Jeffrey Stanyer

ELECTORAL BEHAVIOUR IN LOCAL GOVERNMENT A MODEL OF A TWO‐PARTY SYSTEM*

AMONG the many features of local government which have been the subject of criticism in recent years problems of local democracy have been prominent. None of the discussions have seemed at all conclusive and many of them are not even plausible. The role of political parties, the characteristics of local electoral systems and the aldermanic system are major aspects of local democracy but it is hard to point to a single authoritative analysis of any one of them.1 The purpose of this article is to show how the sort of problems encountered in this field can best be investigated with the aid of a theoretical model. The word ‘model’ is used in political and social science in a number of different ways, but only if it is given the following meaning will individual models have the intellectual virtues here ascribed to them. In this context a model is a deductive system, specified in terms of a set of abstract symbols, rules for forming statements containing them and rules for the transformation of well-formed statements. The latter rules are always rules of deductive inference. Statements in the model are divided into two classes: those which are premises and the rest. A model contains a finite number of nonredundant premises as its axioms, the fewer the better. All other statements in the model must be deducible from the axioms, directly or indirectly. A theoretical model is one which can be used as the basis of theory and this occurs when some of the symbols of the model are given a meaning in the data to which it is being applied. This is often referred to as ‘providing operational definitions’ of the symbols. Thus a model is abstract (that is, contains only uninterpreted symbols for variables and constants), whilst a theory contains the names of classes of things found inthe appropriate subject matter. Thus the conclusions in a theory are predictions about the way the

IRRESISTIBLE FORCES: THE PRESSURES FOR A SCIENCE OF POLITICS*

AMONGST the signs of the growth of a discipline is the appearance of two related ‘crises’-of identity and legitimacy-experiencd by those who believe they belong to the same intellectual fraternity.’ When a subject reaches a degree of sophistication and complexity those who study it begin to question the nature of the enterprise in which they are engaged and its status in comparison with others. Questions about identity are those concerning its relationship with practical affairs and its ‘autonomy’ from longer and better established subjects. Does it give a better understanding than commonsense? Should it be taught to fledgling citizens and potential administrators and politicians ? Questions about legitimacy are those that relate to the scope, rigour and adequacy of its paradigm methods. Does it pass the tests that ought to be applied to all intellectual subjects? If the answer to the second type of question is wholly or partly in the negative individual students do not surrender their claim to a place in the academic sun but engage in search behaviour. The purpose of this paper is not to describe the recurrence of these two crises in the study of politics in the twentieth century but to offer an interpretation of the apparently dialectic nature of the evolution of the discipline? If those who seek to advance political science are explorers, then this is a special sort of map, derived from the experiences of those who have been lost. Like all explorers’ maps it is a picture of the partly unknown.

LOCAL SUPPORT FOR NATIONAL POLITICAL PARTIES: A THEORETICAL ANALYSIS1

A PREVIOUS article outlined a logical model of a two-party system in local government. This model is a deductive system specified in terms of abstract symbols and is intended to serve as the basis for a theory ofelectoral behaviour in party systems in local government. The only exogenous variable in the original model is nationalsupport for the two parties and the analysis is founded on an examination of the consequences of this assumption.2 The fourth axiom of the model assumed that the relationship between local and national support for a political party is a linear function of the form L1 = bN1 + d where L1 and N1 are local and national support for partyl respectively and b and d are constants whose value may differ from authority to authority but which do not change within an individual authority. It was further assumed that there are limits to the range of variation that national support can show. Thereason for this is that if the full range of variation logically possible is permitted the function relating local and national support cannot be linear unless b = 1 and d = 0. The line generated by the function must necessarily pass through the points (0, 0) and (100, 100); if the country is heterogeneous (as assumed by the model) then b# 1 and dZO.3 The true line must therefore diverge from the straight in at least some sections of its total length-in particular at the two extremes. The greater the heterogeneity, the greater the divergence. Thus the very convenient assumption of linearity can only be retained if the variability of national support is restricted to the central section of its total possible range. Both the assumption and the restriction might be defended on empirical grounds. The relationship between the two factors does appear to be approximately linear in its middle section and there do appear to be forces which prevent either party in a two-party situation in a system of free elections from being annihilated.4 It is preferable, however, to give a theoretical account of the relationship which is central to the model-which in fact was the fourth axiom in the system originally presented. A justification on other than empirical grounds is thus highly de~irable.~ It is of course necessary that the theoretical account be consistent with the assumptions of the original model and this is best achieved by using the latter as far as possible. Such a policy also