I hold, then, that the rational way to decide what to do is to decide to perform that one of those alternative actions open to us (including the null-action, the doing of nothing) which is likely to bring about the total situation now and in the future which is the best for the happiness or well-being of humanity as a whole, or more accurately, of all sentient beings. The utilitarian position is here put forward as a criterion of rational choice. It is true that we may choose to habituate ourselves to behave in accordance with certain rules, such as to keep promises, in the belief that behaving in accordance with these rules is generally optimific, and in the knowledge that we most often just do not have time to work out individual pros and cons. When we act in such an habitual fashion we do not of course deliberate or make a choice. The act utilitarian will, however, regard these rules as mere rules of thumb, and will use them only as rough guides. Normally he will act in accordance with them when he has no time for considering probable consequences or when the advantages of such a consideration of consequences are likely to be outweighed by the disadvantage of the waste of time involved. He acts in accordance with rules, in short, when there is no time to think, and since he does not think, the actions which he does habitually are not the outcome of moral thinking. When he has to think what to do, then there is a question of deliberation or choice, and it is precisely for such situations that the utilitarian criterion is intended.
(J. J. C. Smart, An Outline of a System of Utilitarian Ethics [Melbourne: Melbourne University Press, 1961], 29-30)
Note from KBJ: Twenty years after this was written, R. M. Hare developed the idea (of two levels of moral thinking) at book length. See here. Hare of course listed Smart's essay (or rather, the 1973 version of same) in his bibliography.
