Printer Friendly

Counterfactuals and the proportionality criterion.

In order for a resort to war or military force to be morally justified, it is widely held that in addition to there being a just cause, the good achieved by the war must be proportionate to the harm caused. As I write this, the importance of the proportionality test is being exemplified in the ongoing war between Israel and Hezbollah. In that war, many agree that Israel has the right to defend itself in light of Hezbollah's acts of aggression. (1) What has loomed large in public debate is whether Israel's response has been proportionate. In this essay, I will address certain general theoretical issues surrounding the proportionality criterion. My focus will be developing and defending what I call a baseline account of proportionality and explaining the role counterfactual considerations play in that account. At the end of the essay, I will use that account to provide a proper framework for debating the proportionality of the conflict between Israel and Hezbollah.


The role of the proportionality criterion is to weigh the relevant consequences of resorting to war. More precisely, the criterion requires that the relevant good effects of a war be proportional to the relevant bad effects the war will cause. (2) That just war theorists, using the proportionality criterion, concern themselves with the consequences of war is hardly surprising, given that most of us, whether or not we subscribe to just war theory, accept that consequences are a central consideration in judging the morality of actions and decisions, including those made about and in wars. It is surprising, however, that despite the proportionality criterion's long history, there continues to be a significant lack of clarity, precision, and agreement about how it should be defined and applied in practice.

What is clear is that applying the proportionality criterion requires making comparisons in order to arrive at a judgment. In particular, one must compare relevant good and bad effects, which raises difficult questions. For example, should one count only the good effect of achieving the sufficient just causes of a war--those causes that are of sufficient importance or gravity (and of the correct type) to justify one state (or appropriate body) intervening militarily in, or declaring war on, another state--or should one count all good effects? And with respect to bad effects, should all bad effects of a war count, or, for example, should some bad effects of an unjust enemy's actions not count or at least be somewhat discounted? Another complexity arises in making the needed comparison, however--a complexity that has not been adequately addressed by just war theorists. This is the issue of what baseline a (possible) war should be compared against when making a proportionality calculation.


The Diachronic Approach

In dealing with this issue, it will help to distinguish two possible temporal approaches to identifying the baseline--a diachronic approach and a synchronic approach. On the diachronic approach, the relevant baseline is temporally situated prior to the war. More specifically, those using the diachronic approach often use, as a baseline, harms, injustices, or general states of affairs that occurred (or were occurring) prior to the war. In the proportionality calculation, the prewar baseline is then compared with the harms and/or goods that are caused (or expected to be caused) by the war and its aftermath. To highlight the problems with using the diachronic approach, it will help to consider a couple of examples based on the recent wars in Afghanistan and Iraq.

In the 2001 war in Afghanistan, it was sometimes emphasized in general discourse that the number of civilian deaths in the war at some stage surpassed the number of people killed in the terrorist attacks on September 11, 2001. Individual reasons for highlighting this fact surely varied, but it is likely that at least sometimes the purpose was to implicitly or explicitly make the case that the war was, as a result, disproportionate. Such a conclusion, however, would not be warranted. First, on any credible view of proportionality, one must do much more than compare a set of prewar (unjust) harms to the harms a war causes. In particular, there are a variety of relevant good effects that must be included when making proportionality calculations. Relevant good effects in the war in Afghanistan arguably included stopping or reducing future attacks by al-Qaeda and its supporters, punishing or bringing to justice those responsible for the original terrorist attacks, deterring other terrorist groups and their supporters from undertaking or promoting terrorist acts, and achieving the humanitarian objective of overturning a horrifically repressive regime in Afghanistan. There is room for debate both about the war's effectiveness in achieving these goods and about whether they are all relevant, but the point remains that there are surely some relevant good effects that must be included in proportionality calculations, and the simplistic diachronic approach just described fails to take account of them.

There is a further and more significant problem with this example of a diachronic approach. The problem is that in directly including prewar casualties and injustices, the diachronic approach includes harms that are not directly relevant in proportionality calculations. As was noted earlier, the proportionality calculation's function is to weigh the relevant consequences of the war. Clearly, prewar harms and injustices are neither good nor bad consequences of the war. (3)

(Prewar harms and injustices do play a significant moral role in just war theory and in proportionality calculations; it is simply that the role is not as direct as the proposed diachronic approach would suggest. For example, the September 11 attacks did underpin a number of the suggested sufficient just causes for going to war in Afghanistan, such as preventing further attacks, bringing to justice members of the group that perpetrated the attacks, and deterring or stopping supporters of terrorist groups. And in making the proportionality calculation, the September 11 terrorist attacks also were a crucial determinant in deciding the moral value attached to many of the relevant good effects the war was intended to achieve, such as punishment, deterrence, and prevention.)

Sometimes, rather than comparing a single or small set of prewar injustices or harms to the harms caused by a war, a diachronic approach to proportionality will instead compare the general prewar state of a people or country (the prewar baseline) to both the harms incurred in the war and to the general postwar state of the people or country. For example, in evaluating the 1999 NATO intervention in Kosovo, one might contend that the appropriate test involves comparing the state of the Kosovars prior to the bombing campaign with the state of (and harm caused to) the Kosovars both during and after the bombing campaign. This approach, however, is misguided. The problem with this approach, and more generally with any diachronic approach, is that any prewar baseline, no matter how specific or general, will not properly capture the relevant consequences of the war. (And, to repeat myself, the proportionality criterion's function is precisely to consider the relevant consequences of a war.) The reason prewar baselines do not properly capture those consequences is that the relevant consequences are determined not by comparing the war (and its results) with prewar conditions, but rather, roughly speaking, by considering what would otherwise occur if the war is not (or had not been) undertaken. Relevant consequences of a war crucially include not only what we might call the direct harms inflicted and goods achieved, but also the less direct consequences of, for example, preventing certain harms or injustices from coming to pass. The prewar condition of a people or state is simply not an adequate stand-in for what conditions would have been like had the war not occurred and so cannot be used to evaluate the proportionality of the war. Thus, in the Kosovo war, the appropriate comparison, and one that was surely central in the thinking of the NATO leaders, was what would happen to the Kosovars if the NATO forces did not take military action. (4)

The Synchronic (Counterfactual) Approach

The preceding discussion suggests that the proper baseline to use when making proportionality calculations should not be diachronic, but instead should begin at the same time that the decision is made (or will be made) to resort to war. Furthermore, the baseline must invoke counterfactual considerations. Here I find myself in agreement with the late Gregory Kavka, who observed: "We must also compare the war and its effects to what the world would have been like had the war not occurred, i.e., to a counterfactual situation." (5) For example, in evaluating the proportionality of the recent war in Iraq, we surely need to make some judgment about what would have happened if the United States and its allies had not gone to war. Would Saddam Hussein's regime have been able to maintain power, or would it have been overthrown by an internal uprising? If the regime was likely to remain in power, would it have continued or even accelerated its systematic oppression, torture, and murder of the Iraqi people? Would the regime have attempted to acquire weapons of mass destruction, and, if so, what steps would the United States and its allies have taken to thwart such efforts? Answering these counterfactual questions is essential if we are to place a value on the relevant goods achieved by fighting the war.

While Kavka recognized the vital role of counterfactual considerations, he also observed that "it is not obvious which counterfactual situation we are to consider in making our comparison." (6) In what follows, I will argue that a certain counterfactual approach actually does stand out.

At first glance, the appropriate counterfactual may seem straightforward--leaders should compare a policy of going to war with whatever policy they would enact if they do not go to war. This "what they would otherwise have done" baseline is not tenable, however.

Imagine, for example, that a junta is grievously and systematically violating the human rights of its people to an extent that represents a sufficient just cause. Another state's leaders are deciding whether to intervene militarily--an intervention that will require the overthrow of the offending junta. Also imagine that the potentially intervening state is heavily dependent on the unjust junta for oil, and that they lack moral scruples. The leaders contemplating intervention recognize that if they do not intervene, they will--in order to help ensure that the junta is willing and able to continue supplying them with oil--actually help prop up the junta and facilitate the junta's systematic oppression of their people. In such a situation, surely the avoidance of the extra oppression and injustice the leaders contemplating intervention would otherwise facilitate should not potentially be a deciding factor in the proportionality test. Generalizing, we can conclude that the "what they would otherwise have done" proposal is implausibly dependent on the whim and character of state leaders. It entails that a state with leaders who are willing to act immorally to further the injustice that grounds the war's sufficient just cause have a better chance of passing the proportionality test. Something has obviously gone wrong when one's willingness to act immorally may sometimes be a determining factor in morally justifying one's resort to war.

This discussion suggests that the relevant synchronic counterfactual baseline must, unlike the "what they would otherwise have done" proposal, contain a moral qualifier. As a first suggestion in this direction, let me propose the following.

The Morally Qualified Counterfactual Baseline: The appropriate counterfactual baseline is the nonwar course of action that the leaders would undertake if they did not engage in actions that immorally furthered the harms or injustices that underlie the sufficient just cause.

Acting on this morally qualified baseline, a state would both forgo any future activity that immorally furthers the injustices underlying the just cause and stop any current actions that immorally further those injustices. As an example of how this baseline would operate in practice, consider the following. In the case of the Iraq war, many argue that the prewar implementation of sanctions was an immoral contributor to the oppression and suffering inflicted upon the Iraqi people by Saddam's regime. Furthermore, some, including myself, contend that it was the humanitarian injustices that Saddam's regime was inflicting in Iraq and not the threats posed by terrorism and weapons of mass destruction that provided the grounds for a sufficient just cause. Assuming for argument's sake that these two claims are correct, then the United States' contribution to sustaining the sanctions immorally furthered the injustices underlying the sufficient just cause. Thus, on the morally qualified baseline proposal, the proportionality calculation would compare the resort to war with a nonwar counterfactual in which the leaders of the United States (to the best of their abilities) stopped participating in the sanctions.

In developing a morally qualified counterfactual baseline, the focus so far has been on acts that immorally further the injustices underlying the sufficient just cause. There is, however, a further possible concern--namely, how should we deal with the possibility that a state will forgo war and yet undertake nonwar actions that go some ways toward achieving the sufficient just causes? An example based on the 1991 Gulf War well illustrates this concern.

One of the reasons given for going to war against Iraq in 1991--a reason that many just war theorists accept as a relevant good in proportionality calculations--was to deter Iraq from future aggression against Saudi Arabia. If the leaders of the United States had chosen not to go to war, however, it is very plausible that, as an alternative, they would have worked to greatly strengthen the forces aligned against Iraq, especially in neighboring states like Saudi Arabia. This increase in defensive forces would thus be part of the baseline counterfactual case, and it would go some ways in offsetting the deterrence value of going to war, since some deterrence effect would be present in both the baseline and war scenarios. Thus, under the proposed counterfactual baseline, certain likely future choices of the leaders under the nonwar counterfactual could actually be a determining factor in deciding the proportionality of the resort to war, and would thus partly determine whether the resort to war is moral.

I suspect, however, that many people will find it implausible that the proportionality of resorting to war may depend on the willingness of a state's leaders to, short of war, devote resources or make sacrifices that would help achieve the sufficient just causes. (7) To put this in slightly different words, it is counterintuitive that in certain cases it will be moral to resort to war only if a state's leaders are, on behalf of their people, relatively unwilling to otherwise make certain virtuous sacrifices.

I must admit to presently being agnostic on this issue. It may just be that those who are willing to take one virtuous and supererogatory course of action close off, as an ethical matter, other avenues of action--avenues that remain open to the less virtuous. If the consequence just described is implausible, however, this would suggest the need for a further amendment to the baseline counterfactual proposal.

The Morally Qualified Counterfactual Baseline (Version Two): The appropriate counterfactual baseline is that course of action the leaders would undertake if they neither engaged in action that immorally furthered the harms or injustices that underlie the sufficient just causes nor undertook action that was directed at achieving the sufficient just causes.

With the baseline account so amended, the proper counterfactual baseline for the 1991 war in Iraq would be one in which the United States does not take steps to help Saudi Arabia build up its defensive position in order to deter Iraq from further aggression.

It is worth mentioning that this second version of the baseline does set up an attractively straightforward contrast between the role played by the proportionality criterion and the last resort criterion. For if, as the second version requires, the proportionality criterion compares the possible war with a nonwar baseline in which no action is directed at achieving the sufficient just causes, then, by way of contrast, the last resort criterion can be viewed as requiring that the possible war be compared with a (suitably qualified) range of non-war "baselines" that involve some action that is directed at achieving the sufficient just causes to some degree. (8) Despite this attractive point in favor of version two of the baseline, I remain agnostic with respect to the two alternatives. In what follows I will support my proposal in that light.


Despite its apparent merits, many have argued that there are good reasons to reject the kind of counterfactual proposal I have presented.

The Need for Counterfactuals In their very insightful article, "The Just War and the Gulf War," Jeff McMahan and Robert McKim contend that proportionality calculations do not need to rely on counterfactual considerations. They provide the following example:
 Suppose ... that a child is in the path of a speeding truck. A
 passerby snatches the child out of the way. It is not necessary
 to know what would have happened had the passerby not acted in
 order to know that she saved the child. Indeed, suppose that
 someone else would have snatched the child out of the way if the
 passerby had not. It remains true that the passerby saved the
 child--that her act prevented the death--even though it is not
 true that the child would have been killed if she had not acted,
 since in that case another cause would have operated to produce
 the same effect. (9)

This example fails to capture the complicated nature of military conflict. In particular, most military acts not only achieve good, but also cause harm to civilians and soldiers. Thus, a more appropriate sort of example would be one in which the passerby does not merely save the child but also, in the process of doing so, is forced to harm an innocent bystander--for example, by being forced to push a frail old lady to the ground in order to reach the child, when doing so may well break the lady's hip. In such a case, surely the passerby (if practically possible) is obligated to consider, by way of comparison, the harms that will occur if she refrains from saving the child. Thus, if the passerby sees that another person is preparing to rescue the child and is unlikely to injure the child or other bystanders, then it is wrong for the passerby to act. Clearly, despite McMahan and McKim's claim, a counterfactual comparison is needed to determine the morality of the action.

Relevant Effects

Besides general objections to the use of counterfactuals, there are more specific concerns that may be raised regarding my proposal. One concern raises the issue of what good and bad effects should be included in the baseline. Consider the following example. (10) Imagine two leaders, A and B, who are considering intervening militarily in a third state. If leader A does not intervene, she will spend the money the intervention would have cost on income-support programs for the poor, which will do a lot of good, whereas if leader B does not intervene, he will spend the money on tax breaks for the rich, which will do much less good.

If we assume that the relevant harms in proportionality calculations include the financial costs of fighting the war, and that those costs are "valued" by considering the good forgone by incurring the cost, then under my baseline proposal it seems to be the case that the more moral leader, leader A, is less likely to pass the proportionality test than the less moral leader, leader B. Such an outcome appears to run counter to the earlier concern raised in presenting the second version of my baseline proposal--namely, that in developing a counterfactual baseline we should avoid enabling less virtuous leaders to have a better chance of satisfying the proportionality criterion.

In responding to this objection, let me begin by observing that it crucially relies on the view (which I will shortly dispute) that a relevant harm in the proportionality calculation includes the full financial cost of fighting a war. This observation suggests a much broader problem in defending and testing my baseline proposal--namely, that the proposal's plausibility may well depend on what count as the relevant good and bad effects in proportionality calculations. A full discussion of the type and range of effects that are relevant would take us beyond the scope of this essay, but I will nonetheless argue that neither the relevant goods nor the relevant evils are open-ended. (11)

First, with respect to the relevant goods, there is reason to believe they should be limited to the value of achieving the sufficient just causes for the war. McMahan supports this claim in "Just Cause for War." To follow McMahan's argument, one must first recognize that the sufficient just cause criterion not only plays the obvious role of identifying the types of conditions that must be present for a resort to war to be justified, but in doing so is identifying the only types (range) of goods that may be permissibly pursued in fighting the war. McMahan provides a succinct argument for limiting the relevant good effects to the achievement of the sufficient just causes:
 If just cause indicates the range of goods that may permissibly be
 pursued by war, then no goods that fail to come within the scope
 of the just cause, or are instrumental to achieving it, can count
 in the proportionality calculation. If they did, that would imply
 that a war is justified, at least in part, by the fact that it
 would achieve certain goods that cannot permissibly be achieved
 by means of war. (12)

With respect to evils, unfortunately, there doesn't seem to be an equally straightforward way of identifying and qualifying those that are relevant. Instead, there seem to be a variety of principled and seemingly unconnected considerations that will determine whether an evil in a war or even the postwar period counts as a relevant consequence. (13) Still, the following example suggests that the relevant evil consequences are restricted (rather than being open-ended), and restricted in one way that speaks to the earlier objection involving leaders A and B.

Imagine that a generous volunteer is taking initial steps to establish a program that would benefit poor children in his community. Before he can complete the process, a friend calls him to see if he would, instead, be willing to help establish a program for land mine victims in Cambodia. Our volunteer would prefer to work with his friend on the Cambodian program. He recognizes, however, that he only has the time to develop one of the programs, and thus that if he chooses to help his friend he will have to forgo developing the program for the local community. Furthermore, let us assume that the program in the local community would actually generate much more good than the program in Cambodia. In making the decision whether or not to pursue the Cambodian program, would it be plausible to say that a choice to pursue that program would be wrong because the good lost by forgoing the development of the local program outweighs the good the Cambodian program would achieve? Surely the answer is no. The volunteer may want to consider the loss of the local program in making his decision, and may even offer it as a relevant explanation to his friend if he declined his friend's offer, but the loss does not make it wrong (that is, morally impermissible) to choose his friend's offer over the local program.

The reason why the forgone loss to the local community does not make it wrong to pursue the Cambodian program lies, at least in part, in the fact that the devotion of time and energy to help the local children would not be an obligatory act, but rather would be a supererogatory act--a meritorious act that the volunteer is permitted to forgo if he so chooses. While it does seem plausible that certain losses resulting from forgoing one's (prima facie) duties often should count in determining the rightness or the wrongness of an act (more on this momentarily), the same is not true of losses that are the result of forgoing virtuous acts that are merely permissions and not duties.

The proportionality calculation is one of a number of criteria employed collectively to determine whether the resort to war is wrong--in other words, to determine if the resort to war is morally impermissible. In light of this observation, and combined with the preceding observation--namely, that general losses that are the result of forgoing supererogatory acts should not be a determining factor in deciding whether an act is impermissible--it is clear that the relevant goods and harms in the proportionality calculation should not include those general losses. That is, general losses from forgoing an alternative course of action should not be included in proportionality calculations unless that alternative course of action is one to which one is in some sense duty-bound (again, more on this shortly). (14)

Consider again the example of leader A, who, if she does not go to war, would use the resources to assist the poor in her country, and leader B, who, in contrast, would use the resources to provide tax breaks to the rich. Notice that this example appeared to cause problems for my counterfactual proposal on the assumption that proportionality calculations include the intervening state's financial cost of fighting the war. The preceding discussion, however, suggests that those costs--the redirection of resources, with associated losses--should only count in the proportionality calculation if that redirection of resources would be a violation of the leader's duties. The key question is therefore whether leader A has a duty to implement those programs for the poor or has instead a permission to use these public resources in alternative ways. I contend that the commonly held view is that a leader, assuming she is duly elected or appointed, and is acting within the scope of her role, generally has the permission to use (on behalf of her fellow citizens) public resources in a wide range of ways. Spending decisions by political leaders are often rightly criticized, even on moral grounds, but it is generally accepted that a leader has broad discretion in spending public resources, and, in particular, does not generally violate her duties if she spends those resources in ways that do not maximize the benefit for the poor. If this is right, then neither leader A nor leader B is acting impermissibly by reallocating public resources to the war, and so the forgone goods--whether they be goods that help the poor or give tax breaks to the rich--are not relevant in the proportionality calculation. And, hence, leader A's more virtuous use of the resources under the nonwar counterfactual does not affect the likelihood of her meeting the proportionality test. In other words, the counterfactual baseline proposal I have offered does not provide the less moral leader with a better chance of satisfying the proportionality criterion.

(Even though I've argued that losses resulting from forgoing nonobligatory acts should not generally count in proportionality calculations, it still may be the case that when those nonobligatory actions involve very good consequences, they provide a leader with relevant reasons for choosing not to go to war. Thus, if leader A decides not to go to war, her moral case for the appropriateness of that decision may well be stronger if not going to war enables the leader to use those resources to provide domestic or international programs for the poor.)

Having made the case that leaders are generally permitted to allocate financial public resources across a range of programs and activities, including war, it is important to recognize that this may not always be the case. Sometimes a reallocation of resources might involve a failure to fulfill one's duties and, hence, affect the proportionality calculation.

In the aftermath of the recent war in Afghanistan, for example, many, including myself, would contend that the United States has a postwar (post bellum) duty to provide the people with security; help them rebuild their infrastructure; and help them to implement democratic institutions and the rule of law. If this is right, then if the leaders of the United States also had limited access to military resources and failed to carry out their post bellum duties in Afghanistan in order to use those military resources to fight another nonobligatory war, like the war in Iraq, then the losses experienced by the Afghan people due to lessened security would count in the proportionality calculation for the Iraq war. Furthermore, those losses should count under the war scenario regardless of whether, counterfactually, the leaders of the United States would have used those military resources in Afghanistan. This is because the leaders of the United States were duty-bound to do so, and so the losses should count regardless of whether they would, counterfactually, act on that duty.

It is important to note, however, that this example is premised on the leaders of the United States having limited military resources that forced them to choose between providing security in Afghanistan or fighting the war in Iraq. If, as was arguably the case, the United States had sufficient military and financial resources to both fight in Iraq and provide adequate security and support in Afghanistan, but chose instead to reduce or minimize resources in Afghanistan as they entered the Iraq war, then the corresponding losses incurred by the Afghan people due to inadequate security should not count as a relevant harm in the proportionality calculation for the Iraq war. Even if it is the case that in the non-Iraq war counterfactual the United States would provide more security in Afghanistan, the actual choice to not provide that added security in Afghanistan while fighting the Iraq war is not a forced result of the leaders deciding to fight the war in Iraq. Instead, the reduction in military resources in Afghanistan is the result of an additional, separate, and optional decision not to use additional available military resources to provide the morally required security in Afghanistan. In other words, if sufficient resources were available, then the decision to use military resources in Iraq is not "necessarily tied" to the decision to not provide adequate security in Afghanistan, and, therefore, the associated losses and harms experienced by the Afghan people should not count in the proportionality calculation for the Iraq war.

So far I have only discussed the allocation of financial resources in relation to questions of duty. When we turn to human resources, however--whether they be soldiers or civilians, allies or enemies--there are clearly duties present that require leaders considering a resort to war to not put those individuals at risk of serious harm unless the cause is sufficient. A precise analysis of those duties, including both their general and special or partial dimensions, is a complicated matter that I will not pursue here. It is clear, however, that since leaders have these duties, the expected harm that individuals would face in war is a relevant bad effect that must be included in proportionality calculations. (The harms I have in mind here are not only the harms of personal injury and death, but also serious harms to one's community, one's way of life, and associated infrastructure and institutions.) Of course, determining the value to attach to various lives and to the serious disruptions and damage a war causes in communities is also a complicated matter. (15) A clear principled line can and should be drawn, however, between a leader placing people in harm's way (in a war) and a leader choosing to allocate financial resources in a certain way. The former but not the latter should be included in proportionality calculations as a relevant harm.

There is a one further complication facing my proposal. It becomes apparent when one recognizes that there is likely to be a range of nonwar counterfactual actions that the leaders considering war may take and that would neither immorally further the harms or injustices that underlie the sufficient just causes nor be directed at achieving the sufficient just causes. If none of these alternative courses of action affected goods or evils that were relevant in the proportionality calculation, this would not be a concern. Given that my baseline proposal only requires that the counterfactual not immorally further the harms or injustices that underlie the sufficient just causes, however, this implicitly suggests the possibility that some of the counterfactual scenarios could permissibly (morally) further those harms and injustices--harms and injustices that clearly are relevant in the proportionality calculation.

As a hypothetical example of a case involving a permissible furthering of relevant harms and injustices, imagine that leaders in the United States are considering intervening in Iraq on the (assumed) grounds that there is a sufficient humanitarian just cause. Next, imagine that if the United States chooses to forgo war with Iraq, there is some possibility that the leaders of the United States will legitimately undertake some arms sales to Turkey. If they were to do so, however, the leaders of the United States also foresee that in making such a sale, it is likely that some of the arms will illegitimately make their way into Iraq, thereby furthering Saddam Hussein's ability to oppress the Iraqi people. Here we seemingly have a counterfactual case where the leaders of the United States are not immorally furthering the relevant injustices (on the assumption that such foreseen but not intended consequences are sometimes not immoral), but nonetheless are indirectly exacerbating those injustices.

Let us assume then that it is possible for the leaders of a state to permissibly further the relevant injustices. Furthermore, let us recognize that there often are a variety of nonwar counterfactual scenarios a state's leaders would have some likelihood of undertaking--some of which may well (permissibly) further the relevant injustices and some that may not. The result of these two possibilities is that in a given case we may be faced with a range of possible counterfactual baselines that produce different proportionality outcomes. What to do? There is probably no uncontroversial answer here. (16) The approach that perhaps best captures the spirit of the proportionality criterion, however, is to simply do a proportionality calculation for each possible counterfactual nonwar course of action and then amalgamate the results based on the likelihood of each course of action being taken. Thus, building on the example just considered, if there was some chance that the leaders of the United States would sell weapons to Turkey, and some chance that they would not, then one would make a rough judgment as to the likelihood of each counterfactual scenario, do a proportionality calculation for each scenario, and then amalgamate those calculations based on the likelihood of each scenario. Obviously any such probability calculations would be very rough and uncertain. This roughness and uncertainty, however, should not in itself dissuade one from taking this approach. Similar roughness and uncertainty is already present in proportionality calculations, since those calculations require us to evaluate the likelihood of various outcomes of resorting to war--a task that is obviously full of uncertainty in the best of times. And more generally, as moral agents with limited knowledge, we are often forced to make important moral decisions in the face of uncertainty regarding what effects our actions will have and what future courses of action other moral agents (including our future selves) will undertake. Uncertainty and the need to make rough judgments is an unfortunate moral reality that extends far beyond the realm of proportionality calculations.


In closing this essay I will highlight two examples that show the ongoing relevance of the proportionality calculation and the importance of understanding it in the counterfactual terms I have suggested. The first example is based on intense ongoing discussions about the circumstances in which humanitarian military interventions are justified. Examples such as Rwanda, Kosovo, Myanmar, Darfur, and so on provide a strong case for there being humanitarian just causes. It is therefore all the more important to consider the proportionality of proposed military interventions in such cases. We want to make the situation better, not worse. This essay has shown that, in not wanting to "make things worse," we must not simply contrast the state of affairs prior to a possible military intervention with the likely state of affairs after such an intervention. Instead, we should compare the state of affairs that would come to pass if nothing is done with the state of affairs that would likely result from a humanitarian intervention.

As a second example, I return to the conflict mentioned at the start of this essay, the current war between Israel and Hezbollah. Given the widely held assumption that Israel has a just cause in this conflict, the importance of determining whether its response is proportionate (as well as meeting other conditions, such as last resort, minimal force, and so on) is evident. Such proportionality considerations are not as simple as many contend, however. One must not simply weigh the injustice and harm done to Israeli soldiers prior to the war, in the initial Hezbollah raid, against the casualties and destruction that have occurred in the war. Nor is it appropriate to simply compare the state of the people in both Israel and Lebanon prior to the war to their state both in and after the current conflict. As shown above, such simple diachronic comparisons are fundamentally misguided. We must instead ask the more difficult question of what state of affairs would have obtained had the war not taken place--what relevant injustices and wrongs would have gone unpunished (perhaps thereby emboldening and encouraging other terrorist groups and their state supporters); to what extent would the military threat posed by Hezbollah have increased even further; what likely injustices and harms would Hezbollah have inflicted on Israel in the future; and what would all these considerations suggest about the likelihood of a future war in Lebanon, more horrific even than the current one? It is only by first answering these tough counterfactual questions that we can provide a baseline against which to compare the effects of the war, both in terms of the sufficient just causes the war may achieve and the horrific casualties and destructions it inflicts. With respect to this conflict I do not think the proportionality question is easy to answer. It is, however, important that in our discussion we ask and attempt to answer the right questions.

David Mellow *

* I am very grateful to Thomas Hurka, Christian Barry, Dennis McKerlie, and the journal's anonymous reviewers, all of whom took time to comment in detail on earlier drafts of this essay. My thanks also to Jeff McMahan and Alex Sager for discussions I have had with them on this and related topics.

(1) See Editorial, Globe and Mail, July 20, 2006, p. A12.

(2) See Jeff McMahan and Robert McKim, "The Just War and the Gulf War," Canadian Journal of Philosophy 23, no. 4 (1993), p. 507. In requiring that the relevant goods be proportional to the relevant harms, this leaves open the question of whether the proportionality criterion simply requires that the relevant goods on balance outweigh the relevant harms, or, instead, makes some other proportionate demand. For example, perhaps the relevant goods have to outweigh the relevant harms to a large extent, or, alternatively, perhaps the proportionality requirement could be satisfied even if the relevant goods are slightly outweighed by the relevant harms. Just war theorists are not in agreement on this issue, and it can be bracketed for the sake of the present discussion.

(3) In making a similar point, Thomas Hurka observes, "In the Afghan case the relevant U.S. number is not that of civilians killed on September n; their lives were already lost" ("Proportionality in the Morality of War," Philosophy and Public Affairs 33 [2005], p. 59).

(4) See Aryeh Neier, "Inconvenient Facts," Dissent 47, no. 2 (2000), pp. 109-12. Neier argues for the same basis of comparison as suggested in the main text in his telling criticism of Noam Chomsky's argument against NATO's actions in Kosovo.

(5) Gregory Kavka, "Was the Gulf War a Just War?" Journal of Social Philosophy 22, no. 1 (1991), p. 24.

(6) Ibid.

(7) An anonymous reviewer raised this concern in response to an earlier draft of this essay.

(8) See Hurka, "Proportionality in the Morality of War," pp. 37-38, for a concise and clear discussion of the connection between the proportionality criterion and last resort criterion.

(9) McMahan and McKim, "The Just War and the Gulf War," p. 509.

(10) An anonymous reviewer offered this example in response to an earlier draft of the essay.

(11) The best discussions of type and range issues with respect to relevant effects are found in Jeff McMahan and Robert McKim, "The Just War and the Gulf War"; Jeff McMahan, "Just Cause for War," Ethics and International Affairs 19, no. 3 (2005); and Thomas Hurka, "Proportionality in the Morality of War."

(12) McMahan, "Just Cause for War," p. 5.

(13) A point I develop in David Mellow, "A Critique of Just War Theory" (Ph.D. dissertation, University of Calgary, 2003).

(14) Thus, the proportionality condition is not a simple utilitarian calculation undertaken to determine whether the resort to war has the best overall consequences.

(15) For more on this see Hurka, "Proportionality in the Morality of War."

(16) Compare the extensive debate over John Rawls's "maximin" solution to the problem of social justice.
COPYRIGHT 2006 Carnegie Council on Ethics and International Affairs
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Mellow, David
Publication:Ethics & International Affairs
Geographic Code:7ISRA
Date:Dec 1, 2006
Previous Article:The legitimacy of global governance institutions.
Next Article:Killing soldiers.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters