A man stands on the shore. In the distance, two people are drowning. One of them is a stranger, the other is his wife. He can only save one. Who should that be?
According the the philosopher Bernard Williams, the two dominant moral traditions of his day - Kantianism and utilitarianism - both gave the wrong answer to this question.
But this is not because the man should have saved X, and the theories said he should save Y. Rather, the mistake the traditions made was to think of it as a question at all.
To explain. According to Williams, both moral theories required the man to think about what morality prescribed. If he was a consequentialist, for example, he should stand there and figure out which act would have the best consequences - saving his wife or saving the stranger. If he was a Kantian, he should stand there and figure out whether the categorical imperative recommended saving his wife or saving the stranger. And then, once he had figured out what morality demanded, he could dive in and do the morally right thing.
But! But, says Williams - that’s his ****ing wife! He shouldn’t stand there thinking about what the morally right thing to do is, he should just dive in and save his wife.
In this situation, it wasn’t the mark of a morally good person to deliberate before acting. It was rather the mark of some kind of psychopath; or at least, a sign that this man hadn’t the slightest idea of love or of loving his wife.
All this is summed up in a phrase - to think about the demands of morality in order to decide whether to save your wife or the stranger is to have one thought too many.