The Development of Comparative Politics 1.1 The Learned Tradition
Several characteristics marked the understanding of government in Europe and the United States as it had developed by the beginning of this century'.
First, there was a strong normative overtone in discussions on government. Normative approaches were traditionally strong in fields like law, philosophy or theology in which problems of government were discussed at the time. Different ideological traditions, whether Conservative, Catholic or Protestant, Liberal, Radical or Socialist, inevitably had their impact on political discourse. So had more specific traditions of political theory which nourished debates on crucial themes like sovereignty, community, authority, liberty, constitutionalism, rule of taw, democracy and so forth.
Second, discussions of government often reflected particular conceptions of history. In the hands of some, this might lead to the elaboration of 'historical laws',often couched in terms of different 'stages'through which societies would develop. 'Diachronic' comparisons thus came naturally. Models of social change often showed a clear evolutionary or even Ideological bias.
Third, there was generally a strong emphasis on political institutions, which were thought to be not only the results of past political strife, but also factors which could control present and future political developments.
Fourth, 'comparative* politics generally assumed specific country perspectives. Thus, in Britain 'cross-channel' dialogues easily developed into a contrast between (stable) British "cabinet government' and (unstable) French 'gouvemement d'assemblee" (or for that matter British 'rule of law' versus French 'droitadministratif). Cross-Atlantic debates resulted into the conflicting typologies of a 'parliamentary' versus a 'presidential' system of government. Perennial debates in France on the merits^ or lack of merits, of the French revolution strongly colored political discussions on problems of constitutionalism and popular sovereignty. Debates in what was to become Germany had a powerful impact on the analysis of state and nation, of the exercise of power, of 'organicist' versus 'liberal' modes of social and economic development, and of

the comparative role of bureaucracies - subjects which were to become the concerns of future social science also outside German borders. Comparisons of European countries with the United States underscored the early nature of American democracy and stressed the importance of voluntary groups in a free society, but the United States could also be held up as a negative yardstick for alleged abuses, for its spoils system, the role of lobbies or a yellow press, or more generally the dangers of 'mass society'.
Typically, smaller European countries tended to be neglected in the reasoning of learned men outside the borders of the particular country itself. Linguistic frontiers may partly explain this. But probably more important was the assumption, typical of 19th and early 20th century power politics, that small countries hardly mattered. At best they might be of little more than folkloristic interest, at worst they were seen as no more than transient players in a world in which Ihe larger countries determined history.
"Comparative polities' then went generally not much beyond speculation and the study of 'foreign government'. Other states were generally seen as entities all on their own, or at most as possible yardsticks against which to measure developments jn one's own society, and then often as negative yardsticks at that.
1.2 The Political Shocks of the 20lh Century and the Erosion of Institutional Certainties
All this was to change drastically in the wake of three fundamental 20th century shocks: the breakdown of democracy in Weimar Germany, the rise of totalitarian systems and the turn towards aulhoritarianism of most of the new states which were established following Ihe demise of European colonialism.
The formally legal 'MachlUhernahme' in Weimar Germany in 1933 shattered democratic hopes and self-confidence. The Weimar constitution had been heralded as the perfect model of democratic constitutionalism. Its fall destroyed the trust in political institutions as sufficient guarantees of democratic rule. Admittedly, some theoreticians attempted to retain ' institutional ist' explanations, singling out 'faulty'inslitutions such as proportional representation

(e.g. Hennena. 1941), the presence of a directly elected President next to a 'normal' but thereby weakened Kamler, or the absence of judicial review, as major factor* in the destruction of democratic rule. But generally, institutional ist analye" stood discredited. A growing awareness of the patent discrepancy between the proniises of the Soviet constitution of 1936 and the realities of naked power relations in the USSR reinforced this tendency, as were events in France in 1940.
The rise of totalitarian political systems massively changed (he perceptions of politics. Their development, in some countries and not in others, raised new problems of comparative enquiry. Earlier beliefs about the 'natural* development of democracy foundered. 'Autocracy' had been a time-honored category of political analysis, and 'absolutism' had been the natural counterpoint of constitutionalism and later of democracy. But total itarianism seemed to represent an entirely new political phenomenon. Problems of power and leadership, of propaganda and mass publics, of repressive one-party systems and police rule, came to dominate political discussion. Sociological and psychological explanations seemed to offer better insights in the realities of totalitarian rule than traditional political theory or institutional analysis.
The post-1945 world was soon to see also the rise of many new states from what had been colonial dependencies. Such states had generally been equipped with constitutional arrangements, which in most cases proved ineffective to stem developments of authoritarian regimes, whether in the hands of traditional elites, military or bureaucratic governors, or revolutionary party leaders. Such developments further undermined a belief in institutional approaches, and called for alternative modes of analysis.
One effect of the great political shocks of i the 20th century was a massive migration of scholars, notably to the United States, but to a lesser extent also the United Kingdom. One needs only list prominent names such as Karl W. Deutsch, Henry W. Ehrmann, Otto Kirchheimer. Paul Lazarsfeld, Kari LoewCTStein, Hans Morgcnthau, Franz Neumann, Sigmund Neumann, and Joseph Schumpeter, to make clear the importance of this factor for new developments in the study of politics. That field was also to attract the progeny of European refugees who as a typical 'second generation* turned to the analysis of comparative and international politics in great numbers- Exiles from Hitler were followed by migrants from Communist repression, and later still by a

growing number of Third World scholars who opted to stay in the First World. A desire for the systematic study of comparative politics came naturally in such circumstances. It heightened concern with the realities of political power, both within and between states. It made for a characteristically ambivalent attitude about democracy: if anything the belief in democratic values became stronger. but expectations about its chances turned more pessimist.
1.3 Academe and a Changing Political Universe
If migrant scholars looked back naturally on developments in continents they had left, the world was changing, and so was the role of the United States in what was rapidly becoming global politics for policy-makers and students of politics alike. Although Europe remained a key area, other parts of the world, including notably the evolving Communist block, Japan and a rapidly growing number of new states, became matters of urgent political and intellectual concern. So did Latin America, long regarded as a backyard of a Monroe doctrine America. Comparative politics saw the number of its possible units of analysis grow beyond recognition. At the same time problems of political stability and legitimacy, of social and economic development, of competing political regimes and ideologies assumed an entirely new importance.
The need to understand this new world could be met in a variety of ways. It underscored the importance of experts on single countries, notably those which became the object of particular policy concern. It increased the relevance of traditional area studies which it released from their (sometimes almost museum-like) preoccupation with the unique features of 'other* civilizations; in the process cultural anthropology became a more central field in contemporary social science. At the same time, older beliefs about inevitable - and presumably static - differences gave way to concerns with political and social change and beyond this: to discussions to what extent such changes could, and should, be engineered.
All this fitted in well with the traditional temper of American academe. The lure of 'science* had traditionally been strong and had expanded much beyond (he 'natural sciences' into the social sciences and even the humanities.

So had tfae uMUDption that 'science' could and should lead to practical policy results. There wa" a strong belief that the academic enterprise should center on (he elaboration of testable theories. At the same time, the idea of interdisciplinary study stood in high esteem. It was given a strong impetus within some of the great univenities (the University of Chicago being a particularly important center). Such ioterdisciplinarity was moreover reinforced by new agencies, including government research councilg, the newly established (American) Social Science Research Council and a growing number of private foundations all becoming increasingly involved in sponsoring 'relevant* research. This in turn facilitated a massive expansion of graduate schools, and fostered collaborative research between senior and junior scholars, the latter being called upon to 'test' particular theories elaborated by the former through detailed empirical research.
All this came to coincide with the development of new research tools, which helped to foster what was soon to become known as the 'behavioral revolution'. Next to library research and field work in a participatory setting, the survey became a powerful research tool2. Governments began to develop more and more important statistical data to monitor the effects of new policies. A rapidly growing number of international organizations, whether global (such as the United Nations and its specialized agencies, the World Bank or the International Monetary Fund), or regional (the OECD growing from the efforts of the Marshall Plan, and the European Communities being particularly important), came to collect statistical data on many countries. To the extent that they were presented in standardized form, this facilitated comparative inquiry. More and more efforts also went into the construction of time-series data, necessary for the study of developments over time. This massive growth of quantitative data (initially developed mainly in the context of economic and social policies and used in particular by economists and experts in social policies) also found their way into data handbooks and data archives. The computer revolution was concurrently to facilitate the storage, analysis and access to such data. The efforts of individual scholars first, research sponsoring agencies later, made the pooling and preservation of research data (including the

products of survey research for secondary analysis) increasingly common practice. All this occurred at a time of a massive expansion of academic enrolment, which increased facilities not only for graduate research, but also for publishing research findings. Both university presses and specialized commercial publishers massively expanded. Journals proliferated. So did professional associations and the number and specialization of workshop and panels at academic conferences.
If both the mass, and the sophistication*, of such developments in social science were taking place initially mainly in the United States, they soon became an international reality. Early in the post-1945 period deliberate efforts were made to foster international comparative research. One powerful stimulus came from UNESCO, which established its own International Social Science Research Council, and which provided a powerful stimulus for the establishment of international professional bodies such as the International Political Science Association (IPSA) or (he International Sociological Association (ISA). Many national governments expanded their research councils. The idea of international exchange and research cooperation found increasing favor, with the fellowship programs of a number of American Foundations, the Fulbright program, and to a lesser extent agencies like the British Council setting a pattern. In the process. English became increasingly the lingua franca of modern social science.
1.4 The New Comparative Politics
Against this general background of political change on the one hand, a massive expansion of international and national policy-making and research on the other hand, 'comparative polities' developed rapidly. The shift in terminology from the older term of 'comparative government' to 'comparative polities' was symbolic for what was in fact a conscious desire to move away from the traditional concern with political institutions towards a preoccupation with political and social developments generally.
There are some particular landmarks in the development of modem 'comparative politics*. One of these was the Evanston seminar at Northwestern University in 1952 which brought together a group of then younger scholars

including Samuel Beer, George Blanksten, Richard Cox, Karl. W. Deutsch, Hurry Eckrtein, Kenneth Thompson and Robert E. Ward under the chairmar.-"hip of Roy Macridis. In a statement, published in the American Political Science Review, they branded the existing study of comparative government as parochial in being mainly concerned with Europe only, as being merely descriptive instead of analytical, as overly concerned with institutions rather than processes, and as insufficiently comparative being wedded above all to case method approaches (Macridis/Cox. 1953). Some of the members of the Evanston group vigorously clashed with stalwart representatives of an older generation, including such luminaries as Carl J. Friedrich, Maurice Duverger, Dolf Stemberger and William A. Robson during a colloquium of IPSA in Florence in 1954 (Heckscher, 1956). Such older practitioners of comparative government were not readily persuaded by the new gospel. They were to note gleefully that the most irascible proponent of the new 'comparative polities', Roy Macridis, was soon to publish work on France and other countries along what seemed after all rather traditional lines. The continuing need to take account of specific country perspectives was also to become apparent in the work of other scholar* of the group, who after all became editors and authors of influential textbooks organized on the basis of country studies (covering again mainly the larger countries; e.g. Beer/Ulam. 1958; Macridis/Ward, 1963).
In the meantime, a group of scholars (including some members of the Evanston Seminar) was being formed who as a group would have a lasting influence on the development of comparative politics. Many of them were. or would be, active in what was soon to become known as 'the Committee' (i.e. the Committee on Comparative Politics of the (American) Social Science Research Council), In the second half of the 1950s, this Committee deliberately brought together a number of leading area experts. With Gabriel A. Almond as ils highly influential chairman, it set itself to recasting the analysis of comparative politics along mainly structural-functionalist lines. As Almond explicitly stated in the influential volume edited by himself and James S, Coleman. The Politics of the Developing Areas (1960), the ambition was to find "a common framework and set of categories to be used in ...area political- analysis'; to this end Almond himself engaged in "experiments in the application of sociological and anthropological concepts in the comparison of political systems", irrespective of time or area. This work was eventually to lead to the famous 'crises of political

development' model, which sought to analyze political systems in terms of the character and sequence of six major processes: legitimacy, identity, penetration, integration, participation and distribution (see also: Pye, 1965). One manner in which to validate such approaches was to bring together members of the Committee with experts on areas, particular institutions or social processes for a series of books on different aspects of political development, including communications (Pye, 1963), bureaucracies (LaPalombara, 1963), political culture (Pye/Verba, 1965), education (Coleman, 1965), parties (LaPalomba-ra/Weiner, 1966), and (belatedly) stale-formation (Tilly, 1975). Two works were intended to cap the approach: a book offered mainly as a textbook (Atmond/Po-well, 1966), and a co-authored volume on Crises and Sequences in Political Development (Binder et al., 1971). Whereas the first seemed to proclaim certainty, the latter revealed considerable self-doubt and disagreement in the Committee. Clearly, its members did not see eye to eye on such fundamental matters as the existence or not of a linear development from tradition to modernity, and the possibility to engineer social change or not.
Of course, such debates were not restricted to members of the Committee. A great many scholars, in different disciplines, tried their hand defining processes of political development and modernization3. For all their diversity and disagreement, such writings had in common an attempt to understand processes of social change, conceived as in principle comparable over different areas and time-periods, and tackled with instruments from whatever social science discipline seemed appropriate. Such approaches also led to a reconsideralion of past patterns of political and social change in nations already seen as fully or mainly modernized, including the United Stales itself and Western Europe. Historians were asked to join in such efforts at comparative understanding (see for example; Black, 1968; Grew. 1978; Tilly, 1975; 1990).
The impact of these approaches on the discipline was substantial. All manner of Ph.D. candidates swarmed out to study processes of social and political modernization in countries all over the world. They did so with

different interests and intent. Some became thoroughly intrigued with the persistent role of traditional utructures and beliefs, nuking them eager novices in the ranks of area specialists and cultural anthropologists. Others concentrated rather on the other end of the presumed tradition-modernity continuum, identifying largely with the search of economista and experts on public administration for * development'. Yet others felt more happy with the work of various international organizations which sought to monitor and stimulate social and economic developments with the aid of statistical indicators, regarding the universe of nations, or some particular sample of it, as a laboratory in which to test particular development models.
1.5 Inevitable Reactions
For all its exhilaration the political development boom was to create its own mctiocu, in rather different way.
One reaction consisted in the development of counter-models of development which treated the prosperous West not as the prototype of a modern society which others were naturally to attain at some later stage, but as the root cause of an inequitable distribution of the world's goods. Marxist theories of (neo)imperialism held capitalist development responsible for the exploitation of the Third World, and regarded the so-called 'independence* of former colonies as a thin guise for what was in practice 'neo-colonialism'. Notably from the background of Latin America, which had much older independent states than Africa and parts of Asia, developed the various brands of'dependency* theory which emphasized the co-existence of traditional sectors of society and the economy with modem economic sectors which were in practice little more than the emporia of the advanced economies in the USA and Europe. Such models were given a more elaborate treatment in Wallerstein's World System approach, which became in many ways an academic industry of its own.
A second reaction came from those who had difficulty fitting Communist systems into the framework of general development theories. To many such ft problem did not seem particularly urgent: the comparative study of Commu-

nist societies was to a considerable extent a world on to itself and many were happy to leave it at that. The idea of a possible convergence of systems in the West and the East seemed to most observers bereft of reality, perhaps a matter of speculation for economists, not for those who knew the patent differences in political life from direct physical experience or historical analogy. But developments of Communist states did yet enter the field of general comparative politics for at least two reasons. Communist models might and did serve as example and inspiration for Third World countries, notably in their Chinese and Cuban variety. And in a more theoretical vein, a debate arose on the issue to what degree totalitarian systems were themselves a product of modernity. This point had been strongly argued by Carl J. Friedrich, who saw in that characteristic the fundamental difference between older systems of autocracy and royal absolutism and modem totalitarian systems (Friedrich, 1954; Friedrich/Brzezinski, 1956), but was denied by scholars like Wittfogel who saw many common features between the system,': described by him in his 'Oriental Despotism' (1957) and systems of modem totalitarian rule. Nevertheless, whether seen as possible models of modernization, or as alternative expressions of modernity itself, the study of totalitarian systems remained on the whole outside the scope of general comparative politics writing. At least one reason for this was the tendency to equate political modernity with democracy, in systems already existing or as the natural end-product of political development.
A third reaction to the political development literature consisted in the allegation that it rode roughshod over the uniqueness of particular areas or countries. Such was the natural reaction of scholars nurtured in a tradition of 'con figurative* studies, whether of a particular local culture, or a particular political system. Such scholars were not comfortable with what they regarded as overly general categories of analysis. They emphasized that the essence of political and social systems lay in the complicated interaction of many variables which could only be disentangled by destroying the uniqueness of Ihe whole. And they tended to deny the possibility of real comparative study given the inability of scholars to really know more than one or two cases sufficiently well.

The Development of Comparative Politics 1.6 Rethinking Europe
For a time Europe became a somewhat ambiguous area in the development of the new comparative politics. The Third World seemed to attract most of the theorizing and field research, as did to a lesser and more specialist sense the development of Communist systems. Europe seemed possibly somewhat old-fashioned, a world of staid democracies about which all was known and where little happened. The very concept of Europe had become somewhat hazy, moreover. The erection of the Iron Curtain had lopped off a number of countries which had formerly formed a natural part of the European universe. ]f one saw Western Europe as for all practical purposes identical with 'democratic Europe', then certain European countries (including some members of NATO like Greece, or Portugal, not to speak of Spain) presumably did not belong. If democracy were the defining characteristic, why Ihen not study all modem democracies together, thus abandoning the very existence of 'Europe' as a distinct area (a conclusion drawn for example by Lijphart, 1984)?
Whatever such qualms, 'Europe* was soon to figure prominently on the map of comparative politics again, through a variety of circumstances. The persistent concern about 'totalitarianism' naturally made for comparative enquiry into past events: what after all had caused the breakdown of democratic regimes in some countries, and nol others (see notably the consciously comparative study ofLinz/Stepan, 1978). When much later Greece, Spain and Portugal alt returned to democratic rule, the reverse question arose: what were the causes for such transitions from authoritarian rule (see: O'Donnetl/Schmitter/Whitehead, 1986;
Diamond/Linz/Lipset, 1988). The failure of imposed constitutional regimes in many former colonies raised the issue whether alternative models of democracy might have done better; where was one to find these but in Europe (the British dominions usually being regarded as mere offshoots of a British system)? The general concern with development posed many questions for which the history of different European countries might provide possible answers, whatever the dangers of historical analogies. There was a rich literature on European countries, and access to sources was relatively easy. Europe provided, moreover, a variety of cases vital for comparative analysis with a generalizing intent,

provided one really knew the specific cases that made up Europe, and went beyond the exclusive concentration on a few larger countries only.
Much of the history of the development of comparative politics writing in and on Europe can in fact be written in terms of a desire to take account of the political experience of particular countries (for a fuller elaboration of this theme, see Daalder, 1987). As a special subdiscipline, European comparative politics grew largely from the efforts of a new/postwar generation of younger scholars who engaged in a massive trek, to some extent to the United Kingdom, but particularly to the United Stales. They found there an exhilarating world of scholarship, with all manner of theoretical speculation and rich empirical research. This was in strong contrast with the paucity of 'modem' social science literature in their own country, and led naturally to a desire to emulate and replicate studies on America with comparable studies at home. At the same time, a confrontation with Anglo-Saxon scholarship also provoked a natural reaction against what were often fell to be too specifically 'British* or 'American* theories, typologies or models, and fostered a desire to develop alternative theories and typologies which were more in line with the understanding of one's own country. At a minimum, more countries should be brought onto Ihe map of European comparative politics, which somewhat ironically required 'translating' their experience into Anglo-American concepts.
Thus, some of the most innovative comparative politics writing by European scholars betrays, on closer analysis, a strong influence of particular country perspectives. This had been irritatingly clear from what purported to be a general study of political parties by Maurice Duverger (1951), which for all the help the author received in data collection from an early IPSA network of European political scientists, was shot through with French perspectives and prejudices. But one can also document the impact of Italian concerns in the much more sophisticated analyses of party systems by Giovanni Sartori (1976). There is the disappointments of a left-socialist German emigre-scholar about postwar developments in Germany and Austria in the work of Otto Kirchheimer (see the collection by: Burin/Shell, 1969), just as Scandinavia provided the undoubted background of the development of a center-periphery model in the rich work of Stein Rokkan (1970). An even clearer example is the deliberate development of the consociationalist model against the background of The Netherlands, Belgium. Austria and Switzerland, to counter the massive impact

of what eeemed too easy an identification of Anglo-American models of government with democracy perse*.
From the mosaic of such parallel studies a much more sophisticated picture emerged of the diversities of European experience which could be otudied both in a diachronic and a synchronic manner, culminating in what is as yet the most satisfactory attempt at understanding the complexities of European political developments contained in Stein Rokkan's so-called 'topological-typological' map, or 'macro-model' of Europe (Rokkan, 1970; Rokkan. 1975;

1.7 Different Research Strategies
Taking developments in the study of 'Europe' as an example, the considerable variety of modes of comparative study becomes readily apparent.
A seeming paradox is provided by the country monograph. To the extent xuch a monograph is written to elucidate particular political experiences for a more general public, it may offer insights of comparative importance. Fhis is much more true if the monograph seeks lo prove, or disprove, specific theoretical propositions first developed with one or more other countries in mind. The most telling example, however, is the consciously theory-based analysis of a single country case (e,g. Eckstein, 1966; Lijphart, 1968b). Moving o to a somewhat higher level of abstraction are comparative analyses of two, or a few, particular countries'. Most 'comparativists* must confess that their real knowledge of different countries tapers off quickly beyond a rather limited number of cases. One obvious way to overcome such limitations is collaborative research, in which for any given research question experts on different countries
The value of this strategy had not been lost by members of the Committee on Comparative Politics which sponsored as one of its first projects a comparison of Japan and Turkey, see Ward/Rustow eds. 1964

are asked to join in a common research effort. Much the roost frequent books on (European) comparative politics consist of edited volumes of this kind. Such volumes bring much needed information on different countries together and testify to the fruits of cross-fertilization- But most of them suffer the natural defects of group enterprises. The choice of countries is often a function of the availability, or even the reliability, of individual country experts. Even the most rigorous attempt at editorial guidance rarely wsolts in an even quality, let alone genuine comparability, of country chapters. Introductory and concluding chaplers very often are of a rather ad hoc and impressionistic nature (but see for impressive examples volumes of a lasting nature, such as Neumann, 1956; Dahl, 1966; Rose, 1974; Budge/Robertson/Hearl, 1987).
This strengthens the case for attempting individual syntheses after all. The difficulty of such an enterprise becomes readily apparent, however, if one seeks for post-war equivalents of the great comparative government treatises of the past (e.g.Friedrich, 1937; 1941 ;H. Finer, 1932; 1949. not to speak of earlier classics such as Lowell, 1896 or Bryce, 1929). These are very hard to find (e.g. Blondel, 1969; Finer, 1970; for later attempts to analyze 'European polities'see Smith, 1972; 1989; Steiner, 1986; Lane/Errson, 1987; Pelassy, 1992; and the three-author volume of Gallagher, Laver and Mair, 1992), and encounter the obvious problem of an increased number of countries to be treated, with many more empirical research findings of potential relevance to be covered.
Ralher than on analysis at the level of countries as a whole, work has tended to focus on particular institutions such as monarchy (e.g. Fusilier, 1960). heads of state (Kaltefleiter, 1970), the formation of cabinets (e.g. Bogdanor. 1983; Pridham, 1986), parliaments (Wheare, 1963; Von Beyme. 1970). electoral systems (Lijphart, 1993), parties in general (Sartori, 1976; Von Beyme, 1985;
Panehianco, 1988), particular party families, interest groups, bureaucratic structures, and so on. In studies focusing on particular institutions or groups, there is always real danger of analyses out of political and social context.
Alternatively, there is the massive growth of quantitative 'cross-national studies*. As stated before, both the quantity and the quality of data has increased massively in the last decades, through the efforts of governments, international

organizations, the gallant work of those who prepare "data handbooks'^), and organize data archives. Such data invite cross-naliona) studies, in a large number of fields. Thus one need only inspect the guide to journal articles in 'Electoral Studies', not to speak of important collaborative volumes (ranging from Rose, 1974 to Franklin ct al., 1992). to see the richness of studies on electoral behavior, and of elections (cf. Bartolini and Mair, 1990). We have important studies oa political participation (influenced notably by the works of Ver-ba/Nie/Kim, 1978; and Barnes/Kaase, 1979) and on the impact of changing values (an area dominated by the highly debated analyses of Inglehart, 1977;
1990). The study of cabinet coalitions has offered a fertile a testing-ground of formal theories (for a useful survey and discussion see: Laver/Schofield, 1990). As we shall see presently, the data revolution has also had a great impact on the study of the development and problems of modem welfare states and public policy. Not all such cross-national studies are really comparative, however. Although they draw on data from many countries, they are often directed more to problems of general political sociology or psychology than lo a systematic inspection of country variables. 'Contextual' knowledge is often neglected, and with it possibly the essence of comparative politics itself, which in the words of Sidney Verba presupposes that one
"tries to generalize - using that term loosely - about nations, or to generalize about subnational entities like bureaucracies, parties, armies and interests groups in ways that use national variation as part of the expl(inalion''(ital'ics HD; Verba, 1986:28)
A lack of knowledge of (he countries studied has made some such 'cross-
Mtional' analyse! verge on what Stein Rokkan once dubbed mere 'numerologi-cal nonsense*.
See for some notable examples covering rather different variables and countries TayIor/Jodice, 1983; Flora, 1983; Mackie/Rose, 1991; Lane/
Mackay/Newton, (991; Katz/Mair, 1992; Woldendorp/Keman/Budge, 1993

1.8 New Approaches
Developments in modem comparative politics, then, were largely the result of a greater knowledge of individual countries on the one hand, and of a true revolution in data collection and analysis techniques on the other. But at the same time, new political problems appeared^-on the political agenda, which resulted in something like a paradigmatic shift. If comparative politics had concentrated thus far mainly on problems of regime change, political institutions, and what in systems theory one calls 'input' structures, a new concern developed with problems of public policy and political 'output'. Various factors contributed to this development.
One cause was the (renewed?) 'left' revolution in social science in the 1960s and 1970s, which faced the question why 'capitalist' systems endured, once-confident prophesies to the contrary notwithstanding. This led to a new concern with the role of the state which seemed somewhat forgotten in otherwise rival approaches of systems theory and economic determinism7. A parallel debate arose on the extent to which political parties - notably Socialist ones (see Castles, 1978; Schmidt. 1982; Scharpf, 1992;Keman, 1988;1990) - did affect government policies or not. A major element in the discussion became the degree to which slates differed in their dependence on external economic forces (e.g. Katzcnslein, 1985), which could only be solved by comparative inquiry. Even when such studies related to European countries only, the obvious relevance of international economic structures and events brought scholars closer to those who had long been preoccupied with world economic realities (e.g. the proponents of a World Systems approach mentioned earlier).
A second major factor was the development of 'neo-corporatism'. Originating to some extent from a transposition of an approach found useful in the sludy of Latin America (e.g. Schmitler, 1974, reprinted in Schirutler/Lehm-bruch, 1979), it won great acclaim in attempts to explain 'Europe', and possible differences within it. By emphasizing the close interaction between public and
In that light the famous title of Evans, Rueschemeyer and Skocpol, 'Bringing the State Back In' (1985) would seem to testify as much to a new vision of those who had been strangely blind, as to the real record of political studies they criticized.

private actors, the neo-corporatist approach seemed successfully to bridge input and output structures, and to present a more realistic picture of power relation! and policy-making than either those who had spoken uncritically of *the' state, or those who had embraced a naive 'pluralism', had been able to provide. Neo-corporatism became in Schmitter's words 'something like a growth industry'. But the gap between 'general' theory and empirical validation remained substantial, to the detriment of the value of the approach as a tool for general comparative analyses as distinct from the study of specific policy areas.
A third major contribution came from those who set out to analyze the development of the welfare slate in comparative terms. On the one hand, this work fitted in well with the concerns of older development theorists: one should note the link between stale expansion, economic policies and processes of political development which had characterized the work of German Kathedersoti-afistm md Nationaloekonomen; (redistribution had been one of the paramount concerns of the Committee on Comparative Politics; and the leading empirical scholar in this field, Peter Flora (1974; 1975; Flora/Heidenheimer, 1981: Flora, 1986 and following years) saw his work as filling a gap Rokkan's macro-model of Europe. On the other hand, comparative work on the welfare state was to encounter what was soon to become the major debate on its 'fiscal crisis', and on possible limits of state intervention more generally (Lehner et al., 1987;
Kenuui et al., 1987; Lana/Ersson, 1990). The label 'political economy' was to cover a wide variety of concerns, ranging from rational choice paradigms based on individualist self-interest, to studies of specific policy areas, competing models of general economic and monetary & fiscal policy, and renewed debates on political legitimacy. The full weight of such new approaches on the study of comparative politics is discussed by other contributors in this volume.
1.9 The Great New Challenges
But such challenges would seem to pale before the momentous changes taking place in what had been thought of as the Communist world, and the attendant shifts in contemporary international relations. In addition, the progress of European integration, however halting, is affecting the very basis of independent

states as the unit of analysis cm which so much of comparative politics has rested.
The long-standing assumption ofa natural division of labor between the study of international relations engaged in analyzing the interaction of states, and comparative politics concerned with the study of processes within states, always rested on somewhat dubious ground. It left unclear what scholars were to handle the formation of (new) slates; it glossed over the great importance of domestic political processes on the making of foreign policies; it belittled what became known in the international relations literature as 'transnational' politics;
and it postulated a degree of political independence for 'sovereign* states which never completely fitted the realities of an interdependent world (as advocates of a World System approach, dependency theorists and other political economy theorists had long maintained).
The division of the world in rival blocs had arguably permitted a certain separation of international relations and comparative politics. The assumption that existing states within a bloc remained distinct units of analysis seemed tenable in a world of relatively stable alliances (the necessary ceferis paribus qualification being as easily forgotten as it was given). The much more fluid international scene of to-day makes such an assumption rather more questionable.
At the same time, developments within the European Community increasingly undermine the role of member stales as independent units, even though international modes of decision-making remain juxtaposed to supranational ones. Powers of decision in vital matters are either shared or transferred to organs 'beyond the nation-stale', while at the same time states also lose formal or effective powers to regional or local units. The 'national' power to control citizens, groups and enterprises becomes more dubious in a world of increased mobility and communication, affecting the status of individual 'states' as realistic units for comparative analysis.
But the greatest, if generally unexpected, challenge to comparative politics comes from events in Central and Eastern Europe. We mentioned earlier that the study of Communist states had become mainly the concern of a specialist group of (cholars. Experts on Communism have largely lost their "subject', although they have retained their knowledge of language and area. Scholars who were mainly concerned with the study of the development and the

working of democracies, on the other hand, stand before an entirely new universe. Their concern had generally been with the comparative treatment of ecuf/n^ democratic states, which is a far cry from the making of new democracies in societies which have not known democratic rule for two political generations or more. For all the words spoken by pundits at symposia, in newspaper columns or journal articles, the extent to which proven knowledge exists is unclear.
The future of democracy presupposes at a minimum the creation of new institutions, but the brunt of comparative politics teaching since Weimar has tended to discount the independent effect of political institutions. Seemingly abstract debates on the merits of presidential, semi-presidential or parliamentary systems of government, on unicameral or bicameral legislatures, on electoral systems and their effect on the politicization of cleavages and the formation of party systems, on the proper role of Judicial bodies, become suddenly matters of crucial importance again. But they must function in areas with all the remnants of a totalitarian past, rival claims for political control and citizenship, possibly severe disagreements on the nature of the political unit itself - and all this amidst economic ruin and change. It is as if all major issues in the study of comparative politics are chaotically thrown together: the formation of states, the working of institutions, the rivalry of parties and groups, competing ideologies, the provision of state services and their limits, issues of economic interdependence, international power politics, and what not. Against this, one must ruefully acknowledge that basic political phenomena such as civil war, terror, ethnic conflict or the shattering effects of ideological strife, have traditionally tended to fell in the interstices of the study of international relations, comparative politics and political theory, rather than forming their core.
Comparative politics, then, stands before its greatest challenge yet. Never before were so many fundamental questions raised at one and the same time. In all honesty one should acknowledge that it provides few definite answers. o

2 Comparative Politics:
A Distinctive Approach to Political Science?
Hans Keman
There have been several attempts to delineate the boundaries of comparative politics, yet there is little agreement at present on its distincliveness. Essentially, one could argue that there exist four different ways of defining comparative politics; firstly, those who distinguish it from other approaches to political science by referring to certain concepts employed which can only be properly understood by means of comparative analysis; secondly, those who take as a point of departure the central features of the political process which can be analyzed for all political systems; thirdly, there are those who maintain that politics can only be understood by employing a macro-scopical perspective;
fourthly, and finally, there are many who define comparative politics by means of its method: i.e. the art of comparing, and who justify this by referring to the famous quote of Kipling: what know they of England, who only England know!
Although the last way of delineating comparative politics is purely methodological, it is the most prevalent one and not the worst way to define this field. However, 1 do not wholly concur with this view, for it would mean that the domain of a discipline is defined by its method, rather than by either its core subject, i.e. the study of politics, which is then, of course, still in want of a definition itself, or by its mode of explanation that is supposed to advance our knowledge of the core subject. In this chapter our concern will be therefore to demonstrate what comparative politics can add to political science by means of its use of attributes of macro-social units in explanatory statements (Ragin, 1987:
5). This calls for an elaboration of the core subject in terms of an identifiable object of study and how this relates to various levels of analysis. In addition, a

second concern will be to develop a mcta-theoretical approach that is capable of explaining or lnterpretingmu\ti-\eve\ variations of the subject under review (the so-called micro-macro linkage). A final concern is a resultant of (he preceding ones, and involves scrutinizing existing logics of comparative inquiry to account for the observed variation by means of testing empirical hypotheses, thereby either corroborating or falsifying them (LiJphart, 1975: 159).
All these concerns are in itself worthy of serious discussion and deliberation, yet the main issue at hand is that comparative politics, as a field within political science, lacks coherence in terms of a set of theoretical references and related logics of inquiry. In short this chapter is not only an attempt to delineate comparative politics as a separate field, but most of all must be seen as an argument to relate theory and method of this approach of political science in order to gain a viable and feasible approach to explain political processes.
In order to clarify my point of view, I shall first elaborate some of the existing definitions of just what comparative politics is (section 1). In addition I shall argue in section 2 that comparativists should focus on the political process rather than on the context of 'polities' In the sections 3 and 4,1 shall delve into the matter whether or not comparative politics can be characterized by a 'core subject' of its own, and if so. whether or not it can be distinguished by a theory? In section 5,1 enter into the most disputed domain of comparative politics, i.e. the comparative method and its implications fora 'proper* research design. In section 6,1 shall tie together the different arguments regarding views on comparative politics as a more or less distinctive field within political science. The central argument will be that a coherent framework of theoretical references and a corresponding logic of inquiry is required. If il is not possible to do this, comparative politics will still remain a valuable asset to political ocieoce, yet any claim of being a distinctive approach should then be put to rest.
2.1 Delineating the Field of Comparative Politics:
Definitions and Justifications
Comparative politics has grown out of the wish to know more about one's own political system by comparing it with others. In particular, it was previously believed that knowledge about the institutional framework of politics would not

only help to understand the peculiar rites of one's own polity, but would also enable one to draw conclusions about its merits and disadvantages. However, this approach was dominated by the idea that supplying comparable information on the structure and working of a political system in toto would be sufficient to further knowledge about the political process and that the analysis should therefore remain of a descriptive nature based on facts. For example, James Bryce (1929) stated his task as a comparativist-as follows:
"What I desire is, not to impress upon my readers views of my own, but to supply them with facts, and (so far as I can) with explanations of facts, on which they can reflect and from which they can draw Iheir own conclusions" (p. IX).
This institutional and empiricist approach ('it is facts that are needed: Facts, facts, facts. When facts are supplied, each of us can try to reason from them", Bryce, 1929: 13) lost its appeal after the Second World War when a more analytical perspective was introduced (Blondel, 1981: 173-178). A good example of this change is offered by Roy Macridis (1955), who argues strongly favor of a more 'scientific' approach to politics in general and considers the comparative approach as the most promising, if not revolutionary way to go:
"Comparative analysis is an integral part of the study of politics. The comparative study suggests immediately the laboratory of a scientist. It provides us with the opportunity to discuss specific phenomena in the light of different historical and social backgrounds. It suggests variables of a rather complex order that can be dissociated from the cultural background and studied comparatively. (.....)The comparative study of politics is beginning only now to enter a new stage which reflects in essence the progressive systematic orientation in the study of politics. Il is beginning to assume a central role in empirically oriented study" (Macridis, 1955: 1 & 3)
In sum, Macridis believed that comparative politics as a distinctive field within political science, as he saw it, would be able to bridge the growing cleavage between political theory and the empirically based study of politics. This so-called 'revolution* of political science did not materialize and became a "stalled revolution" (Mayer, 1989; 20) being reduced to a 'movement' and not o distinctive field, or sub-discipline within political science. Almond contended

that comparative politics has indeed revolutionized political science, but that this development should be considered as a stage in the development of political acieoce. He concludes:
"It is difficult to see therefore, that comparative politics has a long-run future as a sub-discipline of political science. Rather, it would appear that, like the political behavior movement which preceded it, its promises lies in enriching the discipline of political science as a whole" (Almond, 1968: 336).
Almond's 'developmental' explanation, of course, would mean an integration of the comparative movement into political science. His view is not wholly shared by all of his contemporaries such as Macridis, Rokkan, Daalder and Verba (compare: Daalder, 1986). Although they are critical with respect to the development of comparative politics as a distinctive field within Political Science, they instead stress the fact that the 'stalled revolution' might be seen os the paradoxical result of the growth of comparative research. Thus, instead of ft 'paradigmatic' development there has been a 'pluralistic* one. Verba therefore concludes his views on "where have we been, where are we going?" u follows:
"For the future, one can expect to sec more of the same in the comparative politics field. It will remain fragmented and appear disorganized. (,..,)My guess, nonetheless is, that the discipline will maintain its heterogeneity of styles and theories and that most of its practitioner* will continue lo view that as healthy." (Verba, 1989: 36)
I do not concur with this view, for such a liberal approach to a sub-discipline ultimately evades the question what distinguishes it from other Fields within political science. Furthermore, it avoids Ihe important question of what it has to offer to a student of politics in his, or her, quest for understanding and explaining the politica process in a society. It is remarkable to nole that those who assess the 'slate of the art' in comparative politics choose either lo view it in an evolutionary way. or to see it as a pluralistic development.
For example, Holt and Turner represent the evolutionary, not to say a 'positivist', view. In the early seventies they stressed the pre-paradigmatic Mtuntion and sought a solution in 'scieotism' in order to move beyond heuristic

schemata (Holt/Turner, 1970: 70). Hence the ftiture of comparative politics depended, according to them (and many others; e.g. Mayor, 1972; Lasswell, 1968:
Memtt/Rokkan, 1966), by and large on the development of a proper methodology and genuinely comparable datacol lections. There is nothing wrong with such an empirical analytical approach, but one wonders to what extent this 'solution' is in fact different from Lord Bryce's optimistic view that factual description would render deeper knowledge about thc'Tiature of politics.
The 'later' Macridis does indeed not share this way of thinking any more. Reappraising his own optimism of the 1950ies in 1986 led him to a rather ambivalent posilion. On the one hand. he maintains that "Comparative politics is emerging as the most comprehensive and theoretical branch of political science" (p.49), and on the other hand, he suggests that the search for a grand theory should indeed be abandoned and political scientists should rely entirely on the development of subfields or specializations (p. 21). Hence heterogeneity and compartmentalizalion suddenly becomes a positive development, for:
"clustering, as opposed to the search for grand theory, has given to the field a new vitality, and may ultimately pave the way to the development of some unifying models and priorities" (Macridis, 1986: 22),
These lines of reasoning to justify comparative politics as a distinctive approach are not appear fruitful, nor do they help to develop a description of this field within political science. Such reasoning only leads to a situation, which Wiarda (1986) has described as follows:
"Indeed, the field itself seems presently so fragmented and dispersed as to raise the question in some minds of whether comparative politics can still properly be called a field at all" (Wiarda, 1986:5).
Clearly the lack of a general theoretical framework and the continuing debate on both the method and the principal concepts to be used is most disturbing. This is not necessarily a consequence of using different approaches within comparative politics, or of employing a wide variety of concepts alone (see foi this: Sartori, 1970). However, such a pluralistic attitude precisely produces a situation that most protagonist of this field with to avoid. For this 'live-and-let-

live' attitude implies an abandonment of the search for a more coherent approach to comparative politics and, in fact, robs it from any substantial meaning and theoretical rigor. It will only lead to sacrificing substance to method, or to raising only those questions that can be empirically answered, but do not relate to the critical problems of mankind (also: Mayer, 1989; 21 .Castles, 1987: 222). It is also unproductive to focus on partial problems within the field, such as choosing the correct method, selecting the right concepts, and finding the proper data. However, do not let me be misunderstood, it is not my aim to suggest an alternative 'grand theory' or an overarching 'concept' of politics, that would save the day for the comparative approach and retaliate its critics. Nor do I think that by solving the methodological issues comparative politics will become a 'paradigm' of political science. Instead. I propose the following guidelines to define comparative politics as an separate field within political science:
1. describe the core subject of comparative politics. In other words to question how do we recognize a comparativist, what distinguishes him or her?
develop a view on which theoretical concepts can 'travel'comparatively as well as possess a unifying capacity for explaining political processes? discuss the logic of the comparative method as a means to a goal, rather than as an end in itself. In other words, which method fits the (research) questions asked by a political comparativist best?
2.
3.
This is, of course, a rather ambitious agenda, but if we do not at least attempt to investigate the possibilities of new directions in comparative politics to journey toward a more or less integrated and distinctive field within political science, we had better leave the field as a separate approach. Yet, I wish to uphold the original motives and intentions of comparativists to develop this approach to political science to enhance its rigor and claim as a discipline. Whatever the today's skepticism of the erstwhile protagonists may be, it is worthwhile to investigate the possibilities of comparative politics thoroughly. Let us therefore now turn to the first point on the agenda: the description of the field.

2.2 Comparatively Researching the Political Process

Comparative political research is generally defined in two ways: either on the basis of its supposed core subject, which is almost always defined at the level oJ political system (Kalleberg, 1966; Wallerstein, 1974; Almond/Powell, 1978), or by means of descriptive features that claim -to'enhance knowledge about politics as a process (e.g.: Apter/Andrain, 1972; Roberts, 1978; Dogan/Petassy, 1990) These descriptions are generally considered to differentiate comparative politics from other fields within political science. Although it is a useful starting point. K is nol sufficient. Some authors are more specific in their description and add to this general point of departure that comparative politics concerns nations and their political systems (Wiarda, 1986), or the study of geographic areas. Finally some authors deliver a more or less exhaustive definition in which "thi comparative study of political phenomena against the background of cultural. sociological and economic features of different societies" is the focus o;
comparative politics (Macridis, 1986; Berg-Schlosser et al., 1987).
All these descriptions may be useful up to a point, but they do not help to mark off the field and require greater specification. Comparative Politics must be defined in terms of its theoretical design and its research strategy on the basi;
of a goal-oriented point of reference, i.e. what exactly is to be explained. A wa;
of accomplishing this is to argue for a more refined concept of 'polities' anr develop concepts that 'travel' and can thus be related to the political process ii various societies. In addition, a set of rules must be developed that direct thi research strategy, aiming al explanations rather than at a complete description of political phenomena by comparing them across systems, through time, o cross-nationally. At this point most comparativists stop elaborating then approach and start investigating, often however, without realizing that theon and method are interdependent modes of explanation.
In contrast, comparative politics should be seen as an approach lha' aims at explaining the political process in a society by means of a (meta-i theoretical framework of reference and where explanations are validated h comparing units of analysis (see also: Roberts, 1978; Ragin, 1987; Przeworski 1987; Castles, 1989). The goal of comparative politics is to explain thos. 'puzzles' which cannot be studied without comparing and are derived fron

logical reasoning. Hence, no comparative research without an extensive theoretical argument underlying it, nor without a methodologically adequate research design to undertake it.
In most discussions of comparative politics, it appears to me, that both theoretical and methodological aspects arc divorced, or - at least - treated separately. For example, Ragin (1987) and Przeworski (1987) emphasize predominantly the methodological aspects of the art of comparison as a 'logic of inquiry', which is often underdeveloped or incompletely elaborated. At the same time these authors argue their case by mean;; of examples that are seemingly picked at random. Worse even, it seems that some of the examples are selected to demonstrate the tenability of their view. Theoretical progress and explanatory value appears then to emanate from their 'logic'tsee: Przeworski, 1987: 45ff; Ragin. 1987: 125ff). Yet. the comparative analysis of the political process must be founded a priori in theory and then related to the best fitting 'logic of inquiry'".
Another example of such a separation of theory and method can be found in the study of electoral behavior. This vital part of the political process can be explained fairly well on the basis of deductive reasoning. To validate its micro-level founded hypotheses a comparative research design is not necessary. It can be done, but it is only genuinely comparative, if the explanatory concepts are analyzed by examining the variation in the political properties on both the micro- and macro-level. Electoral behavior or party behavior that is explained by means of the working of electoral systems, features of a party system, or the existing rules of government formation are in need of a comparative anlysis(see. for example, Lijphart, 1984; Sartori, 1976; von Beyme. 1985; Bogdanor/Buller. 1983; Budge/Keman, 1990). However, studies which focus on inlra-systemic variation or micro-level variation are, notwithstanding their quality per se and

usefulness as sources of information, not genuinely comparative in nature (see, for example, Blondel, 1985; Henig. 1979).
This conclusion seems to hold for other types of cross-national research too: since the seventies the study of 'electoral volatility' in Western Europe gained momentum, when it appeared that the division of party systems and the gtructure of voting patterns was less stabilized than originally assumed (Crewe/Denver, 1985; Daalder/Mair. 1984). It is interesting to note (with the help of hindsight) that most analyses were, in fact, based on country-based analytical descriptions with little comparative information. What was lacking was a truly comparative set of theoretical references concerning - in this case - the expticandum i.e. 'political stability' that at the same time is consciously linked to a comparable set of operational terms (see: Bartolini/Mair, 1990:35-46).
The same observation can be made with respect to the study of government formation. On the one hand, there are collections of country-studies (more often than not, I hasten to add, developed on a shared list of elements present in each case description; e.g.Pridham, 1986) that stress the idiosyncratic nature of a country's political process, rather than the communality of the development under review. On the other hand, a development can be observed with respect to the politics of coalition-building in which an underlying theoretical argument has been developed that directs the research, where countries are not the principal focus, but a collection of comparable cases that show variation concerning what is to be explained (e.g. Laver/Schofield, 1990. Budge/Keman, 1990). Other examples could be mentioned to support this point regarding the relation between theory and method in comparative politics (such as the comparative research into the relation between "politics & policy'; see:
Castles, 1989;Schmidl, 1991; Keman/Lehner. 1984; Chapter 7 & 8 in this book). Yet. the principal message is that much of the research that is labeled as comparative, either lacks theoretical foundation, or is based on a research design that is not comparative.
In fact, what one often sees, as Sigelman and Gadbois (1983) have found for the United States, is that the majority of studies published as being comparative are single nation studies. Even iflhis criterion is relaxed (i.e. case-studies with a comparative focus) the non-comparative studies are one-third ol the total reviewed. The same applies to textbooks and courses on comparative politics. Mayer (1983) concludes that nearly half of these are basically country-

focussed presentations without a clear overarching theme or comparative focus. Again, however useful and interesting these studies, textbooks and courses actually may be, they are not genuinely representing comparative politics nor, for that matter, enhancing it as a distinctive field within political science. What is needed, therefore, is to examine the possibility of developing a theoretical perspective of the 'core subject' of comparative politics.

2.3
Modes of Explanation in Comparative Politics: Topics and Approaches

The 'history* of the theoretical development has by and large already been told in Chapter 1 and can be characterized by the development from 'grand theories* spilling over in central topics and related concepts which gradually evolved in more or less separate approaches (see also: Nohleo, 1983: 1077-1079; Wiarda, 1986: Chapter l;Macridis, 1986; Berg-Schlosser et al., 1987). What is remarkable is that in most contributions the core subject under review is either taken for granted, or assumed to be self-evident (e.g. Blondel, 1981). 'Polities' is what governments do (or do not do), the actions of politicians and (their) parties, the institutions that make up a political system and also the process of policy-formation. Yet, little is said about what the nature of politics is, apart from referring to it as an inevitably 'contested notion' (Cf. Connolly, 1983).
I shall not endeavor to put forward here a neatly packaged definition of politics, but rather attempt to show that it is possible to circumscribe it in terms of reference that enables us to distinguish 'polities'from other phenomena in society. I feel this to be necessary, since many of the theories that have been used in comparative politics have failed to recognize that without a coherent circumscription of its core subject, i.e. the 'Political', explanations have tended to become de-politicized. Hence, the theory was not using politics as an explicandum, but as what had to be explained by other non-political features. This development can be amply illustrated by means of the 'grand theories' which were in essence cultural, or functional explanations of political behavior, or conversely remained 'economistic* in nature,
It is not relevant for our discourse to repeat here the criticism;; on these socio-cultural inclined 'grand'theories of comparative politics. Our aim is rather

to demonstrate that the developments in this field led to a kind of theoretical 'escapism' by an attempt to explain politics by means of non-political features of a society and to create an almost ideal-typical concept of political behavior:
the concept of the "proper* citizen, who would enhance the (value-related) concept of the 'best' political system, namely sustaining the working of democracy. Such an approach, however, will often not lead to a theory about the political process, but rather will offer an explanation about its socio-cultural preconditions. Hence my point here is not whether the theory or methodology is correct or not, but rather that this approach is missing the important point of explaining what the political process itself is about (see; Blondel, 1981; Roberts. 1978; Keman. l992a).
In the same mold is, albeit from another angle, the 'economislic' approach (see for this distinction; Barry, 1970). The 'rational choice* school is less guilty of 'escapism' and confusing research designs. Yet, in this approach the main problem is the extent to which the concept of rationality is narrowly defined or not (see: the Chapters 4 & 5 in this book; for an excellent overview consult: Renwick Monroe, 1991), the degree to which the topic under review is essentially still formulated in terms of individual behavior (i.e. on the microlevel of observation), and thirdly, whether or not these theories are - implicitly or explicitly - equating the 'logic' of the market with the 'logic' of the political arena. These criticisms point to a central problem of this approach in terms of theory and research: an 'economistic' concept of rationality suffers from an ecological bias (both on the level of formulating hypotheses and of measuring beyond the level of individual behavior; Dogan/Rokkan, 1966), on the one hand, and from the fact that there is more to politics than can be explained by economic rationality alone, on the other. However, it must be noted that protagonists of the rational choice-approach have attempted to solve this problem. Both the 'public choice' literature (e.g. Mueller, 1989; van den Doel/van Velthoven, 1989) and the development of the so-called 'positive political economy' (e.g. Laver, 1986; Alt/Shepsle, 1990) elaborate the political process per se by means of the institutions that direct the scope of rational behavior of the participants and the consequences in terms of interdependent relations (Scharpf, 1987; North, 1990; Chapter 7 in this book). Notwithstanding the fact that these developments are promising, it is not yet clear to what extent

this type of explanation will eventually lead to a genuine definition of Ihe political process as a core subject of investigation in comparative politics.
In sum; whatever the merits of these approaches are as potential explanations of politics, they do not tell us much about the nature of politics, and therefore are not really theories of the political process, but are instead possible explanations about political phenomena and (series of) political events. How then should we proceed from here if socio-cultural and socio-economic properties are to be considered merely relevant as contextual variables in explaining variations of politics, but not always adding knowledge about the 'political' itself?
Generally speaking comparativisis have proceeded from this point either in the direction of focusing on central concepts, or they have resorted to a 'scientific'mode of explanation by applying the canons of empirical-analytical methodology (including high-powered statistical techniques) to political topics.
The first approach, for example, concentrated on various topics like democracy, political regimes, parliaments, revolutions, welfare states, war & peace, etc. (see for an extensive list of such topics, including the literature: Berg-Schlosser et al., 1987: 271-304). The 'scientific'approach did not differ from the first in this respect, but concentrated much more on the question whether or not hypotheses could be corroborated by available, quantifiable data. Again, this certainly added to the body of knowledge of political situations, events and societal development, but hardly answered the question what politics was all about and how politics itself may have an impact on societal developments. In other words: it appears that we know a lot about politics in the real world, and due to comparative research how different its complexion can be, but at the same time we know preciously little about its nature and the extent to which it is an independent explanatory factor.
An example of this latter issue .is a debate within comparative politics on the question whether or not 'Politics Does Matter' (see: Castles. 1982: 4-15;
von Beyme, 1985: 334-359; Keman, 1988: 71-75; Gallagher et al, 1991: 236ff). It has been one of the first attempts to re-install the 'political' (here: parties and governments) in comparative politics as a variable influencing societal developments. On the one hand, the debate focussed on a topical relation, i.e. democratic decision-making and public policy-formation; on the other hand, it searched for the method to prove right from wrong. Both the theoretical

conceptualization and the correct comparative method were al issue. Regardless of the result of this debate, it demonstrated that seemingly endless debates were possible on concepts itself, the way of analyzing it and the data and techniques used and so on. At the end of the day it appeared, dependent on the countries studied, the time-period under review, the level of measurement and the operational izations employed, that all contenders were sometimes right and sometimes wrong (see also: Castles, 1987; 1989; Keman. 1990). Undoubtedly, this debate within the realm of comparative public policy analysis has helped to further our knowledge of political processes and the usage of the comparative methods. However, it also shows that without a clear view of what the core subject actually is, both the methods used and the data collected cannot really enhance our knowledge of 'polities' as such. Moreover, it does not help us to define and describe comparative politics as a distinctive approach within political
2.4 The Political as the Core Subject of Comparative Politics
The 'political' in a society can be described on the basis of three dimensions:
politics, polity and policy (Schmidt, 1991; Keman, 1992a). Politics is then what I would like to call the political process. On this level actors (mostly aggregates of individuals organized in parties, movements, or groups) interact with each other when they have conflicting interests or views regarding societal issues that cannot be solved by themselves (i.e. deficiency of self-regulation). The process of solving those problems which make actors clash, is more often than not visible through the institutions that have emerged in order to facilitate conflict resolution. Institutions help to develop coalescence and to achieve a consensus among conflicting actors through compromising alternative preferences. These institutions manifest themselves in the rules of the game in a society. This is what is meant with the 'polity'. To put it more formally, rules are humanly devised constraints that shape political interaction. Institutions are here considered to be both formal, like for instance in a constitution, and can be enforced,

wad informal, i.e. they evolve over time and are respected ss a code of conduct by moot actors involved3.
The necessity to solve deepsealed societal conflicts and thus the need for effective 'rules' which enable a political consensus among contestants can be understood as follows: firstly, rules reduce uncertainty among the actors involved, hence they can act strategically (see Chapter 5 in this book); secondly, they provide room for exchange and compromises for those problems thai are aptly circumscribed by Ostrom (1990) as "common pool requirements". The options chosen or decided upon for political action to solve the problem (in whatever form, i.e. this also includes non-decisions and non-actions) is what we shall call policy formation. This process is equivalent to what others would call 'state-intervention' or the 'authoritative allocation of values in a society'. The actions of the state or the allocating agency are in this conceptualization of the 'political'viewed as relatively independent from societal interests (Skocpol, 1985:
45), That is to say: political action, i.e. the relation between politics and policy-making, requires a degree of autonomy in order to be feasible and effective. If this is not the case then the political process is merely 'ritual' and indeed simply a reflection of societal features and developments. In short, a theory of the political process must assume that there exists a mutual and interdependent relation between politics and society, and that its organization is to a large extent independent from society. The issue at hand is then to investigate to what extent and in what way this process can be observed and affects social and economic developments.
It should be kept in mind, that the triad of "polilics-potiiy-policy" in itself is not a theory of the political process. It is instead a heuristic device to delineate the 'political' from the 'non-political'. This description of the ^political', however, makes it possible to elaborate on the core subject ofcompa-

rative politics. That is to say that all those processes that can be defined by means of these three dimensions are worthy of our attention as long as the analysis requires comparison in order to explain the process. The next step therefore is to specify the unit of analysis for comparative purposes.
This is a more complex undertaking than it appears since, as Ragin (1987: 7-9) and Przeworski (1987: 2-4) correctly point out, the unit of analysis is not by definition the same as the properties under review, but may also indicate the theoretical categories that direct the research question at the same time. The term 'unit of analysis' can have two meanings therefore; on the one hand it signifies an elaboration of the theoretical argument, on the other hand it concerns the translation of the theory in properties to be observed empirically.
A number of comparative researchers have drawn attention to this confusing way of using the term 'unit of analysis', which easily leads to equating description with explanation. However, it is quite important to know exactly what is under discussion, if we wish to validate theoretical statements by means of empirical knowledge. Przeworski and Teune propose a distinction between 'levels of observation' and 'levels of analysis' (1970: 50), whereas Ragin introduces the terms 'observational unit' and 'explanatory unit' (Ragin, 1987: 8-9). Both these distinctions between respectively empirical knowledge and theoretical statements appear useful, but I still find them confusing.
Levels or units of observation and analysis may either appear to correspond, as is the case with, for example, studies of political parties and party systems, or in fact do not intersect (e.g. with respect to theories of the world system, where the explanatory variable, the system, by definition equals the observational units, i.e. the nation-states). However, in both these examples, both units are, in one way or another, supposed to be conceptually interlinked. Whatever name one gives it, it does not solve the problem at hand. The comparative analysis of the 'political' always involves a multi-level type of argument. Hence, it involves the observations of comparable parts that are considered to be an analytical whole. Of course, both Ragin and Przeworski realize this also, but instead of specifying the core subject of comparative politics, they resort to a methodological refinement as a solution. I therefore propose to consider the unit of analysis to encompass every topic that can he formulated in terms of ihe heuristic triad of politics, polity and policy. The unit of observation is then simply the operational izal ion of the 'triad'in comparative

perspective, making them belong to the field of comparative politics. That is, if (he question under review can only be explained in terms of mocro-scopical properties which vary from one system to another.
To give a concrete example; the study of the development of the welfare stale is not, by definition, a subject of comparative political research. In my view, it becomes a comparative topic only, if an attempt is made to explain this development by means of macro-political properties such as conflicting interests between economic classes. These conflicts are, depending on the existing features of the liberal democratic state, fought out and subsequently may result in a patterned variation of public policy-formation at the system-level of the state. Hence the core subject is not the welfare state, but instead the extent to which politics, polity and policy can be identified as properties of me political process. This being the case, the extent to which elements of this process are relevant, is explaining the political developmeni of the welfare state.
Alternatively, if one focusses on 'classes'it is not their existence perse that matters, but their degree of political action which may, for instance, explain the emergence, or the change of political institutions (e.g. the rule of universal suffrage, or the role of trade unions in the decision-making process). In the latter case, i.e. institutional change, the "political' is regarded as consequential, whereas in the first example, i.e. the development of the welfare state the 'political' is explanatory. In both examples the unit of analysis is expressed in terms of macro-scopical properties related to the political process, and measured or observed across systems or across time within a system at a level that is representing a coherent political unit. Politics, polity, and policy are thus in my view a set of theoretical references as well as an empirical point of reference, which form together the core subject of comparative politics.
Although here the core subject of comparative politics has been specified in relation to the 'political' in a society, it does not imply that we have o theory too. This implies that an appropriate (meta-)lheorelical approach should be related to the macro-scopical properties, which represent the features of politics, polity and policy. This means, in my view, that approaches which are basically formulated on the micro-level are less likely to be useful to explain political processes (like, for instance, economistic explanations of government formation; see: Lijphart, 1984:46-66). The same applies to 'grand theories' that depart from one overarching concept (like. for instance, political development;

see: Mayer, 1989: 64-74; Lane/Ersson, 1990: 90-126), or theories that are in fac based on socio-economic or socio-cultural features (see: Rokkan, 1970 Dogan/Pelassy, 1990: 3-44). The problem is namely that these macro-societal approaches cannot separate (he political from the non-political in then explanation of events, situations or developments.
Rather than looking for broad concepts or micro-based theories i appears to me that a theoretical approach in comparative politics should h. focusing on the interaction between political actors and institutions, in whici way this interaction influences a system's capacity to perform in accordance will the needs and demands of a society. A viable trajectory to follow is the politica economy of institutionalized behavior of interdependent actors in relation to th< political actions which are, in turn, molded by the political organization of th< slate in a society (see: Laver, 1986; Krasner, 1988; North. 1990; Scharpf, 1987 1991; Strom, 1990; All/Shepsle. 1990; Budge/Keman. 1990; Ostrom, 199( Czada/Windhoff-Heritier, 1991; Grafstein, 1991; Schmidt, 1991; Olsen. 1991 Keman, 1992a). What these authors have in common, is the idea that actor related behavior (i.e. 'politics') must be understood within the context o political institutions (i.e. 'polity'). In addition, thai the variation of the polic performance of a political system can to a large extent be accounted for h examining Ihe patterned interaction between the working of institutions and th. related room for manoeuver of political actors.
Hence the theory-guided question of comparative politics is to wha extent the 'political', the core subject of comparative politics, can indeed accoun* for, and is shaped by the political actions in one system compared to another It is this process and the attempts to explain it by systematic comparison tha distinguishes comparative politics from other fields in political science and it a the same time makes it a field within political science. Which brings us to lh-next question this esny seeks to answer: how and when to compare in a mctho dologically sound fashion?
2.5 The Meaning and Use of the Comparative Method
There IE little dispute about the comparative method being the most distinctly. feature of compntive politics. Yet. at the Mime time there has been a conti nuing debate about what, when, why and how lo compare (e.g. Lijphart, 1975;
Roberta, 1978; Dogan/Pelassy. 1990; Rueschemeyer et al., 1991). There is no need here to go into the comparative method as such (see; Przeworski/Teune. 1970; Merrit/Rokkan, 1966; Maycr, 1972). In view of the theoretical perspective developed in the previous section, I would instead like to focus on the extant methodological controversies which this debate provoked.
What to compare? Rather than focusing on 'macro-social', 'societal' or 'contextual* entities, it should now be clear thai I propose to study the 'political' m a society. This further implies that the conceptualization of 'politics, polity and policy* as a comparative tool is the major methodological concern. The social and economic configuration of a society, or situation is not the primary goal or meaning of comparison, instead capturing the specified differentia of the 'political' across situations and across time will have my concern.
By taking this point seriously, there are a number of implications for the controversies on the comparative method. It concerns questions like:
whether or not political science per se is comparative by nature (e.g. Almond, 1966; Lasswell, 1968; PrzeworskiTTeune, 1970; Lewis er al., 1978; Mayer, 1989),
whether or not the substance, i.e. the relationship between theory and reality, defines the correct approach w terms of case-studies, cross-sectional analysis, variable-oriented (often equated with statistics) research designs (e.g. Lijphart, 1971; 1975; Przeworski, 1987; Ragin, 1987).
whether or not comparisons are only meaningful by applying the longitudinal dimension and confine the number of relevant cases to analyze (e.g. O'Donnel, 1979; Castles, 1989; Bartolini. 1991; Ruesche-meyer ct al., 1991),
whether or not causal explanation is achieved by means of empirical-statistical corroboralion (Ragin, 1987; Lijphart, 1975; Holt/Turner, 1970; Smelser, 1976).
The first issue is more or less reminiscent of the transition from the 'behavioral' dominance in political science and its attempt to achieve 'scientific* status (Mayer, 1972). The comparative method was considered to be the ideal platform, if executed on the basis of statistical techniques of data-collection, variable construction and causal modelling, to reach this status (e.g. Holt/Turner, 1970).

This position strongly coincided with the search for a 'grand theory' of politics. Apart from the fact that for various reasons 'scienticism' in the social sciences has lost its appeal, it simply induced a situation in which we lost track of whal the substance, i.e. the 'political', is (Mayer. 1989: 56-57). Francis Castles (1987) has succinctly pointed out that "the major incongruity is not a matter of theory not fitting the facts, but of the facts fitting too many theories' (p. 198).
Thus, even if one knew what to compare, the refined techniques cannot really help us in deciding what is right or wrong, since what we often lack is a adequate theoretical perspective that is consciously elaborated in proper conceptualizations of the 'political'.
The latter point, i.e. the relation between conceptualization and operational izalion (Sartori, 1970; Lijphart, 1975; Scheuch, 1980; Przeworski. 1987) has been taken up since the early 1970s. Even if one thinks one knows what to compare, the question remains of how to translate it into proper terms for empirical research.
As I pointed out in the preceding section the use of the term 'unit ot analysis' is confusing, moreover I concluded that this issue cannot and should not be solved by means of a methodological point of view alone, but instead ought to be primarely formulated in terms of the core subject of comparative politics. However, this task remains unresolved by the definition of the core subject of the 'political'. It essentially means that one has to chose, on the basis of a topical research question that is formulated in terms of the 'political', the correct research design (see for this: Robertson, 1978; Schmidt, 1987; Kernan, 1988; Rueschemeyer el al., 1991). The question what to compare leads to the matter of how to compare.
Generally speaking, the 'logic'of comparative research goes back to the famous predicament of John Stuart Mill (1806-1873) which has led to the equally well-known distinction between the 'most similar" and the 'most different systems' design for comparing (Przeworski/Teune, 1970; 32). Most comparativists agree on this distinction, but differ on the question of whether or not the research design should be based on as many similar cases as possible, or upon a small number of dissimilar cases. The first approach leads easily to the so-called 'Gallons problem' (Lijphart, 1975: 171), that is to say; few cases, many variables, which make it difficult to arrive at conclusions of a causal nature. This is ultimately the result of'diffusion'(i.e. processes of historical learning), which

leads to spurious relationships, or to 'overdetermination* (i.e. even if cases are, to o large degree, similar, the remaining differences will be large due to the use of concepts that are too broadly operational ized) and in turn this situation will affect the relation between apparently independent variables and the dependent phenomenon (Przeworski/Teune, 1970; 34. Przeworski, 1987: 38-39). As far as I can see, there is, as of yet, no proper solution to this problem. It is by and large due to the dynamic nature of social reality, which cannot be captured by means of controlled experiments.
This problem, inherent to the substance of social sciences, is thus also an inevitable problem of comparative research of the 'political'. Rather than abandoning empirical research aimed at explanatory results altogether (like, for example, Maclntyre, 1971 concludes). Instead 1 suggest to take into account the limitations and constraints of the comparative method and to attempt to develop a research design that is suitable for the research question under review (see also: Ragin, 1987: 9-10).
Four issues with regard to method are to be observed when a research design is developed. Firstly, the context of what is compared; secondly, the role of 'time' with respect to the problem under scrutiny; thirdly, the level of inquiry, Le. the micro-macro link; and fourthly, the number of cases involved. These issues are equally important and decisions made upon them will have a great impact on the plausibility, validity and quality of the outcomes of a comparative research project.
Contextual variables are those variables that make up the environment of the core subject, i.e. of the 'political'. A 'most similar' design is intended to reduce them to the barest possible minimum by means of selecting cases (or:
systems) that are by and large identical, except for the relations between variables under review that represent the research question, i.e. what is to be explained. The debate on 'Does Politics Matter' which 1 referred to. or the analysis of the development of welfare stales that I mentioned, are both examples of how important the choice of cases is in relation to the analytical conclusions based on them. For instance, comparing non-democratic regimes with democratic ones, or 'young'democratic systems with long established ones could lead to conclusions of whether or not parties do play a role in political decision-making. However, such a comparison would not render information regarding about the question to what extent political parties in parliamentary

democracies do matter with respect to policy formation in general, or whether party differences are relevant for the development of welfare statism (Castles. 1982; von Beyme. 1985; Schmidt, 1987; Keman, 1990).
Moreover, the number of variables not controlled for in such a research design would be immense and would thus engender Gallon's problem. Obviously, this is an erroneous path to lake, although it is worth noting it ha,1 been suggested that the problem could be by-passed by using functional equivalents, i.e. assuming that particular actors and institutions in one systen are identical to certain phenomena in an other system. For instance the role o 'leaders' in the Third World and 'traditional' modes of decision-making couk' be compared with the ways party elites and concomitant bargaining take plact in a parliamentary democracy (Dogan/Pelassy. 1990: 37-43), However, given Iht already existing problems of comparing on the basis of similarity I do nol consider this a sound trajectory to follow. The number of contextual variable' must be tow in the eventuality of a research question that is akin to the mosi similar approach, and even in a most different design the number of 'contextual variables should not be excessive, otherwise one would end up with conclusions that everything is indeed different and all situations are peculiar by definition. It is the enduring paradox of Scylla and Charybdis, and there is no easy solution.
Much comparative research is characterized by a //me-dimension. Thi-poses a number of problems which are related to the consequentality of tinu itself, the number of cases that can be studied and Finally, the measurement o time in terms of variation (Grew, 1978: Flora. 1974; Bartolini. 1992). Bartolini notes that, surprisingly enough, the historical method is rarely disputed by social scientists. However, there exists a large body of literature within historiography that discusses .the complexities of temporal variance and its explanatory valm (Romein, 1971; Braudel, 1977; Althusser, 1983; Burke, 1978) and has revolve^ around the so-called 'Annales' school. Like the pervasive discussion withn social science on levels of measurement, i.e. the nucro-meso-macro linkages, the 'Annales' (and Braudel in particular) attempted to differentiate 'time* h;
distinguishing three levels of diachronic development: the long term (macro). the cyclical movement (meso), and the occurrence of events (micro). The lonj term development structures the other levels and makes it possible to relate events to cycles and. according to Braudel, events can be understood in theil proper historical context. In this way, it was claimed, objects of study, such a'

the political development of a society, can be compared as if they were synchronical. Yet, a problem remains that one is implicitly assuming that the structuration of time is a result of a few, universal factors (for instance, the impact of processes of'modernization'). Hence, time remains uni-dimensionally defined and is therefore potentially an overdetermining factor.
The alternative route, which is often advocated, is to incorporate time in the research-design by means of a case study design (Ahrams, 1982; Tilty, 1984; Skocpol, 1985; Ragin, 1991). Apart from the problem of the time dimension in relation to consequentality, another problem is contained in the conceptualization of the 'political': are we looking at the same phenomenon through time or at functional equivalents? In other words: is the development of ft political process captured over time, or does it concern the cross-time variation in a political phenomenon3?
As Bartolini correctly points out, there is no fundamental (or logical) difference between using a synchronical or a diachronical research-design. In both cases the comparativist has to grapple with the fact of whether or not the observed variation is part and parcel of both the independent and dependent variables. Hence, the so-called qualitative comparative case approach (Ragin, 1991; O'Donnel, 1979) which claims to be superior to the quantitative comparative spatial approach is wrong-footed, as long as its proponents do not supply us with a logical argument that the time dimension can only be applied in a comparative analysis based on case studies.
Again, I would instead suggest to reach a decision on the use of time in a research-design on the basis of the research-question. If one investigates, for instance, economic policy-making in OECD-countries, one may make use of a perioditotion which represents a similar incident or event in all the cases under review (see Chapter 8 in this book). If one wishes to analyze the development
.

of 'welfare statism" one could decide to use time-series analysis, whether or not it concerns one or more countries. A final example of choosing a research design may be the study of processes ofdemocratization ofa society. To do this on the basis ofa single case-study is very well justified and useful (e.g. Daalder, 1966;Rokkan, 1966), but a comparative investigation of this process diachroni-cally can be equally justified and useful (e.g. Lipset, 1963; Rueschemeyer et al. 1991). All these examples demonstrate that time can be explicitly taken into account with respect to the research-design. However, the choice in what wa) this is done has more to do with the research question than with a specific approach of including time into comparative research.
Contrary to the time dimension, as has been pointed out here, the problems with spatial analysis have been discussed at great length. Spatial analysis has to do with the level of measurement in relation to the selection o) cases under review. Lijphart (1971) in his seminal article, distinguishes three types of spatial analysis, namely: statistically based, case-oriented, and the comparable case approach (see also: Ragin, 1987; Rueschemeyer et al., 1991) Ragin, in particular, overstates the differences between the various methodological approaches. He develops a dichotomy that is separates the 'case-oriented' from the 'variable-related' research design. The first approach would enable the comparativist to analyze the 'political' more comprehensively than would be possible by means ofa 'few variables, many cases' approach. The latter method is, in Ragin's view, inferior to the 'comparable case' approach because the relationships observed are bound to be biased or 'overdetermined' as a result ol empirical indicators which are either too generally constructed or measured at a highly aggregated level. Hence conclusions based on (quantitative) broad comparative analyses are often only seemingly (statistically) robust and the causal interpretations not always as valid as is pretended they are (see also:
Robertson, 1978; Przeworski, 1987).
However valuable these insights may be and how important a reflection on these issues is, it concerns an argument, which is a false one. In my viev behind these arguments there is essentially a difference in epistimological tradition and even in 'belief systems' or paradigmata. The differences between research designs is often exaggerated and often not based on logical arguments. It is quality (i.e. historical knowledge) versus quantity (i.e. analytical empiricism), holistic explanations versus parsimonious modelling, interpreting

patterned diversity (e.g. on the basis of a 'most different' design) versus judging patterned variation (by means of" 'most similar' design), detailed knowledge of the cases versus theoretical knowledge from relations and so on (Ragin, 1987:
Ch. 2 and 3). Yet. is there really such a difference between the two approaches that warrants such strong views on the rights and wrongs of either approach? It is obvious, that I do not think this to be the case nor that it is necessary (see also: Rueschemeyer et al.. l99I:27ffon this point). Elsewhere I have attempted to clarify this point about applying the logic of comparison to a research question *s a means to develop a theory within the field of comparative politics:
"to construct a theory at all one has to simplify and generalize, rather than describe. There is no point in constructing a general explanation clogged up with minutiae of time and place. The purpose of a theory is to catch and specify general tendencies, even at the cost of not fitting all cases (hence one can check it only statistically, and it is no disproof to cite one or two counter-examples). The theory should, however, fit the majority of cases at least in a general way, and provide a sensible and above all an applicable starling-point for discussion of any particular situation, even one which in the end it turns out not to explain - here it can at any rate serve as the basis of a special analysis which shows which (presumably unique or idiosyncratic) factors prevent it from fitting.
A general theory of this kind serves the historian by providing him with an entry point and starling-ideas. These, we would argue, he always brings to the case anyway; with a validated theory he knows they are reasonably founded and has a context within which he can make comparisons with greater confidence. As we suggested at the outset, there is no inherent conflict between historical analysis and general theory. Each can, indeed must, be informed by the other and supplement the other's efforts. Theory is therefore a necessary simplification and generalization of particular motives and influences, not simply a restatement of them, though complete loss of contact with historical reality will render it too abstract and ultimately ilTelevant.'(Budge/Keman, 1990: 194)

2.6 Comparative Politics as a Distinctive Approach in Political Science
The main argument throughout this essay has been that the purpose of applying the comparative method in political science is to identify regularities regarding the relationship between societal and political actors, the accompanying processes of Jnstitutionalization of political life, and the societal change that emerged simultaneouly. In addition the logic of comparison is seen as the 'royal way' to establish theoretical and empirically refutable propositions that explain these regularities in terms of causality.
To this end. comparative politics as a distinctive approach selects and compares the'political'in a variety of different situations. Comparative analysis is considered to provide a greater opportunity to analyze a greater variety of political behavior and institutions, within and among political systems. Assuming, one knows what to compare and foremost why, the proper research design can be developed to allow for an analysis that accounts in a plausible way for the research question. Thus, it is not a matter of discovering a particular method that can be considered as quasi-paradigmatic for comparative politics.
There are a number of options open to the comparative researcher. In the process of selecting the correct research design one should be aware of the caveats and pitfalls of the several approaches that are in existence. Instead, comparative political science may be defined by the use of a particular core subject, i.e. the triad of polilics-polity'policy, which involves understanding of:
firstly, how concepts are derived from the 'political' in relation to the
research question posed;
secondly, how these concepts can be nude to "travel" from one system (understood here in relation to the unit of analysis) to another;
thirdly, how a set of units of observation can be developed within which systems may be compared and classified;
finally, how and when one compares similar and dissimilar systems. synchronically and/or diachronically; and how and when case-studies may contribute to more general theoretical knowledge.
The understanding of these 'rules' is vitally important for every student of comparative politics and distinguishes it from other fields within political science. They characterize the dislinctiveness of comparative politics. In the

56
following table the main distinguishing features, which have been discussed in this chapter, of comparative politics ssn distinctive fie\d are summarized (see for this also: Nohlen, 1983: 1083):
Core Subject: the political Level of Observation Time Dimension N. of Cases Contextual Variables
1. Politics, Polity & Policy Political System & Society SynchronicalDiachronical Many Few HeterogeneousHomogeneous
2. Politics & Polity Intra-System SynchronicalDiachronical Many Few Homogeneous Homogeneous
3. Polity & Policy Inter-System Synchronical Periodical Many Fewer Heterogeneous Homogeneous
4. Politics WorldSystem Diachronical One Heterogeneous

Of course, other classifications are also possible, but the main point is that a student of comparative politics leams how to develop his or her research design by systematically assessing which options are available on the basis of the research question under review. Such a research design must then be conceptualized in terms of the 'political' that is competent not only answering the specific question under review, but also enhances our (meta-)theorelical understanding of the political process, the 'core subject', for instance in'rational institutional ist1 terms.
Examples of comparative research which can be categorized within this framework are the cross-national analysis of political performance (core subject H\) throughout the world by Powell (1982), the comparative analysis of the politics of government formation (core subject #2), the development of the welfare stale (core subject #3), and the 'world system' approach (core subject )C4). In all these instances important choices were made to connect the research question with a research design. Budge and Keman, for instance, consciously choose to explain the process of government formation in terms of actors (i.e. political parties) in relation to their room for manoeuver due to existing modes of institutionalized behavior (the 'rules of the game').

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