Author's Note: This chapter owes much to participants in a Seminar on European "tty Systems that I directed at the European University Institute in Florence between 1976 and 1979. as well as to sagacious advice from Rudolf Wildenmann.

plan. First, section 5 discusses the singular contribution of Duverger, who for a generation dominated the field of party studies - positively but also negatively, through the critical reactions his book provoked. Then follows a discussion of numerical approaches, which have varied from simple classifications to more complex quantitative analyses and theories of cabinet stability and cabinet coalitions, to culminate in Sartori's fully fledged typology of party systems (section 6). The succeeding section on parties and social cleavages is followed by a discussion of the rather different approaches of 'spatial* writers whose formal models paradoxically raise problems not dissimilar to the cleavages literature: to what degree parties and voters can act in complete freedom, and to what extent they are bound by historical conditions and expectations (section 8).
Most of the authors treated by this stage in the essay proceed from the assumption that, in principle, parties interact with one another as unitary actors. This assumption is dropped in section 9, which focuses on analyses of intra-party role conflicts as between ministers, parliamentary groups, party activists, party members and party voters. It underscores the lack of reliable comparative analyses of internal party processes, which leaves the field free to overly-general propositions, alleging general but unproven trends; e.g., the development of mass parties, of parties of integration, or catch-all parties.
Such general issues logically spill over, in the final section, into a discussion of the actual role of parties in government. While some authors see the role of parties increasing, others speak of an unavoidable decline, giving as reasons the growth of plebiscitary tendencies and manipulative political techniques, the increase of government bureaucratization, and a possible development towards a neo-corporatist society. This discussion merges into the question of the degree to which parties 'count* in government policies, leading further to the complex issue of how to assess the performance of different party systems. This raises problems of both an analytical and normative nature, including that of the different yardsticks by which performance may be 'measured'.
This overview is the work of one author, which implies that its coverage - though broad and hopefully concise - is also selective and unavoidably subjective. It grants from the outset that certain important areas are either not treated or are done so only in passing. Thus, the chapter does not deal with the literature on totalitarian movements or parties (whether fascist or communist), although this was perhaps the single most important concern of writers on parties from the 1930s

onwards. It deliberately leaves aside the flourishing literature of comparative electoral analysis, although that extensive body of writing throws considerable light on the manner in which parties are perceived by voters, and indirectly on the way party strategists seek to act and react on voter preferences. It offers little on internal party processes, which to this day remain very much a blank spot in comparative analyses - whether between different types of parties in any one country or in cross-national studies. Even in all other cases, this essay can do little more than point to problems, rather than indicate their solutions. The chapter, in other words, remains introductory. Its sub-title, 'an overview*, could also have read: *one author's private interim assessment'.
2. The precursors .-o .
The study and evaluation of political parties hardly represents a new field. Were not Bolingbroke, Hume and Burke the real pioneers of the study of party? These authors could become its first discussants, given the existence in Britain of conditions favourable for the legitimate entry of party: a well established representative regime, which acknowledged the plurality of interests and the inevitability of faction, and which came increasingly to regard a formal opposition (Bolingbroke) and joint endeavours to promote the national interest upon some particular principle (Burke) as appropriate characteristics of a "good* political system (for a valuable review, see Sartori 1976, ch. 1).
That relation also holds true in a reverse manner. In societies where such favourable conditions and concurrent political doctrines did not exist, the recognition of the legitimacy of party took much longer. At least two bodies of thought - which one might term the 'statist' and the 'democratist* traditions-for long resisted the full recognition of party as an accepted part of decision-making in parliament, let alone as large-scale organizations mobilizing an increasing part of the people outside it.
In many European countries - though not in all - the state was widely regarded as an institution of a higher moral order. Its sovereignty was to be well protected from the encroachment of special interest. Theorists of the state and constitutional lawyers refused to acknowledge the role of parties as contributing in any way to the formation of authoritative decisions, or at most assigned parties a restricted place, 'og., in elections and in the organization of parliamentary work, holding

that the state and its government should remain unfettered. A tradition of aversion to parties, or of at most a grudging hesitancy, still lingers - particularly among writers of constitutional law, but also among those for whom terms like 'Parteienstaat' or Tartitocrazia" retain an unmistakably negative connotation. (Against that view, see the polemic Tarteienstaat - Oder Was Sonsf in Grewe 1951.)
A similar reluctance to grant a place to party is also found in the democratist tradition. Rousseau's desire to safeguard the direct expression of popular will from representation and interference by *associations partielles* came to exercise a persistent influence. Durable links between members of a political community are thought to hinder the spontaneous processes of opinion formation and to foster political inequality. Such sentiments were clearly expressed by some of the great early writers on parties, for instance Ostrogorski and Michels. Ostrogorski (1902), who provided a first extensive analysis of the rise of the 'caucus* in British and American politics in the latter part of the nineteenth century, hardly hid his dislike of the phenomenon, and advocated the resort to ad hoc issue associations as the only way to allow the people really to determine their own fate. The norm of direct democracy is a strong element in Michels's classical elaboration of an 'iron law of oligarchy' in modern mass organizations (Michels 1911). The ideal of direct democracy could in fact become an alibi for Michels personally to embrace the 'heroic' elitism of Mussolini's Italy over the 'bureaucrat-ized' party politics which he himself had analysed in such colourful detail (for a brilliant and merciless critique of Michels, see Bonger 1934). In the same vein, a belief in the need for spontaneous elites to fight for socialism as expressed by Sorel, and more generally in the syndicalist tradition, contributed strongly towards an aversion against organized party development in a parliamentary setting (see the important introduction by Juan Linz (1965) to the Italian edition of Michels's Political Parties, in which he traces links between Weber, Michels. Sorel, Mussolini and indirectly Lenin). The sentiment was to revive much later in modem New Left thought, with its desire for new forms of spontaneous, collective, yet in practice often highly elite-centred. decision-making. (A reaction against party has also been a constant characteristic of American reform politics; e.g., in the movement for . non-partisan elections, the wish to take nominations out of the hands of professional politicians through primaries, the adoption of special rules to secure 'representative conventions', etc.)
Parties, then, were recognized earlier and more readily and universally in some European countries than in others. In countries with

old pluralist roots and slow evolutionary growth of representative
government, they fell in readily with existing diversity, m countries where the state was largely imposed on society, with a claim for absolute and undivided sovereignty, both statist and democratist doctrines made recognition relatively late and far from secure. The legitimation of parties is, in fact, a crucial variable for the comparative study of the development of mass democracy in Europe (cf. the four-thresholds analysis in Rokkan 1970a, ch. 3; also Dahl 1966, pp. 360ff; Daalder 1966a;andDahl 1971).
Early work on the role of parties forms usually an indissoluble part of more general works on the working of representative government by such writers as John Stuart Mill, Johann Caspar Bluntschi, Walter Bagehot, James Bryce or A.L. Lowell (for an excellent assessment of this genre of comparative politics writing, see Eckstein 1963). Characteristic of their work was a wide-ranging erudition, not limited to any special discipline. They used history, political theory, institutional analysis and social generalizations with little distinction between description and normative judgement. Whatever differences were found could be explained by 'history' or presumed 'national* character. Inevitably, they reasoned on the basis of a small number of countries. As these countries generally had stable institutions, institutions could easily be regarded as major causal variables, hence also as devices suitable for adoption elsewhere. The effect of these writers was to place a limited number of countries on the map of comparative analysis. They contributed mainly to general theories of parnamentarianism and electoral systems which seemed applicable across European countries (as they actually often proved to be in practice, as responsible
parliamentary government developed in one European country after another).
3. The impact of Weimar
The German Weimar Constitution of 1919 represented the culmination of a belief in the possibility of institutional engineering. But there was substantial disagreement among German writers about 'correct' theories of parliamentarism and the role parties were to play therein. Older theories about the need to preserve ihe state as an institution above society and party (as given a new influential interpretation by Cart ochmitt 1931), theories about the need to secure a correct 'balance* "ctween government and parliament so as to have a 'real parliament-

"rism' (as put forward in the influential book of Redslob 1918), and opecific complaints about the (mal)functioning of self-centred parties in an increasingly fragmented party system battled with new theories, which saw merit in an exact representation of the many social forces in parliament where the need for majority decisions would automatically enforce compromise and integration (for an excellent overview see vonAlemann 1973; also Glum 1965).
The rapid 'deconfiture* of Weimar came with a vengeance, and left the field never the same again. A number of writers admittedly continued to trace specific institutional defects which in their view facilitated Hitler's advent to power; e.g., a faulty electoral system like proportional representation (PR) (Hermens 1941); or the opportunity for negative majorities to topple a government they could not themselves replace (leading in 1949 to the introduction of the 'motion of constructive non-confidence', for which both James Pollock and Carl J. Friedrich have claimed rival patents, but which had already been broached by Alexander Rustow as early as 1929; see Wildenmann 1963, p. 80); the undesirability of a president with a separate electoral mandate next to a parliamentary system of government; or insufficient legal powers to ban anti-democratic parties as given later in the 1949 Constitution to the West German Bundesverfassungsgericht.
At the same time, however, other writers began to blame older legal doctrines for being at least partly responsible for the inability of democratically elected party governments to rule effectively. Did not the old bureaucracy and the army remain outside democratic control because of an exaggerated belief in the need to preserve their status as (higher) state organs from direct partisan interference? Should not democratic parties have subjected these Wilhelminian legacies to their authority, as shortly afterwards Hitler and his party were to do successfully? This implied that the function of parties had to be re-assessed. It was clearly not sufficient to regard them merely as instruments of representation; they should also be seen as instruments of effective political control. In that light, for parties to be mainly *Wcltanschau-ungsparteien* was clearly not enough, and possibly one reason for their impotence in practice. This revised stand was to exercise a considerable Influence on attempts to arrive at more functional typologies of parties, as best represented in, e.g., Sigmund Neumann (1956b). going back to an analysis Neumann already gave in his Die deutschen Parteien (1932).
The very basis of institutional analysis was to be eroded, however, by the course taken by Nazi Germany, by the grotesque discrepancy between Stalin's 1936 constitution and his actual regime, and by the

rapid decline in constitutional government, first in the Eastern European countries and later in the large majority of new states. Explanations for such developments required different approaches such as, for example, studies on leadership, propaganda, mass behaviour, political culture and longer-term developmental analysis, which sought to explain why some European countries fell to fascism and others not.
4. The development of political science into a truly international discipline
Since the 1930s, political science developed in a few decades into a much more genuinely international discipline. The large-scale migration of European scholars to the hospitable American social science community massively contributed to this development - not least by merging traditions of German 'Sozialwissenschaften'withautochthonous American social science concerns. In a more interdependent world, where many erstwhile political certainties had been brutally destroyed. political science developed rapidly in country after country. The sheer size and richness of American political science exercised a powerful impact, not least on the study of comparative European politics, in which, for at least one generation, Hitler*s exiles represented the predominant influence. The generosity of American foundations in promoting a busy transatlantic traffic greatly assisted the development of a new, international community of scholars. Paradoxically, many European political scientists of the post-1945 generation not only learned much of their political science theories and techniques in America, but also began, later, to discover one another across European frontiers through the intermediary of American scholarly publications and initiatives. Often, such younger European scholars experienced parallel reactions. For all the stimulus they derived from American academia, not least in prompting them towards much new research work on their own countries, they often felt that prevalent theories and typologies did not really fit the experience of these countries. This forced them to attempt to bring their own country on to the initially mainly American-drawn map of comparative politics, using what was 1" fact becoming the lingua franca of the new social science, English.
Inevitably, such efforts often led to what one might term 'single-country extrapolations*. American writings had been hardly free from "^t tendency. Typically, for all its attempt to present a new structural unctional approach, such a new typology of political regimes as

Gabriel Almond published in his famous article 'Comparative Political Systems' (Almond 1956), in which he juxtaposed an 'Anglo-American type* and a 'continental European* type, seemed to many European readers faulted by many a prioristic value judgements. Yet European scholarship had for long shown a very similar trend, and was to do so for decades to come. One example is the traditional British habit of describing European countries as having one common characteristic - proportional representation - hence many parties, hence instable government, hence immobilism, hence ineffectiveness alternating with authoritarian interludes. Although he set out consciously to interpret party politics in more functional terms. Almond's generalizations were probably substantially influenced by such traditional stereotypes. A more functionalist approach to parties in terms of system and political culture was itself a reaction to stereotyped dichotomies in which (British) parliamentarism and (American) presidentialism had been contrasted, usually to the advantage of the first. Only six years separated the publication of Almond's typology from the influential and much-debated APSA report, Towards a More Responsible Two-Party System*, which probably represents the last major example of an older American desire to make their system more "European* (APSA 1950; cf. Ranney 1954).
If one surveys the development of rival typologies of party systems (or political regimes generally) since the 1950s, the reasoning from the background of specific countries remains a marked characteristic. Many of Duverger's generalizations really fit France only. It is not difficult to discover Norwegian roots in Rokkan's Macro-Model of Europe - undoubtedly the finest and most accurate map of Europe so far developed in our profession (for a bibliography of Rokkan's writings see Daalder 1979a; for a first-rate overview of Rokkan's theories, see Flora 1980). In the same way, Italy is the undoubted starting point of Sartori's much refined typology of European party systems (Sartori 1966, 1970, 1976), as Holland is explicitly behind Ujphart's model of consociational democracy, developed in direct response to Almond's original challenge (Ujphart 1968a, 1968b, 1977a).
In the eager race to articulate the experiences of one's own country for the attention of an English-reading political science public, not all were initially equally successful. Larger countries could long remain a world unto themselves, their scholars caring little whether their countries figured as sui generis or archetypes in foreign writings. Conversely, smaller countries had a triple handicap; non-natives wishing to study their experience often met with a linguistic barrier; there

were fewer natives available to articulate their experiences for foreign audiences, and, traditionally, their experience was thought to be less relevant. It is easy to chronicle an attitude of benign neglect towards smaller states on the part of scholars of larger countries: countries in Scandinavia or the Benelux area might be given credit for having some virtues, e.g., for being sober monarchies (Fusilier I960), for having 'working multiparty systems' (Rustow 1956), or being a "mixed type' between the Anglo-American and the continental European system (Almond 1956). Yet their possible importance was widely discounted:
they were thought to offer at most 'derived' experiences (Moore 1966);
they did not carry a real burden in international affairs (a view dating back at least to Hermens (1941) and Friedrich(1941), who had reasoned that small countries could afford proportional representation and its ensuing fragmentation of political will in a way that large countries could not). Essentially, small countries seemed to many to be of folk-loristic interest only.
In recent years this situation has improved. The increased material brought forward by nationals of smaller countries themselves no longer provides the easy alibi for the disdain of yesterday. The flock of American PhD candidates in search of a country has massively assisted the charting of what to large-nation scholars remained for long terra incognita. Paradoxically, the principle of equality in international organizations (whether the UN, its specialized agencies, the OECD or the EEC) has assisted this process further by making sure that data were assembled on all member countries alike. The turn towards quantification in comparative studies - putting a premium on a larger number of cases -has worked in the same direction.
The growing internationalization of the discipline therefore contributes to comparative study in at least two ways. There is, on the one hand, an increasing rapprochement in theoretical concerns, in vocabulary and in questions asked. On the other hand, more and more monographic literature is becoming available, including now most European states. These may be country studies - either of a descriptive nature, e.g., on the model of the classic studies on France by Philip Williams (1964), on Sweden by Rustow (1955) and on the German party system by Wildenmann (1963) and Kaack (1971)-or single-country case studies in the context of a particular theoretical problem, e.g., Eckstein (1966) on Norway, and the series of books forming part of the literature on consociational democracy reviewed by Daalder (1974). These could "so be what one might term descriptive overviews as given earlier in Herman Finer (1932) and Friedrich (1941), and provided again and

again in a number of thorough American textbooks: e.g., Cole (1953);
Beer and Ulam (1958); Macridis and Ward (1963); in the two PEP-sponsored volumes of Henig and Finder (1969) and Henig (1979); in Raschke (1978) and in the recent massive volume combining country and thematic papers edited by Merki (1980). We now also have the useful series sponsored by the American Enterprise Institute, reviewing different European countries *At the Polls'. There are comparative treatises on parties in a particular 'family': e.g., Paterson and Thomas (1977) on the socialists; Irving (1979) on the Christian Democrats;
Mclnnes (1975), Tannahill (1978) and Timmermann (1981) on the corrmiunists; Rogger and Weber (1965) on the Right; Urwin (1980) on agrarian interests; and Seller (1980) on party political families in general. We now have even world-wide overviews of parties, as in Janda (1980) and Day and Degenhardt (1981).
5. Duverger's Les partis politiques (1951)- a battered classic
Any survey of the more systematic and analytical comparative study of European parties and party systems must start with a tribute to Duverger's Les partis politiques (1951). The book starts, as the author says in his preface,
from 3 basic contradiction: it is at the present time impossible to give a valid description of the comparative functioning of political parties; yet it is essential to do so. We Find ourselves in a vicious circle: a general theory of parties will eventually be constructed upon the preliminary work of many profound studies; but these studies cannot be truly profound so long as there exists no general theory of parties.
Hence, he presents
a preliminary general theory of parties, vague, conjectural and of necessity approximate, which may yet serve as a basis and a guide (bi detailed study, (Duverger, 1964 ed., preface, p. xiii)
The book remains a feat, indeed. Drawing on data from numerous countries, it covers party structure (with sections on organization, membership and leadership), as well as party systems (discussing the effect of numbers, strength and alliances, and the relation of parties to political regimes). It shows a remarkable power of synthesis and it offers a wealth of illustrative material.

Yet, from the very beginning the book has rightly come in'for considerable criticism. For all his intellectual grasp and articulate presentation, Duverger clearly lacked real knowledge of the countries on which he drew for alleged proof - with the result that country experts often were left gasping at his evident misinterpretations. The book has been thoroughly faulted on methodological grounds; see, e.g., Engelmann (1957), Leys (1959). Indeed, the book remains remarkably French-centred even when analysing presumed empirical material from countries beyond the French border. This is especially true in his treatment of the relation between institutions and parties - a major emphasis of the book. Thus, Duverger confidently presented laws' on the relationship between specific electoral systems and particular types of party systems, holding that the former basically determined the latter (see also Duverger 1950). His propositions were soon strongly contested by Lavau in a book significantly entitled Partis politiques et realties socviles (Lavau 1953), and effectively reversed by Rokkan, who traced the adoption of particular electoral regimes to the party systems as they had crystallized at the advent of mass democracy, and who saw existing party systems choosing electoral laws that suited them, rather than the electoral systems manufacturing particular party systems (Rokkan 1968a, but see also Sartori's 1966 argument about feeble and strong electoral systems).
A similar institutional bias in Duverger is apparent in his proposition that in two-party systems institutional relations are really absorbed into intra-party political processes, while multi-party systems leave room for intervention by institutional arrangements and reforms. This represents little but an alibi for a desire to tinker with the institutions of the Fourth Republic, as indeed Duverger's laws' on electoral systems had done equally. Duverger was not alone in that respect. I well remember a session on Executive-Legislative Relations, held at an IPSA Congress in Rome in September 1958 under the presidency of Georges Vedel, when Frenchman after Frenchman held forth on the 'right' way to arrange for a correct distribution of power between government and parliament, without anyone ever mentioning General Charles de Gaulle. It is, of course, rather logical that German and French writers tend to ascribe causal effect to particular institutional arrangements - given that a break in the continuity of their systems (not really caused by institutional factors) effectively allowed them to engage in an overhaul of political institutions in 1949 and 1958 respectively.
Apart from a very few factual revisions of his book in the 1950s, Duverger never saw fit to revise, update and improve on his first pre-

sentation of Les partis politiques. His real interest was with French politics, and notably with his preoccupation to do away with the 'eternel marais* of the centre, through introducing some manner of alternative electoral choice for a supreme executive (see Duverger 1974).
Yet Duverger's book remains a classic, however battered. It was 25 years before Sartori's Parties and Party Systems (1976) effectively made DuveFger's work little but a relic - and then only because the new volume was the work of an author who remained aware throughout of the nature of his theoretical quest, and found it essential to invest another decade of his life fully to familiarize himself with the developments in many European countries on which he sought to base his argument. (For an earlier valiant attempt by one author to bring mainly English language evidence from a number of countries together, see Epstein 1967.)
6. Numerical approaches
Analysing European party systems through numerical relations among parties has its roots in two seemingly rather different approaches.
One is the traditional dichotomizing between two-party systems and multi-party systems, in which the first stood mostly for the British system, and the latter for the party systems of the French Third and Fourth Republics, often thrown together with Weimar Germany and post-war Italy. The approach derives a substantial number of logical consequences from the difference of two and more than two: e.g., in relation to the formation of majorities, the chances of voters directly to determine the choice of government, the likelihood of alternation in office, the opportunity for clear decisions and definite electoral accountability, and the degree to which parties must choose a moderate course. Stemming mainly from traditional, highly normative 'cross-Channel' comparisons, the approach was given a considerable boost by the development of formal models (notably Downs 1957, but also Duverger 1951, Sjoblom 1968 and Sartori 1976).
The second approach had its origin in rather different attempts to validate specific propositions with the aid of cross-national data on election results and the compositions of parliaments and cabinets. Such data were relatively easily accessible - not least because of the useful data collections by scholars like Meyriat and Rokkan (1967) and Mackie and Rose (1974) - and lent themselves to statistical treatment

without a need for real knowledge of each national context. Thus, a rapidly growing number of studies could probe the relation between votes and seats in different electoral systems (Rae 1971); the duration of cabinets (e.g., Daalder 1971; Taytor and Herman 1971; Dodd 1976;
Sanders and Herman 1977); the chances of specific parties on office and offices (see Blondel 1968; Browne and Franklin 1973); and, more generally, the validity of different formal coalition theories (e.g., Browne 1970; De Swaan 1973; T^ylor and Laver 1973; Budge and Herman 1978; Franklin and Mackie 1978). Only few scholars have successfully bridged the rather substantial gap between the traditional kind of systemic comparisons of government-parliament relations in Europe, and the often highly skilful but not always politically well founded cross-national statistical analyses (but see von Beyme 1973;
Smith 1972; De Swaan 1973;Pappalardo 1978).
The traditional two-party versus multi-party dichotomy has of course come in for substantial empirical criticism. Thus, queries have been raised on the applicability of the two-party model in Europe (where at most a dubious case for a two-party Austria could be juxtaposed to the increasingly doubtful case of a two-party Britain). Doubts have been expressed on whether manufactured majorities in seats may be treated on a par with actual majorities among voters. The presumed logic of both alternation and moderation have come increasingly under fire (e.g.. Finer 1975, 1980; also see the treatment in Sartori 1976, chs 6 and 9).
The credibility of the multi-party category, on the other hand, has been also effectively undermined. Early on K.C. Wheare had put forward the view that France was the exception on the continent rather than the rule (cf. Wheare 1963) and Dankwart Rustow had analysed Scandinavian systems under the label 'working multi-party system' (Rustow 1956). The Dahl Oppositions volume (1966) added a considerable amount of new material, as did the elegant attempt by its editor both to offer empirical generalizations and to spell out more qualitative criteria for understanding the working of competitive systems (Dahl 1966, chs 12 and 13). UJphart's successful coining of the consociational democracy model powerfully affected the debate (e.g., Ujphart 1968b, 1969, 1977a; see also the collection of writings in McRae 1974 and the review by Daalder 1974). Since the 1960s the discussions came to be dominated effectively by Sartori's attempt to distinguish between moderate and polarized pluralism - a discussion much refined when published in more definite form in Parties and Party Systems (1976), and further pursued by Sani and Sartori in chapter 11 below.

As a qualitative re-assessment of the traditional dichotomy between two-party versus multi-party systems continued, a variety of counting schemes sought to grapple with the rather varied numerical relations actually existing in European party systems. Thus Wildenmann introduced the category of a two -and-a-half-party system, which was taken up by Jean Blondel in a general typology. In constructing that typology Blondel also made a distinction as to whether one party was dominant or not (see Blondel 1968), later changed into an accounting scheme of party systems in terms of the presence or not of large, casu quo small parties in particular political families (see Blondel 1978, appendix I). Ujphart, in a debate with Sartori, had added together parties in order of declining size (Ujphart 1968b; cf. also Sartori 1976, ch. 9), to see whether particular systems had indeed a high degree of fragmentation. Before this Rokkan had sought to classify party systems in Europe according to the distance which in the inter-war and post-war periods respectively separated the largest party from the majority point, the second party from the largest, the third party from the second, and so on (cf. Rokkan 1968b; later included in Rokkan 1970a, ch. 3). Soon afterwards the discussion came to be dominated by the Rae fractional-ization index -a measure that was widely used (e.g., Rae 1971;Dodd 1976; Wolinetz 1979), and also increasingly contested (note the highly effective onslaught in Sartori 1976, ch. 9, and Pedersen 1980). These various attempts have undoubtedly increased our awareness of the very different numerical relations existing in different European countries, as well as of the changes that have occurred in them. But their value is lessened by their limited analytical bearing and a clear lack of the translation of findings into actual political analysis.
Compared to the earlier numerical classifications, the attempt to test formal coalition theories against data on the composition of European cabinets represents a much more ambitious theoretical exercise. In a short general overview one can hardly do justice to the many theoretical ramifications of that approach (which is most completely reviewed in De Swaan 1973). The various theories generally make a series of (limiting) assumptions: e.g., parties seek office; they are unitary actors; they enter coalition bargaining on the basis of the proportion of parliamentary seats they command; for a coalition a majority of the seats is required. Tests soon revealed that theories basing themselves on numbers alone (e.g.. Piker's 'minimum size* theory) performed poorly, and that successful predictions could be made only if information was available for the location of parties along one or more ideological dimensions. But then, the problem arose

how to obtain these dimensions - in a manner that theoretically should remain free from any knowledge of actual coalition behaviour. Often solutions were proposed that were hardly free from a prioristic judgement. Relying as they did on a placement by authoritative observers who did know the (past) coalition behaviour of parties, the practice of assigning parties a place on a particular dimension led, in fact, to rather tautological results. Later authors have attempted to turn this defect into an advantage by introducing considerations of 'inertia* and 'familiarity' (Franklin and Mackie 1978), or speaking of 'normal* governing parties (Budge and Herman 1978) - thus explaining new coalitions precisely by the predisposition of parties to continue coalitions found practicable before. At the same time, some of the assumptions mentioned earlier have been set against deviant political practice, e.g. where parties are not unitary actors, or where governments form without an assured parliamentary majority (see Daalder 1971'.Herman and Pope 1973; Smith 1978).
Compared with these various approaches, the work of Sartori (1976) stands out in its theoretical and political relevance. Sartori, too, sets out from numbers, holding that the interaction of given numbers of parties provides indeed a major independent variable, and therefore does justice to a truthful political sociology rather than a reductionist sociology of politics (Sartori 1968, 1969). But he adds his well-known counting rules to see whether parties are relevant or not; takes account of definite ideological divides, e.g., those separating system and anti-system parties and of ideology more generally, which to him (in an argument that reminds one of Hermens as much as of Downs) is to a considerable degree also a logical consequence of an increasing number of parties (his view of 'ideological stretching'). Of vital importance is his insistence on the difference between classification and type, and of format and mechanics, and his rejection exactly for this reason of continuous variables in favour of discontinuous classes. His work is too well known to need further treatment in this overview, but readers might turn to chapter 11 below, which, on the one hand, provides new empirical data on the degree of polarization in different European party systems, and on the other hand grapples with the problem - most clearly posed in the literature on consociational democracy - of multidimensional identifications. Sani and Sartori seek to reduce a multidimensional world of party identifications to a uni-dimensional universe of electoral competition.

7. Parties and social cleavages
Uke the approach through numbers, analysis of party systems through the cleavage structures they are thought to represent has an old lineage in political science. To Hume, *parties of interest' were much more normal than 'the most extraordinary and unaccountable phenomenon that has yet appeared in human affairs': 'parties of principle". Traditional analyses described party conflict in terms of specific cleavage divisions as a matter of course; e.g., town versus country. Church versus anti-clericals, one estate against another; and later of classes which were thought to be inevitably in conflict with one another.
More specific research efforts also sought to spell out the relation of parties to particular cleavage divisions from an early moment, as in inventories of the overall hold of religious parties, and the actual or expected strength of radical or socialist parties as the suffrage expanded. Such approaches were given a more precise tool in ecological analysis as pioneered by Andre Siegfried (1913). His Tableau politique de la France de t'Ouest not only set a style for a 'geographic electorale*. in which French scholars were to, and still do, excel; it also underlined the specific regional setting in which political conflicts should be interpreted, and by implication pointed to regional diversities as one possible source of conflict alignment. Siegfried's example was followed and extended in the great work of Rudolf Heberle (1963) on the national-socialist vote in Schleswig-Holstein, and on the role of cleavages generally. Since 1945 ecological analyses have been widely used to probe the persistence of a left' and a 'right', independent of specific partisan expressions in French politics (e.g., Goguel 1951), in studying the politicization of cleavages in successive waves of mass democratization in Norway (Rokkan and Valen 1962, 1964), in tracing the background of communist votes in Finland (AUardt 1964), and in cross-national inquiry of the degree of worker support for left-wing parties (Dogan 1967; see also Dogan and Rokkan 1969). It remains the only tool to analyse historical elections with any degree of precision.
But the advent of survey research also contributed to the popularity of explaining partisan divisions in Europe through social cleavages, for at least three reasons. The first American election surveys strongly emphasized group associations, and they were soon replicated in Europe; group cues such as class, religion, age and sex found their way almost automatically in the publication of survey research, without as yet undergoing clear statistical tests for significance or multicorrelinear-ity; and even in the application of more psychologically oriented

approaches in survey research that spread from Ann Arbor, Michigan, European scholars found it necessary to insist on the rather different structural alternatives that European voters faced in their party systems compared with American voters. Thus, Rokkan and Valen deliberately introduced ecological controls in their survey designs, and in fact began their long historical quest on how persisting cleavage divisions had first gained political importance, precisely in order to explain the very different patterns of electoral preference they found in contemporary surveys.
What was true for within-country variations also applied to cross-national analyses. A clear awareness of the rather different salience of particular cleavages in European party systems (e.g., in the degree of class voting, on divisions within bourgeois politics, and the very different role of religion in politics) led Rokkan to the famous introductory chapter in the Upset-Rokkan volume on Party Systems and Voter Alignments (1967b). In this seminal work Rokkan accounted for the differences in party systems by concentrating on the manner in which particular cleavages had become politicized during the advent of mass democracy. In that same chapter, he also put forward the famous freezing proposition that 'the party systems of the 1960s reflect with few but significant exceptions the cleavage structures of the 1920s' ^ (Upset and Rokkan 1967b, p. 50). The statement was soon to be ^, further empirically validated by Rose and Urwin in two studies of the ^ persistence of party systems in the inter-war and post-war periods -\ respectively (Urwin and Rose 1970; Rose and Urwin 1970); while the 3 extent to which that proposition no longer holds true in the light of the experience of the 1970s is the subject of chapter 3 below by Maria Maguire.
We have now a large volume of cross-national literature on the significance of specific cleavages for European party systems and voting. Already in the 1950s Upset had collected comparative evidence in his Political Man (1960). In a well-known article he sought to analyse to what degree a changing class structure was likely to influence the future of the Left in Europe (Upset 196 7)-a theme for which Stefano Bartolini presents new and more complete data in chapter 6 below. Alford (1963) also stimulated a host of studies on the extent to which parties represented what he termed 'class polarization*. In another seminal article, Rose and Urwin (1969) analysed the electoral fate of socially homogeneous and heterogeneous parties, and they also wrote a separate paper on the impact of regional factors (Rose and Urwin concern with the extent to which specific cleavages translate in partisan preferences in different countries is the core of a large-scale comparative project directed by Rose and published as Electoral Behavior: A Comparative Handbook (Rose 19?4a), while the question of the relative salience of different cleavages (e.g., religion, class and, later, language) is a persistent preoccupation in the work of Arend Ujphart (see Ujphart 1971, 1980). Important work in the field has been done also by the German sociologist Birgitte Nedelmann (1970,1975, 1979).
The intensive preoccupation with social cleavages in the comparative study of party systems has come in for wide-scale criticism, as representing a crude sociological reductionism (cf. Sartori 1968). The point is often substantiated by more modern methods of electoral analysis. New sophisticated multivariate analysis techniques have generally undermined the credibility of simple determinist models of voting 'choice, and often also document a decline in the degree of variance that can be explained through social characteristics in more recent periods (for the most elaborate comparative data see Rose 1974a; Budge and Farlie 1977).
The point is of interest, not only for a careful understanding of the relation between parties and social cleavages in modern times, but also in the light of normative democratic concerns. Can one effectively speak of democratic accountability if parties are little but the automatic expression of forgone choices on the basis of cleavages not necessarily relevant for contemporary politics? The point is one criticism levelled against the consociational democracy literature, which must logically assume that elections are more of a census than the expression of current opinions. Are parties in an ideal-type consociational democracy anything more than routine sentinels guarding opposed camps? Could one not say that Ujphart,for instance,is typicallymore interested in cleavages and in elites exempt from electoral accountability than in parties as independent actors or in the actual impact of elections on political decision-making?
And yet, the criticism is not really fully persuasive. Even if one regards with Rokkan the party systems of the 1960s as mainly a reflection of cleavage structures of another era, this does not gainsay their evident ability to handle a host of new issues and changes, including depression, war, a changing demography, economic growth, increased geographic mobility, the education explosion, the modern communications revolutions and so on. This very ability of 'soi-disant' frozen party systems to handle changes and challenges makes Sartori absolve Rokkan from his charge of sociological reductionism (see Sartori

1968 and Mair, chapter 14 below). Even when parties do reflect older cleavages, this does not protect them from changes occurring in existing cleavage patterns (e.g., in class alignments,religious observance,increased geographic mobility, upsetting regional alignments). Moreover, even in systems with a strong effect of historical cleavage divisions, these exist not in isolation but in interaction and possible rivalry with one another (e.g., of class versus religion). It was one great merit of Rokkan and Valen's cleavages analyses that they remained alert to the possible resuscitation of older cleavage alignments which superflcially seemed to have been superseded by new ones, as indeed they lived to see happen in the Norwegian EEC referendum of 1972, which shattered the apparent stability of the Norwegian party system (see also Converse and Valen 1971; Valen 1976). A great deal of modem analysis is couched again in cleavage terms, e.g.,in the work oflnglehart(1977)and of the authors in the Bames and Kaase volume (1979) - authors who see the rise of new cleavage divisions opposing older alignments in a new post-industrial world. Parties indeed find themselves constantly faced with the necessity to make strategic choices between older and newer cleavages. They engage in rival attempts to politicize those that iare most favourable to them and seek to depoliticize those that are not. In this context they are neither passive agents nor free actors. Some of the resulting problems are traced in chapters 10 and 13 below by Budge and Farlie, and by Sjoblom respectively.
8. Spatial approaches
At this point the cleavages literature paradoxically meets with the rather different paradigm of 'economic' approaches. In a flourishing literature, basically stemming from Schumpeter (1942) and Downs (1957), parties are portrayed as unfettered entrepreneurs, rationally calculating their strategic chances of electoral success by embracing programmes that should win them a maximum number of voters. In such views parties, clearly, roust not be tied down to any bond. Since Downs's pioneering study authors in this tradition have generally reasoned in spatial terms - locating voters and parties in a presumed common space in which both parties and voters move to obtain the maximum possible success. The literature is highly complex and often very abstract, and hence neither easily assessed nor easily applied in practice (see however the overviews in Sjobloro 1968; Barry 1970;
Budge, Crewe and Farlie 1976; and particularly Budge and Farlie 1977).

Apart from the construction of formal models, numerous authors have attempted to derive actual political spaces of European party systems. For this they have used a great variety of methods - surveys in which respondents were asked to place themselves and/or all parties on a number of imposed scales, factor analyses and scaling methods to probe dimensions underlying survey responses to attitude and issue questions, and actual behavioural data such as voter change or successive parliamentary votes. Such studies have given us a large variety of spatial representations of European party systems - rarely of a simple uni-dimensional nature, more often rather complex multi-dimensional ones (for the relevant literature see Budge and Farlie 1977).
These spatial representations leave us with a number of questions. To what extent are spaces constant over time (a point earlier indicated as vital for the empirical testing of coalition theories against data on cabinet-making in European multi-party systems)? Do different actors - e.g., all relevant parties in a given system and all voters - interpret spaces in a similar manner? Do spaces actually represent some kind of past perceptual order or alignments that are effectively salient at any given moment or specific policy preferences? To what degree do spaces effectively constrain political actors - whether parties or voters? One generally finds the assumption that one or at most a few dimensions are really the dominant and hence meaningful ones. This argument is strengthened empirically by the finding that many voters tend to stereotype their views on party stands - an attitude regarded as rational by Downs and others who see in ideology a means to cut information costs and avoid uncertainty. In an extreme version, the rational actions of both voters and parties could be so constrained by the space in which they must operate that little is left of the freedom for all on which the approach was initially based. This would paradoxically expose the spatial approach to the same criticism as that levelled against the cleavages approach: namely, that it is static and deterministic. And in as far as (traditional) cleavages do determine the observed spaces, the difference between the two approaches also logically washes out. But then, the same questions remain: how far can parties manipulate dimensions most favourable to them, and to what degree do new issues lead to new cleavage lines?
At this point the reader should turn to the impressive overview and sophisticated analyses, presented by lan Budge and Dennis Farlie (1977) in their Voting and Party Competition-^ far the most exhaustive review of the problems and promises of this literature. Further contributions to this literature can be found in the chapters in

this volume by Sani and Sartori; Budge and Farlie; and Sjoblom. One element worth noting, however, is that, for all the trappings of formal mathematical approaches, there is now also an increasing turn towards actual political data: e.g., analysis of issue preferences and perceptions, the comparative study of party programmes and their development over time (in a large project directed by lan Budge and David Robertson), and a renewed attention to voting of parties in parliament (for early important examples, see MacRae 1967;Pedersen 1967a;Pedersenetal. 1971; Damgaard 1973; for a good general evaluation of the dimensional approach, see Pesonen 1973).
9. Dropping the assumption of party as a unitary actor
In most formal approaches the party is retained as a unitary actor. This is acceptable in a theoretical model. It is also politically relevant whenever a party does act as one actor; e.g., when it presents an election programme, or decides to enter a cabinet on the basis of a decision that commits the entire party. (One might note that Riker and others see a necessity for a broader than minimum size coalition, whenever there is uncertainty about full party support.) However, in the actual world of politics, it is hardly defensible to regard party as a unitary actor - a point recognized, e.g., by Robertson (1976), who takes functional differences in particular leader groups into account in constructing his model of party competition. Even in the examples just given - presenting an election programme or deciding on a coalition - there is bound to be disagreement before the actual decision is taken, as well as on its later application in practice. In the application of election programmes there are likely to be disagreements on priorities and problems of translation into concrete action. And a decision on the investiture of a new coalition cabinet does not pre-empt a need for continuous decision-making on concrete decisions to follow. Did not Duverger suggest a definite tendency in French politics for government coalitions to slide to the right during a parliamentary period, with the point of gravitation returning towards the left when a new election is in sight (Duverger 1964 ed., p. 330ff.)?
Clearly, then, the study of both party and party systems must enter "ito the internal structural and policy-making processes of parties. It is easy to single out major problems that have been or must be tackled in the literature.
One problem - relevant in many multi-party systems - is the

degree of congruence or (functional) disagreement between the cabinet coalition, as represented by ministers of different parties, and the parliamentary coalition, which may not always follow a cabinet in specific actions or retain the same parties in any particular decision. Countries such as Italy, Denmark or Finland provide a particularly fruitful field of study along these lines. Epstein (1967) has suggested that parliamentary government of necessity engenders cohesive parliamentary parties, but we have too few studies to substantiate that
proposition.
Secondly, there is the major problem of the party in parliament and that outside it (cf. Wildenmann, Partei und Fraktion 1954). The once confident finding of Robert McKenzie (1955) that parliamentary government forces a concentration of power in the hands of parliamentary leaders, whatever internal party rules exist, seems to be increasingly falsified in contemporary Britain (e.g. Finer 1980). General impressions of the more ardent ideological orientations of party activists as compared with parliamentary leaders on the one side and non-active party members and party voters on the other (as formalized in May's 'Special Law of Curvilinear Disparity1 - May 1973) must be validated in further research, while the strains following from the different functions that parties must of necessity combine has led to the proposition that one could analyse parties in alternative terms of a 'party-government' and a 'party-democracy' model. The relation between ministers, members of parliament and parties outside parliament is treated in this volume in a general comparative overview by von Beyme (chapter 12), whose own book on government-parliament relations (von Beyme 1973) is still the most complete comparative study of European countries - diachronically and synchronically - in terms of the interaction between institutions and party cooperation and conflict that anyone has produced since 1945.
A third Held of essential importance is the nomination processes within parties - a subject on which we have far too few studies, and Which in fact merges into the study of elite recruitment, on the one hand, and legislative behaviour research on the other (but see Valen andKatz 1964;Ranney 1965).
One considerable drawback of the absence of detailed, systematic studies of decision-making within parties is that it leaves the field free for sweeping generalizations pointing to secular changes that parties are said to have undergone. Some of tlie older propositions, such as a possible trend towards 'parties of integration' (Neumann 1932, 1956b) or Duverger's forecasts of a shift towards mass parties through

a contagion from the left where parties originated from outside parliament, do not seem to fare so well in the light of actual developments since 1945. Otto Kirchheimer's view about the transformation of European parties into 'catch-all' parties (Kirchheimer 1966b) remains a seductive one, but it has met with a variety of attempts at empirical falsification (e.g., Wolinetz 1979; Zuckerman and Lichbach 1977), and poses large-scale methodological problems when one attempts to test it effectively (see Dittrich's chapter 9 below). Does one witness a trend towards catch-all parties, or rather a resort to catch-all propaganda strategies as a supplement to attempts to retain ties to traditional clienteles? Are we really seeing the Waning of Oppositions (Kirchheimer 1957) and me End of Ideology, or must we rather speak of a revival of new ideological concerns (e.g., Inglehart 1977; Barnes, Kaase et al. 1979)?
A germane subject is the one of the presumed professionalization of party politics. Under the impact of such diverse forces as the natural bureaucratization of large organizations, the inevitable entry of professional public relations specialists in a time of modern mass media politics, and the degree of state financial support for a range of party political activities in a number of European countries, the days of the ordinary membership party are proclaimed to be over. Is that presumed trend gainsaid - or actually confirmed - by the equal signs towards a widespread revival of a new amateur politics? How much do we really know of trends over time in the membership of different parties in different countries, a subject most readily studied in the case of the socialist parties, (where Stefano Bartolini now presents a conclusive overview in chapter 7 below), but which needs supplementary' evidence also for other parties? And even if we have adequate over-time membership figures, wliat is the actual meaning of membership in real decision-making? Important new insights are- coming from the Middle-level Elite project directed from Mannheim by Karlheinz Reif-a large-scale comparative study of delegates to party congresses. Yet even this material will be only one possible beginning for a systematic analysis of decision-making within parties, which is indeed the major gap in our study of parties. We *speak' Michels; we rarely seek to emulate him. We have as yet little else to go on in the elementary analysis of party structure other than the categories - now 30 years old - of Duverger. This leaves the field free for generalities, as exemplified by Epstein's defence of the use and future of *weak parties' (Epstein 967), and Kirchheimer's nostalgia for parties that would not be catch-parties - let alone the bland indictments of anti-party commentators,

uttering uncritical complaints about the Parteienstaat or Partitocrazia, and hurling accusations, as general as they are unsubstantiated, against party bureaucrats and party activists alike.
10. Parties in government
At this point, the discussion merges into the more general subject of the role of parties in government. The subject is a momentous one. It is also a subject that is riddled with impressionistic and a prioristic judgements, rather than one for which we have clear empirical evidence or articulate systematic evaluations.
One major point at issue is whether parties have increased their hold on government over the post-war decades or are suffering from what is termed *a loss in functional relevance*. Proponents of the first view point to a number of developments. On the whole, European party systems have succeeded in holding on to their voters. Data on turnout in elections (see the extensive discussion by Dittrich and Johansen in chapter 4 below) hardly prove a universal, secular decline - let alone one that is easily reduceable to an overall sense of alienation from party systems. As Pedersen shows in chapter 2 below, there are considerable differences in the degree to which European party systems have become destabilized, rather than one definite trend over time. Vital processes in democratic politics, like nominating candidates for elections and controlling recruitment to cabinet office, are generally no less party-monopolized than before. In many European countries, parties seem to loom larger also in below-cabinet appointments, including those to senior bureaucratic positions. One sees few signs that party patronage is in decline. A world of expanded government has created a new realm of very different 'available' institutions and agencies, known in Italy by the telling term of 'sottogoverno*. Appointment powers are a major reason why parties insist on a degree of proportionality in government, and inveigh against systems where apparently perennial governing parties cannot be effectively dislodged through normal electoral channels. (One might add. however, that parties long excluded from national government may very well have rather different opportunities in lower-level governments, which also show little sign of becoming less party-controlled.)
However, for those who see a waning of the powers of parties, such arguments are not conclusive. They interpret individual-level electoral volatility as a sign of increasing disenchantment on the part of the

voters. They point to phenomena of ^elf-starting* candidates and a
more or less institutionalized factionalism to argue that one should not use blanket terms like 'party* if one is to understand the real world of politics behind that label. Again, they insist that party may be a channel through which people must move in their march to ministerial or appointive government office - but that, once these people arrive in office, they assume functional roles which are substantially free from party control (possibly giving them even leverage to reverse the process and to increase their hold on the party in turn),
Those who assume a secular waning of parties also point to the increased weight of alternative channels of decision-making. One factor is the logical consequence of the need for parties in an age of modern television to personalize their electoral appeal - which must strengthen the hands of a few party leaders against the membership base of a party. This element is clearly strongest in systems where voters decide directly on a chief executive or at least a premier, but is far from absent in systems in which parties cannot hope for a direct electoral mandate. The recourse to referenda can be another factor of taking authority away from parties and possibly also their leaders (e.g.. Butler and Ranney 1978), as indeed might be the use of constitutional courts to settle issues on which the parties are unwilling or even unable to take a decision. In a similar manner, scholars have regarded the rise of so-called action groups or *Burgerinitiative* as an unmistakable sign of alienation from existing partisan politics and of a wish of the 'people* to have direct control over decisions affecting them. (In this interpretation a number of obvious counter-arguments are not considered: e.g., the existence of similar groups throughout the history of mass democratization; the rather favourable cost-benefit ratio, which a short cut through direct action by a relatively small and relatively elite group may show in an era of television and tired, overloaded government; and the empirical finding that the new activists often are at the same time disproportionately active within both political parties and well entrenched institutionalized groups. For thorough empirical analyses, see however Verba et al. 1978; Bames, Kaase etal.1979).
Explanations of the presumed functional loss of party systems through increased direct pressure by voters or action groups on the centres of authoritative power would at first sight seem to be rather ar ^moved from the neo-corporatist proposition. In a few well-known articles, Rokkan argued in terms of a clear dichotomy between 'numerical democracy* and 'corporate decision-making* (Rokkan 1966),else-

where summed up in the short phrase "votes count, resources decide' (Rokkan 19756). The argument is one of a substantial displacement of political decision-making from the 'electoral partisan* channel to the corporate network, in which specialized bureaucrats jointly with related and recognized interest groups determine most of the vital political decisions (see Schmitter 1974; Schmitter and Lehmbruch 1979). While even corporate groups may engage in direct action tactics, both developments are seen as eroding the very hold of parties on decision-making processes in modern society. (For an attempt to construct a new typology of European party systems on the basis of the cohesiveness of the political system on the one hand and society at large on the other, see also Smith 1979.)
From there the argument logically shades over into the controversial issue of whether parties do have an unmistakable influence on government policies - or whether policies evolve rather as the inevitable response of any government (whatever its party orientation) to the exigencies of modern social and economic development. The debate first developed in analyses of the impact of partisan politics on policy outputs in American states (e .g.. Dye 1966). The discussion has long since spilled over into comparative and European politics (e .g., Rimlinger 1971; Hecio 1974; Heidenheimer et al. 1975;Wilensky 1976; Stephens 1979; Castles 1979, for an excellent review see Maguire 1980). The debate and research still very much go on, not least because one deals with a variety of policy areas with diverse techniques of analysis. The major conclusion emerging, however, seems to be that parties do indeed
count.
It is one thing to prove a specific relation between the control of government by a particular party and particular policy outputs. To move on from there towards the much more general problem of the relative performance of different governments and party systems represents yet another gigantic step. From chapter 5 below, in which Rose and Mackie deal with the problem of whether incumbency pays electorally or not, it is clear that voters show little regularity in rewarding or punishing governments in power. How should outside analysts judge the performance of specific governments, or of the party system as a whole? What indicators must we use to make qualitative assessments? Can we make effective use of the survey data on relative citizen satisfaction in the regular Eurobarometer project? What additional behavioural data can we find on the degree of legitimacy accorded to different systems? Comparable cross-national social indicators of a more objective kind? How can we resolve the problem of collecting and

analysing reliable comparative data, and account for the possible interaction of variables at issue? Perhaps we should have more studies along the line of chapter 13 below, in which Sjoblom presents a number of coherent propositions on the relation between political changes and political accountability. ,
In assessing the working of party systems in a comparative perspective, not the least problem is the lack of agreed criteria for measurement. Years ago, Anthony King (1969) rightly argued that for any real judgement we must begin to get a clear opinion about the specific functions parties are supposed to fulfil. To assess the performance of parties in any one of them is already difficult; to disentangle each function from forces working elsewhere in the political system, and the socioeconomic system generally, represents an almost insurmountable task. The literature is moreover shot through with a priori assumptions about what a 'good' party system should do. Ensure stable government or regular alternation? Do justice to an electoral mandate, or permit flexible, continuous government? Think in terms of undivided responsibility, or the need to ensure consensus and compromises?
In evaluating systems, we are likely to meet with at least three rather different normative strands: the traditional political formulae in any given system, which are strongly influenced by existing institutions and experiences in their working; presumed functional relationships in the context of a structural-functional theory of politics; and general democratic values, which are rather more the special province of political theory. Perhaps, then, we have not progressed so far after all from the concerns of the traditional writers of representative government - but that at a time when social science methodology feeds a not always easy or even warranted sense that we can and should now rigorously specify our theories and variables.

• [ На главную ]
• [ Библиотека ]
• [ Reader ]
• [ Программы ]
• [ Абитуриенту ]
• [ Фотоальбом ]
• [ О ФП ]
• [ Гостевая книга ]
• [ Ссылки ]
• [ Обратная связь ]