20  Contextualism and Ability

We used the distinction between changing the future and affecting the future in order to defuse the fatalist threat. From the assumption that the future is fixed, the fatalist inferred that our actions are causally inert. But we pointed out that even if the future is fixed, our actions may still play a genuine causal role in how it unfolds.

But the discussion left a residual question on the table. Suppose it is now true that I will not write a book over the summer. I may well affect the future in that my actions will play some causal role in how the future develops. But the question remains: can I write a book over the summer? The fatalist argument failed, but one may now wonder whether I have genuine abilities that extend beyond what I will actually do? This question matters because the debate over free will is often framed in these terms.

20.1 Ability and Free Will

(Inwagen 1975), whose argument we will examine in detail, makes the connection explicit:

It seems to be generally agreed that the concept of free will should be understood in terms of the power or ability of agents to act otherwise than they in fact do. To deny that men have free will is to assert that what a man does do and what he can do coincide. And almost all philosophers agree that a necessary condition for holding an agent responsible for an act is believing that that agent could have refrained from performing that act.

If you have free will, then faced with a simple choice such as, for example, to raise your hand or to put it back on the table, you can raise your hand and you can put it back on the table even if it is now true that you will do one and not the other. That is, given the choice, you raise your hand freely just in case:

  1. You raise your hand.
  2. You can raise your hand.
  3. You can put your hand back ont he table.

But what exactly is for you to have the ability to \(\varphi\) or to not- \(\varphi\) given that you will do one but not the other? To clarify this, it will be helpful to look at David Lewis’ treatment of the time travel scenario, since it delivers a remarkably helpful account of ability.

20.1.1 Tim and Grandfather, Revisited

Compare again with the scenario (Lewis 1976) describes. Consider a time at which conditions are optimal and Tim faces the decision to kill or not kill Grandfather. Since we know that Grandfather died in bed in 1957, we know this:

  • Tim did not kill Grandfather.

  • Tim can fail to kill Grandfather.

But on the face of it, for Tim’s action to count as free, we should be able to attribute to Tim the ability to kill Grandfather:

  • Tim can kill Grandfather.

This seems doubtful, however, given that it is part of the historical record that Grandfather survived. To kill Grandfather would be to change the past, which we have established to be impossible. So it seems that the last claim is false. But then Tim’s situation looks very much like someone with no free will at all. There is no difference between what he can do and what he does.

This is why the time travel scenario is of help here. It provides a vivid, controlled version of the very pressure that fatalism and determinism apply to ordinary agents. If we manage to explain what it means for Tim to can kill Grandfather despite the historical record, then we will have made progress on the general question.

20.2 Contextualism

When (Lewis 1976) describes the case, he sets the reader up for two different verdicts.

  • By ordinary standards, Tim can kill Grandfather.

    Tim has all the necessary skills, the opportunity, the best rifle money can buy, perfect weather conditions, etc.

  • By stricter standards, Tim cannot kill Grandfather.

    It is part of the historical record that Grandfather died in bed in 1957, and therefore it is part of the historical record that Tim did not kill Grandfather. To kill Grandfather would require Tim to change the past, which is impossible.

So, we have reason to accept to apparently contradictory statements:

  1. Tim can kill Grandfather.

  2. Tim cannot kill Grandfather.

(Lewis 1976) responds that there is no context in which we are tempted by both judgments, and the reason for this is that ability attributions are context-sensitive. In this respect at least, ‘can’ behaves much like other contextually sensitive expressions. Compare with the behavior of quantifier expressions such as ‘everything’:

  • By ordinary standards, I packed everything for my flight.

    I packed all my clothes, my computer, and every other item I intended to bring with me for my flight.

  • By stricter standards, I did not pack everything for my flight.

    I did not pack my books, my furniture, my bicycle etc.

That means that there are contexts in which we would accept one but not the other statement:

  1. I packed everything for my flight.

  2. I did not pack everything for my flight.

When we evaluate statements like 3 or 4, we should be mindful of the context \(C\) in which they are made, since the context will generally determine the appropriate domain of quantification. In a context in which we review my travel plans, the domain of quantification will be restricted to the items I intend to bring with me in my travels, but other contexts may determine a broader domain of quantification.

There is a link between quantification and modality, since you may take modal statements to implicitly quantify over possible worlds:

  1. David can speak Finnish.

    It is possible that David speaks Finnish.

    There are possible worlds in which David speaks Finish.

  2. David cannot speak Finnish.

    It is not possible that David speaks Finnish.

    There are no possible worlds in which David speaks Finnish.

The relevant domain of worlds is made salient by the context in which the statements are made. In a context \(C_1\) in which we compare David’s abilities to those of other animals, we may hold fixed certain physiological and anatomical facts and facts to do with the human larynx and nervous system. Because there are worlds in which those facts obtain and David does indeed speak Finnish, we are inclined to judge that he can speak Finnish.

On the other hand, in a context \(C_2\) in which we compare David’s abilities to those of a native Finnish speaker, we may hold fixed the fact that David has no experience whatever with Finnish and the fact that he has no facility for new languages. Because there are no worlds in which those facts obtain and David speaks Finnish, we are inclined to judge that David cannot speak Finnish.

The general principle (Lewis 1976) deploys is this:

Contextualism and Ability

\(S\) can \(\varphi\) is true at a context \(C\) if, and only if, the fact that \(S\) performs \(\varphi\) is compossible with the set of facts that we hold fixed at \(C\).

There is similarly no contradiction involved in our judgments that Tim can and cannot kill Grandfather, since they are made in different contexts relative to different sets of contextually salient facts.

  1. Tim can kill Grandfather.

    It is possible that Tim kills Grandfather.

    There are possible worlds in which Tim kills Grandfather.

This is true in a context \(C_1\) in which we hold fixed the fact that Tim has all the necessary skills, the opportunity, the best rifle money can buy, perfect weather conditions, etc. Crucially, we do not hold fixed the fact that Tim did not kill Grandfather.

  1. Tim cannot kill Grandfather.

    It is not possible that Tim kills Grandfather.

    There are no possible worlds in which Tim kills Grandfather.

This is true in a context \(C_2\) in which we hold fixed the fact that Tim did not kill Grandfather. No worlds in which Tim killed Grandfather are compossible with worlds in which Tim did not kill Grandfather.

This provides another line of response to the fatalist argument. For we are now in a position to explain our ambivalence toward the statements below when we hold fixed certain future facts.

  1. I can write a book over the summer.

    It is possible that I write a book over the summer.

    There are possible worlds in which I write a book over the summer.

This is true in a context \(C_1\) in which we hold fixed the fact that I have all the necessary skills, the time and opportunity to write a book, etc. Crucially, we do not hold fixed the fact that I will not write a book over the summer.

  1. I cannot write a book over the summer.

    It is not possible that I write a book over the summer.

    There are possible worlds in which I write a book over the summer.

This is true in a context \(C_2\) in which we hold fixed the fact that I will not write a book over the summer. No worlds in which I write a book over the summer are compossible with worlds in which I will not write a book over the summer.

20.3 Vihvelin’s Challenge: The Counterfactual Principle

(Vihvelin 1996) raises a concern for the contextual account of ability. She continues to focus on a time travel scenario:

Suppose that [a] time traveler, unhappy with her life, decides that it would have been better never to have lived it. She packs a gun and travels back through time, determined to kill her infant self. She picks a time when she knows that the baby will be alone. She checks carefully to make sure her gun is loaded. She fires. She misses. She fires again. This time the gun jams.

We know that the time traveler’s efforts will all fail. Call her Suzy, and her infant self Baby Suzy. Since Suzy’s survival is what makes the time journey possible in the first place, Baby Suzy must live. Suzy cannot kill Baby Suzy, since this would be a self-defeating act. Yet by ordinary standards, Suzy has the ability to shoot close-range targets and plenty of opportunity. So it seems both that she can and that she cannot kill Baby Suzy.

In premise-conclusion form, this appears to be a variant argument against the possibility of backward time travel:

  1. If If backward time travel were possible, then it would be possible for Suzy to kill Baby Suzy.
  2. It is impossible to for Suzy to kill Baby Suzy.
  3. Backward time travel is not possible.

We know what Lewis’s contextualist reply to this argument is supposed to be. There are contexts in which we judge that it is possible for Suzy to kill Baby Suzy, namely, contexts in which we hold fixed Suzy’s necessary skills and the fact that she had opportunity to pull the trigger. The fact that Suzy kills Baby Suzy is indeed compossible with the contextually salient facts. So, if backward time travel were possible, then it would be possible for Suzy to kill Baby Suzy by ordinary standards. But by ordinary standards, 2 would be false.

On the other hand, there are contexts in which we judge that it is impossible for Suzy to kill Baby Suzy because they are contexts in which we hold fixed the fact that Baby Suzy will eventually grow to become Suzy and travel back in time. No worlds in which Suzy kills Baby Suzy are worlds in which Baby Suzy eventually grows into Suzy. So, it is not true that if backward time travel were possible, then it would be possible for Suzy to kill Baby Suzy all things considered. On the other hand, 2 is true all things considered. At the end of the day, there is no uniform interpretation of ‘can’ on which premise 1 and 2 both come out true. One interpretation makes 1 true and 2 false, whereas the other makes 1 false and 2 true.

(Vihvelin 1996) rejects this response. Her objection is not to Lewis’ characterization of the relevant contexts, it is rather that she rejects his account of ability even by ordinary standards. She relies on a perfectly general principle:

We should agree that someone can do something, in the relevant sense, only if it’s true that if she tried to do it, she would or at least might succeed. And everyone should agree that if someone would fail to do something, no matter how hard or how many times she tried, then she cannot do it.

The objection takes the form of a completely general principle:

Counterfactual Principle

Necessarily, if someone would fail to \(\varphi\) no matter how hard or how many times she tried, then she cannot \(\varphi\).

The motivation is clear. I cannot walk on water because I would never walk on water no matter how hard or how many times I tried. This appears to underlie much of our practical deliberation. Whether something is an option for you depends on whether you think you would do (or at least have a reasonable chance of doing it) if you tried.

Now apply this to Suzy:

  1. If Suzy were to try to kill Baby Suzy, she would fail (no matter how hard or how many times she tried.)

So, given the counterfactual principle, it follows:

  1. Suzy cannot kill Baby Suzy.

But the counterfactual principle applies to interpretations of ‘can’ relative to ordinary contexts, which means that by Vihvelin’s lights, premise 2 is true even by ordinary standards.

(Vihvelin 1996) rejects the argument from premises 1 and 2 to the impossibility of backward time travel, but her reason is importantly different from Lewis, namely, she takes premise 1 to be false by ordinary standards and premise 2 to be true by ordinary standards of ‘can’.