An Introduction to Making Sense of Customary International Law

[Monica Hakimi is the James V. Campbell Professor of Law at Michigan Law School.]

I’m extremely grateful to Opinio Juris for hosting this symposium on my recent article on customary international law (CIL). And I’m honored by the fantastic group of contributors.

In the article, I aim to dismantle a common conception of CIL—an idea that shapes how many international lawyers think about and approach the enterprise. Like any big idea, this one has many variants. But the core premise is that a given proposition can be CIL only if it operates like a rule. Let me unpack what I mean. In using the word “rule” here, I don’t mean to draw the common distinction between rules and standards. A rule can be precise and rigid, allowing little discretion in its application (e.g.,“chemical weapons are banned”). Or it can be more flexible and fact dependent (“civilian harms must not be excessive relative to the direct and concrete military advantage”). What makes it a rule for my purposes is that its main content is both discernible and generalizable. The idea that I want to refute is that CIL consists entirely of rules—that CIL applies more or less in the same way in all cases of a given type, rather than fluctuates without established criteria from one situation to the next.

This conception of CIL—what I call the “rulebook conception”—is most pronounced in CIL orthodoxy. Take the 2018 Conclusions on Identification of Customary International Law that the International Law Commission (ILC) adopted. These Conclusions depict CIL as a body of rules. Some CIL rules are primary rules. They directly regulate behavior by defining who must do what, in what circumstances, and subject to what consequences. The ban on chemical weapons is an example. Other CIL rules are secondary. Secondary rules establish criteria for creating or changing the primary rules. In CIL, they also are how we identify the primary rules. According to the ILC Conclusions, the main secondary rule for CIL is the two-element test, under which a normative position can be CIL only if states widely support it in their: (1) practice, and (2) opinio juris. The Conclusions then identify other secondary rules that elaborate on that test. Their driving premise is that a given proposition can be CIL only if it is both intelligible and generalizable. That premise is what defines the rulebook conception. Although it is particularly evident in CIL orthodoxy, it shapes, in deep and subtle ways, how most international lawyers approach CIL. And it is not only wrong but pernicious.

My argument begins with a straightforward descriptive claim: in the everyday practice of international law, CIL works nothing like a rulebook. The normative material that global actors in the ordinary course treat and recognize as CIL does not derive from secondary rules and does not manifest only as primary rules. To be clear, I am not saying that CIL never displays rule-like levels of clarity and stability. Some of its conduct norms do function like rules. The chemical weapons ban is illustrative. But many CIL norms do not. And even when they do, it is not because they satisfy certain secondary rules.

In fact, CIL does not have any secondary rules. Although orthodox accounts insist that it does, any serious effort to identify these rules ends up exposing that they are nothing like rules. The ILC Conclusions are a case in point. Conclusion 8 posits that, in order to satisfy the two-element test, a practice “must be sufficiently widespread and representative, as well as consistent.” That sounds like a rule. And the ILC clearly intends for it to be one. The ILC asserts that its Conclusions apply across the board—“in any given case” in which CIL is at issue—in order “to ensure that a [primary] rule of customary international law is properly identified.” But digging just a little beneath the surface exposes that Conclusion 8 is not a rule. The accompanying commentary recognizes that what counts as sufficient for purposes of establishing a supporting practice “does not lend itself to exact formulations, as circumstances may vary greatly from one case to another.” In some cases (the ILC doesn’t give us any guidance for identifying which ones), the supporting practice “may have to be widely exhibited,” while in others (again, we don’t know which) it “may well be less.” What Conclusion 8 tells us is that the practice might or might not have to be “widely exhibited.” This is not a rule. Its core content is not discernible or generalizable, like a rule’s would be. It is highly malleable and contingent. And I’m not just cherry-picking the one ILC Conclusion that happens to be rudderless. The two-element test is turtles all the way down.

International lawyers to some extent know this. Global actors do not systematically use the same criteria for validating particular positions as CIL. They use different criteria. Their grounds for “finding” that a proposition is CIL are not fixed or stable but variable and contestable. They are not rules. This doesn’t mean that “anything goes” in CIL. It just means that what goes does not derive from secondary rules.

Instead, CIL’s content depends on how global actors engage with different normative positions in their many interactions. If they broadly, deeply, and consistently treat a particular position as CIL, it will operate like a rule. But because they usually do not act as a coordinated bloc, the ways in which they engage with different normative positions is often not uniform or stable. It varies. Many normative positions that are presented in the CIL process are neither collectively endorsed nor summarily dismissed. They stay in circulation for extended periods with noticeable support and real competition. These positions are not across-the-board rules. But neither do they live “outside” of CIL. They have traction as CIL in some settings. Where they do, they have the look, the feel, and the effect of CIL; they for all intents and purposes are CIL.

Consider the ICRC’s 2009 Interpretive Guidance on the Notion of Direct Participation in Hostilities. This document aimed to define the targeting norms that apply, mostly as a matter of CIL, in conflicts against nonstate groups. The ICRC’s positions on targeting are controversial and incompatible with the known views of a number of militarily active states. But because of the ICRC’s standing in IHL, its positions have been legally salient. They are not always treated as CIL, but they sometimes are, and even when they are not, they are “the main reference point for any discussion on the subject.” For example, the ICRC itself invokes its positions as the best iterations of CIL when it discusses or tries to educate people on IHL. Some national prosecutors, courts, and militaries have also treated these positions as CIL. To claim that they cannot be CIL, just because they do not operate as rules, is to obscure the various ways in which they are actually used and received as CIL in the practice of law.

So, this is my descriptive claim: the rulebook conception describes what many people imagine CIL to be, but it does not describe the material that global actors actually treat as CIL in the day-to-day practice of law. It does not reflect what CIL “is” as a real-world sociological phenomenon. I anticipate an objection. Some readers will say that the social practice that I am describing is not “really” CIL—that the “true” CIL is as the rulebook says. This objection is not responsive. I recognize that the rulebook is one way to conceive of CIL. I’m arguing that it’s a problematic conception.

This brings me to my analytic and normative claims. Because the rulebook conception doesn’t describe the actual practice of CIL, it doesn’t help us understand, advise people on, or appraise the real-world operation of CIL. And it’s not just unhelpful here. It is for two reasons insidious.

First, the rulebook conception itself generates confusion about CIL. It pushes lawyers to use analytic methods that are inapt for working with the real-life practice of CIL. Lawyers who want to engage with or advise people on this practice need to move past the question of whether a particular normative position is in all circumstances a rule. They need to focus on a broader set of questions: to what degrees, in what settings, in what ways, and for whom is the position likely to resonate as CIL? The rulebook’s methods do not assist us with these questions. CIL’s real-life content cannot be determined by applying any secondary rule (there are none to apply) or by assuming that CIL will be in one context what it was in another (it might not be). The legal analysis that lawyers must use to identify CIL’s real-life content requires more craft and expertise than the mechanical application of a bunch of rules.

Consider again the Interpretive Guidance. Trying to feed the ICRC’s positions on targeting through the two-element test to determine whether they “really” qualify as CIL would be (to put it bluntly) a colossal waste of time. No matter how one runs that analysis, it would not tell us what we need to know—which is that the ICRC’s positions sometimes have the force of CIL and other times do not. Worse, the rulebook suggests that this conclusion is somehow out of bounds. After all, to reach it, we have to do almost the opposite what the rulebook instructs: not to distill all of the practice into a fixed rule (“CIL” or “not-CIL”) but to examine the discrepancies in the practice that give the ICRC’s positions some but not all of the traction of a rule. The rulebook directs lawyers toward analytic methods that are at best a distraction and at worst misleading.

Second, the rulebook conception contributes to the enormous skepticism about CIL. Normative appraisals that assume that CIL must be like a rulebook do not even pause to consider how it might do good in the global order by not conforming to a rulebook. But as I explain in the article, CIL’s non-rule-like qualities are part of how it serves to: (1) limit bias in international law, (2) achieve normative settlements through law, and (3) advance values that are associated with the rule of law. To be clear, my claim here is not that CIL’s malleability always does every one of these things or that what it does is always, on balance, desirable. My claim is that CIL’s malleability does these things often enough that we ought to account for them when appraising or trying to reform it. The blinders of the rulebook conception cause us to lose sight of them. We have, as a result of this conception, systematically devalued CIL as a form of international law.

So, by the end of the article, I challenge adherents to the rulebook conception to defend it. Given that this conception fails to describe the real-life operation of CIL, and given that it directs lawyers toward methods that distort rather than sharpen their legal analysis, and given that it systematically discounts a lot of the good that CIL does, why are we holding onto it? Why shouldn’t we discard that way of thinking about CIL and recognize that CIL is an inherently contingent and variable kind of law? I look forward to hearing their answers.