Over at Cydonia Mensae, James McConnell asks (among other things) for atheists to explain where our morals come from and how we can distinguish right from wrong. I noted that there are two broad approaches that many atheists take that seem to me to provide a sufficient, reasonable explanation:

First, moral behavior may have an evolutionary explanation — and there is a robust field of study called evolutionary psychology that sets out to answer that question. Under this view, of course, morality is “universal” only in the sense that it might be true of all humans, but not all potential hypothetical abstract beings.

This means that (for example) if hyperintelligent evolved ants landed on earth, we might not have an independent moral code to which we can appeal in order to convince them not to exterminate us. Indeed, evolved ants might have no sense of individual identity from which to derive a sense of morals that would value, say, individual rights.

The alternative view is to contend that morality is absolute in the same sense that, say, the rules of logic are absolute, and that morality is therefore a component of all rational beings. In that scenario, regardless of the evolutionary history, once a race of hyperintelligent ants developed the capacity for reason, they would roughly arrive at the same moral code that humans share.

In that case, atheism remains compatible with this kind of absolute morality, and there are a wide variety of exclusively secular moral theories that nevertheless arrive at moral absolutes. Probably the starting point for all of these is the work of Immanuel Kant, who shows in his Prolegomena to Any Future Metaphysics how synthetic propositions (that is, propositions that reveal to us new knowledge rather than just restating the predicate) can nevertheless be a priori in the human brain (that is, prior to and independent of our experiences). In other words: Kant — and many others — show that atheism does not require one to adopt some sort of subjectivist moral code.

After expressing this opinion on Cydonia Mensae, in the comment section, Dylan Barry offers three criticisms of Kant. I think they’re worth answering:

[First,] Deontological ethics (Kant) does not adequately account for situations that may result in the conflicting of two moral rules and the duty to keep them. I think of the following example. Say I am a father whose children are starving and who without food will most certainly die, but I am poor and do not have the money to purchase the food. The only possible way that I could feed my children is by stealing, but the act of stealing would be wrong while letting my children starve to death would be wrong too. Neither would be wished for as an categorical imperative. What do I do?

This is absolutely a fair criticism of Kant with which all proponents of deontological ethics struggle. One popular solution is to distinguish between acts and omissions, which is to say that the locus of evaluation for moral behavior is with the agent and not the object (the “patient,” as the SEP calls it).

On the agent-centered view, then, universal laws are negative in character; they require you to refrain from doing certain things (stealing) but not to affirmatively perform actions designed to support others. Under that view, the father should not steal the food even if the consequences are that the children die. This view is, I think, most consistent with Kant’s philosophy. It is also powerfully embedded into our legal system; the concept of “unclean hands,” for example, can prevent a person in whole or in part from suing someone else for damages even if they have been wronged by the other person.

In addition to the agent-centered view, the SEP gives six additional responses by neo-Kantians that I think are worth considering, although ultimately I do not find them as faithful to Kant’s original work.

Second, Kant’s ethic discounts as ethical any action that is not done solely out of duty. I see this as restrictive. If I carry out acts of charity by offering time at the food bank because I like helping people and I like the feeling of happiness that accompanies helping others, Kant would say that I was wrong in using acts of charity as a means of happiness and should have seen it merely as a duty. Such a theory I think fails to take into full account of humans and their intentions in doing good.

Respectfully: I think this misapprehends Kant. As set forth in the Groundwork of the Metaphysics of Morals, our obligations to others are not to use them as a means only. Here’s how Kant puts it:

Now I say: man and generally any rational being exists as an end in himself, not merely as a means to be arbitrarily used by this or that will, but in all his actions, whether they concern himself or other rational beings, must be always regarded at the same time as an end.

[emphasis added]

In other words, we do not violate the categorical imperative every time we use someone as a means — if that were true, Kantianism would be absurd. I “use” the parking garage attendant as a means to park my car at work every day; I “use” the deli counter clerk as a means of securing my lunch, and so on. This is perfectly fine. The problem arises if I were to use these people as a means only; i.e., if I were to act towards them only to achieve my desires and not at the same time regard them as individuals with intrinsic worth, valuable in and of themselves.

This misconception of Kantian duty leads to the compound error suggesting that Kantianism is incompatible with supererogatory acts; that is, with acting above and beyond our minimal negative duties. Not so. Indeed, whether one gives to charity because it makes you happy or because you feel an obligation to affirmatively aid others, such actions are permissible (and morally praiseworthy) within the Kantian framework so long as those actions do not themselves treat the objects of your charity as a means only. So all Kant would rule out would be something like stealing from the rich to give to the poor, or forced organ transplants, or other utilitarian horrors.

[Third,] Another objection to deontological ethical theory is the failure to take into due consideration the consequences of one’s moral decision. Say I were a Swiss man that had decided to harbor Jews during World War II from the Nazis. I receive a knock at the door and it is a Nazi soldier inquiring if whether I have any Jews in my house. Moral duty would dictate that I ought not to lie and say that I am not harboring them and so I must say that I am. However, the consequences would mean those under my care would be shipped off to the concentration camps or maybe even killed. I know that we could argue for the situation to be judged based on its utility, however, this is not what I think Kant would agree with.

I think there is an easier solution to this dilemma, although it requires parsing Kant particularly carefully. There is — as far as I know — no general injunction against lying within the Categorical Imperative, because lies per se do not generate a contradiction of the will. Instead, there’s a prohibition against making a lying promise. because we recognize that promises themselves depend upon the inherent notion that they will be kept to be of value. Here’s how Kant puts it:

He who is thinking of making a lying promise to others will see at once that he would be using another man merely as a mean, without the latter containing at the same time the end in himself. For he whom I propose by such a promise to use for my own purposes cannot possibly assent to my mode of acting towards him and, therefore, cannot himself contain the end of this action.

In other words: so long as you have not promised to the Nazi guards that you will reveal Jews to them, I don’t see why you have an independent duty to volunteer information to guards who knock on your door and demand that you do so. Moreover, I assume that by harboring the Jews in the first place you have explicitly promised to them that you would not reveal their hiding place, so it seems to me that a proper understanding of Kant actually solves this dilemma in the most intuitive manner without resorting to a consideration of hypothetical ends.

Finally, let me add that I do not think one needs to be convinced that Kantianism is true; all one need do is recognize that it is a plausible accounting of objective morality without resorting to some divine command.