On May 7, 1999, a US aircraft flying under a NATO mission bombed the Chinese embassy in Belgrade, killing three people and injuring dozens more. The incident sparked condemnation from Beijing and anti-American protests among the Chinese public. Although Washington and Beijing officials agreed on the material facts of the bombing, they disagreed about the motives behind it. According to the US and NATO, “there was no intent to harm civilians.”Footnote 1 However, Chinese officials, and even some Westerners, were skeptical. One British newspaper, for example, stated that “NATO bombed Chinese deliberately.”Footnote 2 Lamenting these responses, NATO spokesman Jamie Shea observed that “when something bad happens everybody thinks there has to be a secret reason – not a cock-up, but a conspiracy.”Footnote 3 What, then, makes people perceive wartime actions as intentional?
Drawing on research in moral psychology, we argue that moral intuitions shape attributions of intent: people are more likely to ascribe intentionality to wartime actions with negative consequences compared to otherwise equivalent actions with positive consequences. This asymmetry in intention attributions has been dubbed the “side effect” effect or “Knobe effect” (KE) (Knobe Reference Knobe2003). We illustrate these dynamics through two survey experiments (S1/S2). Beyond the main effect, the results show that the severity of the consequence does not substantially affect intentionality ascriptions and that the KE is consistent across demographic subgroups.
Theory and Empirics
Many people believe that intended harms are a deeper moral problem than incidental harms (Traven Reference Traven2015). This view is not only reflected in just war doctrine but also forms a key part of International Humanitarian Law. For example, in his Summa Theologica, St. Thomas Aquinas famously argued that killing in self-defense is morally permissible only if one only intends the good effect of self-defense, and if the bad effect of killing is necessary and proportionate to the good effect of self-defense (Aquinas Reference Aquinas2002, 170). However, recent research in cognitive science and moral psychology shows that the following is also true: moral intuitions about an action’s consequences shape whether people believe the action was intentional. Joshua Knobe famously demonstrated that intentionality judgments are subject to significant biases: when someone causes an outcome that produces a negative side effect, observers are more likely to believe that the agent acted intentionally than they are when the side effect is positive. In the canonical experiment, echoing research on framing from behavioral economics (Kahneman and Tversky Reference Kahneman and Tversky1979), Knobe presented subjects with the following vignette:
The vice-president of a company went to the chairman of the board and asked, “We are thinking of starting a new program. It will help us increase profits, but it will also harm the environment.” The chairman of the board answered: “I don’t care at all about harming the environment, I just want to make as much profit as I can. Let’s start the new program.” They started the new program. Sure enough, the environment was harmed (emphasis added, Knobe Reference Knobe2003, 191).
When replacing “help” with “harm,” people were more likely to believe that the vice-president acted intentionally (2003, 192). Not only has the KE been well-replicated but research also shows that it may hold across cultures, and thus may reflect an evolved bias (Knobe and Burra Reference Knobe and Burra2006).Footnote 4
Advancing this literature, we hypothesize that wartime actions that cause negative, as opposed to positive, consequences are more likely to be perceived as intentional. In applying this literature to the domain of wartime conduct, our surveys make three innovations: first, we examine whether the severity of an outcome changes how observers ascribe intentionality; second, we examine whether the KE applies across issue areas in International Humanitarian Law (IHL) (civilian immunity/cultural sites); and, finally, whether it varies across demographic subgroups.
In S1, respondents read a vignette in which a US officer ordered an operation to destroy a weapons facility. In addition to destroying the facility, the operation caused a side effect that was randomly assigned. Treatment 1 randomized whether the operation affected civilians or UNESCO cultural sites. Treatment 2 randomized whether the side effect was positive or negative. This created four possible treatment groups: (1) civilians killed; (2) civilians saved; (3) UNESCO site destroyed; and (4) UNESCO site saved. S1 was administered at an American university to 328 students from October to December 2018. S2 was administered to 783 American adults through MTurk in April 2019. Since S2 both replicates and expands upon S1, we present the results from S2.
The results in Figure 1 confirm our hypothesis: respondents ascribe greater intentionality to military operations that cause negative (as opposed to positive) effects. In the entire sample, the effect on the dependent variable, Intentional, was 1.7 out of a 6-point scale. When the 6-point scale is collapsed into a binary variable, the negative treatment caused a colossal effect: a 43-percentage point increase in Intentional. Further, despite the plausible intuition that killing more civilians (or priceless artifacts) is morally worse, the severity of the side effect has a far less substantial impact on intentionality judgments. Figure 1 also demonstrates that the KE is consistent across IHL issue areas: both civilian immunity and cultural sites.
Figure 1 Actions with Negative Consequences Cause Greater Attributions of Intentionality.
Figure 2 demonstrates that the KE exists across subgroups, with a small difference between Republicans and Democrats. The dearth of heterogenous effects may be because people’s sense of what is morally right or wrong in foreign policy does not map onto the moral intuitions measured here (cf., Kertzer et al. Reference Kertzer, Powers, Rathbun and Iyer2014). More research is required to reach definitive conclusions, but our results support the claim that the KE stems from a fundamental cognitive bias that is largely consistent across demographic groups (Chu, Holmes, and Traven Reference Chu, Holmes and Traven2020).
Figure 2 Negative Effects Increase Intentionality Attributions Across Demographic Groups.
Conclusion
Our findings have two main implications. First, though we are not the first to bring moral psychology into International Relations (IR) (Kertzer et al. Reference Kertzer, Powers, Rathbun and Iyer2014), our findings suggest that moral intuitions play an important role in how people assess wartime conduct. Recent research examines the extent to which just war principles affect public opinion (Sagan and Valentino Reference Sagan and Valentino2018). Turning just war doctrine (slightly) on its head, our results show that implicit moral reactions shape how people perceive the intent to target civilians and cultural sites.
Second, our findings have implications for work on cooperation during war. Existing work looks at how IHL promotes common understandings that facilitate the reciprocal enforcement of wartime norms (Morrow Reference Morrow2014). While assessments of intent and norm compliance are driven by numerous factors, our results suggest that cognitive biases also play an important, yet understudied role in shaping intention understanding, an issue that future work should address.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/XPS.2020.7
