To the eyes of a European historian, the United States in 2007 vividly resembles England in 1640 and France in 1787. In 1640s England, King Charles I was trying to raise funds for his wars by manipulating and evading the will of Parliament, a bastion of privilege and special interests. Reinforced by disputes over religion and royal prerogative, the confrontation soon escalated into civil war. The king lost his crown and his head in 1649. In the France of 1787, King Louis XVI was trying to raise funds to meet war-accumulated debt by manipulating and evading the will of the country’s sovereign courts, bastions of privilege and special interests. The confrontation soon escalated into revolution. The king lost his crown in 1792 and his head in 1793.

In early 2007, U.S. President George W. Bush was trying to raise funds for a disastrous war in Iraq by manipulating and evading the will of Congress, a bastion of privilege and special interests. The conflict looked to escalate into a great confrontation. As in 17th-century England and 18th-century France, each side claimed to be defending time-honored rights. President Bush will no doubt relinquish his crown, and save his head, in January 2009. He’s no king and this is no monarchy. But in the meantime we rehearse an ancient drama: Rulers demand funds for military activity from reluctant, powerful institutions that claim to speak for a wronged nation. Under these circumstances, the most a ruler can hope for is grudging consent.

Let’s not underestimate the political value of grudging consent, however. In the short run, it means that citizens and their representatives remain properly wary about the harm that rulers may do. In the short run, democracy thrives on bargained compliance rather than on either passive acceptance or uncompromising resistance. In the long historical run, furthermore, grudging consent opens the path to democratization itself, a path that depends critically on how rulers acquire the means to rule. But how—how do the conditions giving rise to grudging consent do that?

Democracy as a Verb

All systems of rule, whether democratic or undemocratic, survive by finding stable supplies of the basic resources it takes to run a government: means of coercion, administration and patronage. Of course, the mix of crucial resources changes over time. For millennia, kingdoms and empires sustained themselves mainly with soldiers, weapons, food, animals, labor power and strategic information. These days, powerful states necessarily add to that array a much richer supply of information, scientific-technical knowledge and communications media. But the principle remains the same: Effective rule depends on the continuous production of crucial resources. If the resources dry up, rulers lose the means of enforcing whatever decisions they make and state capacity collapses. This has happened often in history, as an anthropologist studying Central America or the Near East could readily show us.

The execution of King Charles I

The main difference between non- or pre-democratic regimes and democratic ones is that the former tend to commandeer resources under threat of coercion, whereas democratic ones draw essential resources mainly from subject populations that have substantial power to accept or reject their demands—populations with what we call “voice.” This is grudging consent at work.

Democracy gives popular voice (however grudging) a significant influence over rulers’ performance. Until widespread rebellion broke out in 17th-century Britain, a totally undemocratic King Charles had no need to worry about what sort of government most of his subjects preferred. In today’s democratic regimes, rulers worry constantly about which possible government actions subjects will or won’t stand for. A lot of democratization has occurred since 1640 because it has developed more voice to sustain it.

When it comes to how rulers actually use available resources, it is helpful to think about democracy as a verb rather than as a noun—in short, to think about how democracy takes shape and operates rather than statically what it “is.” If we do this, two otherwise blurry facts about democracy come clear.

The first is that democracy is not a yes/no, on/off affair: It is a matter of degree, and the degree is virtually always changing. The second fact follows from the first: Democracy is reversible. Just as polities can become democratic according to some agreed standard, they can cease to be democratic according to that same standard. Even relatively democratic regimes fluctuate in the extent to which they are democratic, and, as we will see, the same democracy can be growing both more and less democratic simultaneously, along different tracks of democratic development.

That can be so because the concept of voice turns out to be more variegated than is at first apparent. If we focus more intently on voice, we can reasonably think of democratization as proceeding along four more or less parallel tracks: toward a broader range of popular voices; toward greater equality among these voices; toward the increased binding of rulers’ actions by popular voice; and toward greater protection of popular voice from arbitrary action by rulers and their agents. We can summarize these four parallel democratizing “voice tracks” as growing breadth, equality, binding and protection. By the same token, “de-democratization” consists of declining breadth, equality, binding and protection.

It is necessary to think this way because advocates of democracy commonly forget how often de-democratization has occurred, even in regimes they take as models of accomplished democratization. The historical record shows that democratization ordinarily occurs in waves that are followed by lesser but still substantial waves of de-democratization. Over the past few centuries of Western history, for example, whole clusters of regimes underwent a momentous shift from millennia of being simply unavailable for democratic change to frequent oscillation between democratization and de-democratization. Take the period from 1900 to 1949, during which 17 European regimes underwent at least one period of accelerated democratization. Twelve of those countries (Austria, Finland, France, Germany, Greece, Hungary, Italy, Portugal, Romania, Spain, Russia and the Netherlands, if we count the Nazi occupation) also underwent even more rapid de-democratization at least once.

As compared to the undemocratic 18th century, from the 19th century onward European regimes became available for both democratization and de-democratization as they never were before. Latin American regimes became similarly available for movement in both directions once they declared independence from Spain during the early decades of the 19th century. History threw a parallel switch with the decolonization of Asia and Africa after World War II. The Philippines, Thailand and Senegal, for example, all look like recent democratizers that could well move back into de-democratization, precisely because powerful elites feel threatened by further movement down along the four voice tracks of democracy.

De-democratization can happen anywhere, even in the United States. When the Bush Administration responded to 9/11 by narrowing legal protection for people it suspected of complicity with terrorists, it partially de-democratized the United States. De-democratization thrives today far more widely outside of North America, however. Russia provides a sensational example of recent de-democratization. By the late 1990s, Boris Yeltsin’s Russia had become a semi-democratic regime at constant risk of imploding, but it at least preserved room for dissident voices. Arriving at the top in 1999, veteran KGB officer Vladimir Putin rapidly began top-down reforms that narrowed popular voice, made it less equal, rendered it less binding, and removed some of its protections from arbitrary state action. He de-democratized ruthlessly.

Russia has not been the only country to do so. Every year since 1972, Freedom House has rated each of the world’s independent regimes on political rights and civil liberties. Its political-rights ratings combine observations (using my terminology) on breadth, equality and binding, while its civil liberties ratings concentrate on protections. Together, they provide a rough index of change and variation in breadth, equality, binding and protection—in short, of democracy (as a verb) or its absence. As of 2006, Freedom House reported medium-term downward trends in political rights and civil liberties not only for Russia, but also for Argentina, Azerbaijan, Bahrain, Brazil, Burma, Egypt, Eritrea, Ethiopia, Gambia, Hungary, Iran, Kenya, Kyrgyzstan, Malaysia, the Philippines, Seychelles and Zimbabwe.

Origins of Democracy

It is comforting to think of democratization as a switch that, once turned on, can never be turned off. That no doubt explains why so many people conceive of it in precisely that way, without ever having thought about it at all. As we have seen, however, democracy operates more like a thermometer than a switch, rising and falling depending on the outcomes of continuous negotiations between rulers and ruled over how resources for governance are acquired and subsequently how they are used. But the key prior question is how does democracy arise in the first place?

In mature democracies, most negotiation between leaders and citizens centers on government’s performance—how resources are used. But how rulers acquire their means of rule—soldiers, arms, food and so on—turns out to be more basic to the origins of regime types. In general, rulers have three ways of acquiring their means of rule. They can produce those means on their own, as did kings who ran their kingdoms on rents from their own domains, and as 20th-century communist regimes sometimes managed to do. They can seize control over high-value resources that can easily be exchanged for means of rule, as small clients of large powers have sometimes done, as tribute-taking raiders such as the Mongols did regularly, and as rulers of energy-rich states commonly do today. Or they can extract the means from their subject populations through taxation, conscription, confiscation and borrowing.

The first two of these three paths almost entirely bypass consent, even grudging consent. Rulers who support their regimes from their own property or state-controlled production need only have agents who deliver the proceeds without taking excessive cuts for themselves. Rulers who sell off precious commodities, including their loyalty to great powers, need only hold off competitors for those commodities.

The path of extraction, however, runs differently. However coercive or even tyrannical it may be, extraction always involves some bargaining out of agreements between rulers and their subject populations. Even rulers who crush popular resistance to their extractive activities inevitably make some sort of peace with the rebels, state what obligations the rebels have violated, and announce what future forms of compliance they expect from all concerned. Over the long run, most rulers have sustained their rule with some version of the extraction strategy.

Any of the three paths can in some circumstances move a regime closer to democracy. A self-reproducing regime still depends on compliance by the people who run its farms, mills and armies. If a crisis at the bottom or a split at the top occurs, compliance can turn into rebellion or into fragmentation, during which new rulers have no choice but to bargain for control. To some extent, that’s what happened in the Soviet Union and its successor states between 1985 and Boris Yeltsin’s later years. The system’s capacity for self-reproduction collapsed, and wild competition for control of its fragments followed.

In the case of a regime that barters state-controlled resources for the means of rule, the market for its precious resources can collapse, forcing rulers to seek new support directly from its own citizens. Although oil potentates such as Algeria’s Abdelaziz Bouteflika and Venezuela’s Hugo Chávez currently ride high by selling their precious goods to energy-greedy capitalist countries, a rapid drop in the price of oil would confront them with unpleasant choices between cutting back on government expenditures (thus alienating their many clients) and turning to their own citizens for the means of rule (thus opening a round of bargaining over the conditions of extraction). Neither alternative would guarantee democratization; new strongmen could arise to topple the old ones, for example. But either would open negotiations between rulers and ruled that could lead toward democracy, at least in principle.

Still, the royal road toward democracy tends overwhelmingly to be an extractive one. The American Revolution began, after all, with colonial resistance to the British crown’s attempt to extract from colonists a larger share of the costs incurred in the Seven Years War (1756–63), the conquest of Canada and the subsequent financial burden of running an enlarged North American empire. In 1765, imposition of the Stamp Act crystallized resistance across the colonies, escalating a round of bargaining into rebellion and eventual independence. American revolutionaries by no means invented the idea of democracy, but they did supply visible evidence that united citizens could overthrow tyranny. Dutch, French, Swiss, Irish and even English populists were soon appealing directly to the American model of popular sovereignty.

Public Politics: Civil Society and Democracy Revisited

Extraction, resistance and bargaining do not by any means exhaust the processes that promote democratization, though taken together they presage those other processes. An historically grounded explanation of democratization and de-democratization needs to dig still deeper, to center on the dynamics of what I call “public politics.”

Public politics concerns not all transactions between rulers and citizens, only those that visibly engage state power and performance. There are three main processes whose interaction in public politics reliably pushes regimes toward democracy and whose reversal pushes them away from it. The first is the insulation of public politics from categorical inequality. The second is the partial integration of popular trust networks into public politics. The third is the reduction of autonomous power centers. Before further describing these dynamics, it is worth pointing out that, just as the concept of voice is more variegated and interesting than it may first appear, the notion that civil society and democracy sustain one another is too. If we look more closely at how groups in society interact with government, we find significant differences in how each of these three processes contribute to democratization.

Patrick Henry opposing the Stamp Act

The first process involves the insulation of public politics from categorical inequality by rank, gender, race, ethnicity and religion. Building such distinctions directly into a regime (for example, by installing a religious requirement for full citizenship) de-democratizes it. The elimination of property qualifications for suffrage and office-holding in most Western semi-democracies during the 19th century erased the direct inscription of class differences into those regimes’ public politics. Similarly, as Western semi-democracies enfranchised women between 1900 and the 1940s, they by no means eliminated gender inequality, but they did help insulate public politics from it.

As to the second process: Up to a point the integration of popular trust networks into public politics promote democracy. (By “popular trust networks” I mean kinship groups, diaspora communities, craft cultures and religious solidarities that ordinary people use to pursue high-risk, long-term activities such as raising children, conducting long-distance trade and following unorthodox religious beliefs.) A point of maximum influence occurs when members of trust networks have both voice and strong stakes in governmental performance, but the state has not yet incorporated them. At that point, trust networks that straddle the boundaries of governmental inclusion maximize their capacity to give or withhold grudging consent. If correct, this analysis casts an unusual light on recent debates concerning the place of trust and civic participation in democratization.

Historically, most trust networks have grown up outside of public politics. Lineages, religious sects and commercial diasporas typically form far from the state. Sometimes, however, they originate within major political actors (trade unions, for example) or in government itself (as with military veterans’ networks). Many recent analysts of democracy have thought that a high density of civic associations in itself promotes both trust in government and high levels of political participation.¹ This is not so, however. The ways in which various forms of association connect local trust networks to public politics matter far more than the sheer frequency of associations. That is because the forms of relations between trust networks and public politics govern the possibility of contingent consent. Trust networks that segregate themselves entirely from public politics (as do some religious communities) may provide participants with comfort and mutual aid, but they inhibit voice. Relations between trust networks and public politics mediate the effective translation of citizens’ expressed collective will into state action; if those relations are scant, nearly everything will get lost in translation.

Surprisingly, a kind of distrust therefore becomes a necessary condition of democracy. Contingent consent entails, in principle, an unreadiness to offer rulers, however they were elected, blank checks to do as they please with society’s resources. It implies the threat that if they do not perform in accordance with citizens’ expressed collective will, citizens might not only turn them out of office but also withdraw compliance from such risky government-run activities as military service, jury duty and tax collection. In Albert Hirschman’s terms, democratic citizens may display loyalty during recognized state crises, but ordinarily they employ voice backed by the threat of “exit.”² In other words, they offer grudging consent, and the very fact that it is grudging makes clear that it should not be presumed automatic.

Civic associations sometimes magnify popular voice, but their presence does not guarantee the contingent integration of trust networks into public politics. We high-minded citizens often deplore the corruption of 19th-century American urban political machines. We also wince at the frequency with which rival groups of ethnic workers and their associations (for example, lodges, athletic clubs, militias and fire companies) literally fought each other on the streets of American cities. But America’s ward heelers performed the indispensable service of building immigrant and trade networks into public politics while defending their powers of contingent consent to government performance. Immigrants acquired an interest in citizenship as workers acquired an interest in representation, all of which directly connected them to government. These proved to be enormous services to the vitality of American democracy.

As to the third process: Reducing coercive power centers outside of the state itself (for example, warlords, militias and self-governing enclaves) moves public politics away from rivalries among those power centers and toward direct bargaining between rulers and citizens, thus advancing democracy. So long as landlords, drug lords and lords of capital control their own independent means of coercion, they can block or reverse democratization simply by withdrawing grudging consent from public politics. Within the state, furthermore, the creation of ruler-controlled coercive instruments that escape public politics (secret police, clandestine prisons, state-backed paramilitary forces and the like) has a similar effect. It impedes democratization, or reverses it if it has already taken shape.

Only after the humiliation of the 1982 war with Britain over the Malvinas/Falkland Islands, for example, did a semi-autonomous Argentine military and ruthless secret services finally retreat from Argentina’s public politics. A delicious irony then occurred: Juan Perón had risen to power during the 1940s as an ambitious colonel who broadened the military’s role in Argentine politics. But the military soon rejected Perón, and his marvelously effective party machine concentrated on making clients of both organized workers and poor people. By the 1980s, the civilian patron-client chains that the Peronistas built had rapidly expanded to become the most important connectors between ordinary Argentines’ trust networks and public politics. Democracy gained, and it did so through the advancement of grudging consent.

Taken together, the insulation of public politics from categorical inequality, the contingent integration of popular trust networks into public politics, and the reduction of autonomous power centers—all of these promote democratization. Their reversal promotes de-democratization. The three processes matter precisely because they expand the possibility and impact of grudging consent.

Think back to England in 1640 and France in 1787 for illustration. At those historical points, categorical inequality dominated political rights and participation in the two regimes. The same distinctions among royalty, nobility, gentry, upper clergy, lower clergy, urban oligarchs and other groups of commoners who organized routine social life also marked the boundaries of political rights and obligations. Except at the bottom ends of patron-client chains, popular trust networks exercised no influence whatsoever on English and French public politics. Indeed, most ordinary English and French people did their best to keep the precious resources embedded in their trust networks—such as dowries, trade secrets and heterodox religious practices—as far away from the grasp of predatory authorities as possible.

Autonomous centers of coercive power, to be sure, offer a more complex picture. In England, 16th-century Tudor monarchs, including Henry VIII, had relentlessly tamed the great arms-bearing lords who had essentially run their own smaller regimes, as well as the powerful churchmen who had enjoyed riches and autonomy under Catholic regimes. In 17th-century France, Louis XIV had similarly brought the military nobility to heel. Containing those autonomous powers did not bring about a sudden transition to democracy; in England and France both nobles and higher clergy retained substantial autonomy and some local coercive power. But the consolidation of state control opened the way for democracy, and when civil war and revolution consolidated it further, that way became a highway. Those rounds of violence, whatever else may be said about them, gave democratization a chance it had never had before.

Of course, states can go too far, snuffing out or incorporating all autonomous sites of deliberation and solidarity; if so, they de-democratize and earn the name totalitarian. For democratization, the crucial processes eliminate or contain autonomous centers of coercive power that block the negotiation of grudging consent between the state and all its citizens. Warlords do not disband willingly, but they must disband before democracy can develop.

Do these ancient histories mean anything for democratization in today’s very different world? Yes, they certainly do. In the first place, it should be clear from the seminal examples of Britain and France that no sort of stable democracy is likely to emerge where there is no strong state. The idea that democracy will take root in places like Afghanistan and Iraq on the basis of a few elections lacks any grounding in historical reality.

Even more elemental, perhaps, is that democracy never arises from a quick fix. It always depends on longer-run changes in relations among public politics, categorical inequality, trust networks and autonomous power centers. And how rulers acquire their means of rule strongly affects those longer-run changes. In the shorter run, the changes that matter most center on the production of grudging consent from the natural tensions in any society between rulers and ruled—consent that manages those tensions and binds the two sides together. That, in turn, has to do with the way people develop the breadth, equality, binding and protection of their voices. This sort of thing does not happen overnight; but it is happening, little by little, all the time.

1.
For a sample of reviews and critiques, see Bob Edwards, Michael W. Foley and Mario Diani, eds., Beyond Tocqueville: Civil Society and the Social Capital Debate in Comparative Perspective (University Press of New England, 2001); Jason Kaufman, For the Common Good? American Civic Life and the Golden Age of Fraternity (Oxford University Press, 2002); and Tilly, Trust and Rule (Cambridge University Press, 2005).
2.
Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States (Harvard University Press, 1970).