to be resolved in .1 compromise between the two positions. Even the most stable kinds of political situations contained conflicting tendencies and potentialities. Thus we liad to describe our before-and-after situation by using recurrent interactions and processes that tended to maintain tlic system, and ttie connictual tendencies and potentialities tliat gave us insight into possible directions of systemic change and adaptation. Similarly, in the postcrisis stage in our case studies-even though the level and heat of the crisis had subsided-we have also to indicate tlie con-flictual tendencies and potentialities that poiniecl toward future crises.
When it came to applying social mobilization theory in our case studies it immediately became apparent tliai we had to complicate this way of treating the impact of the environment on the political system in several ways. We bad to distinguish between domestic and international environmental impacts on politics. We also bad to distinguish between security or military and economic impingement from the international environment, as well as sucli political-psychological occurrences as "demonstration effect." .Similarly, we had to distinguish between long- and short-term impingements, as some of our case studies demonstrated that long-term social changes might accumulate without significantly affecting levels of political demand. Short-term fluctuations seemed to be the immediate triggers for political crises, whether military or economic. It was necessary to complicate this environment-system interaction analysis by pointing out that environmental fluctuations can work both ways, that is, precipitate crises or allay them. Thus we assimilated social mobilization theory as developed in that body of literature to a much more complex set of theoretical variables that separated environments, discriminated long-term from short-term impingements, and lield open the
question of the direction of environmental impacts, i.e., crisis-inducing or crisis-reducing.
In applying coalition theory to tlie analysis of internal political changes and adjustments, we confronted tlie problem of turning a formal and simple model into an analytic scheme that could effectively be applied to historical reality. Tlie problem of loosening up coalition theory, separating its elements into actors, resources, and issues, and developing ways of measuring resources and issue distances turned out to be tlie most intellectually challenging undertaking of tlie whole research program. When we had finally worked these problems out at least to our partial satisfaction we discovered that this did not enable us to explain the coalition and policy outcome, but only to generate the set of logically possible coalitions from given sets of actors, resources, and issues, and to rank them according to tlieir resources, cohesion, ;ind policy propensities. To predict tlie winning coalition we liad 10 turn to leadership and decision phenomena, as well a.s 10 chiiiice.
When we finally confronted ilic leadership variable as an approach to

developmental causation we derided that we could only deal impres-sionisiically with it. The reader will therefore observe that leadership phenomena tend to be treated residually in our case studies. We decided that dealing with leadership more systematically would call for another major research undertaking. Thus we placed this issue on our agenda for future research and make no claims of having done anything more than to specify ways in which one could handle the leadership variable in such a way as to link it systematically with tlie other three approaches to developmental causation. However leadership phenomena clearly manifest themselves in developmental processes both individually and collectively, and positively and negatively. Thus our overall strategy of research might be best described as one of trial and error, and what we are offering ow readers in this book is an opportunity to share in this exploratory process.
We offer three sets of tentative conclusions in this final chapter. The first is a cross-case-study analysis of the causal constellations and sequences which our particular set of case studies manifests. Second our case studies all in some way suggest explanations for differing patterns of modernization and democratization. Third, and perhaps most important, this experience takes us a substantial way forward in showing us how to draw systematically from historic experience in generating better theories of tlie causes and conditions of political stability and change.
CAUSAL CONSTELLATIONS IN OUR CASE STUDIES
Eacli study we have presented manifests a different constellation of causal components. Eacli case study, however, assesses the value of these causal components even wlien it approaches zero, when its absence may be a part of die explanation of tlie outcome. Thus it is possible in our conclusion to compare causal constellations across all our case studies.
STRUCTURAL-FUNCTIONALISM AND CAUSATION
For each case presented, we can characterize the antecedent system in terms of its system-maintenance and crisis propensities, its "stabilities" and "instabilities." In all tliese cases the preexistent system includes structural or cultural features rendering aspects of it unacceptable to one or another of tlie contenders, which enables us to identify a potential for crisis, in tlie event of an increase in distance on issues or a wider distribution of political resources. This, it now seems obvious, is universal in political systems. There are no "final solutions," only more or less stable adjustments to more or less stable environments. Challenge is inevitable, and we liave only to locale (lie weak points that wilt be tlie first objects of attack. But in some cases, tlie preexisting system manifested little evidence of overt cliallcnge to tlie authority structure. It required
substantial environmental pressures to shake up tlie system. In other cases tlie conditions of crisis were obviously present in the preexisting situation, as was evidence of political mobilization against the ruling xoalition on certain issues. Rut tlie-crisis was rontnined, eitlier through -- the effect of some highly salient "valence" issue (e.g., threat in tlie international environment), or by tcni|X)rarily effective measures of repression or accommodation.
Our range of cases includes examples of tlie successful resolution of difficult problems below tlie crisis level - whether that level is defined in terms of our quantitative measures or by the criterion of whether or not the "constitution" itself became an issue. Britain in 1832, and especially in 1931, resolved serious questions below tlie crisis level as we have defined it. as did Mexico and India. Specific manifestations of structural and cultural features lhat were to affect tlie system's capability for non-crisis response to tlie substantive issues that became salient in each case examined Iiere are sliown in Table X.I.
In Table X.I we have ordered our cases according to tlie polarization index, that is, tlie degree of crisis as measured by our authors on the scale described in Chapter II. A score of 1 to 2 was an estimate of stability; a score of 2 to 3 estimated manageable within-system conflicts; a score of 3 to 4, crisis and confrontation; and 4 to 5, violence and revolution- Tlie range of difference here is quite substantial-from one extreme of tlie scale for France after the defeat in the war of 1870 and during tlie Paris Commune, to tlie other extreme for India.
Tile two cases highest in polarization - France of 1871 and Germany of 1920- were lowest in the estimate of political system legitimacy before the crisis, and were most dependent on effective system performance, taking tlie form of some combination of accommodation, distribution of welfare benefits, and repression of dissent (see columns 1 and 5, Table IX.1). Tlie countries lowest in (lie polari/aiion index (e.g., Britain in 1931 and India) were higher in tlie estimate of system legitimacy- The four cases in between are both at midpoint in the polarization index and in tlie extent of system legitimacy.
When we discuss the developmental outcomes in our cases later in this chapter it will he apparent that tlie liiglier (lie legitimacy of the preexisting system, tlie more limited was tlie extent of structural and cultural change in [lie outcome. Thus India and Britain (1931) came out of their crises with only moderate changes in political structure. Mexico and Britain (1832) weathered tlieir crises through changes in electoral arrangements, |X)litical reorganization, and changes in public policy. In both Japanese cases the constitutional order changed, and in the German and French cases llie changes were revolutionary, i
TIlis first comparison across our cases allows us to distinguish two

models of political development: first, tlic gradual, limited, and perhaps "insiifUcient" response o[ luglily legitimate systems; and seconil, the more extreme adjustments characteristic of systems relying more heavily on short-term effectiveness of performance.
Table X.I also indicates that the ".simultaneity" or "cumulativeness of crisis" hypothesis, which holds tlial the number of salient issue areas and (lie severity of ilie crisis are related (Almond and Powell, 1966:315 ft.), does not liold up well on tlie basis of our case studies. The third column of Table X.I suggests a more complex pattern. India, lowest on the polarization index, lias a twofold crisis of nation-building (language crisis) and welfare, but the issues were separated sliglitly in time and were cross-cutting rather llian cumulative. Mexico had a threefold crisis of welfare, participation, and nation building. However, the salience of these tliree issue areas was relatively low on the eve of the Cardenas era. Germany and France had multiple crises, they tended to be salient issues, and they tended to be cumulative rather than cross-cutting. Before the Meiji crisis Japan had only a state-building (centralization-bureaucrati-zation) problem, but it readied a higlier point in polarization than Mexico with three issue areas. The Japanese issue of state-building, however, constituted a tar greater threat to the structure of the Tokugawa constitution, whereas Mexico cannot really be said to have been confronted by a systemic crisis during the Calles era, though she was troubled by multiple issues.
Severity of crisis, therefore, seems to be explained by some combination of salience, multiplicity, and cumulativeness of issues, so tliat the same political elites and support groups are polari/ed on one highly salient issue or combination of issues. Tlie key is salience and intensity of antagonism, rather than number of issues. Multiple issues can accentuate polarization or reduce it, depending on tlieir incidence,
ENVIRONMENTAL FLUCTUATIONS AND DEVELOPMENTAL CAUSATION
Jn Table X.2 we compare tlie environmental impacts that occurred in our historical case studies. The numbers in tlie cells, ranging from 0 to 3, are judgmental estimates of magnitude of tlie impact of external pressures on tlie stability of the system. Tlie sign tells us whether the impact is stabilizing (+) or destabilizing (-). Tlie columns broadly divide these impacts by environment (international and domestic) and type of impact (military, economic, demonstration effect, social mobilization). Thus in a very summary way we can differentiate among our cases according to tlie magnitude and composition of tlie environmental aspects of causation.
But before examining tlie individual cases reported in the rows of Table X.2, the columns indicate something about tlie different ways

environmental change can enter into developmental causal sequences. By far the largest numbers are in tlie international military column, suggesting that at least in lliese cases war or threat of war lias been [lie most important political system destabilizer or stabiliser. Tlie first column of Table X.2 shows some of the various ways this military variable can affect stability. Thus. in Britain in 1832. a decline in military threat after tlie Napoleonic wars unsettled the British polity. In France and Germany, it was catastrophic military defeat; in tlie Meiji restoration it was tlie threat of invasion and colonization by the Western powers that triggered revolutionary change. In Japan in 1930-36 it was llie opportunity for expansion in China tliat united (lie Japanese behind militaristic goals, contributing to the fall of parly government but stabilizing the polity under military control. In India in tlie 1960's tlie military threat of China and Pakistan reduced internal cleavages in Indian politics.
Thus the interchanges between tlie international security environment and political systems would seem to liave been the most powerful exogenous variable in explaining system stability and change. The international economic environment enters into tlie picture only in tlie British crisis of tlie 1930's, when a generally deteriorating international trade situation peaked into a severe balance of payments crisis, and in the Indian case, when tlie reduction of foreign aid intensified famine risks but pressed politicians toward a new production policy; and indirectly in the second Japanese case, when the collapse in foreign trade set off a depression. Tlie third column, international demonstration effects, has four entries: (1) in tlie first British case, when tlie 1830 French revolution had a mobilizing impact on Britisli opinion; (2) in Germany in 1918, wlien the bolshevik revolution contributed to the polarization of German politics; (3) in mid-nineteenth century Japan, when tlie invasion and colonization of China dramatized tlie foreign tlireat; (4) and in Japan in tlie 1930's, when tlie international trend toward military and totalitarian regimes seemed to suggest an effective answer to tlie country's problems.
On tlie domestic side almost all entries are negative, and all eight cases contain declining economic performance components (depression, unemployment, food shortages, and famine), wliereas only six of tlie eight contain negative entries for social mobilization. The exceptions are the Mexican and Indian cases. Nevertheless tlie social mobilization variables played a role in tliese two crises as well. In tlie Mexican case tlie enormous social turbulence of the earlier pliases of llie Mexican revolution had created expectations and attitudes among Mexican peasants and workers that Cardenas could mobili/e. In India tlie relatively low level of social mobilization and the continued liold of such traditional social structures as caste, village, and ethnic groups gave tlie Indian elites a

deeree of bargaining freedom that enabled them to form ad hoc problem-solving coalitions and thereby avoid liigli political polarization.
At least for these cases, ttie international security variable and the economic performance variable are clearly tlie environmental triggers of crisis and system change. They may also act cumulatively to accentuate crisis or, at counter purpose, as simultaneous polarizers and consensus makers. The social mobilization variables are important transformers of social structure and culture in (lie long run, but it takes sharp deterioration in economic performance to convert these structural changes into political crisis and political change. But these eight cases are hardly representative in time or place. A study of crises elsewhere in Latin America, Africa, and Asia would undoubtedly give us a different set of mixes of environmental causation. We have shown various ways in which the international and domestic environments can impinge on politics. The values of tlie individual variables will vary from one area to the
next, from one time to tlie next.
In Table X.2 it is clear that some of our cases were heavily loaded by their environments, others far less so. France, Germany, and Meiji Japan stand out with catastrophic defeat or acute military threat compelling system response. Both British episodes were depression cases, with social mobilization and deteriorating terms o( trade as the long-run destabiliz-ers. As military threat and invasion are inherently more disruptive of stability than economic threat, it is not an exaggeration to say that the French and German cases reHect a larger environmental destabilizai-ion than tlie two British cases. It would therefore appear that Britain in 1832 and 1931 had higher system legitimacy to draw upon in crises, and was less heavily loaded by international and domestic change than France and Germany in tlieir crises where legitimacy was lower and environmental pressure more acme. These differences in legitimacy and burden go far toward explaining the incremental problem-solving pattern ot Britain and the revolutionary outcomes in France and Germany.
In tlie Mexican case tlie low level of environmental destabilization is striking- Here tlie explanation of tlie outcome lias to rest upon leadership behavior and prcexistent issues and conflicts ratfier than environmental destabilization. Tlie transformation of Mexican politics in tlie 1930's was attributable predominately to leadership will and decision:
llie populistic, imaginative, and resolute behavior ot Cardenas in mobilizing labor and tlie peasantry.
COALITION OPTIONS AND -SEQUENCES
Our analysis so far lias enabled us to explain crises and their outcomes in terms of constraints, pressures, and opmriunities. in other words, "necessary" conditions. Tims in England tlie conclusion ot the Napo-

630 CMJI'J, Choice, anil Change: Some Tentative Conclusions
Iconic wars deflated (lie international security issue and permitted the reform issues lo surface. A depression mobili2ed new political resources created by industrialization ;ind urbanization. With a new mix of resources and salient issues (lie old ruling coalition lost its cohesion, and new options became possible. We move from the necessary conditions of change lo the sufficient ones, Iroin constraining conditions to clioice. To analyze clioice we need options and decisions. Coalition analysis yields the options; leadership and clioice analysis gives us llic decision.
Table X.3 charts the coalition sequences in our case studies. Each coalition outcome in these sequences represents a "power-policy package" put logctiier by the actors to solve the problems presented by tlie crisis. The individual case studies show oilier coalition possibilities at each decision point. Here we only review tlie historic patli of decision, relating it to our earlier discussions of systemic and environmenial constraints, projecting it forward to tlie crisis outcome and its consequences.
Tlie coalition sequence in nineteenth-century Britain follows a steady course of cliange in coalition composition from right-center to center-left. Tlie post-Napoleonic coalition of 1819 is a Tory government following a policy of repressing demands for welfare and electoral reform. Tlie 1826 coalition adds tlie Whigs to this constellation thereby increasing its resources, but reducing its cohesion. Tlie coalition of 1829 moves one step left dropping tlie Tory "Ultras" and picking up the Radicals. It generates tlie first reform- that o[ Catholic emancipation. Almost the same coalition produces tlie Reform Act of 1832. Tlie crucial clioice point was tlie one between 1826 and 1829. A conservative coalition might have lield its ground, even lost some Whig support, and moved toward a repressive policy and coercive governmental apparatus.
It was, in Powell's words, a "complex transition period (in which) no single coalition is really dominant" (Chapter III). The coalition analysis tor 1826 and 1829 produces a broad selection of possible outcomes within tlie "preferred set," witli policy propensities ranging [rom repression to reform. Wlial tlie coalition analysis does in relation to our preceding discussion of structural, cultural, and environmental causation is test (lie limits of their explanatory cowers. From (lie range of coalition possibilities - and more important, from the variety of policy propensities which that range ofteis-we are given tlie limits of tlie determinacy of these environmental factors. Where tlie winning coalition is llie only one included in tlie preferred set, we must conclude that preconditions, whether in the pi-c'cxisling system or as a result of environmental change, were sufficient predictors ol the trend of cliange. But in many of our analyses, tliose preconditions merely created propensities for change that could be counteracted by choice and leadership factors; and in some cases, tlie preconditions left pivotal decision makers with radically ditfcr-

Gabriei A. Almond and Robert j. Mundt

enl alternatives, different mixes of policy ami power that had roughly equal chances of success. To belter explain this argument, it will be helpful to consider llie range of coalition possibilities case by case.
The second British case illustrates an incremental shift to tlic right with the MacDonald Labour faction joining (lie Liberals and Conservatives. A Labour-Liberal coalition built around a public works program was possible in 1930, however, which might have moved Britisli politics moderately left a decade sooner tlian such a sliift actually occurred. The environment liad created the conditions for a new coalition that did not emerge, an outcome we must explain in terms of leadership characteristics.
Decision making could decisively affect the outcome in conditions of substantial redistribution of resources, such as France experienced in 1870 and Germany in 1918. In France, Conservatives and Moderate Republicans occupied the center of llie ideological range. In 1871 they derived no particular advantage from this fact. Resources were so distributed tliat the preferred set of coalition outcomes included only (tie conservative-monarchist alliance tliat actually formed. Tliere was, then, no room for choice. By 1876, however, tlie repeated demonstrations in by-elections of Republican popular supmrl had evened out tlie distribution of resources, and made the pivotal position of tlie center group highly significant. Tlie Republicans then could clioose between shaky control on their own, or greater stability by sacrincing policy preferences tlirough an alliance with llie Conservatives; as we liave seen, they chose the latter power-policy combination. Its emergence depended on (lie agreement of tlie Conservatives, for whom the decision to abandon erstwhile colleagues to (lie right was by no means foregone.
Tlie German case, like the French, began with a national coalition that broke up over war reverses; unlike tlie French case, defeat resulted in the emergence of a moderate left government. Tlie dominant Social Democrats in this coalition also faced a policy-power trade-oft: a shift to tlie left based on reform of the bureaucracy and armed forces might liave given the republic a belter start, by reducing the traditional, authoritarian, and nationalist resources and increasing those of the democratic center and socialist left. This would liave been a high-risk decision with some undeterminable probability of civil war and Allied intervention. Ttie actual Social Democratic choice, a move back toward the right with large resources and high coalition dissonance produced the "pseudo-democratization" of llie Weimar Republic, preserving tlie antidemocratic resources of llie conservative and nationalist parties, tlie bureaucracy, and the military officer corps.
Mexico in 1935 providc-s us the truly exceptional case in which the range for clioicc suggested by our coalition analysis is narrowed to a

single outcome - which did not occurl Tlie "rational" outcome at that time, based on considerations of resources and issue distances, would liave been an alliance of Callistas with the revolutionary generals, which would have required a coup against Cardenas. Coalition analysis could not have predicted the strong policy preferences or leadership skills of Cardenas, who chose, outside the preferred set of outcomes, to jettison the Callistas, mobilize tlie urban workers and peasants, and tlien form a cohesive and effective coalition consisting of tlie Cardenistas and the radical labor and agrarian leaders, based on resources which he himself had created. This outcome clearly cannot be explained except by a high loading on the factor we liave yet to consider, leadership.
In the Japanese crisis of the 1860's, the Satsuma-Aizu coup in 1863 destroyed the old reliable patterns of political behavior, leaving an extremely fluid coalition in which resources were much more evenly distributed than before. The Shogunate could have maintained a preeminence in three of tlie tour most viable coalitions, but rejected all these options that miglit have left them with at least half a loaf. The solution they chose-going it alone-was outside the range of preferred coalitions, and was thus, in effect, preordained to fail.
Flanagan's study of tlie interwar crisis in japan points up a longer range clioice situation, in which the viability of emerging political parties depended on the mobilization and organization of mass support. But party leaders failed to shake off the effects of their traditional origins and, by not developing an independent resource base, they were unable to move toward a ministry responsible to the Diet, or to depolitize the military. The shallow base and lack of popular roots of the party system made it relatively easy tor tlie conservative bureaucratic and military elite to push tlie parties aside and move, in 1932 and 1936, toward a military integration of Japanese government and politics and an expansionist military policy.
Tlie Indian study differs from the others presented liere in its emphasis on explaining persistence under conditions of acute tension and distress-the language crisis and the widespread crop failure and famine. Nation-building, or nation-maintaining crises, focus our attention on the problem of avoiding disintegration. The coalition-making situation, then, is one in whicli there is always a possibility of increased fragmentation and immobilism. Resolution of this fundamental challenge tends toward a coalition of the whole; when that is not possible, then leaders must search for a functional equivalent, as in the Indian pattern of disaggrega-tion of issues and tlie deconcentration of decision sites. This was possible in India because of tlie cohesion of tlie Congress party elite, which made possible this mode of negotiation outside the formal governmental framework.

These examples li;ivc illustrated the limits of coalition possibilities, the varying decree to which actors' decisions in these crisis situations were determined by structural, cultural, and environmental constraints. But it would be a mistake to see coalition and clioice analysis as always on the receiving end of this causal relationship. Although their effect on political change is not independent of those constraints, it is equally important to examine tlie effect of coalition and choice on subsequent outcomes.
In most of our crisis resolution cases, coalition activity shows a unidirectional shift along the contender spectrum, in response to shifts in resources and issue positions. The transition is either to an entirely new set of "winners" (Meiji Japan), or else it is based on the use, by tlie flank group toward which the shift lias taken place, of its pivotal position to maintain its influence in tlie new outcome (tlie Conservatives in France, the Tory Center in 1R32 Britain, MacDonald in 1931 Britain, Cardenas in Mexico, the bureaucracy in Japan of the 1920's and 1930's). But the ability of (lie political system to cope with subsequent challenges or crises is related to this coalition outcome, and at least two cases-Germany and France - show coalition-formation movement first in one direction, then the oilier, as tlie resolution of one challenge changed issue distances and resource distribution to allow a different (and not necessarily adequate) response to tlie next challenge. In France, the repressive solution to distribution and participation demands culminating in the defeat of tlie Commune opened tlie door to a more limited accommodation of participation and the definitive rejection of forms of legitimacy not ultimately based on suffrage. In Germany, war, defeat, and related problems made possible the emergence of a new resource distribution favoring the Social Democrats, which presented tlie contenders with a new set of coalition considerations. At this point, the Social Democrats opted for power over policy, reversing the direction of coalition-formation movement, and attempted to stabilize that movement at a point where they were unable to affect the environment or. consequently, to cope with future crises.
We have paid considerable attention liere to choice as a trade-off between power and policy and resources and issues. From the cases examined, we might speculate on the implications of these trade-offs for political change. It would appear that tlie greater the adjustment to be made in the "constitution" hy a new coalition, tlie lower its ruling potential is likely to be. Tlie more extreme tlie policy preferences of tlie ruling coalition, the greater its polari/ation from other actors, and tlie likelihood that oilier actors will not join or support the coalition. Facing increased uncertainly ;tbout tlie consequences of pursuing its innovative goals, tlie new elites are pressured 10 sacrifice innovation for safety. Here tlie crucial variable i.s risk-taking. Thus llie German Social Democrats,

faced with tlie clioice between a broad coalition to tlie right, or a narrowly sufficient coalition to tlie left. cliose a low risk-taking strategy. Cardenas's option was to proceed with tlie coalition tliat met his policy goals, shoring up its originally high instability in a series of remarkable maneuvers - in other words, a high risk-taking strategy. A more effective response to tlie problems precipitating crisis came from a narrowly based coalition tliat represented sharply denned policy responses. These two examples suggest tlie hypothesis tliat tlie stability of a crisis outcome may be directly related to the size of tlie emergent coalition in the short run, but inversely related to it in tlie long run.
Tlie crucial factor, then, in tlie success of an innovative strategy is the capacity to mobilize and create resources that may require a high input of creative leadership as well as tlie availability of resources.
LEADERSHIP. COALITION SEQUENCES, AND OUTCOMES
Our coalition analysis alerts us to the power-policy options in given crises. Tlie value of this analysis is tliat it forces us to specify all of the logically possible coalitions, and order them by their resources and cohesion, thus taking the logic of explanation one further step. It tells us tliat the winning coalition should be in this set of coalitions, and further tliat it will most likely be in a particular subset that includes those coalitions with "reasonable probabilities" of forming (i.e., the preferred set). The assumption is that freedom of choice is limited to this preferred set.
In judging the strength of llie leadership variable we can employ tins preferred set of options as a kind of measure or scale of tlie magnitude and quality of leadership as a causal variable. It will tell us something of the risk-taking propensities, tlie policy propensities, the strength or resoluteness of leadership, and tlie skill properties of leadersliip. Tlie most striking example of'leadership causation is tlie Mexican case. Cardenas broke from his alliance with Calles, moving in early June 1935 from an immobilist coalition with more than lialf tlie resources to a coalition with the left-labor and peasant leadersliip and the CediUistas. The new grouping at that time controlled less tlian half tlie resources, but supported the reforms tliat Cardenas favored. Within a month from the formation of this coalition its share of resources had increased to over half. Within a year Cardenas was able to drop the CediUistas, thus forming a more unequivocally reformist coalition witli two-tliirds of the resources. By 1938 lie liad shape<|I a powerful left-oriented coalition dominating almost three-fourths of (lie resources. Thus we can say with some confidence that a willingness to take risks, a resoluteness in policy direction, and Cardena&'s skill in resource mobilization form a large part of the explanation of [lie reformist outcome of tlie Mexican case. Our best measure of C;ii"deiias's leadership ability i.s tfie fact tliat his coalition clioice in late June 193f> was outside the preferred set of outcomes; lie

had, in effect, used the uncertainties of real-world decision making to his advantage, producing an outcome that would not have been possible had all (lie actors possessed all the relevant information on that clioice situation. In the case of Ramsay MacDonald we liave seen how, in 1931, he missed the opportunity to form a Labour-Liberal coalition tliat was the most attractive in the preferred set, at least in terms ot working toward policy goals lie had long espoused.
Thus, the effectiveness of leadership must be judged relative to the decision-making context in which it is exercised. To consider, first of all, the constraints of the preexisting system, it seems that where political structures are highly legitimate, individual roles within them tend to be precisely defined. Sometimes, the cultural environment has a constraining effect that transcends given structures or regimes; where, as in the Japanese cases again, value is placed on consensual decisions at all levels of society, the ability of an individual to strike out in new directions is clearly limited. Undoubtedly some level of restraint of this type occurs in all societies, and in all the cases considered here, effective leaders liad to break from the rules of the game - an act that increases tensions independently of the issue positions involved- Thus, Cardenas incurred the special onus of breaking with leading members of the "revolutionary family" - an act seen as a breacli of faitli that could only have added to the issue distances.
Just as the range of coalition options is enhanced by clianges that aggravate potential crises in the preexisting system, so do such changes open new possibilities for the contenders- Even the feudal Shogunate saw a wider range of possible outcomes as a result of Perry's arrival. In the chaos following military defeat in 1870 France and 1918 Germany, contenders were led to consider a much wider range of alternatives than previously would liave occurred to them. Indeed, an innovative leadership role occasionally becomes an institutionalized part of political response patterns. The "man on horseback" is a product of social and political confusion, his chances of success inversely related to the viability of preexisting structures as crisis tlireatens. Some political systems periodically make room for such figures, as when in repeated crises the pattern develops of resorting to a charismatic figure to reduce the polarization among contenders when none of themlhas the resources to prevail. Sucli was the case in nineteenth-century France, where a pattern of this type was set. The redistribution of political resources following the defeat of the Commune eliminated the need for a Caesaristic solution to the problems faced by tlie Third Republic; but tlie tendency endured and was manifested again throughout tlie Third, Fourth, and Fifth Republics.
The constraints of llicsc macro- variables on leadership are quite faith-

fully reflected in tlie clioice range specified in our coalition analysis. Wlien issue distances are low and coalition patterns traditionally flexible, many possible outcomes will be included in the preferred set. When long practice and firmly established ideologies dictate ratiier rigidly which coalition partners are appropriate, or when issue distances are extreme, then even when environmental impact has brought about fundamental reassessments and suggested new realignments, the preferred coalition list may be small and the realistic options limited.
The quality of leadership must be measured against the scope of the leader's discretion- Cardenas wrought dramatic changes in a situation where environmental impact did little to create a "crisis," and where the solution lie chose would not have been evident from the estimated resources available to him at the time of his decision. MacDonald, on the otiier extreme, acting in a period of high insecurity when new solutions were actively being sought, failed to lake a lesser risk.
In Table X-4 we have rated leadership properties for each of the cases, considering only the preeminent actors at crucial decision-making phases. The cases are coded dichotomously by strength of leadership, by which we mean a set of clioices by a leader or leaders which optimizes the power base and tlie policy goals of a coalition. The table also indicates whether leadership was exercised collectively or by an individual. Our distinction here is between those interrelationships of high-placed individuals witliin a given contending group tliat are symmetrical, and tliose that are asymmetrical. Obviously, this distinction is somewhat arbitrary: in 1832 Britain, for example, sucli figures as Grey were especially prominent. But Grey does not compare with, say, Cardenas, in tilis respect, and the Reform Act was actually a product of wider Whig bargaining to tlieir left and right. The Radicals also showed skillful leadership in bringing mass discontents to bear on tlie political process, but tlie Wliigs were equal to tliem, benefiting from tlieir mass support in exchange for little immediate gain for tliose supporters in terms of the conditions causing unrest. Britain in 1931 was a contrasting case in both respects: political decision making tended to be concentrated in a single leader, MacDonald, who, as we have seen, made poor use of rather wide
scope for innovation.
The same can be said of the collective leadership oE the German Social Democrats. As with oilier contenders whose policy propensities require departures from tlie status quo, they can be faulted tor not taking the risk of loss of control necessary to turn the corner. MacDonald's action in June 1931 is even less excusable; lie cliose to hold onto a coalition lliat was less effective than alliance with tlie Liberals would liave been. In other words, his choice at that early point in the crisis increased neither power nor policy. The German Imperial government and the military reflected the weakness of conservative leadership after Bismarck in their stubborn persistence with political forms and military ends that were incompatible with their needs for wide support in a crisis period and with their international situation. Theirs was an unrealistic choice of conservative policy over power. The Social Democrats, as we have seen. gnve up tlieir chances for policy change at a lime when the direction of shift in political resource distribution suggested risk-taking.

In the Frencli case, leadership was decisive in all the various outcomes. Napoleon III no longer showed the decisiveness and resolve that his Caesaristic role required in (tie last years of tlie Second Empire. Thiers' political skill and experience in the struggle with the Communards, in contrast to their lack of direction and misjudgment of tlieir resources. largely explains the shift in resources in his favor in May 1871- His political skills in the formative period of the Third Republic and his perceptive appraisal of resource values explain his preference shift on the form-of-governmem issue. He put his policy preference for societal stability over the short-run utility of his own faction, in moving to the ultimately more effective coalition based on republican institutions.
The Shogunate in the last days of Tokugawa Japan acted on political concepts that had been deeply implanted in that long and stable period of feudalism. Their strategy of going it alone in the early 1860-s on an insufficient resource base was only possible because of the lack of information available to oilier contenders. Once the information lag was overcome, the Shogunate-s refusal to accept a wider coalition meant that in time coalition options would appear that would exclude it- The Meiji regime's vigor and openness 10 new information made it a classic example of effective coping with an initially threatening environment. The party leaders in Japan of the 1930-s never fully appreciated that a responsible parliamentary government required an appropriate resource base in popular support. Rather, the military-bureaucracy alliance developed that popular support once the parties had missed their opportunities and
been thrust aside.
This sketching in of the leadership variable goes beyond the data base
of our program. Our judgments are impressionistic though they seem plausible in the light of historical evidence. Though it leaves our work incomplete, it may be appropriate to pause at tins point and give the choice and decision aspect of developmental causation the kind of research and analytic effort it deserves. We have developed systematic ways of establishing the constraints and options in political situations and the ways in which tlieir outcomes are determined: What ranges of freedom of choice there may have been in historical contexts, or may be in contemporary situations, and how leadership and clioice qualities may effect the exploitation of tliese opportunities, are the problems of future research.
THEORETICAL AND METHODOLOGICAL IMPLICATIONS
Although the principal purpose of this book is exploratory, it provides some demonstration of the advantages of this approacli to develop_ mental causation over earlier attempts at historical developmental explanation. Surely, Barrington Moore-s landmark study of alternative

routes to industrialization (1966) is one of tlie most im)K)i-tant of tliese attempts. Yet it would appear to be substantially oversimplified and over-determinist in the light of our examination of four cases that are either included or treated in Ins comparative analysis - Britain and France on one hand, and Germany and Japan on the other. Though lie qualifies these hypotheses here and (here in Ins work, it is not an exaggeration to impute the view to him tliat (lie democratic-capitalist route to modernization had been substantially determined in Britain in tlie seventeenth century with tlie commercialization of agriculture, and in France in the late eighteenth and nineteenth centuries with the destruction of the aristocracy and tlie emergence of an independent peasantry. Moore argues that tlie German fascist-authoritarian route to industrial modernization was predetermined by the fusion of royal bureaucracy and landed aristocracy in tlie seventeenth and eighteenth centuries, and in Japan by tlie alliance of bureaucracy and landliolders in the latter half of the nineteenth century. His effort at a parsimonious social-structural explanation of tlie differences in tlie patterns of political and economic development leads to a loss of discrimination in his interpretation, and a sense of inevitable drift once a course lias been set.
For example, if we illuminate the interplay between social structural change and political change in Britain in tlie eighteenth and nineteenth centuries it would appear that Moore lias overestimated the determinacy of the British developmental pattern, Our 1832 case study suggests that Britain in the 1820's had an option for a responsive "undemocratic" solution to tlie destabilizalion of British politics resulting from industrialization and tlie termination of the Napoleonic wars. Close examination ot this episode suggests tliat from tlie perspective of this period in British history the Whigs and tlie Tories might have moved either way-toward reform or repression. Tlie memories of members of (tie political elite and their risk-taking propensities provide us with tlie "sufficient" explanation of tlie reform option.
Our three nineteenth-century European cases raise some rather serious questions about Moore's interpretation of the causes of democratic and fascist modernization. Tlie possible British option in the 1820's of a center-right coalition in favor of repression might have turned the British developmental path somewhat in the direction of tlie French and even German pattern. And a France led by a more effective Napoleon III in 1870, or by a more effective conservative-monarchist coalition in 1872, might have come out with a constitutional monarchist solution and perhaps moved toward a more British type of developmental patli. A Prussia, even with Bismarck but without tlie Frcncli gift of tlie triumph of 1870, might have been unable to produce a Second Reicli witli tlie particular mix of institutions and classes that stood in the way of effective democrali/ation. Although the m;i(To-variablcs or class structure give us a

part of the explanation, it would seem that the micro-variables of coalition options and human clioices had considerable independent causal
value.
Though Moore's treatment of German development is briefer, it seems
to be particularly oversimplified and overdeterministic. He not only seems to have ruled out other options in the nineteenth century, he overlooks the opportunities for choice that may still have existed in the twentieth. If our case study is correct the left-center option in the 191&-1920 period might have reduced tlie conflict propensities of the Weimar
period and avoided National Socialism.
We feel a bit less confident in asserting that Japanese political development liad another option during the rise of political parties in Japanese politics during the first two decades of the twentieth century. The party elites may have had a choice of consolidating a democratic party system through the development of party organization and the legitimation of trade union organization. This would have provided the resources that might have enabled them to resist the pressure of the conservative forces of Japanese politics and carry through a democratic integration of Japanese government and politics rather than collapsing under an authoritarian traditional one.
Similarly, Moore's reductionist method conceals some fundamental differences in the German and the Japanese developmental patterns. In the German case industrialization and social mobilization were associated with the development of mass party organization and the effective formation of interest groups. Hence a populistic component had to be built into the authoritarian solution. In Japan industrialization was not accompanied by significant political mobilization and organization. Consequently, the Japanese solution was more traditional and authoritarian, lacking tlie mass mobilized parly that we usually associate with fascism.
Robert A. Dalil (1971) has also formulated a theoretical model of alternate historical pathways, viewed from the probability that they will result in stable democracy (or "polyarchy," the closest empirical approximation ot the democratic ideal). Tlie value and limitations of the insights he provides emerge rather clearly from an attempt to fit them to
the examples in our case studies.
Dalil suggests two dimensions along which to measure democrat! za-tion: liberatiialion, or tlie legitimation of opposition and competition tor authority positions; and mcltt'.ivcness, "tlie proportion oE tlie population entitled to participate on a more or less equal plane in controlling and contesting tlie conduct of government" (1971:4). Dahl identifies tour logically possible combinations of extreme highs and lows on each measure: (1) closed hegemony, witli neither competition nor wide participation; (2) competitive oligarchy, wliere contestation is restricted to a small elite; (3) inclusive hegemony, witli wide participation but little oppor-

tunity for public contest; and (oI) polyarchy, the combination of inclusive-ness and competition. He then identifies three possible paths from hegemony to polyarchy: direct inmsition, or through one of the two "intermediate" paths (i.e., competitive oligarcliy to polyarchy or inclusive liegemony to polyarchy).
A major difficulty with Dalil's approacli is its treatment of the demo-cratization process in nondialectical terms. Tlius in's model suggests that stability and democracy must he sought in a one- or two-stage sequence in which liberalisation and inclusiveness must be ordered. His polyarchic solution is to proceed first u'itli competition, providing for inclusiveness only when a "gentlemen's agreement" on the terms of political competition lias been firmly established.
It is perhaps unfortunate tliat in the development literature which deals witli democratizaiion, the prerequisites of stability and of democracy are not usually disaggregated for analytical purposes. To draw on one of our cases, in France in 1875, all contenders accepted, if they did not embrace, the legitimacy of electing office holders by mass suffrage. This was a victory for llie champions of inclusiveness. But it was a victory accorded after the range of allowable competition iiad been greatly reduced. Under these conditions, Tliiers perceptively foresaw the place of mass suffrage in [he Third Republic as a new means of providing symbolic support to the authority structure, while having a minimal impact on policy. Once the range of contestation had been reduced by suppressing the revolutionary left, inclusiveness was much less salient as an issue, and was accepted.
Was this the achievement of stable polyarchy in France? Yes, in the short run, for a mass electorate could choose among competing candidates. But. a more thorough democratization developed only much later, when this inclusive electorate was provided with a range of competing contenders that adequately reflected the variety of political preferences seeking expression in France.
We would conclude from this analysis tliat, in the interplay of elite clioice and mass political mobilization, it is necessary for stability that contestation advance ahead of inclusiveness as Dahl suggests; but we would add that it is necessary for democratization that inclusiveness be established ahead of contestation. The paradox disappears over time, as we have seen in the French case, in the dialectical development of first one and then the oilier.
We can test these propositions witli our British cases. According to Dahl (I971:3'l)> British development followed the path from closed liegemony to com]^etitive oligarchy to poly;irchy. Our contention, however, based on our two Brili'sli studies, and the historical events between them, would be llial British development is more aptly characterized by a

zig-/ag path along die diagonal, i.e., llic path of direct transition (Figure X-l). We liave seen that the earlier emergence of competition led to movement toward inclusiveness in 1832. But this and succeeding moves toward inclusiveness gave impetus in turn to wider competition. The 1832 outcome was possible because the range of competition could be restricted to Tories and Wliigs; the lfi67 suffrage extension could be adopted because the competition could lie restricted to Conservatives and Liberals. In the final extensions of suffrage in (lie twentieth century, the Liberals were replaced by Labour, as a set of alternatives emerged that more closely reflected the distribution of resources and issue positions.
Dahl's preference for the earlier emergence of competition fits the short-term analysis of France in 1871; the high level of inclusiveness that had emerged under Napoleonic hegemony could not sustain the sharp polarization of that year. The level of competition fell, however, with the collapse of the Commune, to a point tliat could support high inclusiveness. Witli tlie principle of wide participation firmly established, the range of competition inevitably increased giving substance to the democratic forms toward the end of the century. However, those contenders who had been included, then forced out of the arena in sharp development reverses along the dimension of competition remained uncommitted to the rules of the game, and created tlie condition of immobilism that marked tlie Third Republic (Figure X.2).
In our German case, the missed opportunity of the Social Democrats in 1918-20 was to reduce competition to a level commensurate with the establishment of a stable polyarchy - but in tills case by eliminating the resource base of the riglit-wing contenders. The outcome was again an immobilism deriving from the existence of contenders that were hostile to the Weimar Republic. Dahl (1971:129) suggests an appropriate axiom:
>
"The greater tlie belief within a given country in the legitimacy of the institutions of polyarcliy, tlie greater the chances for polyarchy"; but, as in Ins discussion of other conditions affecting the attainment of stable polyarchy, he does not systematically connect subculture configurations and elite belief systems to his elaboration of historical sequences.
It would appear that any significant movement along one dimension will induce movement along the oilier; if a "vertical" movement is in some way inhibited, pressure intensifies to slow or roll back movement along tlie horizontal dimension as well, until an appropriate point of equilibrium between tlie two is readied, as in France in May 1871 (Figure X.2). Because tlie two dimensions of movement are associated with eacli other, tlie most stable path of democratization will be along the path of direct transition. As inclusiveness must be accorded in blocs (not one voter at a time), a perfectly smooth curve cannot be attained. Hence the zig-zag path along the diagonal may be the route to polyarcliy that keeps competition and inclusiveness in sufficient balance to avoid crisis
and instability.
Dahl's elaboration is, in our view. restricted by his description of historical patterns in terms of four ideal types, whereas an adequate explanation requires measurement at intervening points. Here our use of the polarization index in measuring competition and our quantification of the more important varieties of |x>litical resources in measuring inclusiveness give us a basis for moving along Dalil's dimensions with something like tlie precision required.
These criticisms of tlie work of Moore and Dalil are not intended to detract from their creative insigliis, but to suggest how we may get beyond the limited explanatory power of their formulations. Tliese limits result in the fust instance from a deterministic treatment of politics (plus a failure to consider llic effect of iiilcrnation.it politics), and in (lie scc-

ond from an ahistorical, nondialectic approach. We would argue that our approach enables us to win back a bit of the autonomy of the political sphere, rehabilitate tlie role of human choice and creativity in developmental problem solving, and chart the outcome as a complex and continuous historical pattern.
Furthermore, these critiques of Moore and Dahl are not to suggest that
the rest of us who have engaged in developmental theorizing have done any better. They suggest the advantages o( a method tliat deals more effectively with the interaction of politics and its environments and that illuminates tlie relationship between constraint and choice in political problem solving. We have learned in (.his study bow to do better research on comparative political development. And it may be useful to spell this
out a bit more explicitly by way of conclusion.
The purpose of scientific work is to enhance our capacity to explain and predict, and insofar as we succeed in doing this, to attain our own goals more surety, more efficiently, more economically. Though some of us in political science may view our problems with the disinterested curiosity of a pliysicist, a discipline that deals with human conflict, coercion, and welfare can hardly escape moral responsibility whether through the mechanism of professional conscience, or through tlie expectations of
our constituencies and audiences. '
"We suggest that the approach illustrated in our case studies is a more
effective way of gelling at explanation, prediction, and control of political development than the approaches so tar employed or recommended in the field. To begin with explanation, ttie model we use in our explanatory efforts satisfies the requirements of a good model. Having reaped the heuristic value of mechanical, biological, and cybernetic analogies, we have dropped these analogies, and liave constructed a theoretical framework appropriate for analyzing politics and political change- In developing and testing the applicability of this model we have used a "most different case study strategy of comparison" (Przeworski and Teune, 1970:33; Lijphan, 1971:687), a strategy that requires us to sample a wide range of variation of the phenomena being investigated to determine whether we have included all tlie important tlieoretical variables, and have some understanding of their modes of interaction. The selection of cases in this book illustrates tlie uses of this kind of exploratory strategy. In future developmental research we will be better able to deal with such variables as tlie impact of tlie international environment, or of leadership on political development. Our understanding of tlie various ways these and other variables may affect development has been greatly enhanced by tlie dissimilarities of tlie cases we liave examined. It lias given us more confidence that we liave a set of related concepts "able to travel" in time
and space (Sartori, 1970:1034).
Having formulated and tested this analytical model on a wide range o(

cases, we suggest that we can now benefit from a "most similar comparative strategy" (Pr/eworski and Tenne, lOPO:?!; Lijphart. 1971:688). Indeed, as Lijphart argues, tlie strategy of comparison that comes closest to the scientific ideal of experimental control is tlie comparison of [lie same political system at adjoining points in lime. Our two British and two Japanese cases demonstrate tlie advantages of a diachronic comparative case study strategy, particularly the Japanese case studies that were done "back to back" so to speak.
If we want to be taken seriously in our efforts to explain how Britain, France, Germany, Italy, the Netherlands, Spain, and tlie like, got to be the way they are (and our literature abounds in such propositions) we shall have to move now to longitudinal studies of political development tliat combine lime series analysis witii "crisis, choice, and change" analysis. Tlie field seems generally to agree on tlie desirability of longitudinal analysis of political development. Our work suggests that a set of time series measuring environment and systemic change at the macrolevel for any political system will leave tlie developmental path unexplained unless the microlevel of coalition options and leader choices are "factored in." Tims, as Rittberger (Chapter VI) suggests, an adequate explanation of German political development wonki consist of a set of curves over time describing environmental change and systemic response. When these curves fluctuate or "kink" (in other words, in crisis periods) we may assume that political resources and issues have "opened up," [hat new coalition possibilities have emerged, and that political leaders Iiave clioices which will effect tlie future developmental path. Thus Rittberger suggests tliat a combination of time series analyses plus a set of crisis case studies - including tlie Siein-Hardenberg reforms in Prussia of 1807-19, tlie revolution of 1848-49, the constitutional conflict in Prussia in 1862-66. and [he November (1918) revolution - would give us a constellation of causes of German development which would "explain" it more rigorously and definitively than is now possible. Simila*- research strategies for British and Frencli development might begin to transform the comparative political history of modern Europe into a more systematic and reliable body of propositions.
We also argue tliat tin's approach to developmental explanation enhances our capacity to predict developmental paths and outcomes. When we reacli that point in the analysis at which we can specify coalition and policy options, we are in a position to predict that the outcomes are narrowed down to ;i specifmble and limited set of clioices with particular power and policy propensities. This is a lot, but it is not prediction in any precise sense. We have found in our case studies [hat we Iiave to predict more than one possibly winning option, which in most cases leaves llic actual outcome indeterminate.

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