Interpreting Treaties in Domestic Courts

Conflicting Rulings

The application of public international law in domestic courts is a theme that causes much trepidation in litigants yet has been a treasure trove for academics. This is no less true for treaty interpretation, which domestic courts, like other actors, are supposed to approach using Articles 31–33 of the Vienna Convention on the Law of Treaties (VCLT), which bind as provisions some 120 states and the rest as customary international law. Yet, practice varies. In some states, like the US (not a VCLT party) and Japan (a party), courts have often deferred interpretation to the ministry of foreign affairs, and reference to the rules has been sporadic at best. In others, like the UK and Switzerland (both parties), courts have often endorsed the rules but not necessarily followed them to the letter. Overall, the rules seem to be used mostly as an implicit methodology alongside domestic canons of interpretation.

Such observations are quite pertinent in the recent Finnish case of the Eagle S, a now–scrapped tanker, which was registered in the Cook Islands and used to export Russian oil products. On 25 December 2024, the ship damaged several submarine cables by dragging its anchor for about 50 nautical miles in the Finnish EEZ. Upon discovering this, Finnish authorities requested the ship’s crew to navigate to the Finnish territorial sea. After the crew complied, the ship was seized, and its leading crew members were arrested and put to trial. A year later, on New Year’s Eve, the same thing happened with another vessel, the cargo ship Fitburg.

The question is whether Finnish courts may exercise jurisdiction over the alleged offences. Since the jurisdictional provisions of the Criminal Code of Finland allow it, the question boils down to the meaning of Article 97 of the United Nations Convention on the Law of the Sea (LOSC), to which Finland is a party. The provision sets forth an exclusive right to institute penal or disciplinary proceedings concerning any person in the service of a ship to the ship’s flag State and the person’s state of nationality “[i]n the event of a collision or any other incident of navigation concerning a ship on the high seas”.

So, was damaging the cables by dragging the anchor a collision or any other incident of navigation? If it was, Finnish courts could not exercise jurisdiction, that is, institute penal or disciplinary proceedings concerning the crew; if it was not, Finnish courts could do so.

In August 2025, the Helsinki District Court (DC) deemed the Eagle S occurrence an incident of navigation, over which Finnish courts could not exercise jurisdiction (R 706/2025/12270; available on request). The ruling raised some eyebrows in the Finnish press, and it was not applauded internationally either. In a 2025 blog post, it was deemed “[f]or coastal States … an unwelcome precedent”. The implications of the ruling appeared dissatisfying, even dangerous, given Russia’s increasing reliance on hybrid warfare involving a shadow fleet. Unsurprisingly, the judgment was appealed by the prosecutor, and a year later, the Helsinki Court of Appeal (CoA) returned with the opposite verdict (R 16/2025/2572; for now, to my knowledge, only in Finnish on request): it was not an incident of navigation, so Finnish courts could exercise jurisdiction, and the case was sent back to the DC.

Domestic Canons

As of now, the matter of Eagle S (and Fitburg) remains unresolved in the courts. Regardless, at this stage it seems useful to ask how the Helsinki Courts arrived at their distinctive conclusions. Finland being bound by the VCLT, its Courts should interpret treaties “in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose” (Art. 31.1), as accompanied by all the other compulsory and supplementary specifications and factors that follow (Arts. 31.2–4, 32 & 33). But what did the Helsinki Courts actually do?

The DC began its examination of “incident of navigation” from the travaux of the High Seas Convention (precursor to the LOSC), which provide damage to cables as an example of an incident. It then considered the Enrica Lexie Award of 15 May 2020, according to which firing firearms from one ship to another is, to no one’s surprise, not an incident of navigation; incidents must be connected to the movement or maneuvering of the vessel (Award, paras 649–656). The third factor was an ILA Interim Report of 2024, which suggests that intentional damaging of cables might not be an incident. Ultimately, the Court based its view on the fact that the wording of Article 97 does not exclude intentional acts, and that the damage was connected to the maneuvering of the vessel. The DC also noted that intentionality had never been alleged in the charges in the first place.

The CoA forayed first into the intentions of the drafters of LOSC per PCIJ’s Lotus judgment of 7 September 1929, finding the basis of the provision to have been a collision between ships. It then considered the Enrica Lexie Award, noting the factual discrepancy between that case and the one at hand. Finally, it stated that in context, the term incident only includes coincidental occurrences. Regarding the Eagle S, only the initial falling of the anchor may have been such, but dragging it thereafter could not. The travaux in the CoA’s view did not alter this reading, since the statements therein were only exemplary. Article 113 on breaking submarine cables and pipelines was also seen as irrelevant because it is about establishing such acts as crimes.

Both Courts’ approach diverges quite starkly from the VCLT rules. The selection of elements seems eclectic, as does the order of their appraisal. There appears to be heavy (Nordic) emphasis on the travaux, which in the VCLT scheme are rather afforded a secondary rule.

Incident of Navigation

So, what if we followed the VCLT rules, starting with the ordinary meaning of “incident of navigation” (disregarding here the term collision, which could also be discussed)? The term incident, in isolation, covers virtually anything that happens at sea. However, the provision refers specifically to an incident of navigation, of getting from place to place. This is more restrictive: the incident must be of navigation, not just while navigating or even related to navigation.

Considering the Eagle S case, the difference is crucial. Dragging the ship’s anchor for hours and causing damage is perhaps an incident. However, first, the anchor’s purpose is to stop movement, so it is not navigational equipment. Second, the dragging maneuver is not of navigation. Rather, it is something irrational and possibly harmful that can be superfluously done while navigating. Viewing the end result as an incident of navigation goes against, in good faith, the ordinary meaning of the terms, likening to navigation the weaponizing of the vessel’s equipment.

By simply considering the provision’s ordinary meaning, one need not even tackle the discussion (entertained in the Finnish press and by the Helsinki Courts) whether the ordinary meaning covers both intentional and unintentional acts. This latter distinction, which flows from fixating on the concept of incident alone, is a bit problematic. Does not determining intent already involve an exercise of penal jurisdiction (i.e. questioning the crew)? That said, incidents of navigation are coincidentally often unintentional, as anything outside the purview of navigation strongly implies intent.

Now, the ordinary meaning of “incident of navigation” must be determined in its context. The Helsinki CoA refers thereto but seems to overlook that, in treaty interpretation, the context refers to the treaty text, including its preamble and annexes, and certain other instruments made by the parties (Art. 31.2) – not some abstract context. In this sense, the judgments only address Article 113 of the LOSC, leaving out Articles 94.7 and 221.2, which employ the same concept, and the possibility of other relevant contextual agreements or instruments. This is perhaps understandable as leaving out dead ends.

However, the Courts probably should have addressed the immediate context, that the provision only concerns “penal or disciplinary proceedings”. Does not the requirement above that the incident be of navigation unnaturally circumscribe the provision’s scope to a very narrow set of criminal acts? This point is addressed below with reference to the travaux.

Both Courts also seemingly disregarded factors in addition to the context (VCLT, Art. 31.3). This is understandable with regard to subsequent agreements (none) and state practice (lacking an agreement). However, the Courts could have considered other instances of the term incident as possibly “relevant rules of international law applicable in the relations between the parties” (lit. c). The 1973 MARPOL Convention defines incident as comprising discharges of harmful substances into the sea, which does not exclude sabotage. In the Annexes to the Chicago Convention on civil aviation, i.e. ICAO SARPs, incident refers to a lesser kind of accident (Annex 13) that can result from unlawful interference. In my view, the appropriate reason to reject such as determining factors is the provision’s whole wording (incident vs. incident of navigation) and context (the LOSC vs. other regimes), as described above.

That leaves the interpreter with supplementary means (VCLT, Art. 32) like the travaux, case law and legal literature. In the VCLT, their use is subject to certain, somewhat debatable strict criteria, which essentially establish their secondary nature. Here, we arrive at factors that the Helsinki Courts prioritized, such as the travaux, the Enrica Lexie Award, and literature. The main difference of opinion between the Courts here was seemingly the Award’s finding that something not involving movement and maneuvering of a ship cannot be an incident of navigation. However, the Award set this forth only as a minimum threshold (“at least”, para. 639). Something involving movement and maneuvering may thus still be a non-incident, given other factors.

Above, it was noted that requiring that the incident must be of navigation excludes from its scope many criminal acts. This seems odd from a contextual viewpoint, as the provision only concerns penal or disciplinary proceedings. However, the travaux seem to confirm such a narrow reading. The provision was, after all, created in response to the Lotus–Boz-Kourt collision incident of navigation, which involved one ship crashing into another during their regular movement.

Finally, to give credit, both Courts did recognize that the meaning ought to be sought pursuant to the authentic texts (VCLT, Art. 33; LOSC, Art. 320). Though the Courts only utilized the English text, there does not appear to be a difference in meaning between the texts that would have to be reconciled.

In the end, we conclude that dragging an anchor for several hours to harm submarine cables is not an incident of navigation; not because it is supposedly intentional (not coincidental) or due to other factors, but simply because it is not of navigation.

A Call for Coherence

For domestic courts, matters of public international law are often taxing. Cases rarely pop up, but when they do, they tend to be tricky – just ask Italy. General courts handle a broad workload and face pressure to resolve cases expediently. No wonder some courts still defer treaty interpretation to the executive, and others that do not still approach the task somewhat eclectically. Divergence is to be expected, since auto-interpretation of international obligations by states and adjusting them to a domestic context are features of international law, though this has varied. It may also seem sanctimonious to declare that by simply following the VCLT rules courts would reach the exact conclusion. According to one Belgian judgment, violating the VCLT rules is per se not a reason to annul a decision: what really matters is whether the interpreted treaty was contravened. Interpretation is an art, not a science, as the cliché goes.

Yet, the chosen method of reasoning often determines the interpreter’s outlook on the question at hand and affects the soundness of the end result. As argued, using international rules on interpretation would probably improve the effectiveness of international obligations domestically and internationally. Fragmented approaches easily lead to fragmented views on material law. Difficulties in the doctrine of sources are quite understandable in matters that are governed by customary rules. But where a convention prescribes a particular method of interpretation, would it not be most straightforward simply to follow such rules? As another cliché goes, there is no reason to reinvent the wheel.

Explaining why courts forego the VCLT rules is a complex endeavor. Some might simply not know about them, and some might entertain critical legal thinking, regarding them as essentially circular. Some may implicitly accept them but treat them buffet-style. Regardless, the Eagle S judgments discussed here provide an ounce of evidence that treaty interpretation in domestic courts is still an unsettled practice that calls for a broader coherence. Substantively, of course, the saga, whichever way it ends, seems just the beginning of the debate over the thin line between incidents and non-incidents of navigation.