INTRODUCTION

The very re-appearance of political science in Italy is in itself a major and highly deserved tribute to Giovanni Sartori's¹ intellectual and organisational activities. To cut a long story short, Fascism destroyed what had existed in terms of Italian political science. The founders, Gaetano Mosca, Vilfredo Pareto and, though German by birth, Roberto Michels, had not, after all, created a 'school'. After 1945, there were no heirs to the scholars who had launched the theory of the ruling class, elaborated the analysis of elites, and formulated the iron law of oligarchy. There was only one department of Political Sciences (in the plural), the already famous 'Cesare Alfieri' of Florence, but not a single chair in Political Science.

Sartori was obliged to start his career studying and teaching political philosophy and writing on Hegel and Marx, Croce and Kant. In retrospect, this experience proved to be very useful because it gave him the masterful capability in logic and the analysis of concepts that he has put to such very good use. Although already a member of the international circuit of political scientists, the non-existence of political science as a discipline in Italian universities meant that he first had to become full professor of Sociology. By 1966, following a reform of the departments of Political Sciences that he himself had forcefully advocated, he was able to move to Political Science. From 1969 to 1972 he served as Dean of the department in Florence. In addition to having been instrumental in promoting reform of the political science curriculum, he founded and directed the Centro Studi Politica Comparata, from which have come all second-, and most third-generation Italian political scientists. Then, in 1971, Sartori founded the Rivista Italiana di Scienza Politica. By that time, he had already published widely in Italian and in English and was well established as a top political scientist. Though having also received an offer from Oxford, he moved to Stanford in 1976 and in 1979 became Albert Schweitzer Professor in the Humanities at Columbia University, where he is now Professor Emeritus. Sartori, born in May 1924, has just turned eighty, but he has lost none of his energy. He divides his time between New York and Italy and is a frequent, active and provocative participant in debates on Italian current affairs. His knowledge of political science continues to serve him (and many of us in Italy) well.

As a scholar, Sartori is not the author of just one book, written and re-written. Nor has he confined himself to just one topic. He has left important and long-lasting marks in three major fields of political science: the theory of democracy; party systems; and constitutional engineering. Above all, Sartori is a truly comparative political scientist. He has always thought, and has tried to prove, that progress in research goes hand in hand with progress in theory – and, consequently, that if there is no progress in theory, progress in research is highly unlikely. Finally, to Sartori, progress is not just measurement and quantification. As he has written: 'Words alone beat numbers alone. Words with numbers beat words alone. And numbers make sense, or much greater sense, within verbal theory' (Sartori, 1976: 319). For what it is worth, I could not agree more.

DEMOCRACY

Sartori's first major book was Democrazia e definizioni. First published in 1957, it went through several editions and Sartori

'He divides his time between New York and Italy and is a frequent, active and provocative participant in debates on Italian current affairs.'

himself translated and partly reformulated it for the American edition: Democratic Theory (Sartori, 1962, 1965). He never stopped writing about democracy. The synthesis of his ideas can be found in his two-volume work, The Theory of Democracy Revisited, part one of which covers 'The Contemporary Debate', and part two 'The Classical Issues' (Sartori, 1987a). A condensed, largely reorganised version with a large appendix discussing the state of democracy after the fall of the Berlin wall was later published in Italian as Democrazia. Cosa è (Sartori, 1993).

What is it that makes Sartori's many writings about democracy so important and lasting? First of all, he is never parochial. His books are about the ideas and the theories of democracy as they have been formulated and revised through time by all authors of significance. Secondly, he analyses, assesses, and criticises all the contributions that impinge upon his subject, and does so with a tremendous amount of knowledge. By discussing, criticising, and, whenever appropriate, praising their interpretations, he takes seriously all the authors who have contributed to our knowledge of democracy. For a taste of Sartori's keen attention to developments in the field, I would suggest, in particular, his splendid chapter on 'Vertical democracy' where the arguments of the elitists and the anti-elitists are thoroughly revisited (Sartori, 1987a). Finally, Sartori also developed his own theory of democracy, one that is intelligently constructed around a combination of Schumpeter's competitive theory and Friedrich's anticipated reactions. As is well known, Schumpeter has often been accused of advocating a situation in which teams of leaders compete for the government of the country and in which, after the victory of one team, there is no space for voters or voter participation. However, in his 'feedback theory of democracy' (Sartori, 1987a: 152), which obviously fell within competitive theory, Sartori suggested that neither the winning nor the losing teams would stop paying attention to the interests and preferences of the voters. To quote directly, 'Large-scale democracy is a procedure and/or a mechanism that (a) generates an open polyarchy whose competition on the electoral market (b) attributes power to the people and (c) specifically enforces the responsiveness of the leaders to the led' (Sartori, 1987a: 156, italics in original). The latter concept refers to what is now the very fashionable analysis of accountability that is better qualified in the sentence: 'representation intrinsically consists of two ingredients: responsiveness and independent responsibility' (Sartori, 1987a: 170).

The Theory of Democracy Revisited is remarkable for several other reasons. It is simultaneously a book about the different definitions of democracy provided through time – a book that therefore offers a splendid introduction to, and analysis of, the history of democratic ideas – and also a contribution to the construction and analysis of political concepts, the clarification of which has always been paramount in Sartori's work. A valiant attempt at this can be found in his edited collection, Social Science Concepts. A Systematic Analysis (Sartori, 1984). Sartori's work on classification has been his least successful undertaking, and there is no denying that the discipline moved on and continues to move in other, more confusing and confused, directions. The 'Tower of Babel', as Sartori has defined it, still stands, as too many political scientists prove unable and/or unwilling to be precise and consistent in their terminological and conceptual analyses. In any case, Sartori's unceasing interest, accompanied by his willingness to provide sound foundations for the study of politics, culminated in a small classic, really a concise encyclopaedia: Elementi di teoria politica (Sartori, 1987b, 1990, 1995).

While in The Theory of Democracy Revisited, Sartori aims at creating what he calls the theory of 'mainstream political democracy', his presentation and criticism of all the other theories and theses remains outstanding even today. Indeed, with surprising foresight, he effectively identified and, even more importantly, fundamentally clarified almost all the contemporary issues, such as electoral democracy, participatory democracy, referendum democracy, and even the construction of democracy and democratisation. Finally, the book is not only about democracy. It is about politics, and political concepts – liberty, equality, ideology – and the political alternatives to democracy: autocracy, authoritarianism, totalitarianism. It is a summa of what is known and should be known, that can still guide, cum grano salis, empirical research.

There are two additional, particularly important, aspects of Sartori's theory of democracy. The first is a permanent lesson for all scholars and practitioners dealing with existing or 'real' democracies. No analysis of democracy can be complete without drawing upon both descriptive and normative views of democracy and then combining them – as happens in practice. Often accused of being 'just a realist' – a label he accepts provided it is accompanied by the adjective 'cognitive' – Sartori does not refrain from stating his preference for a specific type of democracy: 'democracy should be a [selective] polyarchy of merit' (Sartori, 1987a: 169, italics in original). There is no need to stress how many theoretical and empirical implications follow from this definition, many of them dealt with and clarified by Sartori himself. The second, oft-forgotten, aspect has to do with the role of ideas and with democratic mentalities. From this point of view, his 1969 article, 'Politics, Ideology, and Belief Systems' (Sartori, 1969b), is especially important. His argument is rather easy and straightforward: ideological mentalities are less conducive to democratic behaviour; pragmatic mentalities are supportive of democratic behaviour. If one wants to understand why liberal democracies have defeated the challenge of authoritarian and totalitarian regimes and why religious fundamentalism has emerged as the only alternative to liberal democracies, then Sartori's book provides the indispensable analytical tools.

PARTY SYSTEMS

No scholar interested in the analysis of contemporary democracies can afford to neglect or minimise the role of political parties. When Sartori first approached the subject of political parties, he had already written on representation and on the role of parliamentarians. The stimulus to deal specifically with political parties came, as he indicated, from his dissatisfaction with Duverger's (1954) Political Parties, at the time by far the best comparative study of contemporary political parties. Three long essays (Sartori, 1966, 1968, 1970a) stand out in Sartori's production of that period and anticipate the extremely important and, in my opinion, unsurpassed contributions further elaborated in his major book, Parties and Party Systems. A Framework for Analysis (Sartori, 1976). This book was written as the first of a two-volume study, but the draft of the second volume was stolen and never recovered by the author. That volume was meant to deal with the structural–organisational analysis of political parties. The draft of one chapter of that work was recently

'...Sartori does not refrain from stating his preference for a specific type of democracy: 'democracy should be a [selective] polyarchy of merit'.'

found, and is to be published in West European Politics.

Until Sartori's book, political scientists had treated parties only in terms of their number (and in a small number of unconvincing contributions, some continued to do so even afterwards). As a consequence, only three types of party system were identified: one-party systems; two-party systems, and multiparty systems. Several distortions followed because the existence of more than two parties did not necessarily amount to a multiparty system; systems in which only two parties existed could not always be considered two-party systems, and finally, not all multiparty systems worked and performed in the same way. Sartori suggested retaining the numerical criterion, but supplementing it with a criterion referring to the type of competition at work in the party system. That is, he suggested adding to the 'format' of the party system, an evaluation of its 'mechanics'. In so doing, he not only succeeded in formulating precise distinctions among different types of multiparty system, but he also convincingly indicated, in an intellectually very rich and suggestive chapter, the various possibilities for transformation. In the case of Italy, attention was focused on the relationship between the Christian Democrats (DC), the party continuously in government, and the Communists, who comprised its continual opposition. This situation was defined as an 'imperfect two-party system'. Resorting to a comparative perspective, that is, looking also at the cases of the Spanish Republic in the 1930s, at the Weimar Republic, and at the Fourth French Republic (later, at the end of the 1960s, Chile also conformed to the pattern), enabled Sartori to formulate his category of 'polarised pluralism', that is, of a party system characterised by the existence of anti-system oppositions, by a large ideological spread, by the occupation of the centre, by the impossibility of alternation, and by very low overall performance. In my opinion, even more important features of Sartori's classification are its identification of a predominant party system and its distinction between two-party systems and limited multiparty systems. In predominant party systems, one party, always the same, repeatedly wins, over a long period of time, enough seats to enable it to govern alone. Leaving aside all the other criteria, Italy never had a predominant party system because the DC never won enough seats to govern alone (and it never desired to). By contrast, Sweden and Japan can correctly be classified as historically predominant party systems.

The distinction between a two-party system and a limited multiparty system is best exemplified by comparing Britain and Germany up until 1987, when the Greens emerged as a party to be counted because of their relevance in electoral competition and in the formation of governmental coalitions. In Great Britain, only two parties were in a position to win an absolute majority of seats. One of them, not always the same (hence alternation in office), did, and always decided to govern alone. It was a genuine two-party system. In Germany, there was at most only one party capable of winning an absolute majority of seats. It did not govern alone. Alternation was possible and in fact took place because one party (the Liberal party) decided to form coalitions with either one (the CDU/CSU) or the other (the SPD) of the two parties. Hence, though small, often smaller than its British counterpart, the German Liberal party had to be counted. In contrast to the British Liberal party, it was relevant in the making of governmental coalitions. But the German party system should by no means be defined as a two-and-a-half party system. It was a limited multiparty system, as its subsequent evolution has convincingly shown. As for the British party system, it remained a two-party system even during the long period of Conservative rule from 1979 to 1997 and has of course continued to be such since then.

Interestingly, even the party systems of polarised pluralism have evolved according to the transformation rules stated by Sartori (1976: 291), that is following changes in either (i) the electoral system; (ii) the constitutional structure, or (iii) the degree of international autonomy. All three types of change easily apply to the transformation of the polarised pluralist systems of France (1946–1958) and Italy (1947–1989) into somewhat extreme multiparty systems characterised by alternation in office. To conclude on this point, there is, in my opinion, still a lot to be learned and a lot to be done in moving, with Sartori's help, from classification to measurement and, therefore, in providing a dynamic theory of party-system change.

The analysis of political parties has attracted the attention of many scholars, above all many sociologists who are fundamentally interested in the 'social bases of politics' (the subtitle of Lipset's (1960) rightly famous book, Political Man). Sociologists, Sartori (1969a) claims, resort to social variables as independent variables in order to explain the origins, the evolution, and the tasks, in terms of voting and preference representation, of political parties. Political scientists ought to have a different approach/perspective. For political scientists, political parties themselves are the independent variable. Explanations of political phenomena can, and must, be found within politics itself. In any case, all political explanations have to look first for the existence of political variables capable of providing the necessary (and, often, the sufficient) clues.

Sartori's theoretical concern was, at least initially, obviously shaped by the Italian intellectual climate. One way or another, most explanations of Italian political phenomena were sought in economic and sociological variables. In general, what could be called raw Marxist reductionism was largely prevalent among Italian political analysts. Sartori's criticism was directed not simply at the many practitioners of this type of reductionism, but also at its consequences, which appeared to deny the very existence of a science of politics. However, even at the international level, some Marxist and sociological reductionism was present in the field of political science. Sartori strongly objected to it and suggested a different approach and interpretation. He argued that 'class behaviour presupposes a party that not only feeds, incessantly, the 'class image', but also a party that provides the structural cement of 'class reality''(Sartori, 1969a: 84). It is difficult to say whether Sartori really won the battle, but his lesson stands as a reminder that political science should, and indeed, can provide better explanations of political phenomena than sociology, especially sociology of the more or less Marxist variety.

It is no wonder that Sartori has remained faithful to his perspective and has offered many excellent examples of and significant contributions to the analytical power of political science. For instance, the explanatory power gained by utilising simple political variables clearly appears in his analysis of the impact upon party systems of different electoral systems. Sartori's 1968 essay, 'Political Development and Political Engineering', although

'For political scientists, political parties themselves are the independent variable.'

not especially well known, illustrates this point and is important from at least three points of view. First, it went against the tide prevailing at the time when a majority of the scholars studying political development thought, more or less inadvertently, that, once launched, political development was a unilinear, positive, almost teleological process. Not so, wrote Sartori, at about the time when Huntington (1968) published his enormously influential Political Order in Changing Societies. Sartori's essay is also notable because of its attempt to formulate a probabilistic explanation based on purely political and institutional factors: the relationship between strong and weak party systems on which strong or weak electoral systems are superimposed. Finally, whenever a probabilistic explanation is formulated and tested some consequences must be drawn, following the probabilistic procedure. If factors a, b, and c appear then it is likely that consequences x, y, and z will follow. This was a clear and original way of presenting the foundations of political and constitutional engineering. In fact, we find ourselves at the intersection of two of the streams of thought that Sartori subsequently pursued: political and constitutional engineering on the one hand, and the comparative method on the other.

COMPARISON AND CONSTITUTIONAL ENGINEERING

Comparative Constitutional Engineering (Sartori, 1994) represents the culmination of decades of study of the consequences of electoral systems and of the nature and working of institutional systems. Ten years after its publication one can confidently state that it remains the most authoritative single study in its field. Three aspects are especially remarkable. The first is its methodological approach in terms of positive and negative incentives, rewards and penalties. The second aspect is substantive. Significantly, Sartori not only contributes to a more sophisticated appreciation of the impact and the consequences of different electoral systems and suggests which electoral system is most appropriate according to the desired outcomes, but he also proceeds to a very neat classification of presidential, parliamentary, and semi-presidential systems assessing their strengths and weaknesses in an elegant and parsimonious manner. The latter contribution must be stressed given the plethora of disorderly, cumbersome classifications that do little or nothing to advance our knowledge. The third aspect refers to the quality of political knowledge and is extraordinarily significant because it has to do with Sartori's conception of political science as a science. What kind of knowledge can political science acquire and provide? Sartori is convinced – and he says so clearly – that the type of knowledge acquired by political science can be put to work to improve the performance of political systems. Of course, this kind of applicable knowledge must be acquired and utilised through carefully crafted comparative analyses.

The various comparisons currently being offered are to some extent based on the methodological imperatives Sartori had identified and analysed a quarter of a century before in what is probably his most famous article, 'Concept Misformation in Comparative Politics' (Sartori, 1970b) – often revealingly misquoted as 'Concept Misinformation' (sic) – and in the opening article, a true statement of purpose, of the Rivista Italiana di Scienza Politica: 'La politica comparata: premesse

'Sartori's position is ... that scholars ... ought to compare in order to check the validity of their generalisations and middle-range theories.'

e problemi' (Sartori, 1971). There is no way of summarising the important lessons to be learned from these two, more-than-seminal, rigorous and suggestive articles. However, at least two elements must be emphasised. The first one concerns why scholars compare. Sartori's position is not that scholars compare in order to explain specific phenomena. On the contrary, scholars ought to compare in order to check the validity of their generalisations and middle-range theories. Comparison is, above all, a 'method of control'. The second important element Sartori emphasises is that comparison is a demanding method. Therefore, scholars must learn how to compare. Sartori specifies that in order for them to do so satisfactorily, it is indispensable that they acquire two types of learning: (a) how to classify phenomena; and (b) how to move along the ladder of abstraction (Sartori, 1970b: 1044). All this said, the best comparative strategy consists in relying on medium-level categories in order to perform 'intra-area comparisons among relatively homogeneous contexts (middle range theory)'.

Often, scholars interested in methodological issues get lost in the complexity of their techniques and reflections. Sartori is an exception. Indeed, Comparative Constitutional Engineering represents the best example of his ability to remain true to his methodological principles. The 'universe' of cases is represented by democratic political regimes and the analysis moves from electoral systems to institutional systems. Once again it is extremely difficult to summarise the content of this book of 'relatively modest size' because of its sharp, highly parsimonious, and extremely rigorous analysis. Generalisations are formulated and checked against existing cases. If exceptions are found, the existing generalisations are reformulated and followed with the statement of a law that should cover all the cases satisfactorily. This procedure is implemented with extreme care when dealing with the consequences for party systems of electoral systems and in formulating new, more accurate 'laws'. This is a book that contains important lessons for scholars and policy-makers interested in shaping and reforming their political systems. Unfortunately, most of the literature on democratisation can be accused of having underestimated or even totally neglected the importance of mechanisms and institutions for the achievement of whatever desired outcome.

CONCLUSION

In this brief essay, I have at most been able to suggest what I think are the most outstanding and enduring of Giovanni Sartori's contributions to contemporary political science. Four major books in particular – Democrazia e definizioni (Sartori, 1957), Parties and Party Systems (Sartori, 1976), The Theory of Democracy Revisited (Sartori, 1987a), and Comparative Constitutional Engineering (Sartori, 1994) – as well as several articles published in prestigious journals, such as the American Political Science Review, and several chapters in important collective enterprises, mark his special, and from many points of view, exceptional scholarly career. Each of his books has been translated into several languages and they remain essential reading, for scholars and students alike, in a discipline that often seems lacking in 'institutional and professional memory'. All this said and duly emphasised, I think that Sartori has made two additional immensely important contributions to political science. First, he has led the way in the search for true autonomy in political research, accompanied by a solid methodological awareness with specific reference to comparative politics and the comparative method. Secondly, he has convincingly shown that good political science does not simply mean 'pure' political science, focused mainly on formulating neat and polished models, sophisticated paradigms, and the quantification of its findings. In several areas political science has, and has always had, a powerful tendency to become political engineering. This tendency should not be resisted. Rather, it should be channelled, disciplined, and put to work. By so doing, we will witness (and contribute to) a better political science not only for its practitioners, but also for political systems and citizens. As suggested by his powerful editorials (now available as a set of collected essays in Sartori, 2004) in the leading Italian daily, Corriere della Sera, and in the most important Italian weekly, l'Espresso, and by his frequent appearances on TV as a skilful and even abrasive discussant of public issues, Sartori's strongly held belief in the significance and relevance of political and constitutional engineering gives additional meaning to his work in political science.

Notes

¹ A scholarly assessment of Sartori's contribution to political science is provided in La scienza politica di Giovanni Sartori, which I have edited for Il Mulino, Bologna (forthcoming, February–March 2005). This volume contains chapters by Domenico Fisichella, Giorgio Sola, Angelo Panebianco, Adriano Pappalardo, Giuseppe Di Palma, Giacomo Sani, Stefano Passigli and myself, as well as a complete bibliography of his publications, prepared and organised by Oreste Massari.

References

1. Duverger, M. (1954) Political Parties, New York: Wiley (First published in 1951 as Les Partis politiques, Paris: Armand Colin).
2. Huntington, S. (1968) Political Order in Changing Societies, New Haven: Yale University Press.
3. Lipset, S.M. (1960) Political Man: The Social Bases of Politics, Garden City, NY: Doubleday.
4. Sartori, G. (1957) Democrazia e definizioni, Bologna: Il Mulino.
5. Sartori, G. (1962, 1965) Democratic Theory, Detroit/New York: Wayne University Press/Praeger.
6. Sartori, G. (1966) 'European Political Parties: The Case of Polarized Pluralism' in J. LaPalombara and M. Weiner (eds.) Political Parties and Political Development, Princeton: Princeton University Press, pp. 137–176.
7. Sartori, G. (1968) 'Political Development and Political Engineering' in J.D. Montgomery and A.O. Hirschman (eds.) Public Policy, Vol. XVII, Cambridge, MA: Harvard University Press, pp. 261–298.
8. Sartori, G. (1969a) 'From the Sociology of Politics to Political Sociology' in S.M. Lipset (ed.) Politics and the Social Sciences, New York: Oxford University Press, pp. 65–100.
9. Sartori, G. (1969b) 'Politics, Ideology, and Belief Systems', American Political Science Review 63(2): 398–411.
10. Sartori, G. (1970a) 'The Typology of Party Systems — Proposals for Improvement', in E. Allardt and S. Rokkan (eds.), Mass Politics: Studies in Political Sociology, New York: Free Press, pp. 322–352.
11. Sartori, G. (1970b) 'Concept Misformation in Comparative Politics', American Political Science Review 64(4): 1033–1053.
12. Sartori, G. (1971) 'La politica comparata: premesse e problemi', Rivista Italiana di Scienza Politica 1(1): 7–66.
13. Sartori, G. (1976) Parties and Party Systems. A Framework for Analysis, Vol. 1., Cambridge: Cambridge University Press.
14. Sartori, G. (ed.) (1984) Social Science Concepts. A Systematic Analysis, Beverly Hills/London/New Delhi: Sage Publications.
15. Sartori, G. (1987a) The Theory of Democracy Revisited, 2 vols (Part One: The Contemporary Debate; Part Two: The Classical Issues), Chatham: Chatham House Publishers.
16. Sartori, G. (1987b, 1990, 1995) Elementi di teoria politica, Bologna: Il Mulino.
17. Sartori, G. (1993) Democrazia. Cosa è, Milan: Rizzoli.
18. Sartori, G. (1994) Comparative Constitutional Engineering. An Inquiry into Structures, Incentives and Outcomes, London: Macmillan.
19. Sartori, G. (2004) Mala Tempora, Rome and Bari: Laterza.

About the Author

Gianfranco Pasquino is Professor of Political Science at the University of Bologna. He has written widely on Italian and comparative politics. His most recent books are: Il sistema politico italiano and Sistemi politici comparati, both published by Bononia University Press, in 2002 and 2004, respectively.