Start Submission

Reading: The ICJ’s Judgement in Somalia v. Kenya and Its Implications for the Law of the Sea


A- A+
Alt. Display

Case Notes

The ICJ’s Judgement in Somalia v. Kenya and Its Implications for the Law of the Sea


Kai-chieh Chan

Université Paris, 2 Panthéon-Assas, Paris, FR
X close


By its judgement of 2 February 2017, the International Court of Justice took up jurisdiction to adjudicate the maritime dispute between Somalia and Kenya. Notwithstanding surrounding controversies, the Court set out important rules concerning the law of treaties. The main implication of the judgment is that the Court embraced a more objective definition of treaties and identified the significance of context as well as travaux préparatoires in treaty interpretation. By doing so, the Court further established itself as the default adjudicator in law of the sea disputes unless the reservation to its jurisdiction is sufficiently precise. This case note summarises the facts and analyses the potential ramifications of this judgement on international dispute resolution. 

How to Cite: Chan, K.-. chieh ., 2018. The ICJ’s Judgement in Somalia v. Kenya and Its Implications for the Law of the Sea. Utrecht Journal of International and European Law, 34(2), pp.195–204. DOI:
  Published on 22 Nov 2018
 Accepted on 01 Nov 2018            Submitted on 18 Dec 2017

I. Background and the Majority Opinion

Somalia initiated the proceedings against Kenya in the International Court of Justice (hereinafter “the Court”) regarding a disputed Exclusive Economic Zone of around 42,000 square kilometers. Somalia based its claim on both parties’ acceptance of the Court’s compulsory jurisdiction under Art. 36(2) of the Court’s Statute, otherwise known as the “optional clause declarations”. In response, Kenya pointed to the reservation it made under the article, which excludes the Court from dealing with disputes ‘in regard to which the parties to the dispute have agreed or shall agree to have recourse to some other method or methods of settlement’.1 Kenya put forward two independent objections to the Court’s jurisdiction and the admissibility of the case: first, it pointed out that a Memorandum of Understanding (MOU) signed between the disputing parties constituted an agreement to have recourse to some other methods of settlement. Second, Kenya argued that the United Nations Convention on the Law of the Sea (UNCLOS), to which both States are parties, contains a dispute resolution mechanism that also amounts to an agreement to have recourse to some other methods.

In relation to the first objection, the Court found that it had to first ascertain the legal status of the MOU before analysing its content. The MOU was signed by the Kenyan Minister of Foreign Affairs and the Somali Minister of National Planning and International Cooperation on 7 April 2009, before subsequently being registered by the Secretariat of the United Nations on 11 June 2009 at Kenya’s request. Despite initially recognising the MOU, the Somali authorities later denied the validity of the instrument, publicly labeling it as “non-actionable” and “void” respectively in October 2009 and February 2014.2 In addition to its constant protests against the MOU, Somalia further argued that the MOU was not ratified by its Parliament and that allowing a Minister to sign binding bilateral agreement was ‘not customary for Somalia’.3 These contentions were, however, rejected by the Court. First, the Court appeared to consider that Somalia’s protests were inadmissible due to laches or acquiescence.4 Second, since the MOU itself provides that it ‘shall enter into force upon its signature’, ratification was unnecessary.5 Lastly, the Court observed that under customary international law, Somalia may not seek to revoke an international law obligations by virtue of internal law provisions regarding competence to conclude treaties; there was no reason to suppose that Kenya was aware that the signature of the Minister may be insufficient.6

Upon characterising the MOU as a treaty that legally bound the parties, the Court moved on to decide whether it constitutes an agreement to some other methods of dispute settlement. If it does, then according to Kenya’s reservation to the Court’s jurisdiction, the Court would not have jurisdiction. The crux lies in paragraph 6 of the MOU, which reads:

‘The delimitation of maritime boundaries in the areas under dispute … shall be agreed between the two coastal States on the basis of international law after the Commission [on the Limits of the Continental Shelf, hereinafter CLCS] … made its recommendations to two coastal States concerning the establishment of the outer limits of the continental shelf beyond 200 nautical miles’.

To fully appreciate the Court’s response to this argument, one needs to first distinguish between delineation and delimitation of the continental shelf beyond 200 nautical miles: although both acts are essentially coastal States drawing lines on their territories, their respective subject matter and the procedure are distinct. Delineation involves drawing the line between the coastal State and the ‘Area’ (the part the High Sea defined as the ‘common heritage of mankind’ by Art. 137 of the UNCLOS) and delimitation involves establishing the line between two coastal States. To prevent coastal States from over-claiming a continental shelf by delineation, Art. 76 of the UNCLOS requires its signatories to make a submission to the CLCS, which would make recommendations to the coastal States. On the contrary, delimitation has no such requirement: communication with the neighbouring state will suffice.7 Although the two acts are indeed distinct, to ensure that its actions do not prejudice matters relating to delimitation, according to its procedural rules the CLCS would not consider a submission on delineation if there was an on-going dispute on delimitation ‘without the consent of all States concerned’.8

Kenya sought to establish the ‘logical’ temporal link that delimitation must come after delineation.9 It argued that since paragraph 6 of the MOU made the unequivocal declaration that issues of delimitation ‘shall be agreed between the two coastal States’ only after the CLCS made its recommendations on delineation, the Court’s decision on delimitation can only come after that. The Court rejected this argument on the basis that according to the wording of both the title and the first five paragraphs of the MOU, the parties intended to keep the process of delineation and delimitation distinct.10 On this matter, the Court made the important observation that the purpose of the MOU is not to set an alternative dispute resolution method, but rather to give the parties’ consent to allow the CLCS to go on with reviewing the submission notwithstanding the existence of a dispute on delimitation. The Court’s interpretation as to the purpose and the parties’ intentions of the MOU is essentially three-fold. First, it analysed the wording of the title as well as the first five paragraphs, before finding that they ‘do not contain any commitments’ or obligations on how the dispute should be resolved.11 Second, in relation to paragraph 6, which contains the word ‘shall’, the Court resorted to the method of interpretation under customary international law as codified by Art. 31(3) (c) of the Vienna Convention on the Law of Treaties (VCLT), which allows it to consider ‘[a]ny relevant rules of international law applicable in the relations between the parties’. Since both Kenya and Somalia are parties to the UNCLOS, the Court observed that there is a level of similarity between paragraph 6 and Art. 83 of the UNCLOS (the latter reads ‘[t]he delimitation of the continental shelf between States with opposite or adjacent coasts shall be effected by agreement on the basis of international law, as referred to in Article 38 of the Statute of the ICJ…’). In reading paragraph 6 of the MOU in light of Art. 83 of the UNCLOS, the Court reasoned that since the latter simply ‘requires that there be negotiations conducted in good faith’ and ‘does not prescribe the method for the settlement of any dispute’, neither should the former.12 In the Court’s view, this interpretation is further substantiated by the ‘subsequent practice’ (Art. 31(3)(b) VCLT) of the parties engaging in negotiations in 2014, even before the CLCS issued any recommendations as to delineation. Had Kenya truly believed that delimitation could only come after delineation, reasoned the Court, it would not have commenced those negotiations.13 Lastly, the Court assessed the travaux préparatoires of the MOU: interestingly, the text of the MOU was drafted by neither party but by the Norwegian Ambassador as part of Norway’s assistance to African countries’ legal development. The Court’s reasoning is that if paragraph 6 had the function of dispute resolution claimed by Kenya, this would have been highlighted by the Norwegian Ambassador.14 But since the Ambassador’s previous talks about the MOU contained nothing about paragraph 6, the Court came to the conclusion that the MOU was insignificant.

In relation to the second objection, Kenya argued that according to Art. 287 paragraph 3 of UNCLOS, State Parties that do not specify their choice of dispute resolution mechanisms will be deemed to have ‘accepted arbitration in accordance with Annex VII’ of UNCLOS. Since neither party has made any choice of dispute resolution mechanism, Kenya argued that Annex VII Arbitration should constitute the parties’ agreement to have recourse to some other methods, which would fall under Kenya’s reservation to the Court’s jurisdiction. On the other hand, Somalia relied on Art. 282 of UNCLOS, which provides that if the signatories ‘have agreed, through a general, regional or bilateral agreement or otherwise, that such dispute shall, at the request of any party to the dispute, be submitted to a procedure that entails a binding decision, that procedure shall apply’ instead of Annex VII Arbitration.15 It claimed that the parties’ acceptance to the jurisdiction of the Court, even in the case of a reservation like Kenya’s, still constitutes a general agreement to settle disputes otherwise and would thus preclude the jurisdiction of UNCLOS tribunals. Kenya’s reply to this was that with its reservation to the Court’s jurisdiction mentioned above, there was no general consent to resolve the dispute in the Court and Art. 282, therefore, did not apply.

The Court’s solution to this conundrum lies in its observation on the structure of UNCLOS and its interpretation of the travaux préparatoires of UNCLOS. As to the structure of UNCLOS, the Court reasserted that according to Art. 286, Art. 287 only matters if ‘no settlement has been reached by recourse to’ articles including Art. 282. Put otherwise, the Annex VII Arbitration and other mechanisms under Art. 287 only play a residual role compared to the general mechanism set out in Art. 282.16 This alone, however, does not solve the question since it was exactly Kenya’s contention that ‘no settlement can been reached’ under Art. 282 with its reservation: the Court thus found it necessary to look into the travaux préparatoires. Most interestingly, the Court found only silence: despite the prevalence of reservation to the Court’s jurisdiction like that of Kenya’s, ‘there is no indication in the travaux préparatoires of an intention to exclude’ the Court’s jurisdiction.17 Hence, the reasoning here is very much the same as the one employed above on the travaux préparatoires of the MOU: if there existed such an intent to exclude the Court’s jurisdiction, there would have been some dispute about it. The Court is also of the opinion that the wording ‘or otherwise’ in Art. 282 was added to make the article as broad as possible in favour of the Court’s jurisdiction, potentially covering reservations that are not particularly specific (like that of Kenya’s).18

The Court, therefore, found itself having jurisdiction since neither the MOU nor Art. 282 of the UNCLOS constitutes alternative dispute resolution mechanisms that fall under Kenya’s reservation to the Court’s jurisdiction. As an ending note, the Court added that taking up the jurisdiction has the benefit of avoiding a negative conflict of jurisdiction.19

II. Analysis

While perhaps not the most discussed judicial decision by the ICJ in recent years, the Judgment on the Preliminary Objections in Somalia v. Kenya may have some implications for the distribution of jurisdiction in law of the sea disputes that cannot be underestimated. Equally controversial is its method of treaty interpretation, which effectively contributes to the Court’s jurisprudence on the definition of treaties, the significance of ‘context’ and applicable relevant rules (Art. 31(3)(c) of the VCLT), and finally, the value of travaux préparatoires.

The judgment has its most immediate influence on the law of the sea. On the one hand, by rejecting Kenya’s argument that a State must first delineate the outer continental shelf before delimitation, the Court’s conclusion on the delimitation of the continental shelf beyond the 200 nautical miles confirms the approach taken by ITLOS (International Tribunal for the Law of the Sea). In the Bay of Bengal case, for instance, the Tribunal stated that:

‘There is a clear distinction between the delimitation of the continental shelf under article 83 and the delineation of its outer limits under Article 76. Under the latter article, the Commission is assigned the function of making recommendations to coastal States on matters relating to the establishment of the outer limits of the continental shelf, but it does so without prejudice to delimitation of maritime boundaries’.20

The current state of law, settled by the ICJ and ITLOS, is therefore, that States are free to choose whether they first follow the CLCS path or instead delimit the continental shelf with their neighboring States.

On the other hand, concerning the jurisdiction of dispute resolution mechanisms, as observed by Prof. Treves, the vast majority of the parties to UNCLOS have not chosen any forum under Art. 287.21 Signatories of UNCLOS that have reservations to the Court’s jurisdiction with one that looks like Kenya’s (e.g. Canada and Australia) may now find themselves bound by the International Court of Justice’s compulsory jurisdiction,22 in case such an objection is raised by one party.23 The Court’s conclusion is arguably consistent with the majority of scholarly opinion:24 for instance, Prof. Alan Boyle comments that: ‘two states which have made declarations in similar terms under Article 36 (2) of the ICJ Statute will remain subject to the compulsory jurisdiction of the ICJ even in the LOS Convention cases’.25 But as confirmed by Prof. Boyle himself in the proceedings of the present case, this time acting as counsel for Kenya, what he wrote did not cover the situation of a reservation to an optional clause declaration. According to Judge Patrick Robinson, the most vehement dissenter in the case, none of the publicists cited by Somalia covered the situation of a reservation.26

If a party wishes to raise an objection of jurisdiction by operation of Art. 282, it will need to demonstrate that there exists an alternative dispute resolution mechanism that is ‘otherwise agreed’ by the parties. The case law of ITLOS provides several clarifications. First, Art. 282 requires that the alternative mechanism ‘entails a binding decision’. This is why the Tribunal denied Japan’s argument in the Southern Bluefin Tuna cases, confirming that a simple commitment to settle the dispute by peaceful and diplomatic means would not amount to such an alternative.27 Second, the alternative dispute resolution mechanism must be shown to be sufficiently general to cover the issues of the law of the sea. In the MOX Plant case (Ireland v. United Kingdom), the United Kingdom raised the objection that the main elements of the dispute were governed by the compulsory dispute settlement procedures of, inter alia, the European Community treaties. The Tribunal rejected the argument by noting that, even if other treaties do contain ‘rights or obligations similar to or identical with the tights or obligations set out in UNCLOS, the rights and obligations under those agreements have a separate existence from those under UNCLOS’.28 In other words, in the Tribunal’s perspective, the jurisdiction of the European Court of Justice is not sufficiently general to cover law of the sea disputes as regulated by UNCLOS.29 In this sense, the ICJ’s ruling in Somalia v. Kenya can be seen as the Court’s declaration that its mandate is sufficiently general to cover law of the sea disputes.

The practical ramifications of Somalia v. Kenya on the law of the sea are, therefore, two-fold. First, arbitral tribunals established under UNCLOS or ITLOS facing similar situations may be more willing to give jurisdiction to the ICJ for reason of judicial comity. Second, States are also advised to form their reservations to the Court’s jurisdiction carefully if they do not want the Court to adjudicate their maritime dispute. As the Court itself puts it in the present case, a better form of reservation to exclude the Court’s jurisdiction here would be to exclude disputes relating to maritime delimitation outright.30

The more general implication of the present case, however, is for the law of treaties. The core issues with using MOUs are the insufficiencies in uniform practice and the lack of terms to distinguish legally binding MOUs from those that are not. For instance, while the United States regards MOUs as valid vehicles for a treaty, the United Kingdom considers otherwise and seeks to avoid words like ‘shall’ or ‘agree’ when negotiating MOUs.31 In Somalia v. Kenya, it can be argued that it is the parties’ intention to make the MOU binding: after all, not only does it contain the word ‘shall’, it is also registered at the UN according to Art. 102 of the UN Charter. On this, although commentators tend to agree that registration at the UN naturally raises a presumption that the instrument is a treaty, many of them argue that an MOU alone does not constitute a treaty.32 There is no consensus, however, on the exact criteria that an MOU needs to be regarded as a treaty. While some consider that every agreement concluded between states is of a normative nature and is not made subject to domestic law is a treaty,33 others fervently argue that more emphasis should be put on the intention of the parties.34 It is possible to relate this academic dispute to the similar schism between subjective and objective theories of interpretation in domestic contract laws,35 or the opposing views on the ‘Lotus rule’ (what is not prohibited is allowed) in the theory of international law.36

In practice, however, there is no settled view amongst arbitral tribunals. Before Somalia v. Kenya, despite the general reluctance of arbitral tribunals to treat MOUs as formal treaties, they are nevertheless generous in granting them residual effect. In the Iron Rhine Railway arbitration, the arbitral tribunal made it clear that an MOU is not binding ‘as a matter of international law’ but is ‘not regarded as being without legal relevance’.37 Tribunals also tended to favour the subjective theory: as the arbitrators in Salini v. Jordan put it, ‘agreements are only binding upon the parties if they intended to create legal relations between themselves’.38 Similarly, in the US-UK Heathrow Airport User Charges arbitration, the tribunal ruled that a 1983 US-UK MOU was not a treaty and, therefore, rejected it as an independent source of rights and duties, as the parties to the MOU did not intend ‘to create independent legally enforceable obligations’.39 The tribunal remarked, however, that the MOU is still a ‘potentially important aid to interpretation’ for it ‘constitutes consensual subsequent practice of the Parties’.40 In comparison, the Court’s reasoning in Somalia v. Kenya may be regarded as a deviation from the subjective approach. It is noteworthy that even though Somalia explicitly asked the Court not to rule on the status of MOU under international law,41 the Court did so anyway by putting forward the brief criteria as follows: ‘under the customary international law of treaties … an international agreement concluded between States in written form and governed by international law constitutes a treaty’.42 In the subsequent paragraphs, the Court focused on the capacity to enter into treaties and left the subjective criteria of intention unexamined.

Two other noteworthy developments pertaining to the general rule of interpretation (i.e. Art. 31 VCLT) are the significance of ‘context’ and the revival of ‘applicable relevant rules’ (Art. 31(3)(c) VCLT). In Somalia v. Kenya, the majority put forward the rule that under Art. 31(1) VCLT, ‘ordinary meaning, context and object and purpose to be considered as a whole’.43 This seemingly harmless principle in fact led the majority to consider paragraph 6 of the MOU (which states unequivocally that delimitation ‘shall’ only ‘be agreed’ by the two countries ‘after’ the CLCS has made its recommendations) as merely entailing a non-binding recommendation to negotiate. This interpretation, as will be discussed later, is achieved through the Court’s operation of context of paragraphs 1–5 of the MOU and Art. 31(3)(c) VCLT. Priority was given to the context of the MOU but not without some sacrifice to the plain meaning of the word ‘shall’. The first and a more philosophical problem with this principle thus concerns the hierarchy of Art. 31(1): should one consider context before considering the plain meaning, for example that of the word ‘shall’? Or should the context only be used to provide assistance in clarifying the ordinary meaning?44 While the majority in Somalia v. Kenya appears to embrace the former approach, the dissenting judges clearly favour the latter approach. As per Judge Bennouna, the majority’s exaggeration of the context ‘is highly unusual’ and ‘ultimately amounts to inverting the order set out in Article 31’.45 Indeed, the Court stated as early as 1950 that:

‘…the first duty of a tribunal which is called upon to interpret and apply the provisions of a treaty, is to endeavour to give effect to them in their natural and ordinary meaning… [if] the words in their natural and ordinary meaning are ambiguous or lead to an unreasonable result, then, and then only, must the Court, by resort to other methods of interpretation, seek to ascertain what the parties really did mean when they used these words’.46

One can observe that the Court in Somalia v. Kenya provided no indication as to why the word ‘shall’ in paragraph 6 of the MOU is so ambiguous and unreasonable that its effet utile was to be deduced by other methods of interpretation.47

More interestingly, the plain meaning of the word ‘shall’ was also at the centre of the dispute in the recent investment arbitration case Churchill Mining and Planet Mining v. Indonesia. In Churchill Mining, the tribunal had to interpret Art. 7 of the UK-Indonesia BIT, which states that Indonesia ‘shall assent’ to all requests for arbitration. On the one hand, the tribunal recognised that the word ‘shall’ may denote an obligation, in that there is no discretion on the part of the host State but to accept a request for arbitration. But on the other hand, as Indonesia contended, ‘shall’ can also be understood as implying a future action, meaning that while a subsequent reply by the host State is required, that reply may not necessarily be a positive one.48 It was the exact same contrast between obligation and future action in the meaning of the word ‘shall’ that was at stake in Somalia v. Kenya. While the Court resorted to context to solve the conundrum in Somalia v. Kenya, the tribunal in Churchill Mining went further to abandon Art. 31 VCLT and embrace the supplementary means in Art. 32 VCLT, only to conclude that ‘shall assent’ functionally equals to ‘hereby consents’.49 This interpretation is regarded by some commentators as unusual or done in bad faith.50 Without estimating any future impact that Somalia v. Kenya may have on the definition of the word, the word ‘shall’ itself no longer guarantees the legal certainty one may imagine.

Equally significant to the Court’s mention of context is its use of Art. 31(3)(c) VCLT, which allows interpretation of treaties in light of ‘relevant rules of international law applicable in the relations between the parties’. The rule allows a court or a tribunal in charge of interpreting a treaty to refer to general international law or other treaties, either for the purpose of gap-filling or for the future development of international law.51 The Court has previously used Art. 31(3)(c) VCLT in the Military and Paramilitary Activities in and against Nicaragua case of 1986, the Oil Platform case of 2003, and the Djibouti v. France case of 2008. In Oil Platform, the Court famously interpreted the US- Iran Treaty of Amity of 1955 using the general international law on self-defense.52 This approach attracted considerable criticism: for Judge Higgins and Judge Kooijmans, instead of evaluating the legality of U.S. actions under general international law, the real problem should have been whether the US had violated the treaty at hand.53 In all three cases, the Court was asked whether declarative provisions in other treaties (such as the agreement to deal on the basis of ‘equality, mutual respect and peace’) could have any impact on interpretation. On this point, the Court’s view is consistent, that although these general provisions do not themselves constitute legal obligations that could be independently violated, they do have ‘a certain bearing on the interpretation and application’ of the treaty at hand.54 The Court’s approach to Art. 31(3)(c) VCLT in Somalia v. Kenya, however, was significantly different: it ruled that paragraph 6 is not an agreement on dispute settlement due to its similarity with Art. 83(1) of UNCLOS, the latter of which does not prescribe a method of dispute settlement.55 In the eyes of dissenting judges, this method is unprecedented and highly unsatisfactory: the wording of the two texts is not exactly the same and their objects, as well as their contexts, are arguably different. The Court is also accused of intentionally neglecting Art. 83(2) of UNCLOS, which comes immediately after Art. 83(1) and requires the parties to resort to Annex II arbitration instead of the ICJ.56 Interpretation of treaties based on similar terms, or provisions in other treaties applicable between the parties, is not unheard of: in the WTO US-Shrimp case, the Appellate Body has famously read that the term ‘exhaustible natural resources’ in light of similar provisions within UNCLOS.57 While assuming that similar terms imply similar intentions across different treaties, however, one is well advised to take into consideration the differences in context, object, purposes, subsequent practices, and even travaux among them.

Last but not least, one can readily discern the extensive reliance on travaux préparatoires to interpret the intention of the drafters in Somalia v. Kenya. Two major observations can be made. First, in the present case the Court considered the relevant travaux préparatoires not for what it revealed, but instead for its silence on a particular matter. The Court assumed a similar stance in the Oil Platform case, in which it invoked the silence in the U.S. domestic ratification process to dismiss part of Iran’s interpretation of the Treaty of Friendship between the U.S. and Iran.58 But this method is not without some controversy: as Judge Bennouna put it, ‘[h]ow can one interpret the silence of a text in such a way?’59 In Cruz Varas v. Sweden, the European Court of Human Rights had to decide whether it was authorised by the Convention to order interim measures; upon recognising the silence of travaux of the Convention on the issue, the Court readily confirmed the the travaux as irrelevant.60 Judge Robinson, himself a former Ambassador involved in the negotiation of UNCLOS, disagreed with the majority’s conclusion that the silence in the travaux préparatoires of UNCLOS implies the parties’ intention to allow the ICJ’s jurisdiction even in the case of a reservation. He argued that if the majority’s interpretation was true, it would suggest that State Parties of UNCLOS ‘treated reservations as having no legal significance. It is improbable that this could have been their approach’.61 A more accurate description of the intention of the drafters of UNCLOS, according to Judge Robinson, was that they simply had not ‘given any thought whatsoever to those reservations’ since ‘there was no need to elaborate because it was obvious…that Article 282 would not apply to an optional clause declaration with the Kenyan-type reservation’.62 At the end of his dissent, Judge Robinson was even able to identify segments of travaux which showed that proposals to make the ICJ at the head of the list of fora in Art. 287 were rejected. International lawyers should be particularly prudent when interpreting silence, not only due to the often fragmentary nature of preparatory works, but also because of the established doctrine that reliance on preparatory works can only be supplementary.63

A second observation can be made on the opposability of travaux préparatoires to parties that are not present in the negotiations. As pointed out by Judge Bennouna, in relation to the MOU, ‘there are simply no such travaux in the relations between the two States parties to the MOU. At most, there are elements concerning the assistance extended by the Norwegian Ambassador Longva to the Parties to conclude this agreement’.64 Since neither Kenya nor Somalia was involved in the drafting of the MOU, can the travaux of the MOU be invoked to illustrate the parties’ intentions? In the Territorial Jurisdiction of the International Commission of the River Oder case, the Permanent Court of International Justice refused to admit the part of preparatory works in which three of the parties to the dispute had not participated;65 some commentators share the same opinion.66 According to the International Law Commission (ILC), nevertheless, the River Oder case does not constitute jurisprudence constante, since ‘a State acceding to a treaty in the drafting of which it did not participate is perfectly entitled to request to see the travaux préparatoires’.67 A possible compromise seems to be that the travaux can be invoked only if they have been publicly available.68 In Somalia v. Kenya, the Court mentioned the two available ‘travaux’ of the MOU: one is a presentation given by the Norwegian Ambassador at a conference and the other comments on the protection of Somali natural resources written by the Permanent Mission of Norway. Although this information is indeed publicly available, in light of the dissents of Judge Bennouna one may question the ILC’s decision to abstain from any definition of travaux. Commentators like Charles de Visscher have notably refused to accept documents not emanating from contracting States to be classified as travaux.69

III. Concluding Remarks

The core lesson of the Court’s judgment on the preliminary objection in Somalia v. Kenya is to be prudent. If a party wishes to insist on the point that no negotiation should commence before a certain date, it should take care not to carry out any negotiations at all, whatever the excuse may be. Prudence is also required in the process of reservations to the Court’s jurisdiction, which must be unequivocal as to which kinds of dispute are to be excluded. The corollary of this requirement is that States are not encouraged to draft ambiguous reservations in the hope of escaping from the commitment to the compulsory jurisdiction of the World Court. The Court has made it clear that, in the case of an optional clause declaration that contains the words that the country accepts its jurisdiction on ‘all disputes’, it will only deny jurisdiction if the parties have explicitly agreed on resorting to another method of settling the dispute.70 In the same vein, States are not encouraged to alleviate their international obligations simply by calling the legal instrument a name other than a treaty (such as a MOU). Prudence is further advised to the drafters, who should clarify whether they intend it to denote an obligation or a future action by using the word ‘shall’. Practice has shown that without some guidance from the text, courts or tribunals will need to venture into the realm of contextual or supplementary means of interpretation, which may not give the word its intended meaning.

On interpretation, we are prone to forget Vattel’s first general maxim: in claris non fit interpretatio (do not interpret when there is no need for interpretation). Context, from the other articles in the same treaty to similarly articulated paragraphs in other treaties, may be just as important or even more important than the plain meaning of the text. Preparatory works, regardless of their origin or the level of their expressiveness, may be considered to support an interpretation reached otherwise.

By virtue of the judgement in Somalia v. Kenya, the Court has established itself as the default adjudicator in the law of the sea disputes unless the reservation to its jurisdiction is sufficiently precise. On the one hand, arbitral tribunals established under UNCLOS facing similar situations may be more willing to give jurisdiction to the ICJ for judicial comity in the future. On the other hand, the concern of a negative conflict of jurisdiction was clearly one of the reasons that motivated the Court’s ruling. Although the issue was not dealt with in detail by the Court, its significance cannot be underestimated: as the Permanent Court of International Justice once stated in the Chorzów Factory case, the Court:

‘…cannot allow its own competency to give way unless confronted with a clause which it considers sufficiently clear to prevent the possibility of a negative conflict of jurisdiction involving the danger of a denial of justice’.71

One can read this along with the Court’s deliberate choice not to interpret the acceptance of its compulsory jurisdiction and the relevant reservations restrictively.72 With some level of flexibility preserved, the Court saves itself from the perplexing issues of jurisdiction and thereby focuses, instead, on achieving substantive justice for the parties.


1United Nations, 531 Treaty Series 114. Emphasis added. 

2Somalia v. Kenya, Judgement on Preliminary Objections (2 February 2017) paras 18–19. 

3ibid. para 38. 

4ibid. para 42 (‘Somalia did not protest that registration until almost five years thereafter’). See also para 49: ‘[Somalia] must by reason of its conduct be considered as having acquiesced in the validity of that treaty’. 

5ibid. para 45. 

6ibid. para 49. See Art. 45 of the Vienna Convention on the Law of Treaties (VCLT); similarly Cameroon v. Nigeria, Equatorial Guinea intervening, Judgement (10 October 2002) para 265 (observing that, under customary international law, any restraint on the capacity to enter into treaty shall be ‘manifest’ and ‘at least properly publicized’). 

7Bjarni Már Magnússon, The Continental Shelf Beyond 200 Nautical Miles (Brill-Nijhoff 2015) 135–136. See the Court’s judgement in Nicaragua v. Colombia: ‘the role of the CLCS relates only to the delineation of the outer limits of the continental shelf, and not delimitation’ (Judgement on Preliminary Objections, 17 March 2016 para 110). 

8Somalia v. Kenya (n 2) para 69. Rules of Procedure of the CLCS, Annex I, Art. 5, para a. 

9ibid. paras 53–54. 

10ibid. paras 75–77. 

11ibid. para 72. 

12ibid. paras 90–91. See also para 95 (‘the fact that the Parties set an objective as to the time for concluding an agreement does not, given that this paragraph is not prescriptive of a method of settlement to be followed’). The interpretation on UNCLOS is consistent with the Court’s previous judgement in Cameroon v. Nigeria, Equatorial Guinea intervening (n 6) para 244 (declaring that Art. 83 of the UNCLOS only requires that there be negotiations conducted in good faith, but not that they should be successful). 

13ibid. paras 92–93. But on this point, both Judge ad hoc Gilbert Guillaume and Judge Mohamed Bennouna argued that the majority ignored the fact that throughout the negotiations Somalia rejected the validity of the MOU. In their view, Kenya’s endeavour to bring Somalia back to the table (under the obligation to negotiate of the MOU) cannot be reasonably treated as the evidence that Kenya has renounced its right under the MOU. See the dissenting opinion of Judge ad hoc Gilbert Guillaume, at para 27 (stating: ‘Cette acceptation de la négociation frontalière peut-elle aujourd’hui être opposée au Kenya? … J’ai les plus grands doutes à cet égard’); see also the dissenting opinion of Judge Mohamed Bennouna, at 3. 

14ibid. paras 103–104. 

15ibid. paras 110–111. Emphasis added. 

16ibid. para 125. This position is prima facie consistent with the doctrinal opinion e.g. Yoshifumi Tanaka, The International Law of the Sea (CUP 2015) 420 (describing the dispute resolution system under UNCLOS as ‘two-tier’ and stating that ‘the compulsory procedures are essentially residual’). 

17ibid. para 129. 

18ibid. paras 128, 130. 

19As some commentators observed, the concern might stem from Kenya’s last minute declaration to exclude maritime delimitation from the Convention’s compulsory system. Marco Benatar and Erik Franckx, ‘The ICJ’s Preliminary Objections Judgment in Somalia v. Kenya’ European Journal of International Law Analysis (22 February 2017) <> accessed 21 September 2017. 

20ITLOS, Delimitation of the Maritime Boundary Between Bangladesh and Myanmar in the Bay of Bengal, Judgment of 14 March 2012, para 376. 

21Tullio Treves, ‘Conflicts Between the International Tribunal for the Law of the Sea and the International Court of Justice’ (1999) 31(4) Journal of International Law and Politics 818. 

22See State Parties’ Declaration Under Article 36(2) of the Statute of the Court <> accessed 21 September 2017. 

23Thus the procedure was able to continue in the Southern Bluefin Tuna cases (New Zealand v. Japan, Australia v. Japan), XXIII RIAA 1, by a tribunal constituted under Annex VII of UNCLOS, despite the fact that all parties have accepted compulsory jurisdiction of the ICJ. 

24The one exception that the author is aware of is Bernard H. Oxman, ‘Courts and Tribunals: The ICJ, ITLOS, and Arbitral Tribunals’, in Rothwell, Elferink, Scott, and Stephens eds., The Oxford Handbook of the Law of the Sea (OUP 2015) 401–402 (suggesting that UNCLOS arbitral tribunals or ITLOS, once considering a case like that in Somalia v. Kenya, may want to construe Article 282 strictly so as to afford the aggrieved party a choice of forum ‘because the ICJ is not wanting for cases.’) Prof. Oxman’s opinion was not considered in the proceedings. 

25Alan E. Boyle, ‘Problems of Compulsory Jurisdiction and the Settlement of Disputes Relating to Straddling Fish Stocks’ (1999) 14 International Journal of Marine and Coastal Law 7. Holding the same position are Tanaka (n 16) at 423–424 and Treves (n 20) at 64. 

26Dissenting opinion of Judge Patrick Robinson para 31. 

27ITLOS, Southern Bluefin Tuna cases (n 23) paras 53–55. 

28ITLOS, MOX Plant case, Request for Provisional Measures Order of 3 December 2001, para. 50. 

29Tanaka (n 16) 423. 

30Somalia v. Kenya (n 2) para 128. 

31John H. McNeill, ‘International Agreements’ (1994) 88 American Journal of International Law 821. 

32Anthony Aust, Modern Treaty Law and Practice (CUP 2007) 50–53 (stating: ‘it would seem that an MOU has effect only in the realm of politics or morals’); Richard Gardiner, Treaty Interpretation (OUP 2015) 90; Sir Robert Jennings and Arthur Watts, Oppenheim’s International Law (CUP 2008) 1202–1203 (stating: ‘Thus [State parties] may conclude a memorandum of understanding … thereby implying that they do not have the intention of entering upon legal rights or obligations’). 

33Jan Klabbers, The Concept of Treaty in International Law (Springer 1996). Klabbers’ theory has its foundation in Qatar v. Bahrain Judgement on Jurisdiction and Admissibility (1 July 1994) (finding exchanges of letters and minutes of discussions as constituting treaties.). 

34Aust (n 32) 49, 52 (depicting judgment in Qatar v. Bahrain as ‘special’.). 

35See e.g. Marcel Fontaine and Filip De Ly, Drafting International Contracts (Transnational Publishers 2006) 107, 138 (opposing the English and French approach which purportedly favours the subjective theory of contract interpretation to the German and Dutch approach of objective interpretation). 

36Robert Kolb, Théorie du droit international (Bruylant 2013) 315 et seq (describing the rule’s origin in the famous Lotus case). 

37Belgium v. Netherlands, Award of 24 May 2005, XXVII RIAA 98 para 156. 

38Salini S.p.A. v. Jordan, ICSID Case No. ARB/02/13, Award (31 Jan 2006, Tribunal: Bernardo Cremades, Ian Sinclair, and Gilbert Guillaume) para 78 (emphasis added). 

39U.S. v. U.K., Award of 30 November 1992, XXIV RIAA 131 para 6.8. 

40ibid. para 6.7. 

41Somalia v. Kenya (n 2) para 34; see also para 41: ‘Somalia has invited the Court to reject Kenya’s preliminary objection without considering the status of the MOU under international law’. 

42ibid. para 42. 

43ibid. para 64. 

44Gardiner (n 32) 210 seems to support the latter view. 

45Dissenting opinion of Judge Bennouna, at 2. 

46Competence of the General Assembly for the Admission of a State to the United Nations, Advisory Opinion (3 March 1950) 8 (emphasis added). This rule was confirmed by the Court in Arbitral Award of 31 July 1989 (Guinea-Bissau v. Senegal), Judgement (12 November 1991) para 69 as reflecting customary international law. 

47Dissenting opinion of Judge ad hoc Gilbert Guillaume, para 9 (stating that ‘l’utilisation du mot "shall" marque qu’il s’agit là d’une obligation’). 

48Churchill Mining PLC and Planet Mining Pty Ltd v. Republic of Indonesia, ICSID Case No. ARB/12/14 and 12/40 (Tribunal: Albert Jan van den Berg, Michael Hwang, Gabrielle Kaufmann-Kohler), Decision on Jurisdiction (24 February 2014) paras 157, 162, 166. 

49ibid. para 230. 

50Andrea Bjorklund, Georgios Petrochilos, Stephan Schill and Diane Desierto, ‘Investment Law at the Crossroads of Public and Private International Law’ (2014) 5 Select Proceedings of the European Society of International Law, 173–174. 

51Gardiner (n 32) 333; Belgium v. Netherlands (n 37) paras 58–59 (treating relevant treaties in environmental law and the laws of development as a whole.). 

52Islamic Republic of Iran v. United States of America, Judgment (6 November 2003) 182 para 41. 

53Separate Opinion of Judge Kooijmans para 42; Separate Opinion of Judge Higgins para 49. 

54Djibouti v. France, Judgment (4 June 2008) 38 para 114; Islamic Republic of Iran v. United States of America, Judgment on Preliminary Objection (12 December 1996) 813; Military and Paramilitary Activities in and against Nicaragua, Judgment (27 June 1986) 136–137. 

55Somalia v. Kenya (n 2) paras 90–91; para 97 (stating ‘the text of the sixth paragraph of the MOU reflects that of Article 83, paragraph 1, of UNCLOS’). 

56Dissenting Opinion of Judge Bennouna 3–4; dissenting opinion of Judge ad hoc Gilbert Guillaume para 12: ‘[o]n notera cependant que ces textes ont des objets fondamentalement différents’. 

57WTIDS58/AB/R (12 October 1998) paras 127 et seq. 

58Islamic Republic of Iran v. United States of America, Judgment on Preliminary Objection (12 December 1996) 815 para 29. 

59Dissenting Opinion of Judge Bennouna 4. See also the dissenting opinion of Judge ad hoc Gilbert Guillaume para 31: ‘les travaux préparatoires sont quasiment muets en ce qui concerne ce paragraphe. Quant aux circonstances dans lesquelles le mémorandum a été conclu, elles ne nous éclairent pas davantage’. 

60Cruz Varas and Others v. Sweden, Judgement (20 March 1991) para 95. 

61Dissenting opinion of Judge Robinson para 27. 

62ibid. paras 30–31. 

63P-M. Dupuy and Yann Kerbrat, Droit international public (13th edn, Dalloz 2016) 357 para 315; Aust (n 32) 247 (stating ‘travaux must therefore always be approached with care. Their investigation is time-consuming, their usefulness often being marginal and very seldom decisive’). 

64Dissenting Opinion of Judge Bennouna 4. 

65Mustafa Kamil Yasseen, ‘L’interprétation des traités d’après la convention de Vienne sur le droit des traités’ (1976) 151 Collected Courses of the Hague Academy of International Law 89–90. 

66Jean Combacau and Serge Sur, Droit international public (L.G.D.J. 2016) 181: ‘ils ne sont pas opposables aux sujets qui n’ont pas participé à la négociation’; Yasseen, ibid. 

67Yearbook of the ILC, 1966–II 222. 

68See Judge Schwebel’s dissenting opinion in Qatar v. Bahrain (n 33), Judgement of 15 February 1995, 38–39 (rejecting Qatar’s objection that part of the draft text prepared by Saudi Arabia as inadmissible as ‘it was never sent the draft in question’. While the Court did not find it necessary to decide on the issue, Judge Schwebel stated that ‘[n]one of the preparatory work at issue was or is secret, or known to one but not another Party’). 

69Yearbook of the ILC, 1966–II, 223 paras 20–22; Yasseen (n 65) 84–86. 

70Somalia v. Kenya (n 2) para 119. 

71PCIJ, Factory at Chorzów, Judgment on Jurisdiction, 26 July 1927, Series A, No. 9, at 30. 

72CIJ, Fisheries Jurisdiction Case (Spain v. Canada), Judgment of 4 December 1998 para 44. (‘Conditions or reservations thus do not by their terms derogate from a wider acceptance already given. Rather, they operate to define the parameters of the State’s acceptance of the compulsory jurisdiction of the Court. There is thus no reason to interpret them restrictively’). 

Competing Interests

The author has no competing interests to declare.

comments powered by Disqus