Paul Starr is professor of sociology and public affairs at PrincetonUniversity. His new book, Freedom's Power: The True Force ofLiberalism, will be published this spring by Basic Books.

War is always a risky enterprise for the political party orideological faction that undertakes it. Like the Vietnam War, theIraq War has broken the grip on national power of a dominant partythat had been confidently reshaping American politics. Democracieswant speedy victories, especially if they were promised one; agovernment that fails in war throws into doubt its whole view ofthe world. Even a party that leads a nation to victory in a warwith overwhelming support may be punished at the polls afterward.Think of the defeat of Winston Churchill's Conservatives in 1945and the reverses suffered by the Democrats after both world wars.

Since Vietnam, though, war has spelled particular trouble forliberals. Torn between competing values, liberals and theDemocratic Party have been prone to divisions between hawks anddoves and to ambivalence and uncertainty among theircross-pressured leaders. Liberals are dogged by charges from theright that they are unserious about national security; but theyalso worry that war endangers everything that they value and allthat they want to accomplish domestically. The first argumentclaims that liberalism is unprepared to fight wars, while thesecond suggests that liberalism unravels in wartime. Either way, itseems, if war looms, liberalism loses.

It is not just about contemporary liberalism that such argumentshave been made. The idea that constitutional government and liberaldemocracy are unsuited to the rigors of war has a long genealogy,and for a time the historical evidence was at least ambiguous.Classical liberalism had its heyday in the mid-1800s, when theconditions of world politics were relatively benign. Well into thetwentieth century it seemed reasonable to suppose that, like aplant that grows only in bright sunshine, liberalism flourishes onlyin peace. Reflecting on his party's decline after World War I,David Lloyd George, Britain's last Liberal prime minister, wrote inhis memoirs that "war has always been fatal to Liberalism."

Lloyd George may have been rationalizing his own failures as a partyleader. If wars were generally fatal to liberalism, it could neverhave survived, given the frequency of war throughout modernhistory. If liberal governments under liberal leadership wereincapable of seeing war through to a successful conclusion, thegreat struggles of the twentieth century against totalitarianismwould have ended in catastrophe, and today we would live in adifferent world. Liberalism has turned out to be stronger and moreeffective in war than its adversaries have expected, and it hasproved to be more resilient under the pressures of war thanliberals themselves have feared. History does not prove thatcontemporary liberalism will have the same strength and the sameresilience, and it certainly does not suggest that liberals shouldwelcome war; but at a time when a conservative government hasfailed in war and thrown into doubt its whole view of the world,liberals would do well to remember a tradition that rightfullybelongs to them and shows why they can do better in matters of bothwar and peace.

`States make war, and war makes states," the sociologist CharlesTilly observed. But if war is so decisive in the making of states,why the growing dominance and proliferation of liberal democraciesduring the past two centuries? Why not states thoroughly devoted tomartial values? And why weren't liberal societies transformed intothe "garrison states" that many in the mid- twentieth centuryfeared they would become in the age of total war?

At the root of such questions is a misapprehension that liberalismis a source of state weakness because it is centrally concernedwith individual liberty. The core principles of liberalism,however, provide not only a theory of freedom, equality, and thepublic good, but also a discipline of power--the means of creatingpower as well as controlling it. This discipline has been asingular achievement of constitutional liberalism, dating from lateseventeenth- and eighteenth-century England and America, and ofmodern democratic liberalism as it has evolved since roughly thelate nineteenth century.

Liberal constitutions impose constraints on the power of any singlepublic official or branch of government as well as the state as awhole. The constraints protect citizens from tyranny, but that isnot all they do. They also serve to protect the state itself fromcapricious, impulsive, or overreaching decisions. By binding thosein power, making their behavior more predictable and reliable, andthereby increasing the trust and the confidence of citizens,creditors, and investors, constitutionalism amplifies the long-term power and wealth of a state. Constitutional liberalism imposesa further discipline by dividing power within the state and betweenstate and society, and requiring public disclosure and discussionof state decisions--all of these serving as limits on the abilityof officials to pursue their own private interests and enabling thecitizens to control their government. Liberalism wagers that astate so constructed can be strong but constrained--strong becauseconstrained.

Modern democratic liberalism extends the same logic, bothconstraining and enlarging the state's power. To make thegovernment accountable to the entire public is a way of bothlimiting the power of officials and strengthening publicresponsibility and patriotism. Rights to education and otherrequirements for human development and security aim to advanceequal opportunity and personal dignity and to promote a morecreative and productive society. To guarantee those rights,liberals have supported a wider social and economic role for thestate, counterbalanced by more robust guarantees of civil libertiesand a wider social system of checks and balances anchored in anindependent press and pluralistic civil society.

Wars have historically fostered the expansion of state capacities,but without necessarily destroying constitutional government. Theimmediate effects in well-established liberal democracies have runin seemingly contrary directions. Wars have tended to makesocieties less liberal but more democratic- -that is, they haveundermined civil liberties while leading to expanded politicalrights. Once wars have ended, however, their illiberal effects havetypically been reversed, but the democratizing and state-buildingeffects have remained. In short, war has been a catalyst in thetransformation of the liberal state, contributing to the featuresnow associated with modern democratic liberalism.

The examples of war's immediate illiberal effects are well known:governmental suspensions of habeas corpus, infringements of freespeech, public hysteria against dissenters and suspect minorities.The democratizing effects of war have an equally impressivelineage. In seventeenth- and eighteenth- century England, the needto raise armies and obtain new revenue led kings to concedeauthority to Parliament; in nineteenth- and twentieth-centuryAmerica and Europe, the demands of large-scale warfare similarlyled governments to concede new rights of citizenship as a way ofgenerating popular loyalty and rewarding sacrifice. In the UnitedStates, the expansion of voting rights to African Americans afterthe Civil War, to women after World War I, and toeighteen-year-olds during the Vietnam War exemplified the pattern.In Europe, too, war led to expanded rights of citizenship,including social as well as political rights.

These domestic political effects are not inherent in war as such,but seem to depend on the form of warfare and the level ofmobilization, especially the "military participation ratio" (theproportion of the population under arms). The higher the ratio, thesociologist Stanislav Andreski has argued, the more likely war willhave a socially leveling impact. Governments in need of popularenlistment and mobilization are especially likely to expandcitizenship rights as a means of securing loyalty and commitment.Mutual dependence and personal sacrifice in wartime may alsopromote social solidarity, especially a sense of obligation tosoldiers--hence the repeated enlargements of the franchise duringand immediately after wars.

The scale of mobilization in total war may therefore have helped topreserve and to extend democracy. Indeed, it seems scarcelysurprising that wars requiring mass conscription and popularparticipation would break down social hierarchies. The puzzle,rather, is why individual liberty and limits on state powersurvived the age of total war at all. What is even more surprisingis that on the whole, despite bouts of collective anxiety andrepression, the liberal democracies grew more liberal as well asmore democratic over the course of the twentieth century. Insteadof collapsing in the face of crisis, the institutions and ideas ofconstitutional liberalism shaped and limited policies to meet thechallenges of war as well as economic depression. And just whenthey mattered most, those choices proved successful, reinforcingthe commitment to liberal democracy and validating confidence inits principles as a basis of security as well as justice.

The explanation for the deepening of liberalism also lies in theparticular adversaries and challenges that liberal states faced inthe twentieth century. Fascism and communism posed threats toliberal democracy that were simultaneously ideological andstrategic. In opposing and fighting totalitarian regimes, theliberal democracies appealed for international as well as domesticsupport on the basis of ideals of freedom and equality, and in theprocess were forced to confront such contradictions as racialinjustice at home and their own role as colonial powers. The globalstruggle for dominance in the twentieth century ended not only withthe defeat of the Nazis and Soviets, but also with the discreditingand repudiation of anti-Semitism, racism, colonialism, and otherideologies denying human equality that had long enjoyedrespectability in Europe and America.

The domestic side effects of wars, hot and cold, would not havemattered if during the past two centuries authoritarian andmilitaristic regimes were able to mobilize greater power and defeatliberal democracies. But this has not been the historical record.Since 1816, democracies have won three-fourths of the wars in whichthey have been involved, according to Dan Reiter and Allan C. Stamin Democracies at War, the most authoritative study of the subject.

Reiter and Stam's analysis points to two sets of factors that havebeen critical in enabling democracies to be more successfulmilitarily than autocratic states have been. The first has to dowith the decision to go to war. Democracies are far less likely toinitiate wars that they end up losing. That is not becausedemocracies are inherently pacific; they often do initiate wars(though during the past two centuries democracies have rarelyattacked each other). When democracies have attacked first, theyhave won 93 percent of the wars, whereas dictatorships that havestruck first have lost four out of ten times. (When attacked,democracies have also been more successful, prevailing in 63percent of the cases, compared with just 34 percent fordictatorships.)

Using historical cases to back up their analysis, Reiter and Stamsuggest that democratic leaders typically refuse to launch a warunless they are virtually certain of victory, because they knowthat if their nation loses they will surely be driven from office,whereas dictators are more willing to gamble when the odds ofvictory are lower because they often survive in power even afterdefeats in war. Moreover, because autocratic regimes make decisionsin secret and refuse to tolerate criticism by a free press andopposition parties, dictatorships are more likely than democraciesto miscalculate the odds of victory, wax overconfident, and start awar in which they are overmatched. The greater military success ofdemocracies exemplifies how the liberal discipline of power worksand why, as Stephen Holmes puts it, constitutionally limited powercan be more powerful than unlimited power.

The second set of factors has to do with how well states fight oncewars have begun. Here Reiter and Stam look at the outcomes ofindividual battles during the past two centuries, using a databaseoriginally created by military historians for other purposes. Thekey factors in democracies' war-fighting advantage appear to havebeen greater initiative among the soldiers of democratic armiesthan among soldiers of autocratic regimes (which Reiter and Stamattribute to differences in political culture) and better militaryleadership (which they attribute in part to the greater ability ofdemocracies to make merit rather than political loyalty the basisof military promotion). They also cite historical cases in supportof the proposition that dictators' armies are quicker to surrender(which may be due to the higher standards of treatment thatdemocracies, at least in the past, have been known to extend toenemy prisoners).

One difficulty with Reiter and Stam's quantitative analysis of warsis that it treats all pairs of warring countries equally. Theconflict between Honduras and El Salvador in the Football War of1969 weighs as much in the statistical results as the conflictbetween Germany and the United States during World War II. Yet ifthe latter had turned out to be among the one-fourth of wars lostby democracies, the entire world would now be different. Accordingto Reiter and Stam's analysis, military support from othercountries is not a significant factor in explaining why democraciesare more likely than dictatorships to win wars that theiradversaries have initiated. But even if that is a validgeneralization when counting all wars the same, the contrary casesof the two world wars simply matter more to the overall outcome ofglobal politics.

Total war could have given totalitarianism an edge. Lackingaccountability to voters, internal checks and balances, a freepress, and independent power centers in civil society, the fascistand communist regimes had a relatively free hand in conscripting,taxing, and otherwise extracting resources from their societies andtherefore seemed to hold advantages in the sheer force they couldgenerate. But by virtue of their political structure, thetotalitarian states also suppressed initiative, lagged in criticaltechnological innovations, and lacked means of self-correction.These deficiencies had fateful consequences.

As it turned out, the modern forms of despotism were not a winningnational strategy in the twentieth century. As before, governmentswith constitutionally limited powers proved to be more powerfulthan governments with unlimited powers. Moreover, by the end ofWorld War II, the liberal democracies had learned that it wasimperative for them to build international alliances andinstitutions to have any chance of stopping aggressive wars andmaintaining peace and security. As the power of modern liberalstates is based on a democratic partnership at home, so the liberaldemocracies have come to pursue power through partnershipinternationally.

The danger of states making war and war making states is a spiral offorce, ending in a thoroughly militarized world. But becauseliberal democracy and liberal internationalism proved an effectivestrategy for creating power and prevailing in conflict, a differentself-reinforcing cycle set in, at least for a time. Liberaldemocracies fought and won wars, which led to furtherdemocratization, which helped to protect individual liberties oncethe war emergencies ended.

Unfortunately, there is nothing inevitable about this cycle. Thereis no guarantee that war will continue to lead to democratization,or that democracies will be able to reverse the immediate illiberalconsequences of war. If the positive effects of war on politicaland social equality depend on a high military participation ratio,the connection may have disappeared, at least in the United States.The kind of technological war that America now wages no longerrequires mass enlistment or popular mobilization and consequentlyseems to generate no pressure to expand rights or benefits. Indeed,the ability to wage war without conscription and with so little callfor personal sacrifice from the public may reduce the highthreshold for starting wars that has been partly responsible fordemocracies' military success. And if reversing the illiberaleffects of war depends on bringing war to a close, what of a global"war on terror," which it will be impossible ever to say has cometo an end? In short, whether liberal democracies can maintain theirdistinctive qualities in the face of war may depend more than everon their leadership.

According to the dominant historical pattern, the Iraq War shouldnever have happened. The governments of democracies, Reiter andStam tell us, rarely initiate unsuccessful wars, in part because ifthey fail to win--and win quickly--they are predictably punished bythe voters. A government that gets itself into this position haseither overestimated the odds of military victory in the firstplace or known the odds were poor but gambled anyway.

The Bush administration's decision to go to war in Iraq was clearlyborn of overconfidence. American triumphalism had been buildingever since the Soviet collapse; a relatively easy victory inAfghanistan after September 11 seemed to confirm that, as theworld's only superpower, the United States was in so dominant aposition that it needed only the resolve to exploit it.Conservatives were particularly intoxicated with delusions thatAmerica could bend the world to its will. The insularity of thepresident's inner circle, the pliability of theRepublican-controlled Congress, and the docility of the press inthe run-up to the war all contributed to the administration'sability to ignore conflicting evidence and opinions, itsunderestimate of the war's complexities and costs, and its failureto prepare and set in motion a coherent plan for the occupation.

Overconfidence in American military power was also crucial to theadministration's disregard for sources of long-term democraticadvantage in international conflict. Openly dismissive of "oldEurope" and international opinion, Donald Rumsfeld and others madeit more difficult not just for this administration, but also forfuture ones to win moral and legal support abroad for Americanpolicy. And by repudiating international standards in the treatmentof prisoners and approving torture either directly or throughrendition to other countries, the administration surrenderedanother significant asset of democratic countries in defeatingtheir adversaries. The war itself has degraded America's hardassets as well as its soft ones, and weakened its deterrent power.

Just as autocratic regimes are more likely to make mistakes becausethey make decisions in secret without checks and balances, so theBush administration's resistance to transparency and its claims ofunilateral executive power have been a threat to a vital source ofAmerica's power as well as its liberties. The president's decisionsto authorize secret surveillance programs, to hold prisonerswithout trial, and to resist any role by Congress and the judiciaryin national security matters all need to be understood in thislight. The case against unaccountable executive power is not simplythat it violates the Constitution, but also that unchecked power ofany kind is more likely to result in catastrophic blunders, and toundermine the purposes it is supposed to serve.

Though accused of being unpatriotic, the critics who were skepticalof the president's claims about Iraq were, in fact, following thelogic of prudence that has historically contributed to democracies'success in war and avoidance of long and costly military disasters.In opposing administration policies that violate the separation ofpowers, norms of governmental transparency, and internationalconventions regarding prisoners, liberals are not merely insistingon legal niceties. These rules serve our long-term interests. Arecent article in The Weekly Standard refers to Republicans as the"power party" and to Democrats as the "peace party" in Americanpolitics today. That may be the way it looks from the right. Butafter squandering and degrading America's power, conservatives haveno claim to be its special guardians.

There is a different way of thinking about power from the one thatconservatives in the Bush era have championed, and that way ofthinking grows out of the liberal tradition and historicalexperience. The crucial historical lesson is not that liberalprinciples and public debate must give way in war for the sake ofnational defense: constitutionally limited power has proved to bemore powerful than unlimited power. Democratic partnerships at homeand abroad are critical to the nation's strength. America has risento its current position partly on the basis of these ideas, andstaying true to them would be a victory in itself.

By Paul Starr