I offer below an incomplete sketch of my approach to morality. It agrees with much of what Gilbert Harman has written.
Writings on ethics tend to treat two different questions under the title of “ethics.” One is a question of prudent behavior; the other is a question of moral behavior. But, I must admit, that some contexts are such as to raise the perplexing question as to which has priority.
To distinguish the question of prudence from that of morality, let us consider the case of a solitary individual cast away on an island — call him Crusoe. His primary problem is one of survival. Given his desire for survival, he will have to follow, what Kant called, hypothetical imperatives. There are things he ought to do in order to survive. If he has tools, there will be right and wrong ways of using them. Again, if he is interested in efficiency, he must do things in certain ways. Will certain dispositions and habits — call them virtues — serve him well? Certainly. He can strive for happiness, which will be a function of leisure time. But I make no claims about what constitutes happiness. Whatever it is, it will be a matter of satisfying his interests. And since interests vary, so will the ways for reaching happiness. What I want to claim is that the question of moral behavior does not make sense for a solitary individual — unless he lives in a mythical world with another person — a god or spirit.
Let us now introduce another person to the island — call him Friday. I will call core morality the behavior which both of them agree to regarding each other. This will be expressed by rules. And these rules can be negative or positive, depending on what they forbid to be done to each other, or required to be done to each other. Justice will consist of abiding by the freely made agreements. (Agreements may be complicated because of tacit assumptions.)
However, there is also the most important question of trust. Can Crusoe trust Friday, and vice versa? Assuming that both are sincere, this does not settle the question of trust. Whether each is trustworthy is in part determined by their character, or, to put it another way, by their possession of virtues.
They can also agree on how to behave towards plants, animals, and the rest of their environment. Such rules I will call marginal morality.
If we introduce a third person on the island — call him Tarzan, with whom we do not interact, then Crusoe’s agreement with Friday in regard to Tarzan can be called peripheral morality. Peripheral morality may allow us to kill, rob, torture, sacrifice, or enslave Tarzan.
Theories of normative morality have been either teleological or deontological, i.e., based on consequences or based on rules. If I were Crusoe entering into an agreement with Friday, I would agree to the rule:
Neither of us will harm the other, except for self-preservation.
This is the element of egoism that is presupposed in normal core morality. But there is also, what I will call abnormal core morality, which takes place in sports in which the requirement of not harming each other, and the exception of self-preservation are suspended, and what is done is done strictly by the rules regardless of consequences.
This code of honor, as it way be called, requires a strict obedience to the rules with no exceptions for self-preservation. In extreme cases, it has the form: I will do X, even if it kills me; and if I do not do X, I will kill myself. Such was the code of the kamikazi and the samurai.
In making rules, one can also agree to punishments for breach of rules, and also for allowable excuses. Given that we have prudential considerations, sometimes one has to make a judgment call about how to act. Or, as C. D. Broad would say: What is the most fitting way to act under the circumstances?
[See my “Concepts of Persons and Morality“]