Trends in Prisoner Abuse
How common is it for prisoners to be abused once they surrender, or are they actually courteously cared for more often than we might think? Although there are several ready examples of both brutal and honorable prisoner treatment, identifying overall trends in the conduct of captors has remained difficult in the absence of a common metric for abuse. With such a measure in place, general patterns in the treatment of prisoners across interstate wars from 1898 to 2003 begin to emerge more clearly. As will become evident, there is substantial variation in the treatment of prisoners by level of abuse, specific type of violation, and over time.
Table 1.2 summarizes the frequencies and corresponding percentages of cases falling into each level of abuse for the six dimensions of violations. The total number of cases included across each component sometimes differs, since for some types of violations or wars sufficient material was harder to locate than for others. For instance, information was generally more readily available for execution, while gathering definitive details on labor practices proved more challenging.
As we can see, the two extreme points for execution are almost equally likely: a little under one-third of all capable captors engaged in high levels of direct prisoner killing and a similar amount engaged in low levels of direct prisoner killing. Medium degrees of execution turn out to represent the most common outcome for prisoners, though the relative risks of capital punishment faced by surrendering soldiers is fairly evenly balanced across the three levels of abuse. The frequency of killings conducted by Somalia against captured Ethiopian soldiers during the 1977-78 Ogaden War might not be the norm, but the more benevolent conduct of India in safekeeping the lives of many Pakistani prisoners during the 1999 Kargil Conflict is not majority practice either.80 The fact that prisoners in over two-thirds of the cases faced pronounced dangers of perishing directly at the hands of their captors points to the life-and-death nature of captivity across many conflicts.
A less even distribution is apparent when it comes to the torture of prisoners. Over 40 percent of captors choose not to engage in torture or sparingly resort to harsh interrogation practices. The relatively lower incidence of torture may be due to the greater time, training, and resources required to organize a concerted torture program. The infamous “Hanoi Hilton" of the North Vietnamese had a complex series of holding and interrogation rooms where captives could be interrogated, often for months on end.81 Torture has also come to be seen as a particularly horrendous violation even compared to other crimes.82 Any notion that conduct has been driven by a strong antitorture norm, however, is confronted by the fact that the next most common outcome is high levels of torture. Given the smaller propensity for captors to engage in medium amounts of torture, it appears that states more frequently choose to either systematically adopt or refrain from this method of torment. Should a captor choose to engage in torture, they appear more likely to do so in a comprehensive manner as was evident by both sides in the Iran-Iraq War, which led to the ICRC's exceptional decision to twice publically declare that the laws of war needed to be enforced in the conflict by the international community.83
Table 1.2 Summary for each dimension of prisoner abuse in interstate wars, 1898-2003
jVo/i's: Capable eaptors only.
Denial of legal rights represente osedium-level abuse. Ported conscription represents high-level abuse.
The dire fortunes facing most prisoners are perhaps best summed up by the more than 70 percent incidence in which captives have their rights denied in some meaningful sense. The frequency with which legal rights are breached is perhaps not completely surprising, since this is the broadest category of abuse and is not completely divorced from many of the other types of violations.84 By contrast, the rarity of forced military conscription shows that captors do not resort to all types of violations equally.85 The reluctance of states to coerce enemy combatants into serving in their armed forces makes some sense. Given U.S. concerns during the Vietnam War of the rise in so-called fragging incidents, where American troops even targeted their own superiors, captors may have good reasons to be wary of placing weapons into the hands of erstwhile enemies.
Although direct use of prisoners in battle is uncommon, more frequent is their placement in hazardous labor conditions, such as mining or large construction projects. Around one-third of all cases involved the presence of labor conditions that reached medium or high levels of abuse. Compared to execution and torture, however, hazardous labor conditions reaching the maximum level of abuse are around half as common. The fact that over two-thirds of captors engage in only low-level abuses in this area suggests that labor has been less of a concern as an
overall source of abuse. Some of the practicalities in putting captives to work may put a brake on some of the worst excesses for this dimension of violence. To effectively exploit prisoner labor, working conditions must meet at least modest levels of safety. Even Nazi Germany realized the need to improve conditions for Soviet prisoners somewhat if captive labor forces were to effectively contribute to the war effort for more than a minimal amount of time.86 There still remain numerous instances, from Imperial Japan to Tsarist Russia, where captors sought to exploit prisoner labor to the greatest extent possible with little care for the wellbeing of their captives.
On the other hand, in well over half the cases, prisoners must cope with subpar housing, nutrition, and medical resources, which greatly heighten risks of dying from injuries, hunger, or disease. While many Ottoman prisoners were summarily executed by Bulgarian captors and their allies during the First Balkan War beginning in 1912, a large number of others suffered a far worse death through malnutrition and ultimately starvation as they languished in scantily equipped prison camps. Similarly, executions were not infrequent on the western front during the First World War, but as the conflict wore on British prisoners were far more likely to die in the sparse conditions of German camps than in the immediate war zone, which one British War Cabinet member likened to a policy of “slow assassination."87 Although food and housing requirements are among the areas receiving the greatest attention under international humanitarian law, this has not necessarily translated into healthier or more comfortable living conditions for the vast majority of prisoners.
Looking at each dimension of violations separately reveals some interesting similarities and differences, but it is instructive to assess in a more comprehensive manner the general risks posed to prisoners across past wars. Table 1.3 reports the relative frequency and percentages of each level of abuse using the summary indicator for prisoner treatment, which is equal to the highest observed extent of violations by a captor across the six components of abuse. Brutality is by no means the only fate awaiting surrendering soldiers in captivity; in more than one in four instances captors behaved fairly humanely and treated prisoners more like guests (albeit under a watchful eye) than victims. The conduct of the United States during the Spanish-American War, Russia during the Russo-Japanese War, or Paraguay during the Chaco War shows that the decent treatment of prisoners is by no means a rarity over the last century of interstate warfare.
Table 1.3 Summary of overall prisoner abuse in interstate wars, 1898-2003
Notes: Capable captors only.
Despite numerous episodes where abuses were limited to relatively low levels, history shows this represents the least common strategy adopted by captors. Far more pervasive are medium or higher levels of abuse, which take place at almost identical frequencies. Seemingly irrational decisions, such as continuing to fight on against low odds, become more comprehensible when troops contemplating surrender may be alarmed by what likely awaits them upon capture. The fear of falling into the hands of the North Vietnamese even led several U.S. soldiers to take such extreme actions as calling down air strikes on their own positions during attacks, believing this offered a greater chance of survival than capitulating to the adversary.88 The prevalence of high levels of abuse, followed in close succession by more modest yet still substantial mistreatment of prisoners, directs attention to the poor prospects faced by many troops finding themselves under enemy control. The distribution across the three categories indicates that no level of prisoner abuse clearly predominates. The fate awaiting many soldiers who surrender is oftentimes certainly harsh but must be considered alongside a significant minority of episodes where the conduct of captors remains more encouraging from a humanitarian standpoint.
Looked at together, the relative frequency of each type of violation and the overall level of abuse demonstrate substantial variation in the treatment of prisoners across wars. To provide a better sense of possible temporal trends in the pervasiveness of prisoner abuse, figure 1.1 illustrates changes in the treatment of prisoners over time. The lines in the figure show the percentage of captors engaging in each level of overall abuse during separate twenty-year periods.89 The final period from 1978 to 2003 covers a slightly longer range, since the time periods do not divide neatly into equal intervals. figure 1.1 reveals some interesting patterns. Low-level abuses took place at a fairly modest relative frequency in the first period from 1898 to 1917, while high-level abuses were more common compared to the others during the same time frame (43 percent). This period most notably includes the First World War and the marked gap between low- and high-level abuses lends some credence to claims that these truly were the bad old days of warfare typified by brutality both on and off the battlefield.90
Figure 1.1 Trends in the levels of prisoner abuse over time in interstate wars, 1898-2003
Notes: Percentage of each level of prisoner abuse across twenty-year periods of interstate war onset for the years 1898-2003. Last period includes twenty-six years of war onset since the time frame does not divide into an equal number of years.
The subsequent interwar period saw low- and high-level abuses declining somewhat together. While instances at either extreme still occurred, such as the horrific treatment of Chinese captives by Japan in the Third Sino-Japanese War versus the fairly decent conditions for prisoners in the Chaco War, both levels of abuse were less prevalent compared to the earlier period. Mid-level violations, such as during the 1919 Hungarian-Allies War, were by far the most common category, making up over half of all episodes of abuse. This trend toward the mean perhaps reflects a degree of war weariness from the lengthy trench warfare of the First World War but also the smaller overall scale of conflicts at this time compared to the adjoining periods.
The next period, 1938-57, is the only time where lower forms of abuse predominated, which may be somewhat unexpected given both World War II and the Korean War broke out over this stretch. World War II was certainly characterized by almost unprecedented levels of brutality on the eastern front. The treatment of prisoners was on the whole much better in the western theater, though even here Nazi Germany engaged in more frequent and severe abuses against captured Allied troops. This was also a time when the distribution across each level of abuse was the most even, which is reflected in fairly comparable numbers of cases falling into each category of abuse.
Moving further into the second half of the twentieth century, the following period shows a general pattern of convergence in the relative frequency of the two more violent levels of conduct with lower-level abuse again becoming less pronounced. This was the heyday of several drawn-out insurgencies, particularly in Indochina. However, there were also numerous fairly quick wars, such as the Six Day War between Israel and several of its Arab neighbors, or the even shorter Football War between El Salvador and Honduras, where prisoners were for the most part spared the worst forms of violence.
If the prior period was characterized by some merging in the tendencies behind the two greater levels of abuse, the final span from 1978 to 2003 points toward a slight polarizing trend. Mid-level abuses declined to their lowest point of the periods studied, while both lower and higher levels of abuse rose relative to the prior interval. The last few decades have witnessed many cases heralding the benevolent ideals espoused by many lawyers and just war theorists, such as U.S. conduct in the Persian Gulf War or the conflict in the Cenepa Valley between Peru and Ecuador in the mid-1990s. However, the most recent era also testifies to the persistent horrors of war where captives have been brutalized, whether on the battlefields separating Iran and Iraq or in the Nagorno-Karabakh region disputed between Armenia and Azerbaijan.
Except for the interwar period, no level of abuse appears to overshadow the others. Focusing solely on lower-level abuses seems to offer some promise for humanitarian proponents, since the relative frequency of these more benevolent cases of prisoner treatment shows some gradual, albeit halting, improvement over time. Counterbalancing this positive swing are the stubbornly high levels exhibited by the most severe forms of violence against prisoners. After dipping around the middle of the twentieth century, the propensity for captors to engage in the most deplorable types of abuse has continued to climb such that levels in the most recent era are actually slightly higher than in the first interval one hundred years earlier.
To be fair, episodes involving medium levels of abuse have shown a general decline, especially since the end of the interwar period. Combined with the modest rise of captors refusing to engage in significant prisoner violations, the average level of abuse demonstrates a small drop over time. It is nonetheless a stark reality that the twentieth century largely ended as it began, with the most common fate for captured soldiers being the infliction of systematic and extreme forms of violence. Although several prominent scholars have pointed to a general improvement across a wide range of metrics related to violence,91 this uplifting current does not seem to have gained much traction for combatants literally caught in the midst of war. The general pattern over more than a century of interstate warfare indicates that the treatment of captured combatants is not becoming more humane-high levels of abuse regrettably persist. This is sobering news for those policymakers or activists hoping belligerents would have progressively embraced prevailing international norms of humane conduct and correspondingly placed a higher priority on respecting the health and well-being of prisoners falling into their hands.
With this picture of the trends involving prisoner treatment in place, left untouched thus far is what accounts for why captives are victimized in some conflicts but not others. The propensity of captors toward prisoner abuse reveals significant variation over more than a century of warfare but also within specific time periods. There appear to be no universal pressures pushing all, or even most, belligerents toward either purely cruel or completely compassionate conduct toward captives. The next chapter lays out the main set of factors captors often take into account when deciding how to care for surrendering troops and develops an explanation for why belligerents turn to the decent versus poor treatment of their prisoners.