Book Reviews

How Not to End War

By Michael J. Glennon
Wednesday, October 18, 2017, 7:30 AM

PDF version

A review of Oona A. Hathaway and Scott J. Shapiro's, The Internationalists: How a Radical Plan to Outlaw War Remade the World (Simon & Schuster, 2017)


Years ago, while I was visiting the State Department legal adviser’s office, a lawyer who handled treaties said he wanted to show me something. He disappeared into the massive vault that safeguards the official copies of all U.S. treaties and emerged holding a document by its corner, between his thumb and index finger, as though it were a rare curiosity. “Look at this,” he said. It was an instrument of ratification from a nation that had only recently approved the 1928 Kellogg-Briand Peace Pact. (The United States is the designated depository.) “Can you believe that?” he asked. In years of working on use-of-force issues on Capitol Hill, I’d never heard anyone refer to the pact. But, as my friend pointed out, it had never been formally terminated, and now and then new instruments of ratification continued to trickle in.

By its terms, the pact should hardly be forgettable. In only two sentences, it lays out an unqualified prohibition against going to war. Article I provides that “[t]he High Contracting Parties solemnly declare in the names of their respective peoples that they condemn recourse to war for the solution of international controversies, and renounce it, as an instrument of national policy in their relations with one another.” Article II adds that “[t]he High Contracting Parties agree that the settlement or solution of all disputes or conflicts of whatever nature or of whatever origin they may be, which may arise among them, shall never be sought except by pacific means.” That’s it. The pact provides no exception for self-defense, whether in response to an actual armed attack or an imminent threat of one. Yet every major belligerent in World War II was a party.

The pact became a laughingstock. Thomas Bailey wrote that it “proved a monument to illusion.” George Kennan described it as “childish, just childish.” Robert Ferrell concluded that the pact showed “that American popular understanding of the great problems and policies of post-1918 international affairs was appallingly naïve.” To Ian Kershaw, it is “singularly vacuous.” James Lindsay called it “the international equivalent of an air kiss.”

Two prominent legal scholars, Oona Hathaway and Scott Shapiro, set out to prove them and a host of others wrong. “Provocative” is hardly the word for their effort. The dust jacket announces that “this book will change the way we view the history of the twentieth century...” The pact’s signing, the authors write, was “one of the most transformative events of human history” (p. xiii). Before its ratification, they suggest, the world was a Darwinian jungle in which states typically addressed wrongs through war. That world came to an end in 1928 with ratification of the pact. “War has been outlawed for nearly a century and the result has been a period of unprecedented peace and cooperation between states” (p. xxi). “[I]nterstate war has declined precipitously, and conquests have almost completely disappeared” (p. 418).

Theirs is quite a pirouette, spinning through not only two centuries of wars of conquest but a maze of interstate wars in general. As a history of ideas, the book is well-written, jargon-free, and informative. It’s filled with lively sketches of the main characters and engrossing historical vignettes.

As a work of international law and relations, however, the book suffers from an array of problems. Its main weakness lies in the shaky scaffolding supporting its thesis. The argument that the pact succeeded overstates the prevalence of war before its ratification and understates it afterwards. The book fudges the pact’s legal obligation. Wars of conquest, the pact’s supposedly indisputable success story, are rendered a thing of the past largely by defining them out of existence. The book fails to establish causality. And its case rests ultimately on a mistaken notion of law as the prime source of revolutionary change.

  1. War Before and After the Pact

Before 1928, the authors write, “[war] was the method for solving disagreements between states” (p. 314). International law posed no constraint on recourse to war anywhere in the world. “The Old World Order—the world in which war was a permissible way to address wrongs—was bloody, brutal, and unjust” (p. xxii).

Although it’s surely true that initiating war was legally permissible in earlier centuries, the level of interstate violence varied considerably from region to region. Henry Kissinger has suggested that in the world today, there is no single international system; multiple systems exist side-by-side. The same was true in earlier times. During the 19th century, war wracked Asia and Africa as European nations waged vicious campaigns to establish and maintain colonies by subduing native peoples. Within 19th-century Europe itself, however, the picture was different. The authors’ observation concerning the international legal institutions that have grown up since 1928—that they “have brought seven decades of unprecedented prosperity and peace” (p. 419)—could be made more accurately concerning the geopolitical institutions that emerged from the Congress of Vienna in 1815. Klemens von Metternich, Lord Castlereagh, and the Prince of Talleyrand created a balance-of-power system aimed at preventing any one nation from gaining enough power to overwhelm others. Their system worked: The Concert of Europe was, to be sure, ruthless in putting down internal rebellion, but within Europe, the century following 1815 was remarkably peaceful compared to the preceding century. There was, as Eric Hobsbawm noted,

[N]o general European war, nor any conflict in which one great power opposed another on the battlefield, between the defeat of Napoleon and the Crimean War of 1854-6. Indeed, apart from the Crimean War, there was no war involving more than two great powers between 1815 and 1914. The citizen of the twentieth century ought to appreciate the magnitude of this achievement.

The catastrophic breakdown in August 1914 devastated Europe on a scale not seen since the religious wars 300 years earlier, traumatizing the survivors and fostering intense anti-war sentiment throughout Europe and North America during the 1920s and into the 1930s. It was from this zeitgeist that the Kellogg-Briand Pact emerged.

The pact, the authors argue, “was a decisive break from the past” (p. xiv) that “remov[ed] war from states’ legal toolkit...” (pp. xiv-xv). It “proved remarkably successful” (p. xiv). The outlawry of war “precipitated the end of international war itself” (p. 334). Its “New World Order is not simply the law. States actually obey it” (p. xvii). In modern times, “[i]nterstate wars are rare” (p. 368). “The deadliest [conflicts], those that kill at least 1,000 people, have declined even further—by half” (p. 334).

On its own, of course, the number of conflicts means little. An increased number of conflicts could still make for a more peaceful world if the conflicts result in fewer casualties; conversely, it’s possible to have a less peaceful world with fewer conflicts, because the conflicts result in enormous casualties. Merely counting conflicts—even those that brought about 1,000 or more deaths, which is the authors’ threshold—doesn’t begin to capture the scope of violence and the gravity of suffering that wars wreak. The “Correlates of War” project codes World War II as a single war, as it does the Ifni War of 1957-58. World War II, it reports, had 16,634,907 battlefield casualties; the Ifni War, 1,122. Many international relations scholars who study these issues therefore examine the number of war deaths, inexact though they may sometimes be. Hathaway and Shapiro do not. Without doing so, it’s not possible to know whether the bloody, brutal, and unjust “Old World Order” that they decry has in fact disappeared. War is a horror, after all, because of the massive, untold human suffering it causes, which is not reflected in conflict numbers alone.

Assume, nonetheless, that it’s the number of conflicts that matters. The authors contend that the total number of conflicts has dropped by 40 percent over the last several decades (p. 334). The work of other scholars suggests otherwise. Meredith Reid Sarkees and Frank Wayman summarize their research along with that of J. David Singer and Frank Wayman in the third war-data handbook of the Correlates of War project, published in 2010. They write as follows:

In the late 1980s, interstate wars had been occurring at their usual frequency, rather than declining, and were still at higher levels than in earlier periods in our history. If these post-Cold War patterns continued until the new millennium, we predicted that the 1990s, which were supposed to be the first decade of the “New World Order,” would witness more wars than any decade since 1816. When we updated this analysis after the twentieth century had ended, our warning against premature optimism was confirmed. Interstate wars in the 1990s (as in the 1980s) were still at higher levels than average... Consequently, the 1990s ended up being the decade with the most war onsets since Napoleon.

This bleaker picture changes little if the temporal scope is broadened. According to Sarkees and Wayman, the seven decades from 1928 to 1998 saw the onset of 45 interstate wars (compared with 37 in the seven decades before ratification of the Kellogg-Briand Pact). The authors write that the pact was “reaffirmed” by the United Nations Charter (p. 419), but the U.N. Secretary-General’s High-Level Panel, in 2004, found violations of the charter’s use-of-force rules so numerous as to defy quantification. “For the first 44 years of the United Nations,” the panel observed, “Member States often violated [the charter] rules and used military force literally hundreds of times...” By one count, the panel said, from 1945 to 1989 “force was employed 200 times, and by another count, 680 times.” Other studies have reported similar results. Arthur M. Weisburd, for example, has counted 100 interstate wars between 1945 and 1991. Thomas M. Franck wrote in 1987 that the use-of-force rules “are adhered to, at best, only by some states, in some instances, and have been ignored, alas, with impunity in at least two hundred instances of military conflict since the end of World War II.” (This was seventeen years after publication of his seminal article, Who Killed Article 2(4)?) Herbert K. Tillema counted 690 overt foreign military interventions between 1945 and 1996. The legal justification states offered for using force in these instances—when they were moved to give one—was almost invariably self-defense. In the final pages of the book, the authors themselves acknowledge the regression permitted by that rationale. “Unfortunately,” they write, “the growing reliance upon self-defense as a justification for using force…threatens to make self-defense the exception that swallows the rule against war” (p. 416).

The return to pre-pact levels of violence, in any event, is borne out by the Correlates of War project. In the authors’ view, it is “[t]he best dataset of military conflict ever developed” (p. 312). According to it, the 47 interstate wars that started between 1820 and 1929 resulted in a total of 10,541,597 deaths—an average of 224,289 deaths per war. The 48 interstate wars that started between 1930 and 2004 resulted in a total of 21,539,303 deaths—an average of 448,735 deaths per war, an increase of over 100 percent. And these are only battlefield deaths; the numbers do not include tens of millions of civilians who died since 1928 of war-related violence, disease, or mass-liquidation.

This doesn’t seem like a decisive break from the past.

  1. The Pact’s Legal Obligation

Assume, however, that the worlds before and after the pact were as starkly different as the authors contend. Assume also that state parties to the pact intended to comply with its requirements in good faith. What did it demand of them?

The first question is what, specifically, the pact outlawed. Everyone, including the authors, seems to rule out one interpretation—that the pact was intended to prevent only wars of conquest. The book finds nothing in the pact’s text or negotiating history to support that interpretation. The pact was intended to put an end to all war, period. War was broadly conceived, as commonly understood; the pact required that all disputes be resolved by peaceful means. The only remaining question was whether a state could lawfully respond militarily—in self-defense—to a war that was thrust upon it by an aggressor. It’s possible that the answer to that question could have changed over time depending on how much weight is given to subsequent state practice. If state practice is accepted as affecting the pact’s meaning, then it’s also possible that practice might not merely have modified the pact’s meaning—it’s possible that widespread, contrary practice might have eviscerated the pact altogether. The answers to two questions become crucial: First, did the pact mean what it said and ban all war, without exception? Second, if it did, did that meaning change over time to reflect state practice?

The book answers neither question clearly. The pact’s original meaning, the authors suggest at the outset of the book, was, just as it said, to flatly prohibit all war. The authors write that “[t]he Pact was aimed at ending war between states...” (p. xiv). “By outlawing war,” they write, “states renounced the principal means they had for resolving their disputes. They demolished the existing system...” (p. xvi). States foresaw that recognizing a self-defense exception would provide a ready-made pretext to go to war. “Indeed, it was precisely this concern that led authors of the Pact to omit an express exception for what they called ‘defensive wars’...” (p. 416). Its blanket prohibition was therefore meant to be taken seriously: “If states can always invoke self-defense as a justification to use force, then the prohibition on war becomes meaningless” (p. 416). For similar reasons, the authors seem to suggest that the pact ruled out humanitarian intervention. “In the New World Order,” they write, “the only legitimate way for one sovereign state to get another sovereign state to do what it wants is by offering to cooperate in ways that benefit both” (p. xviii). “It is better to live in a world where war is not a permissible mechanism for righting wrongs, even if that means some wrongs remain unaddressed” (p. 422). The authors thus appear to believe that the pact meant what it said: It prohibited all war, however compelling the justification. Every dispute had to be resolved peacefully, even one arising from an armed attack or a humanitarian crisis.

Yet elsewhere in the book they point out that, in testifying before the Senate, Kellogg “explained that the treaty would not interfere with the right of self-defense” (p. 159). Kellogg viewed the right of self-defense as so central to the notion of statehood that it cannot be contracted away, advising his French partners that there was no need to make the right explicit in the pact. In fact, Kellogg indicated that the pact would no longer bind a party that was attacked by another. Kellogg even proclaimed that the agreement would not preclude U.S. enforcement of the Monroe Doctrine (p. 159). The government of Japan, the authors observe, relied on these assurances (to its detriment) in later military actions in China (p. 159). Nowhere do the authors suggest that Kellogg was mistaken or disingenuous in interpreting the pact as permitting defensive war.

Nor is the book clear on whether the meaning of the pact changed over time. At some points, the authors embrace a kind of originalism, with non-compliant state action seen as a violation, not a modification, of the treaty. “The postwar consensus on the illegality of war is under greater assault today than it has been in seven decades,” they write (pp. 415-16). Humanitarian intervention, like the “corrosive” U.S. air campaign against the Islamic State, represents a “challenge” to the system, not evidence that the system has adapted to behavioral changes (p. 416). “What matters is not whether the law is sometimes broken,” the authors assert. “In 2014…there were 1,165,383 violent crimes reported by law enforcement in the United States. But that does not mean that the laws against violent crimes in the United States are ineffectual” (p. 418). Earlier in the book the authors contended that the pact came to impose only a narrow “rule against conquest” (pp. 332, 358), which is discussed below. One wonders how that transformation could have occurred if the pact’s original ban against all war was unaffected by violation.

Yet at other points the authors appear to accept that deviant state practice, or even contrary opinio juris, could have altered the ban’s meaning. “It is difficult to imagine war serving any useful function other than a defensive one,” they write (p. xiv). “In the New World Order, states cannot act unilaterally to address violence in other states unless they have been attacked or face imminent attack...” (p. 369). It’s not clear whether Hathaway and Shapiro are referring to the pact, the U.N. Charter, or both. They don’t address whether the pact is still in effect—or, if so, how to reconcile it with the charter—beyond the cryptic observation that the charter “reaffirmed” the pact (pp. 331, 419). But they seem to accept that a treaty’s meaning can change over time, since the text of neither the U.N. Charter nor the pact authorizes use of preemptive force. (Article 51 of the charter draws a bright line by requiring an “armed attack” as a predicate to using defensive force. Scholars such as Louis Henkin, Ian Brownlie, and Phillip Jessup, who understood the law to reflect that bright line, knew that its placement was intentional; the charter was, after all, premised on the assumption that a standing, or stand-by, constabulary force under the direction of the Security Council would swiftly come to the defense of a nation that reported that it had been the victim of an armed attack.) The authors’ suggestion that the pact was reaffirmed by the charter implies that, somewhere along the line, the pact may have come to include the self-defense exception that the charter explicitly recognized. If the proscription of the charter changed after 1945 to permit defensive force to meet an imminent threat, could the proscription of the pact not have changed as well?

The book’s failure to come to grips with whether and how the meaning of the pact might have adjusted to contrary state practice is a puzzle—relating, as it does, to the book’s perplexing comparison of the international legal system to domestic legal systems. The two differ fundamentally in their treatment of non-compliant conduct. In the international legal order, actors within the system—states—are lawmakers; their actions have legal import. Not so in domestic systems, where the conduct of actors—individual citizens—does not shape legal rules. Within the international system, repeated, widespread violations can alter a rule’s meaning—or they can disembowel a rule altogether. At some point, ineffectuality shades into desuetude. The authors seem, again, to embrace some notion of adaptation through practice when they present the pact as having come to impose merely a “rule against conquest” (pp. 332, 358), given that they proffer no evidence for such a rule from either its text or travaux préparatoires. Yet they don’t confront the underlying question, which is precisely what’s given rise to the pact’s widespread dismissal: Why have numerous violations of the rule outlawing war not pushed it across the line that divides adaptive modification from desuetude?

  1. Wars of Conquest

The authors assert, nonetheless, that the rule against conquest has worked. Conquests, they write, are nowadays “practically nonexistent” (p. 368). An average state is now likely to suffer a war of conquest “only once or twice a millennium” (p. 314). That would be an astonishing achievement indeed. How could the authors have arrived at such a remarkable figure?

The answer is: semantically, by defining wars of conquest virtually out of existence.

Here’s a short, incomplete list of wars that they either omit from their analysis or define as not constituting wars of conquest: the German occupation of most of Europe before and during World War II, and Soviet domination of the Baltics and Eastern Europe afterwards; the Japanese occupation of Manchuria and parts of China in the 1930s; the Arab-Israeli wars at the founding of Israel; the Korean War; the First Indochina War between France and Vietnam; the invasion of Egypt by Britain, France, and Israel in 1956; the Six-Day War between Israel and its neighbors in 1967; North Vietnam’s invasion of South Vietnam, concluded with formal reunification in 1976; Argentina’s invasion of the Falkland Islands in 1982; the Soviet invasion of Afghanistan in 1979; Tanzania’s invasion of Uganda in 1979; Vietnam’s invasion of Cambodia in 1979; the Russian invasion of Georgia, South Ossetia and Abkhazia in 2008, and Crimea in 2014; the United States’ invasions of the Dominican Republic in 1965, Cambodia in 1970, Grenada in 1983, Panama in 1989, Afghanistan in 2001, and Iraq in 2003; NATO’s 1999 bombing campaign against Yugoslavia; the U.S. overthrow of the governments of Iran in 1953 and Guatemala in 1954, and later U.S. efforts to destabilize the governments of Cuba, Chile, and Nicaragua; and many of the violations of the U.N. Charter that the Secretary-General’s High Level Panel, as described above, found too numerous to count.

The authors write these wars out of their narrative either because the conflicts did not involve a formal transfer of territory or by applying a technical, formalistic, highly-qualified definition of conquest that is at odds with commonly-accepted notions of the term. Consider their key criteria. They exclude wars carried out or approved by multinational organizations, such as NATO; wars in which a victorious state refrained from claiming sovereignty over newly occupied territory; wars that resulted in the reversal of earlier, unrecognized seizures of the same territory; wars that resulted in surrender or “debellation,” in which one side is completely vanquished; wars of independence; and all threats of force, even imminent threats backed by military exercises and naval or battlefield preparations that resulted in capitulation and the wholesale domination of a victim state (pp. 312–13, 450).

These exclusionary criteria present multiple problems. First, the cloak of multilateralism often conceals naked self-interest. When the United States invaded Grenada in 1983 and installed a new, friendlier government, for example, it announced that it was acting on behalf of the Organization of Eastern Caribbean States. Nothing in the U.N. Charter (or, earlier, in the pact) exempts multilateral organizations from the rules that govern the use of force. Military alliances can hardly be regarded as benign guardians of those rules merely because they qualify as multilateral organizations. NATO toppled Muammar Qaddafi’s Libyan government in 2011 in a military operation that exceeded the scope of authority conferred upon it by the Security Council. NATO didn’t even pretend that its 1999 attack against Yugoslavia had been authorized by the council, let alone that it constituted a response to an actual or imminent armed attack. During the Vietnam War the Johnson administration argued that it was acting under the auspices of the SEATO Treaty. Soviet interventions in Hungary in 1956 and Czechoslovakia in 1968 were far from legitimated by claimed ties to the Warsaw Pact.

Second, the benefits of acquiring new territory militarily often accrue without having to claim formal sovereignty. Hitler never claimed sovereignty over occupied Poland, France, or other nations in German-occupied Europe. Political and economic domination by an occupying power can be achieved without formal sovereignty.

Third, non-recognition does not imply condemnation or signify the illegitimacy of a territorial claim. International law imposes no requirement that one state recognize another state or its territorial boundaries. Recognition is entirely optional; non-recognition implies nothing.

Fourth, the perceived illegitimacy of territorial claims is not necessarily the reason that such claims sometimes get reversed. The authors applaud the reversal of the Axis powers’ conquests during World War II, but, of course, the Axis powers were forced to surrender their ill-gotten gains for a simple reason: They lost. Stalin didn’t withdraw from Eastern Europe. Even the authors confess that “[t]he outlawry revolution would have failed had the allies not won the war...” (p. 331).

Fifth, recognition and non-recognition are not always neat, binary categories. States sometimes recognize other states de facto but not de jure. While Latvia, Estonia, and Lithuania were under Soviet control, for example, the United States declined to formally recognize the Soviet-installed puppet governments, but decrees of those governments were sometimes given effect within the United States for humanitarian reasons—in fairness to those nations’ citizens and others who were affected by the decrees. Since derecognizing the Republic of China in 1979, the United States has continued, under the Taiwan Relations Act, to grant the authorities on Taiwan many of the privileges that would otherwise be available only to a formally recognized state.

Sixth, what we call war has changed dramatically in recent years. Focusing on traditional, set-piece battles and invasions that formally alter sovereignty misses aggressive behavior such as economic strangulation, targeted drone strikes, incessant paramilitary raids, state-sponsored terrorism, and debilitating cyberattacks that can produce effects which are in every material respect the functional equivalent of an extinction of sovereignty. Yet the authors’ dataset excludes even older coercive techniques amounting to imminent threats of force. They present a riveting account, for example, of Commodore Perry’s ultimatum to the Japanese, delivered at cannon-point: Trade with us or fight us (pp. xvi, 134-38). Yet Perry’s threats—and most of the brutal compulsion that the authors inveigh against in the “Old World Order”—would not run afoul of the rule against conquest they’ve formulated. They worry at the end of the book that the “postwar consensus on the illegality of war is under greater assault today than it has been in seven decades” (pp. 415-16). They believe that increased reliance upon the protean justification of self-defense reflects this backslide. But like gunboat diplomacy, the use of even pretextual defensive force alters no borders; if no conquest results, why the concern?

Let there be no mistake, therefore, about the effect of the cramped definition the authors devise: It immunizes state conduct that contravenes the plain text of the pact’s requirement to resolve all disputes by pacific means only. If these criteria define the pact’s meaning—if that’s all it meant—then the tears that Frank Kellogg shed when he signed it were shed for the wrong reason.

  1. Causation

The greatest challenge that Hathaway and Shapiro confront, of course, is establishing a causal link between the pact and the changes that they claim have occurred in state behavior. Even if they establish a transformation in state conduct, the possibility remains that other factors, alone or together, were at least equally causative. Consider a few plausible, alternative explanations.

One obvious possibility is that a deepening and broadening global social norm has emerged that has rendered the world generally and gradually less violent. Steven Pinker has popularized this theory in The Better Angels of Our Nature. It’s a powerful argument. He may be right. If he is, it counts against Hathaway and Shapiro’s case, not for it, because Pinker sees the strengthening of that norm as a centuries-long trend, not the consequence of a single historical pivot-point. The authors do not explain why the putative decline of war has not been the result of such a norm, rather than war’s sudden illegalization in 1928.

The authors do acknowledge three potential alternative explanations. The first is nuclear weapons. Some international relations scholars believe that nuclear weapons have discouraged the use of conventional force in disputes between nuclear powers because those powers fear that such use could escalate into a nuclear exchange. The second is the spread of democracy. In 1928, there were fewer than twenty democracies in the world; today there are at least a hundred. Some scholars believe, with Immanuel Kant, that a tendency of democracies not to fight one another has reduced interstate strife. A third potential conflict suppressant is international trade: It makes little economic sense for a state to attack a trading partner that supplies essential imports or provides jobs and cash to domestic exporters. None of these theories is uncontroversial. Critics respond, for example, that while democracies may not attack each other, they don’t resist laying into non-democracies when the spirit moves them; that nuclear weapons don’t seem to have dissuaded nuclear powers from promoting bloody proxy wars in and among client states; and that Norman Angell’s trade-promotes-peace theory didn’t prove exactly right in 1914.

The authors could have elaborated all these counter-arguments, and then some, had they been willing to engage on the merits with proponents of these different accounts. But they don’t. Instead, they offer two side-stepping responses. The first is that neither the advent of nuclear weapons, nor the spread of democracy, nor increased global trade “offers a convincing answer to [the] basic question” of “[w]hy did most of the borders after the Second World War snap back to the lines that existed when the Pact was signed?” (p. 332). Well, the answer, again, isn’t buried in econometric arcana: The answer is, because the Axis powers lost.

The authors’ second response is their idée fixe, the explanatory poster on every wall of the book: It’s the pact, of course! The chain of causation always ends on August 27, 1928, the date the pact was signed. Any becalming effect nuclear weapons may have had lies in a “shared internationalist commitment…to keep the peace,” which is, naturally, “rooted in the Pact” (p. 332). If democracy is causative, it’s because “conquest no longer ‘counts’ as a legitimate reason” (p. 332) within democracies, where electorates are presumed to understand and share the authors’ definition of conquest. And it was of course “the end of conquest” that “helped unleash greater economic cooperation by making it safe to trade...” (p. 333). “The missing element in all of these explanations, in other words, is the outlawry of war that began with the Peace Pact” (p. 333). Other factors had some influence, the authors acknowledge, but it was the pact that “sparked a series of events that would lead to the construction of a new global order” (p. xviii).

No doubt the pact did have some effect. The question is, how much? It’s extremely difficult in the social sciences to establish that X causes non-Y. Unobserved variables that might have influenced states’ proclivity to honor the pact might also have produced an independent proclivity to avoid war. As Judge Richard Posner put it, “[o]ne can never be certain whether observations that confirm (that is, are consistent with) theory A are not really confirming theory B instead, which overlaps with or includes A.”

Instead of recognizing the burden they confront, however, the authors set an almost impossibly high hurdle for themselves. They write that the pact, again, “was among the most transformative events in human history, one that has, ultimately, made our world far more peaceful” (p. xiii). The history of the 20th century needs re-writing. Extraordinary claims such as these pervade the book—and extraordinary claims, as Carl Sagan used to say, require extraordinary evidence.

But the authors present none. What they present, instead, is a series of unsubstantiated assertions that the salutary trends they extoll all find their genesis in the pact. Nothing in the authors’ dataset and nothing in their constructed narrative even begins to rule out the possibility that the same forces that led to the pact’s ratification could also have been responsible for behavior consistent with the pact, or the possibility that later, unrelated forces have had a far broader impact on state behavior than did a decades-old treaty mutilated by the Second World War. The authors’ belated realization that the pact’s supposed effects didn’t kick in until after World War II should have presented a flashing yellow light warning of a collision with other oncoming, supervening causes. But few are even noted.

Consider a brief list of additional, potentially relevant factors. The age of colonization has come to an end, meaning an end to the gruesome wars of subjugation waged by Western empires against peoples of the South and East. Increased international travel, educational exchanges, news, and communication have given governing elites a sense of shared values and made it more difficult to dehumanize adversaries when disputes arise. Nearly three times as many states exist in the world today as existed in 1928 (195 vs. 73), meaning potentially greater internal homogeneity and less incentive to forcibly unite or separate territory for reasons of religious, racial or ethnic propinquity. It’s possible that more territory was up for grabs before the 1920s, when fewer nations were possessed of the requisite military technology, economic resources, political cohesion and loyal allies needed to establish and successfully defend their borders. Moreover, without the capriciousness with which colonial powers drew many international borders, some intrastate wars—which have been on the upswing since 1928—would today be interstate wars. Even accepting those boundaries, the line between inter- and intrastate conflict is notoriously difficult to draw, as with the fighting today in and around Syria and ISIS-controlled parts of Iraq. In addition, would-be aggressors today may be weaker, and potential victims stronger, relative to one another. Smaller power differentials could, in themselves, account for claimed decreases in the amount of land taken in wars of conquest. Many of these trends accelerated in the period following ratification of the pact, leading to the easy-but-mistaken inference that the pact has been responsible for more amicable interstate relations.

Law also, of course, has had an influence. That includes the pact. Begin, however, with what one might expect to have been the first agreement to catch the eye of juridical archeologists looking to unearth a fossilized treaty for restoration—the League of Nations Covenant. Why select ratification of the Kellogg-Briand Pact rather than the 1919 ratification of the covenant as the historical pivot point? Momentum from the covenant’s ratification, after all, led directly to the pact. The terms “territorial integrity” and “political independence”—deployed in the U.N. Charter’s keystone Article 2(4) provision outlawing non-defensive war—were lifted directly from Article 10 of the covenant. “[T]he similarities between the League and the United Nations,” Brierly observed, “are many and fundamental.” Like the U.N., the League had an organizational structure. It had an assembly that met yearly. Each member state had one vote. It had a council, consisting of permanent members and nonpermanent members elected by the assembly. Decisions of the council required the unanimous agreement of all members. It had a secretary-general. It’s no exaggeration to say that the United Nations was constructed from the burnt timbers of the League. The pact, in contrast, established no organizational structure. It did impose a normative constraint, but so did the covenant. Under Article 12 of the covenant, state parties agreed not to go to war until three months after submitting a dispute for peaceful resolution, which they were obliged to do. Brownlie wrote that the League presumed the impropriety of war as a means of self-help; that surely was President Woodrow Wilson’s expectation. “It’s not too much to hope that this is the last Balkan war,” he said, “and maybe the last European war for a long, long time.” (Actually, a hyper-legalist might hypothesize that the real legal pivot-point was the 1945 ratification of the U.N. Charter. It was only after its ratification, after all, that peace supposedly has broken out.) 

Other unsupported assertions of causality suffuse the book. The authors see the doctrine of neutrality as having been gutted by the pact, for example, rather than (as many scholars believe) by the battering the doctrine took in World War I, the Spanish Civil War, and the difficulty in reconciling it with the covenant and the U.N. Charter. Perceptions of justice (or the absence thereof) also are deemed causative in the authors’ discussion of manifestos. The authors identify some 400 manifestos issued by governments over the last five centuries to justify their resort to war. From these, they deduce that “wars could not be entered into for any reason whatsoever. They had to be made in the name of justice” (p. 44). But it hardly follows that the practice of issuing manifestos was a cause of restraint. That a war was claimed to be just does not mean that it was required to be just. Hitler justified Germany’s 1939 attack on Poland under international law by claiming self-defense; Mussolini’s representatives argued for weeks in the League of Nations that Italy’s 1936 invasion of Abyssinia was justified as a response to Abyssinian aggression. The authors present no evidence that any war was averted by a never-issued manifesto that wouldn’t write—like the alternate, never-delivered speech for John F. Kennedy that Ted Sorensen supposedly was unable to compose announcing an attack on Cuba during the missile crisis. All that the pile of manifestos shows is that justice has been such a malleable concept that throughout all of the gory conflicts the authors decry, leaders had no difficulty finding some way to frame their every cause as just.

In sum, even if the authors managed to show a clear change in the incidence of war before and after the signing of the pact, they have not established causality. Too many other factors might have been responsible.

Finally, the book is, alas, punctuated by errors of law and fact that are not immaterial to its causation claims. It’s not true, for example, that in 1949 Chiang Kai-shek “proclaimed the Republic of China” and “establish[ed] it as an entity separate from mainland China” (p. 324), which constituted “a lasting transfer of territory…recognized by the rest of the world” (p. 328). The Republic of China (ROC) was established in 1912. After being forced to retreat to Taiwan during the Chinese Civil War, Chiang, who had led the government since 1928, claimed that all of China was and continued to be the Republic of China, not that the ROC was a sovereign entity separate from Mao’s mainland. Only 20 states today recognize the ROC, which still declines to declare itself a separate state.

It’s not true that under international law, “any member of the state can be held liable for the violation of their state’s legal obligations. If Germany violates international law, any German is responsible, regardless of whether it is his or her fault” (p. 269). Under principles of state responsibility, the whole of a state is accountable for its international wrongs—international judgments against it are payable out of common funds from its treasury, for example—but individual citizens do not incur personal liability for the state’s international transgressions. That the United States violated international law in torturing detainees does not mean that every individual American citizen is potentially subject to punishment because it did so.

It’s not true that visitors to the U.N. headquarters in New York “are leaving the United States and passing into international territory...” (p. 336). Under applicable treaties, U.S. authorities have limited access to specified facilities, but the United States has hardly ceded sovereign territory to the United Nations. Rand McNally need not use a different color for Turtle Bay.

It’s not true that “the crime of aggressive war…is now one of the four crimes that can be prosecuted before the International Criminal Court in The Hague” (p. 415). As of this date, the 2010 Kampala amendment authorizing such prosecutions has not become operational. No one can currently be prosecuted before the court for the crime of aggression.

Most importantly, as noted earlier, the central premise of the authors’ argument concerning wars of conquest—that non-recognition signifies disapproval—is not true. Recognition in international law is wholly discretionary. States can lawfully decline to recognize other states or their territorial boundaries for any reason, or for no reason. Non-recognition connotes nothing.

  1. Conclusion

“Legal revolutions,” the authors tell us, “begin with” the “passing of a law” (p. 331). In fact, that’s rarely the sequence of events. The legal revolution in civil rights in the United States did not begin with the passing of the Civil Rights Act of 1964, nor did the legal revolution in environmental protection begin with the passing of the Environmental Protection Act or the Clean Air Act or the Endangered Species Act. These laws, and other landmark statutes like them, capped passionate, antecedent political movements, just as the Kellogg-Briand Pact capped the ardent pacifism of the post-war era. 

Reformist zeal, however, is not enough to forge laws that work. Law’s efficacy depends ultimately upon propitious, background political conditions. In this important respect, the international and domestic legal systems are alike: The conditions required for effective legal rules are much the same in both. Legal systems work well when actors within the system are relatively equal in power. Future dealings are expected. Trust is high. A consensus exists concerning foundational values. The price of non-cooperation is steep. Individual and collective interests align. Underlying social norms reinforce legal norms. Free riders and transgressors are easily spotted and penalized. The authors insist that law creates power, but it does not: Law’s force derives from underlying political conditions such as these. Absent such conditions, law goes dark.

In the international legal system, the requisite background conditions have not been present in sufficient degree to invigorate international rules limiting the use of force. Nations’ efforts to “abbreviate historical transitions,” in Hobsbawm’s phrase, have produced paper rules, not working rules, that govern the use of force. The authors write that “rules rise and fall together” (p. 421). They’re right. “Laws, like houses, lean on one another,” as Burke put it. The effect of inefficacy is contagion: The entire legal system is discredited when prominent rules are flagrantly violated. Legal reform is set back—and international law is weakened, not strengthened—by pretending that failed rules work. That is the untold story of the Kellogg-Briand Pact and its shattering effect on international law. 


Cite as Michael J. Glennon, How Not to End War (October 17, 2017),