When the international lawyers get to be heard – the story of tax treaty interpretation as told in Sweden

At the time of writing, Sweden is a party to approximately 80 bilateral tax treaties.

Cf. Skatteverket, Rättslig vägledning, http://www4.skatteverket.se/rattsligvagledning, Search path: “Regler och ställningstaganden”, last visited on 19 September 2015.

No tax treaty concluded by Sweden establishes rights for physical and legal persons resident in the country. According to the approach taken in Sweden to the issue of the relationship between international and domestic law, tax treaties acquire effect domestically due only to an act of incorporation. This is to say, Parliament has to adopt a law containing a text similar to that of the Act (1998:258) concerning the Double Taxation Agreement between Sweden and Macedonia:

The agreement for the avoidance of double taxation on income and capital signed by Sweden and Macedonia, on 17 February 1998, shall apply as law in this country.

Modified up until SFS 2011:1390; italics are added. Translation to English by the author. In Swedish, the provision reads: “Det avtal för undvikande av dubbelbeskattning beträffande skatter på inkomst och förmögenhet som Sverige och Makedonien undertecknade den 17 februari 1998 skall gälla som lag här i landet.”

Thus, irrespective of the obligations owed by Sweden and Macedonia to each other under international law, it was not until the 1998 Act of Parliament entered into force that physical and legal persons resident in Sweden could claim tax relief for dividends on shares paid by a company resident in Macedonia.

Obviously, incorporation entails consequences only on the condition that Swedish tax authorities are entitled to impose tax on such dividends according to other tax legislation.

When a tax treaty has been incorporated into the Swedish legal system, however, an intimate relation of substance is created between the treaty and domestic law. This is perfectly illustrated by the quote from the 1998 Act concerning the Double Taxation Agreement between Sweden and Macedonia. Given how Parliament normally chooses to construct a law on incorporation of a tax treaty, the meaning and significance of that law will be wholly dependent on Sweden’s international obligations. This means that taxpayers and the Swedish Tax Agency – and by extension, Swedish administrative courts – will sooner or later find themselves engaged in tax treaty interpretation. When the meaning of a tax treaty cannot be established simply by reading its text, no person or authority can say for sure, without having first interpreted the treaty, what precise legal substance was incorporated into Swedish law.

The Vienna Convention makes a distinction between interpreting and understanding a treaty. In the terminology of the Convention, it is fully possible for a person to confer a clear meaning upon the text of a treaty without first engaging in interpretation. I will return to this issue in Section 4.

Scholars at Swedish universities specializing in international taxation seem to agree that when interpreting tax treaties concluded between Sweden and other states, consideration must be had to Articles 31–33 of the 1969 Vienna Convention on the Law of Treaties (VCLT).

UNTS, Vol. 1155, p. 331. The VCLT applies to treaties concluded between states. It is to be noted that another Vienna Convention on the Law of Treaties was adopted in 1986; it applies to treaties concluded between states and international organizations and between international organizations inter se. UN Doc. A/CONF.129/15. The two Vienna Conventions are sometimes confused. See, for example, Bjuvberg (2015a, p. 113, n. 6) and Bjuvberg (2015b, p. 428, n. 3 and 5). The 1986 Vienna Convention lacks relevance for the interpretation of tax treaties concluded by Sweden with other states, for obvious reasons. In addition, it has not yet entered into force.

Consequently, tax-law scholars have had reason to comment repeatedly on the meaning of Articles 31–33 of the VCLT. Their commentary has appeared in various publications. It has covered topics ranging from the significance of party intention for the application of Articles 31–33 and the status of those articles under international law to the possibility of ambulatory interpretation and the extension of fundamental concepts, such as “the object and purpose of a treaty” (Article 31, paragraph 1), and “the special meaning” (Article 31, paragraph 4).

See, for example, Berglund (2013, pp. 63–84), Bjuvberg (2015a, pp. 111–130), Bjuvberg (2015b, pp. 427–443), Cejie (2010, pp. 79–98), Dahlberg (2003, pp. 137–155), Dahlberg (2014, pp. 249–271), Hilling (2014, pp. 322–340), Kleist (2012, pp. 47–123), and Sallander (2013, pp. 51–61). The Story of Tax Treaty Interpretation as Told in Sweden

First is the observation that I feel little at home with the picture of the Vienna Convention painted by the tax law literature. The understanding of Articles 31–33 represented by the body of Swedish tax-law expertise sits ill with my own collected experience of treaty interpretation in an international law context generally. Second, as a rule, tax-law scholars apply a methodology that differs substantially from the methodology that international lawyers would normally use for finding themselves in a corresponding situation. While international lawyers are trained to establish their assertions regarding the meaning of Articles 31–33 based on the practice of international judicial bodies and international legal literature bearing on the topic, tax-law scholars, to a considerable extent, base their assertions on findings of colleagues active in the same field. Reading Swedish academic literature on international taxation leaves me with the clear impression that the system suffers from something similar to an acoustic feedback. Even if no one seems prepared to supply the basis for recommended understandings of the Vienna Convention, years of repetition have generated widespread acceptance of those same understandings, making criticism exceptionally difficult.

There is strong reason to break this negative pattern and let international law doctrine take its rightful place on the scene. As I firmly believe, cross-disciplinary dialogue is needed to bring the discussion on tax treaty interpretation entertained in the camp of Swedish tax-law expertise on a more constructive course. It is precisely for this reason that I have written this article. Consequently, I will use this article to highlight a total of 11 assumptions about Articles 31–33 that I have found to bewell represented and/or influential in Swedish tax-law literature. I will offer assessment of all 11 assumptions viewed from the perspective of the current practice of international courts and tribunals and the work of international legal scholars. It is my sincere hope that this will help open the door to a more general dialogue between tax-law and international law expertise.

The article will be organized so that each of the 11 assumption corresponds to a separate section. I will use a methodology tailored to suit its purpose. It should be clearly understood that in writing this article, I have no intention to criticise the work of any single tax-law scholar. What I wish to stress is what I conceive to be a failure of an entire system. For this reason, I will approach my task as typically legal scholars do when analyzing and assessing a highlighted proposition rather than its utterance by some particular person or institution at some particular occasion. I will introduce each section with a statement, which in a verbally condensed form captures the well-represented and/or influential assumption to be commented upon. To ensure that statements are not perceived as direct quotes, they will be put using italics.

Articles 31–33 are relevant for the interpretation of tax treaties concluded by Sweden with other states because they reflect customary international law.

See Berglund (2013, p. 65, n. 130), Cejie (2010, p. 88), Dahlberg (2003, p. 139, n. 9), and Kleist (2012, p. 71).

Before Articles 31–33 of the VCLT can be applied for the purpose of the interpretation of a tax treaty between any two states (A and B), two conditions must be met: A and B must be parties to the VCLT and the tax treaty must have been concluded by A and B after the point in time when the VCLT entered into force between them.

Cf. VCLT Article 4. It should be noted that among international legal scholars there has been some disagreement as to precisely how to understand the second condition. See, for example, Thirlway (1972, p. 108) and McDade (1986, pp. 499–511).

Since many years, international judicial practice takes for granted that customary rules on treaty interpretation are identical to those contained in the Vienna Convention.

See, for example, Maritime Dispute (Peru v. Chile), Judgment of 27 January 2014, http://www.icj-cij.org, pp. 26–27, para. 57.

While Article 31 is applied for the purpose of establishing the meaning of the text of a treaty (objective interpretation), the purpose of Article 32 is to establish the intentions of the treaty parties (subjective interpretation). This is to say that the text of a treaty shall have priority over the intentions of its parties.

See Bjuvberg (2015a, pp. 113, 117), Berglund (2013, p. 67), Cejie (2010, p. 89), Dahlberg (2014, p. 252), and Kleist (2012, pp. 72–73).

The ordinary meaning conferred by Swedish lawyers on the terms “objective” and “subjective” interpretation is commented upon by Professor Stig Strömholm:

An important distinction that should be noted in this context, as it is still commonly upheld in discussion on legal interpretation, is that between ’subjective’ and ’objective’ method of interpretation. In very general terms, ’subjective interpretation’ can be said to refer to a method characterized by the aim of establishing and finding guidance in the purposes and intentions that actually existed at the time of the creation of a law (it is this emphasis on the ’historical law-maker’ that justifies the use of the term ’subjective’). An ’objective’ method, on the other hand, is characterized by its proponents attempting to find guidance, not in any ”law-makers intentions, but in ’objective’ circumstances related to the law as such.

See Strömholm (1996, p. 453). Translation to English by the author. In Swedish, the passage reads: “En begreppsbildning, som bör omnämnas i detta sammanhang och som alltjämt ofta användes i lagtolkningsdiskussionen, är uttrycken ’subjektiv’ och ’objektiv’ tolkningsmetod.Mycket generellt kan det sägas, att uttrycket subjektiv brukar ges åt en metod som kännetecknas av strävan att finna och söka ledning i de syften och avsikter som faktiskt förelegat vid lagstiftningens tillkomst (det är detta betonande av den ’historiska lagstiftaren’ som rättfärdigar termen ’subjektiv’). En ’objektiv metod’ däremot kännetecknas av att dess förespråkare vid tolkningsproblemens lösning vill söka ledning icke i lagstiftarens avsikter utan i sådana (’objektiva’) omständigheter som hänför sig till själva lagen.”

So interpreted, the italicized proposition introducing section 3 bears on one of the founding ideas of Articles 31–33. It has far-reaching practical implications: any tax-law expert who embraces the claim without due reflection will inevitably often go astray when trying to come to grips with more concrete issues of interpretation–such as the question of whether a static or an ambulatory approach should be taken to the interpretation of tax treaties.

Cf. Linderfalk and Hilling (2015, pp. 34–59).

The ultimate purpose when interpreting a treaty, as described in Articles 31–33 of the VCLT, is to establish the intentions of the treaty parties–or more accurately, the meaning that the parties intended the treaty to communicate.

See, for example, Navigational and Related Rights (n. 18), p. 242, para. 43.

Articles 31 and 32 serve to limit the interpretation process in two ways. The first limitation lies in the fact that only certain types of interpretative data may be used. Acceptable data are organized into categories; in the terminology of the Vienna Convention, categories of data correspond to means of interpretation. According to Article 31, consequently, interpreters may draw upon conventional language, the context,

The context is considered to comprise not only the categories of data defined in Article 31, paragraph 2, but also those defined in paragraph 3. See, for example, Oppenheim’s International Law (1992, p. 1274, n. 17). Cf. the ILC Commentary on the Final Draft Articles on the Law of Treaties: “[T]he word ‘context’ in the opening phrase of paragraph 2 is designed to link all the elements of interpretation mentioned in this paragraph to the word ‘context’ in the first paragraph and thereby incorporate them in the provision contained in that paragraph. Equally, the opening phrase of paragraph 3 ‘There shall be taken into account together with the context’ is designed to incorporate in paragraph 1 the elements of interpretation set out in paragraph 3.” Report of the International LawCommission to the United Nations General Assembly, covering the work of the second part of its seventeenth and eighteenth sessions (UN Doc. A/6309/Rev.1), Yearbook of the International Law Commission (1966:2), p. 169, at 220.

When international lawyers interpret a treaty in accordance with Article 31, they naturally start by determining the ordinary meaning of the terms of the treaty. When the ordinary meaning is ambiguous, international lawyers turn to determining which one of the alternative ordinary meanings best coheres with the context and the object and purpose of the treaty. In other words, no interpretation according to Article 31may possibly exceed the boundaries of conventional usage. In this respect, Article 31 differs from Article 32, which allows for more extensive interpretations, although only under strictly limited conditions; the condition is that the application of Article 31 leads to a “manifestly absurd or unreasonable” result. This is not to say that, ultimately, the application of Article 31 is intended to serve a different purpose than that of Article 32. For both articles, the purpose is to establish the intentions of the treaty parties.

See, for example, Oppenheim’s International Law (1992, p. 1267). Cf. the ILC Commentary on the Final Draft Articles on the Law of Treaties (n. 29), p. 169, at 220: “The article as already indicated is based on the view that the text must be presumed to be the authentic expression of the intention of the parties; and that, in consequence, the starting point of interpretation is the elucidation of the meaning of the text, not an investigation ab initio into the intentions of the parties.”

The second limitation introduced by Articles 31 and 32 is the hierarchical organization of interpretative data. If application of Article 31 leads to a result other than that ensuing from application of Article 32, then normally the treaty shall be understood in accordance with Article 31. It is only when the application of Article 31 leads to a result, which is either ambiguous or obscure, or manifestly absurd or unreasonable, then the treaty shall be understood in accordancewith Article 32. This is not to say that the text of a treaty takes priority over the intentions of the treaty parties, as all means of interpretation are used to establish party intention. The hierarchical organization of means of interpretation is instead based on the idea that when seeking to establish the intentions of the treaty parties, some means of interpretation normally provide better guidance than others.

Cf. The ILC Commentary on the Final Draft Articles on the Law of Treaties (n. 29), pp. 217–223.

Articles 31–33 do not provide any rules of interpretation proper. They provide only guidelines for the interpretation process.

See Berglund (2013, p. 67) and Bjuvberg (2015a, p. 130.)

“Interpretation” is a tricky word inasmuch as it can be used in several different senses. First, “interpretation” can be used to refer either to just any process of understanding of a text or to understanding obtained or attempted only under limited conditions.

See Dascal and Wróblewski (1988, pp. 203–205).

If interpretation is the only way to understand the text of a treaty, it is impossible to disclose a difference of the meaning of two or more authenticated versions without having previously applied Articles 31–32. Article 33, paragraph 4 apparently assumes that treaties can be understood even without interpretation.

Second, international lawyers can talk about the “interpretation” of treaties and be interested in two different things: knowing how a specific reader or group of readers actually came to understand the treaty in some particular way or knowing whether and how this same understanding can be justified.

On the distinction between the discovery and justification of legal propositions generally, see, for example, Wasserstrom (1961, pp. 25–31) and Wroblewski (1992, pp. 14–16). On the significance of this distinction for the understanding of Articles 31–33 of the VCLT, see Linderfalk (2014b) and Linderfalk (2007b, pp. 133–154).

In any case, in the particular context of Articles 31–33 of the VCLT, there is no doubt as to what is meant by “interpretation”. When states and international judicial bodies apply Articles 31–33, it is because they are interested in establishing whether or not an assertion about the meaning of a treaty can be justified. The motive that actually generated this same assertion is for them of no material consequence.

The proposition that Articles 31–33 provide only guidelines of interpretation is based on a dual misconception. It presupposes that in the context of the Vienna Convention, “interpretation” means, first, any process of understanding of the text of a treaty and, second, the discovery of its meaning. This dependence of the proposition on a particular understanding of the word “interpretation” comes out very clearly in Martin Berglund’s monograph on the avoidance of international double taxation. It is no mere coincidence that this is the work that most forcefully argues the idea that Articles 31–33 are to be seen only as a collection of guidelines. Berglund writes:

If, prior to interpretation, you wish to determine whether or not the intention was expressed in the text, this requires that you can distinguish between the text before and after interpretation. However, such a distinction cannot possibly be made.

See Berglund (2013, p. 68). Translation to English by the author. In Swedish, the passage reads: “Om man före tolkningen ska kunna bedömaomavsikten har uttryckts i texten krävs det att man kan skilja mellan texten före tolkning och texten efter tolkning. En sådan strict åtskillnad låter sig dock inte genomföras.”

He continues a few pages later:

When you ask the question whether some subsequent legal development must be taken into account for purposes of interpretation, it has in fact been taken into account already, in the sense that it forms an element of your pre-understanding of the issue. Therefore, interpretation is necessarily dynamic, because the prevailing relationship between meaning and context is in continual flux.

Ibid, p. 76. Translation to English by the author. In Swedish, the passage reads: “När man ställer sig frågan om en viss senare rättsutveckling ska beaktas i den aktuella tolkningssituationen är den redan beaktad i den bemärkelsen att den utgör en del av den rättsliga bakgrundsförståelsen. Tolkning är därför med nödvändighet dynamisk eftersom meningssammanhanget ständigt ändras.”

In a universe where “interpretation” is thought to mean just any process of understanding of the text of a treaty and Articles 31–33 of the VCLT are seen as a means to explain the discovery of its meaning, Berglund’s observations are completely conceivable and possibly even correct. In the international legal universe that actually prevails, his observations are irrelevant.

The ILC Commentary adopted in 1966 is of good help when attempting to understand Articles 31–33.

See Dahlberg (2003, p. 150), Kleist (2012, p. 73 ff) and Sallander (2013, p. 54 ff).

As we can see, Articles 31–33 are not entirely easy to understand and apply correctly. This seems to be in part a result of the human element: Swedish lawyers understand Articles 31–33 in the light of their collected experience of the interpretation of Swedish legislation and private law contracts. Partly, it seems to be a result of the particular design of the VCLT. To enable the adoption of a legal regime, which can be applied to treaties without having further regard to details such as subject–matter or scope, or number of parties, and which allows at the same time scope for necessary adaptation of the regulation to the particular interests and needs that may have to be satisfied in concrete cases of interpretation, Articles 31–33 have been drafted with a primary focus on means of interpretation. No one means of interpretation allows itself any inferences about the meaning of a treaty. To enable successful analysis, interpreters must combine means of interpretation with one or other communicative assumptions. The Articles, however, provide only sporadic reference that helps identify the particular communication assumptions that international law finds acceptable. For more detailed guidance, interpreters must fall back on international judicial practice, which of course remains a less accessible means for the determination of law.

Because of this, Articles 31–33 themselves must often be subjected to interpretation. When tax-law scholars interpret Articles 31–33, they sometimes look for support in the extensive Commentary that the International Law Commission adopted in 1966;

See ILC Commentary on the Final Draft Articles on the Law of Treaties (n. 29).

The ILC is a subsidiary organ of the UN General Assembly. Its particular task is to codify and progressively develop international law. Typically, for a very long time, this task was fulfilled following an established procedure: the Commission would list topics suitable for codification; on the basis of careful studies of relevant practice, following lengthy discussions within the Commission, the Commission would adopt a set of final draft articles, which together with an accompanying Commentary would be passed on to the General Assembly for further action. The General Assembly in turnwould typically convene a diplomatic conference, which would negotiate on the basis of the ILC draft articles; it would make the necessary amendments or additions; and possibly, it would adopt a treaty that could be signed and ratified or acceded to by states. This was the procedure followed in the case of the conclusion of the VCLT.

The 1966 Commentary, together with the final draft articles adopted by the ILC, was part of the background material supplied to all Vienna Conference participants.

Documents of the Conference, United Nations Conference on the Law of Treaties, First and second sessions, Vienna, 26 March to 24May 1968 and 9 April to 22 May 1969, Official Records, p. 7.

When you determine the ordinary meaning of a treaty, this is done primarily with the resort to the specific terminology practiced in the relevant sphere of activity governed by the treaty, such as international taxation.

See Dahlberg (2003, p. 151) and Kleist (2012, pp. 74–75, 89).

In the terminology of Article 31 of the VCLT, the “ordinary meaning” refers to the meaning conferred on the terms of a treaty in conventional language.

Note that conventional language comprises not only a lexicon and a grammar, but also rules of pragmatics. On the importance of this observation, see, for example, Linderfalk (2008a, pp. 343–364) and Linderfalk (2013, pp. 241–250).

If the parties to a tax treaty reach an agreement concerning its interpretation, then the agreement can be used as a means of interpretation, based either on Article 31, paragraph 2, lit. a or b, or Article 31, paragraph 3, lit. a; this applies even when the agreement is binding under international law.

See Kleist (2012, p. 108 ff) and Sallander (2013, p. 55).

There is something contradictory about this assumption. If, on the one hand, the parties to a tax treaty enter into an agreement concerning its interpretation, and this agreement is binding under international law, naturally, the agreement applies. There is no need to invoke any particular provision of the Vienna Convention. The parties to the treaty must apply the interpretative agreement, irrespective of whether or not the agreement can be used as a means of interpretation, in the sense of Articles 31–33, based on any of the particular provisions that they contain.

See, for example, Legal Consequences for States of the Continued Presence of South Africa in Namibia (SouthWest Africa) notwithstanding Security Council Resolution 276 (1970), Advisory Opinion of 21 June 1971, ICJ Reports 1971, p. 22, paras. 21-22.

The special meaning of the terms of a treaty is a means of interpretation that can be used in exactly the same way as any other means of interpretation.

See Bjuvberg (2015a, p. 117), Dahlberg (2014, p. 252), Dahlberg (2003, p. 153), Kleist (2012, p. 90) and Sallander (2013, p. 55).

According to Article 31, paragraph 4, “a special meaning shall be given to a term if it is established that the parties so intended.”Many tax-lawscholarswould seem to think that the special meaning of a term in a treaty amounts to very much the same as the meaning conferred upon it in the relevant technical language, in this case, the parlance of international tax lawyers. If, for example, a commentary to the Organization for Economic Cooperation and Development (OECD) Model Convention indicates that among international tax lawyers, “the exercise of an employment” means something else than it does among users of everyday English, then, as the argument goes, Article 31, paragraph 4, provides a basis for interpreting the term in the light of tax-law parlance rather than everyday language.

This argument is based on a misunderstanding. In the terminology of the Vienna Convention, “special meaning” refers to just any meaning that cannot immediately be inferred from conventional language.

Cf. Draft Article 71 as provisionally adopted by the ILC in 1964: “Article 71. Terms having a special meaning. Not withstanding the provisions of paragraph 1 of Article 69, a meaning other than its ordinary meaning may be given to a term if it is established conclusively that the parties intended the term to have that special meaning.” Draft Articles on the Law of Treaties, Report of the International Law Commission covering the work of its sixteenth session, Yearbook of the International Law Commission, 1964, Vol. 2, p. 174, at 199.

Article 32 contains a short list of the means of interpretation that can be used based on this same provision. The list is not exhaustive, which means that in principle, any interpretative data may be used.

See Berglund (2013, pp. 67–68), Bjuvberg (2015a, p. 117) and Dahlberg (2014, p. 252).

While Article 32 provides examples of what shall be regarded as coming within the scope of “supplementary means of interpretation,” it does not provide any exhaustive definition of the concept–the wording of the provision makes this abundantly clear. According to Article 32, consequently, lawyers may apply for the purpose of the interpretation of a treaty: supplementary means of interpretation, “including the preparatory work of the treaty and the circumstances of its conclusion.”

Italics are added.

In the interpretation of tax treaties concluded by Sweden with other states, interpreters may use as a means of interpretation the OECD Model Tax Convention on Income and on Capital with Commentaries, based on either Article 31, paragraph 2, lit. b, or Article 31, paragraph 3, lit. c.

Dahlberg (2003, pp. 152–153).

The OECD works to promote increased harmonization of international tax regulations. As early as 1963, the OECD Committee on Fiscal Affairs (CFA) adopted a Model Tax Convention on Income and on Capital, including a set of extensive Commentaries. Since then, both the Model and the Commentaries have been updated and revised, with increasing frequency.

That the OECD Model and Commentaries can be used to inform interpretation of tax treaties is nowadays taken for granted by Swedish administrative courts.

See, for example, RÅ 1996 ref. 84.

Professor Mattias Dahlberg has suggested two additional grounds that may serve to justify a use of the OECD Model Convention and Commentaries. First, as Dahlberg proposes, “the OECD Commentaries could possibly, at least partly, be considered as belonging to the relevant rules of international law referred to in [Art. 31, para. 3] lit. c.”

See Dahlberg (2003, p. 153): “Enligt min mening skulle kommentaren till OECD:s modellavtal åtminstone i någon mån kunna hänföras till sådana internationella rättsregler som det refereras till i c).”

Second, as Dahlberg submits, the OECD Model and Commentaries could possibly be used as a means of interpretation based on Article 31, paragraph 2, lit. b, which instructs interpreters to consider, for the purpose of the interpretation of a treaty, “any instrument which was made by one or more parties in connexion with the conclusion of the treaty and accepted by the others parties as an instrument related to the treaty.” This suggestion fails to convince for three reasons. First, the OECD Model and Commentaries are not made by any party to a tax treaty but by an organ of an international organization. Second, the OECD Model and Commentaries are not made in connection with the conclusion of a tax treaty, in the sense of Article 31, paragraph 2, lit. b. Third–and perhaps most important–irrespective of the particular tax treaty considered, it will be exceptionally difficult to demonstrate that the OECD Model and Commentaries were accepted by the parties as an instrument related to the treaty. Comparison can be made with treaty reservations, which serve as a classic example of instruments coming within the scope of application of Article 31, paragraph 2, lit. b.

See, for example, McRae (1979, pp. 189-170) and Villiger (1986, p. 344).

There is no corresponding presumption that can be applied to the OECD Model and Commentaries.

Whether an ambulatory or a static approach shall be taken to the interpretation of tax treaties can be stated in the form of one or several general principle.

See Cejie (2010, pp. 93–96) and Dahlberg (2003, pp. 144–145, 151– 152).

The question whether an ambulatory or a static approach shall be taken by interpreters when applying Articles 31– 33 has long been discussed among international lawyers.

See, for example, Bjørge (2014), Higgins (1996, pp. 173–181), Linderfalk (2008b, pp. 109–141) and Waldock (1981, pp. 535–547).

How do you approach open-ended treaty language such as this? Tax-law literature published in Sweden demonstrates attempts to answer this question by the laying down of general principles of shifting contents, directed often as a response to a very particular issue– that concerning which of several consecutive versions of the Commentaries to the OECD Model Convention should be used for the purpose of the interpretation of a tax treaty. For example, in line with express recommendations of the CFA, literature often makes a distinction between “changes to the Commentaries that are a direct consequence of amendments of articles in the Model Convention,” on the one hand, and “changes and additions to the Commentaries that serve either as clarifications or additions to unchanged articles in the Model,” on the other hand.

See Dahlberg (2003, pp. 144–145).

When the Vienna Convention uses terms that do not provide a direct answer to the question of whether modern or historical language shall be consulted when determining the ordinary meaning of a treaty, for example, this is the result of a deliberate drafting strategy. The parties to the Vienna Convention have decided to leave this matter to be decided by interpreters at their discretion, on a case-by-case basis.

Cf. the ILC Commentary on the Final Draft Articles on the Law of Treaties (n. 29), p. 222, with regard to the provision that then corresponded to final Article 31, paragraph 3, lit. c: “[The Commission] considered that, in any event, the relevance of rules of international law for the interpretation of treaties in any given case was dependent on the intentions of the parties, and that to attempt to formulate a rule covering comprehensively the temporal element would present difficulties. It further considered that correct application of the temporal element would normally be indicated by interpretation of the term in good faith.”

Articles 31–33 do not allow consideration of work done within Sweden in the course of preparing for incorporation of tax treaties into domestic law.

See Cejie (2010, p. 89) and Sallander (2013, p. 55).

Because Article 32 of the VCLT defines the concept of “supplementary means of interpretation” by bluntly and generically referring back to whatever customary rules of treaty interpretation exist at each and every moment of interpretation, it can be of interest to know that international courts and tribunals sometimes use as supplementary means of interpretation preparatory work produced in the course of domestic law-making processes.

See, for example, Case Concerning Oil Platforms (Islamic Republic of Iran v.United States of America), PreliminaryObjections, Judgment of 12 December 1996, ICJ Reports 1996 (II), p. 803, at 814–815, para. 29; Beagle Channel Arbitration (Argentina v. Chile), Award of 18 February 1977, International Law Reports, Vol. 52, p. 93, at 186;Maritime Delimitation in the Area between Greenland (Denmark v. Norway), Judgment of 14 June 1993, ICJ Reports 1993, p. 51, para. 29; Guinea – Guinea- Bissau Maritime Delimitation (n. 49), p. 669, para. 70.

As traditionally conceived, law consists of many different branches: civil law, constitutional law, penal law, procedural law, international law, and so on. It is appropriate to think of these different branches of law as different communities of practice. Different branches of law recognize the relevance of partly different sets of data and use partly different methodologies. Similarly, each branch of law has a partly different ethos–if you specialise in criminal law, for example, you will take a partly different perspective on legal data than will civil law colleagues; you will probably also adopt a partly different set of fundamental assumptions. At the same time, of course, branches of law are not tightly separated units, considering what they purport to govern and aim to achieve. As lawyers sometimes tend to forget, the habit of referring to the law that courts and administrative authorities bring to bear on the resolution of disputes in terms such as “criminal law” or “civil law” is merely a way of simplifying structured legal thinking. In reality, legal elements interact in the sense that often any proper understanding of, say, criminal law elements presupposes an understanding of elements traditionally identified with other branches of law. This is why, occasionally, the ambition of legal scholars to understand law and legally relevant data gives them reason to think across branches. Accepting the identity of tax law and international law as different communities of practice, when tax-law research is brought to bear on international law, as for example, when it raises issues of treaty interpretation, that research is cross-disciplinary in nature. It is not terribly different from legal research drawing heavily on non legal disciplines such as linguistics, psychology, sociology, or feminist theory. I think it is important that this nature of the research is fully acknowledged, for the sake of its success. Thinking across disciplines can certainly be a very good way of bringing about new knowledge. At the same time, as thinking will have to be done in a context that researchers are not perfectly at home with, it can be terribly challenging. Cross-disciplinary research, to be truly successful, requires a willingness and an ability of the scholar performing it to adapt to a different mind-set; it requires at least some appreciation of data, with which previously he or she may have been wholly unfamiliar. In the case of taxlaw research drawing heavily on the international law on treaty interpretation, a successful outcome assumes more particularly that scholars allow themselves necessary time to fully assimilate both the theory of treaty interpretation and the rich practice developed by international judicial bodies over the past 50 or so years. This observation brings me back to my earlier call for cross-disciplinary dialogue. As I see it, tax-law scholars interested in learning more about tax treaty interpretation have clear reason to collaborate with colleagues specializing in international law. Conversely, of course, international law scholars interested in learning more about treaty interpretation in a tax-law context have reason to consult with tax-law expertise. The reason for such new initiatives lies partly in the virtues of cross-disciplinary research and partly in the virtues of dialogue. Consequently, cross-disciplinary dialogue will not only help to bring the discussion on tax treaty interpretation entertained in the camp of Swedish tax-law expertise on a more constructive course. In an increasingly perplexing legal universe characterized by rapid developments, bringing more or less daily a new set of issues to the researchers’ agenda, crossdisciplinary dialogue will also be cost-efficient. It will bring new knowledge at the minimum cost or effort.