this does not mean that the correct answer to the question “is X possible?” is “yes and no!” Suppose I promised my friend I would attend his birthday party but, through no fault of my own, found myself far away from his home 30 minutes before the party was supposed to start. I am 20 minutes away by bike and 40 minutes away if I walk, but unfortunately I do not know how to ride a bike. I trudge, sullen, up to my friends front door after my walk, and knock. My friend answers, furious, and asks for an explanation for my tardiness. Suppose he asks, in specific, “could you have made it on time?” What should I reply? Clearly, I claim, I should not say that I could have made it, nor should I say that I could and couldn’t. I should say something like: “I couldn’t have made it on time! I had another commitment that ran long and was 40 minutes across town.” My friend could say many things to justify his anger at this point, but one thing he shouldn’t say is “you could have biked” because I will point out I don’t know how. But suppose he insisted “But your legs are working, your proprioception isn’t shot, you have decent hand eye coordination. Clearly you could have biked here to be on time for my party, which is why I’m mad at you!” My friend is being ridiculous. Of course there is some delineation of the facts on which I could have made it to his house (e.g. by not holding fixed my lack of knowledge) but that delineation is just never appropriate for the context at issue. It would be just as strange to insist in similar situations e.g. while describing my capabilities to a coach for the basketball team, or when deliberating about whether I should go to finland. My claim is that saying Tim can kill his grandfather is like saying I could have made it to my friends birthday party. It requires just as much (if not more) of a stretch from our ordinary frame of modal evaluation as it takes to say that I could have made it to my friends party because I could have biked there. So Tim can’t kill grandfather, and Suzy can’t kill her younger self, given what we ordinarily mean by ‘can’.
What have we learned? In some sense, nothing we didn’t already know. You knew before reading this article what would happen if Tim, or Sally, tried to commit their respective decade-hopping-homicide. They would miss, or slip, or wince, or (…). One important lesson is that the laws of nature constrain. Although my conclusion in this article is compatible with the popular view that we do not hold fixed the laws of nature when evaluating counterfactual claims, the argument of this article is certainly far neater if you grant that we do hold the laws fixed. In so far as the pattern of reasoning in this article strikes the reader as plausible it should give them some reason for thinking that the laws of nature are one of the principle means by which we adjudicate questions of ability. A second important lesson is that the shiftiness of modal contexts does not, on its own, give us any reason to abandon insisting on an answer to modal questions. I’m asking right now whether so-and-so is possible, and pointing out that my asking in a different context might elicit a different response gives me no reason not to insist on receiving an answer now. Finally, it’s worth noting that the laws of nature might be different in ways that deprive us of certain abilities in certain circumstances, but not always. Determinism, on my view, does deprive us of some abilities (I’ll expand on in the future), but none of our abilities we think we have in ordinary circumstances. Sometimes we can make the mistake of thinking that the question is all or nothing: either a change in the laws threatens all of our claims about abilities or none, but the time travel case shows this is false.
So in addition to what we have learned about the abilities of the characters in the next Christopher Nolan original project, we get some interesting insights of philosophical significance.