H.L.A Hart devotes chapter 10 of The Concept of Law to international law, and in particular to the question of whether international law counts as a genuine case of law. Though I’m open to persuasion on the topic, I’m somewhat skeptical of the idea that international law is genuine law, and find Hart’s arguments in favor of its validity as law rather confused. That said, this post is devoted to a small and all-things-considered inconsequential confusion in Hart’s discussion, not the larger issue at the center of the chapter. So the point I’m making is a semi-pedantic one, but I’m going to make it anyway.
According to Hart, international law, unlike municipal law, lacks anything like “an international legislature, courts with compulsory jurisdiction, and centrally organized sanctions” (CL, p. 214). It also lacks the structure of primary and secondary rules that Hart takes to be a central feature (though, somewhat confusingly, not the defining feature) of law (CL, p. 214).
Given these facts, we might be inclined to infer that international law is too different from the paradigm cases of “ordinary” law to count as law at all. But we might equally be inclined to infer that international law is just a different kind of law than the sort we regard as paradigmatic. Given this stark and apparently irresolvable disagreement, anyone making either inference might be inclined to leave the matter there, each stipulating on the basis of sheer preference that the scope of the term “law” includes or excludes international law as a matter of semantic fiat. But Hart rejects this “stipulate and let stipulate” approach. Semantic fiat, he argues, won’t suffice. We need a unified account that includes both municipal and international law.
What I dispute here is not so much Hart’s conclusion as his argument for it in the passage excerpted below.
The short way suggested [i.e., dismissal of the issue by semantic stipulation] would indeed be appropriate if we were dealing with a proper name. If someone were to ask whether the place called ‘London’ is really London, all we could do would be to remind him of the convention and leave him to abide by it or choose another name to suit his taste. It would be absurd, in such a case, to ask on what principle London was so called and whether this principle was acceptable. This would be absurd because, whereas the allotment of proper names rests only on an ad hoc convention, the extension of general terms of any serious discipline is never without its principle or rationale, though it may not be obvious what that is (CL, p. 215).
Hart’s argument presupposes that the allotment of proper names for places rests on a single unequivocal convention that decides each case in a determinate and conclusive way. In other words, if someone asks whether a given place in England is “really London,” we are, Hart suggests, to “remind him of the convention” (emphasis mine) that settles the matter, as though a single convention did so, and was the only thing that did so. The “allotment of proper names,” Hart writes, “rests only on an ad hoc convention,” (my emphasis on “an”) the implication being that a single convention suffices to resolve disputes about any and all place names. Though I agree that this is sometimes true–and may well be true of London–it isn’t always true. Where it isn’t true, the “absurdity” to which Hart alludes does not arise.
Consider some examples. I grew up in the St. Cloud section of West Orange, New Jersey, “St. Cloud” being the unofficial name of a vaguely-defined neighborhood in West Orange.* One convention (not necessarily the only one) that defines the borders of this neighborhood is the school district map created by the township’s Board of Education. The purpose of the map is to define the set of families eligible to send their children to St. Cloud Elementary School: all and only those who live in the St. Cloud school district may attend St. Cloud Elementary School; those who live outside of the St. Cloud district must attend other schools. (I don’t have access to the actual map that the School Board uses, so consider what I say about a mere just-so story about it.)
The School Board’s convention is bound to be fairly well defined, but may not be perfectly so. We can imagine, at least hypothetically, that the convention does not dictate the status of land that was undeveloped at the time when any given version of the map was drawn up. Those who defined the boundaries of the St Cloud school district may simply have ignored then-undeveloped land as irrelevant to the goal of determining membership in the school district, leaving the matter undetermined for the time being.
Once previously-undeveloped land is developed, of course, a decision has to be made about its disposition as regards membership in the school district. In this context, it’s not absurd to ask whether or not a newly developed (previously undefined) parcel of land “really is” part of the St. Cloud neighborhood/school district. Granted, it might be more natural to ask whether or not the parcel should become part of the St. Cloud district. But it’s not illogical to ask whether it is part. The latter question presupposes that while that while the existing boundaries of the St. Cloud district don’t directly or explicitly address the disposition of this particular parcel of land, they might tacitly imply one.
For instance, a given parcel might be so close to the borders of the St Cloud district, and so far from the borders of any other, that it might just make intuitive sense to regard it, for all practical purposes, as existing “within” St. Cloud. In cases like this, “whether it is within” and “whether it should be within” are conceptually distinct but practically identical inquiries. The more obvious the answer, the less conceptual space between the one sort of inquiry and the other.
I’ve so far been treating the School Board’s map as the only relevant convention defining the boundaries of the district, but there’s no reason to assume that. Other entirely distinct conventions might apply. Notoriously, real estate agents and street gangs have their own definitions of where neighborhoods begin and end, not necessarily ones that correspond exactly to the maps drawn up by the local School Board.
Or there might just be an informal tradition (or traditions), separate from the decisions of the school board (or the real estate brokers and street gangs), of regarding the St. Cloud district as having such-and-such boundaries. The traditions in question might be vague or contested. If so, while disputes about whether a certain plot of land was “in” or “outside of” St. Cloud might seem (or might be) petty and pointless, they wouldn’t be logically absurd. One tradition might set the boundary at one place for one reason; another might set it somewhere else for another. It might be that, on the traditional understanding, some public schools technically outside of the St. Cloud district fall directly into it. Likewise, some St. Cloud students might be attending St. Cloud School while residing outside of what locals consider the St. Cloud district. And so on.
What’s true of the St. Cloud district of West Orange is true of many of the other places where I’ve lived. I lived for a few years in Whitehouse Station, New Jersey, an unincorporated community within Readington Township in Hunterdon County, New Jersey. Whitehouse Station was one of several such communities within Readington–including Whitehouse proper, Dreahook, Stanton, etc. School board decisions aside, the “boundaries” of these communities were set by local traditions, some stretching back a century or more. That fact made it logically intelligible (though practically speaking, a waste of time) to ask where Whitehouse Station began and where it ended–what was “really” in Whitehouse Station, and what was in Whitehouse proper or Dreahook, or Stanton, or whatever.
Another example: I now happen to live in Princeton, New Jersey. Strictly speaking, Princeton’s boundaries are set by municipal law and include Princeton Borough and Princeton Township, but informally, there is a tradition of referring to the “Greater Princeton Area” well beyond those boundaries as “Princeton.” Thus, businesses in nearby Montgomery Township bear the name “Princeton Volkswagen,” or “Princeton Deli,” often for crassly commercial reasons, but sometimes out of a sense of sentimental attachment to the locality, along with the belief that Princeton somehow extends miles beyond its official boundaries. Are such businesses “really in” Princeton? It depends on how you look at it. Something similar is true of Princeton University, part of which, technically speaking, juts into adjacent townships. It seems self-evident that (all of) Princeton University should be located (entirely) in Princeton. One convention says that it does. But a few say otherwise.
Before I lived in Princeton, I lived on a 70 acre farm outside of Lambertville, New Jersey. Going by the map, by the school district, and by the routing for 911 calls, I lived in West Amwell Township, a rural pro-Trump enclave five miles inland from the Delaware River. But according to the US Postal Service, I lived in zip code 08530, the zip code for Lambertville, a trendy, leftish tourist trap directly on the Delaware. And I had a Lambertville street address. So was the farm really in Lambertville or was it really in West Amwell? It depends on which convention you adopt. In my case, that depended on who I was talking to at any given time.
Washington Heights, viewed from New Jersey. Clearly a dated photo, as I can barely recognize my old neighborhood from it, to the left of the bridge. (Wikipedia Commons)
Before that, I lived in the Washington Heights neighborhood of Manhattan, north of Harlem, and west of Inwood. The same thing could be said of its boundaries, but I won’t belabor the point any more than I have.
Contrary to Hart, there’s no single convention at work in these examples, but several potentially competing ones. Where there is practical conflict between them (and there sometimes is), there is a need for the adjudication of the rival claims involved. In cases like that, settling place names is not as simple a matter as Hart suggests. We can’t just settle the matter by reminding the contestants of some single convention that does all of the normative work. We have to do some normative work to figure out what convention is the right one to apply.
For a more radical and politically consequential example of this phenomenon, consider the allotment of place names to and within Jerusalem and the West Bank. I’ve written about this before at PoT, so I won’t repeat myself at length here. Suffice it to say that it makes perfect sense to ask whether a given parcel of land is “really” in Jerusalem or better yet, really in East Jerusalem; or whether a given village or refugee camp is really “in” or actually “outside of” the West Bank, or “Judea and Samaria.” Though not quite the issue Hart raises, it can make sense to ask whether the “real” Tekoa is the one inhabited by Palestinian Arabs or the newly-fabricated one of the same name nearby, inhabited by Jewish Israeli settlers, and intended to supersede the first. These questions may seem petty to outsiders but to insiders are live questions often resolved with live ammunition.
Even more fraught is the general question, for any location between Gaza and Jericho, or Golan and the Negev, whether one is “really” in Israel, or “really” in Palestine, or really in neither of the two, but rather in Israel/Palestine or (as Muammar Gaddafi used to say) really in “Isratine.” Once, when a Palestinian friend and I were standing on the western shore of Lake Tiberius (or do I mean the Sea of Galilee?), he pointed wistfully at the eastern shore, referring to it unself-consciously as “Syria.” I was about to correct him–to tell him that it wasn’t “really” Syria–when it occurred to me that there was no point in doing so, any more than there was a point to correcting him when he referred to Israeli Golan as “Syria.” We were, momentarily, operating by different place-name conventions. So I left it at that. We remain friends.
Once you get started with examples of this sort, it becomes hard to stop. Is the shrine located at the center of Hebron “really” the Cave of Machpela, or is it really Masjid al Ibrahimi? Is the shrine located in the Old City of Jerusalem “really” the Temple Mount or is it really Haram Sharif? If someone refers to Masjid al Aqsa as the “Temple Mount” can that “really” be so if there’s no actual temple there, except as a matter of historical conjecture and/or utopian imagination? It’s not clear. If conventions really settled these matters, no one would fight over them. But try telling the people doing the fighting.
What’s true of the places I’ve described is likely to be true of many other places as well. Is Taiwan really in China? How about Hong Kong and Tibet? Is Catalonia really in Spain? Or is that precisely not so? Is Ukraine really in Russia? Or does it name a place decidedly outside of Russia? Is the Pakistani-controlled part of Kashmir really Azad Kashmir? Or is it just Kashmir? Or is it Pakistan? Is Pine Ridge Indian Reservation really a part of the United States? Or is it, as some Lakota nationalists would tell you, a sovereign nation unto itself? No single convention answers any of these questions.
So what’s the upshot for Hart’s overall argument? There isn’t one, really–at least as far as international law is concerned. All that my argument here implies is that even when it comes to place names, semantic fiat is not the simple matter that Hart makes of it. So the contrast he draws between place names and theoretical concepts is somewhat (though not wholly) ill-chosen. “[W]hereas the allotment of proper names rests only on an ad hoc convention,” he writes, “the extension of general terms of any serious discipline is never without its principle or rationale, though it may not be obvious what it is” (CL, p. 215). Contrary to Hart, the allotment of proper place names doesn’t always rest on a single convention, and sometimes none of the conventions involved are literally ad hoc. Like general terms in serious academic disciplines, there is often a deep rationale behind the allotment of a given place name, “though it may not be obvious what it is.”
Academics tend not to regard the allotment of place names as being on par with the considerations one encounters in “serious academic disciplines,” but the thing is, non-academics do. Indeed, they sometimes even go beyond that. Ironically enough, that tendency may well be a fundamental rationale for the existence of international law–assuming, of course, that it exists.
*There’s a brief, interesting discussion of the concept of “neighborhood” in Michael Walzer’s Spheres of Justice: A Defense of Pluralism and Equality, pp. 35-42, one of the few I’ve encountered in political philosophy.