Now it must be remembered that, in a sense, our original theory does hold and even insists that this is the case. We have, for instance, frequently referred to it in the last chapter as holding that an action is only right, if it produces the best possible consequences; and by “the best possible consequences” was meant “consequences at least as good as would have followed from any action which the agent could have done instead.” It does, therefore, hold that the question whether an action is right or wrong does always depend upon a comparison of its consequences with those of all the other actions which the agent could have done instead. It assumes, therefore, that wherever a voluntary action is right or wrong (and we have throughout only been talking of voluntary actions), it is true that the agent could, in a sense, have done something else instead. This is an absolutely essential part of the theory.
But the reader must now be reminded that all along we have been using the words “can,” “could,” and “possible” in a special sense. It was explained in Chapter I (pp. 29-31), that
