If that's what you think justifies punishment, it doesn't seem to be necessary that the criminal acted freely when he committed his crime, for us to be justified in punishing him. There could be a case where, even though the criminal was coerced or hypnotized when he committed the crime, still, punishing him would bring about good consequences of the sort just mentioned. It might serve as a good deterrent to other crooks; it might keep him from committing crimes later in his life; and so on. (In fact, it's not clear that the person even needs to be guilty of the crime he's being punished for. In some cases, punishing him might still have those good consequences, even though he is innocent.)
If this is the right story about what justifies punishment, then it does seem to matter that you performed the act you're being punished for, and it seems to matter that you performed that act of your own free will. It's not fair to punish you for something that happened by accident, or if you were forced to do it, or if it was unavoidable for some reason. For example, if I push you out a window and as a result you land on your roommate's pet cat, it doesn't seem fair to punish you for crushing and killing the cat. (Even if you're happy that the cat is dead, it still seems like it would be unfair to punish you.)
So on these accounts, if you're to be morally responsible for doing X, one necessary condition for this is that you did X freely. (Perhaps there are other necessary conditions, too: for example, you may also need to appreciate what the morally relevant effects of your actions will be.)
Now, it is possible to enter into a contract where you consent to be held responsible for certain outcomes, even when you don't freely bring those outcomes about. For instance, I may permit you to look at my butterfly collection only if you agree to be held responsible for any damages you cause. Suppose you accept these terms. Then, while you're looking at my butterfly collection, you involuntarily sneeze, and damage all my butterflies. I can fairly hold you responsible for the damages you caused here, even though you did not freely bring those damages about.
But even in cases like the one I just described, there is still some connection between questions about freedom and questions about responsibility. You're responsible for the damages to my butterfly collection because you entered into the contract with me of your own free will. If you had been forced to enter into that contract, maybe it would no longer be fair for me to hold you responsible for the damages.
So according to accounts of sort II and IV, there is some connection between questions about moral responsibility, and the legitimacy of blaming or punishing someone, on the one hand, and questions about a certain kind of freedom, on the other hand. That is the kind of freedom we will be interested in, in this course. We will be asking: What does it take to have this kind of freedom? Do people have the kind of control over their actions which is necessary for them to be morally responsible, and for us to be justified in punishing and rewarding them? If all our actions are result of our genes and our upbringing, or if they're all the necessary unfolding of how the world was long before we were born, does that show that we lack that kind of freedom or control?
First let's get clear on the distinction between bringing a certain outcome about on purpose, and merely bringing that outcome about.
Suppose your baby is in a pit that contains a savage tiger. The pit also contains another baby, but the tiger hasn't noticed this second baby yet. It's impossible for you to save both babies. However, you know that the help is on its way, and the tiger will only have time to eat one baby, before help arrives.
Now, if you somehow manage to pull your own baby out of the pit, you will thereby bring it about that the tiger eats the second baby. And you know that you will bring that about. But you wouldn't be bringing it about on purpose. You wish you could save both babies, you just know that you can't. And if, after you pulled your baby out of the pit, the tiger somehow failed to notice the second baby, or the second baby escaped somehow, that wouldn't thwart any plan of yours. Contrast a case where, rather than pulling your baby out of the pit, you instead attract the tiger's attention over to where the second baby is. There you are trying to get the tiger to eat the second baby, in order that it will be too busy to eat your baby. In that case you are intentionally trying to bring it about that the tiger eats the second baby. That seems to be importantly different from the case where you just pull your baby out of the pit, knowing that this will have the unfortunate effect that the tiger will eat the second baby.
The cases I just described were meant to illustrate the difference between bringing some outcome about knowingly but unintentionally, and bringing that outcome about intentionally, or "on purpose." They weren't supposed to provide an example of someone being morally blameworthy for an outcome which he knowingly but unintentionally brought about. In the tiger case, if you pull your baby out of the pit, and thereby knowingly but unintentionally bring it about that the tiger eats the second baby, there does not seem to be anything morally awful about what you've done.
However, in some cases, you can be morally blameworthy for things that you knowingly but unintentionally bring about.
Suppose we're at war, and you're in charge of where our planes drop their bombs. You decide to bomb one of the enemy's ammunition factories. This factory stands right next to a big children's hospital, hence many children will be killed in the explosion.
You regret this, but you say, "Hey, I'm not destroying the children's hospital on purpose. It's just that I know that that will be an effect of bombing the ammunition plant. The death of all those children is merely something that I knowingly bring about, not something I'm bringing about intentionally." Clearly this would not excuse you from the deaths of all those children. In this case, you are morally blameworthy for their deaths, despite the fact that you did not bring their deaths about intentionally or on purpose.
My point here is that you can be morally responsible for the effects you know your actions will have, even when you don't bring those effects about on purpose. The tiger case shows that you're not automatically responsible for all the effects you know your actions will have. Whether you are responsible will depend on the details of the case. The bombing case is one example where you do seem to be morally responsible for the foreseen but unintended consequences of your actions.
So in summary:
Surely we think this is legitimate. We do think it's OK to allow luck to play some role here. We may not feel the same way about lots of luck: If a whirlwind bursts onto the field, pins your team to the wall, and pulls the ball out of your opponents' hands and into the goal, you probably would protest against allowing this goal to count. But it does seem to be OK if some kinds of luck play a role in determining who wins the game.
Now what about moral praise and blame? Is it also legitimate to allow luck to play a role in how these get dished out? We'll understand the thesis that There is moral luck as claiming that your moral status can depend in part on facts outside your control, or on facts you don't have complete control over. Is this right? Can our moral status depend on facts outside our control?
There is a Kantian intuition which denies that our moral status can depend on any facts outside our control. Consider two criminals who attempt to murder their respective enemies. One of the criminals succeeds and the other fails. We can suppose that whether their murder attempts succeeded depended on facts outside the criminals' control. If so, then according to the Kantian intuition, this attempted murderer and this successful murderer would be just as bad, morally, and they would deserve the same kinds of blame.
Our ordinary moral assessments, on the other hand, do seem to allow for some kinds of moral luck. The extent to which we praise and blame people does depend in part on circumstances beyond their control. For example, we would punish the successful murderer more severely than the attempted murderer.
Or consider a truck driver who neglects to get his brakes checked, and a result he loses control of his truck and it goes up on the sidewalk while he's turning a certain corner. If the truck driver is lucky, then no one will be walking on the sidewalk, and so no one will get hurt. If the truck driver is unlucky, then someone will be walking on the sidewalk, and will get injured by his truck. As a matter of fact, we would blame and punish the unlucky truck driver much more than we would the lucky truck driver, who was just as negligent.
Here's another sort of case: in action movies, you'll often see the hero running after the criminal, firing shots at him, despite the fact that there are lots of innocent bystanders running around. This is extremely risky. Those shots might hit an innocent bystander. If the hero is lucky, as he usually is, no innocent people will get killed by his stray bullets; and he'll be praised as a hero. If the hero is unlucky, on the other hand, and he ends up killing an innocent bystander, then we'll criticize him and blame him for his risky behavior. And he will also blame himself. So it seems like whether he turns out to deserve praise or blame will depend on facts outside his control.
Here's yet another sort of case: consider the ordinary German citizens during the Nazi regime. They had the opportunity to behave heroically and oppose the regime (at great risk to their own safety), or to keep quiet and do nothing. Most of them did the latter. We, and contemporary Germans, regard them as morally blameworthy for this. But we shouldn't be too proud of ourselves, for we haven't faced the same test that they faced. No doubt many of us would also have failed to be heroes. We're just "lucky" that we weren't placed in the same circumstances as they were. This illustrates how the kinds of blame or praise one turns out to deserve can depend on the kinds of circumstances one faces. And that is another example of a fact which is largely outside one's control.
So there are a lot of examples illustrating that the ways we naturally dish out moral praise and blame are sensitive in part to facts which are outside the agent's control.
Nagel thinks we have both intuitions: both the intuition that there is moral luck; and also the nagging Kantian feeling that there ought not to be any moral luck, that we really ought to hold people responsible only for the things which are fully under the control. These intuitions conflict with each other.
What should we do about the conflict? Should we go along with the Kantian intuition, and eliminate all elements of luck from our practices of praise and blame--in the way we discussed eliminating all elements of luck from sports games?
In his article, Nagel argues that this is not a satisfactory response. (See pp. 180 and the middle of p. 183. On pp. 183-184, he argues that simply accepting moral luck is also an unsatisfactory response. So Nagel thinks in the end we have a paradox. But let's focus on the one part of his argument, where he says that we can't simply accept the Kantian intuition.) When you think about it, Nagel argues, it turns out that almost all, if not all, the things we do are dependent in part on facts outside our control. For example, the bomber in the case we described before can say he's no more responsible for the deaths of all the children than anybody else is. It was just his "bad luck" that he found himself placed in the middle of a war (he didn't have any control over that), and just his "bad luck" that there was a hospital next to his bomb target (he didn't have any control over that), and so on. And the murderer can say that he didn't have any control over the fact that he hated his enemy. All he did, anyway, was pull the trigger on a gun. It was just his "good luck" that the gun fired properly and that his enemy got struck by the bullet. The murderer didn't have full control over those facts, either. So if we go along with the Kantian intuition, it will turn out that we're morally responsible for very little if anything. Or alternatively, that we're all just as morally blameworthy as the bomber and the murderer. Both of these seem to be unacceptable results.
So it looks like we have to find a way to live with some element of moral luck, if we're to make any discriminations at all with respect to how morally blameworthy people are. We may not want to accept all kinds of moral luck. For instance, you may want to insist that the attempted murderer and the successful murderer deserve the same kinds of blame and punishment. But your argument for this can't just be: whether the bad result comes about is outside these agents' control. For that's not enough, by itself, to show that the end-result of their actions is morally irrelevant. We have to allow for some kinds of moral luck, if we're to make any moral discriminations at all. Unless you think that nobody is morally responsible for anything (or you think that everybody is equally bad), the fact that your actions and their outcomes depend to some extent on facts outside your control can't be enough, by itself, to absolve you from moral responsibility.