This essay is the relic of a projected history of ‘individualism in politics’. That enterprise – the reasons for abandoning it will be transparent from what follows – prompted some reflections which may be of more general interest. For no habit is better established in the humanities and social sciences, as well as in political speech, than that of constituting a subject-matter for oneself by means of an -ism. And while there may be instances where the context makes this practice unproblematic, a survey of the past and contemporary record of -isms gives grounds for suspicion and reservations.

Over the past fifteen years, there has been a steady stream of books and articles on business-government relations describing the ‘privileged position’ occupied by the business corporation in the American political system. Taking issue with the pluralist paradigm that dominated writing and research on American national politics in the two decades after the Second World War, these writers have argued that business is not simply another interest group. Instead, they have suggested that its role in American society is more akin to that of a dominant class, power elite or private government: it thus possesses a degree of influence that invariably exceeds that of any other class or interest group. This appraisal of the political dominance of business in contemporary American society primarily rests on four sets of interrelated observations. These include the ability of business to define the political agenda; the extent to which business gains disproportionate benefits from the political process; the need for elected officials to maintain a high degree of ‘business confidence’; and the superior capacity of business interests to mobilize political resources, work closely with each other and shape the climate of public and elite opinion.

Elections in democratic societies are open, well-publicized events. But preceding the glamour of modern election campaigns and the drama of election-night broadcasts there is a process which usually receives little attention or publicity. Political parties, which dominate elections, choose their candidates.

A concern with the relationships between different levels of centralization and the management of change is long-standing in many of the social sciences. The French educational system has always stood in the minds of many as a monument to over-centralization and red tape. In her recent magisterial study, The Social Origins of Educational Systems, Archer has examined the consequences for educational change of different levels of centralization in the educational system in a variety of national settings. In her book she argues with great care and a wealth of historical example that it is possible to establish certain broad generalizations which describe the manner in which educational change is produced, and that sharp distinctions in this regard must be made between the centralized systems and the more decentralized systems. Her study covers the period up to 1974. Since then there have been very important changes in the French educational system. These changes and the political conflict which surrounded them, particularly between 1974 and 1978, are of considerable interest in their own right, to students of both comparative public policy and French politics. The principal purpose of this discussion, however, is to examine Archer's propositions about the modes of change typical of centralized systems in the light of the reforms introduced in these years. My concern is therefore with the administrative and political consequences of centralization and this inquiry leads in turn to some discussion of the role of pressure groups in the Fifth French Republic.

David Miller's article ‘Market Neutrality and the Failure of Co-operatives’ has two principal theses. The first of these, with which I have no quarrel, is that co-operatives will have a difficult time surviving in a free market economy. The second is that the truth of the first thesis suggests that a crucial argument in favour of libertarianism is wrong.