4
Talking to Fight
In the previous chapter, I provided evidence that a combination of latent external pressures and battlefield trends explains a great deal of negotiation strategy. I found consistent evidence that higher latent pressures for peace lead to more frequent negotiations, regardless of belligerents’ actual desires to either stop or continue fighting. Simultaneously, a more distinct battlefield trend in one side’s favor increases the likelihood that any negotiations that do take place are sincere efforts to resolve conflict. These findings support Hypotheses 1, 2, and 3 as characterized at the end of chapter 1.
While chapter 3 analyzed how battlefield activity can influence negotiation behavior, this chapter explores how negotiations can reshape the battlefield. Hypotheses 4 and 5 speak to these dynamics. This is an important test of my theory. If the only consequence of unsuccessful peace negotiations is that war continues on the same path as before talks began, then it is easier to dismiss diplomacy as a mechanical reflection of the battlefield. A core innovative claim I have made is that insincere negotiations not only fail to end hostilities but can be exploited to reshape the subsequent trajectory of conflict. Evidence that talks can affect the nature of fighting itself is therefore essential to painting a more complete picture of my conception of wartime negotiation.
This chapter provides that evidence. I first return to my quantitative data to substantiate the remaining two hypotheses derived from my theory. In support of Hypothesis 4, I confirm that periods of negotiation are associated with lower levels of active hostilities on the battlefield. The effect is especially pronounced for internal negotiations organized by the belligerents themselves. My theory suggests that these moments of relative quiet on the battlefield are natural for belligerents that sincerely want to stop fighting but also necessary for belligerents that merely want to abuse negotiations to improve their political and military position. On that note, I demonstrate that negotiations that do not result in peace are followed by battlefield trends that systematically differ from the state of fighting prior to the onset of talks. This result, which is consistent with Hypothesis 5, is novel and underscores the importance of viewing negotiations as a very real extension of war.
In the latter half of the chapter, I walk through a brief qualitative analysis of the relationship between fighting and negotiating in two conflicts: the War of the Roman Republic of 1849 and the Cenepa Valley War of 1995. Both support Hypotheses 4 and 5 but also more fully depict how belligerents can use diplomacy to directly sculpt the political and military battlefields in their favor.
Quantitative Analysis
Chapter 2 described the negotiation and battle data I have amassed to analyze ninety-two interstate conflicts between 1823 and 2003. I now use these data to investigate the relationship that negotiations have with both contemporaneous and subsequent battlefield activity.
A Quieter Battlefield
Hypothesis 4 as described in chapter 1 avers that negotiations should be associated with a lower level of active hostilities on the battlefield. I can use my quantitative negotiation and battle data to test this claim across dozens of interstate conflicts. To do so, I treat the number of active battles per war-day as the primary outcome of interest. This is a count variable, which is most appropriately analyzed using a Poisson model.1 The key explanatory variable is the occurrence of negotiations on a given war-day. Standard errors are clustered by war.
Many of the control variables I include in these models are identical to those described in chapter 3, so they will not be repeated here. Three new variables, however, become relevant when considering the impact of negotiations on active hostilities. The first is my measure of recent imbalance. My own theory and analysis suggest that it is appropriate to account for contemporaneous battlefield trends when analyzing how diplomacy affects the state of the battlefield.
The second is the existence of a formal ceasefire. Many periods of negotiation also feature ceasefires that hold alongside diplomatic bargaining. For example, during the final phases of the Saudi-Yemeni War of 1934, Ibn Saud announced a ceasefire on all fronts when also indicating his willingness to negotiate with Yemen. Saudi Arabia had the upper hand on the battlefield at the time, and talks that took place alongside a ceasefire promptly forged an agreement filled with concessions by Yemen.2 Ceasefires, by definition, are designed to pacify the battlefield. It is important to ensure that any effects I find regarding diplomacy are not simply consequences of ceasefires. I therefore include a binary variable for whether a ceasefire was active on a given war-day.
The third is the trend in the number of recent battles. One may be concerned about reverse causation: instead of negotiations leading to reduced hostilities, perhaps reduced hostilities precede negotiations. To account for this possibility, I take the number of battles that have taken place over the previous sixty days of war (or the entirety of the war to that point, whichever is closer) and determine the slope of the best-fitting line across those data. The estimated slope is a measure of an active battle trend.
Figure 4.1 reports the estimated effects from two models of active battles. The first model analyzes negotiations as a whole. The negative and statistically significant coefficient for the negotiation variable makes clear that periods of diplomacy feature quieter battlefields. Stated in more intuitive terms, the occurrence of talks is associated with a 30 percent decrease in the number of contemporaneous battles.
Other important patterns emerge upon disaggregating negotiations by whether they are internal or external. The disaggregated model shows that much of the pacifying effect of negotiation is driven by internal negotiations. Internal talks have a markedly negative and statistically significant effect, while external negotiations are associated with an estimated coefficient that is close to zero and is not statistically significant. Moreover, external talks are associated with a 6 percent decrease in the number of active battles from a baseline rate, while internal talks are associated with a much larger 43 percent decrease—a sevenfold jump. A linear hypothesis test verifies that this difference between the two coefficient estimates is statistically significant as well (p << 0.01). Negotiations primarily pursued by belligerents are more likely to be reflections of sincere intentions to stop fighting. Meanwhile, talks resulting from third-party efforts will create some drawdown in hostilities but reflect a more hollow effort to find a lasting peace. This is a natural implication of my theory.
Active battle trends over the previous sixty days exhibit a positive and statistically significant relationship with the number of active battles on a given war-day. This is hardly unexpected, yet recall that the purpose of including this variable was to address concerns of reverse causality. Even after adding this measure to the regressions, we still see that negotiations have a meaningful impact on contemporaneous levels of fighting.
The coefficient for ceasefires across both models is negative and slightly larger in magnitude than what I find for internal negotiations, but it is statistically significant only at the 90 percent level. Even if ceasefires were to exhibit a highly statistically significant effect, the fact that negotiations can have an impact on fighting that is comparable to that of ceasefires—agreements explicitly made to pause hostilities—merits attention. Notably, despite the common belief that negotiations are tethered to ceasefires, my data suggest that the two are only weakly related across two centuries of war.3 About 9 percent of negotiations include ceasefires, and only 30 percent of ceasefires include negotiations.
The pacifying impact of negotiations on the battlefield could be construed as evidence that the back-and-forth construct of fighting and negotiating in bargaining models of war is reasonable.4 There are three reasons this conclusion is not completely justified. First, the effect identified in my models represents a decrease in fighting rather than a complete cessation. Second, the differential effects I find for internal and external negotiations are not captured by formal models. Third, the alternation process assumes that no other aspect of the war could change during or as a consequence of a negotiation period. A core contention of my theory is that negotiations can change a belligerent’s reversion outcome even when negotiations do not yield agreement. This leads to my last set of statistical tests.
Changing Battlefield Trends
Negotiations that lead to conflict settlement reveal themselves to have been sincere efforts to forge peace. A fair share of negotiations that do not end conflicts may also be sincere, albeit misguided, efforts, but it is also likely that many of these talks are insincere ploys to buy time and revitalize one’s political and military fortunes. If insincere negotiations generally help to realize these cynical goals, then we should see evidence that the trajectory of the battlefield changes before and after talks that do not lead to war termination. This prediction is at the heart of Hypothesis 5.
To assess this prediction, I focus on a subset of my data that analyzes only war-days surrounding the 143 negotiation periods that do not terminate conflict. For each of these negotiations, I extract the sixty days prior to the start of the negotiation period and the sixty days following the end of the negotiation period.5 I compare battlefield trends on the prenegotiation days with those on postnegotiation days to see whether the battlefield has systematically changed trajectory. My main explanatory variable is a straightforward indicator for whether a war-day occurs in the sixty-day postnegotiation window following the cessation of talks.6 Since I am making direct comparisons of temporal windows within wars, it is not necessary to include a complete battery of controls. I do, however, add two. The first captures any concurrent negotiation taking place within a sixty-day window. The second is a simple linear time trend variable, which accounts for the possibility that some changes on the battlefield are simply consequences of time. I return to this point later in this section.
Figure 4.2 displays the results of several regression models that analyze differences in battlefield activity on the sixty days before and after all negotiations that do not end a conflict. For now, I focus on estimates depicted in black, which represent models that include all negotiation periods. The first model, which looks at differences surrounding all negotiations, exhibits a positive and statistically significant coefficient for the postnegotiation period. This is consistent with Hypothesis 5. If I calculate the variance of the recent imbalance measure within each war, I find that the median value of these variances is 0.594. As such, the coefficient for the postnegotiation period, 0.225, represents a sizable shift on the battlefield.
The fact that negotiations have this prominent effect on subsequent hostilities is conspicuous. If this dynamic indeed exists, it has significant theoretical and policymaking implications and thus demands further investigation. We can better understand this relationship and how it is manifested on the battlefield by replicating the analysis using two alternate measures of battlefield activity. Each provides additional insight into precisely how fighting changes after talks end.
The War Target’s Rally
The first is my measure of momentum. Recall from chapter 2 that this is simply the recent imbalance measure before taking its absolute value. Momentum takes a positive value when recent battlefield trends favor the war initiator and a negative value when trends favor the target. The second model in figure 4.2 indicates that the upswing in recent imbalance is reflective of battlefield trends that tilt toward the war target’s advantage. The point estimate for the postnegotiation period is –0.232. For reference, the median value for momentum within wars is 0.805.
The fact that battlefield trends undergo a reversal of fortune for the target dovetails with an array of research on the importance of the first-mover advantage in war. By having the ability to select the time, place, and manner in which to begin hostilities, war initiators have a vital chance to choose conditions that optimize their likelihood of success.7 Targets may not expect to fight, or they may be unable to predict when, where, or how an initial attack will occur. Initiators can exploit this window of opportunity during which their opponent is not fully prepared or mobilized.8 Clausewitz highlights the importance of initially waylaying the opponent through “plans and dispositions, especially those concerning the distribution of forces” so that opening combat “confuses the enemy and lowers its morale.”9 Sun-Tzu makes a similar point in The Art of War: “Whoever is first in the field and awaits the coming of the enemy, will be fresh for the fight; whoever is second in the field and has to hasten to battle will arrive exhausted. Therefore the clever combatant imposes his will on the enemy, but does not allow the enemy’s will to be imposed on him.”10 In other words, the first-mover advantage exploits the initial and temporary disparity between an opponent’s potential capabilities and its mobilized capabilities. Negotiations can therefore play a crucial role in allowing some war targets to mitigate their second-mover disadvantage by buying time to more fully mobilize their military forces and adjust both their political and military strategies.
Lining Up with Prewar Expectations
Results based on recent imbalance and momentum show that the battlefield tends to shift in one side’s favor following failed negotiations. Another factor we may want to consider is the extent to which the outcomes of fighting are consistent with belligerents’ expectations from before the war began. Scholars frequently note that actors’ strategic choices are influenced by prewar beliefs.11 My ability to gauge this consistently across dozens of interstate wars is limited, but one plausible measure is the CINC ratio, which I use as a control variable in the previous chapter. Recall that the CINC ratio is based on the Composite Index of National Capability measure from the National Material Capabilities dataset.12 I sum the annual CINC measures for all active belligerents on each side in the war, accounting for the entry or exit of individual states. I then divide the war initiator’s total score by the sum of both sides’ scores. Values close to 0 reflect wars with more capable war targets, while values close to 1 indicate more capable war initiators. To be sure, actors do not have complete information regarding all parties’ material capabilities, but we can at least assume that there is enough commonly observable information for both sides to have a somewhat similar understanding of each side’s relative strength.13
These values must then be compared to a measure of observed battlefield outcomes. To produce this, I use my battle data to calculate what proportion of battles have been won by the war initiator up to that day in the conflict. As with the CINC ratio, proportions close to 0 or 1 reflect fighting that favors the war target or war initiator, respectively. By taking the absolute value of the difference between the CINC ratio and observed proportions from fighting, I can measure the degree of inconsistency between expected and observed outcomes. When inconsistency is close to 0, the proportion of battles won by each side perfectly aligns with what we might expect based on their material capabilities. Increasingly higher values suggest a widening gap between expectations and reality.
The third model in figure 4.2 reflects how levels of inconsistency change around failed negotiations. The results are noteworthy. Talks are followed by battlefield developments that decrease the level of inconsistency. In other words, fighting becomes more consistent with prewar expectations. The average variance in the inconsistency measure within wars is 0.005. The estimated coefficient of –0.013 for all negotiations, despite sounding small, is substantively significant.
These findings underscore the importance that material capabilities indeed have on the consequences of fighting. Pauses from diplomacy allow more of the belligerents’ dormant abilities to be realized, which leads subsequent battlefield activity to more closely reflect each side’s underlying strength. The average level of my inconsistency measure on the last day of war, however, is 0.318—a sizable number on a 0-to-1 scale. Many conflicts end on the back of fighting that does not align very closely with the material resources each side can bring to bear. This discrepancy could reflect the fact that the CINC measure is a flawed predictor of war outcomes.14 It may also indicate how the bargaining range expands due to the costs of fighting, permitting agreements that are not directly aligned with prewar expectations. Yet this still provides evidence that a costly lottery approach to understanding war may not be sufficient. The ebbs and flows of intrawar activity are highly influential in determining whether and why conflicts come to an end.
Negotiation’s Fading Effects
The results to this point have treated all negotiations equally, assuming that their potential effect on the battlefield remains constant over time. My argument offers at least two reasons we may want to differentiate between negotiations over the course of hostilities. First, any first-mover advantage that a war initiator wields at the beginning of hostilities should eventually diminish. Past studies have found that war initiators tend to fare worse the longer a conflict lasts, pointing to the weakening of their initial strategic and operational advantages.15 Beyond the advantage that war initiators often enjoy when starting wars, both sides should also be able to narrow the gap between their underlying capabilities and their ability to mobilize those capabilities for war. Second, the more frequently negotiations fail to end conflict and are instead followed by a shift in battlefield trends, the more likely it should be that at least one belligerent would avoid negotiations until it feels more confident that talks will be sincere.
Returning to figure 4.2, I now focus on the results depicted in gray, which disaggregate negotiations by whether they were the first failed negotiation in a war. The first period of negotiations appears to drive much of the postnegotiation effect for all three measures of battlefield activity. This distinction is especially prominent for momentum and inconsistency; the estimated coefficient for the postnegotiation effect for these first rounds of talks is between double and quadruple the magnitude of the estimates for all subsequent negotiations.
The strength of these postnegotiation effects may raise two connected concerns. One is that these changes on the battlefield are simply a function of time and would occur regardless of whether negotiations ever took place. Another is that my finding is a statistical phenomenon based on a reversion to the mean. A couple of considerations settle these matters. First, the models reported in figure 4.2 include a variable for a linear time trend, and these do not undermine the main finding. Second, I perform a series of placebo tests where I create random periods of negotiation in the data and assess whether the postnegotiation effect appears in these simulated analyses. All of my placebo tests produce no evidence of a meaningful effect.16 These null results lend further support to my argument that negotiations share a nontrivial relationship with belligerents’ reversion outcomes and thus with their abilities to regather themselves and redouble their efforts on the battlefield. The rebalancing of the battlefield certainly requires the passage of time, but time on its own does not sufficiently identify the moments in which battlefield trends actually change. Negotiations provide the mechanism through which these shifts occur.
Put together, the quantitative results tell a coherent story about how negotiations that do not end conflicts can instead shape belligerents’ abilities to prosecute the continuing war. Negotiations provide a moment of relative quiet and opportunity for belligerents to better mobilize their forces and harness their underlying material capabilities, which may not be used as fully or efficiently at the outset of hostilities. A main effect of failed negotiations is that they bring battlefield outcomes closer in line with prewar expectations, which are defined by each side’s observed material capabilities. In many but not all cases, the side that benefits more from this respite is the side that did not have as much time to prepare for the conflict: the war target. These stabilizing effects are most salient in the earliest rounds of negotiation but become less prominent as wars proceed.
Talks and Fights in Two Wars
Despite the strength and stability of my statistical findings, one may believe they capture an average effect that does not reflect changes actually seen in any individual war. The quantitative analysis also captures only the effect that failed negotiations have on the battlefield, but my theory also suggests that diplomacy has political ramifications as well.
To address this, I now turn to two extended cases that more clearly illustrate how negotiations are used to influence the state of the battlefield in a manner consistent with my large-N quantitative results. Both conflicts offer meaningful evidence in favor of Hypotheses 4 and 5.
The War of the Roman Republic of 1849
The War of the Roman Republic presents an archetypal example of how a state can use negotiations to reshape the battlefield for its own benefit. In this particular conflict, insincere negotiations were important and successful enough not only to defeat a belligerent but to functionally terminate its existence as a political entity.
The Fall and Rise of Rome
The year 1848 was a tumultuous one for the Italian states. Revolutionaries in several Italian states in the North, including Lombardy and Venetia, sought to wrest themselves free of the Austrian Empire’s rule and to create a unified, independent, and liberal Italian republic. King Charles Albert of Piedmont, receiving the apparent support of Pope Pius IX, declared war against Austria in March 1848. This revolutionary war, however, proved unsuccessful against Austria’s superior forces.
The failed revolution did not discourage the underlying cause. Indeed, many nationalist Italians directed their ire at Pius IX, the ruler of the Papal States (see figure 4.3). Upon his election as pope in 1846, Pius IX—an Italian himself—was hailed by those in Italy as a liberal champion of reform, republicanism, and progressive causes. Yet during the revolutions two years later, Pius IX reversed course, embracing conservative values and refusing to confront the Austrian Empire.17
Enraged nationalists orchestrated the assassination of the prime minister of the Papal States, Pellegrino Rossi, on November 15, 1848. Days later, the Swiss Guard was replaced by the “Civic Guard,” and Pius IX was effectively imprisoned in the Papal Palace. He managed to escape on November 24 by dressing as a priest and fled to the port of Gaeta in Naples, itself the capital of the Kingdom of the Two Sicilies.18 With the pope gone, liberals claimed power over the Papal States, taking control of the center of the Italian peninsula. The new government instituted a bevy of desired political reforms, including the establishment of a democratic government, religious freedoms, the abolition of the death penalty, and key social reforms to help the poor and unemployed. The ascendant liberals also sought to declare war against Austria once more.
On February 9, 1849, a republican government backed by a new constitution declared itself the Roman Republic. Modeling itself after the governing structure of ancient Rome, the new republic was led by a triumvirate: Carlo Armellini, Aurelio Saffi, and Giuseppe Mazzini. Of these three, Mazzini—a politician, activist, and leader of several failed uprisings—arguably wielded the most influence in the Italian revolutionary movement and the broader democratic campaign to take hold across Europe. The armed forces of the Roman Republic were commandeered by Giuseppe Garibaldi, who himself was a follower of Mazzini and a fierce proponent of a unified Italian republic.
Exiled in Gaeta, Pius IX sought the military assistance of other states that were sympathetic to Catholicism. One of these was France, where Louis Napoleon had been elected the first president of the Second Republic on December 20, 1848.19 Louis Napoleon was caught between two difficult crosswinds. On one hand, he had personally participated in an insurrection against the Papal States in 1831 and personally supported the republican project in Rome. On the other hand, his recent electoral victory rested heavily on the votes of French Catholics, who obviously desired a restoration of the pope and Papal States.20 Napoleon ultimately chose to prioritize electoral politics and deployed an expeditionary force to reclaim Rome for the papacy.
Fighting, Talking, and Seeking the High Ground
French troops landed at the city and seaport of Civitavecchia, about forty miles northwest of Rome, on April 25. Led by General Nicholas Charles Oudinot, the forces, numbering between seven and eight thousand, headed toward Rome, believing they would be welcomed as liberators. Anticipating the arrival of the French, Garibaldi and his men prepared defenses on the Janiculum—a hill in western Rome—between April 27 and 29.
Hostilities, and the overall war, began in earnest on April 30. Oudinot and the French Army were not prepared to meet much, if any, resistance. They were misinformed, and in more than one manner. Oudinot’s map of the Janiculum was out of date, causing him to lead his troops to a gate that had been closed. A daily cannon that fired from the Janiculum at noon, meant simply as a way to publicly communicate time, alarmed the French, who thought they were being attacked. Garibaldi took advantage of these deficiencies and successfully repulsed the French attackers, inflicting approximately 850 casualties and diminishing 11 or 12 percent of the French troops’ overall strength.21 Oudinot withdrew his forces to Civitavecchia and requested further reinforcements.22
Garibaldi sought to press this advantage and pursue the French all the way to the sea. Mazzini, however, decided to avoid further confrontation, hoping to instead establish a relationship with the French. Mazzini believed that the French Republic was fundamentally sympathetic to the Roman Republic’s liberal cause and that it would also be willing to fight against the Austrians.23
France’s loss at the Janiculum was deeply humiliating and politically precarious for Louis Napoleon. French officials had not expected Roman resistance, much less to be driven back by an outnumbered opponent. Napoleon feared that news of this loss would tarnish his reputation as a military leader. Oudinot and Napoleon went as far as to egregiously misrepresent the events at the Janiculum, informing the French Assembly that the “affair of April 30 is one of the most brilliant in which the French troops have taken part.”24 Oudinot also bragged that no French troops had been taken prisoner, conveniently leaving out the fact that the Romans had freed all their French prisoners as a sign of goodwill.25 But the actual outcome could not be concealed; news of the loss at the Janiculum spread in Paris by May 7.26 Citizens and members of the French Assembly were upset to have been misled but ultimately seemed uninterested in further hostilities. Public sentiment bent toward peace with the Roman Republic.
These developments on their own were already troublesome, but another complication loomed: French legislative elections, meant to elect the first National Assembly of the Second Republic, were scheduled for May 13 and 14. The Party of Order, which had been instrumental in Louis Napoleon’s electoral victory the previous year, had only 23 percent of seats in the Constituent Assembly that had created the new republic. A larger loss against the Roman Republic would sink Louis Napoleon’s chances of obtaining a favorable assembly.
He therefore had two significant reasons to initiate negotiations. The first was that based on battlefield developments, the relative costs of unabated fighting were high. Another assault on the Roman Republic in the current state of affairs would likely be a failure, and any move by Garibaldi to counterattack could prove dangerous. Additional French losses not only would be harmful militarily but would ripple through upcoming legislative elections. The second was that the costs of negotiating were quite low compared to the potential benefits from extracting side effects. By stalling for time, Louis Napoleon—belatedly realizing the broad support for the new republic in Rome—could send reinforcements to Oudinot and mount a much more serious attack, all while satisfying the general public’s desire for a suspension of hostilities during the election period. The French government would have much to gain from insincere negotiations.
On May 8, Louis Napoleon charged the career diplomat Ferdinand Lesseps with a mission to go to Rome with the ostensible aim of forging a ceasefire. We now know, however, that the French government had no intention of negotiating a meaningful peace agreement and that Lesseps was unaware of this ploy. On the exact same day, Louis Napoleon quietly sent a message to Oudinot, stating that “our military honor is at stake. I will not suffer it to be compromised. You may rely on being reinforced.”27 It was also around this time that Austria and the Two Sicilies involved themselves in the war, sending their own troops to Rome.
Lesseps left Paris on May 9 and arrived in Rome on May 15. Despite Mazzini’s interest in speaking with the French government, it appears that France initiated the actual negotiations that began on this date. Lesseps received an enthusiastic welcome and was enraptured by Mazzini’s affability and warmth.28 By May 17, Lesseps and Mazzini arranged a two-week armistice during which the two sides would continue negotiating to forge a more permanent agreement.
This two-week ceasefire and negotiations proved critical for France. In that brief time, the Party of Order won an outright majority of seats in the first National Assembly, and Louis Napoleon’s promised reinforcements arrived. During talks, Oudinot had asked Lesseps to request permission for French troops to occupy several hills overlooking the city. Oudinot promised that this was only for the health of the troops and would in no way endanger the Romans. The triumvirate granted permission, believing that discussions with Lesseps were going well.29 But Oudinot and others in Paris had made this request in order to prepare for additional hostilities once reinforcements arrived. Seemingly earnest diplomatic bargaining had thus allowed Oudinot to place his troops, now including Austrian soldiers and numbering twenty thousand, on advantageous high ground just two to four miles outside Rome.
On May 31, as the ceasefire period came to a close, the Roman triumvirs and Lesseps signed the Treaty of Peace and Alliance. The agreement stated that the French would become allies with the Roman Republic against Austria and Naples. French troops would remain outside Rome and aid in its defense, but they would not join any offensives against the Austrians or Neapolitans. Lesseps, perhaps realizing that he had ceded more than he should have, insisted on adding a clause saying that the treaty had to be ratified by the French Republic.
When Lesseps returned to the French camp with the agreement in hand, Oudinot refused to sign. He instead gave Lesseps a note from Paris: “The Government of the Republic has put a stop to your mission. You will be good enough to start upon your return to France as soon as you have received this despatch.”30 Lesseps soon learned of rumors spread by his own government that the mission had “affected his brain.”31 Louis Napoleon’s government had successfully used diplomacy—and Lesseps—to buy time. Now wielding far greater political and military power than he did mere weeks earlier, Louis Napoleon immediately nullified the treaty and initiated a siege against Rome on June 3. French forces eventually peaked at 34,000, while Garibaldi could collect a maximum of only 18,600 less well-trained and less well-equipped men.32
The Roman Republic was unable to mount a proper defense against such a tightly orchestrated siege. By the end of June, the Roman Assembly gathered to discuss their next steps. Three options became part of the discussion: they could surrender, they could continue fighting, or their government and troops could withdraw from Rome and continue their resistance from the Apennine Mountains. The Roman Republic ultimately settled for the first option. A truce was negotiated on July 1, and the Roman Republic was disestablished, allowing for the reinstatement of Pope Pius IX the following year. While Garibaldi and some of his forces opted to fall back to San Marino and fight another day, the present war and the Roman Republic were finished.
Talking One’s Way to Victory
The events of the War of the Roman Republic offer a direct example of the strategy and potential of insincere negotiations. Louis Napoleon faced an unexpected loss at the Janiculum and recognized that continued fighting would be perilous not only in a militaristic sense but also in a political one. As such, Louis Napoleon foresaw relatively high costs to continuing hostilities without making any changes. His government thus made an initial offer to open negotiations. In support of Hypothesis 4, which predicts that periods of negotiations should be linked with lower levels of active hostilities, the battlefield was dramatically quieter over the two weeks during which negotiations took place. Lesseps the diplomat was unaware that his own government was using talks to create a reversion outcome that was far more favorable to the French Republic when negotiations would be intentionally sunk. The war target, fully capable of mobilizing itself to fight more effectively and to adjust strategy, used the veil of diplomacy to facilitate that process. This case also vividly illustrates Hypothesis 5, which anticipates that negotiations that do not result in settlement will be followed by a battlefield trend that differs from what existed prior to talks. The striking reversal of fortune in France’s favor following the abandonment of peace negotiations forced an end not only to the war but to the Roman Republic.
One may believe that the war would have unfolded in the exact same manner even if no negotiations had taken place. The Roman Republic largely accepted the French diplomatic overture because it had few other options. Its primary goal was not only to defend its new existence but to garner the French Republic’s political and military support, especially against the Austrian Empire. The victory at the Janiculum was, in the triumvirate’s view, a necessary step for the Roman Republic to defend itself. But every choice the republicans made from that point was designed to ingratiate themselves with the French. The triumvirate took the necessary gamble of negotiating with France in hopes that Louis Napoleon would agree to defend the Roman Republic. They knew that the agreement they had made with Lesseps was risky, but all they could do was hope that France would fulfill it. In the moments before Mazzini took the proposal to the Roman Assembly for approval, he told Lesseps, “The positions with regard to which we are about to facilitate your occupation, and the privilege which you reserve for yourselves of only repulsing our foreign enemies if they come directly in contact with yourselves, leave our political existence at the mercy of your good faith.”33 It was also clear that Garibaldi’s forces would not win an outright war against the French, and Oudinot’s military forces would have been reinforced even in the absence of any diplomatic interactions between the two sides. Recall that both Lesseps and military reinforcements were deployed on the same day. Perhaps the war would have ended in exactly the same manner whether or not Lesseps and Mazzini engaged in diplomatic discussions.
This conclusion, however, overlooks several benefits that Louis Napoleon could not have accrued if he had adopted a strategy based solely on military force. By opening diplomatic channels, he created an opportunity for General Oudinot’s forces not only to be reinforced but to relocate themselves to a strategically advantageous spot. This relocation process could have faced armed resistance and created undue costs for the French military if it were not permitted by the Roman Republic over the course of negotiations. Moreover, due to the apparent success of negotiations, Garibaldi’s forces were directed to return to Rome, and they did so between May 31 and June 2—immediately before France would renege on its agreement—even though they had the ability to continue fighting. The fact that Garibaldi’s men were concentrated in Rome made General Oudinot’s task of neutralizing the Roman Republic considerably easier. The republican army would resume their armed resistance once the siege began, but their efforts were curtailed by their disadvantageous position. Consistent with the findings from the quantitative analysis, the battlefield tilted more dramatically and in the war target’s favor and became more consistent with prewar expectations after negotiations were halted. The occurrence of talks also quelled domestic opposition to the war in France on the eve of legislative elections. Louis Napoleon managed to deflect political pressures for peace by engaging in the theater of diplomacy but with the ultimate intent of stalling for time until his political fortunes were more secure. These benefits outweighed any costs Napoleon may have perceived from being willing to negotiate with a largely unrecognized state.
Mazzini’s writings after the war indicate his indignation over being betrayed. A letter he wrote on August 7, 1849, two months after the Roman Republic dissolved, bemoaned that “France had been guilty of falsehood” and that the Roman Republic had been a “victim [who] was then on the ground, with a dagger in her throat.”34 Regardless of whether members of the Roman Republic were deceived, two facts remain. The first is that the republic had no other tractable option but to negotiate and hope that the other party would agree to the triumvirate’s requests. The second was that one belligerent strategically used negotiations to promote and redirect its war aims rather than settle them.
In sum, France would have ultimately prevailed in this conflict, and the Roman Republic had very little chance of survival. Yet the costs and outcomes of war are not binary. Louis Napoleon used diplomatic talks to create a more favorable reversion outcome that mitigated both political and military liabilities. By saving some French lives and resources, the French Republic accomplished an easier victory, and Louis Napoleon more effectively navigated complicated political waters. None of these would have been possible without the exploitation of negotiations.
The Cenepa Valley War of 1995
The Cenepa Valley War was a brief but sharp struggle between Peru and Ecuador.35 By the war’s peak, approximately five thousand troops were deployed to an area of jungle approximately twenty-seven square miles in size.36 The conflict offers an effective illustration of how belligerents on both sides can use diplomacy to strategically manipulate not only battlefield activity but also the broader political environment. Through negotiations, both Peru and Ecuador extracted side effects that enhanced their gains from conflict.
A Watershed Moment
The dispute had roots in the administrative borders drawn by the Spanish Empire in the eighteenth century. Many of the lines drawn at the time were somewhat arbitrary, as they were all under Spanish dominion and often ran through uninhabited swaths of jungle. But as South American states began to claim independence in the early 1800s, questions about territorial control became more pressing. Peru and Ecuador had political and economic reasons to determine a clear boundary line between them. Between 1802 and 1936, both states engaged in at least thirteen concrete but failed diplomatic efforts to define a border.37 Some of these discussions involved third-party assistance, with the United States and Spain being the most active mediators. None of these projects succeeded in determining a mutually acceptable line of demarcation.38
After more than a century of deepening dissatisfaction and nationalistic fervor, the dispute came to a head in July 1941, when Peru and Ecuador fought a monthlong war over their border.39 The conflict quickly revealed Peru’s superiority; its military was thirteen times the size of Ecuador’s,40 and its invading force was five times larger than the Peruvian contingent it faced on the battlefield.41 In light of the decisive Peruvian victory, four nations—Argentina, Brazil, Chile, and the United States—helped the two countries arrange a settlement. The eventual agreement, called the Rio de Janeiro Protocol, more commonly referred to as the Rio Protocol, was signed on January 29, 1942. Peru was to ultimately withdraw from Ecuadorian territory. In exchange, Ecuador was to acknowledge that seventy-seven thousand square miles of disputed land would be granted to Peru, and Ecuador would end its claim for direct land access to the Amazon River. The four states that helped forge the protocol cosigned as guarantors of the agreement.
The protocol quickly resolved about 95 percent of existing territorial disputes between Peru and Ecuador. In 1946, the US Army Air Corps attempted to map the disputed area to help finalize a territorial arrangement. The aerial survey sparked a new and unexpected controversy: the Cenepa River, previously thought to be short and inconsequential, was actually 120 miles long and created two watersheds. The Rio Protocol spoke of only a single watershed, and its description of the revised border no longer applied properly to approximately fifty miles of disputed territory in the Cordillera del Cóndor (Condor Mountain Range).42 Ecuador claimed this meant the protocol was not executable and that a new deal had to be made. Peru argued that the protocol could remain intact because the boundaries had already been determined, regardless of the number of watersheds. With neither side being willing to cede ground to the other, Ecuador withdrew from the demarcation process in 1948. By 1960, Ecuadorian president-elect José María Velasco Ibarra adopted a hard-line stance that the entire Rio Protocol was null and void because it was signed under duress following the 1941 war and because it could not be fully executed due to the discovery of new geographical features not addressed in it. Subsequent Ecuadorian governments all perpetuated this claim.
Disagreements regarding this territory continued for decades, and both states gradually increased their military presence in the Cenepa Valley (see figure 4.4). Armed forces briefly clashed in another limited conflict in January 1981. This dispute is often called the Paquisha War, named after the Ecuadorian base that Peru claimed was established in Peruvian territory.43 Peru was again quick to demonstrate its military superiority and seized the Paquisha outpost in two weeks. Due to Ecuador’s unwillingness to recognize the Rio Protocol, the Organization of American States (OAS), rather than the guarantor states, oversaw an agreement where both states pulled their forces back to clearly defined territory. Further bloodshed was avoided, but the underlying issue was left unresolved.
At the start of the 1990s, Ecuador began to deploy military units near the Cordillera del Cóndor and Cenepa River. Ecuadorian forces built three new bases—named Tiwinza, Base Sur, and Cueva de los Tayos—on high ground in the disputed territory. These outposts were heavily fortified and surrounded with numerous land mines. Peru’s internal struggles with the Shining Path party and guerrilla organizations led the government in Lima to do little in response to these developments. Once President Alberto Fujimori managed to capture the Shining Path’s leaders in 1992, however, he began to focus more on Ecuador’s activities in the Cenepa Valley. Peruvian forces had a few minor encounters with Ecuadorians in 1994. January 9 and 11 saw two brief incidents between the two sides’ troops, during which troops were captured and sent back to their commanding officers.44 The gradual increase in the number of run-ins between soldiers raised concerns for Ecuadorian forces, who believed the Peruvians were lurching toward a larger offensive.
With these concerns in mind, on January 26, Ecuador launched a surprise attack on a Peruvian helipad. The rocket and mortar fire killed three Peruvian soldiers and forced the rest to flee into the jungle. Ecuadorian troops renamed the helipad Base Norte.45 Armed hostilities were underway.46
Multiple “Guarantees” of Peace
Third parties cast a large shadow over Peru and Ecuador. Knowing the international community would likely disapprove of conflict, both belligerent states quickly and readily attempted to argue the justness of their cause to third-party actors, making the case that the other was the aggressor and violator of international law. In a letter dated January 27, the Ecuadorian minister of foreign affairs wrote to the president of the United Nations Security Council, claiming that “Peru had launched military operations against Ecuadorian army positions located in Ecuadorian territory. In exercise of the right of self-defence recognized in Article 51 of the Charter of the United Nations, my country was naturally compelled to respond to Peru’s attacks.”47 One day later, the Peruvian deputy minister of international policy also wrote to the president of the Security Council, claiming that Peru “has been the victim of armed aggression by Ecuador.”48
As a response to its retreat on January 26, Peru initiated a series of ground assaults between January 28 and 31. These developments triggered growing global concern. On January 29, César Gaviria Trujillo, secretary general of the OAS, started flying between Quito and Lima to try to resolve the dispute.49 Ecuador expressed interest in working with the OAS, but Peru refused OAS mediation on the basis that any mediation needed to follow the framework of the Rio Protocol and thus be led by the guarantor states. Indeed, the international community agreed with Peru’s stance. Ernesto Cardenas, the president of the Security Council (and coincidentally the UN ambassador from Argentina), indicated that the United Nations would stay out of the matter and directed the guarantor states to address the crisis.50
On January 31, representatives from the guarantor states gathered in Rio de Janeiro, two days after the fifty-third anniversary of the original protocol’s signing, and invited both Peru and Ecuador to come discuss peace. Both belligerent states agreed to participate. One day before talks began, reports from Lima indicated that a large number of well-equipped Peruvian troops were headed to the Cordillera del Cóndor.51
Negotiations at Rio de Janeiro initially appeared promising. By February 1, delegates from Peru and Ecuador announced that they had reached the terms of a ceasefire. Argentinian president Carlos Menem publicly expressed confidence that the war would soon end, and President Fujimori of Peru even publicly announced that his military would “immediately begin the withdrawal of tanks” from the disputed land.52
But this optimism did not last. The ceasefire declared on February 1 appeared to fail within one or two days, and the guarantors were unable to develop a written agreement that satisfied both belligerent states. The primary disagreement involved the size and scope of military drawdowns. Peru asked for a limited withdrawal that would create a demilitarized zone approximately five miles wide. Ecuador, fearing that this would allow Peruvians to later retake any disputed outposts, sought an unconditional ceasefire and complete demobilization of troops.53 Talks were ended on February 5, with the guarantor states planning on another round of mediation in Brasilia on some undetermined date.
Two days before the Rio de Janeiro negotiations officially ended, Peruvian helicopters began a new round of attacks that bombed Ecuador’s outposts in Cueva de los Tayos and Base Sur.54 Despite Ecuador’s insistence that it had not lost any posts, Peru appeared to take these two bases within a handful of days.
Hostilities continued in bursts over the next week. Because the thick jungle environment made ground operations difficult, an increasing amount of fighting took place from the air.55 Ecuador shot down three Peruvian planes, while Peru shot down two Ecuadorian warplanes.56 At least one Peruvian helicopter was also taken down during an attempted attack on the Coangos outpost.57
The outpost at Tiwinza, which was one of the disputed positions closest to Ecuador, became a focal point for both sides.58 Ecuadorian forces had secured the area with mines, artillery, and other traps prior to the war and made significant efforts to maintain control over it. One notable aerial photograph shows Ecuadorian troops at Tiwinza; they had used stones to write out a phrase that Ecuadorian president Sixto Durán Ballén would popularize during the war: “ni un paso atrás” (not one step back). Multiple Peruvian assaults against Tiwinza had failed to make headway.59 On February 12, Fujimori stated on national television that additional Peruvian forces were headed toward Tiwinza, which he claimed was the last outpost on Peruvian soil that remained in Ecuadorian hands.60 A large-scale Peruvian assault began in the early hours of February 13.61
Fujimori was quick to declare victory that evening, proclaiming on television that “the Peruvian flag waves in Tiwinza” and declaring a unilateral ceasefire that would come into effect the next day at noon.62 Ecuadorian officials denied Peru’s claim and invited both the media and the Red Cross to come visit Tiwinza, Base Sur, and Cueva de los Tayos to see that they remained in Ecuadorian hands.63 We now know in retrospect that Peru had already taken Base Sur and Cueva de los Tayos earlier in the month but that their latest assault on Tiwinza failed.64 Fujimori likely made his announcement to mislead his public in order to increase his standing before elections would take place in April. In any case, Ecuador agreed to reciprocate Peru’s ceasefire, but this arrangement collapsed within hours.65 Between the fourteenth and sixteenth, the two sides continued fighting but failed to appreciably change the situation on the ground.
On February 17, the four guarantor states brought Peru and Ecuador back together for talks, leading to the Declaration of Itamaraty. The agreement called for a ceasefire, gradual withdrawal, and creation of a demilitarized zone that would allow military observers from the guarantor nations—designated as the Military Observer Mission, Ecuador Peru (MOMEP)—to monitor the withdrawal. The agreement left some key questions unanswered. For one, it did not specify when the two sides would begin their troop withdrawals. Moreover, it did not address the actual border dispute. The demilitarized zone was meant only to stop the fighting and not to define new boundaries.66 The primary issue of demarcation would be saved for talks, which would happen at some point in the future.
Both sides violated the ceasefire repeatedly, with most infractions taking place around Tiwinza. Peruvian battalions staged an attack on Tiwinza on February 18 but were thwarted by mines and artillery fire.67 Some of the heavier fighting of the conflict took place on February 22, with Peruvian forces attempting to penetrate Tiwinza yet another time. This was the one and only moment Peruvians reached the outpost, but they were soon pushed back out.68 Ecuador’s military command indicated that fourteen soldiers died as a result of this successful defense.69 Despite the fact that Ecuador successfully repulsed the Peruvian attack, this clash was the country’s worst day in terms of casualties, leading Ecuadorians to call February 22 “Black Wednesday.”70 These continuous flare-ups prevented the team of approximately twenty MOMEP military observers called on by the Declaration of Itamaraty from entering the disputed areas.71 During a visit to the Cueva de los Tayos outpost on February 24, Fujimori briefly came under mortar fire, with one mortar exploding about 150 feet from his group of soldiers and reporters.72
Yet over the next few days, skirmishes faded away, and the guarantor states pressed Peru and Ecuador to adhere to the Declaration of Itamaraty. On February 28, the six parties gathered once again in Montevideo, Uruguay. As part of the inauguration ceremony of incoming Uruguayan president Julio María Sanguinetti, Peru and Ecuador signed the Montevideo Declaration. According to the declaration, the belligerents “reiterate[d] their commitment to the immediate and effective cease-fire established in the Declaration [of Itamaraty],” and the “Guarantor Countries reiterate[d] their commitment to continue to carry out their obligations under the Rio de Janeiro Protocol.”73 Both countries’ military forces began withdrawing from the area at the end of March, and MOMEP verified that the process was completed by the middle of May.74
The monthlong war in the Cenepa Valley was estimated to have cost a total of $500 million across both states.75 Ecuador lost and could not recover control of the outposts at Cueva de los Tayos or Base Sur. Fujimori would claim that “there was not an Ecuadorean soldier left on Peruvian soil,”76 but Ecuadorian troops did manage to hold Tiwinza, some other bases slightly in the periphery of the disputed area, and the Cóndor Mirador base further south where they had successfully infiltrated the Peruvian rear.77
Peru, Ecuador, and the guarantor states eventually formulated the Act of Brasilia on October 24, 1998. The agreement designated some limited territorial allocations to each side, guaranteed Peru the right of free navigation on the Amazon River, and established an ecological park that would be co-operated by both nations. The agreement was not too dissimilar from the Rio Protocol fifty-six years before, but it also granted Ecuador a small patch of land at Tiwinza (in Peruvian territory) to recognize Ecuador’s human losses there.78
Exchanging Fire and Accusations
Militarily speaking, the war in the upper Cenepa Valley represented a form of reserved victory for Ecuador. The Peruvians’ significantly larger military, as well as Peru’s history of easily subduing Ecuador on the battlefield, would have suggested an easy win for Peru. Ecuador’s ability to keep control of Tiwinza, despite Fujimori’s claims, represented the first Ecuadorian military triumph over Ecuador since the battle of Tarqui in 1829.79 Peruvian political observers also expressed disappointment about the conflict’s outcome, noting that this was Peru’s “worst defeat since 1879” (alluding to Peru’s loss against Chile in the War of the Pacific of 1879–83) and that “anything less than a swift annihilation of Ecuadorean forces represents a first-ever defeat.”80
Ecuador’s successes in claiming Base Norte on January 26 and maintaining control of Tiwinza rested on its first-mover advantage. The Ecuadorian military enjoyed a series of strategic benefits, many of which were accrued by learning lessons from its embarrassing loss in 1981, professionalizing its military, and engaging in long-term preparations for a potential conflict.81 Ecuador’s outposts, which had been constructed starting in 1991, were on high and familiar ground that would limit the effectiveness of Peruvian aircraft. Short supply lines and a system of footpaths facilitated the movement of soldiers, supplies, and information. Thousands of land mines and guns defended Ecuador’s outposts.82 Ecuador also picked an “ideal” time and place to start the fight. Peru had so little expectation of an international conflict that it redirected much of its military and intelligence apparatus to deal with the Shining Path and the Túpac Amaru Revolutionary Movement at home.83 Even when Peru started paying more attention to events in the Cenepa Valley in the months before hostilities began, it did not fully understand Ecuador’s military presence or capabilities, assuming little had changed since the 1981 dispute. As a result, Peruvian forces were few in number, deeply undersupplied, and not prepared for a surprise attack anywhere near the Cenepa Valley. Logistical operations were poor and Peruvian roads in the area were terrible, forcing Peru to use helicopters to transport supplies and support to its inadequately prepared soldiers—all in airspace that Ecuadorians more effectively controlled.84 But Peru also had the ability to alleviate some of these handicaps if it could find more time to prepare.
Thus, when talks began in Rio de Janeiro on January 31, neither side was ready to settle. The belligerents had not yet exchanged or learned enough information to consider a lasting settlement. Peru wanted to reverse these unfavorable conditions, while Ecuador wanted to capitalize on them. Insincere diplomacy played a significant role in both belligerents’ attempts to promote these goals. The need and latitude to extract side effects from negotiations were facilitated by the immense pressure that the guarantor states placed on Peru and Ecuador to engage in discussions. Diplomats from the guarantor countries did not appear to understand this situation, believing that Peru and Ecuador would easily accept peace. When the first round of talks formally ended on February 5, Alexander Watson, the chief US representative in Rio de Janeiro, admitted that he originally “was planning to be here only six hours.”85
For Peru, talks created a chance to mobilize more of its forces and send them to the Cenepa Valley. Discussions for a potential ceasefire agreement were active on January 31 but began to fall flat. By February 2, diplomats indicated there were “slim hopes” of any negotiated solution. But later that evening, the two sides presented new proposals and spent the night attempting to consolidate them into a single document. On February 3, Durán Ballén of Ecuador announced a willingness to accept a ceasefire on these terms. That same day, however, Peru asked for a series of changes to the proposed agreement. Importantly, a report on February 4 noted that “some analysts here [in Rio de Janeiro] say Peru is prolonging the peace negotiations to give its troops time to advance on disputed borderlands.”86 Peru thus focused on using negotiations insincerely to regroup and reorganize more of its military potential.
For Ecuador, where many were riding a wave of nationalistic fervor and optimism, negotiations were an opportunity to curry international favor for its cause while directing ire toward Peru. Durán Ballén reportedly told his National Security Council that he would request the guarantor countries’ diplomatic assistance because Ecuador would gain an advantage by apparently acting in good faith while Peruvian forces repeatedly attempted to take Tiwinza and the other bases by military means.87 Indeed, Durán Ballén widely communicated that Ecuador had accepted the guarantors’ ceasefire proposal on January 31, only for Peru to ask for numerous unacceptable modifications that sank the deal.88 Ecuadorian communications secretary Enrique Proano also spoke to the press on February 5, claiming that Peru was attempting to look accommodating in negotiations “to make it appear that Ecuador has taken an intransigent position, and to justify the escalation of an armed aggression against a pacifist country.”89 Durán Ballén also went on series of trips to the South American guarantor states between February 6 and 7 to show Ecuador’s professed interest in peace and to seek a general demobilization (instead of Peru’s request to create a temporary demilitarized zone).90
Hypothesis 4—that periods of negotiation co-occur with lower intensity of battlefield activity—finds partial support here. Levels of fighting settled slightly at the outset of Rio talks when both sides were initially discussing the possibility of a formal ceasefire, as well as when the belligerents’ delegates claimed to have reached a ceasefire on February 1.91 Yet Peru bombed both Cueva de los Tayos and Base Sur on February 3, two days before talks ended. Notably, these attacks all used aircraft; ground forces were not involved, as they were still waiting to be reinforced.92 Regardless, consistent with the theoretical basis for Hypothesis 4, Peru and Ecuador frequently declared a willingness to halt fighting over the course of the war, blamed their opponent for any violations, and used these activities to further both their political and their military causes.93
For whatever caveats exist with respect to Hypothesis 4, the Cenepa War illustrates Hypothesis 5 well: conditions on the battlefield became markedly different following the end of failed negotiations at Rio. The additional time allowed both sides to activate more of their latent military power and send it to the Cenepa Valley. Approximately 1,500 troops were in the valley at the start of the war, and this number quadrupled to 6,000 by the end.94 As a result, most of the overall hostilities and casualties during the war took place in the last two weeks. This delayed uptick shows how much the chance to mobilize mattered, as well as how much early fighting may not have been sufficient to convince either belligerent to revise its beliefs or strategies.95 This opportunity was especially useful for Peru, which seized Cueva de los Tayos and Base Sur in the immediate aftermath of the Rio talks falling apart. These battle victories significantly undermined Ecuador’s position in the Cenepa Valley and better reflected the reality of Peru’s overall military capabilities. By early February, these military realities pressed President Durán Ballén to publicly acknowledge that the Rio Protocol was in force but to argue that it was not executable in the disputed area due to its inaccurate description of the geography there.96 His government also began to call Argentina, Brazil, Chile, and the United States “guarantors” instead of “friendly countries.”97 The Itamaraty and Montevideo Declarations were both built on top of the Rio framework. Ecuador’s recognition of the 1942 Rio Protocol after three decades of deeming it null and void was a major policy victory for Peru.
Consistent with the quantitative evidence for Hypothesis 5, the first round of talks had the greatest impact on changing the battlefield’s trajectory. The additional negotiation that produced the Declaration of Itamaraty on February 17 was followed by additional fighting, including Peru’s deadliest attempt to take Tiwinza on February 22, but none of these efforts appreciably changed the states’ control over the outposts. Both countries had already mobilized much of what they would mobilize to fight the war.
One may argue that the war’s ultimate outcome would not have been very different in the absence of negotiations. Ecuador almost certainly could not have clinched a swift or total military victory against Peru, and Peru would have eventually mobilized more troops to fight Ecuadorian troops in the upper Cenepa Valley. It is nonetheless notable that Peru’s battlefield turnaround took place right at the tail end of the first talks in Rio—and that similar shifts did not occur at any other moment. Moreover, recall that insincere negotiations produce, beyond the battlefield, other side effects that are not captured by my quantitative analysis. Both belligerents engaged in public posturing after negotiations in order to frame themselves as the peace-seeking party and their opponent as the aggressor. Ecuador’s government was especially adept at manipulating the public’s understanding of the conflict. Durán Ballén’s trips to Argentina, Brazil, and Chile following the Rio talks were widely reported and drew sympathy to Ecuador’s cause. Some days later, Peru also started sending former ambassadors and other notable people across Latin America and the United States to make the Peruvian case. A report from that time stated that this action “was launched partly in response to the belief that Ecuador was ahead in the public relations war.”98 Ecuador’s diplomatic offensive, which projected an aura of strength and confidence based on the military’s successes, may have helped Durán Ballén acknowledge the Rio Protocol—which significantly expedited a settlement—while also creating conditions for the guarantor states to recognize Ecuador’s control over Tiwinza.99
Tiwinza notwithstanding, Ecuador fell short of an ideal outcome. The Ecuadorean plan in the Cenepa Valley essentially involved a fait accompli: seizing the disputed territory and forcing the international community and guarantor states to accept the reality of Ecuador’s de facto control.100 This goal was partially realized at best. Yet the fact that both sides could celebrate some measure of success rested on how they used the guarantors’ diplomatic efforts to reshape the military and political battlefields in their favor.
Externally motivated negotiations had cross-cutting impacts on the war’s outcome. In one respect, guarantors may have facilitated insincere negotiations that furthered each side’s war aims. Both Ecuador and Peru realized political and military victories, as limited as they may have been, by exploiting the guarantors’ diplomatic endeavors. Yet in another respect, the high latent pressure the belligerents felt to settle the conflict may have allowed them to back out of fighting and enter negotiations more easily. Neither state wanted the war to escalate dramatically,101 but it was not clear whether the war would remain limited. Over one hundred thousand men were mobilized by the two nations to potentially fight for their country.102 The war claimed only hundreds rather than thousands of lives, and diplomacy likely limited this fallout.
Words Can Kill
Few would deny that fighting, in some form or another, helps dictate when and why belligerents choose to talk to one another during war. Yet far more people may be more resistant to the idea of that story in reverse: talking can directly influence the nature of fighting.
In this chapter, I have presented both quantitative and qualitative analysis attesting to the effect that wartime negotiations have on the subsequent trajectory of the conflict. My quantitative study of interstate wars between 1823 and 2003 shows that periods of diplomatic bargaining between the belligerents are consistently associated with lower levels of hostilities on the battlefield—even when we also account for formal ceasefires. This pattern is consistent with Hypothesis 4 in chapter 1. But my findings go one step further in showing that these temporary lulls in fighting provide moments of opportunity for belligerents that can and want to fight more effectively. In the immediate aftermath of negotiations that do not resolve conflicts, the battlefield swings in one side’s favor more dramatically than was the case before talks began. Additional tests show that these shifts are directly tied to negotiations and are not simply statistical artifacts where the battlefield inevitably changes over time. These results lend strong support to Hypothesis 5.
The War of the Roman Republic in 1849 and the Cenepa Valley War of 1995 illustrate how these dynamics play out in individual conflicts over discrete negotiation periods. In the former case, France used talks with the Roman triumvirate to arrange a deal that Louis Napoleon immediately nullified as soon as his troops had relocated to more advantageous high ground and once his party had prevailed in legislative elections. The republicans, who had been unsure of France’s intentions, paid the price of losing their inchoate nation while Louis Napoleon minimized his own nation’s losses. In the Cenepa conflict, both belligerent states leveraged the enormous diplomatic pressure that outside states placed on them to extract benefits for themselves. Peru used the slowdown in hostilities during mediation to reinforce woefully unprepared troops in and around the battlefield. Ecuador used talks to play the role of the peace-seeking party and went on a diplomatic offensive afterward to continue appealing to international audiences. Both countries profited, in slightly different manners, from negotiating insincerely with one another.
These results stress the importance of viewing negotiations as a strategic act that, in specific circumstances, belligerents exploit in service of their war objectives. Scholars and policymakers frequently think of negotiations as producing one of two outcomes: the war ends, or the war proceeds as if nothing meaningful had happened. Neither ripeness theory nor the bargaining model of war fully acknowledges or furnishes a theoretical explanation for whether, when, and why diplomacy is leveraged to produce side effects that change belligerents’ reversion outcomes and the dynamics of the subsequent hostilities. Such a binary form of thinking reflects an overly restrictive view of how diplomacy can affect conflict, potentially producing dubious and possibly harmful policy prescriptions about how to bring wars to a peaceable end.
Put together, this and the previous chapter present a comprehensive empirical case in support of the five primary hypotheses underlying my theory of wartime negotiations. But the four brief historical case studies across these two chapters each speak to separate parts of my argument. The next chapter delves into a war—and, specifically, its diplomatic activity—that illustrates and substantiates the entirety of my theory.