The ICJ and Evolutionary Treaty Interpretation

The ICJ and Evolutionary Treaty Interpretation

On Monday, the ICJ delivered its judgment in the Costa Rica v. Nicaragua case, concerning navigational and related rights on the river San Juan (Registry summary; judgment). The case itself is not terribly important in the grand scheme of things, but upon reading the judgment I came across several questions of broader import that our readers might find of interest.

But first just to say a bit about the facts of the case for the sake of greater comprehension. The river San Juan is on the border between Costa Rica and Nicaragua. The border, however, does not run along the thalweg of the river, as is usually the case with river borders, but is set by treaty to run along the Costa Rican coast. Thus, the entire river belongs to Nicaragua as sovereign. The question raised by the case is what are the navigational and related rights on the river of Costa Rica and its riparian communities, under the terms of the 1858 Treaty of Limits between Costa Rica and Nicaragua. The principal of these rights was set by Article VI of the Treaty, giving Costa Rica a perpetual right of free navigation ‘con objectos de commercio’.

There was a dispute between the parties as to the meaning of this last phrase, with Costa Rica claiming that it meant ‘for the purposes of commerce’, while Nicaragua maintained that it meant solely ‘with the articles of commerce’, i.e. with merchandise (para. 45). On this issue, the Court sided with Costa Rica (paras. 50-56), and the judgment is generally speaking more in its favor than in Nicaragua’s. The background to this dispute is at least partly one of economic self-interest: Costa Rica claimed that this phrase covered not merely trade in goods, but also trade in services, such as the transportation of tourists and passengers on the river by Costa Rican ships. Nicaragua, on the other hand, asserted that the phrase covered only the trade in goods.

This brings me to the issues of general import raised in the case. First, and most important, is the Court’s approach to treaty interpretation. In that regard the Court gave short shrift to its long-standing dicta that restrictions on state sovereignty are not to be presumed lightly (paras. 47-48), and it also made several fascinating remarks on evolutionary treaty interpretation, to which I will devote the bulk of this post.

Second, the Court was quite ‘activist’ in basically creating bright line rules out of whole cloth from treaty provisions that were silent on the concrete matter at hand, but which the Court considered to have exhaustively regulated the relations between the parties. For instance, the Court held that Nicaragua may not impose any visa requirements on passengers on Costa Rican ships on the river, though it could have held instead that Nicaragua may not unjustifiably deny visas to such passengers, but that it could do so in some exceptional circumstances. This willingness to set bright line rules probably emanated from the Court’s desire to preclude further disputes between the parties and definitively settle the matter.

Third, there is the Court’s willingness the establish, with very little evidence, the existence of a bilateral customary rule binding the parties. Namely, it found that the riparian communities in Costa Rica, through the Costa Rican state, had a customary right to subsistence fishing from the river. The Court derived such a customary right solely from the absence of protest by Nicaragua to fishing by Costa Ricans, without in any way searching for opinio juris by Nicaragua and Costa Rica that this absence of protest derived not from the lack of practical significance of the very small amount of fishing involved, but from Nicaragua’s sense of legal obligation, its view that the riparian Costa Ricans had a right to fish as they did.

Indeed, the Court found such a customary right even though Costa Rica never claimed, until the case came to the Court, that it had such a right in its relations with Nicaragua. The Court justified its position by saying that the paucity of evidence of custom is explained by the remoteness of the area and its small population (para. 141). Now, we all know that in some cases the Court does not apply the criteria for formation of custom in a very rigorous way (e.g. regarding the use of force in the 1986 Nicaragua case), while in others it is conversely extremely hostile to the existence of custom (e.g. in the Asylum or the North Sea Continental Shelf cases). But I don’t think that I’ve ever seen the Court establish a customary rule, even a bilateral one, in quite so off-hand a way as it did here. In that regard, I find the separate opinion of Judge Sepulveda to be more persuasive.

But now onto the juiciest part of the judgment, the Court’s approach to evolutionary treaty interpretation. (Warning and apologies in advance for any overly-theoretical rambling).

The readers will recall that the Court found Article VI of the 1858 Treaty to give Costa Rica perpetual freedom of navigation on the river ‘for the purposes of commerce’. But this still left the meaning of the word ‘commerce’ to be determined. According to Costa Rica, this term encompassed everything and anything, any activity or dealing between persons, including the transportation of passengers and the use of the river by Costa Rican officials providing public services (para. 59). For Nicaragua, in 1858 the term ‘commerce’ meant only trade in goods, and this was the meaning that governed (para. 58):

In Nicaragua’s view, for purposes of the Treaty, “commerce” covers solely the purchase and sale of merchandise, of physical goods, and excludes all services, such as passenger transport. This interpretation is clearly consistent with Nicaragua’s contention, just rejected, that “con objetos” means “with merchandise”. But, Nicaragua argues, even if the phrase is translated as “for the purposes of commerce”, the result is the same, because in 1858 the word “commerce” necessarily meant trade in goods and did not extend to services, the inclusion of services being a very recent development. Nicaragua admits that passengers were already being transported on the San Juan in 1858, and even that this was an especially profitable activity, but it adds that this activity did not fall within the scope of what was commonly called “commerce” at that time. As for the transport of tourists, there was no such activity at the time in the area in question.
Nicaragua contends that it is important to give the words used in the Treaty the meaning they had at the time the Treaty was concluded, not their current meaning, which can be quite different, because this is the only way to remain true to the intent of the drafters of the Treaty; and determining that intent is the main task in the work of interpretation.

Thus, in Nicaragua’s view, it is the 1858 meaning of the word ‘commerce’, not its 2009 meaning, that was dispositive. The Court disagreed (paras. 63-64):

It is true that the terms used in a treaty must be interpreted in light of what is determined to have been the parties’ common intention, which is, by definition, contemporaneous with the treaty’s conclusion. That may lead a court seised of a dispute, or the parties themselves, when they seek to determine the meaning of a treaty for purposes of good-faith compliance with it, to ascertain the meaning a term had when the treaty was drafted, since doing so can shed light on the parties’ common intention. The Court has so proceeded in certain cases requiring it to interpret a term whose meaning had evolved since the conclusion of the treaty at issue, and in those cases the Court adhered to the original meaning (to this effect, see, for example, the Judgment of 27 August 1952 in the case concerning Rights of Nationals of the United States of America in Morocco (France v. United States of America) (I.C.J. Reports 1952, p. 176), on the question of the meaning of “dispute” in the context of a treaty concluded in 1836, the Court having determined the meaning of this term in Morocco when the treaty was concluded; the Judgment of 13 December 1999 in the case concerning Kasikili/Sedudu Island (Botswana/Namibia) (I.C.J. Reports 1999 (II), p. 1062, para. 25) in respect of the meaning of “centre of the main channel” and “thalweg” when the Anglo-German Agreement of 1890 was concluded).

This does not however signify that, where a term’s meaning is no longer the same as it was at the date of conclusion, no account should ever be taken of its meaning at the time when the treaty is to be interpreted for purposes of applying it.

On the one hand, the subsequent practice of the parties, within the meaning of Article 31 (3) (b) of the Vienna Convention, can result in a departure from the original intent on the basis of a tacit agreement between the parties. On the other hand, there are situations in which the parties’ intent upon conclusion of the treaty was, or may be presumed to have been, to give the terms used ? or some of them ? a meaning or content capable of evolving, not one fixed once and for all, so as to make allowance for, among other things, developments in international law. In such instances it is indeed in order to respect the parties’ common intention at the time the treaty was concluded, not to depart from it, that account should be taken of the meaning acquired by the terms in question upon each occasion on which the treaty is to be applied.

Our readers familiar with US debates on constitutional interpretation, originalism etc. must have already picked on the references to ‘original meaning’ and ‘original intent,’ and the equally familiar confusion and oscillations between (supposedly objective) meaning and the intentions of the parties that have plagued the law for a couple of millennia at least. Further (paras. 66-68),

[W]here the parties have used generic terms in a treaty, the
parties necessarily having been aware that the meaning of the terms was likely to evolve over time, and where the treaty has been entered into for a very long period or is “of continuing duration”, the parties must be presumed, as a general rule, to have intended those terms to have an evolving
meaning.

This is so in the present case in respect of the term “comercio” as used in Article VI of the 1858 Treaty. First, this is a generic term, referring to a class of activity. Second, the 1858 Treaty was entered into for an unlimited duration; from the outset it was intended to create a legal régime characterized by its perpetuity.

This last observation is buttressed by the object itself of the Treaty, which was to achieve a permanent settlement between the parties of their territorial disputes.

Therefore (paras. 70-71),

The Court concludes from the foregoing that the terms by which the extent of Costa Rica’s right of free navigation has been defined, including in particular the term “comercio”, must be understood to have the meaning they bear on each occasion on which the Treaty is to be applied, and not necessarily their original meaning.

Thus, even assuming that the notion of “commerce” does not have the same meaning today as it did in the mid-nineteenth century, it is the present meaning which must be accepted for purposes of applying the Treaty.

Accordingly, the Court finds that the right of free navigation in question applies to the transport of persons as well as the transport of goods, as the activity of transporting persons can be commercial in nature nowadays. This is the case if the carrier engages in the activity for profit-making purposes. A decisive consideration in this respect is whether a price (other than a token price) is paid to the carrier ? the boat operator ? by the passengers or on their behalf. If so, then the carrier’s activity is commercial in nature and the navigation in question must be regarded as “for the purposes of commerce” within the meaning of Article VI. The Court sees no persuasive reason to exclude the transport of tourists from this category, subject to fulfillment of the same condition.

On the other hand, any navigation not carried out either to transport goods intended to form the subject of commercial transactions or to transport passengers in exchange for money paid by them or on their behalf cannot be regarded as falling within the “purposes of commerce” under Article VI. That is the case, in particular, of navigation by vessels used in the performance of governmental activities or to provide public services which are not commercial in nature.

So, to sum up: because the contracting parties used a generic term, commerce, and created a perpetual regime, the meaning of the term ‘commerce’ is to be updated every time that the treaty is applied. It is its 2009, not its 1858, meaning that governs.

This approach of the Court is of course quite familiar – it is used all the time in many domestic legal systems in interpreting constitutional provisions, as well as in international human rights law. Cf. the European Court’s view of the ECHR as a ‘living instrument,’ etc., etc. Indeed, a problem closely analogous to this one is the US Supreme Court’s interpretation of the Commerce Clause of the US Constitution, giving Congress the power to regulate foreign and inter-state commerce. But this method of evolutionary interpretation is, of course, controversial, especially in the US, with the various incarnations of the originalist movement arguing that the Constitution should be interpreted in accord with the original intentions of its framers, or (as is now the case) in accord with its original public meaning.

The chief problem with many of these debates on interpretation is naturally that the various discussants usually speak past each other. It is indeed quite apt that these debates frequently degenerate because the speakers use different meanings of ‘meaning’, and so forth. So, though I’d like to try to comment on the ICJ’s ruling above, I first have to define my starting points. And in that regard, it is the interpretation/construction or the interpretation/application distinction that I find to be most useful (see more here at Larry Solum’s blog). The former is the activity of establishing the linguistic or semantic meaning of a text; the latter the activity of translating that text into workable legal rules to be applied in a given case.

The emphasis on this distinction is particularly helpful because the distinction is so frequently disregarded. Art. 31 VCLT does sort of distinguish between interpretation and application, but perhaps not strongly enough. The ICJ’s ruling above, of course, elides the distinction completely, and that is my chief point of criticism. Let me try to explain this more fully.

Take, for example, Article 7 ICCPR which prohibits, inter alia, ‘cruel, inhuman or degrading treatment or punishment’. Similarly, see Art. 3 ECHR or the Eighth Amendment to the US Constitution, prohibiting cruel and unusual punishments. As stated above, interpretation is the activity of establishing the (linguistic or semantic, ordinary or plain) meaning of these words. The word ‘cruel’ thus means ‘disposed to inflict pain or suffering : devoid of humane feeling; causing or conducive to injury, grief, or pain.’ This semantic meaning of the word ‘cruel’ is the same today in 2009, as it was after the Second World War when the ECHR and the ICCPR were being drafted, as it was in 1789 when the US Bill of Rights was drafted. The meaning of the word has NOT changed with the passing of time.

The application of that meaning, however, its construction into rules governing particular situations, HAS changed. Thus, in 1945 the community of states might not have thought that subjecting juveniles to the death penalty was ‘cruel’; they think so – and we as a community of lawyers think so – today. Likewise, an execution by hanging in a public square might not have been thought of as ‘cruel’ in 1789, but it is thought of as ‘cruel’ in today’s America. This is not because the meaning of the word ‘cruel’ is different, or because its interpretation has changed; it is because our application of that meaning has been altered. The challenge of evolutionary interpretation, be it of of a constitution or of a treaty, is not really one of interpretation, but one of construction, and of establishing the process through which a change in construction can happen yet remain broadly legitimate. Are we, for instance, in our application of the word ‘cruel’ in any way bound by what previous generations considered to be cruel? If so, to what extent? (One such process, comparatively widely used, is for a court to assess the growing consensus of the community bound by a particular instrument. Cf. the European Court’s approach in searching for a European consensus on particularly sensitive moral issues, and its relationship with the margin of appreciation doctrine.)

In some cases, the distinction between interpretation and construction may be hard to draw, and this indeed may be true for the present Costa Rica v. Nicaragua case. But the distinction is nonetheless indispensable, as it is the only way of assuring the fixation of the core of a legal norm, and thus the only way of assuring a level of legal certainty and predictability, but to at the same time still allow for non-legislative change in the law. (And, to paraphrase JHHW’s last EJIL editorial, I am well aware of the whole radical indeterminacy thing, but please do spare me).

So, back to the ICJ’s ruling: it is troublesome precisely because it does not acknowledge the interpretation/construction or application distinction, in whatever form, and because the Court holds that the term ‘commerce’ was generic, and therefore intended by states to have a changing meaning. Even if the term ‘commerce’ meant only trade in goods in 1858, says the Court, it also means trade in services, such as transportation and tourism, in 2009, and this is the meaning that we will use. Of course, the peril with such reasoning is that in 2050 or 2100 or whenever the word ‘commerce’ might lose any relation to the meaning of the word today or in 1858. Merriam-Webster tells me, for instance, that today ‘commerce’ can also mean any kind of ’social intercourse’, particularly ’sexual intercourse.’ What are we to do then if the meaning of the word becomes solely the latter – will Costa Rica then only have the perpetual right to freedom of navigation for purposes of sexual intercourse, or will it retain its previous rights regarding trade in goods or transportation of passengers? If so, why, when the meaning of the word has changed, and when the parties have intended, as the Court says, for the current meaning to be applied?

The answer to this question lies, again, in the interpretation/construction distinction, and in the fixation of some core semantic meaning of a word to the meaning that it had when it was written or uttered. For instance, if in the future the word ‘cruel’ came to mean only something mildly unpleasant, we would still apply the meaning it had when the text was drafted, that of causing injury, grief or pain. The same goes, in principle, for the term ‘commerce.’ If in 1858 the term truly semantically denoted only trade in goods, as Nicaragua contended, then I honestly don’t see why a subsequent change in the meaning of the word, be it to ‘trade in goods and services’ or ’sexual intercourse’ should change its interpretation.

If, however, the term commerce in 1858, then as now, semantically denoted an activity for profit that involves an exchange for money, but the term as interpreted was only applied to situations involving trade in goods, as only such situations in fact occurred, then there is no obstacle in construing the term ‘commerce’ more expansively to take into account the changing circumstances. This, however, has nothing to do with the changing meaning of the word, but with a change in the application or construction of that meaning.

It is only this second possible original meaning of the word ‘commerce’ that is truly generic, as the Court puts it, and that allows for an evolving interpretation, or better, for an evolving construction. A further factor to be taken into account in such an interpretation is the subsequent practice of the parties under Art. 31(3)(b) VCLT, as Judge Skotnikov well points out in his separate opinion, which may well have shed some more light on the matter. But it cannot be, as the Court appears to say, that the semantic meaning of a word was one in 1858, and another in 2009, and that it is this latter that should be controlling. If a mere change in the dictionary definition of a word in a legal text could lead to its reinterpretation, then radical indeterminacy will truly become the only game in town.