Saturday, February 1, 2014

Conscience and intending the impossible

One of the toughest problems is what to do about cases of mistaken conscience. Let's say Samantha has a justified false belief that it is right to kill one innocent that ten might live (whether anyone can be really justified in thinking this is a question to bracket), and acts on it. Then it seems: Samantha did wrong and she did right. She did wrong in killing the innocent but she did right in following her conscience. But that shouldn't be the whole story. For suppose that Samantha had done the opposite—refrained from killing the innocent. Then she would have done wrong in disobeying conscience and right in refraining from killing the innocent. So whatever she does, she does wrong and she does right. And yet the two cases are not analogous. For when she kills, then she is inculpable of the murder by reason of her justified false belief. And when she refrains, she is culpable for violating her conscience.

We could leave it at this. But that would leave unexplained why it is that the duties of conscience are what culpability aligns with.

For years I've been trying to explore a story here, and I am never quite happy with it, and I am still not happy with it, but let me give it one more try. Intention, permissibility and impermissibility applies to action types, and not just to action tokens. And while there are no impossible action tokens, there are impossible action types that can be the objects of one's intentions. Many people, some sane and some not, have intended to trisect an angle (with ruler and straightedge). In so doing, their intentions had an object, the action type trisecting an angle. Moreover, their intended action type was permissible, albeit also impossible. We might say that per impossibile had they succeeded, they would have done something permissible and worthwhile.

Now go back to Samantha. Samantha intends a consequentially justified killing of an innocent. This action type, just like the trisecting of an angle, is impossible. It is impossible for consequences to justify the killing of an innocent. But if per impossibile she succeeded, she would have done something permissible and worthwhile. Plausibly, Samantha's intended action type while impossible is permissible, just as trisecting an angle is.

So on this story Samantha intended to do something permissible, but failed. She ended up doing something other than she intended. Samantha's action was an attempt at a consequentially justified killing of an innocent, and at least if she was justified in thinking that the attempt would succeed, she did right to make the attempt. And had she refrained from the attempt, she would have done wrong. Compare the case of someone who is ignorant of the impossibility of trisecting an angle and is told that an innocent will die if he does not trisect an angle. He acts well by trying to trisect and would be we doing wrong by refraining from trying.

On this story, if she kills, Samantha doesn't do both right and wrong. She simply does the right thing. But this right thing is an attempt at the impossible, and hence fails. And its failure, tragically, results in the death of an innocent (though if indeed the ten are saved, there is a silver lining, though not a justification). And if she believes that the killing would be consequentially justified, then in refraining from trying to kill, she simply does wrong.

But don't we want to say that Samantha unjustifiedly killed an innocent, and that's wrong? We need to be cautious here. The experienced surgeon who does her very best but who nonetheless kills the innocent patient does not do wrong. Her performance of the surgery is, indeed, a killing. And it's not a justified killing. But the surgeon's action is not intentional under the description killing the patient, and to say that the surgeon did wrong or right in killing the patient jars in the same way that it jars to ask whether my stumbling over a bump while walking to work was right or wrong. The stumbling was a part of my attempt to get to work, and hence was a part of a right action. But it was an accidental part as far as my intentions go. The same goes for the surgeon. It is harder to say this in Samantha's case, but perhaps not impossible. She did not intend a killing simpliciter, but a justified one. I would be inclined to say that both the surgeon's and Samantha's killings are non-justified, rather than unjustified.

The case of Samantha is particularly striking because it is impossible for a killing of an innocent to be consequentially justified. But one can also have similar cases where what is intended is possible. Suppose I reveal a secret that I promised to keep silent because I justifiedly but falsely believe that I ought to. Then I intend to break confidence as I ought. I fail—my breaking confidence is not justified. But what I intend is in fact a possible action type—there are times when one ought to break confidence. I do the right thing simpliciter: I attempt to do what I ought.

In the cases of Samantha and of confidence breaking, mistaken conscience enters into the story by making it possible for the agent to intend what otherwise the agent could not intend. Thus mistaken conscience functions much as the attempted trisector's false belief that one can trisect an angle.

There are probably some really serious problems with the above as a general proposal of what happens in conflicts of conscience. Here is one that particularly bothers me. I had breakfast today. Suppose, however, that I had promised someone not to have breakfast today (say, in order to experience solidarity with the less fortunate) but I completely innocently forgot the promise (imagine someone slipped me a forgetfulness pill). The analogue to the Samantha story would be that I intended to have a breakfast that I did not promise not to have. But of course I am exceedingly unlikely to have intended this while eating breakfast (wouldn't thinking about promises have brought my promise to mind?). Do you ever have such intentions when eating breakfast? (I suppose if one was in a habit of making promises to skip breakfast, one might. But one shouldn't make a habit of skipping breakfast—it's not healthy.)

Perhaps, though, whenever we do anything, we should be intending to do it rightly, or to glorify God through it, or the like. And if I tried to, say, rightly eat breakfast, while bound by promise not to eat, my action was a failure. So we do have the same pattern as in the Samantha story.

But is it really the case that whenever we act we need to have some such intention? Personally, I find this a plausible proposal. After all, we are to love God with our whole heart, soul, might and mind, and St Paul tells us to pray always and to take captive every thought (noema) for Christ. Thus every action of ours should be at least implicitly directed (perhaps in a way that even an atheist can) at the glory of God. When we fail to have it so directed, we do wrong.

This sounds right, but I don't know that it solves the breakfast problem. For suppose that I eat breakfast with no such intention, and eat contrary to my innocently forgotten promise. Then indeed I do wrong by not having the right God-glorifying intention in eating breakfast. But my innocent ignorance of my promise is still relevant. I am culpable for not intending to glorify God, but I am not culpable for breaking my promise, it seems. So something has yet to be explained.

But that we can handle a number of cases using the above method suggests that we may be able to do even more if we put our minds to it. Maybe there is something special about the promise case, for instance.


Unknown said...

One thing I find illuminating is your note that Samantha's action falls under two descriptions, one type and one token. By characterizing her action according to a type, one which is permissible though impossible, you can make intuitive the claim that what she does is morally right. That is, you can give a satisfactory error theory for our pretheoretical intuition that what she does is wrong (if consequentialism is false). This is much needed, in my view, if we are to make theoretically attractive a view on which erring conscience binds for the people who would normally shy away from such a view.

n cases when an agent acts to avoid violating a mistaken conscience, like Samantha does, she acts for a good end in one sense, but not in another. In the first sense, she acts for a good end because the end she has in mind is something like “saving the innocent,” or more broadly, “doing what is right” or “doing what I ought” or “obeying God” etc., and these are good ends. Normally, they would justify an action. In the second sense, though, the end she intends is a bad one. The end which is actually doing double duty as a means to the first end, when taken out of that context, is just "killing the innocent,” and this is not a a good end. So, what we need to say is that the second sense gets something wrong. It extracts the end from its context as a means to some further end— the good end that would justify her action— and so misdescribes the action. In your terms, we have put the token into the wrong type.

Recently, I have been arguing a moral reason depends for its “form,” or for “what it is a reason to do," partly on the mental states of the agent whose reason it is. And you seem to be saying that an action depends for its form on the mental states of the agent whose action it is. These may be two sides of the same coin, but here is where things get tricky. What I want to say is that one cannot have a reason to do what is impossible. Or, if there is no way for S to satisfy a reason R (e.g. “a reason to save ten people”), then S does not have a reason R to save ten people at all. The content of S’s reason— what S has reason to do— depends on what is available to her, given her mental states. Now even if Samantha’s mental states make available to her “a consequentially justified killing of an innocent,” it is not possible for her to bring that about. So she does not have a reason to bring that about. Maybe the asymmetry between what action types are possible for her and what reasons there are for her is a good result. It helps us explain why Samantha does not do anything wrong— she doesn’t act in violation of a reason— but when you break confidence as you ought and can, you do have a reason since you can, and so you violate that reason.

The breakfast case nicely illustrates that we should individuate "human acts,” as Aquinas calls them, using the end a person has in mind. When you innocently have breakfast, it’s not a human act. It’s not one that falls into the type that can be appropriately evaluated as morally good or bad because of your ignorance. Instead, we will want to know more about why you forgot your promise to not eat breakfast. Was something you did knowingly that lead to your forgetting a human act? If not, then the ignorance is genuinely innocent. If so, then that action is the one we can evaluate as morally good or bad, but eating breakfast is still not going to be morally evaluable.

Alexander R Pruss said...


This is very helpful.

"It extracts the end from its context as a means to some further end": This suggests that it is different to intend something as a means or as an end. When you intend it as a means, you intend it-as-a-means. Thus, the man in Socrates' example in the Lysis who learns that his son's health will benefit from wine does not intend that his son drink wine, but intends that his son drink health-inducing wine. (This helps explain Socrates' claim that the man is a not a "friend" to the wine, but only a "friend" to his son's health.)

"What I want to say is that one cannot have a reason to do what is impossible."

Maybe what we should say that while one cannot have a reason to do what is impossible, one can have a reason to try to do what is impossible. Suppose it hasn't yet been discovered that angles cannot be trisected. You'll be killed unless you trisect an angle. Maybe you're right and you don't have reason to trisect an angle. But you sure have reason to try! Normally, one has reason to try because one has reason to do, and there is no difference between what gives reason for both. (Moreover, in the case of successful action, I think the token trying is identical with the token doing.)

(By the way, what do you think of Albritton's example in his presidential address of how you can try to do something you see is impossible? "Coach, I can't do that." "Of course you can't do it yet. But *try*, and one day you'll succeed." (Paraphrase.) I don't know if this is a real trying...)

Samantha has a reason to try to permissibly save ten people.

"When you innocently have breakfast, it’s not a human act."

I don't understand that. It's an action chosen in response to reasons, and hence morally evaluable. If, for instance, you ate too much or too pickily, you may be guilty of gluttony (Aquinas says that both too much and too picky are species of gluttony), and if you ate in just the right way, you may be praiseworthy for enjoying the gifts of creation. It is only in the aspect of the nonfulfillment of the promise that the action is accidental.

Maybe we can say that the eating of breakfast is a human act. But the contra-promise eating of breakfast is not a human act?

Or maybe the "human act" distinction is too coarse-grained to capture this. Following Davidson, the eating of breakfast is the same event as the contra-promise eating of breakfast which may be the same event as the alerting of burglars (let's say that they hear the slurp-slurp). So all of these are human acts. But only one of the descriptions describes it according to its actish essence ("essence" in the medieval sense, not in the modern modal sense--Mike Gorman has very good papers on "essence"), and that's the eating of breakfast.

Compare: Alex the human is the same substance as the dark shape in the corner behind a lit screen. But only one of the two descriptions describes that substance according to its essence.

Maybe in the first instance moral evaluation is evaluation of an act according to its essence?