Fifteen years have elapsed since the British Parliament enacted legislation designed to curb air pollution. The 1956 Clean Air Act appeared seven years in advance of comparable American legislation and well before the quality of the environment became a matter of world-wide concern. In view of the fact that Queen Elizabeth I had complained about smoke pollution in London, or that nineteenth-century novelists had penned vivid portraits of London's smoke and fog, the 1956 Act should perhaps be viewed as being at least a century late in appearing. Nevertheless, the Act did mark a distinctive policy innovation, particularly with regard to those provisions relating to smoke pollution from domestic sources. An analysis of the impact of those provisions will be the subject of this paper.

Nobody, least of all anyone intending to allude to sociology and stay respectable, should nowadays dare employ a generalization from a single instance. Nonetheless (since this is an introductory paragraph, and therefore only rhetoric), 1 I shall permit myself to do so and excuse the lapse by calling it an illustration. A professor of my acquaintance recently observed that reading an essay on Montesquieu (by one of my better students) had almost persuaded him to go himself and read Esprit des lois. In this country, I suspect, Montesquieu is less read than merely acknowledged as ‘a precursor of sociology’ or even as one of the ‘great theorists’ of sociology in general and political sociology in particular. One of the contributions that I hope this article will make is to persuade others, not merely historians of ideas, that Montesquieu should actually be read, because his ideas and his difficulties are important, interesting and instructive.

Do novelists have something to say about politics? By raising the question in such a starkly rhetorical way it is tempting to respond, ‘Yes, of course, novelists (at least some of them) do have something to say on a wide variety of subjects which traditionally fall under the rubric of political thought’. One could quickly give examples: Tolstoy's devastating critique of military leadership in War and Peace, George Orwell's penetrating analysis of the reality of totalitarianism in 1984 and Jean Paul Sartre's portrait of a whole society in a state of imminent collapse in Le Sursis. With a little time for reflection it would surely not be difficult to expand this list to several hundred titles. But if the novelist is credited with having political thoughts or ideas it is rather odd to find so little systematic analysis and interpretation of those ideas by political scientists.² Why is this the case? For some political scientists perhaps the reason lies in their explicit acceptance of models of explanation which approximate those of the physical sciences. They wish to constrict their range of inquiry into politics in the interest of developing better models for explaining and predicting political phenomena. There is no need to quarrel with this perfectly sound approach, except where it leads its proponents into a dogmatic rejection of other modes of inquiry.

Several attempts have been made recently to explain the formation of coalitions, especially government coalitions. Most of these theories assume that each of the actors involved (political parties in the case of government coalitions) uses only one criterion — such as the ‘size’ of the coalition or its ‘ideological diversity’ — in choosing between alternative coalitions. The criterion yields for each actor a preference ranking of all the possible coalitions. It is then assumed that the coalitions which are most likely to form are those which are ‘in equilibrium’ or ‘undominated’, a coalition T being undominated if there is no other coalition which is preferred by all its members to T.

In recent years the discipline of political science has focused much of its attention on political parties. In 1967 Professor L. D. Epstein noted ‘Writing about Political Parties in Western democracies is not a novel enterprise. It is several decades since political scientists, once preoccupied with constitutional forms, ceased to neglect parties’. In Britain, however, there remains much uncharted territory which requires detailed exploration by the political scientist. Much of the research carried out in Britain has concentrated on one of two levels of analysis, either national or local. Basic information about the intermediate branch of party organization, the regional and area structures, is lacking. As Professor J. Blondel has noted: ‘The eleven regions of the Labour Party and the twelve areas of the Conservative Party are rarely examined’. There is, therefore, a gap in our knowledge of political parties in Britain, a gap which Professor R. T. McKenzie readily acknowledged in his own study of British parties.

Are local election campaigns just periodic acts of renewal, seasonal rites, or do they gain extra votes? Whatever may be true at general elections, low turnout is a reason for thinking that the scope for organization at local elections may be considerable.

Mr Ake is not the first mechanical determinist to argue for the possibility of a political science analogous to natural science. He is aware that there are objections to this position, but he scarcely meets them.

It seems odd that in 1972, political scientists still find it necessary to take the scientific pulse of their discipline, to exhort, cajole, and recommend, in the manner of a reproving family physician, their colleagues to ‘pay more serious attention to what the scientific study of a phenomenon entails’. Such an injunction is doubly regrettable when accompanied by a naive and confusing analysis of concept formation and confirmation in science, an account sure to encourage a dogmatic epistemology among social scientists who, in other circumstances (specifically, the opportunity to devote scarce time to careful study of the voluminous literature in the philosophy of science), would likely identify themselves with a more catholic epistemology and methodology. Mr Ake, in his account of the logic of scientific inquiry, glosses major issues in the philosophy of science, and, as a result, tacitly represents the community of scholars in the latter discipline as univocal with regard to fundamental issues of language, logic, and epistemology. I shall briefly try to state more of the case than Mr Ake has done, in a manner so that I might not too be accused of glossing (or, at least, of glossing with a minimum of distortion) the major issues he has raised in his note.