Countering Kant's Categorical Imperative

The categorical imperative is Kant's central contribution to moral philosophy.  It is not really useful in itself for guiding one's moral action but acts as a sort of test for a given ethic:

Act only according to that maxim whereby you can at the same time will that it should become a universal law without contradiction.

It would seem then that we might proceed to list a number of proposed moral rules, think really hard about whether each adheres to this meta-rule, and then write the ones that do in our Book of Moral Laws.

The famous counter-example to Kant's imperative is lying to a murderer (after Kant proclaimed that "never lie" fits the criteria for being a universally valid law).  If you want the example to seem even more egregious, use a Nazi.  For example, Anne Frank is hiding in your attic, a Nazi enters your house and asks you "Is Anne Frank hiding in your house?"  Kant maintains that it would still be wrong to lie; just about everyone else thinks it would be okay.

Why does this counter-example work?  I think there are two reasons.  First, there is the trouble with abstracting the situation to a neat ethical quandary.  Much of philosophy has been haunted by the idea that reality can be expressed in true or false propositions, but such abstraction must fall short.  Words and propositions are tools for vaguely referring to actual experience, not the other way around.  As such, it might be as true (or false) to claim that the decision is whether or not to lie as opposed to helping a murderer or even whether to utter the words "not here".  Or if we persist in claiming the correct abstraction is whether to lie, what exactly constitutes a lie? 

But even if we "play along" that we can meaningfully reduce any scenario to a particular true or false ethical proposition, the counter-example still defeats Kant.  The reason relies on intuition: it just seems offensive to say that telling a lie would not be worth preventing a murder (or kidnapping).  In fact, the counter-example follows a formula that can be used to defeat any specific ethic.  Here is all you have to do: think of a situation where a small transgression is required to prevent a larger one.  Such hypotheticals can get to be pretty silly, but if you are imaginative enough you might come up with a realistic one.  You can get a lot of mileage out of the "man about to open fire on a crowd of people" scenario.  Invent some smaller bad thing you must do to stop him, like:

• push an old woman to the ground
• steal a make-shift weapon from someone
• destroy property
• kill a single innocent by-stander (?)

The last point is to show that there is a hazy line somewhere between what you think would be okay or not okay to do to prevent a catastrophe.  You could argue forever about where this line is.  But then you'd be missing the point, which is that there are no absolute lines.  Any ethic you produce, no matter how refined, can be defeated by inventing yet another scenario that would require breaking it in order to prevent a yet larger catastrophe.

A different approach is warranted for ethics.