During his campaign, Donald Trump publicly endorsed torture, indiscriminate bombing of populated areas, and killing civilians with familial ties to members of ISIS—all of which are illegal under international and domestic law. In a recent interview with the New York Times, however, the President-Elect implied that he might reconsider his position at least on torture. After learning that his nominee for Secretary of Defense, General Mattis, opposes torture because he “never found it to be useful,” Trump told the Times that he was “impressed” by this response. Trump said that Mattis did not necessarily change his mind, but that his decision would depend on whether “Americans feel strongly about bringing back waterboarding and other tactics,” and if so, “I would be guided by that.”  If Trump will be listening to the winds of popular opinion on the specific issue of torture, it is likely that he will keep an ear to the ground on other international norms related to the use of force. Whether or not the American public wants to “take the gloves off” may influence how the U.S. fights and obtains intelligence in the coming years.

Against this backdrop, the International Committee of the Red Cross’s (ICRC) major survey on the laws of war is especially timely. Between June and September 2016, the ICRC asked 17,000 people in 16 countries – Permanent 5 (P-5) countries (in addition to Switzerland) and numerous war-torn countries – about their attitudes toward a range of issues related to the laws of war.  The survey results, which are available on the ICRC’s website, provide insight into American public attitudes toward foreign civilian mistreatmment and U.S. compliance with international law and how American attitudes compare with those in other countries.

In a post for Just Security, Ryan Goodman paints a pessimistic picture of the ICRC’s findings: Americans see a range of military attacks on civilians and civilian objects as “part of war” and are less inclined to consider these attacks as “wrong” than individuals in most of the other countries in the sample, in particular those living in war-torn countries. Moreover, Americans, along with people who live in other countries removed from war, do not find the Geneva Conventions to be terribly effective compared to those living in war-torn countries. The results also indicate that Americans have a fairly elastic attitude about torture, with 54% of Americans saying it is wrong in general and only 30% if that individual is an enemy combatant.

Goodman’s summary offers a helpful overview of the data. He focuses the bulk of his analysis on the differences in attitudes between people living in war-torn countries whose exposure to conflict leads them to have a greater appreciation for the efficacy of the laws of war than people living outside conflict zones for whom violations of humanitarian principles have become normalized over time. We agree with Goodman that the results for American attitudes are quite concerning when compared to opinion in war-torn and war-free countries. However, a closer look points to peculiarities and particularities of the data that may temper some of the pessimism about American attitudes toward civilian mistreatment in war.

First, as is always the case with surveys, it is important to scrutinize the question wording to see how it may have affected how individuals responded to questions. Curiously, in many of the questions about military attacks that could inflict harm to civilians, respondents were given four possible responses:  “wrong,” “just part of war,” “I don’t know” and “prefer not to answer.” The report suggests that individuals in P-5 countries were less “supportive and respectful of humanitarian norms” because they responded with “just part of war” at higher rates.

It would be a mistake, however, to draw the inference that “just part of war” is synonymous with not being “supportive and respectful of humanitarian norms.” While the report makes it seem obvious that “wrong” and “just part of war” are opposites, individuals might not actually see it that way at all. If given the option, many respondents might actually have been tempted to respond “both” or “unfortunately” just before “just part of war.” It would not be surprising if individuals in countries such as the U.S. look at actions taking place in Syria and agree that particular attacks against civilians are wrong, but nonetheless have become part of war, even if they do not actually support particular policies that have led to those outcomes. Given the semantic ambiguity over “just part of war,” the percentage of respondents answering “wrong” is arguably a more meaningful indicator (Goodman also relies on the answer of “wrong” in his analysis).

Second, taking the rates of “wrong” responses, the variation across questions demonstrate some measure of American public sensitivity to ethical considerations reflected in humanitarian principles governing the conduct of hostilities. Whereas 76% of Americans think it is wrong to attack hospitals, ambulances and health care workers to weaken the enemy, 57% said it was wrong to deprive civilian populations of food to weaken the enemy, and 36% said it was wrong to attack “enemy combatants in populated villages or towns in order to weaken the enemy, knowing that many civilians would be killed.” The recent  political science literature debating whether citizens at home care about the treatment of citizens abroad in wartime has been divided into two main camps: those who point to apathy and those who point to empathy.

The ICRC’s observational data suggests a possible third camp, in which individual attitudes are mediated strongly by ethically-salient contextual factors causing them to act like what moral psychologist, John Mikhail, calls  “intuitive lawyers.” That is, individuals implicitly recognize how different human actions comport with or violate basic moral and legal categories, such as ends, means, and side effects. Mikhail has found that individuals intuitively rank a) intentional harm to be morally worse than foreseeable harm (ceteris paribus), b) using harm as a means to be morally worse than harm being a side-effect (ceteris paribus), and c) actions that result in harm to innocents to be morally worse than omissions that result in equivalent harm (“omission bias”).

Aspects of the ICRC data suggest that these intuitions might be operating in the context of public opinion and the treatment of civilians in war. Indeed, moral opposition in all countries—P5 + Switzerland and the war-torn countries in the sample—is highest for the question about attacking hospitals, but drops respectively for the subsequent two questions about depriving civilians of food and attacking combatants in civilian-populated areas in order to weaken the enemy. On the surface, this general pattern appears consistent with moral intuitions based on three ethical distinctions: a) intentional vs. foreseeable harm to innocents, b) harm as a means to an end vs. harm as a anticipated side-effect of an action, and c) acts vs. omissions.

The question related to attacking hospitals and medical workers consisted of intentional attacks against civilians, whereas harm to civilians caused by targeting combatants in heavily populated areas falls under the foreseeable but unintentional category. Moreover, the latter  does not actually consist of an IHL violation since it is legal to carry out armed attacks against combatants inside populated areas as long as it does not inflict disproportionate harm to civilians). Thus, it is not surprising that a larger percentage of individuals considered the former to be “wrong” than the latter. It is also not surprising that a larger percentage of individuals considered depriving civilians of food to be “wrong” than foreseeable civilian harm through attacking combatants. In the former, the deprivation of food (e.g. withholding food aid) is a means through which the enemy is weakened, whereas the foreseeable civilian harm is a side effect of targeting enemies.

Why more individuals considered intentional attacks against hospitals and health care workers “wrong” than depriving civilians of food is more difficult to explain. To the extent that individuals interpreted “depriving” as withholding food, omission bias may explain the variation in judgment. Refraining from giving food constitutes an omission, or failure to action, and therefore less morally problematic than actively killing civilians (a commission). (If individuals interpreted “depriving” as removing or destroying food supplies, then omission bias would not apply by definition). It is also possible that individuals interpreted the intentions behind food deprivation differently. If many did not see it as intentional civilian harm, but instead the foreseeable civilian harm caused by depriving the enemy of life sustenance, then we would expect less moral opposition than intentional attacks on health-care workers and facilities.

Crucially, these ethical distinctions help explain only the pattern of overall attitudes across the questions. They do not account for variation among countries. For example, despite U.S. respondents’ sensitivity to these distinctions, Americans were generally more  tolerant of civilian harm for all three questions. This finding is especially concerning since the U.S. military has been by far the most active around the world during the postwar period. To the extent that public attitudes do influence policy, Americans’ tolerance for civilian harm could cause its military to be more cavalier. Nevertheless, the ICRC data do show that Americans overwhelmingly oppose intentional civilian killing (only 9% believe that it is appropriate to attack combatants and civilians) and generally believe that civilian killing should be avoided as much as possible (78%). It seems the public is not on board with the indiscriminate killing and targeting civilians with familial ties to suspected terrorists that Trump endorsed on the campaign trail.

Third, on the specific question of torture, Americans’ opposition appears to be fairly soft, although not the strong support that Trump suggested would offer the grounds for bringing back waterboarding. Over half (54%) consider torture to be “wrong.” However, when asked whether “a captured enemy combatant [can] be tortured to obtain important military information,” opposition to torture weakened, with only 30% saying no and 46% yes. The two main upshots are that those in favor of torturing enemies for intelligence still constitute less than a majority of Americans and, importantly, within this group 39% said their opinion changed after being informed that torture is illegal. It should also be noted that the premise of the question—that torture “would obtain important military information”—runs counter to what Trump’s own nominee for Defense Secretary has asserted, not to mention a US Senate report on interrogation, which found that enhanced techniques produced either no intelligence or fabricated, faulty information.

Lastly, attitudes about the effectiveness of the Geneva Conventions appear to have soured in the last fifteen years. In 1999, 52% of respondents agreed that the Geneva Conventions had prevented wars from getting worse. By 2016, that number was just 38%. Upon first glance, the decline looks alarming, but the ambiguity of this question again raises some concerns in terms of interpretation. What does it mean to “prevent wars from getting worse”?  While the ICRC likely thought of the question as probing whether people think that the Geneva Conventions are preventing IHL violations, it might not be surprising if respondents were simply responding to aggregate levels of conflict rather than jus in bello rules with which the ICRC concerns itself. We don’t have to scour the planet long to see why the public, and in particular Americans, would look at ongoing conflict in Syria, Afghanistan, Iraq, and Yemen and think the Geneva Conventions have less clout than they did more than a decade and a half ago.

On the other hand, the Geneva Conventions were never supposed to prevent the outbreak of war or the intensity of a war but rather regulate the treatment of civilians once in war.  That there is seemingly ubiquitous war compared to 1999 does not mean that the Geneva Conventions are not working. The distinction suggests that there are pockets of progress when it comes to the laws of war. Just last week, the Pentagon released a 230-page document that in part detailed the ways the laws of war inform American targeting decisions. The standards for protecting civilians articulated in this document certainly  depart from those guiding U.S. “terror bombing” during World War II. To be sure, contemporary targeting standards fall short of protecting civilians to the degree that many international lawyers and human rights activists hope to see.

In short, there appears reason to hope that any attempts by the President-Elect to lower standards on how the U.S. uses force abroad and treats foreign civilians will run up against a skeptical American public—both popular sentiment and organized societal sectors—not to mention opposition from some actors within the military and foreign policy establishment unwilling to carry out those policies.