Prisoners of war hold special status in the Westphalian system. Michael Walzer describes their position as “a strange world this side of the hell of war … deprived of the relative security of political or social membership” (1969, 779). Yet, prisoners of war do not exist on humanity’s periphery but at its center. Mustered from the society of the state, their capture marks a social and legal transition to a new society whose members were titled by a nineteenth century theorist as the “citizens of the world” (Walzer 1969, 778). Prisoners of war simultaneously exist under the jurisdiction of their homeland, their capturer, and humanity at large.
Political science offers many paradigms by which to frame prisoners of war. These theories are the lenses through which we interpret observations and project predictions. We address two of these theories: realism and social constructivism.
Realists view the international system as dominated by interest. State behavior is determined by their want and power to acquire it. International relations regarding prisoners of war in a realist world is based on reciprocity. Prisoners are akin to hostages, held by states to ensure reciprocal treatment toward their own citizens captured by enemy states.
Constructivists view the world as a social construction created through a shared understanding of values. These values are transmitted through communication to create norms of behavior. Constructivists view prisoners of war as existing in a value based world. The actions of states regarding prisoners of war are based on the values, norms, and identities those states hold.
In this paper, we ask whether realism or constructivism offers greater explanatory leverage regarding prisoners of war in the Westphalian system. First, we place the discussion in context by introducing the foundational treaties and major historical phases of prisoners of war. Second, we explore how both realism and constructivism view prisoners of war and what observable predictions those frameworks imply. Finally, we compare those predictions against the real world, using both historical case studies and textual analysis of international treaties. What appears below is not a research paper. We provide no definitive, comprehensive answers. Instead we offer only the crude architecture of what an answer could look like, that constructivism provides the greater explanatory leverage.
Prisoners of War in International Relations
The discussion of prisoners of war treatment has a long history. In early subsistence societies prisoners of war were often killed because low economic productivity limited the number of enemy warriors any society was able to support (Gelb 1973, 71). In civilizations requiring mass labor prisoners of war often become slaves. The most common practice was a combination of these two policies. Male prisoners were slaughtered or sacrificed and female prisoners were enslaved (Gelb 1973, 72). This period is marked with a stark lack of coordination regarding the prisoners of war.
With time, basic coordination developed between warring parties regarding prisoners of war. Evidence of this can be found in Livy’s account of diplomatic exchanges between Rome, Greece, and Macedonia in 202-203 CE. This dialog contains a discussion on the return of Macedonian soldiers and their commander, Sopater imprisoned by the Romans (Dorey 1957, 186). While limited, this is representative of a general movement toward a more organized and universal system on prisoners of war.
The Peace of Westphalia introduced a new system of political organization and through that, a new approach to prisoners of war. The chronic shortage of seamen in the fleets of Spain, France, and England during the 18th century led to the common practice of exchanging or ransoming prisoners during conflict (Anderson 1960, 77). The primary consideration determining whether these transactions took place “was not which power was losing the greater number of men, but which power needed the return of its prisoners more” (Anderson 1960, 78). There was no general international policy regarding exchanges, instead a cartel system was used wherein states agreed to either “all for all” or “man for man” transfers (Anderson 1960, 79). This cartel system represents the prototype of the modern prisoner of war legal and social framework.
The Hague Regulation of 1899, later revised in 1907, offered the first permanent, formal international prisoners of war agreement. The treaty was so highly regarded that American Commander Raymond Stone wrote in 1914:
When … the Great War broke upon an astonished world, we rather took comfort to ourselves in the thought that no matter how swiftly and vigorously military operations might be prosecuted, the Conventions of Geneva and of The Hague would insure humane care and chivalrous treatment to the prisoners of war of both sides. (Stone 1919, 406)
Stone admits basing his idealism on two assumptions. First, the treaties were legally binding on the states in the conflict. Second, that “in this day and generation of high development and in the elements of morality and humanity the belligerents would feel themselves morally if not technically constrained to abide by the principles…” (Stone 1919, 406). Stone’s optimism was not entirely misplaced. World War I saw marked improvements for prisoners of war. Still, during the war the life of prisoners was harsh and many modern prisoner rights were not yet codified. It would be thirty years and another world war before the current legal institution developed.
The Geneva Conventions of 1949 are the core framework dealing with prisoners of war. The rules codified in the treaties apply to all wars between all signatories and leave few issues unaddressed. Unlike the ad hoc agreements of the cartel system which focused on the physical exchange of prisoners, the Geneva Conventions focus on the rights of the prisoners themselves. The Conventions includes detailed agreements on the discipline, diet, treatment, forced labor, housing, services, and the legal definition of prisoners of war. The 1949 treaties are the last major international agreement on prisoners of war and mark the most recent step in a historical trend toward more organized, formalized legal institutions dealing with prisoners of war.
Prisoners of War in Realism
Realism places the heaviest premium on the pursuit of power. The theory also claims to be ruthlessly objective. Fundamentally, realism believes there is a desire for power that motivates the human condition. This drive determines the actions of states. The interest in power is biological and creates a permanent state of insecurity. Since power is relative, the best outcome possible is a balance wherein nobody thinks they can get away with aggression.
The world in realism is inherently non-hierarchical. There is no actor with power over the states and thus the relations between states must but reached through mutual, self-reinforcing agreements rather than universal norms of behavior. In this anarchic world, states only abide by agreements when it is in their interest to do so. This is most often accomplished in reciprocal behavior. That is, states cooperate only in so long as they think the other state is reciprocating the behavior. In this view, the behavior of a state toward prisoners of war held in its territory is only as good as the treatment the enemy provides its own prisoners of war. Positive or negative treatment of prisoners of war by one party is reciprocated by the other.
States in realism contain two primary characteristics of interest. First, they are unitary actors. That is, substate actors have little or no influence on the behavior of states. This poses interesting theoretical questions regarding prisoners of war since they, by definition, are substate actors. The only time negative behavior toward prisoners of war changes state behavior is if it affects the realpolitik power of the state. Similarly, states represent the primary actors in the international system and extra-state actors such as international governmental organizations (IGOs) and non-governmental organizations (NGOs) do not influence state behavior.
Finally, the actions of states are driven by the desire for national security and survival. States work not to spread their ideology or values but to secure themselves in a competitive world through military and economic capacity. The obsession with national interest frames realism’s view of prisoners of war. Prisoners are kept healthy only so long as it is in the national interest to do so. However, when the costs of supporting enemy prisoners become greater than the benefits, states have little reason to maintain them.
While realism does not directly address prisoners of war, we can draw out four predictions. These predictions tell us the kind of information which would prove realism was an accurate and useful paradigm through which to view the world. Later in the paper we will test these predictions against the real world.
First, regime type does not affect state behavior toward prisoners of war. The pressures of international competition in realism’s anarchy are so large as to drown out any influence offered by variations in domestic systems (Morrow 2007, 561). Prisoners of war exist outside the realm of domestic politics. Therefore, the values found in the domestic political structure do not motivate state behavior toward prisoners of war. In realism, brutal autocratic dictators treat prisoners of war the same as stable democracies.
Second, state behavior will be unaffected by treaties, IGOs, and NGOs. Realism’s assumption that substate and extra-state actors have little influence predicts states will comply with treaties, IGOs and NGO if and only if they would do so even if those structures did not exist. Thus, while a state might comply with an international convention “the legal status of the treaty has no real effect on whether they do” and the treaty will be violated as soon as the interests of the state diverge from that of the treaty (Morrow 2007, 560). Therefore, under realism we will observe states ignoring attempts from treaties, IGOs, and NGOs to influence state behavior toward prisoners of war.
Third, prisoner of war treaties will focus on enforceable reciprocity rather than values. Two assumptions will dictate the contents of treaties under realism. First, states will only comply with international treaties if they would have behaved the same without the treaties. Second, the nature of international cooperation in realism is through reciprocity. Following these assumptions, successful treaties should focus not on values (cheap talk in realism) but on the reciprocal, self-reinforcing arrangements. Treaties codifying these reciprocal behaviors will be less likely violated and better appear to withstand the test of time.
Fourth, spirals of retaliation will be observable in history. Morrow points out that “political institutions must be self-enforcing to be sustained. In the language of game theory institutions must form an equilibrium of the game…” (Morrow 2001, 973). Regarding prisoners of war in realism, this equilibrium is created through the reciprocal behavior between warring states. However, when states are unwilling (i.e. not taking prisoners of war) or unable (i.e. lacking the resources to treat prisoners well) to reciprocate, the equilibrium becomes unbalanced. One outcome of such an unbalance can be a spiral of retaliation, wherein a state’s unwillingness or inability to provide sufficient care for prisoners of war leads to retaliation against prisoners held by its enemy, in turn sparking mutual retaliations from both sides. This Axelrod “ALL D” prisoner of war equilibrium should be observable in history.
Prisoners of War in Constructivism
Realism neglects the world of ideas. Constructivists believe international relations is not autonomous to humanity’s structure of ideas, values, and norms. To constructivists “norms trump interests because norms constitute interests” (Morrow 2007, 560). The material power prominent in realism is still important, but the ideas, norms and values are just as important. This material world is only impactful when it is filtered through agents’ ideas of it. Agents are limited in transforming the world to a direction predicated on their thoughts of it. It follows then, that researchers cannot explain the structure of the material world unless understanding the ideational part behind it.
Constructivism gives two primary reasons why states will enter into treaties regarding prisoners of war. These reasons are centered on the values the states hold rather than their interests.
First, states are drawn to joining in organizations with like-valued states. Hawkins (2004) argues that norms fundamentally lodged in the “taken-for-granted beliefs of precedent, international cooperation, and prevention of bodily harm” are likely to be adopted by similarly valued states (799). Constructivism posits that, at its heart, prisoners of war treaties are concerned with the codifying the socialized values of the signatory states on the health and well-being of the prisoners under their care and under its enemies’ care. Thus, like-valued states will join with other states in creating enforceable international agreements on the treatment of prisoners of war. This agreement over the prevention of bodily harm can be shared even between warring states. States can go to war while simultaneously agreeing prisoners of war must be protected.
Second, states might join international prisoner of war agreements to precommit to socialized norms. Precommitment removes the temptation of violating closely held values in the future. The theory, developed by Steven Ratner (2004), argues a state can attempt to “restrain oneself from doing something that one would otherwise do because such restraint will itself directly improve one’s future welfare” (87). During peacetime states holding a strong value of protecting prisoners of war might fear that under the fog of war there could be the temptation to violate those values for some short term benefit. One way to overcome this dilemma is to restrain future action by precommitting to the values through an enforceable international treaty before presented with such temptations.
Paralleling the discussion of realism, we derive four observable predictions on prisoners of war from the constructivist paradigm. These represent the information we expect to find if constructivism represents an accurate lens through which to view the world.
First, the domestic structure of states will affect behavior toward prisoners of war. Specifically, democracies treat prisoners of war better relative to non-democracies. Constructivists argue states comply with international treaties because they have internalized the values and norms underlying such obligations. Since democracies are more open than other forms of government to the civil society norm entrepreneurs advancing these norms, they are more likely to internalize and comply with the norms (Finnemore and Sikkink 1998, Morrow 2007).
Second, states largely comply with treaties, IGOs, and NGOs. These three institutions represent the primary norm entrepreneurs of international civil society and hold, to greater and lesser extents, influence over state behavior through positive or negative social pressures. Even undemocratic states share certain social values (e.g. state legitimacy …) with the international community and these socializations can be exploited by extra-governmental actors to further their goals regarding prisoners of war. Thus, constructivism predicts instances of even autocratic state behavior toward prisoners of war being influenced by treaties, IGOs, and NGOs.
Third, treaties will reflect the shared values of the states rather than reciprocal behaviors. States enter into international prisoners of war treaties because of shared values and norms on their treatment. The text of the treaties should focus on the codification of those shared norms.
Fourth, states which have internalized the norms toward the treatment of prisoners of war will unilaterally comply with those norms even when their enemy is unwilling or unable to do so. State behavior toward prisoners of war under constructivism is driven by internally held values and not external interests. Prisoner of war noncompliance by one state will not effect the compliance of their enemy.
Comparing Realism and Constructivism
The table 1 presents the percent of World War II prisoners of war which died in captivity by state dyad. Realism predicts no relationship between regime type and prisoner of war behavior while constructivism predicts democracies will be more humane toward prisoners of war.
Twenty seven percent of US and Commonwealth prisoners of war held by Japan died in captivity. By comparison “relatively few” Japanese prisoners of war in held by the US and Commonwealth countries died in captivity. Furthermore, four percent and less than one percent of US/ Commonwealth and German prisoners of war died in captivity respectively. The data supports the constructivism prediction. More democratic states (US, Commonwealth countries) systematically treated prisoners of war held by them better than less democratic states (Japan, Germany). These preliminary results open the possibility that the dynamics of the domestic system of states influences their behavior toward prisoners of war.
Treaty, IGO, NGO Compliance
Regarding treaty, IGO, NGO compliance, realism predicts states behave the same whether or not the treaty exists. Constructivism predicts states which internalize the norms of a treaty will follow them.
Some evidence supports the realism predictions. States often wantonly disregard the prisoner of war treaties to which they have ascribed. A memo by the International Committee of the Red Cross (ICRC) written during the Iran-Iraq War bemoans:
…although the Geneva Conventions confers on those prisoners a legal status entitling them to specific rights, the Iranian authorities’ continuous delaying tactics since May 1982, the obstacles and restrictions they have raised and their refusal to allow the ICRC to visit some POW camps have prevented the ICRC from carrying out its work for the prisoners whom the Iranian authorities admit they are holding. (MERIP 1983, 40)
However, this same memo reports that prisoners of war held by Iraq, “have been able for the law few months to correspond with their families in a satisfactory manner” (MERIP 1984, 40). Therefore the evidence suggests the relationship is not so clear cut. The two nondemocratic, warring states, which agreed to same treaties, behave differently regarding compliance toward the Geneva Conventions and the ICRC. Realism provides no explanation. Under realism Iraq has little reason to comply with the Geneva Conventions since Iran is clearly not reciprocating. Constructivism however does offer an explanation. Ropp and Sikkink (1999) note constructivist scholars have theorized that states which accept treaties with no intention of complying with them may find the obligation changes their identity and thus gives them an incentive for compliance. Therefore, while both states agreed to the Geneva Conventions we can see that to some extent (i.e. Iraqi behavior), the treaties have influenced state behavior.
Another example is the Japanese compliance with the treatment of prisoners of war during the Russo-Japanese War and World War II. During the former the Japanese “scrupulously fulfilled its obligations to Russian and German POWs under the treaties at the time” (Morrow 2001, 990). However, this behavior changed during World War II wherein the “Japanese had switched to neglect of prisoners at best and outright abuse of them at worse…” (Morrow 2001, 990). In both cases Japanese enemies complied with the treaties throughout the wars.
Under realism’s interest based paradigm, the Japanese had no reason to switch from compliance to noncompliance. The national interest regarding Japanese prisoners of war did not change. What did change however, was the value Japanese society placed on prisoners of war. Between the two wars Japan developed a “culture of shame” surrounding surrender in battle (Morrow 2001, 990) and this value extended to include US and Commonwealth prisoners. Japanese state behavior change toward prisoners of war is explained by a shift in values, not in interests.
Realism predicts treaties will focus on improving reciprocal, self-reinforcing relationships. Constructivism predicts treaties will focus on the shared socialized values and norms between signatory states.
There are some instances of the reciprocal relationships predicted in realism. Article 4, Chapter Two of the 1923 treaty between Greece and Turkey regarding prisoners of war specifies:
…Greece shall restore to Turkey and shall transport to Smyrna simultaneously all the Turkish prisoners of war detained by her.
Turkey shall thereupon restore to Greece an equivalent number of Greek prisoners of war, officer for officer, soldier for soldier… The remainder of the Greek prisoners of war shall be repatriated by the Turkish Government immediately after the signature of peace… (Agreement Between Greece and Turkey Respecting the Reciprocal Restitution of Interned Civilians and the Exchange of Prisoners of War, 1923)
In the treaty, state behavior is driven by reciprocity. That is, Turkey transfers its prisoners of war only after Greece reciprocates by releasing an equal number of prisoners.
Most major prisoners of war treaties however, deal almost exclusively with the fair and responsible treatment of prisoners. These treaties follow the constructivist predictions by focusing on the shared values states hold. These values range from the use of muzzled dogs to track down escaped prisoners to a mandatory day of rest from forced labor on the Sabbath (Prisoners of War 1919, 14 -19). These types of clauses represent the bulk of prisoners of war treaties and yet are too specific, minor, and/or undetectable to be used in reciprocal relationships. The signatories must have known ex ante that violations to many of these rules would never be uncovered in time to retaliate. Therefore, it is fair to conclude that these treaty contents represent the socialized shared values between the states.
Realism predicts spirals of retaliation whenever one state is unwilling or unable to reciprocate behavior toward prisoners of war. Constructivism instead predicts unilateral compliance when the state has internalized treaty norms.
There is little doubt retaliation is commonplace is war. Morrow calls retaliation “the unstated but recognized tool of enforcement” (2001, 981). After the Dieppe Raid in 1943 German prisoners of war reportedly had their hands tied for a time before being freed, a violation of the treaties in place at the time. German infantry soldiers responded to the news by binding the hands of Commonwealth soldiers they held in the field (Morrow 2001, 981).
However, direct acts of reciprocity occur almost exclusively at the tactical level and do not represent official policies of states. William O’Brien argues state level retaliation has largely discontinued as an accepted practice. He observes that “most forms of reprisal are no longer legal under international customary and conventional law; these include retaliation in kind against [prisoners of war] for mistreatment by the other side, and reprisals against civilians” (1995, 46). If we limit retaliation toward prisoners of war to that of the state levels of analysis we find support for the predictions of constructivism. That is, states which have socialized treaty norms unilaterally comply even in the face enemies unwilling or unable to reciprocate.
In 1914, Commander Stone observed:
Upon the entry of the United States into the war, our government deemed it expedient to utter a pronouncement to the effect that while the Government of the United States, for good and sufficient reasons, did not consider that the actual provisions of the several Conventions of Geneva and the Hague were operative in this war, it did consider that the principles underlying the provisions were of force and that they would be followed by the United States as a general guide in the circumstance. (Stone 1919, 407)
The United States, while acknowledging the letter of the treaties was not in force, announced publicly that it considered the norms codified in the treaties in force and would act accordingly. These actions do not reflect reciprocity but a drive to fulfill the socialized identity and norms the nation had for itself and its actions.
In the 16th through 18th centuries the concept of “parole” was commonplace. Gentlemen and aristocrats captured in battle could be released and allowed to return home if they gave their word not to rejoin the fighting (Walzer 1969, 779). By World War II this ethic was largely abandoned except for one curious and illuminating case. During fighting in the Philippines a small group of Americans was captured by Japanese infantry, who upon receiving the Americans’ word not to fight, paroled and freed them. The US government accepted the paroles of the soldiers and deemed parole as legally prohibiting the men from offering any support to the US war effort against the Japanese (Walzer 1969, 780). Realism offers no explanation for the position of the US government. Since the probability of developing a reciprocal parole system was remote, there was no reason for the US government harm its interest by unilaterally abiding by the parole agreement. However, constructivism offers an explanation. Honoring the parole was in compliance with socialized norms the US had internalized, even when doing so was a unilateral action.
The United States is not alone in unilateral compliance with prisoner of war agreements. The mission statement of the Israeli Defense Force (IDF) refers to the ideal that “IDF soldiers will not use their weapons and force to harm human beings who are not combatants or prisoners of war” (Sucharov 2005, 182). The ideal is held in such high regard that IDF soldiers possess, the option and obligation to disobey blatantly illegal orders (Sucharov 2005, 182). This code of conduct makes reciprocal retaliation by the Israeli government legally impossible. The outlawing of prisoner of war reciprocity defies a realist explanation since the threat of retaliation is what the paradigm holds as the central tenant of prisoner of war interstate relations. Through a constructivist lens however, the actions reveal a state with internalized norms of behavior and a desire to follow them even in the face of an enemy unwilling or unable to do so.
Through a brief overview of prisoners of war in both the realist and constructivism paradigms we have found the former wanting. From each of the two theories we derived four symmetrical, observable predictions. The end result was that time and again realism’s interest based approach to international relations failed to offer an explanatory edge over constructivism’s value based approach. While this paper does not attempt to posit itself as a comprehensive test of the two theories, the results do offer weak support that constructivism is better able to provide explanatory leverage regarding prisoners of war in the Westphalian system. That is, without historicizing and contextualizing state behavior through appreciation of their values and norms, theories and research on prisoners of war offer less ability to interpret, explain, and predict.
“Agreement Between Greece and Turkey Respecting the Reciprocal Restitution of Interned Civilians and the Exchange of Prisoners of War.” 1924. 18 (2, Supplement: Official Documents):90-2.
Anderson, Olive. 1960. “The Establishment of British Supremacy at Sea and the Exchange of Naval Prisoners of War, 1689-1783.” 75 (294):77-89.
Axelrod, Robert M. 1984. The Evolution of Cooperation: Basic Books.
Dorey, T. A. 1957. “Macedonian Troops at the Battle of Zama.” 78 (2):185-7.
Finnemore, Martha, and Kathryn Sikkink. 1998. “International Norm Dynamics and Political Change.” 52 (4, International Organization at Fifty: Exploration and Contestation in the Study of World Politics):887-917.
Gelb, I. J. 1973. “Prisoners of War in Early Mesopotamia.” 32 (1/2):70-98.
Hawkins, Darren. 2004. “Explaining Costly International Institutions: Persuasion and Enforceable Human Rights Norms.” 48 (4):779-804.
Merip. 1984. “Treatment of Prisoners-of-War in the Iran-Iraq Conflict.” (125/126, The Strange War in the Gulf):38-9.
Morrow, James D. 2001. “The Institutional Features of the Prisoners of War Treaties.” 55 (4, The Rational Design of International Institutions):971-91.
—. 2007. “When Do States Follow the Laws of War?” 101 (03):559-72.
O’Brien, William V. 1995. “The Rule of Law in Small Wars.” 541 (Small Wars):36-46.
Ratner, Steven R. 2004. “Overcoming Temptations to Violate Human Dignity in Times of Crisis: On the Possibilities for Meaningful Self-Restraint.” 5 (1).
Ropp, Steve C., Kathryn Sikkink, and Thomas J. 1999. The Power of Human Rights: International Norms and Domestic Change: Cambridge University.
Stone, Raymond. 1919. “The American-German Conference on Prisoners of War.” 13 (3):406-49.
Sucharov, Mira. 2005. “Security Ethics and the Modern Military: The Case of the Israel Defense Forces.” 32 (2):169-99.
Walzer, Michael. 1969. “Prisoners of War: Does the Fight Continue after the Battle?” 63 (3):777-86.
Christopher R. Albon is a political science Ph.D. specializing in armed conflict, public health, human security, and health diplomacy.