The phrase, ‘political obligation’, is far more popular in English than in other European languages. Whether this may be due to historical circumstances, or to a peculiar bent of the English mind, is a fascinating question; but it is not the one which I propose to discuss here today. I am mentioning it only to explain the choice of my subject, a subject which would probably sound rather uncommon to an Italian audience, but which, I am sure, has a familiar ring to English ears. My subject, I would like to emphasise, is the nature of political obligation; my approach will be analytical rather than historical; and indeed, let me say at once that this lecture is a kind of examen de conscience in which I am myself involved, since I feel personally responsible for having introduced the phrase into my own native language, and for having used it during a life-time as if its meaning were self-evident, without ever raising the questions which I want to raise today.

In this Paper I am concerned to argue that the traditionally and contemporarily important arguments for a liberal society do provide a justification for what may fairly be called a liberal society. However, many liberals may wish to deny that the society which these arguments are seen to justify when their various limitations and qualifications are noted, can properly be called liberal, for it is less committed to non-interference with liberty than is the liberal society which those who advance these arguments believe to be justified by them.

A theory of punishment should give some account of mercy and yet it is true to say that very little has been said about it at all. It is commonly regarded as a praiseworthy element in moral behaviour—something to be practised occasionally both for the good of the one who punishes and the one who is punished. The suffering that punishment involves is unpleasant for all concerned, and if it is possible to avoid it or lessen it without moral injustice, then it is desirable to do so. We condemn as hard and unbending the judge who never shows mercy and the suggestion is that the poor unfortunates whose lot it is to be judged by him are poor unfortunates indeed. This is reflected in the fact that the opposite of merciful—merciless—means cruel. Presumably there are occasions when it is appropriate to show mercy to an offender, and other occasions when it is not. What are the conditions for the appropriate exercising of mercy, how do we decide how much mercy is appropriate, and when is a judge morally obliged to be merciful, if ever?

To some extent, perhaps under Moore's chastening influence, eccentric philosophical denials of the existence of physical objects, other people's minds, the past, and so on, have gone out of fashion. All the same there is at least one very common philosophical conclusion which, though not as extravagant as these, is no less paradoxical. This is the dogma that ethical statements could not describe anything at all, and the collateral claim that they could not be true or false. This is, I suggest, as paradoxical as traditional denials of the existence of chairs and tables, since the denial that ethical statements could be descriptive, like the denial that chairs and tables could exist, is opposed by the common observation that ethical statements do describe, just as the other is opposed by the common observation that chairs and tables do exist. I hope to make two things perfectly clear. First of all, that ethical statements do describe states of affairs; and secondly, that the reason ultimately given for saying that they could not describe anything, namely that they differ in verification from statements of fact, is only partly true, and its uncritical acceptance has led to the canonisation of the belief that ethical statements are non-descriptive and have no truth-values.

Line 24 on page 218 in the July number of this volume of Philosophy should read as follows:

naturally out of matter itself lifeless or that consciousness and intelli-