Universally Preferable Behavior and the Maxim Description Problem

[Part of The Molyneux Project; read the main critique here]

So now that I'm no longer going to be communicating with Stefan directly, I'm not sure if there are any real prospect of my actually understanding his book. Still, I think it might be worthwhile, if anyone's interested, to lay out a bit of criticism of some of the main themes that I actually found objectionable, and not just confusing. This post won't represent an attempt to completely tackle that task. Rather I'll focus on one idea which comes up several times in Stefan's book, and which represents a significant problem for any theory taking the form that Stefan's does.

A cursory examination of Stefan's book reveals that he's working in a distinctly Kantian framework, in the sense that he takes morality to be based on maxims which are evaluated for moral goodness or badness by universalizing them and determining whether they come into contradiction with themselves (Kant introduces this framework in his Fundamental Principles of the Metaphysic of Morals).

Now, I have a long track record of being accused of misrepresenting Stefan's positions, and I'm very comfortable with the assumption that I'll be accused of doing this here in describing his methodology. Nevertheless, I believe that the proof is on the paper, and I will take great care to establish beyond what I think is a reasonable doubt that this is the approach that Stefan is using. I apologize in advance if this is tedious.

Stefan writes, on page 43:

...the first test of any scientific theory is universality. Just as a theory of physics must apply to all matter, a moral theory that claims to describe the preferable actions of mankind must apply to all mankind. No moral theory can be valid if it argues that a certain action is right in Syria, but wrong in San Francisco. It cannot say that Person A must do X, but Person B must never do X. It cannot say that what was wrong yesterday is right today - or vice versa. If it does, it is false and must be refined or discarded.

Then, on page 44, he writes:

If I say that gravity affects matter, it must affect all matter. If even one pebble proves immune to gravity, my theory is in trouble. If I propose a moral theory that argues that people should not murder, it must be applicable to all people. If certain people (such as soldiers) are exempt from that rule, then I have to either prove that soldiers are not people, or accept that my moral theory is false. There is no other possibility. On the other hand, if I propose a moral theory that argues that all people should murder, then I have saved certain soldiers, but condemned to evil all those not currently murdering someone (including those being murdered!) - which is surely incorrect.

If, to save the virtue of the soldiers, I alter my theory to argue that it is moral for people to murder if someone else tells them to (a political leader, say), then I must deal with the problem of universality. If Politician A can order a soldier to murder an Iraqi, then the Iraqi must also be able to order the soldier to murder Politician A, and the soldier can also order Politician A to murder the Iraqi. The application of this theory results in a general and amoral paralysis, and thus is proven invalid.

It seems clear that Stefan is using a maxim-based approach here which is more or less identical to the one used by Kant. That approach states moral precepts in the following form: "In all circumstances C, I will do X," universalizes them, and searches for problems. It's worth pointing out that Kant discusses two different ways in which a universal law can come into contradiction with itself: it can be impossible to conceive of a world in which it is followed, or such a world could not be desired by any rational person. Stefan does not explicitly assent to this distinction, but I think that there is evidence that Stefan would accept both sorts of "contradictions" as sufficient reason for rejecting a maxim.

On page 66, Stefan writes:

Raping someone is a positive action that must be initiated, executed, and then completed. If "rape" is a moral good, then "not raping" must be a moral evil - thus it is impossible for two men in a single room to both be moral at the same time, since only one of them can be a rapist at any given moment - and he can only be a rapist if the other man becomes his victim.

He continues:

...two men in a room must be considered to be in the same situation. If only one of them can be good, because goodness is defined as rape, and only one of them can rape at any time, then we have a logical contradiction that cannot be resolved.

I take this to be pretty good evidence that Stefan accepts Kant's first sort of contradiction as a reason for rejecting a maxim. The maxim, "I will always rape," cannot be adopted as a universal law; it would be impossible to conceive of a world in which it was. Stefan also seems to accept that Kant's second sort of contradiction, that a reasonable person would be unable to will the universal adoption of the rule, would be problematic for a maxim. On page 80, Stefan writes:

...if stealing is good, then goodness becomes a state achievable only in the instant that Doug steals Bob's lighter. In that instant, only Doug can be moral, and Bob cannot be. After that, goodness becomes impossible to achieve for either party, unless Doug keeps giving Bob's lighter back and then snatching it away again.

Of course, it seems patently ridiculous to imagine that the ideal moral state is for one man to keep giving another back the property he has stolen, and then immediately stealing it again. Thus logic seems to validate our instinctual understanding of the foolishness of this as a moral ideal...

Stefan goes on to argue that a maxim advocating the practice of stealing would be contradictory for the reason discussed above (it would be impossible to conceive of a world in which everyone consistently advocated theft), but his statements here suggest that Kant's second variety of contradiction would also provide some grounds for rejecting a maxim.

It might have been noticed that I set up the phrasing of a maxim to explicitly include circumstances, and yet in many of the examples above, Stefan did not make any mention of circumstances. I would point out that in any situation in which circumstances are not explicitly set out, it is not unreasonable to think that the maxim could simply apply to all circumstances. Further, there are examples in which Stefan does make reference to circumstances as potentially providing basis for making moral distinctions. For example, returning to the soldier example on page 44, Stefan suggests that we could alter a moral theory to say that "...it is moral for people to murder if someone else tells them to (a political leader, say)..." Here we would need to make use of the maxim structure outlined earlier: "In circumstances in which another person tells me to murder, I will murder."

There is further reason to believe that circumstances should be relevant in the exposition of a moral theory: It could be that in some circumstances, an otherwise impermissible action would be perfectly acceptable. For example, it is generally impermissible to kill another person. But if that person is attacking you, then it seems that it would be acceptable to use lethal force in self-defense. The maxim, "I will kill people" seems obviously problematic; this is not so with the circumstance-dependent maxim, "In circumstances in which I am being attacked and can save myself only by killing my attacker, I will kill my attacker."

But none of this seems particularly problematic for Stefan's theory. So far, I haven't really done anything but clarify how I think the theory is supposed to work. So what's my problem? The objection is one that has been commonly raised against Kant; I'll call it the Maxim Description Problem (there's probably an "official" name for it somewhere, but I don't know what it is). The problem is this: Our evaluation of an action will depend on the maxim on which we are acting when we perform it. But any individual action can be justified by a number of different possible maxims. Accordingly, we could reach the conclusion that an act is permissible or impermissible depending on the maxim that is being used to justify it.

To invoke a classic example, we might imagine that a person has come to your door and is asking where your friend is. Unbeknownst to the person at the door, you know that the reason that she is asking is because she is a murderer who wants to kill your friend. You know where your friend is, and know that if you tell the murderer his location, she will certainly kill him. You also know that if you lie, nothing bad will happen to you, and that your friend will be able to escape with his life. What should you do?

On one hand, you can consider the maxim, "Whenever I can promote the outcomes I desire by lying, I will do so." We can imagine that if everyone adopted this maxim, no one would ever accept anyone's word for anything, and there would never be any occasion to lie in the first place. Such a maxim would simply make no sense as a universal law. On the other hand, as Christine Korsgaard points out in her essay, "The Right to Lie: Kant on Dealing with Evil," it would not be absurd for everyone to adopt a maxim by which they would lie in order to keep a murderer from her victim. Here we might imagine the maxim to be "Whenever I can prevent someone from murdering a person by lying, and I know that no other negative consequences will occur on account of my actions, I will lie."

This brings into focus an important problem, however. The fact that a particular maxim is unacceptable for justifying a particular action does not mean that there is no acceptable maxim which could justify it. So it could be that an action is morally permissible, even if we can show how a number of maxims which could justify it are unacceptable. This becomes a problem for Stefan in a number of different places in his book, where he seems to try to argue that certain actions are morally impermissible by rejecting particular maxims on which they could be based.

For instance, returning to the soldier example, Stefan wrote:

If, to save the virtue of the soldiers, I alter my theory to argue that it is moral for people to murder if someone else tells them to (a political leader, say), then I must deal with the problem of universality. If Politician A can order a soldier to murder an Iraqi, then the Iraqi must also be able to order the soldier to murder Politician A, and the soldier can also order Politician A to murder the Iraqi. The application of this theory results in a general and amoral paralysis, and thus is proven invalid.

This is fine, but as I've noted before, it does not prove that soldiers are wrong in killing people (murder is a bad word to use here, because the definition of the word "murder" is "wrongful killing," and therefore it is conceptually impossible for murder to be permissible). It only proves that soldiers' killing people cannot be justified by the maxim in question. It is logically possible that there is some other maxim which would justify soldiers' actions.

And indeed, such a maxim is not too difficult to come up with. One example would be the maxim, "Whenever I have declared myself to be a combatant of a particular group in a universally recognized manner, and I can kill a recognized combatant of another group who has not surrendered in a universally recognized manner, I will do so." That maxim could be adopted as a universal law without any contradiction that I can think of. But if this is the case, then how is the maxim-based approach supposed to serve as a moral guide? It might seem like any time we come across an action that cannot be justified by a particular maxim, the most we would be able to say is that we simply aren't sure if it's permissible or not. In order to effectively put Stefan's methodology to work, we need a way to determine what the right maxim is.

This problem is brought into focus by another factor in the equation, known as the Principle of Formal Equality. This principle states that in order to treat two things as ethically different, there must be a ethically significant difference between them. We have already seen evidence that Stefan accepts this principle, when he wrote:

Just as a theory of physics must apply to all matter, a moral theory that claims to describe the preferable actions of mankind must apply to all mankind. No moral theory can be valid if it argues that a certain action is right in Syria, but wrong in San Francisco. It cannot say that Person A must do X, but Person B must never do X. It cannot say that what was wrong yesterday is right today - or vice versa. If it does, it is false and must be refined or discarded.

The implication here is that being located in Syria as opposed to San Francisco cannot represent an ethically significant difference between two scenarios; being Person A instead of Person B is not a morally relevant way to distinguish a circumstance; taking place today instead of tomorrow cannot matter to an ethical theory. For example, notice that the maxim, "Whenever it is May 6, 2008, and I am Danny Shahar, and I can steal a pen from the University Book Store without anyone ever noticing, I will do so," could be adopted as a universal law without coming into any sort of contradiction with anything. But surely that doesn't mean that it would be okay for me to steal the pen. The Principle of Formal Equality helps us explain why: the maxim I've offered is unacceptable because it only works because of distinctions that aren't morally relevant in any way.

However, as we have seen, maxims can coherently contain certain distinctions which allow them to apply only to actions which occur in specific circumstances. So, for example, I am justified in killing in self-defense when I'm being attacked, even though there are other scenarios in which I'm not justified in killing people. What is needed, then, is an account of what kinds of features of a set of circumstances are morally relevant. If we had such an account, we could conceivably come up with a proper description of the set of circumstances in which your action was taking place, and then determine whether the maxim based on that set of circumstances could be acceptably adopted as universal law.

Unfortunately, as far as I can tell, Stefan doesn't provide any such account. And perhaps more unfortunately, no one else has either. Until someone does, it seems like any maxim-based approach to ethics is going to be hampered by the Maxim Description Problem. Annnnddd...that's part of the reason why I take a rights-based approach to ethics! [Added later: turns out rights-based approaches to ethics aren't that great either :-P]

So that's one problem with Stefan's view. When I get a chance, I'll try to touch on some more problems in future posts. But for now, I hope this was helpful!