INTRODUCTION
If those of us who began to write about political development some fifteen years ago had been fully aware of what was at the end of the . tail we held in our hands, we might have let it go. We knew that the/ existing body of political theory - our concepts of political structure and process and our classification schemes - was inadequate to cope with the problems of discriminating and explaining the varieties of political phenomena that began to dominate our attention in the 1950's. Our theoretical efforts took the form of improvisations. We "theorized" because in some sense we had to- In exploring new terrain we felt that a poor map was better than no map at all.
Our theory-building and modeling first took a simple dichotomous form. Working from the classic formulations of Max Weber (1947:324 ft.), Ferdinand Toennies (1940), and Talcott Parsons and Edward Shils (1951:
53 ft.), several innovative social scientists including David Apter (1955:
8ff.), Francis Sutton (1963:67 ff.), Lucian'Pye (1962:32 ff.), and Fred Riggs (1957:23 ff.), constructed ^nodels of traditional and modern forms of society and polity. All four drew on Parsons' "pattern-variable" explication of the differences between modernity and traditionallty in forms of culture and social structure. A structural-functionalist approach began to take shape, deriving from anthropological functional theory and the work of Talcott Parsons as elaborated by Marion Levy (1952:
149 ff.) and several associates (Levy et al., 1968:77). This approach was based on a set of logically derived functional "requisites" of any society. One could therefore compare societies, and particularly modern and tra-

diu'onal ones, by examining the ways in which different kinds of social structures performed these functions.
Using a somewhat different perspective some of those associated with the work of the Committee on Comparative Politics experimented with ways of discriminating among modern and nonmodern political systems, drawing on the Weber-Toennies-Parsons tradition, but building on the traditional functional concepts of political llieory, particularly separation of powers, and more recent political science research dealing with the emergence of the democratic political infrastructure of political parties, interest groups, and communication media. This effort culminated in the functional classification employed in tlie Politics of the Developing Areas (Almond and Coleman, 1960:3 ft-, 532 ff.) and in the classification scliemes proposed in James Coleman's concluding essay to that volume, and in Edward Shils's Political Development in the New States (I960). The concepts we elaborated and tlie classification scliemes that we experimented with were very definitely of an ad hoc sort. The new nations were terra incognita. We needed some preliminary mapping operation;
some way of guiding field research. Most of us engaged in this work were part theorists and part empirical researchers. We were not pure theorists. We simply had enougli sense to know that empirical observation by itself would be insufficient to guide us to the important phenomena and relationships of the newly emerging nations. We were trying to work our way out of an intellectual quandary, and innovating and experimenting seemed to be tlie only creative way of confronting the intellectual challenges which tlie new nations presented.
This ad hoc, experimental approach to theory seemed to some of us more profitable than engaging in efforts at wliat might very well have turned out to be premature theoretical formalizations at that stage of the game. Though we were concerned with political development and change from tlie very outset (tlie instability of the new nations made this inescapable) we were quite aware that tliese models and classification schemes which we were busy constructing in tlie late 1950's and early 1960's were not developmental or causal theories. We were also aware of the fact that we were comparing non-Western political systems according to Western categories and from a Western perspective. After all, we were Westerners, beginning with tlie knowledge and concerns of the West, trying to understand how tlie newly emerging or rapidly changing political systems of Asia, Africa, and Latin America were similar to or different from our more familiar institutional systems and processes.
If this early effort at conceptualization and classification represented our first efforts to come to grips with development theory, a more historical, longitudinal tlirust began in the late 1950's and early 1960's. Again we started with tlie relatively well-known Iiistorical experience of

iet A. ff>nd
the West with political development, attempting to draw from it a set of hypotheses that might enable us to explain the pressures, conflicts, and problems which the new nations were experiencing.
The logic of our undertaking was elementary- As the Western nations were in some sense modern, and the non-Western ones were in almost all cases not modern but seeking to become so, the historical experience of the modern nations had some relevance for our understanding of the problems and prospects of modernizing efforts among the new nations. We did not assume that the new nations would follow in older developmental "paths," but rather that the historical experience with modernization, and the various hypotheses suggested in Western historiography, might give us an initial grip on the developmental prospects of the countries of Asia, Africa, and Latin America.
Out of this historical speculation came the "crises" or "system development problems" hypothesis of developmental paths and their consequences. The political histories of Britain, France, Germany, Italy, and other countries in Europe contained many generalizations and explanations of the different historical paths taken by these countries in the last several hundred years. Thus historians have explained the intensity and violence of German and Italian nationalism in terms of the late arrival of these two countries to nationhood, and the coincidence of these efforts at nation-building with the simultaneous crises of social mobilization, participation, and welfare during industrialization. The contrast with such countries as Britain and France was obvious. These countries had attained settled borders and legitimate national identities long before the onset of the participation and welfare revolutions ot the industrial
period.
From the perspective ot their political history of the last three or tour
centuries it appeared that all the European nations had experienced common problems. Thus in the "age of absolutism" the central institutions of the emerging European states were formed, and penetrated the local structures and peripheries of tlieir societies. This centralizing process took form with important consequences for the subsequent development of these nations. Thus in Prussia centralization and bureaucratiza-tion were far more thoroughgoing than in Britain. These differences were presumed to explain the strikingly different experiences with de-mocratization in Germany and Britain. The French pattern of state-building again differed from both the Prussian and British patterns, and was similarly assumed to liave affected French democratization patterns. A second related problem had to do with forming a sense of national identity. Whether this national identity occurred early or late, whether it spread gradually from upper classes to middle classes to lower classes, or exploded, so to speak, as it did in France during the Frencli Revolution

and the Napoleonic period, whether it preceded or came after the industrial revolution, whether it encountered or failed to encounter resistant local, religious, and ethnolinguistic particularisms had serious consequences for the cohesion of these Western national societies, the structure of their party systems (Lipset and Rokkan, 1967:1 ff.), commitment to the nation, and compliance with state authority.
During Lhe nineteenth and twentieth centuries the common problems confronting Western European political systems have been participation and welfare. The countries that had more or less resolved their state- and nation-building problems encountered and solved their participation and welfare problems differently from those nations that carried state- and
nation-building problems into the nineteenth and twentieth centuries still significantly unresolved.
This system development theory took the form of the assertion that the relationships in contemporary political systems between central and local political organs, the homogeneity and heterogeneity of the political culture, the structure of the party and interest-group systems, the characteristics of the bureaucracy, and the kinds of public policies produced by these political systems, could be explained in part by the particular ways in which these common environmental challenges had affected the political system historically - the order in which they were experienced, their magnitude and intensity, their separate or simultaneous incidence, and the ways in which elite groups in these political systems responded
to these challenges (Almond, 1970:159 ff.; Pye, 1966:63 ff.; Binder et al., 1971).
These two thrusts in developmental theorizing, the first cross-sectional and classificatory, the second longitudinal and explanatory, reflect the principle directions of theoretical experimentation during the 1960's. As this literature grew and concepts and definitions proliferated, a polemic began to take shape at two levels: the first, concerned with differences in detail, classification schemes, and definitions of concepts; and the second,
concerned with more fundamental problems of causality, choice, and determinacy in political development.
The first polemic was largely internal, carried on principally among the political development theorists. It was concerned with definitions, trait lists, classifications, and assumptions. There are many different lists of modern or "developed" political traits, including the eight properties referred to by Ward and Rustow (1964:3 ff.), the threefold "development syndrome" first formulated by Coleman (1965:15 ff-), the criteria of political modernization suggested by LaPalombara (1963:39 ff.), the four measures of effective institutionalization listed by Huntington (1968:
12 ff.), and three characteristics of political modernization listed by

Claude Welch (1967:7), the criteria listed by Dankwart Rustow (1967:
35 ff.), David Apter (1965:43 ff.), and S. N. Eisenstadt (1968:256 ff.), and many others. This confusion in definitions and lists of traits, however, is more apparent than real. The lists overlap and the differences in the numbers of traits and the particular ones itemized are the consequence of different levels of generalization and kinds of emphasis. The generic themes of cultural secularization, structural differentiation and specialization, and increasing capacity or performance ran through all of them.
Classification schemes also differed. Almond and Powell (1966:213 ff.), Coleman (1960:532 ff.). Apter ("1965:24 ff.), Shils (I960), Dahl (1963:
25 ff.), Lijphart (1968), and others offered different ways of sorting out political systems. The differences were largely the result of the criteria used, whether the focus was on the non-Western areas, included European historical cases, and the like.
More serious was the debate about whether the definitions of development or modernization included democratization, or whether "unilinear," ideological, or ethnocentric assumptions were behind our developmental theories. Were the developed Communist countries less modern than the democratic countries, or simply modern in different ways? Also, did the concept of development imply indefinite progress along certain lines, and fail to take into account the occurrence and probability of "decay" or developmental breakdown as stressed by Huntington (1968:1 ff.) and Eisenstadt (1964, 1966), or the possibility of significant cultural and value innovations that might change tlie very content of modernization (Goulet, 1968:295 ff.)
SYSTEM FUNCT10NAUSM AS CAUSAL THEORY
More serious criticism of tliese intellectual efforts came from outside the circle of developmental theorists, from logicians and philosophers of science, from political sociologists, and from economists or political scientists applying economic models to political analysis. Perhaps the earliest and most cogent critique came from the philosopher and logician Carl Hempel (1959), who argued that the anthropological-sociological functionalism theory was a logically defective form of explanation. His critique was directed principally against the notion that all social traits and institutions were in some sense functional, or adaptive from the point of view of the society, the culture, or tlie social system as a whole. Functionality and adaptiveness are so loosely defined that any cultural trait may be said to contribute to them. Hempel argued that the more generic system-funclionalism, which simply affirmed the interdependence of the various components of social and political systems, was not a causal

explanation but simply a set of hypotheses about the "association" of social structures and patterns, hypotheses of a probabilistic sort, and not an explanatory theory (Cancian, 1968).
The logical critique of system functional ism as falling short of "explanation" is, on the whole, correct. Yet it misses the point of the importance of the introduction of the system metaphor into the social sciences and political science, which places the individual components of political processes in context (Parsons, 1968: vol. 15; 458). A reciprocal association between tlie elements of political systems and their environments is assumed-a statistical or probabilistic assumption about the interactions of political institutions and processes. If it falls short of explanation in the tight sense of experimental science then it suffers the difficulties encountered by all science "that nature is capricious at the quantum level" and that "The order of systems emerges as a statistical reality out of the disorder of particles" (Ingle:410).
In another important critique, Alvin Gouldner argued that the assumption of reciprocity among the components of social systems has been based too much on analogies with mechanical and biological systems where a tight meshing of parts and close interdependencies are typical, and where liomeostatic mechanisms clearly function to maintain the system at efficient or adaptive operating levels. These mechanical and biological analogies are criticized by Gouldner (1959:241 ff.) as being misleading when describing and explaining social systems. In social systems, he argues, components may operate in relative isolation, or the interactive adjustments may be long delayed, and less predictable.
System tunctionalism is a theoretical framework for contextual mapping. It requires the analyst to place any particular relationship in a larger context of association and meaning, even though the specific features of this context have not been fully proven or demonstrated. Thus, using a system-functional framework it is possible to explain in a preliminary way the immobilism and instability of the French Fourth oRepublic. Voting data and opinion surveys in the France of the 1950's suggest the existence of a fragmented and polarized political culture. This in turn may be related to an interest group subsystem dominated by political parlies or by the Catholic church and therefore unable to articulate the demands of the various functional groupings in French society in a pragmatic, negotiable form. In a similar way one may establish a relationship between the fragmented patterns of French political culture and a party system unable to form stable coalitions in the legislature and cabinet. This pattern of coalition instability and weakness may in turn be related to a pattern of political performance not effectively responsive to popular needs and demands, it also may be related to a

^l A.i
developmental pattern over time in which tliere was a high probability
of political system breakdown and revolution.
By contrast, in tlie British political system it was possible to relate a homogeneous and largely allegiant political culture to a subsystem of interest groups capable of effectively articulating demands, and a party system capable of aggregating these demands into alternatives of political leadership and public policy. This in turn may be related to a relatively neutral, politically subordinated bureaucracy, capable of implementing public policy decisions. In addition, the overall responsive pattern of the British political system may be related to a developmental pattern involving incremental-adaptive forms of change.
Preliminary political-system mapping operations of this kind, although not constituting explanation in the scientific sense, are essential parts of an explanatory strategy. The discovery of structural-cultural associations of these kinds generates important hypotheses for more rigorous investigation. Furthermore, it ensures that more rigorous analysis and research on specific relations will not lose sight of the larger contexts of causation and meaning. Rigorous studies of particular institutions or processes, such as voting, legislative processes, party activities, and bureaucratic decisions, can only be meaningful if they are located in the larger systems of which they are parts: larger systems that specify the variables
necessary for explanation and evaluation.
At the present our use of the system concept is metaphorical (Deutsch, 1963:3). We have taken it from the harder sciences where the probability of reciprocality is higher than in the social sciences and where the use of experimental methods makes lawlike generalizations possible. We need a concept of social and political system more empirically grounded and applicable to social processes. From this perspective much of the current logic-chopping about the system metaphor is a case of tilting with figures of speech (Mowse, 1966:607 ff.; Finer, 1969:5 ff.; A, James Gregor,
1971:575 ff.).
The "new political science" critiques of system functionalism as ideologically conservative make a useful point. The mechanical, cybernetic. and biological system models carry equilibrium and homeostatic assumptions that connote stability and persistence. Although societies have analogous processes - socialization and social and legal controls of one kind or another - the dynamic and confiictual aspects of social systems tend to be underestimated, even concealed by using these analogies. Yet the implied remedy of eliminating the system concept is worse tlian the problems with the present model. A good theory of development-incremental as well as revolutionary - that deals with cause and effect as it occurs in nature, will have to be systemic. Marxism with its inter-

relation of economic, stratification, and political variables is a good example- Studies of the common consequences of industrialization and social mobilization for culture, social structure, political mobilization, and political organization are another.
SOCIAL MOBILIZATION AS CAUSAL THEORY
The earlier work in political development theory was systemic and classificaiory. Although it was assumed that political systems existed in environmental settings and that their stability depended on stability in the environmental setting, relatively little was said about those settings or about the interactions of system and environment.
But as development theorists tried to explain the causes of political change, this intrasystem emphasis became increasingly untenable. Soon political development tfieorists were speaking about the crises that political systems encountered and that explained the ways in which they developed. They began to talk about the performance of the political system in its environment, about the capabilities, or capacity of political systems as these were affected by different sequences of environmental change, political system change, and performance change. The system development or crisis theory was a product of this effort at placing the political system in its environmental and historical context.
Tin's emphasis developed its own momentum and method in the 1950's and 60's. Some of those who contributed to this kind of theory were also system-functionalists of one sort or another. Thus Karl Deutsch (1963) and Daniel Lerner (1963) contributed to the system-functional literature as well as to the "social-mobilization" literature. What explains the fact that this social-mobilization school has tended to take off in its own direction was its increasing reliance on aggregate national-level quantitative data and increasingly sophisticated statistical analysis.
Though relating social structure and stratification to political structure and processes was a very ancient tradition in political theory, continuing all the way from classical theory up to Marx, Weber, Barrington Moore (1966) and Karl de Schweinitz (1964), the more recent efforts differed from the earlier formulations by stressing empirical quantitative analysis. The model of the approach is to be found in the early work of Karl Deutsch, Nationalism and Social Communication (1953), but perhaps the first full-scale developmental theory was Lerner's Passing of Traditional Society (1958). Lerner, like most associated with the social-mobilization aggregate statistical school of development theory, was concerned with the causes of "democratization" or tlie development of participant societies. Using a combination of aggregate quantitative data gathered for some seventy countries and an attitude survey administered in several

Middle Eastern countries, Lerner proposed a sequential pattern of modernization. This involved four phases: first, urbanization and industrialization; second, education and literacy; third, exposure to the mass media; and fourth, politicization or mobilization. This phasing of modernization was based not on longitudinal or historical investigation of sequences, but rather on the strength of the correlations among these
variables in contemporary nations.
Until quite recently this work has sought to derive developmental explanations from correlational analysis of contemporary nations. The assumption has been that if these processes are strongly correlated, they must have sequential causal connections. The logic of the sequence is derived from the strength of the correlations and the apparent relations of the processes themselves. Thus industrialization and urbanization are clearly related, and create the conditions for exposure to the mass media and the development of literacy. Through his attitude survey in the Middle Eastern countries, Lerner sought to strengthen his causal explanation of political participation by adding to the necessary conditions for participation, a sufficrent cause in the form of his "empathy" theory. As people were mobilized from their rural, traditional, relatively static settings to the dynamic industrial, literate, media-saturated urban environment, they had an increased capacity to identity with others, to imagine themselves improving their social statuses, and mastering their environments. This growing capacity for empathy, according to Lerner, lead to demands for political involvement and participation.
Karl Deutsch, who was the first to speak of social mobilization, commented about the process, "The first and main thing about social mobilization is, however, that it does assume a single underlying process of which particular indicators represent only particular aspects; that these indicators are correlated and to a limited extent interchangeable; and that this complex of processes of social change is significantly correlated with major changes in politics" (1961:495). Deutsch was concerned with roughly the same set of mobilization indicators that Lerner treated, and he gave similar stress to exposure to the mass media. Deutsch points out that social mobilization increases the potential level of political tensions or demands and "also brings about a change in the quality of politics by changing the range of human needs that impinge upon politics." This change in the range and intensity of human needs and demands increases political participation that may "express itself informally in greater numbers of people taking part in crowds and riots, in meetings and demonstrations, in strikes and uprisings; or less dramatically, as members of a growing host of organizations" (1961:496).
Although both Lerner and Deutsch were interested in explaining increased pressures for political participation, they tended to let the
analysis rest with the explanation of rising levels of political demand, without seeking to explain the institutionalization and stabilization of democracy. A second social-mobilization approach to developmental-explanation lias stressed (lie relationship between tliese variables and democratic stability. James S. Coleman (1960:532) classified the seventy-odd independent nations of Asia, Africa, and Latin America into three categories according to the degree of democratic competitiveness: competitive, semicompetitive, and noncompetitive. Cases were assigned to these classes judgmen tally, based on the more or less authoritative literature describing these systems. He discovered a strong association between level of socioeconomic development or modernization and these political characteristics. Thus on indicators of wealth, industrialization, urbanization, and education the countries with competitive political systems came out highest, the semicompetitive next highest, and the authoritarian countries lowest.
Lipset (1959:69 ff.) broke tliese countries down into two classes: stable democracies; and the unstable democracies, popular dictatorships, and elite-based dictatorships. He also found that with one or two exceptions the stable democracies came out higher on the indices of wealth, industrialization, education, and urbanization.
Two more statistically sophisticated analyses of these relationships followed the publication of Lipset's paper. The first, by Pliilips Cutright (1963:253 ff.), was essentially a criticism of Lipset's analysis, particularly its simple dichotomi/ation of political .systems into two major categories - stable democracies and all tlic ottiers- Cutright developed a measure (his "political development scale") with a score tor tlie number of years during the twenty-year period between 1940 and 1961 that each of seventy-six countries had had a competitive party situation in the legislature, and a popularly elected executive. Tlius lie offered a more explicit and continuous measure of democracy and stability than had Lipset's analysis. Cutright then related positions in his political development scale to such socioeconomic variables as educational level, urbanization, spread o? the communication media, and economic growth and labor force characteristics. He found that he could explain most variations in the level of democratic stability by covariation in tliese four independent variables. Like Deutscli and Lerner, he found that the communication variable had the strongest correlation. Cutriglit's analysis represented a refinement of the Lipset and Coleman hypothesis regarding the relationship between socioeconomic modernization and democratic stability.
Deane Neubauer (1967:1002 ff.), in turn, undertook to improve Cut-right's analysis. He argued that a more refined index of democratic political development shows that beyond a given threshold of democrati-zation, social mobilization variables lose their strength. He concerned

himself with the question of how to explain differences in degree ot democratic participation among countries that have already attained democratic stability. He confined his analysis to twenty-three relatively stable democracies, and used a more discriminating scale than Cut-right. Thus his democratic score is based on the percentage of the adult population eligible to vote, the ratio of popular votes tor political parties to tlie percentage of seats in representative bodies, the competitiveness of the communication media, and the degree to which the citizenry have multiple sources of information available to it, and the like. He found that at this level of democratic stability differences could no longer be explained by level of socioeconomic modernization, but by the historical and cultural characteristics of societies, and the accidents of leadership
and collective choice.
This social mobilization theory of political mobilization and participation is an excellent example of cumulativeness in research. Each piece of research takes a step in the direction of increased rigor and sophistication. Thus McCrone and Cnudde (1967:72 ff.) support the emphasis on exposure to communications as explaining the largest amount of variance in the dependent variable - political participation - through the
use of causal modeling.
The aggregate statistical school of development theory includes other tendencies than those treated here. Another group of researchers has been concerned with explaining the incidence of "internal war" or collective violence, and with the economic, social, political, and psychological conditions associated with this extreme form of political instability. Ted Gurr, the most sophisticated analyst of the causes of civil strife, demonstrated that the sense of relative deprivation (differences between expected and actual social benefits) as interred from a variety of social and political indicators, had the strongest causal relationship to the incidence and severity of civil violence in contemporary nations. Other variables that affected the level of civil violence but in less direct and substantial ways were the coercive capacity of the society, tlie extent and effectiveness of social institutionalization and organization, the legitimacy of the regime, and characteristics of social structure and political culture which facilitate internal violence (Gurr, 1968:1104 ff.).
The virtue of the social-mobilization aggregate statistical school of development theory is that it forced scholars working in this general field to confront explicitly and rigorously the problem of explaining how aspects of political systems change or develop in relation to changes in the social and economic environment. In addition it has stressed the importance of operationalization, particularly in finding quantitative indicators for political variables and their relationships, tlius moving political development studies toward empirically grounded theory.

However, this approach to development theory has several problems. First, it tends to treat politics as a dependent variable, as being caused by societal change, and not as an initiator, or causer of change. Second, by focusing on such specific political variables as participation, stability, or collective violence, its treatment of politics as a dependent variable is lacking in explanatory breadth. Treating politics as a system interacting with its environments both in a dependent and independent sense, and consisting itself of interacting structural-functional and cultural elements might have avoided these analytic distortions. Third, the international environment has been almost completely neglected as a factor influencing stability and change, a most serious omission indeed. Fourth, through the reliance primarily on cross-sectional contemporary data, causal inferences regarding- sequences in time are essentially heuristic and still require testing against historic time series.
The more recent work in one way or another seeks to overcome these shortcomings. Thus Gurr (1970) builds in more political system variables in his efforts to explain the level and type of civil violence. Zapf and Flora (1970) have begun to test and modify the Lerner-Deutsch social-mobilization theory against European time-series data. Brunner and Brewer (1971) have begun to place the social-mobilization approach in a system-theoretic context, thus treating political change both as consequence of socioeconomic change, and as a cause of sociopolitical change. They test their approach against a ten-year historical record (the 1950's) of these relationships in Turkey and the Pliilippines. By far the most interesting aspect of their work is their computer simulation of these variables for Turkey and the Pliilippines, and their computer experiments with manipulating the values of these variables. Though the possibilities of this operationalizing of system-environment interactions, and of using computer simulation as a way of experimenting with different kinds of transformations and public policies, is only sketched in the Brunner and Brewer book. it is ;the most promising product of this research tradition. It goes far toward combining the rigor of the social-mobilization approach with the scope and concern for context of the system-theoretic approach, and adds the possibility of experimentation
to test hypotheses and to explore the implications of differing public policies.
Ricliard Pride's study (1970) of the relations between patterns of social mobilization, parly system characteristics, and stable democracy is a good example of aggregate quantitative analysis of political development patterns that relies on historical time series data (Flanagan and Fogelman, 1968). The relations he establishes between the rate of social mobilization, the timing of democratization, the incremental character of party system development, and democratic stability are more convincing tests

of these hypotheses than those based on cross-sectional data. Flanagan and Fogelman (1970:1 ff.) promise to build in international variables more explicitly in their work on developmental patterns.
RATIONAL CHOICE THEORIES AND DEVELOPMENTAL CAUSATION
It seems evident that the system-functional and social-mobilization approaches to tlie explanation of political development ought to be combined. In other words, it is as possible to develop indicators of international threat and opportunity as it is to develop indicators of the various components of the so-called social-mobilization syndrome. Similarly it is quite possible to develop indicators for the interaction among the institutions and processes within the political system- the characteristics and operations of pressure groups, political parties, coalition-making in the legislature, and the organization and performance of the bureaucracy and courts. The rates, levels, and patterns of output or performance of the political system in the domestic and international environment can similarly be operation all zed and related to the other variables. The two approaches thus are complementary.
Their combination suggests the prospect of expressing political development over time as a set of related curves - curves reflecting the impact of the domestic and international environment on the political system;
curves reflecting the responses of the political system to these influences, and the related interactions within the political system; and curves reflecting changes in the activities of the political system in the domestic
and foreign environments. Brunner and Brewer (1971) hint at these possibilities as well as the difficulties.
But both approaches to developmental theory tend to be deterministic; that is, the choice or decision aspect of developmental processes tends to be treated in aggregate terms-volume of demands, voting rates, incidence of collective violence, or of governmental repression. Curves moving in the same direction, or in inverse directions, rising or falling;
by increments or suddenly "kinking" in patterned ways, give us statistical associations, but not explanations. The individual acts or events that add up to aggregate measures or that by themselves have causal value, are / acts of human decision and choice.
These system and social-mobilization approaches to developmental explanation have a mechanistic or organismic bias. Human choice, exchanges, bargains, and the frequent intervention into developmental processes of outstanding leaders without whom historical outcomes of great importance cannot be satisfactorily explained, are hardly the central perspective of system functionalism and aggregate statistical analysis of social processes and change. Tlius the attack on these two approaches

is justified, bin the attack on their deterministic quality comes from two different intellectual traditions. The first, coming from economic theory and from psychological learning theory, uses rational-choice models and assumptions in explaining (lie structural and developmental patterns of political systems; and the second, stemming from many sources including personality theory and historical insight, stresses the unusual quality of individual political leaders or the cultural patterns of political elites and groups.
As James Buchanan defines tlie rational choice approach, "political structure is conceived as something that emerges from the choice processes of individual participants. ... It involves the empirical judgment tliat political process can be 'factored down' to the level of individual choices" (1966:26-27), A systemic description establishes a set of associations, a structure of relationships. The explanation of this structure requires the explication of its choice and decision aspects. Buchanan is arguing that a systemic associational analysis of politics is incomplete unless it makes explicit the individual clioice basis of particular structures and patterns of interaction. Similarly Herbert Simon argues that the various functional processes sucli as "voting, legislating, adjudicating, and administrating have always been conceived as decision-making processes" (1966:15). But Simon also argues that rational choice is not the only relevant approach: "In identifying approaches to political research, one should not regard the various approaches as mutually exclusive, much less as antithetical" (1966:15).
More extreme claims for the rational-choice approach are made by John Harsanyi who rejects all systemic approaches, and who seems to suggest that social stability and change can be fully explained by individual rational choice (1969:513 tf.). But this extreme view does not seem to be sliared by any other contributors to this approach. Thus Peter Blau, in his elaboration of a sociological version of the rational-choice approach - "exchange theory" - as a way of explaining social institutions and processes argues that it "must be complemented by other theoretical principles focused on complex structures with institutionalized values" (1968, vol. 7:452; 1964).
An early demonstration of the uses of rational choice to explain political structure and process was Anthony Downs's Economic Theory of Democracy (1957). Downs applied the market model of economic theory to democratic voting and electoral processes, explaining some important characteristics of political party systems by referring the issue preferences and their distributions among voters and party leaders. He assumed that voters seek to maximize their rewards in terms of policy pay offs beneficial to them. Party leaders, on the oilier liand, wish to gain or to maintain power or office, and to do this they must get a certain proportion of

the votes, depending on the nature of the electoral system. Making different assumptions about how policy preferences are distributed among the voting population. Downs derived deductively three different kinds of party systems. Assuming a normal curve of preference distribution among the voters, Downs showed that the most rational strategy for party leaders would be to move in policy toward the mode of the curve where the great majority of the voters tended to concentrate. Thus the result would be, depending on the electoral system, a two-party or a multi-party coalition system, in which the issue distances between the major political competitors would be relatively small. Such a situation would contribute to a relatively peaceful and orderly alternation of competing parties. If the voter distribution preferences are bimodal (i.e., the voters tend to be polarized in two antagonistic political camps), a two-party or a multiparty system would tend to be unstable. Public policy would tend to be discontinuous and the probability of violence and revolution would be high. Downs suggested a third example in which voters' preferences were multimodal, witli voter concentrations on the right, on the left, and at several points in the center. Here he argues that the only possible majority coalitions to form governments would have to be made up of combinations of center parties. In view of the fragmentation of the party system under these circumstances, however, this kind of multimodal distribution of voters would be associated with political instability and stalemate.
Dahl, in his study of political oppositions in Western democracies (1966), tested some of these hypotheses in an analysis of party systems and electoral patterns, principally in Europe and North America. He concluded "that when voters' opinions are . . . unimodal, both the two-party system and the multi-party system are likely to lead to moderation and compromise among the leading parties. When, on the other hand, opinion is strongly polarized in a bimodal pattern, two parties each striving to retain the support of the extremists on its flank, will only exacerbate a conflict, and in multi-party systems will decline in votes and influence" (1966:376 ff.). In confronting party systems and voters' preferences in modern democracies, however, Dahl found it necessary to use more complex assumptions about the distribution of opinions on political questions among voters in different democratic countries. It could not be assumed that there was only one summary distribution o( voter preferences, as many distributions of such preferences as there were salient issues would have to be included. The same voters might be conservative on religious and social issues, liberal on economic issues, and moderate or indifferent on other questions. The real complexity of preference distributions among voters would have different consequences for the structure and composition of the party system, and tlie "coalition-

ability'* of its parts. Similarly, the salience or intensity of voter preferences would have important effects on the structure and process of the party system. Nevertheless both these deductive and empirical efforts have clearly demonstrated the importance for political theory of factoring down structural-functional patterns to individual choice behavior,
or working up from individual choice patterns to structural and process patterns.
Several others have illustrated the uses of rational-choice theories for analyzing political process and change (Gamson, 1961; Riker, 1962;
Groennings, Kelley, and Leiserson, 1970; Browne, 1971:39Iff.). Leiser-son predicted the distribution of cabinet posts among (lie factions of the Liberal Democratic party of Japan using a modified version of games theory (1968:770 ft.). In an analysis of Latin American politics in coalition-theory terms, Eldon Kenworthy (1970:130 ff.) both illuminates aspects of Latin American politics and suggests useful ways of adapting coalition theory to the special complexities of third-world countries. Other examples in Groennings, Kelley, and Leiserson (1970) suggest the advantages of using coalition-theory concepts in empirical political analysis. Tlie capacity of several versions of formal coalition theory to predict real coalition outcomes in European countries proved to be quite limited, though the Leiserson version, which builds in some realistic assumptions, proved to be more useful than the Gamson and Riker versions (Browne, 1971:391 ff.).
One of the principal innovators in this work concludes quite modestly i that the application of rational-choice models and, in particular, mathematical games and coalition theory, is just in its beginnings (Riker, 1968:2; 529). Perhaps some difficulties arise because the formal-mathematical model of games and coalitions requires conditions rarely if ever approximated in political reality. In most political conflicts the actors or players are changing, the roles are asymmetrical, the utilities of the players or their preferences are complex and changing, the values of their political resources are uncertain and changing, and the information available to the political contenders is incomplete and not equally distributed. To apply these rational-choice theories to politics and political development will require some sacrifice of rigor and simplicity-which seems to be the trend of some recent work. There is no intrinsic reason why the interactions of voters and party leaders, interest group and party leaders, legislative blocs, interest groups and bureaucrats, bureaucrats and citizens, and the rest cannot be factored down into individual decisions, bargains and exchanges, or coalition patterns. Ilchman and
Uphoff (1969) and Mitchell and Mitchell (1969) suggest some of these possibilities in their recent work.

LEADERSHIP AND DEVELOPMENTAL CAUSATION
The most recent contender in the political development literature is
the leadership approach; its principal protagonist is Dankwart Rustow. After pointing to the shortcomings of the functional, cultural, and decision-making approaches to developmental theory, he argues, "The leader as a figure omnipresent in any political process, as the maker of decisions, originator and recipient of messages, performer of functions, wielder of power, and creator or operator of institutions can bring these disparate elements into a single, visible focus. The study of leadership, moreover, can readily be supplemented with an examination of the social and political organization that he founds and transforms, with an analysis of the psychological appeals and political sanctions that give leader and organization a hold on their mass following. In short these may be the elements for a new theoretical view, both comprehensive and dynamic, of the political process as a whole" (1970:7).
Rustow is a bit uncertain on the place of leadership in the theory of political development. At some points he seems to be treating it as the most important element of developmental causation, and at others he places it in a broader strategy of developmental explanation, but he is quite unequivocal in stressing its contemporary importance: "In short, in a world embarked on rapid technological change and involved in an unprecedented degree of global interdependence, the call tor leadership has been continual and ubiquitous. Still, there has not yet developed an impressive or sophisticated theory of political leadership-whether in democracies or under dictatorship, whether in old nations or new" (1967:136).
None of its recent advocates argue that leadership is the only meaningful approach to development theory. As Robert Tucker (1965:574) states:
"We do not face a choice between explaining history by reference to leader-personalities or assigning them no importance at all." But despite the obvious importance of the leadership' variable in developmental explanation and its growing literature, no systematic effort has yet been made to place it in a broader strategy of developmental explanation. Wriggins (1969) makes a valuable first movement in this direction, but is mainly concerned with the kinds of strategies that leaders in new nations may employ to maintain themselves in power.
Rustow has brought together a valuable collection of case studies illustrating the importance of leadership in political development and other areas of social innovation. Rustow's own contribution to this collection, "Ataturk as Founder of a State" (1970:208 ft.), the interpretation of De Gaulle contributed by Stanley and Inge Hoffmann (1970:248 ff.), and Henry Kissinger's illuminating analysis of Bismarck's role in the formation of Germany (1970:317 (T.) are good illustrations of the importance of leadersllip in shaping and transforming political systems. Of course, there are many studies illustrating llie importance of leadership in developmental causation. The Georges' study o( Woodrow Wilson (1956) is an excellent example of the impact of a leader personality on foreign policy. Perhaps the most celebrated recent products of this literature are Erikson's full-length studies of Luther (1958) and Gandhi (1969).
Relating the leadersllip variable to other approaches to developmental explanation raises many questions. It is quite evident that in many significant historical episodes leadership in its more dramatic manifestations simply doesn't appear. Thus the Reform Act of 1832 in Britain, although of obvious importance for British political development, did not produce a dominant leader, or unusual leader personalities. The Meiji Restoration usually is not associated with the names of particular leaders who were responsible for this fundamental transformation of the Japanese political system. In Britain, particularly in the last hundred years or so, the existence of a kind of "counter-leadership" mechanism might be argued. Thus Ostrogorski's [ear that the extraparliamentary party organization would overwhelm the British parliament toward the end of tlie nineteenth century proved unfounded as the ambitious Randolph Churchill had his career cut short when he sought to dominate the Parliamentary party. Even more recently, R. A. Butler, who did so much after the Conservative defeat following the end of World War II to renovate and modernize the Conservative party organization, found that his very innovativeness was counted against him in his struggle for party leadership. Similarly in modern Japan a kind of "antileadership" cultural mechanism has persisted.
In many cases of significant developmental outcomes leadership appears somewhat less dramatically than Rustow's formulation suggests. But in many other cases it is just that kind of dramatic individual performance which explains political innovation. The Lenins and Stalins, the Napoleons and De Gaulles, the Bismarcks and Hitlers, the Ataturks, Nyereres, and Cardenases all are obvious demonstrations of the importance of individual revolutionary, charismatic, and reconstitutive leaders in the explanation of political change. Perhaps the confusion arises because of the fact tliat in the current defense of the leadership approach to developmental explanation the focus is on the leader as the "causer," and not what is being caused - the developmental or innovative outcome.
Once the leadership variable is separated from the dependent variable that it is supposed to explain, the theoretical problem becomes a bit

clearer. Thus in some situations significant outcomes follow upon more anonymous kinds of political processes than the leadership school usually considers. But even what precedes significant changes in political systems may not take the form of salient political leadership behavior, surely individual and collective decisions in all cases precede such innovative outcomes, just as they precede more everyday, recurrent outcomes.
Clearly, then, leadership as an aspect of developmental causation is a species of a larger category of causation that may perhaps be appropriately called "decision making." Those decisions which lead to significant changes may indeed be attributable predominantly to a single individual, or they may be attributable to a group or several groups or individuals engaged in bargaining. But whether they are predominantly individual or collective they have to be viewed as human decisions, bar.
gains, or choices.
Furthermore, the leadership school of developmental theorists includes
those aspects of choice behavior which are not rational, but which are rooted in culture and personality - culturally conditioned belief systems, operational codes, and the like. personality qualities of temperament, strength or weakness of character, special gifts, or talents. In other words, the leadership approach combines rational and nonrational aspects of human choices, whereas the rational-choice approach assumes rational actors making choices with full information about the preferences and
resources of other actors.
Thus the leadership school can be criticized for its failure to place
high salience leadership under a larger category of choice phenomena that includes the collective and individual, the rational and non-rational, as subcategories. In other words, the rational and nonrational aspects of choice behavior-whether individual or collective-are complementary aspects of the human decision side of developmental causation, just as the social-mobilization and system-functional approaches are complementary aspects of the determined environmental constraints and opportunity aspects of developmental causation.
Leadership therefore is important in the strategy of developmental explanation, not as a substitute tor other forms of developmental causation, but as an essential approach. One may even go further and accept Rustow's argument that individual political leadership is particularly important for explaining significant transformations of political systems. Other approaches to causation, such as the conception of interdependence and reciprocal causation of system-functional theory, the issue-forming and resource-redistributing consequences of social mobilization, and the coalition formation of rational-clioice and coalition theory, may be viewed as constraints or opportunities that limit or define tlie options of leaders. The unusual and innovative leader is the individual who dis-

covers or creates new options, mobilizes and combines new and old resources in creative ways, or arrives at policy formulations that change the issue distances among the competing political actors.
THE LOGIC OF DEVELOPMENTAL EXPLANATION
It is quite evident that all four of the approaches to developmental causation that we have discussed have something to contribute to the theory of political development. We might have followed the example of the early Mao and settled for a kind of "Hundred Flowers" eclecticism, but a connective logic slowly began to suggest itself-
It seemed self-evident that if each approach gets at an aspect of developmental causation, it must be possible to relate and cumulate each aspect of causation with the others in a strategy which increases the strength and specificity of our explanations of varieties of political change. Our first insight was that these four approaches in their contemporary polemical formulation tend to sort out in a fourfold table with determinacy and choice on the rows and stability and development on the columns. Table 1.1 makes this simple point.
TABLE I.I. Approaches to Developmental Causation Stability

Although system-functional theories attempt to deal with change, development, and growth, they tend to be more concerned with the explanation of persistence and system maintenance. Change tends to be treated as a movement from one equilibrium to another (Parsons and Shils, 1951:107). And although values and personality enter into the model, choice and decision are not brought into clear focus as catalysts and shapers of developmental outcomes.
The social mobilization school, on the other hand, falls rather neatly into the determinacy-development cell. Though both Deutsch and Ler-ner include psychological variables, at; least in this part of their work, they hardly spell out the "choosing," "deciding" aspect of development. Deutsch comes closer to such a view in his cybernetic system model which emphasizes "steering," but at best here we are dealing with a kind of computer program witli the creative programmer and his "wet-ware"

left out. In his international relations theories, Deutsch quite selfconsciously adopts a decision, bargaining, coalition-theoretic frame of reference. Nevertheless the principal thrust of the social mobilization school is this explanation of development in deterministic terms.
The various approaches to choice and decision explanations of political development, whether they are based on rational-choice assumptions or emphasize empirically discoverable propensities for choice and derision - whether they take the form of market, games, or coalition theories; or political culture, subculture, role culture, operational codes, or belief systems - tend to explain the persistence of structural and functional patterns- Hence they may most appropriately be located in the choice-stability cell, though they, too, have attempted to explain development. The leadership school on the other hand quite clearly falls into the choice-development cell.
A second step in discovering the logic of the interrelationships of these four approaches to developmental explanation was recognizing that historical development has an aspect of continuity as well as change, and an aspect of determinacy as well as decision and choice. No matter how complete and thoroughgoing a revolution may be, no scholar would argue that nothing persists of the antecedent state of the society or the political system. Scholarly interpretation of the most violent and trans-formative revolutions of our time-such as the Bolshevik Revolution in Russia or the Nazi revolution in Germany-treats these developments in terms of continuity and change, determinacy and choice. No self-respecting historian would offer an interpretation of the Russian Revolution that would leave out the roles of Lenin and Stalin on the one hand, or Russian economic and social backwardness and defeat in war on the other. Similarly, in the case of German political development, no self-respecting historian would omit the German defeat in World War I and the depression on the one hand, or Adolf Hitler on the other. Similarly, any historian dealing with modern Russian and German political development would emphasize environmental, structural, and cultural features that carry right through from prerevolutionary to postrevolu-tionary days.
These thoughts are humbling. Some very powerful minds in the behavioral sciences wrestling with problems of fundamental importance appear to have come up with extraordinarily one-sided causal explanations, which any self-respecting historian would have discarded without hesitation.
It would, of course, be an exaggeration to argue that all the contributions to political development theory can be sorted out neatly under these four approaches. Huntington (1970, 1971), Deutsch (1961. 1963, 1970), Apter (1971), Rustow (1970) and others have made eclectic use of

these approaclies. Nevertheless, each approach has its polemical advocates, and the literature of political development lias a tendency to move off
in tliese monocausal directions, isolated by their methods and conceptual vocabularies.
TAKING THE HISTORICAL CURE
Tlie exercise out of which this book emerged started in an impulse to "return to historical nature." The polemic was resounding all about us, and disillusionment was rife (Verba, 1967). The contrast with the mood that characterized this field of interest at the time of its beginnings some fifteen years earlier was striking. Tlie search for development tlieory had begun in the same era that produced the Peace Corps. Many naive hopes and expectations of the end of ideology and of revolution in the West, and the discovery of low-cost high-yield approaches to modernization in tlie new nations had been dashed in the interval.
Our search for a cure in liistory now took a more modest, empirically grounded, form. The logic of our inquiry was simple. Since the development that we were seeking to explain occurred in history, why not select several historical episodes, examine tliem in great detail, try out our varieties of developmental explanation, and see how they fit? What uses could we make of system-functional tlieory, of social-mobilization theory, of rational-choice and coalition theory, of leadership theory? The process of selecting and carrying out our studies of these historical episodes has continued for more than a three-year period, and lias involved a constant, moving from theorizing to empirical study and back again.
Of all the general approaclies to developmental explanation that we had encountered in development tlieory, Albert Hirschman's seemed to have the best "fit." His Strategy of Economic Development (1958) was an attack on the general trend of the theory of economic development, arguing that the searcli for the primum mobile of developmental causation was fruitless, that any one candidate for "first cause" might trigger off a developmental impulse and introduce a sequence of opportunities tor change in a modernizing direction. His criticism of "balanced growth" and argument that development occurs as a chain of disequi-libria contributed to our concept of "crisis" in the developmental process. His treatment of developmental policy making in coalition-theoretic terms (1965:327 ff.) represented a fruitful combination of system-functional and rational-choice explanation. And his conception of developmental entrepreneurship and its problems in underdeveloped countries recognized the importance of leadership in the economic sphere. His eclecticism and skepticism regarding total solutions encour-

need us to seek a similar open tlieory of developmental explanation
(1971:1-39).
Although the purpose of our research was to draw directly from history, we wanted to avoid being drawn into the particularities of history. Each significant historical episode has its own library of historical monographs, and one might proceed almost indefinitely to illuminate aspects of such episodes depending on the questions one asks- Our purpose was essentially theoretical. We wanted to develop and try out a framework of explanation that would enable us to compare and contrast the causes and consequences of different kinds of historical sequences. A satisfactory theory of political development would specify the important variables necessary for developmental explanation, and provide an analytic structure that would enable us to discover how different constellations oE these variables produce different outcomes. A theory of development should make it possible for us to compare across developmental episodes, and sliould help us avoid sinking into the particularities and uniquenesses of history. We had no quarrel with the ideographic propensities of history. On the contrary, such an approach is evidently important, and from a theoretic view as well- Studies of individual cases in depth serve as a check on the tendency o? theory to abstract, simplify,
and distort historical reality.
The case studies included in this book are a "selection" and not a "sample." Because our work was exploratory and concerned with the varieties of ways in which significant political changes are caused, we were interested in selecting cases that might illustrate this variety. The British Reform Act o( 1832 (Chapter III) is generally viewed as the exemplar of incremental democratization, a largely peaceful adaptation of a political system to basic changes in economy and social structure. The crisis of the British party system of 1931 (Chapter IV) is generally seen as another example of incremental adaptation, but one that brings out some of the costs of this accommodative form of democratic crisis management. The events leading to the formation of the French Third Republic (Chapter V) and the Weimar Republic of Germany (Chapter VI) illuminate patterns and sequences that contribute to democratic instability, immobilism, and breakdown.
The Cardenas phase of the Mexican Revolution (Chapter VII) and the Meiji Restoration in Japan (Chapter VIII) are two contrasting cases ot political modernization in preindustrial societies. Tlie Meiji Restoration succeeded in suppressing and postponing the political mobilization commonly associated with industrialization and social mobilization. On the other hand, the Cardenas phase of the Mexican Revolution deliberately 261 in motion a political mobilization and organized institutions capable

of containing and controlling it. The two strategies had sharply different consequences for the processes of democrat! zation in Japan and Mexico. In Japan the party system was weakly rooted in the population and collapsed under the pressure of authoritarian forces. Jn Mexico the political mobilization of workers and peasants led to the co-optative, controlled pluralistic structure of the Party of the Institutionalized Revolution-a coalition of the whole given to political stagnation with intermittent spurts of conflict and decision.
Tlie Indian crisis of the 1960's (Chapter IX) is a limiting case-one in which such environmental "crises" as famine, international war, and internal violence occur but fail to cumulate in such a way as to produce a major structural change- The focus is on the interplay and counterpoint of different kinds of environmental impacts, and the bargaining
skill and propensities of political leadership, in producing a system-maintenance outcome.
All seven cases contribute to our understanding of the processes and problems of democratization, as shall be discussed in Chapter X. These historical cases provide abundant illustration of the varying ways in which important political outcomes have been conditioned and caused. We shall consider our purposes well served if our work puts to rest efforts at monocausal explanation, remedies the neglect of the crucial role of international events in tlie shaping of polities, and redresses the deterministic imbalance in our developmental explanations. Who can view the events leading to the formation of the Third Republic, the Weimar Republic, and the formation of the Meiji regime, and fail to appreciate the powerful impact on political development of the international system? Wlio can view the Cardenas episode, or the events leading to the French catastrophes of 1870, and fail to register the crucial role of leadership and decision? Finally, who can view any of these historical cases and still cling to a single approach-however elegant,
however parsimonious-and still claim to contribute to an explanatory theory of political development?
TRANSFORMING HISTORICAL EPISODES INTO ANALYTICAL EPISODES
The first framework we used in our early efforts to analyze these episodes was rather simple and temporal. Most episodes we examined seemed to fall into a pattern, beginning with a relatively stable system, through a period of change occurring outside the system and within the system, and then culminating through a series of linked changes to a consequent system. Each episode was therefore divided into five time periods; (1) the preexistent system, (2) environmental changes, (3)

"within-system" changes, (4) linked changes, and (5) the resulta'i: 6"-tern. As our analysis and comparison of episodes continued, wf ^.furi'S. that this temporal notation served the useful function of keeping our analysis empirically grounded, in other words, it prevented us from taking off prematurely into theoretical abstractions. But as time went on our temporal categories began to fade into analytic categories.
It appeared at first that a different approach to developmental causation was appropriate for each temporal phase in our episodes. Thus in describing the preexistent and subsequent state of the political system, a system-functional approach seemed to fit. In the second phase, which stressed environmental impacts on the stability of the preexistent system, we felt that the social-mobilization approach was appropriate but modified to include international environmental impacts. In the third phase of our developmental episodes it appeared that choice and decision approaches were appropriate, both coalition-theoretic analyses and leadership theory. This enabled us to handle the problem-solving aspects of developmental episodes as the system went into crisis as a consequence of environmental fluctuations. In the fourth phase, we used a kind of "moving equilibrium" system-functional approach to the linked sequence of changes that followed the crisis resolution of the preceding phase.
For a long time we persisted in this mixed analytic-temporal frame of reference. In a sense our fingers had been burned by earlier theoretical experimentation, and we clung to history in an effort to avoid theoretical traps. With increasing confidence we moved toward the analytic frame of reference summarized in Table 1.2. The first and fifth columns of Table 1.2 show that the analysis of the antecedent and the consequent political system requires the application of all four approaches to developmental causation. This reflects our ultimate realization of the obvious point that any theory which seeks to explain change or development must be the same theory that explains stability. There are the recurrent patterns of system-functionalism in stability; there is system-environment interaction in stability; there is decision and coalition-making in stability;
and there is leadership in stability. Furthermore our original model of developmental episodes implying a movement from equilibrium through disequilibrium to equilibrium again simply does not correspond to historical reality. In some of our cases we begin with an antecedent po-htical system that is quite unstable, and we end up with systems that can hardly be called stable, as for example the Weimar and Third Republics.
But whatever the shape of the developmental curve might be in our individual cases, the analysis of developmental causation requires specifying the properties of the preexistent and the resultant states of the

Gabriel A. Almond ^7

system. And the description of these properties requires that we apply
all tour approaches to developmental causation. That is, we must look
at the Wilhelminian Reich, the Tokugawa Shogunate, the regime of
Napoleon III, the Mexican Maximato under Ptutarco Calles, the "old
Unity" of pre-Reform Act Britain, and the like, in terms ot its decision,
exchange, coalition, and leadership properties, its structural-functional
characteristics, and its system-environmental-interaction patterns.
It became increasingly clear to those of us participating in this intel-
lectual enterprise that as we were groping toward a more adequate way
of explaining developmental causation, we were really fumbling with a
more adequate way of approaching political theory in the very broadest
sense. If our premonitions are correct the concern that began with e?-
torts at developmental explanation some ten or fifteen years ago, cul-
minating in the polemic regarding developmental causation of the
current period, now shows the way toward a general theory of politics;
one that enables us to deal in a more balanced way with varieties of
political systems, aspects of choice and constraint which affect their
characteristics, and their propensities tor persistence and for change.
Table 1.2 suggests that in the contrast between the antecedent politi-
cal system described under column I and the consequent political sys-
tern described under column V, we have what the logicians refer to as
the "explanandum" or the theoretical dependent variable that our analy-
sis is supposed to explain. And columns II, III, and IV list what the
logicians refer to as the "explanans." More specifically it is a grid or a
framework of variables that can enable us to pick the particular constel-
lation explaining the differences between the state of the system at the
outset of the developmental episode and the state of the system at. its
culmination.
'"^ne antecedent and the consequent state of the political system have
to be described in structural-functional, in system-environmental-inter-
action, in decision-coalition, and leadership terms. These four different
^y8 0^ looking at the political system are complementary, and the logic
of this complementarity needs to be spelled out in general terms and in
our specific cases, that is to say, the structural-functional characteristics of
the political system have to be related to environmental characteristics.
And these structural-functional-performance properties of the political
system have to be related to the typical decision patterns and tendencies.
Table 1.2 shows that the transformation of the system may be explained
^y Ganges in the values of exogenous (environmental) and endogenous
('ntrasystem) variables. We are not suggesting here that the process of
change is only set in motion as a consequence of changes in environ-
mental variables. They may indeed, as we will show later, be set in mo-
tlon through changes in the values and properties of endogenous

variables (such as political leadership). We want to show in Table 1.2 that what began as temporal or historical episodes have taken their final form as essentially analytic formulations in which historic time is converted into clianges in the values and properties of our variables.
EXOGENOUS CAUSATION
Table 1.3 shows how we handle exogenous causation in developmental episodes, using a similar approach to that of the social-mobilization school of developmental theory except that we systematically include the international as well as the domestic social environment- We graph over time various indicators of social structural and cultural change, and international system change. We also distinguish between such long-term trends in the domestic and international environments as industrialization, urbanization, communication, literacy, and education, and short-term changes that may trigger political developmental sequences. In other words, the relationship between social and international change and politics is not of a direct one-to-one order. Rather, it appears that social and international change may continue for a long period and only begin to trigger cliange in tlie political system when a short-term kink or set of kinks occurs in the curve or curves. Consider a set of curves over time reflecting increases in the proportion of the labor force engaged in industrial pursuits, in the literacy rates, and the like. There may be considerable periods during which such changes in the values of these social environmental variables may be unaccompanied by changes in the level of demand being brought to bear on the political system. A bad crop year or years, or an industrial depression, or both,

mav suddenly trigger off a substantial increase in the level of demand and may also complicate the demand structure by introducing new kinds of issues. We call these short-term fluctuations in the environment "accelerators" or "decelerators," for the short-term fluctuations may work both ways. Thus good crop years and increases in industrial productivity may be short-term influences that reduce and simplify the level and structure of political demand.
Under each heading of long-term and short-term trends we specify the domestic and international environments. Thus long-term changes in the international environment may constitute either threats or opportunities to the elite of the political system. But in international environmental change, just as in domestic environmental change, it may more frequently be the case that some short-run international development, some outbreak of war, some defeat, some victory, triggers the political system change. Thus, for example, under some circumstances a high or increasing international threat level culminating-in war may decelerate political demand. Patriotic values may become uppermost and dampen down social and political conflict. There are of course other circumstances in which increases in the level of international threat, may sharpen political conflict, or have other effects on the political system. Here, we simply wish to make the point that international or domestic environmental change may be of a long-term character or of the shorter-term kind, and that it is a curve or set of curves of change with short-term "kinks" that produce the crises that usually accompany political development.
It should be evident by now that in Tables 1.3 and 1,4 we have stretched across the rows the relations that appear in columns I and V of Table 1.2. In our analyses of the antecedent and consequent political systems we describe and explain not only the structural and functional properties of the political system itself, but the patterns of input and output, and the decision, coalition, and leadership patterns with which they are associated. In Tables 1.3 and 1.4 we want to make it possible to specify the particular sequence of transformation that results in the political system in column V of Table 1.2. Changes in system-environmental properties are specified as the independent variables in Table 1.3 dealing with exogenous causation. Changes in decision and leadership properties are specified as the independent variables in Table 1.4. It is these changes in decision and coalition properties which trigger the linked sequence that results in the consequent system.
A change sequence of the sort schematized in our tables may be initiated externally or internally. Thus some significant break in continu-^y in the domestic or international environment or both may trigger a developmental sequence, or some significant internal transformation

may trigger the sequence, for example through the emergence of a new political leader or some breakdown in decision and coalition patterns. In our exposition, the fact that we present Table 1-3 first, treating exogenous causation, does not mean that all sequences of cliange necessarily
must be initiated in the environment of tlie system ratlier than within the system itself.
To return to Table 1.3, long-term and short-term international and domestic changes effect change in the political system via two sets of dependent variables-changes in the structure and composition of political demand, and changes in the distribution of political resources. As changes in social or international structure or botli occur, the rates and the composition of political demand are affected. Thus industrialization transforms the class structure and may significantly increase the demand for political participation on the parts of tlie middle class and the emerging working class. A slowing of tlie rate of economic growth or an economic depression may bring a sudden increase in the demand for welfare. Similarly, clianges in tlie international environment may complicate, intensify, or reduce tlie intensity of political demands.
Clianges of these kinds may also be viewed in their effects on the magnitude and the distribution of political resources available to political elites botli within and outside of the political system. To illustrate, we may have a situation in which tlie index of social mobilization increases and is converted into an increase in demands for political participation. Those members of tlie working class wlio become politically mobilized enter into and affect the distribution of political re. sources among tlie various political actors. Workers even without the franchise may come to the support of the more liberal elements operating within tlie political system, or indeed they may be recruited to the support of these elites. Whether the newly mobilized workers enter into resource calculations as voters, as strikers, as demonstrators, or as rebels, tlie resource balance previously characteristic of the political system lias
been altered and new opportunities for political maneuvering and coalition making have been created.

Gabriel A. Almond

ENDOGENOUS CAUSATION
Table 1.3 provides an analytic bridge to Table 1.4, which moves us into the problem-solving aspect of developmental episodes. The dependent variables in Table 1-3 become converted into the independent variables in Table 1.4. Changes in the structure and composition of political demand have now become converted into the positions or preferences of the political actors in relation to these issues: the changes in the magnitude and kinds o? political resources that have become available and their distribution botli within and outside the political system now become converted into tlie resources available to, or competed for by, the
various political actors.
With the new sets o? preferences or issue positions of the political actors and tlie resources available to them, we can generate a set of logically possible coalitions and specify their policy propensities. And if we assume rationality and perfect information we can predict the winning coalition. But particularly in times of crisis, both elements of coalition making - preferences and resources - are quite uncertain. To explain why a particular coalition and policy outcome occurred, we must rely on imperfect information and the nonrational aspects ot the decision process, either individual or collective. In Table 1.4 we refer to the decision and coalition propensities of actors. Previous bargaining and coalition patterns may be "sticky." We cannot make the assumption of complete alliance mobility. In addition, in situations of this sort a skillful and powerful leader or set of leaders may mobilize latent resources and affect the resource balance, or may manipulate preferences and issue distances in such a way as to significantly affect the coalition-ability of actors. A charismatic leader may emerge and convert "charisma hunger" into a powerful political movement.
Table 1.4 indicates that the explanation of the winning coalition and policy outcome requires an analysis of the preferences of the actors, and their resources - e.g., their votes, their seat's, their positions in the bureaucracy, or their control over instruments of coercion. It we were to assume that each elite actor can freely form a coalition with every other actor, and it we were to assume that perfect information is available regarding the issue positions of tlie various actors and the value of their resources, we could spell out, as the coalition theorists do, the set of logically possible coalitions specifying tlie preference or policy characteristics of these various coalitions, and their ranking according to the value of their resources. Though tlie winning coalition will be in the top few of this set of logically possible coalitions we cannot predict it simply on the basis of this information.

Dependent variables

The reason for [his is clear. Information, particularly in crises, is not fully available to all the actors; issue distances tend to fluctuate in crisis situations; and the value of resources and their distribution also tend to be uncertain, particularly in crisis situations. Longer term, habitual bargaining, and coalition propensities must be taken into account and leadership decisions may radically affect preferences and resources. In situations in which the intensity of demand is high the crucial or salient resource may suddenly turn out to be violence rather than votes. A sudden shift in tile principal resource from votes to guns may radically upset the expectations of actors. A skillful leader sensitive to mass moods may sense the sliift in the rules of the game, and in the relative values of different kinds of political resources. Similarily, a skillful and imaginative leader may discover the "policy package" and symbolism capable of mobilizing effective support. Thus we must have among our independent variables in this problem-solving phase of the developmental episode, the decision-bargaining-coalition propensities of both group and individual political actors. It will take all three of the independent variables listed in Table 1.4 to explain the winning coalition and policy outcomes of the developmental episode.
Great leaders are great coalition makers. We are not the first to stress this point. To say that the leader is a coalition maker is to specify the elements of leadership in such terms that they are congruent with coalition theory. Thus preferences and resources are the elements that set limits on the kinds of coalitions which are possible. But insofar as preferences and resources are not fully fixed - as they are in games - leadership may have certain ranges of maneuver, opportunities for affecting preferences and resources. Surely leadership implies an ability to sense latent preferences, reorder the priorities of followers, accentuate or reduce their intensities, and the like. Leadership also implies a capacity to impute relatively correct values to resources, sense the point at which and the extent to which the legitimacy of a regime is being built up or drained away, and to mobilize and bring to bear upon politics latent resources that other political actors may neglect or fail to appreciate.
Bismarck's greatness lay in that he, almost alone among the Prussian elite, could entertain the idea of breaking the traditional coalition with Austria, and cultivating a detente with France. Having disposed of Austria, he could then turn against France. In his internal policy Bismarck was able to place the liberal constitutional impulses of the German middle classes into conflict with their nationalism, and to bring them into coalition with the traditional authoritarian elites who, at an earlier time, had been their enemies in politics. Cardenas, hand-picked for (lie Mexican presidency by Calles with the intention of continuing the exploitative and authoritarian regime of the Maximato, knew that

the disappointed and frustrated Mexican peasantry held latent and powerful grievances and demands. He knew how to articulate these demands, how to mobilize the aroused peasantry and bring them into coalition with the urban working class, and thus outmaneuver the military regime. Not every significant developmental episode has a Bismarck or a Cardenas but all contain a leadership variable of some sort-individual or collective, dramatic or pedestrian. Any theoretical framework intended to help us explain the onset, tlie course, and the outcome of significant historical developmental episodes should alert the analyst to leadership phenomena.
DEVELOPMENTAL LINKAGES
The outcome of developmental episodes usually involves a slow, linked denouement. This sequence of linked developments is described in Table 1,5 in terms of our four principal approaches. The coalition and policy outcomes described in Table 1.4 generate disturbances throughout the political system and its environments. Having developed the logic of the preexistent system and the sequence of disturbances culminating in the new coalition and policy outcomes, we now have to follow these disturbances through environmental-system linkages, structural-functional linkages, coalition linkages, and leadership linkages.
For example, consider a developmental episode such as the British Reform Act of 1832. A Whig-Radical coalition forms the cabinet and enacts electoral legislation eliminating many of the "rotten boroughs," and lowering and standardizing suffrage requirements. In the short run, antisystem pressure is reduced, but in the longer run the introduction of electoral reform triggers demands for further extensions of the suffrage to enfranchise the working class, and for welfare legislation. Public policy in the next decade or two alternates between welfare measures intended to alleviate working conditions, the lowering of food prices by eliminating agricultural protection, and repressive measures.

The structural-functional consequences or linkages with these coalition-policy changes gradually transform the party system, the governmental bureaucracy, and tlie cabinet system. Tlius the increase in the suffrage is linked witli the development of more effective and differentiated extraparliamentary party organization, which then is linked witli increasing party discipline in the House of Commons. In turn, cabinets become more clearly responsible to and dependent upon their majorities in the House of Commons. Tlie enactment of welfare legislation and the need to control popular disorder lead to the emergence of a welfare inspectorate and specialized police forces.
The coalition linkages set in motion through these changes lead to a gradual setting aside of tlie old Whig aristocracy, which in the course of subsequent decades gradually shifts into tlie Conservative party. As the old Whigs pass out of tlie picture, middle-class leaders with the demo-gogic and organizational skills essential witli mass suffrage come to the fore. Despite the persistence of aristocrats in tlie Conservative party in the mid-nineteenth century, tlie distinctive new leadership patterns of British politics are innovative, sucli as Joseph Cliamberlain, Gladstone, Disraeli, and Randolpli Churchill represent. The changes that the Reform Act have triggered take some thirty years to settle down into a more or less stable system of interaction, witli the party, cabinet, and modern bureaucratic system emerging during the second half of the nineteenth century.
Our case studies follow these linked developments in the careers of the Third Republic, tlie Weimar Republic, the Meiji constitution, the interwar Japanese autlioritarian regime, the Cardenas phase of the Mexican Revolution, and others. Our analysis of linked development will help us identify the significance of our developmental episode, which is not fully contained in tlie main innovative coalition and policy outcome specified in Table 1.4. Tlius the formation of the Frencli Third Republic is significant only when we can link it with the typically immobilist. politically polarized, and unstable coalition and leadership system that obtained in France in the late nineteenth century and the decades prior to World War II. Tlie linkage analysis is more speculative, arguing hypothetical causality from tlie association of structural, environmental, coalition, and leadership patterns. To test these hypotheses about the relations among these linked aspects of development would call for the execution of other developmental episodes in detail. Thus, for example, it would be essential to make similarly detailed case studies of the Reform Act of 1867 in Britain, or perhaps of tlie Leon Bourgeois regime of 1895 in France to test tlie hypotheses regarding linked developments formulated in our analyses of tlie causes and consequences of the Reform Act of 1832 and tlie formation of the Third Republic in tlie

CONVERTIBILITY AND THRESHOLDS
i
It we are to move constructively beyond the current theoretical polemic regarding developmental causation and explanation, we must be able to establish that these four approaches are convertible, one into the other - that they are indeed parts of a general theory o? developmental causation, as our explanation of the analytic framework used in studying our historical episodes indicates. We have shown that it is possible to convert changes in social and international structure into changes in the level, the intensity, and the composition of political demand. If the class structure of a society changes substantially, and if the level of threat or of opportunity in the international environment fluctuates significantly, the kinds of demands and pressures being brought to bear on the political elites are bound to change. We can test these hypotheses regarding the conversion of structural clianges of these sorts on tlie structure of political demand by measuring and relating changes in demand to changes in the characteristics of the social and international environment. Thus we can convert environmental changes of different kinds to
political system pressures of related kinds.
But we cannot do this with the certainty that characterizes "hard"
science. We cannot establish a "political boiling point," even one that varies with altitude. The "boiling points" of political systems vary substantially with the structure and culture of the system. In some that are high on responsiveness and low on repressiveness, the conversion point may be quite low. In others, characterized by a structure and culture of repression and compliance, the "boiling point" may be very high indeed. Nevertheless, for any system, important environmental changes will affect changes in the level and composition of political demand.
Similarly, changes in the social and international structure environing the political system will affect the magnitude and kinds of political re. sources available in that system. Urbanization, industrialization, increasing literacy, and communication - in other words the social-mobilization syndrome-create a high probability that new social groupings will become available for political mobilization. Developments of this sort shake up the resource pattern on which the governing coalition has been based. Thus we can convert from changes in the level and composition of political demand, and the magnitude and kinds of political resources, to the destabilization of the preexisting coalition pattern.
But even though we can convert from demand and resource curves to coalition phenomena, we liave a problem again with the notion of a threshold. Political systems differ in the speed witli which changes in demands and resources are converted into coalition clianges. Presumably,

in a democratic system the conversion point would be relatively low. Coalition changes would occur through electoral processes reflecting these shifts in demands and resources. Though this notion of the low-convertibility thresholds of democratic systems has some value, we know from bitter experience that social pressures and extralegal as well as legal repression of various kinds can depress the level of political demand and can devalue political resources. Nevertheless, in authoritarian systems the conversion threshold from changes in demands and resources to changes in coalition and policy outcomes is higher, and of a different sort, than occurs in a democratic system.
Finally, we can convert leadership aspects of development into coalition and policy outcomes. The level and composition of political demand and the magnitude and distribution of political resources can be converted into a set of logically possible coalitions with different policy propensities. This is done by following changes in the level and content of political demands through to the policy preferences of different elite actors; just as we can follow changes in the magnitude and relative importance of different kinds of political resources through to the distribution of the resources among the different elite groups. Calculations of the logically possible combinations of actors based upon their issue orientations and their resources will tell us something about the relative cohesiveness, strength, and policy propensities of these possible combinations. But lack of information among the actors, and the volatility of demands and resources, particularly in crises, makes it impossible for us to predict the winner, though we can pick the possible winners. To explain the winning coalition we have to examine the decision, bargaining, and coalition-making tendencies of individual and collective leadership. The properties of leadership and decision, whether of a collective or an individual sort, can be converted into the coalition-making process. Leadership can be translated into skills-energy, imagination, etc.- in handling preferences and resources.
Conversion and threshold among these variables deserve a far more detailed analysis. We merely wish to elaborate these points sufficiently to make it clear that we now have to move beyond the level of theoretical polemic, and recognize that all four approaches are parts of a common logic of explanation of political stability and development,
MANIPULATING THE SCENARIOS
A great advantage of this comparative approach to political history is that over time and as we compare from one episode to another, we begin to develop a certain virtuosity, an ability to draw out hypotheses other than those exemplified in the episode itself. We can, in other

words, rewrite the scenario. It in a specific case we ascert^-.i-L dia.? the _",". coalition changes were brought about by a particular co'ni.i-uun]^ .it' -? international and domestic trends, we can assume a difffrfciv, r-^-isi^-'la-tion, and then speculate with some discipline as to ho- oJi;s diiiercai;
constellation might have affected the coalition and poh'-y process 'Jf, having spelled out other possible coalitions than the one thai. ai:cuc'ly was victorious, we can speculate about what might have happened in h" event of the victory of other possible coalitions.
In doing this we shall not be departing too far from the speculations of historians. Thus it has been pointed out that the Whigs in 1831 might have sought a coalition with the Tories organized around the policy of repression, which was one possible coalition and policy change that the British political elites considered as violence erupted in Britain prior to the Reform Act. Similarly historians have speculated that the Russian Revolution might have followed a different course it Kerenski had been less committed to fully maintaining Russia's commitments to the Allied alliance, and had not embarked on the fatal June offensive. There has been speculation that the French Revolutionary elite might have followed a course that would have enabled it to include the lower clergy in their coalition; it this had been done the full alienation of the Catholic church might have been avoided and a more moderate revolutionary course followed. Others have speculated that a more skillful and resourceful Leon Bourgeois a century later might have been able to overcome Senate opposition to his program of welfare and fiscal reforms, thus avoiding the alienation of a substantial part of the French working class which flawed French democracy throughout the Third Republic.
Who knows but that more resolute liberal bargainers in Prussia and the other German states in the mid-nineteenth century might have made a better deal with Bismarck, establishing the Reichstag and the political parties on a more powerful basis vis-a-vis the monarchy, the army, and the bureaucracy, thereby giving a stronger footing to a civilian and democratizing Germany? Or again in 1918 the German Social Democratic elite might have moved toward the 'left rather than toward the discredited military-bureaucratic establishment, clipped the powers of the old regime, and founded a more viable Weimar Republic.
All these scenario rewritings can be found in the speculations of historians and other observers. A more systematic comparative historical approach may enable us to do this simulative experimentation more thoroughly and with better controls. We would do this not in the "Monday-morning quarterbacking" mood of the unhappy exile allocating blame and responsibility, but in the mood of the ttieorist concerned with extracting the maximum knowledge about developmental processes from
historical experience, or in the mood of the statesman seeking to 4earn
from the past how to ascertain the constraints and tile opportunities of the present. We do this in a limited sort of way with our episodes, but surely more can be done with this "experimental" approach to history. The first steps are already being taken toward the computer simulation of political systems and developmental processes (Brunner and Brewer, 1971; Bonilla et al., 1971). The great promise here is that as we learn how to approximate more closely the real complexity of political processes, the experimenter may play the role of policy maker in a developmental game, and test the costs and benefits of alternative public policies. Coming from the side of historical reality, the case studies in this book may help in decisions as to how to develop such models, make them more realistic, lay out the variables that might be included, and develop the logic of their interaction.

.

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