Acerca del autor

Jack A. Goldstone is Professor of Sociology and Director of the Center for Comparative Research in History, Society, and Culture at the University of California, Davis.

Fragmento. © Reproducción autorizada. Todos los derechos reservados.

Revolution and Rebellion in the Early Modern World

University of California Press

I think God Almighty hath a quarrel lately with all Mankind ... for within these twelve years there have the strangest Revolutions and horridist Things happen'd not only in Europe, but all the world over.
—James Howell

For nearly 350 years scholars have wrestled with the "general crisis of the seventeenth century." Certain facts are not in dispute. The first half of the seventeenth century saw a widespread slowing and eventual halt to the steady increases in population and prices that began around 1500. In addition, rebellions and revolutions shook regimes across the Eurasian continent, most notably the English Revolution; the Fronde in France; the anti-Hapsburg revolts in Catalonia, Naples, Sicily, and Bohemia; the Khmelnitsky revolt in the Ukraine; the celali revolts in the Ottoman Empire; and the collapse of Ming rule in China (Parker and Smith 1978; Aston 1967). But there is no agreement on how to explain these facts. Historians contest fiercely over the causes, the connections, and the significance of these events.

Some have argued that the seventeenth-century crisis marks a turning point in the history of capitalism. Others have suggested that what occurred was mainly political, a crisis of absolutism. Certain scholars have found a worldwide depression in the seventeenth century and have sought to explain it. Yet others have noted that trade grew rapidly in some markets, and thus claim that a general depression is illusory. These debates have proved inconclusive, despite marshaling some of the keenest minds ever applied to historical studies (most of the key contribu-

tions are presented in the collections edited by Aston 1967; Forster and Greene 1970; and Parker and Smith 1978; other important works are Anderson 1974; de Vries 1976; and Wallerstein 1974, 1980).

A fundamental defect in these debates is that all arguments tend to be strongly Eurocentric. Reflecting the biases of both Marx and Weber, on the one hand they assume that events in the West marked important structural changes, demonstrating the dynamism of Europe; on the other hand, the equally major political rebellions and regime changes in the East are given only minor attention and are often dismissed as mere peasant uprisings or dynastic changes. In fact, the political rebellions in the Ottoman Empire and China involved the same broad spectrum of elite, urban, cultivator, and heterodox ideological struggles against a fiscally weakened state that characterized Western political crises (for China, especially, see Wakeman 1986). Moreover, the Eastern crises resulted in arguably greater changes in state power, class structure, and local government than most Western crises of the period. Thus, at the very least, the history of the general crisis in the East should be more closely examined by Western scholars. At most, and I shall argue the point strongly, Eurocentric notions of the seventeenth-century crisis, which posit a crisis of capitalism that led to dynamic structural changes in Europe and extended to, but caused far less change in, peripheral areas in Asia, should be discarded altogether. To grasp the nature of social change in the seventeenth century, we need to recognize a worldwide crisis of agrarian absolutist states that affected both Eastern empires and Western monarchies.

The same is true of the turbulent period from 1770 to 1850, which had its most memorable moments in the French Revolution but also included Pugachev's revolt in Russia; the European rebellions and revolutions of 1820-1821, 1830, and 1848; the Greek, Balkan, and Egyptian revolts within the Ottoman Empire; and the start of the bloodiest revolt in history, the Taiping rebellion in China. Hobsbawm (1962) has quite reasonably christened this period the "Age of Revolution." Viewing all of these events, one might reasonably conclude that the entire early modern period was simply one of successive crises.

Yet in the years from 1660 to 1760, full-scale political breakdown was curiously rare (Rabb 1975). There were isolated peasant revolts, elite rebellions, and dynastic changes, but revolutionary civil wars were absent from Europe and the Ottoman and Chinese empires. This stability occurred even though the two main alleged engines of revolution— war and the growth of capitalism—roared even more strongly than in

the preceding century. This period was one of massive wars, from the "Second Thirty Years' War" (1688-1714), spurred by the ambitions of Louis XIV, to the Seven Years' War (1756-1763), employing the largest and most sophisticated armies that Europe had yet seen. This period was also one of enormous development in capitalist enterprise: in England, the vast majority of enclosures—perhaps two-thirds or more of all enclosures between 1500 and 1850—took place in this century (Wordie 1983); in England, France, and the Netherlands, the rural putting-out industry, which rendered the worker dependent on the suppliers of capital, grew rapidly, displacing the independent artisan in many crafts (de Vries 1976); and throughout the world, the volume of overseas trade grew exponentially, as intra-European trade was joined to vast colonial and imperial trading networks in the Americas, the Levant, and the Far East. I shall further detail, and attempt to measure, these trends later on. Yet it should be evident that the presence of war and the growth of capitalism persisted from 1660 to 1760. State breakdown did not persist. We thus face a problem of historical explanation: why was state breakdown, not merely in Europe but on a worldwide scale, clustered in two marked "waves," the first culminating in the mid-seventeenth century, the second in the mid-nineteenth, and separated by roughly a century, from 1660 to 1760, of stability?

Attempts at comparative study of Asian and European state breakdown in the early modern period have been rare (Mousnier 1970b, 1984). But as we shall see, there are remarkable similarities in the originating patterns of state breakdown in Eastern and Western absolutisms: state fiscal crises linked to inflation; intra-elite divisions over social mobility; and popular uprisings, partly autonomous and partly elite orchestrated, that pressed basic economic demands so fiercely as to lead to changes in political, social, and economic organization. In seventeenth-century China, just as in seventeenth-century England, there was even a faction called the "Levellers," who proclaimed a new age of equality of all men and influenced the popular revolts that accompanied the fall of the Ming. Nonetheless, after the state breakdowns of the mid-sixteenth century, Europe's economic advance accelerated through discoveries and innovations, while Asian economies grew only quantitatively, becoming oddly fixed in their economic and administrative techniques. Thus two more historical questions arise: Why did the near simultaneous state breakdowns in major Asian empires resemble, in their origins, the state breakdowns in Europe? And to the extent that they had similar origins, why were the long-term outcomes so different, lead-

ing to that divergence in development commonly known as the "rise of the West?"

States and State Breakdown

No doubt, some readers are already impatient with my juxtaposition of vastly different societies and the ill-defined term "state breakdown." Yet the states of early modern Eurasia—including the European monarchies, Russia, China, the Ottoman Empire, and Japan, in roughly the sixteenth through the nineteenth centuries—were not greatly different from each other.1 The Eurasian states comprised more or less bureaucratic assemblages of officials, under the direction of a single hereditary ruler, and claimed sovereignty over well-defined and usually contiguous territories. Their economies were overwhelmingly agrarian, and the income of states and elites mainly depended, either directly or indirectly, on taxes and rents on land.

By "the state" I mean the institutions of centralized national-level rule-making and rule-enforcing power, including the individuals who controlled those institutions when acting in their official capacities. For early modern states, this included the hereditary ruler and the ministers, legislative assemblies with national powers, royal or imperial tax collectors and judges, and those military forces directly subject to central control. Thus by "the state" I denote only a part, although usually the dominant part, of the total set of political actors and institutions. In early modern societies a variety of other groups and individuals also might have exercised judicial or administrative powers, such as the ability to collect taxes or fees, enact regulations, and administer punishments for violations. Such groups—including regional parliaments or estates; county, municipal, and village authorities; churches and their religious judges; and lay and religious seigneurs—though subordinate, were somewhat independent of the central executive authority and on occasion collided with it. Thus state breakdown generally involved the collapse of the central authority's ability to dominate in a confrontation

Throughout this book, I use "Eurasia" to denote only the temperate regions north of a line drawn roughly from the easternmost end of the Mediterranean Sea through the Caspian Sea, the Himalayas, and the southern border of China. I thus exclude most of Syria, Mesopotamia, Persia, the Indian subcontinent, and Southeast Asia. These latter regions were often differently administered than the regions north of this line; moreover, their economic and demographic histories are less complete. For these reasons, and to keep this study within manageable bounds, I make no claims about whether the argument developed in this book applies to these southern lands.

with other politically powerful actors, rather than the breakdown of all political institutions.2

These societies also exhibited sufficiently separate political and religious authorities that the relationship between the two was problematic. Although specific individuals, above all, rulers, might exercise both political and religious authority, and though the state might seek to determine religious practice and appoint religious leaders, there existed a semiautonomous elite outside of official state office who acted as custodians of a well-articulated moral tradition (Eisenstadt 1980b). In Europe this condition arose because of the presence of the Catholic priesthood; in the Ottoman Empire it arose because of the presence of the Greek Orthodox Church and of the many Islamic religious schools, pious foundations, and their scholars and mullahs; in China it arose because the Chinese literati, although providing the empire's officials, never placed more than a fraction of their members in office and retained an informal role in local education, administration, and ceremonial functions that gave them a partly autonomous role in Chinese society; in Japan it arose because of the separation of the roles of the shogunal administration (which governed) and the imperial court (which remained the sacral center of society).

"Elites" were families of exceptional wealth or status, usually literate, but not necessarily state officials. They were somewhat differentiated by their pursuit of various paths to wealth and status, such as through land, arms, trade, administration, or religion. Such differentiation never constituted a clear-cut division of labor, however, and most elite families showed diversified portfolios of pursuits.

Because of the diversity of elite interests, the relationship between the state and elites was always problematic. Elites expected states to perform certain tasks of government that provided diffuse benefits: orchestration of successful military campaigns; maintenance of internal order; regulation or provision of certain collective goods, such as coinage and transportation networks; and provision of status- and wealth-enhancing opportunities to elites through regulated access to military and

M. Weber (1978) and Mann (1988) ascribe to states a monopoly of legitimate force; but this is a false characterization of early modern states, which existed in tension with semiautonomous sources of legitimate authority at the regional level or among groups subject to religious law. My definition follows Weber and Mann in ascribing to states the centralized national rule-making and rule-enforcing authority, but it differs in recognizing that the state shares political space with other actors and authorities. My definition therefore allows for a routine degree of tension and potential conflict that fuels political dynamics.

civil offices (sometimes including regulated access to elite education and training). To perform these tasks, the state required resources gained by taxation, land-ownership, or other sources of revenue. Yet the concentration of resources in the hands of the state posed certain problems. If the level of resource provision was too low, the state could not perform its tasks; if the level of resource provision was too high, the excess might be used to increase the wealth and power of rulers at the expense of broader elites. Elites therefore generally supported state actions that appeared consistent with the performance of expected state tasks, but opposed state actions that threatened an excessive concentration of power in rulers at the expense of broader elite interests. For its part, the state generally sought to perform its expected tasks in a manner sufficient to avoid widespread elite attacks, while seeking to increase its resources to enhance its wealth and power relative to domestic elites and international competitors (Levi 1988). This continuous tension over state resource control was usually managed by negotiation between the state and elites. State crises occurred when these latent conflicts escalated into overt struggles.

Of course, there was also routine tension between the population at large and the state. Rural and urban groups in early modern society looked to the state to ensure maintenance of roads, bridges, and irrigation works; bread at affordable prices and adequate employment; protection from banditry, foreign conquest, and elite exploitation; and administration of local justice. A certain amount of taxation was generally accepted as the price of these collective goods. In fact, these services were usually provided by local or religious authorities rather than by state officials, often for fees collected in addition to state taxation. Nonetheless, common people generally regarded their king or emperor or sultan as the ultimate guarantor of local order. The state was generally blamed when these services deteriorated, or when taxation reached levels that threatened traditional peasant living standards. Under these conditions, popular groups might revolt to demonstrate their grievances against the state; or they might revolt in the name of the king against local elites, decrying the latters' "failure" to comply with the presumed wishes of the king to protect the common folk.

Popular groups also contested with elites for resources, while being dependent on elites for guidance and protection in matters of politics, justice, religion, and material welfare. As with the state, popular groups ordinarily accepted status distinctions and resource extraction by elites as consonant with elite protection and guidance. Yet as with the state,

this relationship should be considered one of "routine tension," for elites generally sought to maximize resource extraction and popular groups to minimize it, and elite definitions of adequate and effective guidance and protection often differed from popular views.

The societies studied in this book differed from most earlier agrarian empires in having widespread internal markets, so that the prices and distribution of goods were widely affected by supply and demand. Prices were meaningful in that they affected the access to goods and services for a wide variety of actors. Thus rulers had to acquire food for armies and many other services at prices set by the market; elites' earnings from rents or productive enterprises depended on the market price of their products; urban workers' incomes depended on a market-influenced wage; and peasants' access to land depended on market-determined rents. Of course, many actors used political power to influence the market and attempted to hedge or evade market restrictions. And peasants who had secure landholdings adequate to provide for their families might have participated in the market only marginally, to acquire occasional tools or luxuries or to earn small amounts of cash to pay taxes. However, in each of these societies the market was so widespread that it substantially influenced the incomes and opportunities of many actors, including all of the politically important ones.

In sum, in all of these societies there was (1) an agrarian economic base; (2) a hereditary ruler and officials who administered a territorial state, but who remained in some tension with semiautonomous local, regional, and religious and cultural authorities; (3) a literate elite extending beyond the circle of officialdom who followed a variety of pursuits, and who remained in some tension with the state over the adequacy of state performance and the level of state resource extraction; (4) urban and rural popular groups who were subject to cross-cutting allegiances and resource extraction from both elites and the state; and (5) reasonably well functioning internal markets wherein prices affected the access to goods for a wide variety of social actors.

This is not to overlook such matters as wide differences in the resources of states, in the autonomy of elites, and in the content of religion, all matters to be addressed in due course. It is only to say that for all of these societies one can reasonably speak of "states," of "elites," of conflict between "secular rulers" and "religious groups," and of "prices" and refer to things not impossibly dissimilar across societies.

I shall refer to the revolutions and rebellions examined in this book as cases of state breakdown . By state breakdown I mean a particular

combination of events, but not quite a revolution. The term "revolution" is frequently used with little care; as a result, it has become vague and slippery. Moreover, since the French Revolution, this term has picked up overtones of radical political change that are out of place in referring to many political crises that occurred before 1789. Let me therefore carefully define the subject of this book.

Some authors (e.g., Huntington 1968) define revolution in the extreme sense of a violent and permanent overthrow of government and elites, based on mass participation, that establishes new political and economic institutions, a new status structure with new personnel, and a new set of legitimating symbols and beliefs. This model fits the twentieth-century Russian and Chinese Communist revolutions reasonably well, the French Revolution of 1789 somewhat, and most early modern state crises hardly at all. It is thus not very useful for discussions of the early modern period. The use of the term "revolution" has been further muddled by such phrases as "unsuccessful revolution," "social revolution," "abortive revolution," and "elite revolution," all of which are meant to imply something distinct both from each other and from such events as coups, civil wars, and popular rebellions. Because there have been few events that fit the extreme sense of "revolution," yet many cases of governments being overthrown or temporarily disabled, there have developed almost as many varied definitions of "revolution" as there are analysts and cases of state crises.

Of course, the term "crisis" is equally ambiguous. Scholars often use the term to denote any threat to law and order and thus have found "crises" in a bewildering variety of times and places. As the term is loosely used, a "political crisis" may denote a crisis of confidence in the government, a state bankruptcy, a coup d'état, an elite revolt, a peasant rebellion, an urban riot, or a civil war. I shall use state crisis specifically to describe a shift in elite or popular attitudes toward the state. That is, a state crisis is a situation in which politically significant numbers of elites, or popular groups, or both, consider the central state to be operating in a manner that is ineffective, unjust, or obsolete. Such a state crisis may stem from actual failures of governmental performance, such as bankruptcy, defeat in war, or inability to suppress local disorders. Or it may stem from changing economic conditions or reckless governmental actions that cause elites or popular groups to lose confidence in, or withdraw their allegiance from, the state. Either way, elite or popular groups consider the state incapable of performing necessary tasks of governance; the result is that the state loses the allegiance necessary to

govern. But a state crisis is not bankruptcy or military defeat per se ; states often survive such problems without losing the allegiance or confidence of elites and ordinary subjects. For example, Spain in the late sixteenth century, and France at the end of Louis XIV's reign, suffered military setbacks and state bankruptcies. But elites rallied around the state, and widespread popular disorders did not arise. A state crisis occurs when there has been a standing suspicion that the state is ineffective or unjust; such concrete events as a fiscal crisis or military reversal are then taken as the definitive proof of this suspicion. It is thus the shift in elite or popular attitudes toward the state, not a particular event, that marks the crisis of state authority.

Much of the political science literature describes such a shift in attitudes as a crisis of "legitimacy" (Zimmerman 1979, 1983). I prefer to avoid this term; it does not capture my meaning—in part because legitimacy has legalistic connotations that suggest a "legitimate" government is duly constituted whereas an "illegitimate" government is illegally constituted. These considerations are often irrelevant. Moreover, "legitimacy" is inadequate because this term generally focuses attention only on matters of justice and ignores state effectiveness. Political allegiance depends on the extent to which elite and popular groups view the state as just and effective in performing the duties of governance. Even legally constituted governments can become perceived as ineffective or unjust, while even illegally formed governments (such as those produced by revolutions or coups) can be perceived as effective and just in their actions. Moreover, even an unjust state may be perceived as so effective that it maintains the allegiance of key groups; and even a just state may be perceived as so ineffective that it loses that allegiance. In sum, a state crisis exists when politically significant numbers of elites, or popular groups, or both, no longer maintain allegiance to the existing state (as it operates as a set of institutions, not just its incumbents), regardless of how the state was formed and regardless of whether this shift in allegiance is due to the state acting unjustly or merely ineffectively, or to changes in the economy or international environment of which the state is victim.

A state crisis usually indicates a situation of imbalance—in the eyes of influential elites and of large numbers of ordinary people, the state is either failing to perform the expected tasks of governance, demanding too many resources for that task, or both. Thus the routine tension over state performance and state revenue escalates to unusually high levels. What will result from this crisis, whether it will pass lightly or lead to

overt conflict, depends on the flexibility of state authorities, on the unity and organization of elites, on the mobilization potential of popular groups, and on the precise relationships among these actors, including their financial, organizational, military, and ideological resources. A state crisis may lead to revolution; but it may also lead to an unsuccessful attempt at revolution (as in Prussia in 1848) or to a successful reform (as in the English reform crisis of 1830-1832).

This book seeks to explain a particularly severe kind of state crisis, which, as noted above, I call state breakdown . State breakdown occurs when a state crisis leads to widespread overt conflict, including a combination of elite revolts, intra-elite struggles, and popular uprisings. A state crisis may be resolved peacefully if elites shore up state power, or if reformers succeed in rectifying state injustices. Or a state crisis may be resolved with a coup d'état if important elites are united and able to achieve desired changes without popular mobilization. But if elites are highly alienated from the state, command substantial resources, and are divided among themselves, a state crisis may lead to elite revolts and sharp intra-elite conflicts. And if popular unrest is waiting in the wings, conflict between the state and elites may open the doors to popular uprisings or to mobilization of the population to support competing factions. Struggles for power among different groups may then lead to civil war.

State breakdown thus refers to a condition of grave disorder, with a collapse of state authority. Such a situation is sometimes called a revolution. However, I shall reserve the term "revolution" for those cases where state breakdown is followed by substantial changes in political and social institutions and in the ideology used to justify those institutions.

To sharpen these definitions, and to clarify the differences between the wide variety of political crises that have occurred in history, it is useful to adopt some simple notation borrowed from vector algebra.3 Let us divide the notions of state breakdown and revolution mentioned above into several constituent elements. We then have a range of factors that might be present or absent in a political crisis: (1) widespread elite or popular belief that the state is ineffective, unjust, or obsolete, producing widespread loss of confidence in, or allegiance to, the state; (2) an elite revolt against the state; (3) popular revolts—either urban, rural,

The application of this technique to holistic comparisons in the social sciences is developed at greater length in Ragin (1987).

or both—against state or elite authority; (4) widespread violence or civil war; (5) a change in political institutions; (6) a change in the status and power of traditional elites, chiefly landlords in agrarian societies; (7) a change in basic forms of economic organization and property ownership; and (8) a change in the symbols and beliefs that justify the distribution of power, status, and wealth.

We can then describe a given historical event by a string of ones and zeros, which designate whether a given element among the eight factors above was or was not present. Thus the Russian and Chinese Communist revolutions (counting the revolt of the Dumas in Russia as an elite revolt) would merit a complete row of ones: (1,1,1,1,1,1,1,1). The French Revolution differs slightly, for even if one counts the loss of seigneurial power of French landlords as a significant change in landlord power, there was no significant change in property ownership or in the forms of economic organization. Thus we would designate the French Revolution (1,1,1,1, 1,1,0,1). Stability would be noted by a complete row of zeros: (0,0,0, 0,0,0,0,0). A peaceful reform of government could be designated (1,0,0, 0,1,0,0,0). This definitional system is extremely versatile. Thus one can describe an ordinary coup (0,1,0,0,0,0,0,0) or an elite revolution (1,1, 0,0,1,1,1,1) without confusion. A dynastic civil war that posed no challenge to the state per se, and was only a contest among elite leaders for succession, as frequently occurred in the Ottoman and Mughal empires, could be specified as (0,1,0,1,0,0,0,0), while a secessionary civil war, such as the American Civil War, could be designated (1,1,0,1,1,0,0,1). More importantly, one can express the differences between such diverse historical events as the English Revolution of 1640 (1,1,1,1,1,0,0,1), the Ming-Qing transition (1,1,1,1,1,1,0,0), and the French Revolution (1,1, 1,1,1,1,0,1) more precisely than by arguing whether or not the first two were "truly" revolutions. In fact neither was exactly like the French Revolution, nor like the other, although all shared many of the same features.

This definitional system is capable of specifying 128 different kinds of events, ranging from stability to extreme revolution. Of course, many of these definitionally possible events have no attested empirical contents. For instance, I know of no popular rebellion that succeeded by itself without associated elite revolts or elite leadership in creating institutional change (1,0,1,1,1,1,1,1). But this notation allows us to speak more precisely of the varieties of political change without being bound by the accumulation of meanings that have surrounded the word "revolution." In particular, we can specify that by state breakdown is meant

any event that involves a crisis of central state authority, elite revolts, popular uprisings, and widespread violence or civil war: (1,1,1,1,x,x,x,x). The x's in the last four slots mean that these may be zeros or ones. Thus the object of explanation in this book is a severe crisis involving the first four elements in combination, whether or not this crisis produces the kind of changes we might wish to call a revolution. In particular, my goal is to explain why this severe kind of state crisis periodically erupted all across northern Eurasia in certain decades but was generally absent in others.

In this book I examine, in varying degrees of detail, a number of instances of state breakdown in the early modern world. I consider mainly four cases: the English Revolution in 1639-1642; the French Revolution in 1789-1792; the Anatolian rebellions in the Ottoman Empire (starting roughly in the 1590s, with disorder peaking at the assassination of the sultan in 1648 and ending with the suppression of Abaza Hasan Papa's revolt in 1658); and the fall of the Ming dynasty in China (c. 1644). In addition, I more briefly examine the Fronde and the revolts in the Spanish Hapsburg Empire in the mid-seventeenth century; the revolutions of 1830 in France and of 1848 in France and Germany; the Taiping rebellion in Qing China, and the Meiji Restoration in Japan. Perhaps only one or two of these events, depending on one's point of view, deserve to be called "revolutions." However, all were instances of state breakdown. I also explore the relative tranquility of the late seventeenth and early eighteenth centuries in Europe and Asia. My primary goal is to show how a common causal pattern lay behind all these events. But in chapter 5 I shall return to important differences.

Before laying out the causal framework, it is worth noting some of the problems in current historiography, and some lacunae in the theory of revolutions, that this framework seeks to redress.

Problems in Historiography

At present, the historiography of early modern political crises is in turmoil. For both the English and French revolutions, "revisionists" have blasted once widely accepted explanations based on long-term social change. For both cases, an influential and learned school of thought now insists that these revolutions in fact had no long-term social causes; instead they were brought on by purely political conflicts and exacerbated by chance conjunctions of unfortunate circumstances.

Until recently, most scholars viewed the English Revolution of 1640

as the culmination of extensive changes in English society. In an influential synthesis, L. Stone (1972) argued that the Stuart state was financially and administratively weak; that social mobility in the preceding century had augmented the importance of the gentry and lessened that of the peerage, creating struggles to broaden access to political power while weakening one of the supports of the throne; and that unresolved religious conflicts left England with a Church that had limited appeal to an increasingly independent gentry, creating a deepening problem of Puritan disaffection. Given the weakness of the state and the growing range of political, social, and religious conflicts, some form of major crisis was inevitable. Although Stone rejected arguments that traced the revolution to the growth of capitalism and the triumph of an English bourgeoisie, able proponents, including Hill (1961), B. Moore (1966), Anderson (1974), and Wallerstein (1980), put forth versions of this Marxist view. Though differing in emphasis, these authors shared Stone's view that the revolutionary crisis could be understood only by tracing its long-term social causes.

In recent years, however, key elements of both these approaches have come under attack. Examinations of parliamentary debates have shown that the political conflicts of the revolution were hardly a long-term running battle over royal versus parliamentary prerogatives; instead Parliament showed little tendency to criticize the Crown or to limit its authority before 1629 (Russell 1979). Similarly, studies of Puritanism show that the Puritan clergy, following their disciplining and co-optation by Elizabeth and James I, were quite cooperative with the Crown until the Arminian assault of the 1630s (Richardson 1973). Work on rural capitalism has uncovered evidence that by the early seventeenth century the Crown was a leader in raising rents, enclosure, and the search for profits (Thirsk 1967a); the image of an anticommercial Crown facing an aggressively entrepreneurial "bourgeois" gentry thus appears false. Moreover, the House of Lords appears to have played a major, even dominating, role in many of the parliamentary maneuvers of 1640-1642, casting doubt on the "rise of the gentry" (C. Roberts 1977a). In short, the old synthesis appears eroded. Still, all that has been accomplished is criticism; no alternative long-term explanation has appeared in its place. Revisionist scholars thus conclude that there were no long-term causes, and that the crisis of the 1640s, far from being inevitable, was a simple product of policy errors of the 1630s in the context of England's unique political institutions, namely, a dependence on the cooperation of Crown and gentry in local and national admin-

istration. This system could not survive a monarch that systematically antagonized the gentry and bungled its political appointments and policies. It was thus simply Charles I's mistakes that produced the revolution (K. Sharpe 1978b).

A strikingly similar trend in scholarship has affected studies of the French Revolution. The dominant synthesis for the last century has been the Marxist view that this was a "bourgeois" revolution in which an emerging capitalist class asserted its rights against a conservative Crown and nobility. This synthesis was and still is vigorously defended by an influential group of French scholars—Mathiez (1928), Lefebvre (1947), Soboul (1975), Mazauric (1970), Godechot (1970, 1971), and Vovelle (1984). Yet a new generation of American and French scholars has largely abandoned this view, for its elements have been undermined by critical scrutiny. Most damaging is the evidence that the capitalist bourgeoisie played a relatively limited, largely provincial, role in the revolution; the revolution's leaders in Paris were drawn from professionals and, to a rather large degree, the nobility (Eisenstein 1968; Egret 1968). G. Taylor (1972b) has shown that both nobles and prosperous non-nobles shared similar economic profiles. Doyle (1972) has presented evidence to question the old notion that eighteenth-century France was marked by a "feudal reaction." C. Lucas (1973) has suggested that conflicts within the nobility were as important, if not more important, than conflicts between nobles and bourgeoisie. Furet and Richet (1970) have argued that the conflicts that fatally weakened the French state were not between distinct classes but rather among the varied elements of the absolutist administration, drawn from different status groups and increasingly divided in their access to power and social prominence.

French absolutism relied on a congeries of separate but overlapping institutions—parlements and provincial estates, aristocratic provincial governors and recently ennobled intendants, court officers and private financiers, an army dominated by gentlemen officers, and a civil administration dominated by judges and lawyers. Keeping this administration in operation required a skillful hand in both economic and political policies. Furet (1981) has argued that a combination of severe weather that weakened the agricultural economy and the inability of Louis XVI to handle the multiple conflicts within the ramshackle but extensive and expensive French state led somewhat fortuitously to state breakdown. Not a crisis of capitalism but rather a crisis of absolutist administration brought down the old order. Thus the French Revolution, like the English, has been reduced from the culmination of long-term underlying

social changes and stresses to the bad luck and misdirection of an unhappy monarch, unable to meet the managerial demands of France's uniquely complex administrative system.

While the sharp criticisms of the revisionists have satisfied the urge to unearth new facts and to uncover false impressions, the total effect has been deeply dissatisfying to many historians, who now see the most momentous events in English and French history construed as mere accidents. Unwilling to abandon the satisfying old syntheses, yet shaken by the revisionist attacks, many scholars find themselves in a quandary. In regard to the English Revolution, Sayer (1985, 3) notes Marxists' troubled response to critiques of the Marxist interpretation: "The 'bourgeois revolution' is extraordinarily difficult to pin down in England. The 'classic ground' obstinately, infuriatingly, refuses to fit the classic models. The usual strategy for dealing with this is to retain the models... whilst variously 'ad hoc-ing' away England's 'peculiarities.'" Non-Marxists complain that "at present there seems to be no generally accepted interpretation of the English Civil War, and little has been done in the search for a new synthesis other than to demolish the old" (Carlton 1980, 168). What remains is a "chaotic and centrifugal state of social history today, which is flying off in all directions with no large structure to hold it together .... Nothing is left but a ragbag of miscellaneous topics, all fascinating in themselves, but without anything to bind them together" (L. Stone 1984, 47). Regarding the French Revolution, Behrens (1974, 637) similarly has noted that "it has often been pointed out, and is now publicly admitted, that the orthodox explanation of the Revolution in terms of a class struggle will not stand the test of facts. [But] now that the orthodoxy is discredited we do not know what to believe .... In the days of Mathiez and Lefebvre's ascendancy we thought we knew. Now there is no coherent explanation."

Old certainties of Asian history have also fallen under a barrage of new research. Traditional Asian history, from Chinese dynastic historians to the Arab historian Ibn Khaldûn, explained recurrent state breakdowns through the dynastic cycle: imperial families eventually produced weak heirs and corrupt followers, and then succumbed to more vigorous challengers. Modern historians, under the influence of Marxist comparative history, have long reiterated this theme, adding that the cyclic history of the dynastic period was due to the "feudal" character of traditional Asian societies and the absence of Western "rationalization" and progress.

However, recent scholarship on the Ming and Qing economy has

undermined many historians' most cherished myths regarding Asia's lack of change and of economic rationalization. Evidence of advanced commercial enterprise in the sixteenth and seventeenth centuries in China is so great that Chinese historians now speak of this period as containing visible "sprouts of capitalism." Under the Qing, Chinese trade was evidently regulated by local merchant associations whose enforcement of contracts, regulation of weights and measures, and freedom from state interference in commerce equaled or exceeded anything found in the absolutist West (Myers 1982). As early as the sixteenth century, China exhibited regionally specialized agriculture, particularly in rice and cotton, serving vast markets. Moreover, the fall of the Ming dynasty was marked by changes in administration, in land law, and in rural class relations. Specifically, increased centralization, uniformity of taxation, collapse of the power of local landlords and transfer of their authority to the bureaucracy, and abolition of serfdom and its replacement by formally free tenant cultivators (Wakeman 1985) exceeded the changes produced by the English Revolution and paralleled those in France centuries later. Progress in commerce and rationalization may have differed from that of the West, yet the difference can no longer be framed in terms of a complete absence of economic change and rationalization in China. Similar changes have been noted in Ottoman history. These events are described in more detail later. However, it should be evident that the old shibboleths of the "unchanging East" are done for.

In sum, once widely accepted explanations of early modern history are now in tatters. The English and French revolutions, it is argued, were little more than large-scale historical accidents. The evidence for early sophistication and growth in Asian economies leaves it a mystery why the Asian empires experienced simple "dynastic" crises, and why they later fell behind Europe in economic and political development. The basic contrast that once made sense of world history—between an early modern West that progressed by inevitable revolutions and an early modern East mired in traditional stagnation—no longer receives support from historical research.

One might expect that historians searching for better explanations of the seventeenth-century crises would undertake careful comparisons of Eastern and Western political history. Yet at present, such comparisons are almost entirely lacking. A number of modern scholars have endeavored comparative studies of Asia and the West: Braudel, Chirot, Collins, Elvin, Hajnal, G. Hamilton, E. L. Jones, Laslett, McNeill, Mousnier, North, Skocpol, Wallerstein, and Wong. However, all of these studies

have limitations and many deal only tangentially with politics. Braudel (1967), Elvin (1973, 1984), Hajnal (1965, 1982), E. L. Jones (1981b), Laslett (1971), and North (1981) deal chiefly with economic and demographic history; Mousnier (1970b) deals with peasant revolts, and Wong (1983) with administrative capabilities. Chirot (1985), Collins (1980), and McNeill (1982) deal with overall political and economic development but, despite their significant insights, still rest within an essentially Weberian framework of Western rationalization. Skocpol (1979) compares the French Revolution with the modern Chinese Revolution of 1911-1949; she thus pulls these events out of temporal context and fails to shed light on the coincidence of major political upheavals and divergent political and economic developments in the early modern period. G. Hamilton (1984) maps the divergent paths of cultural development in China and Europe and traces their implications for long-range economic change. However, the political upheavals of the Ming and Qing do not figure in his accounts. Wallerstein (1980, 1989) and his colleagues have examined the process of the Ottoman Empire's incorporation into the European trade patterns from the seventeenth to the nineteenth centuries; however, they have not directly addressed the problem of Ottoman state breakdown. In short, historians have yet to examine the major political breakdowns in imperial China and Turkey, and their connections to long-term economic and political development, through comparison with state breakdowns of early modern Europe.

Another curiosity is also apparent. Historians have long recognized that early modern European history shows a concentration of political crises in the period 1550-1650, followed by a century of stability, then another concentration of crises in the period 1750-1850. And they have drawn freely on theoretical accounts of revolution by sociologists and political scientists in their attempts to explain these events (Forster and Greene 1970; Zagorin 1982). Yet the reverse has not occurred; no theorist of revolutions has begun with the empirical clustering of early modern revolutions and attempted to build a theory of political crises consistent with this fact.

The Limits of Theory

Historians of early modern revolutions who turned to social science for explanations (L. Stone 1972; Zagorin 1982) found theories designed chiefly to explain the widespread violence that occurred in the developing world after World War II, violence that was considered symptomatic

of the strains of "modernization." From the 1950s to the early 1970s, most social scientists viewed revolutions as merely the largest events on a scale of political violence that ranged from individual illegal acts, through riots and rebellions, and ultimately to revolutions. They thus focused on generalized models of social "strain" (Smelser 1963). They differed over whether these strains stemmed from individual frustration and discontent (C. Davies 1964; Gurr 1970) or from systemic "disequilibrium" (C. Johnson 1966) or "imbalance" (Huntington 1968) in the development of different social and political institutions. However, analysis of individual-level angst and systemwide disequilibria left little room for an examination of how specific political institutions had worked, or failed to work, in particular historical contexts. Thus social theorists simply plucked elements of complex historical narratives and used them illustratively, often out of context, in building and testing their theories. In turn, historians simply plucked elements of these theories and used the terms to couch their descriptions of particular crises. With a few outstanding exceptions (B. Moore 1966; Wolf 1969), careful comparisons of how political crises had developed in various historical settings, and how they were similar or different, were not undertaken.

This situation changed markedly in the 1970s. The historically grounded work of C. Tilly and his collaborators (Tilly et al. 1975; C. Tilly 1978) challenged the validity of general theories of violence, arguing that more attention needed to be paid to how conflicts developed and how resources for opposition were mobilized in specific historical contexts. Paige (1975), Eisenstadt (1978), Skocpol (1979), and a host of other scholars developed a new "social-structural" perspective on revolutions (see Goldstone 1980, 1982, for a detailed survey). These works generally focused on a few cases and presented extended narratives and analyses. Their goal was not to provide a universally applicable theory of revolutions or of political violence; instead they explicitly sought to understand how different episodes of political violence varied and how historical context mattered to the causes and outcomes of state breakdown. Despite the richness of this work, however, it is not quite applicable to the problem of explaining the "waves" of political crises in the early modern period.

Charles Tilly's (1967, 1978, 1986) work on rebellion and revolution in Europe examines the organizational basis for popular protest and how changes in state-civil society relations affected the form and intensity of collective behavior. Yet Tilly treats these explanatory changes as long-term continuous processes—the growth of state power and the spread

of capitalist organization of production. Though his work brims with insights on which I shall draw, it thus does not address the periodic "waves" of state breakdown.

Skocpol's (1979) work is among the most influential of recent theories of revolution. Comparing the French, Russian, and Chinese revolutions, she noted that each revolution occurred when the state faced pressures from wars with more advanced capitalist states. She also observed that in each of these countries there was some combination of structural weaknesses that created the potential for revolutionary crisis: backward agriculture that could not support a competitive military (Russia); autonomous elites who could block the state's attempts to raise taxes and centralize power (France and China); and strong peasant villages that could readily mobilize for attacks on landlords in the event of a weakening of the central government, under either traditional village leaders (France and Russia) or communist party organizers (China). The conjunction of pressure from abroad, structural constraints on state actions, and peasant organization that abetted effective rural uprisings produced revolutions. Skocpol's theory has not escaped criticism. Gugler (1982) and Dix (1983) have suggested that Skocpol understates the role of urban workers in bringing about revolutions. Eisenstadt (1978) and Sewell (1985b) have argued that cultural differences, which Skocpol neglects, played a large role in shaping revolutionary possibilities. However, Skocpol must be given credit for bringing three observations to the forefront of current studies of revolution: (1) revolutions arise from a conjunction of events—state crises, popular uprisings, and elite actions—each of which may have different causes and therefore must be separately investigated and explained; (2) states are not merely objects of revolutionary struggles but also actors in the social drama whose actions and options are crucial in precipitating revolutionary crises; and (3) revolutions are often the product of international forces that impinge on particular states and interact with their particular institutions.

Unfortunately, this work so far has had little impact on comparative studies of early modern history. The major comparative studies by Anderson and Wallerstein are still wrought in a solidly Marxist framework that subordinates state actions to class struggles. Skocpol's work, which does focus attention on the state and on international military competition, is a valuable corrective. Yet it turns out to be poorly suited to explain the pattern of state breakdowns in the early modern world.

The incidence of war itself provides little guidance to the long-term recurrence of revolution. Much has been made of the impact of particu-

lar wars, such as the Thirty Years' War, which preceded the seventeenth-century state breakdowns in England, France, and Spain, and World War I, which led to the Russian Revolution of 1917. Yet little attention has been paid to the even larger wars that did not produce state breakdown. From 1688 to 1714, Louis XIV brought Europe into almost continual armed conflict, leading some historians to label this period the "Second Thirty Years' War." France suffered extensive defeats in these wars, which were fought with larger armies and cost far more than the first Thirty Years' War. Yet despite defeat and bankruptcy, neither France nor any other of the combatants experienced revolutions. Similarly, the immediate outcome of the enormous conflicts of the Napoleonic Wars (1799-1815) was not revolutions in the major European powers but rather increases in the authority of conservative states in England, Prussia, Austria-Hungary, and Russia. Furthermore, insufficient attention has been paid to those revolutions that erupted in times relatively free of war, such as the revolutions that erupted in Belgium, France, Italy, Poland, Switzerland, Germany, and Austria-Hungary in 1830 and 1848. As these cases make clear, the incidence of war is neither a necessary nor a sufficient answer to the question of the causes of state breakdown.

From 1550 to 1815, there were few decades in which Europe was free from major wars. Moreover, the scale and cost of warfare was constantly growing. Yet in these centuries state breakdown was sharply cyclic, including a peak during the relatively peaceful (in terms of interstate conflict) mid-nineteenth century. How then should we view the connections between war and revolution?

Military pressure, of course, depends on the nature of war, as well as its incidence. The "military revolution" of the sixteenth century has been blamed for increasing the costs of war to a ruinous degree. Yet martial technology was almost unchanged from 1550 to 1850; once earthwork fortifications had countered the invention of the siege cannon, and the musket had replaced the bow, the chief element of war remained the infantryman, equipped with musket or pike. Artillery gradually supplanted cavalry, but the pace of change was slow (van Creveld 1989, 97). What changed most dramatically was the scale and cost of war. These trends, in turn, were related to broader trends in the economy: changes in the size of populations meant changes in the number of men eligible for service, and changes in prices enormously affected the costs of putting those men under arms. And in the periods 1550 to 1650 and 1750 to 1850, European nations experienced unusual bouts

of population growth and price inflation. Thus, for example, the increases in military costs in the sixteenth and seventeenth centuries occurred largely because of a doubling in major states' populations and a fivefold increase in price levels; these trends combined created a tenfold increase in the costs of larger armies and their provisions (G. Parker 1976, 206).

We can therefore follow Skocpol in noting that the pressure of fighting, or preparing for, international conflicts can undermine state finances. Yet we must insist that it is not the mere incidence of such conflict that is crucial; such incidence does not correspond to the cyclic pattern observed for early modern state breakdown. What does matter is how the cost of meeting military pressures impinges on states. And in the early modern era, as I detail in the following chapters, those costs depended chiefly on cyclic trends in population and prices.

Moreover, in the early modern period the comparatively advanced capitalist states—the Netherlands in 1566-1648 and England in 1640-1688—were as likely to suffer revolutions as their adversaries. Thus a theory that sees state breakdown as primarily owing to military pressure from more advanced capitalist states fares poorly for this period. In addition, as noted above, Skocpol largely neglects urban tumults, which were crucial to state breakdown in these centuries. She also neglects the impact of cultural differences, without which one can hardly comprehend the varying effects of state breakdown in early modern Europe and Asia. Crucial deficiencies, therefore, have limited the impact of Skocpol's work on the comparative history of early modern Europe and Asia.

Problems in the theory of revolutions have also been explored in more narrowly focused studies. Trimberger (1978) extended the structural theory of revolutions to cover cases of "revolution from above." She noted that Japan in 1868 and Turkey in 1921 both experienced state breakdown and new institution-building, yet in each case the struggle for power was relatively brief and confined to elites. She maintained that this kind of elite revolution also was due to states coming under pressures from more advanced states abroad, but each occurred where there existed, instead of the structural weaknesses cited by Skocpol, a highly professional bureaucratic elite, devoted to government service rather than to ownership of land, which had the flexibility to reshape institutions to meet external pressures. Goldfrank (1979) has used a structural approach—emphasizing the place of Mexico in the international system, conflicts among the elites, and regional variations in the organization of peasants—to examine the origins of the Mexican Revo-

lution. Abrahamian (1980) has used a similar approach, albeit emphasizing urban organization rather than the organization of peasants, to analyze the recent revolution in Iran.

Wolf (1969), Paige (1975), Migdal (1974), J. Scott (1976), and Popkin (1979) have waged a debate over the factors that govern peasant participation in revolutions, with Scott stressing cultural factors, Paige stressing the economic relations of cultivators to landlords, Midgal and Wolf stressing the penetration of villages by capitalist organization and the impact of population growth, and Popkin emphasizing peasants' calculations of advantage within the traditional village. All of these factors no doubt play a role, but their primacy in specific situations has not been resolved. One thing does appear clear, however: peasant participation in revolutions is rarely a simple protest against traditional exploitation. Instead, peasants act when they have the opportunity, which involves village mobilization and weaknesses in landlord or state control, and when the terms of traditional exactions are changing, which may be caused by population shifts, changes in marketing or agricultural practices, or changes in the state's and elites' opportunities and needs.

Regarding urban actors, Rudé (1964) has studied police records of revolutionary crowds and found that, far from being irrational mobs, these crowds were made up chiefly of established workers and artisans seeking to defend their economic interests. Further studies of workers' movements by Calhoun (1983b) and Aminzade (1981) demonstrate the manner in which revolutionary movements have drawn power from workers' defense of traditional rights. Rejai and Phillips (1983) have examined revolutionary leaders and found that leaders rarely precipitated revolutionary crises; instead revolutionary situations—state breakdown and conflicts over authority—gave scope for individuals who would likely have followed traditional professions to emerge in revolutionary roles.

Increased attention has also been devoted to the outcomes of revolutions. Skocpol (1979), Eckstein (1982), Walton (1984), and Tardanico (1985), examining mostly twentieth-century revolutions, have identified a number of factors that influence revolutionary outcomes. They suggest that socialist governments are most likely to emerge when economic resources are concentrated in a few capital-intensive centers, when mass mobilization is extensive, and when external pressures from capitalist countries are modest; capitalist governments are favored outcomes when the reverse situation holds.

This volume of work has been impressive in its insights. Current theo-

ries of revolution stress variety. Ranging over a large number of cases and problems, scholars have adopted multicausal, conjunctural explanations. They trace variation in revolutionary conflicts and outcomes to differences in military pressures, differences in the autonomy of elites and of peasants, variations in the resources of states, and changes in the opportunities and pressures imposed on nations by shifts in the international economy. Recent work in the theory of revolutions is also impressive for its willingness to depart from traditional views. For example, Skocpol downplays class conflict, denies that the French Revolution was chiefly a "bourgeois" revolution, and stresses the autonomy of the state. (Indeed, Marxist theories of the modern capitalist state have increasingly conceded an autonomous and critical role to state managers.)

Yet most of this effort is concerned with explaining twentieth-century revolutions, in which the collisions of international colonial and economic forces with traditional regimes played a major role. Where theorists of revolution have looked back as far as 1789, they have, like Skocpol, sought comparisons to more modern events. The theory of revolutions has yet to address early modern political crises in their own right. As a result, recent developments in the theory of revolutions have been of limited relevance to studies of early modern history.

Most historians of early modern revolutions and rebellions have retreated from the Marxist synthesis since it too often has been at odds with the findings of research. Yet recent theories of revolution have considerable flaws of their own, and only limited applicability to early modern Europe. As a result, studies of early modern revolutions and rebellions lack a common framework. To some scholars, this lack of synthesis is unsettling. J. Fletcher (1985, 37-38), whose studies of Mongol history span Eurasia, has recently asked, "Is there an early modern history? Or are there only histories? ... Without a macrohistory ... the full significance of the historical peculiarities of a given society cannot be seen." McDougall (1986, 20), contemplating the impressively wide-ranging findings of historians, has asked, "Can the [historical] profession long survive without some overarching structure to house and organize our sprawling warehouse of special knowledge? Or will it collapse under its own weight?" To date, theories of revolution have not provided an adequate framework for examining the state crises of the early modern world. It is no surprise, therefore, that many historians of early modern revolutions and rebellions, feeling far richer in facts than in overall frameworks, criticize social scientists for their misleading or inadequate offerings.

Toward a Common Framework

Large states of the early modern period, whether monarchies or empires, faced certain common constraints. They needed to raise sufficient revenues to support their armies and reward their retainers. They needed sufficient allegiance from the elites to secure loyal officials for government service and, perhaps more importantly, to secure loyal local authorities in an era when centrally appointed officialdom rarely penetrated below the county level. And they needed to provide sufficient stability and sustenance for the working and cultivating population so that the latter could pay their taxes and other obligations and yet not be inclined to support rebellions. Thus any train of events that simultaneously led to fiscal deterioration, elite factionalism and disloyalty, and a major decline in popular living standards or undermining of popular traditional rights, threatened the ability of states to maintain their authority. In the sixteenth century such a train of events did begin, on a worldwide scale.

A Demographic/Structural Model of State Breakdown

Put simply, large agrarian states of this period were not equipped to deal with the impact of the steady growth of population that then began throughout northern Eurasia, eventually amounting to population increases in excess of the productivity gains of the land. The implications of this ecological shift went far beyond mere issues of poverty and population dislocation. Pressure on resources led to persistent price inflation. Because the tax systems of most early modern states were based on fixed rates of taxation on people or land, tax revenues lagged behind prices. States thus had no choice but to seek to expand taxation. This was all the more true as population increases led to the expansion of armies and hence to rising real costs. Yet attempts to increase state revenues met resistance from the elites and the populace and thus rarely succeeded in offsetting spiraling expenses. As a result, most major states in the seventeenth century were rapidly raising taxes but were still headed for fiscal crisis. Moreover, elites were seeking to secure their own relative position. Population growth increased the number of aspirants for elite positions, and their demands were difficult to satisfy given the fiscal strains on the state. Elites thus were riven by increasing rivalry and factionalism, as pursuit of positions and resistance to state demands led to the formation of rival patronage networks in competition for state re-

wards. Finally, population growth led not only to rural misery but also to urban migration and falling real wages, owing to the especially rapid expansion of youth cohorts that accompanied the population growth. Thus both urban workers and rural artisans staged food riots and wage protests.

Ideological battles flared as well, for state weakness, elite competition, and popular discontent combined to fuel religious conflicts. In these periods, the states' fiscal difficulties undermined their financial support for established churches, and the attempts to use the churches to buttress their increased demands for taxes and other resources led to entwined religious and political opposition. Dissident elites and dissatisfied artisans were thus widely recruited into heterodox religious movements. As all of these trends intensified, the results in each case were state bankruptcy and consequent loss of control of the military, elite movements of regional and national rebellion, and a combination of elite-mobilized and popular uprisings that manifested the breakdown of central authority.

Naturally, these trends showed numerous national and regional variations. Yet, as I demonstrate in the following chapters, the evidence for these trends is strong even in diverse settings. Thus this framework offers a way to understand the political crises of the early and mid-seventeenth century across northern Eurasia, and to explain their common features.

The growth in population that occurred throughout the temperate regions of Eurasia from about 1500 is well known but still not adequately explained. Climate records show a distinct warming in the later Middle Ages that lasted until around 1600, at which time population growth rates also slowed (Galloway 1986). In addition, from 1500 onward the recurrent visitations of plague, which reappeared in Europe roughly every thirty years since the fourteenth century, ceased. In discussing specific cases below, I present the specific population data, as well as the controversies over causation, for each nation. But it is generally well established that a combination of favorable climate and receding disease led to a doubling of the population in most regions between 1500 and the early 1600s.

These conditions did not last. The climate turned distinctly cooler and more variable after 1600; plague returned, accompanied by smallpox, typhoid, and other infectious diseases. By 1650, outside of a few exceptional pockets of growth, population increase had halted worldwide. There followed a century of high mortality, in which population

in most regions stagnated or declined. Yet for those who lived, conditions, paradoxically, improved. With population nearly constant, agricultural improvement immediately raised per capita food supplies; food prices thus widely stabilized or fell. With the labor force not growing but trade and output increasing, real wages began to climb upward. Elite composition stabilized, and elite incomes and royal revenues were no longer eroded by inflation. On the contrary, landed elites and states took advantage of favorable conditions to retrench. The age of Louis XIV was one of short lives and high mortality; but it was also an age of stable prices and relative domestic peace for those who survived, not only in France but across the continent.

Some time in the late seventeenth century, the peak of poor climate and epidemics was passed; by the early eighteenth century, the population had recovered its early-seventeenth-century levels and then began to surpass them. By the second half of the eighteenth century, pressure on the land, and accompanying inflation, were evident throughout Europe and China. Renewed expansion of the cities and of the pool of aspirants to elite positions, combined with falling real wages, led to concentrations of the ambitious and the impoverished in European capitals. In nations with weak financial structures and high expenses, such as France and China, administrative structures were crumbling by the end of the eighteenth century. In nations that had used the favorable interlude of 1660-1760 to streamline expenses and expand their incomes, such as England and Prussia, administrative structures held, although they faced recurrent elite and popular protests throughout the first half of the nineteenth century, culminating on the continent in the revolutions of 1848.

After 1850, railroads and steamships combined with cheap American and Russian grain to end the specter of population outstripping food supplies in Europe. However, in Russia and China population growth continued to exceed agricultural expansion, and land hunger contributed to state breakdowns in both countries in the second decade of the twentieth century. But these events take us far beyond the early modern era.

This model of state breakdown emphasizes demographic changes. Yet it does not rest on those changes alone. The key process is how demographic changes affected critical aspects of the social structure. Thus for each case examined in this book, we shall look at evidence pertaining to the following questions: (1) How did population change? (2) How

did shifts in the balance between population and economic output affect prices? (3) How did price changes affect state income? (4) How did population and price changes affect elite incomes and the recruitment of families and individuals to elite positions? (5) How did population and price changes affect the income and employment conditions of the general population, both cultivators and landless laborers? And (6) how did changes in the relative income of states, elites, and popular groups— and their attempts to cope with those changes—affect ideological alignments, including allegiance to established churches or heterodoxies? Because this explanatory framework combines attention to both demographic change and the ability of various social structures to respond to such change, it is not a demographic model but rather a demographic/ structural model of state breakdown.

Moreover, while this framework emphasizes material changes as the cause of state breakdown, these material changes dictated no particular outcomes. I argue in chapter 5 that state reconstruction offered a variety of choices to elites, and that whether such reconstruction involved radical change or the strengthening of traditional institutions depended chiefly on particular cultural frameworks and the development of elite ideologies . Clearly, material and ideological factors influenced both the causes of state breakdown and the outcomes. However, this framework suggests a particular balance of material and cultural forces; it gives a predominant role to material factors in bringing about state breakdown, but a predominant role to culture and ideologies in shaping state reconstruction.

Was Demographic Change an Independent Force in History?

There is increasing awareness that early modern populations controlled their fertility. Restrictions on marriage, and the practice of infanticide, meant that populations were not simply controlled by their "animal instincts." One may thus ask whether it is proper to consider population change as an exogenous factor in early modern history, or whether population change should itself be considered as dependent on changing economic or cultural conditions.

Recent detailed studies of historical populations allow us to firmly answer this question—long-term changes in population size before 1850 were dominated by independent movements of mortality . Populations

did control their fertility, but this was only a secondary process, undertaken in response to mortality changes. Thus a population experiencing low mortality, and hence rapid growth, would try to slow down that growth by restricting marriage and childbirth. Conversely, a population experiencing high mortality, and hence population decline or stagnation, would encourage marriage and try to avert population decline. But until about 1770 in England, and until about 1850 elsewhere, fertility control worked only to moderate the impact of mortality changes. In every case in this book for which there is detailed information—England and France prior to 1800, and Germany in the nineteenth century—independent mortality movements governed whether or not total population grew. As Wrigley and Schofield (1981, 244) note for England, prior to 1750 "mortality was clearly the more important influence on growth rates." The rule for early modern populations was simple: when mortality was low, the population grew; when mortality was high, the population declined or stagnated. Fertility control merely affected the rate of growth or decline.

I present evidence that supports this claim in some detail in the case studies that follow. Yet we may well ask what caused the worldwide mortality shifts that led populations to alternately grow and stagnate in unison all across Eurasia.

The very breadth of population movements—which moved in the same direction, although at different rates, from England to China— suggests that the answer cannot be found in local economic conditions or culture patterns. Indeed, long-run mortality was clearly independent of such economic conditions as prices and food supplies. Wherever the data for population, food prices, and wages exist, we consistently find that periods of population growth were precisely periods of relatively high food prices and low wages; whereas in periods of low food prices and high real wages populations stagnated. The reason for this pattern is simple: if population size moved for its own reasons and food supplies were relatively stable, then a growing population and rising demand would lead to higher prices, and a stable or declining population to lower prices. Similarly, a larger population and more laborers, if they faced relatively stable employment opportunities, would drive down real wages. Thus the patterns in the data clearly suggest, as Wrigley and Schofield (1981) and R. Lee (1980) note for England, as Dupaquier (1979) and Armengaud (1976) note for France, and as W. Lee (1977) notes for Germany, that population moved independently, for reasons

exogenous to the agrarian economy. In R. Lee's (1980, 547) words, "in preindustrial Europe, as far back as records will take us, population swings were largely autonomous, not a response to economic variations."

Mortality did not respect income. The long-term mortality trends of English peers and French dukes were quite similar to those of the English and French populations as a whole. We are thus forced to seek a cause of mortality change that roamed freely across national borders, different cultures, and income groups. Scholars have hence come to believe that the critical determinants of long-run mortality change were long-run movements in the incidence of disease.

A focus on disease helps make sense of the coincidence of European and Asian population trends, for, as McNeill (1977) remarks, by the thirteenth century Europe and Asia had become linked through a confluence of disease pools. From the time of the Black Death in the fourteenth century, disease swept out of central Asia westward to the Middle East and Europe, and eastward to China. At the same time that the Black Death was ravaging Europe and the Middle East, massive epidemics accompanied the fall of the Yuan dynasty in China (1368). And in the late sixteenth and early seventeenth centuries, when plague returned to Europe, this time accompanied by smallpox and typhus, devastating disease epidemics were again reported in the Ottoman and Ming empires (Dols 1979; Dunstan 1975).

Once a new disease entered a vulnerable population, it did not merely kill off a portion of the population and then disappear. Instead it remained endemic, raising mortality levels for generations, until the virulence of the disease diminished as the population gained resistance. Thus the original impact of the Black Death in 1347 was exacerbated by repeat outbreaks every decade or two until the late 1400s. And the incursion of plague and smallpox from the mid-seventeenth century was followed by repeated outbreaks until the early eighteenth century. These waves of disease helped keep population growth low throughout the fifteenth century, and stagnant from the mid-seventeenth century to the early eighteenth.

We know almost nothing, however, about what signals the outbreak of a new virulent disease. We may ask, in our own day, what accounts for the new plague brought by the AIDS virus, and what its effect would have been had modern medicine not identified the virus and its mode of transmission? The origins of such killers as bubonic plague, smallpox,

and syphilis remain something of a mystery. And why such diseases erupt in worldwide pandemics in some centuries and recede in others is unknown.

It is possible that the answer lies in part in worldwide shifts in weather. The periods of high mortality in early modern history occurred during periods of cooler and more variable weather than characterized periods of low mortality, as indicated by tree-ring growth (Galloway 1986, 7). It may be that cooler weather forced people to spend more time indoors, increasing risks of disease transmission (Post 1985), or that cooler weather led to changes in the migratory behavior of disease-carrying insects and rodents. It also may be that more variable weather— violent alternations of hot and cold, irregular seasons, and severe storms—weakened people's resistance to infection.

At present, these weather-related hypotheses are merely speculation. What we do know is that in the early modern period population growth was controlled chiefly by mortality movements, and that long-term mortality was determined chiefly by outbreaks of virulent disease. High long-term mortality levels were not due to high food prices or low real wages and thus appear to have been independent of economic factors. Mortality levels rose and stayed elevated for generations, as disease outbreaks repeated with slowly diminishing severity. Mortality levels then fell and remained low until the next major disease incursion. The rate of population growth or decline in particular regions was of course somewhat affected by local economic and cultural conditions. During good times, population growth was most rapid in areas with plentiful land to settle, such as Eastern Europe and Russia. But good times and bad times came in unison to all with the movement of contagious disease.

Such ebbs and flows in disease outbreaks are also found many centuries earlier. An unknown plague swept the Roman Empire in the late second century and recurred for several generations. No other epidemic is then recorded until a plague, probably bubonic, wreaked havoc in the Mediterranean and Western Europe under Justinian three centuries later (E. Jones 1974, 134; Le Roy Ladurie 1981).

These episodes of disease incursion may be somehow linked to global weather patterns, or caused by some little-guessed mechanism governing the activity of bacterial and viral agents. In either case, we may take the long-term movement of mortality, and hence of population, in the early modern period as a major independent force in history.

Taking the long-term movement of population as a starting point, we may then ask what were its economic, social, and political consequences.

By relating demographic change to prices, and thence to state incomes, elite recruitment, and popular living conditions, we can develop a potent causal framework for analyzing the timing and nature of state crises.

Post-Malthusian and Nonlinear Aspects of Population Change

The notion that population change was at the root of large-scale historical change has often been considered, and just as often been quickly dismissed. This dismissal stems from two conventional modes of thought: a crude Malthusian view of population change, and linear thinking about causation in history. Lest the preceding model be lightly dismissed, both these modes of thought should quickly be debunked.

Malthus's view of population change is often crudely presented as a belief that population increase results simply from the uncontrollable passion between the sexes. Population would therefore increase until food supplies were exhausted, leading to starvation and disease and a terrible reversal of the previous increase. Malthus in fact recognized that individuals could choose to limit births, and that prudence could keep population in check (Wrigley 1969, 35). But in its crude popular version, Malthusian theory predicted recurrent bouts of starvation as shortages of food provided a sharp limit to population increase.

We now recognize that this crude picture does not apply to early modern history. Populations were not constantly reproducing to the point of starvation. Periods of relatively low population pressure alternated with periods of higher pressure, but in general, popular control of fertility through restrictions on marriage or infanticide kept populations well below the Malthusian limit. Those who consider the crude Malthusian view the only way to envisage the role of population in history may therefore believe that since this theory has been overturned, they are free to dismiss any role for population growth in explaining historical change.

Yet historical demographers and economists have developed rather more sophisticated views of the causes and effects of population growth in history. First, it is recognized that population is not driven merely by births; bouts of epidemic disease may raise or lower long-term mortality, leading to periods of population stagnation or growth. Second, it is recognized that food supplies did not pose strict limits to growth; food supplies sometimes grew faster than population increased, sometimes slower, but in the long run the growth of food supplies has generally

exceeded population increase (Boserup 1981). Third, it is recognized that long before food runs out, distributional effects take over. That is, even though food supplies may be adequate for the population as a whole, an increase in the labor supply that lowers wages, or a slight reduction in food supplies that raises prices, may leave some sectors of the population considerably worse off. That is, even though the population as a whole is well within any "Malthusian" limit, changes in the relative supply of food and labor may make it difficult for many to afford to buy enough food or other items of consumption (Sen 1981).

We are all familiar with the last effect—when demand for homes increases, the price of houses goes up long before we "run out" of housing. Similarly, early modern societies experienced periods when population was growing faster than food output; the result was not crude Malthusian starvation, but rather a shift in prices that led to a redistribution of income and opportunities.

This last point sums up what we might call the post-Malthusian view of population dynamics in early modern history, the view that informs this study. That is, I shall not be looking for periods of starvation driven by runaway births. Instead, I begin by noting that exogenous mortality shifts, probably due to incursions and moderations of epidemic disease, led to some periods when population stagnated or fell and to some periods when population grew strongly. What I shall be looking for are the distributional effects as societies adjusted, badly or well, to changes in the relative balance of food and people. These changes—in prices, and in the welfare of different sectors of society—are the politically important effects of population change.

Nonetheless, those tied to linear thinking about historical causation may still dismiss the role of population increase. Such increases, after all, were gradual and not enormously large. Events such as revolutions and rebellions are both large and sudden; their very nature seems to call for a different kind of explanation. Let us consider these problems under two headings: first, the size of historical causes and effects; second, their suddenness.

The matter of size is simple: to someone bound by linear thinking, a ten percent change in a causal factor should lead to a ten percent change in its effect, a doubling of the cause to a doubling of the effect, and so on—simple proportionality. Yet history is notoriously nonlinear : most causal factors affect different groups very differently (Abbott 1988). Moreover, simple causes may combine and interact, greatly multiplying their effects.

Population increases have a particularly nonlinear effect on marginal groups—that is, groups who face some sort of boundary conditions, such as peasants who are seeking to gain new lands, or younger sons of elite families who are seeking new elite positions. In any overall growth of population, those groups pressed outside some boundary grow far more rapidly. To visualize this effect, consider a circle with a boundary region. We might say those inside the boundary are peasants with land, or sons who will inherit elite positions; those outside the boundary are landless peasants, or younger sons with no position to inherit. What happens if the circle expands? If there is no growth in the bounded area, even a slight expansion of total population will lead to a manyfold increase in the numbers outside the boundary. Even if the bounded area grows (new lands are brought into production or new elite positions created), any growth that is less than the total population increase will still lead to a disproportional increase in the marginal groups. Thus, in the example sketched below, we start with only ten percent of the population in the marginal region. Yet if we double the size of the circle, we end up with fifty-five percent of the population in the marginal region. That is, even though the total population has only doubled, the marginal population increased eleven times! Even if we allow the bounded area (the supply of land or elite positions) to expand by half, we still find that doubling the total population results in increasing the excluded group six and one-half times, from ten to sixty-five. In sum, increases in total population generally produce a much, much larger increase in marginal populations—that is, in those groups competing for some relatively scarce resource—than in the population as a whole. The increase in competition is therefore much greater than the increase in overall population, by itself, would suggest.

Image not available.

To this effect, we must add the interaction of multiple causes. As noted earlier, shifts in demand and supply lead to shifts in prices. In particular, growing demand for a scarce item, such as land or food, leads to price inflation. In early modern history a doubling of population usually led

to a substantial increase in prices—anywhere from a doubling to a quadrupling was common. Consider then the financial impact of the growth of marginal populations on a government. What is the impact of a doubling of population on the costs of providing poor relief for workers without land, or of creating additional elite positions for younger sons who lack positions to inherit? Even in the last case in our simple example, where the marginal population increased only six and a half times, when one multiplies this by a doubling or quadrupling of prices, the financial burden imposed by growth of the marginal group increases by a factor of thirteen to twenty-six! This result is worth repeating: when one considers the impact of population growth on the size of marginal groups, combined with the increase in prices that population growth usually brings, the financial costs of supporting such marginal groups may increase thirteen to twenty-six times, even if the total population only doubles. That is nonlinear response with a vengeance!

Failure to consider such responses can lead to foolish dismissals of the impact of inflation or population growth on states and elites. For example, English historians might note that the English monarchy's revenues tripled from 1540 to 1640, a rate about the same as was seen in price increases. Thus, they conclude, the monarchy "kept up" with prices, and inflation was not a problem. But they fail to note that this inflation was accompanied by an increase in population that led to larger armies. If the size of the army doubled in line with population growth, while prices tripled, the cost of keeping the army would rise by a factor of six , leaving the Crown far short of "keeping up." Similarly, some historians might argue that as the population doubled from 1540 to 1640, and the number of peer families slightly more than doubled, there should have been no increase in "unsatisfied demand" for peerages— after all, they increased more than proportionally to population. But as we now realize, simple proportions do not capture historical change. In this example, if the population was growing slowly before 1540, then there might have been only one surviving younger son unable to inherit a title per ten peer families; if the population was growing more rapidly in the next hundred years, then by 1640 there might have been three surviving younger sons with no chance to inherit a title per ten peer families. The doubling of peer families would then create a sixfold increase in the number of younger sons unable to inherit a title, who would be seeking new positions. When these population dynamics are understood, it can be seen that the doubling of peer families would greatly increase, not merely satisfy, the demands for new positions. In short,

linear or proportional thinking overlooks much of the effect of population change.

The problem of the suddenness of revolutions or rebellions likewise derives from proportional thought. That is, one might think that a sudden event must have a sudden cause. To dispell this notion, let us consider the causes of earthquakes, also sudden events. Prior to an earthquake, heat and kinetic energy from the earth's core create pressures on the earth's crust. Fault lines develop, and pressure builds along these faults. At some point, a small block that had been keeping the crust on either side of the fault in place, despite the pressure, breaks off. The earth is then free to move, and a sudden release of pressure occurs—an earthquake. What has caused the quake? The immediate cause was the breaking off of a block along the fault that held the pressure in check. But the breaking off of that block, of course, was itself caused by the cumulating pressure along the fault. Thus the pressure caused both the "trigger," or sudden releasing event, and the movement of the earth that followed. The sudden earthquake, in other words, was the result of long-term cumulating forces acting on a structure (the crust and its faults) that resisted pressure for a time, then suddenly gave way, releasing the pent-up forces.

The causes of revolutions and major rebellions operate in ways that seem remarkably similar to the forces that build up to cause earthquakes. That is, in the years before such a revolution or major rebellion, social pressures for change build. Yet the existing social and political structures for some time resist change (even though pressures and deformations may be visible). Suddenly, however, some response to the mounting pressure—a state bankruptcy, a regional rebellion—occurs which weakens that resistance (like a block breaking off along the fault). At that point, there is a sudden release of the pent-up forces and a crumbling of the old social structures—a revolution or major rebellion. More concretely, the Scots and Irish rebellions in Great Britain in 1637-1641, and the state bankruptcy and calling of the Estates General in France in 1789, were themselves responses to the mounting social and fiscal pressures in those societies. Yet these particular events also served to unleash far greater social pressures, which overwhelmed these states and led to revolutions.

The causes of revolutions and major rebellions therefore need not be sought solely among sudden events. Such events can be considered as "triggers," or "releasers," of pent-up social forces, but they are not the fundamental causes. Indeed, such "releasing" events are themselves gen-

erally the result of cumulating social pressures. A long-term cumulative factor, such as population increase, can therefore readily lead to a sudden event. What matters is whether the existing social and political institutions are flexible enough to move easily in response to such pressures. Where institutions are flexible, as in modern democratic states, pressures can usually be absorbed through electoral realignments and policy changes. Where institutions are relatively inflexible, as in hereditary monarchies or empires with traditional systems of taxation, elite recruitment, and economic organization, the result is more likely to be revolution or rebellion. The key to this study, therefore, will be to trace how cumulating pressures on relatively inflexible regimes led to crises, and conversely, how the cessation of such pressures led to periods of relative stability.

A last impediment to grasping the role of population in history is the view that causes come from a particular direction. That is, the causes of revolution might be sought from the top down (problems in the nature or capacity of states) or from the bottom up (in the discontent or actions of the populace). Yet again, both of these "linear" approaches distort historical reality.

As students of revolutions and major rebellions know well, the reason these are such "big" events is that they are instances where social institutions break down at a variety of levels . That is, there is a crisis of national government, but there are also crises of local government. There are conflicts with the state, but also regional conflicts and even conflicts within families. There are elite rebellions, but also a variety of rural and urban popular movements. To trace all of these events from one direction, whether top down or bottom up, is virtually impossible. What is needed is a search for causes that operate on a variety of social scales — national, regional, local, even familial. Scientists have invented a term for phenomena that show similarity on different levels or scales of operation: structures with this property are called "fractal." Societies do not operate simply from the top down or the bottom up; instead they exhibit similarity of organization on a variety of scales—national governments are complemented by similar state and local governments; national elites are complemented by similar state and local elites; large cities and businesses are complemented by similarly organized smaller cities and businesses. To explain revolutions and major rebellions therefore requires a causal approach that recognizes this fractal character of societies.

A major virtue of focusing on population pressures is that these pres-

sures operate at all levels of society—nationally and locally, even within families. Although the causal pathways are somewhat varied, as I discuss below, by starting with population dynamics we may avoid a top-down or bottom-up approach to causal explanation and instead focus on factors that operate simultaneously across many levels.

In sum, the model of this book takes a post-Malthusian and nonlinear approach to population dynamics. That is, I shall focus on the distributional effects of relative shifts in population and resources, rather than on massive shortages for whole societies. And I shall emphasize that the impact of population growth is not simply proportional to changes in overall population, but often many times greater, particularly for marginal groups or when combined with price changes. Moreover, it is not necessary to look for sudden or unidirectional causes for revolutions and rebellions. Instead, it may be more useful to examine the impact of cumulating forces on inflexible institutions, and to pay attention to such forces at a variety of levels of society.

A New Synthesis

This causal framework offers a new synthesis in several respects. First, it provides a synthesis of economic and political history quite different from the existing Marxist model. The latter, which has come to be called "the social explanation" of early modern revolutions, argues that the motivation for change came from shifts in the relations of production identified as the growth of capitalism, which produced conflicts between social classes that undermined states. I argue instead that the motivation for change came from ecological shifts in the relation of the population size to agricultural output, which produced diverse conflicts between elites and states, among elite factions, and between popular groups and authorities. These conflicts revolved around a variety of axes, including national and local political authority, status, and economic and religious issues. This causal framework is thus a "social" explanation of state breakdown that does not attach primary importance to the growth of capitalism or class conflict.

Second, this work synthesizes the "new" social history of demographic and social history with the "old" history of revolutions and state crises by mapping the diverse links between them. I draw on changes in population, urbanization, prices, income levels and distribution, social mobility, and education and relate these changes causally to major political crises.

Third, this framework is used to account for similar episodes of state breakdown in Europe and in Asia. It thus offers a synthesis of Eastern and Western history in the early modern period.

Fourth, this framework calls attention to cyclic forces in early modern history. Most long-term histories focus on a secular process, for example, the growth of capitalism, of democracy, or of the state. Yet the early modern period shows marked cycles in numerous aspects of social, economic, and political life. I argue that explanation of early modern history therefore requires a synthesis of secular trends and cyclic processes.

Fifth, this framework synthesizes both structural and cultural approaches. It does so not by simply merging them or by pointing to both sorts of explanatory factors. Rather, I argue that material and structural factors played the key role in producing state crises; however, once begun, cultural factors played the key role in producing certain processes and outcomes . Thus structural and cultural factors both played important, but quite distinct, roles.

Sixth, the methodology used is a synthesis of quantitative statistical methods and the case-centered approach typical of qualitative comparative historical research.

I begin by examining the disputes surrounding the origins of the English Revolution. Chapter 2 uses the English case to show in some detail how demographic change affected early modern polities. For comparison, I then briefly examine two contemporary political events: the state breakdown known as the Fronde in France, and the state breakdown at the peripheries of the Spanish Hapsburg Empire, in Catalonia, Portugal, Naples, and Sicily. Chapter 3 presents the controversy over the French Revolution and attempts to resolve several problems by extending the approach used to explain the English Revolution. Chapter 4 then turns to the state breakdowns in seventeenth-century Ottoman Turkey and Ming China, demonstrating the numerous similarities in their causes to the state breakdowns of the West. I also briefly examine two nineteenth-century cases of state breakdown in Asia: the Taiping rebellion in China, and the rather different and unique case of Japan's Meiji Restoration. Chapter 5 considers why the outcome of early modern state breakdowns differed so greatly between Europe and Asia. This chapter, which focuses on the role of ideologies and cultural frameworks in the unfolding of revolutionary crises, attempts to integrate political breakdowns, cultural

frameworks, and economic history to suggest a reason for the rise of the West, and the relative decline of the East, between 1600 and 1800. Finally, chapter 6 draws conclusions for an understanding of state crises in the modern world, including problems of Third World development and the declining international power of the United States.

The claim that it is possible to offer a common causal explanation for the periodic breakdown of states in widely diverse settings may evoke considerable skepticism. It is valid skepticism, given the simple positivistic models of social change that have been offered in the past and the current confusion about the methods and goals of explanatory social science.

It is true that many sociological attempts to discuss historical change have been inadequate. I believe this is because of self-made barriers that are partly the product of twentieth-century social science and partly inherited from the grand masters of sociology, particularly Marx, Durkheim, and Weber. These barriers take two main forms: (1) a focus on secular , to the exclusion of cyclic , processes in history, and (2) a tendency to simplify social variety into exclusive bipolar categories.

Those uninterested in methodological and theoretical controversies may wish to skip directly to chapter 2, where the substantive argument begins. But in light of the confusion generated by past discussions of history and social science, a brief examination of the issues involved in the comparative-historical method may be of interest to some readers.

One does not apply theory to history; rather one uses history to develop theory.
—Arthur Stinchcombe

Social theory has often hindered attempts by sociology to contribute to our understanding of history. It has done so by taking what should be empirical problems and raising them to the level of theoretical disputes. Unfortunately, as these problems are unresolvable at this level, theoretical discussions have tended to be endless and inconclusive and to divert attention from real historical issues. All too often, social theory has misleadingly classified historical change solely in terms of the "progress" of simple secular processes. And far too often, social theory has

misleadingly addressed historical causation in terms of simple dichotomies. It is therefore necessary to show how these theoretical blinders can and should be cast aside.

In addition, both historians and social scientists are often hampered in their pursuit of theory by a misunderstanding of the nature of "science," owing to their unfamiliarity with the practices of the natural sciences. Far too many historians and social scientists ignore the diversity and richness of modern science, believing that the model for "proper science" is eighteenth-century physics. This unduly narrow view of science, inspiring hopes or fears when scholars talk of making history more "scientific," has driven an unnecessary wedge between history and social science. I shall argue that history and sociology can create a distinct sphere of inquiry—comparative history. This endeavor has its own problems and techniques, and must be understood and defended for what it attempts in its own right and not as a variant of other fields.

The Barriers Imposed By Social Theory

Secular and Cyclic Processes in History

Classical social theory saw social change as linear. Struck by the ongoing "great transition" to modern technical and political organization, nineteenth- and early-twentieth-century social theorists—most notably Marx, Durkheim, and Weber—saw all past change as a prelude to the "great" change. The five hundred years that intervened between the passing of manorial and political feudalism in the fourteenth century and the arrival of industrial capitalism and parliamentary democracy in the nineteenth century were thus classified as a long (and consequently troubled) period of "transition." This view is still with us. Charles Tilly (1984a, 1985) recently undertook the courageous task of drawing an agenda for early modern European social history. He suggests (1984a, 49) that we focus on the "growth of national states and the development of capitalism." Yet to speak of "growth" and "development" implies continuous secular processes and reflects the classical obsession with the movement from feudalism to capitalism. Crucial this change certainly was; yet exclusive reliance on this metaphor of continuous movement distorts our view of the key centuries in which this change took place.

The era from the Hundred Years' War to the revolutions of 1848 was not merely a time of transition. It was a period of distinct economic and political formations: an agrarian, urbanizing, commercial economy and

a semicentralized, partly bureaucratic, dynastic state, in which hereditary monarchs ruled with the aid of appointed officials. In this book I refer to such formations, following Skocpol (1979), as agrarian-bureaucratic states . The degree of commercialization and urbanization varied, as did the degree to which officialdom—generally recruited from a distinctively educated and landed stratum—was subject to the monarch's control. But the economic and political system was quite distinct from the largely local, subsistence, nonurban economies and feudal polities that preceded it, and from the industrial economies and democratic polities that followed. The major European states in these centuries had much in common, in political and economic organization, with the agrarian-bureaucratic states of the Ottoman and Chinese empires that were their contemporaries.

In these centuries, social and political changes had their own distinct patterns. While some changes were under way in which industry and democracy later took root, these were not the only changes in society. In fact, social change in the early modern period can only be understood as a product of linear, or secular, processes and cyclic processes .

It is enormously difficult to fit the history of the early modern period into a simple secular framework, for the centuries between 1250 and 1850 show relentlessly cyclic patterns. In regard to state-making on the European continent, the late medieval crisis and early struggles with the popes gave way to the stronger Renaissance states of the fifteenth and sixteenth centuries. Yet by the seventeenth century these states experienced widespread political crises, facing rebellions and revolutions on a vast scale. In the late seventeenth and early eighteenth centuries states grew stronger still, with absolutism strengthening its grip in France and Germany; but the late eighteenth and early nineteenth centuries were again a period of state crises, with absolutism everywhere defeated or forced to make concessions (even in Prussia, if only slight) to some form of constitutional order. In regard to the growth of capitalism, we find an expansion of the market economy and urbanization in the High Middle Ages, followed by a decline with the onset of the Black Death. The sixteenth and early seventeenth centuries were a period of great commercial expansion; the late seventeenth century, on many indices, a time of stagnation. In the eighteenth century, expansion resumed. Even if we examine an index of "capitalism" as precise as the proportion of English farm laborers who were nonhousehold hired wage-earners (as opposed to traditional in-household farm servants), we find that this ratio rose markedly in the sixteenth century, then fell in the late seventeenth and

early eighteenth centuries, before expanding definitively in the late eighteenth and early nineteenth centuries (Kussmaul 1981, 97-98). And if we look at such vital factors as population and price levels, a similar cyclic pattern forcibly presents itself. How, given the fundamentally cyclic character of so much of early modern history, can we be content to focus primarily on secular changes?

Two answers have been offered to this question. Both are unsatisfactory. The first is to assert that all the cyclic movements observed were merely internal dynamics of a secular process. Thus in the growth of capitalism one must be aware of the "crisis of feudalism" in the fourteenth century, the "crisis of capitalism" (sometimes specified as mercantile capitalism) in the seventeenth century, and the "crisis of capitalism" (specified as the birth of industrial capitalism) in the nineteenth century. The periodic crises, visible in the political as well as the economic spheres, were generated by internal contradictions in the development of capitalism itself. This is the answer offered by Hobsbawm (1965), Wallerstein (1974), and Anderson (1974). I deal further with this notion later on, when discussing particular cases. But it is now widely recognized that the mechanisms by which the "crises of capitalism" are supposed to account for political and demographic crises— making politics dependent on class struggles and demographic growth dependent on income levels—are inconsistent with the historical record of European revolutions and of population change. The question of why capitalism should undergo periodic internal crises has thus proved one of the thorniest, least soluble problems in the Marxist analysis of world history.4 Moreover, from 1500 to 1850 China and Ottoman Turkey had

The contortions that arise from viewing the seventeenth-century crisis as a "crisis of capitalism" are evident in authors as different as Ruggerio Romano and Theodore Rabb. Romano (1978, 205) states that "the 'capitalist' experiment of the sixteenth century ended in the return of the feudal type of economy." Rabb (1975, 99) comments that "capitalism had made no striking gains at the expense of feudalism." Both comments are ridiculous. By the end of the seventeenth century, agrarian labor in Western Europe was essentially free, although the land was still under seigneurial (but not feudal) j ustice and dues; political control was exercised by state-appointed officials and centrally commanded and paid armies; cities had expanded to a degree undreamed of in feudal Europe; and internal and overseas trade had reached new heights with the exploitation of sea routes to Southeast Asia and the New World. Yet after 1700—and this is what Romano and Rabb point out— the power of monarchies and the social and economic dominance of the landed nobility increased all over Europe. In short, Europe after 1650 was no longer feudal, nor yet capitalist; it comprised agrarian-bureaucratic states with commercial economies. Since there is no room for such a category with its own dynamics in the Marxist view, debates become meaningless. One can claim that since Europe was no longer feudal after 1400, it was "essentially" capitalist, as argued by Wallerstein and Braudel; one can equally claim that since Europe was not yet capitalist before 1850, it was "essentially" feudal, as argued by Romano, Rabb, and most historians of Old Regime France, following Soboul (1977b). By arguing whether to assimilate early modern European society to one or the other of two social formations that did not exist—except in the most partial, local, and fragmentary manner—after 1400 and before 1850, scholars can only be diverted from understanding the dominant social relations of the period.

Nor is this a problem only for Western historians. For most of its imperial history, China too was an agrarian-bureaucratic state and thus difficult to fit into Marxist categories (unless one accepts "Asiatic despotism" as adequate). As Rowe (1985, 284) relates, the Chinese scholar Tao Xisheng, facing the same predicament as his Western Marxist counterparts, "found himself forced to declare virtually all of China's recorded history [over 2,000 years!] a 'transitional age' suspended between two Marxist social formations."

the same cyclic trends of state building and state breakdown, and of population and prices, as Europe, in almost exact synchrony. Yet their economies were only very weakly integrated into European trade before 1800. (In subsequent chapters I examine how small their links with Europe were, compared to their domestic economies.) In the early modern era Ottoman Turkey and China were powerful, substantially independent, and noncapitalist economies. How then can one explain political and demographic cycles that are found on a worldwide scale by pointing to the dynamics of a process—the growth of capitalism— that occurred only in Europe?

A second answer is to ignore that these cyclic movements occurred on a worldwide, synchronized scale in the early modern period and instead treat each fluctuation or crisis as a local phenomenon. Thus English population movements are explained in terms of English agriculture, French political crises in terms of the development of the absolutist state, and Chinese peasant rebellions as part of the dynastic cycle, even though the timing of population movements, political crises, and peasant rebellions generally coincided almost exactly across all of these states.

While historians might indeed insist on the uniqueness of each crisis, this view gives rise to occasional paradoxes. As J. Clark (1986, 24-25) has rightly pointed out, historians analyzing seventeenth-century England often describe the period from 1540 to 1640 (or to 1688-1689) as marking the final decline of royal and aristocratic prestige and the crucial turning point in the liberation of Parliament from king and aristocracy. At the same time, historians analyzing nineteenth-century England describe the period from 1740 to the Reform Bill of 1832 in precisely the same way. They cannot both be right. Unless, that is, the late seventeenth and early eighteenth centuries marked a revival of the power and prestige of the monarchy and aristocracy, so that the battle fought in the seventeenth century needed to be fought again in the nineteenth. Allowance of cyclic elements in political and social trends resolves the paradox.

As another example, historians of sixteenth-century England trumpet the "discovery" of the poor that led to the Elizabethan poor laws (Beier 1989). But historians of the nineteenth century, discussing the New Poor Law of 1834, similarly note the sudden "discovery" of the poor and attribute it to a rise of social conscience associated with industrialization. These claims would be of some interest to historians of poverty in eighteenth-century France (Forrest 1981; Fairchilds 1976; Hufton 1974; C. Jones 1982), who note that in this period the French "discovered" the problem of poverty, and that indeed it was "perhaps the most serious social question which faced Enlightenment France" (C. Jones 1982, 29). What is going on? The answer to this paradox is again resolved by a focus on cyclic forces. In the sixteenth century, population growth flooded labor markets, leading to previously unknown, widespread unemployment among the able-bodied. This new social problem then attracted attention. It faded after 1640, when population stagnated throughout Europe and labor markets tightened. It returned in eighteenth-century France, and in nineteenth-century England, when population growth renewed and again flooded labor markets, producing again a wave, not seen for several generations, of massive unemployment and underemployment among the able-bodied. In short, the problem of unemployment, not merely poverty, was not "discovered" by this or that society. It was a real problem that periodically recurred as a consequence of cyclic forces—in this case, population growth.

Classical social theory emphasized secular social change and the differences between Western Europe's and Asia's cyclic development. Yet early modern history everywhere shows evidence of cyclic changes. We could strain to keep the theory intact by tracing all the cyclic patterns to some contortions or internal contradictions of secular processes and thus fit history to the theory. But as discussed below, it would be a poor fit indeed. Alternatively, as Arthur Stinchcombe suggests in the epigraph above, we could investigate the history to learn the nature of the cyclic processes and, where needed, use the history to develop better theory, blending secular and cyclic change. That is the course I follow in this book.

Other pitfalls posed by social theory also lie in wait.

The Pernicious "Problem of Order"

One reason for the slow development of social theory is that social theorists often start by seeking to solve "the problem of order": how can

individuals create and maintain patterns of group behavior that transcend the lives and intentions of the individuals involved? (See T. Parsons 1937 for the classic formulation of this problem.) Yet it is absurd to reduce all the complexities of social behavior to a single "problem of order." Do natural scientists attempt to solve "the problem of nature"? No. Though a grand unified theory may be the holy grail of some physicists, most scientists—biologists, chemists, geologists, astronomers, zoologists—get on with the business of solving the myriad of distinct problems presented by nature, using a diversity of models and theories to solve their particular problems. Perhaps, someday, models developed for different problems will be shown to have common elements, and a reduction in complexity can be achieved. But the first order of business is to solve specific problems in research; reduction is an elegant final course, not the initial step.

Yet sociology seems to have gotten it backwards. The interminable arguments over whether "the social order" is based on conflict or consensus, on whether "social change" is founded primarily on material or ideal factors, and on whether "micro" or "macro" behavior is the fundamental object of sociological concern, reflect this notion that there is a problem of social order that, once solved, will allow all social behavior to be explained and understood. It should be evident by now that no such single solution is possible. History and current experience present many, many kinds of social order—that is, sustained patterns of multi-individual behavior. Some are based on conflict, others on consensus; some instances of social change may be primarily rooted in material changes, other instances may be primarily rooted in ideological factors; most show a mixture of both. The only way to find out where a particular social order or social change of interest is on the scale of these factors is by close empirical examination . In other words, the problem of order is not a single problem, and hence not amenable to solution by theoretical clarification and reduction. It is an empirical problem—or rather problems, for there are many actual social orders, and hence many different valid solutions to their description. The useful task, on which cumulative progress can be made, is to identify particular behaviors, or social orders, of interest, and seek to establish how they are developed and maintained, being aware that different answers are likely to hold at different times and places. To seek first a univeral answer is to endorse a fruitless and primitive monism.

To compound its difficulties, sociology often adds the error of Manicheanism to its monist view of the social order. That is, if there is one

problem of order, and a single solution, then everything other than that solution is considered an evil deception. Modern social theory tends to divide its adjectives into opposed pairs, declaring one element of the pair to be essential or fundamental, the other to be secondary at best, mere superstructure or even illusion at worst. Too often, sociologists treat "micro" and "macro," "ideal" and "material," and "conflict" and "consensus," as mutually exclusive categories, rather than as ideal types that denote extreme ends of an empirical scale. Since reality is dominated by various mixtures in the middle, arguing about the polar extremes merely diverts attention from real social processes.

To give an example, consider the terms "micro" and "macro". As Alexander and Giesen (1987) have pointed out, this is a purely conceptual distinction; in the real world, there are not separate "micro" and "macro" behaviors. All social behavior involves both individual actions and socially generated resources—language, symbols, or institutional arrangements.

Yet debates about this distinction have steered sociology on the wrong course. Scholars use exchange theory to explain the microfoundations of macrostructures (Homans 1961), or "structuration" (Giddens 1976, 1979) or "linkage" (Alexander and Giesen 1987) theories to emphasize that microactions are constantly generating and maintaining macro-structures, and vice versa. But they have not escaped from the dichotomous concept itself. And there lies the nub of the problem. For social behavior is not merely an interplay between "micro" and "macro" levels. Society exhibits order on a multiplicity of scales . Indeed, many aspects of social structure are neither micro nor macro; they are fractal .

Geometry has recently embraced the term "fractal" to denote structures that show similar features, regardless of the scale on which they are observed. For example, a shoreline is indented with bays and inlets; but each bay has its own little bays and inlets, and so on. It is the same with society. At the national level one finds national political authority, national business firms, national political parties, national status hierarchies, national unions, national voluntary organizations, and the like. But within the nation are provinces that look and act like little nations; within the provinces are counties and municipalities that are similarly structured; within the municipalities are local voluntary groups, local clubs, local business firms, and the like, which are in some ways similarly structured. Even families—as settings for authority, conflict, coalition formation, and intergenerational conservation of property—have certain characteristics in common with larger societies. Social order is

thus reproduced and maintained, in a largely self-similar fashion, across a variety of scales. A focus on the micro versus macro distinction excludes an awareness of this whole phenomenon .

Of course, society is not fractal in the same way as a natural formation, which may be identically self-similar across different scales. Precisely because social structure is the product of human intent and historical accident, rather than of self-replicating natural forces, a key problem is to establish to what degree social order is similar across different scales and different dimensions (e.g., political, economic, kinship) of a given society. To emphasize this distinction from the natural sciences, and keep social intention in mind, we could describe social structure as "near-fractal": a structure that is ordered—by a combination of conscious intent and historical accident—on a variety of levels, and whose ordering principles on those levels are largely, although problematically, self-similar.

Once this near-fractal character of society is recognized, a host of empirical problems present themselves. To what extent are the structures of society on various scales and dimensions congruent? For example, a nation-state with elected officials usually has state and local governments with elected officials, and voluntary associations and labor unions with elected officials. Conversely, nation-states dominated by a party apparatus and party-appointed officials usually have party-appointed officials in charge of provincial and local governments, unions, and the like. Yet there are exceptions. How is it that precisely at the time political structures of greater democracy were becoming established in nineteenth-century England, economic structures of greater authoritarian-ism—the factories—were simultaneously developing? Even today, although publicly held firms elect their officers, voting is weighted by share ownership, as opposed to the equal vote of all participants that is preferred in the political system. How do these variations manage to subsist? Are they holdovers on the way to a more congruent future? Or are they stable? To what extent does socialization in family units, which are undemocratic, impede later appreciation of individual rights and responsibilities? Is the movement for "children's rights" an inevitable extension of national norms into the lower levels of the fractal hierarchy?

These issues are of crucial importance, for example, in judging the prospects for successful restructuring in the Soviet Union and China: to what extent can each avoid fractal congruence and maintain party-dominated political organizations in the same society with "free" business firms and elective unions? This crucial problem of modern politics im-

mediately springs forth when one considers the fractal structure of societies. Yet it is completely obscured (and hence ignored) by the traditional micro versus macro categories.

Indeed, the key dynamics of social reproduction and social change generally operate at the level of coherent groups—families, peasant villages, urban working classes, regional elites, business firms, and local governments—that mediate between individuals and the society as a whole. Are these dynamics micro or macro? Of course they are neither, and explanation is lost, rather than gained, by trying to fit them into one category or the other. Jettison the theoretical blinders that confine us to one or the other end of the scale, and the rich reality in the middle is revealed for empirical examination.

Casting aside these blinders is of the greatest importance for the task of this book, for the analysis of state crises is greatly handicapped by focusing on only the macro or micro levels. Macroanalysis focuses on the nature of state structures, particularly the relations of various classes to the state and of the state to other states in the international arena. Microanalysis focuses on the mobilization of individuals, particularly the rewards and penalties accruing to ordinary individuals who participate in revolutionary movements. All these factors are important. But they miss the very nature of major state crisis, which is a breakdown of order on a variety of social scales .

Thus state breakdown typically involves not only breakdown of the central government but failure of provincial and municipal governments as well. For example, at the same time that the English and French monarchies were having financial difficulties and struggling with fractious parliaments (or parlements), in 1640 and 1789 respectively, many English and French municipal governments were nearing crises because of their heavy debts and were struggling with fractious freemen and city corporations. Explanations that focus only on the national government level miss the scope and depth of revolutionary crises.

Moreover, state breakdown typically involves not only conflict between the state and elites (the focus of Skocpol's [1979] work), but also conflict within and among different elites . And not only conflict among elites at the national level, but also conflict among local elites at the county, town, and parish levels. Indeed, conflict among elites often reaches down to the level of the family, with fathers and sons, or siblings, divided by the crisis.

Similarly, working upward from the "ordinary individual's" attitude toward the government overlooks the fact that popular conflicts can take

forms other than uprisings against the state; revolutions also generally show a host of conflicts rooted in lower-level social structures that more directly touch popular groups than conflicts involving the national government. For example, peasant conflicts often involve revolutionary and counterrevolutionary peasant villages pitted against each other, while urban conflicts involve various groups of workers and urban elites on opposing sides. A microfocus on the rewards and penalties facing the "ordinary" individual often misses the dynamics of peasant and urban worker activity, which is mediated by different kinds of local organization.

This book approaches the problem of state breakdown as a problem of explaining the breakdown of institutions on a multiplicity of scales . My goal is to explain the crisis of the state as a crisis of both national and local authority; to explain elite conflicts as a set of nested conflicts that occur on local, regional, and national levels; and to explain popular activity as involving conflicts that differ according to their regional and rural or urban settings. Such a widespread breakdown of order is typical of major state crises. Thus the critical problem is to explain why institutions should break down simultaneously on a variety of scales and dimensions at a particular time . Approaches that are either macro or micro at best address only part of this conjunctural problem. The causal framework of this book, which begins with the impact of demographic and price changes on social structures at various levels, is useful precisely because it illuminates how conflicts and instability can develop simultaneously throughout the varied and multi-tiered structures of a society.

Similarly, the question of whether social order and social change are based on material or ideal factors—or some wishy-washy blend of both—is often posed as a theoretical query that can be answered and then used to guide empirical inquiry. The same is true of the question of whether social order is based on conflict or on consensus. With the question posed in this fashion, any answers can only lead to rejection of empirical variation out of hand. It may be that in some cases, or for certain aspects of a complex social process, material factors (or conflicts) do dominate; in other cases, ideological factors (or consensus) may do so. To assume that it is always material factors (or conflicts), or always ideological factors (or consensus), or always both that provide the key to social dynamics is to theorize historical variation out of existence.

In short, theory can never solve "the problem of order," and indeed sociology committed perhaps its biggest error in approaching its task in this fashion. We have erred wherever we have used theoretical categories

as rigid frames to shape our empirical inquiries, rather than as flexible molds to be shaped by the data.

Some sociologists do tend to begin with theoretical categories and then rigidly apply them to history; this understandably confounds and distresses historians. The latter's tendency to treat sociologists as people describing the world by viewing it through the wrong end of a telescope therefore has some foundation. Yet sociology need not operate that way. Social theory that is responsive to historical variation, and can help to explain that variation, gets to the heart of historical issues.

Comparative history is not the application of social theory to history. What then is it?

Comparative History: a Manifesto

Mais ce n'est pas l'histoire!
—Pieter Geyl, commenting on Arnold Toynbee's A Study of History

Variation in History

Eighty-seven years ago, Mantoux (1903,122) wrote that "the particular, whatever occurs only once, is the domain of history." This view, often expressed earlier and echoed since, has remained dear to historians' view of what they do. Yet we must be clear what this means. The "particularity," the "uniqueness," is the property of a sequence or collection of events. The task of all historical inquiry is to explain those particular sequences of events that have occurred. However, this task would be quite impossible if all of the components, if every aspect, of such sequences were totally unique.

Consider the history of a nation, such as England. If every moment in English history were utterly unique, in all aspects , what value is there in talking of England? Presumably, some aspects were common to the society that existed in the southern two-thirds of the island of Britannia over several centuries. Language, form of government, and ethnic heritage all changed somewhat, but not so much that they were totally different at any given moment from what they were at the next.

Even though historical sequences are unique in toto , it is therefore possible to recognize some aspects of those sequences that are similar to those of other times and places. The historians' claim of the uniqueness of particular sequences then generates a useful problem. For we face not

an a priori theoretical assertion of absolute incomparability but rather an empirical task: investigation of the degree of difference between non-identical, but possibly similar, collections or sequences of events. History, in short, shows both continuity and change. It is thus possible to ask questions such as the following: In what respects did England's eighteenth-century monarchy differ from that of the seventeenth century? What changes occurred in England's economy from 1750 to 1850, and were these changes greater than those in the preceding century? If such questions are unanswerable in principle (and no practicing historian would claim this, except in the heat of debate with recalcitrant social scientists), then the "history of England" can have no meaning.

If there is both change and continuity in history over time, the same variation is found over space . An eighteenth-century traveler from England to France would recognize certain features of the landscape— monarchy, nobility, and agricultural techniques—as differing in detail from the English analogues, but they would not be completely unrecognizable. One can therefore ask about the differences between English and French forms of government (or agriculture, or income, or religion) in the eighteenth century. Again, the historical claim of uniqueness is a matter of degree, to be established by empirical inquiry.

Here too, our conceptualization of events has been a slave to a simple, mutually exclusive categorization generated by a priori theorizing. Historians and social scientists tend to argue over methods as if there were only two ways to look at history—either every historical sequence is unique or else there are general principles or "laws." Neither position is quite true. The reality (and again, this is something every good historian knows and practices) is that of historical variation . Elements with greater or lesser continuity over time vary and combine in distinctive ways across time and space. Doing history—that is, reconstructing, describing, and explaining particular sequences of events—is almost always a matter of identifying which particular aspects of a complex situation are changing and which are remaining more or less the same, and then attempting to answer the question of why.

Cases and Critical Differences

What distinguishes comparative history is its use of the case-based method to study historical variation. A historian might present a long-term history of Europe, or of disease, or of income distribution that spans several countries or several centuries. Such general histories, long-

term histories, or world histories are not comparative history. They have in common with it an interest in large-scale historical variation, and a necessary recourse to secondary sources. But they differ in approach. The long-term or general history attempts to recount what happened within a given, possibly very large, time frame.

Comparative history is more focused, and much more thematic, than such general long-term or large-scale histories. Comparative history begins with questions such as the following: Why did Japan respond more successfully than China to Western incursions in the nineteenth century? Why did serfdom vanish after 1500 in Western Europe, while something very much like it expanded in Eastern Europe? Did ideology play different roles in communist and noncommunist revolutions? Such questions arise from noticing a critical difference in the record of historical variation—two sequences of events that seem to share some characteristics but are strikingly different in others. It may be that the apparent similarities are, on close inspection, illusory. Or it may be that only a small, but crucial, factor separated quite similar situations and led to striking differences. What makes comparative history interesting is that the answer in each case is not known in advance. It is an empirical inquiry to discover what happened and why.

The central goal of comparative history is thus not merely to find analogies or generalities in historical experience. It is to find causal explanations of historical events . Given that historical variation reveals both continuity and change, comparative history proceeds by asking which elements of the historical record were crucial. Thus to study merely the history of two cities, or of two countries, is to practice parallel, but not comparative, history. The latter depends on identifying some key difference between the cases and asking which of the many distinct elements in these cases were responsible for the particular difference in question.

It is common to take different countries as cases. But this is certainly not necessary. Traugott (1985) recently analyzed the forces on each side of the barricades in Paris in 1848—his cases are the insurgents, on the one hand, and the members of the Mobile Guard, on the other. His comparison of these two groups is the basis for his argument that the Mobile Guard remained loyal to the government not because of its class makeup but chiefly because of its socialization under military-style training and discipline between February and June of that year. Similarly, Calhoun's (1982) study of the English working class takes as its cases the artisanal workers of the early nineteenth century and the factory

workers of the late nineteenth century; by identifying the differences between these groups, he is able to challenge E. P. Thompson's (1963) earlier claims of continuity in a single English working class. And Brustein (1985, 1986) has used regions as cases to shed light on patterns of French political protest in the seventeenth and eighteenth centuries.

Both historians and sociologists often ask "sociological" questions of the historical record, such as what was the pattern of social mobility in nineteenth-century Marseilles (Sewell 1985a)? Or they may turn to history to question features of the present-day sociological terrain, such as how did physicians become such a powerful profession in America (Starr 1982)? Answers to such questions draw on both the historians' skills of analyzing primary sources and the sociologists' knowledge of social processes. Such inquiries therefore do produce a blend of the sort Abrams (1982) envisaged, which goes under the name of social history or historical sociology. Though enormously valuable, this specialty is distinct from comparative history, which rests on the case-based approach. In this sense, comparative history is only a part, although perhaps the most controversial part, of a broader historical sociology. (More wide-ranging analyses of historical sociology are provided by Skocpol 1984 and by Hamilton and Walton 1988.)

Robust Processes in History

Comparative history often starts with apparently similar situations, among which there are nonetheless striking differences that call for explanation. Yet the reverse pattern of historical variation also sometimes arises—that is, in historical situations that seem markedly different, similar sequences of events unfold. There are striking similarities between the French Revolution of 1789 and the Russian Revolution of 1917, despite their vast separation in time and enormously different historical settings. Thus the comparative historian may also approach historical variation from the angle of significant similarities : why, in contexts that seem to differ widely, are similar patterns of events observed?

Examination of similar sequences of events in different historical settings has provoked enormous controversy and even scorn. Comparative history has been given a bad name by attempts to overextend limited generalities to lawlike findings. Such pretensions marred the otherwise valuable work of Marx and the Hegelians, including, in his own way, Arnold Toynbee. Suspicions have also arisen that, in seeking generalities,

comparative history is disdainful of facts. But good comparative history does the opposite on both counts: it seeks limited generalizations and draws its strength from careful use of historical details.

First, it should be recognized that comparative history cannot simply "use" historical narratives as data. This is because historians do not tell us what happened in a particular time and place; rather, they argue about what happened. At any given time, certain patterns of events may be generally agreed on, others hotly disputed. Given the nature of historical evidence and the problems of reconstruction, it could hardly be otherwise. That is why it is essential for comparative historians to engage fully the secondary literature of the cases they study in order to be aware of the historical arguments, uncertainties, and issues at stake. A comparative historian does not approach historical scholarship as a miner approaches a mine. Instead, he or she approaches the literature as a historian would, to engage in a conversation about what happened.

Second, identifying similar sequences of events in different historical contexts is not the same as seeking general laws independent of historical context. Rather, identifying such sequences is more like a geologist's activity in mapping different regions and discovering similar fossils in similar rock strata at widely divergent places. The geologist will then hypothesize that a common process occurred in both places and will attempt to carefully reconstruct what that process was. But this process is not a "law" in the same sense as the law of gravity.

A law is a relationship that holds regardless of varying initial conditions. It is true that to make predictions, one must know the initial conditions of a system. Thus to predict the motions of two gravitationally bound bodies, one must know their masses and the distance between them. But gravity per se, and the inverse square law of its operation, do not depend on those masses or positions. The initial conditions are merely parameters that can vary without affecting the operation of the law. The law is independent of the initial conditions. The geologist's claim that a similar process occurred in different sites, however, is precisely a statement about initial conditions. It is a claim that at some particular times in history, physical laws, such as gravity leading to the sedimentary deposition of rocks and chemical reactions leading to fos-silization of animal remains, must have acted on sufficiently similar initial conditions in distinct sites to create a similar rock or fossil bed.

What the geologist has found is evidence of a "robust process." Such a process is not a law but a combination of characteristic initial condi-

tions with particular laws that produces a characteristic outcome. It cannot be used to make precise predictions, for it has no parameters to be plugged into a "law" that then predicts precise outcomes. Thus the geologist would not hope to predict the precise number and location of fossils in a particular sort of rock. Geological processes act in historical context, with initial conditions that are never exactly alike. Yet the geologist would predict the types of fossils, and their rough proportions, if he or she knew that the rock was formed at a time when certain species lived. That is, if the initial conditions were similar (even though not identical), then regular laws would produce similar (though not identical or exactly predictable) results.

Similar robust processes also occur in history. This is because most men and women do act in a discernibly consistent or rational fashion. Indeed, we rely on this consistency every day. We count on generals not to betray their countries, and on bankers and bureaucrats not to defraud or destroy their organizations. Of course, there are traitors and saints; yet we label these individuals with special words to indicate their exceptional nature. If most generals were traitors and most bankers were saints, these words would lose their meaning; and, needless to say, history would be very different.

Thus the key to historical explanation is knowing that, in a given situation, most people will react in some consistent fashion. This does not mean that all people have the same goals. In some groups of individuals, honor may be so important that an honorable defeat is valued more than a base victory. In other groups, spiritual status is more valuable than material rewards. The point is that we do not consider all human action to be random or inexplicable, because we believe we can discover what a given person or group of persons values. We then expect, in most cases, a certain consistency in their behavior.

This does not imply that everything that happens occurs because someone intended it. On numerous occasions—such as when so many people simultaneously seek the same quiet stretch of country trail that it ceases to be a quiet country trail—people make choices that, because of the simultaneous or reactive choices of other people, or because of the lack or misunderstanding of information, lead to results other than what the actors originally intended. Rational-choice theory makes this discrepancy explicit and examines the consequences (Schelling 1960; Olson 1965). This kind of analysis can be extremely useful, for it is quite valuable to be aware of the kinds of circumstances in which the intended

acts of individuals, in their interaction, produce collective outcomes that no one intended or foresaw. Yet what makes daily life, and social science, possible, is our belief that if we can identify certain sets of salient initial conditions that confront a particular actor or group, we can expect that they will react in a particular (though not identical) fashion that produces a characteristic (though not completely predictable) outcome.

A splendid example is the work of Skocpol (1979) on revolutions. Skocpol identifies a historical process that led to social revolutions in France, Russia, and China. The process involves the collapse of state power owing to failure to cope with external pressures from competing states, and to peasant rebellions mobilized at the village level. Though monarchs and peasants sought changes in their respective conditions in each case, the combination of their actions produced social revolution of a kind that neither monarchs nor peasant villagers intended.

Yet to merely describe this process is only a beginning. The key parts of Skocpol's work consist of tracing how the processes unfolded in her several cases. She notes that the processes differed in specific details. For example, states failed to cope with external pressures either because of inadequate economic development (as in Russia in 1917) or because powerful elites blocked changes necessary to improve state efficiency (as in Imperial China and Old Regime France). What makes accounts of this kind convincing is that they show how state leaders, faced with the situations of limited resources and external pressures, were trapped into taking actions that angered elites and led to state paralysis. This created openings for peasant rebellions from below. Again, Skocpol demonstrates how villages in Russia and France with local autonomy and organization reacted spontaneously to the opportunity of state paralysis, whereas villages in China, which lacked autonomy from local elites, could not react similarly until the Japanese invasion and mobilization by the Chinese Communist Party had created a village-level organization in China that was autonomous of local landlords. There is no single cause, or single combination of causes, that created social revolutions. Skocpol presents no "law" of revolutions that would apply in all contexts, regardless of initial conditions. Instead, she delineates a specific historical set of conditions that occurred in similar fashion in several places. Under those conditions—a situation of state paralysis and peasant organization, which could be produced by slightly differing combinations of causes—the reasonable actions of state leaders, elites, and peasants would likely result in social revolution. The identification

of the critical elements of their respective situations that led actors to take similar actions, even though the situations were not in all respects identical (nor were the consequences), is the essence of good comparative history.5

A robust process in history is a sequence of events that has unfolded in similar (but neither identical nor fully predictable) fashion in a variety of different historical contexts. Yet, it is not a mere "limited historical generalization" or an analogy between different events. The statement "European monarchs in the seventeenth century were crowned" is a limited historical generalization. But we have no idea what produced the recurrent events—is mere coincidence an adequate explanation? Or did the choice of coronation as a symbol of monarchy have roots in a process of symbol manipulation and inheritance of a common symbolic heritage? If so, then there is a robust process behind and responsible for the limited generalization. Similarly, one may note an analogy in the contestations of power between the French Assembly and French monarchy in 1789, and between the Petrograd Soviet and the Provisional Government in 1917, and label this situation "dual sovereignty." But is the resemblance coincidental, or is it significant? What makes such an analogy historically meaningful? It is meaningful if we can identify a process of initial conditions leading to similar choices by actors in the two situations—that is, a causal process connecting the similar elements in each situation. As Stinchcombe (1978, 117) has pointed out, "the causal forces that make systematic social change go are people figuring out what to do." Thus a robust process is less than a law but more than a limited historical generalization or analogy. It is a causal statement, asserting that a particular kind of historical sequence unfolds because individuals responded to particular, specified, salient characteristics in their respective historical situations. If the salience of those characteristics is great, then one can reasonably expect that in a wide variety of historical contexts, actors will respond somewhat similarly to them, and their likely actions can thereby be predicted or explained.

Skocpol, in her own explicit methodological statements, has claimed to be applying J. S. Mill's "method of differences." Yet as critics (Nichols 1986) have shown, she does not comply with Mill's method, which calls for all cases to be identical except on the one factor being examined. Skocpol's argument points to clusters of conditions, rather than individual factors, as responsible for social revolutions, and her cases are far from identical with respect to their other features. The search for "robust processes," I believe, is a better characterization of Skocpol's method. Skocpol is thus among those many innovative scholars whose actual methods are far superior to their avowed methodologies.

Enormous confusion has arisen because of the belief of many historians and social scientists that pointing to robust similarities in history is akin to alleging the existence of predictive, determinate laws. Historians argue that since historical sequences are unique and subject to accident, the search for causally meaningful regularities is a hopeless, misguided quest. Social scientists, in contrast, argue that the search for causal regularities should be conducted across large numbers of cases, testing for valid predictions and for exceptions that might invalidate these predictions. But both these views err, for they fail to grasp the difference between robust processes and laws.

To give an example from natural science, Darwin's theory of evolution by natural selection presents a process: when certain initial conditions exist—species with variation among individuals who are in reproductive competition—the law that progeny inherit the traits of parents creates a process whereby traits of the reproductively successful individuals are diffused throughout the population, giving rise to new species. But one can adduce exceptions: laboratory mice (or other domesticated species) do not experience natural selection. Their reproductive success is controlled by their owners, and individual variation is purposefully minimized (among laboratory mice) or moved in specific directions (such as particularly attractive or valuable features of animals or crops). These exceptions, though, do not invalidate the theory of evolution. Precise, determinate prediction is also not at issue. The theory of natural selection makes no prediction of any particular new species, or of the precise time of its emergence.

Evolutionary biologists are quite aware that predicting the future is impossible, since, as S. Gould (1988, 32) observes, "the contingencies of history permit such a plethora of sensible outcomes." Nevertheless, Gould continues, given a set of initial conditions, "we can explain after the fact with as much potential confidence as any science can muster." As I. B. Cohen (1984, 96) notes, "Darwin showed that the way of progress in all of the sciences was not necessarily mathematical in style . . .. [His was the] first major scientific theory of modern times that was causal, but not predictive."

The power of Darwin's theory is that, by pointing to a simple process that has unfolded, in similar fashion, in a number of different circumstances, a large number of details of the historical record can be related to each other in a manner that had previously been unrecognized or misunderstood. The validity of the theory of natural selection depends on whether it can be shown that, for a number of species, the variety of

present species did develop through the diffusion of new traits from prior species. This means a close examination of the fossil record to see if the evidence conforms to the alleged process. One tests an alleged process by examining if, in the actual historical sequence that it is intended to explain, the details of what occurred are consistent with that process.

Similarly, the geological theory of plate tectonics—which details how the earth's crust is divided into plates that move along the earth's surface, pushed out by new rock forming in the oceans from midsea ridges and absorbed by deep-sea trenches—does not allow one to predict when or where an earthquake will occur. What it does explain (and predict) is that earthquakes are (and will be) more frequent in certain regions, those sites along the boundaries of moving crustal plates. Again, the theory is not a mathematical relationship; it is the identification of a process— the motions of crustal plates—not previously recognized. Once identified, knowledge of that process can then be used to explain certain phenomena, including the distribution of earthquakes across regions, the existence of seafloor ridges and trenches, and the age of the rocks on the ocean floors. It is not prediction of a particular earthquake that offers "confirmation" of the theory but rather evidence of certain patterns in the geologic record: seafloor rocks are younger than crustal rocks and are older as one examines specimens further from the midsea ridges; and crustal rocks show great deformation and evidence of stress along long fault-line regions that trace an outline of crustal plates. Again, the value of the theory is not that it predicts any one specific event (it does not), or that it explains every detail of the past pattern (it does not). Rather, the value lies in the identification and explanation of the connections among a great many details of the geological record that were not previously recognized or understood.

The essence of historical explanation, even in the natural sciences, is thus relating events to each other through a process . That process may rely on very simple regularities of behavior, or "laws." Yet mere identification of the laws does not provide the theory. The observation that parents pass traits to their offspring is not the same as the theory of natural selection. To identify the process, one must perform the difficult cognitive feat of figuring out which aspects of the initial conditions observed, in conjunction with which simple principles of the many that may be at work, could have combined to generate the observed sequence of events.

What has confused the debate is Hempel's (1942) assertion that if one knows the laws and the sequence of events, one simply applies the

law to obtain explanation. But this is emphatically not so. In history, the observed sequence of events features uncontrolled initial conditions, that is, an endless number of events of unknown relevance. How then can one know which initial conditions are the relevant ones to use in explanation? One can possess all the relevant laws, and all the relevant data, and still be helpless without knowledge of the process by which they are related. Before Darwin, people knew of the fossil record and knew the principle of inheritance. But they did not know how to combine them until Darwin elucidated the process of natural selection.

How then, if not through prediction, do we test an alleged robust process? As with studies of geology or evolution, the allegation that a particular process in human history occurred in specific historical settings requires a detailed examination of the historical record of those cases. A process can be considered robust if its workings can be traced in different contexts—that is, if in different historical circumstances actors nonetheless met with situations that, in certain respects, were similar, and hence acted, in certain respects, in similar ways. The "certain respects" are presented in terms of a model that abstracts and simplifies certain details of the complex and varied reality. But the validity of the demonstration depends on whether those actual historical details are consistent with the alleged process. George (1979) has described this mode of proof as process tracing —showing that observed outcomes could reasonably be expected to result from the sequence of actions likely to be taken in the specified situation.

This emphasis on tracing of causal processes sharply differentiates comparative history from correlational studies that use many cases, such as the cross-national research often conducted in political science or sociology. Any limited generalizations found in this latter fashion, though useful, are only hypotheses whose basis in causal processes remains to be demonstrated. In this sense, comparative histories and large-data-set correlational analyses are complementary, but not competitive or substitutes for one another.

Good comparative history must therefore "sink a huge anchor in details" (S. Gould 1986, 47). By confronting the historical details and demonstrating previously unrecognized connections between them, comparative history makes its mark. Of course, as with geology or evolution, it is not possible to explain all the details of a historical sequence. Elucidation of a process that connects many details of what happened is sufficient to produce a work of great value.

A Summation

Comparative history does not mine history to develop laws. Let us instead list what comparative history does do. (1) Comparative history uses case-based comparisons to investigate historical variation. (2) It seeks to engage in historical debates by offering causal explanations of particular observed sequences of events. (3) It may develop such causal explanations either by identifying critical differences between similar situations or by identifying robust processes that occurred in different settings. Outstanding comparative histories often do both. (4) It uses simplifying models, but this does not mean it is precisely predictive. (5) It validates its findings by process tracing, rather than simply by correlation. (6) It sinks an anchor in historical details, for its validity rests on how well the relationships it describes correspond to, and make sense of, those details. While its findings may have some predictive value, and may be generalizable, that is not the critical test. Instead, (7) the test of the worth of a work of comparative history is whether it identifies and illuminates relationships heretofore unrecognized or misunderstood in particular sequences of historical events that have occurred.

As Geyl said of Toynbee's efforts, in words that apply to all comparative history, "but it is not history." True enough; it is not the reconstruction of what happened in a particular place at a particular time from the remaining evidence. Nor is it a broad chronological narrative, of the sort recognized as general national or world history. It is a different endeavor, with different goals and methods. Yet it shares history's concern for getting the sequence right, and for human action. It seeks to engage in the same conversation with historians as they argue about those sequences. No less than history, comparative history seeks to explain why certain things happened and how the world we live in today got to be as it is.

This book claims that there were robust processes behind the waves of state breakdown that swept across Eurasia in 1560-1660 and 1760-1860. This book also claims that an awareness of those processes helps to resolve many of the problems in current analyses of the English and French revolutions, and of the Ottoman Empire and China. But pointing out these similar aspects is only part of the story. For although similar processes across Eurasia led to state breakdown, there were critical differences that affected the outcome in each case. In Europe, these differences were mostly structural differences that determined whether the

outcome of crisis would be successful, or unsuccessful, attempts at revolution. But between Europe and Asia (or, more accurately, between northwestern Europe and Asia), there were cultural differences that affected the response to crisis—whether state reconstruction would be innovative or traditional. Understanding the robust processes and the critical differences therefore explains, first, why crises took place all across Eurasia in the same chronological periods and, second, why the detailed unfoldings and outcomes of those crises sharply differed.

This claim can only be judged by weighing the substantive argument and the evidence. Let us therefore plunge into the historians' conversation regarding the English Revolution.

Excerpted from Revolution and Rebellion in the Early Modern Worldby Jack Andrew Goldstone Copyright © 1991 by Jack Andrew Goldstone. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.