lar kind of action may produce good effects, in other circumstances a precisely similar action may produce bad ones. And, since the circumstances are always changing, it is extremely unlikely (though not impossible), that actions of any particular class, such as murder or adultery, should absolutely always be right or absolutely always wrong. Our theory, therefore, does not imply that, if an action of a particular class is right once, every other action of the same class must always be right: on the contrary, it follows from its view that this is unlikely to be true. What it does imply, is that if we consider any particular instance of any class, that particular instance cannot ever be both right and wrong, and if once right, must always be right. And it is extremely important to distinguish clearly between these two different questions, because they are liable to be confused. When we ask whether the same action can be both right and wrong we may mean two entirely different things by this question. We may merely mean to ask: Can the same kind of action be right at one time and wrong at another, or right and wrong simultaneously?
