Argumentation in the Interpretation of Statutory Law and International Law: Not Ejusdem Generis

: This contribution bridges three ﬁelds—pragmatics, argumentation, and law. Arguments can be seen as the verbal formulation of inferences that articulate justiﬁcatory relationships, meaning that behind every argument is at least one argumentative inference. As an argumentative activity and verbal practice, legal discourse has gaps to be ﬁlled by pragmatic inference. Neo- and post-Gricean frameworks can thus tentatively be used for its analysis. Based on these frameworks, this contribution asks whether argumentation in the interpretation of statutory law is the same as in international law. More precisely, it looks at judges’ legal interpretations, which function as justifying arguments because they are constrained by rules/canons of interpretation. It is shown that neither a pragma-dialectical hierarchy of statutory canons nor a hierarchy of related presumptions carries over to international law where there is no such hierarchy.


Introduction
This contribution bridges three fields-pragmatics, argumentation, and law-thereby adding to the growing literature on this threefold interface (e.g., Feteris 2017;Macagno et al. 2018;Skoczeń 2013;Walton et al. 2021;Walton et al. 2016). From this perspective, "arguments can be seen as the verbal formulation of (cognitive) inferences which articulate various kinds of justificatory relationships" (Oswald et al. 2018, p. 11). This means that behind every argument is at least one argumentative inference (ibid., p. 11). Many researchers agree that the law has gaps that need to be filled by pragmatic inference (Capone 2015, p. 387;see also, e.g., Carston 2013;Giltrow and Stein 2017;Skoczeń 2019).
Several authors in the field of legal interpretation and argumentation conceive of the law as an "argumentative activity" (Feteris 2017, p. 18). Argumentation is typically a verbal practice; reliance on linguistic or pragmatic approaches is thus, in principle, legitimated (Oswald et al. 2018, p. 11). In other words, since argumentation is one particular type of communication, neo-Gricean (Skoczeń 2019) and post-Gricean frameworks can tentatively be used for pragmatic (Carston 2013;Smolka and Pirker 2016) and rhetorical analysis (Oswald 2007(Oswald , 2016. This contribution takes a descriptive perspective and looks at argument acceptability, effectiveness, and persuasiveness (Oswald et al. 2018, p. 11) in legal argumentation. In this context, it assesses the merits of different pragmatic approaches to argumentation in legal interpretation, identifies problems with neo-Gricean approaches in international law, and suggests a post-Gricean account as an alternative.
The contribution compares argumentation in statutory law-with an illustrative focus on U.S. law-to international law because the latter is often either neglected in favor of the former (see e.g., Carston 2013;Feteris 2008;Macagno et al. 2018;Skoczeń 2019) or viewed through a national law lens (Slocum and Wong 2021). A focus on particular fields of law is important because different fields use specific methods of interpretation (Feteris 2017, p. 14). There are, however, two reasons why the focus of this contribution cannot be entirely reduced to specific jurisdictions or courts. First, this contribution assesses two prominent least from a theoretical normative standpoint. Due to its more ambiguous nature (Tumonis 2012, pp. 132-33), international law might be better modeled by a relevance-theoretic framework than a neo-Gricean framework.
The most important insight this contribution delivers is-as its title indicates-that prominent pragmatic theories of argumentation do not neatly apply to international law. Secondary to this finding is that this particular field of law might, therefore, be better modeled by a different pragmatic theory. For reasons of space, a few principled considerations of why the prominent post-Gricean framework of relevance theory (RT) appears more suitable in this context must suffice. RT claims that (neo-)Gricean maxims are somewhat superficial manifestations of the twofold relevance-theoretic principle of "relevance", that is, first, the claim that human cognition is oriented towards maximizing relevance, and second, that utterances create expectations of optimal relevance (Carston 2013, pp. 16, 28). Optimal relevance refers to "an implicit guarantee the utterance is the most relevant one the speaker could have produced, given her abilities and preferences, and that it is at least relevant enough to be worth processing" (ibid., p. 28). As the rules of interpretation appear to be plausibly translatable into (neo-)Gricean maxims, it appears possible that they may be integrated into RT as additional, institution-specific (or context-specific) constraints on relevance and interpretation.
(Neo-)Gricean approaches do not aim to provide an account of "the processes of online utterance comprehension", but to offer principles or maxims to account for utterance comprehension (Carston 2005, p. 305). RT, by contrast, aims to build a psychologically plausible (Wilson 2017, p. 81) pragmatic theory of meaning and utterance interpretation in context (Sperber and Origgi 2012, p. 331). (Neo-)Griceans and post-Griceans agree that the utterance comprehension is "achieved by means of (defeasible) inferential processes, which are constrained, but not determined, by the linguistic evidence" (Carston 2013, p. 13). However, while prominent neo-Griceans claim that, although defeasible, inferences as to what an utterance means are drawn by default-or in default contexts (Carston 2013, p. 13;Horn 2009, p. 22), RT claims that there are no defaults: every communicative situation is particularized (Noveck 2018, p. 27); that is, the drawing of inferences is always contextsensitive (for an example, see Section 3), guided by the principle of relevance. 4 Which maxim-or rule of interpretation in Article 31(1) VCLT-is applied thus depends not on any hierarchy or defaults, but on its relevance (in RT's technical sense) to the case.

Rules of Interpretation in International Law-Is There a Hierarchy?
There has been growing interest in legal argumentation, not least of all because of the peculiar role and tasks of the judge in legal decision making and justification. Since the 20th century, it has been recognized that the legislator (or international lawmaker) cannot foresee all possible cases and new developments in society, and will thus generally formulate general rules. As a result, legal rules are considered to have an open texture, and they can be indeterminate in a given case. Therefore, a judge's legal reasoning cannot be characterized as mere subsumption and the drawing of a syllogistic conclusion, but rather as the reasoned solution of interpretation problems in applying legal rules (Feteris and Kloosterhuis 2009, pp. 307-8). The mere fact that the judge's arguments or, in other words, the rules of interpretation, are part of a legal system and thus authorized by the conventions of this system cannot suffice as a justification for their use. If this were the case, it would be almost impossible to bring legally relevant criticism against a judge's legal reasoning. Instead, interpretative arguments may be viewed as more or less appropriate (or persuasive) ways to achieve legal determinations. Such arguments can be assessed according to the outcomes that are obtained through their use, relative to the legal and social values at stake. Interpretative arguments may be supported by reasons and subject to criticism, both of which may, in turn, be relevant to the legal use of such arguments (Walton et al. 2021, p. 50).
The role of the judge varies in different forms of legal procedure (Feteris 2012, p. 235). This contribution concentrates on general public international law, that is, the field of international law in general, which designates "the set of norms resulting from legal acts that govern interstate relations" (Ammann 2020, p. 58), rather than any specific subfield such as international criminal law, and compares it with statutory interpretation with a focus on U.S. law. A judge in public international law will normally be one of a group of judges belonging to an international court-such as the International Court of Justice (ICJ) or the European Court of Justice (ECJ)-or an ad hoc arbitral tribunal that typically decides by predetermined majorities. Of course, both the structure of international courts and the structure of international legal orders in which these courts operate differ from one another. For instance, the EU legal order is a highly integrated order of supranational law, which creates rights and obligations for individuals as well as member states, whereas the UN legal order, in which the ICJ operates, functions more like general public international law, which is more focused on rights and obligations of states. While the ECJ generally follows the VCLT (to the extent that it represents customary international law; see below), it "has rejected the proposition that the VCLT applies to the EU's founding treaties"; its use of the VCLT is thus only comparable with that of other international courts-and is thus only of interest to this contribution-with regard to international agreements or treaties (and not with regard to EU Treaties or EU law) (Odermatt 2015, p. 122). This contribution's perspective is similarly applicable to other regional integration courts, such as the Court of the EAEU (EAEU Court).
Two more examples illustrate the comparability of the VCLT application between international courts despite their differences. The first example concerns the EAEU Court. Founded in 2015, its mission is to ensure the uniform interpretation and application of EAEU law (Diyachenko and Entin 2017, pp. 54-55). In a similar fashion to the ECJ in the past when the ECJ deemed traditional methods of interpretation (i.e., those of classical international courts) insufficient, the EAEU Court must now develop its own methods of interpretation (Diyachenko 2019, pp. 77-78). Notwithstanding this need, according to its Statute the Court must apply customary international law, that is, the relevant rules of the VCLT (ibid., p. 80). The second example concerns the ECJ. At first glance, the ECJ's use of the VCLT rules appears to strictly observe international law (Odermatt 2015, p. 122). However, objections might be raised as to what the two courts do in practice. The EAEU Court's practice shows that it most often uses the literal, or textual, method of treaty interpretation (see below) (Diyachenko 2019, p. 81). The ECJ's approach to the application of the VCLT is, in turn, highly influenced by its approach to EU law. Its emphasis on examining the "object and purpose" of a treaty (see below) at the expense of other methods of treaty interpretation "mirrors the approach in EU law that favours more teleological reasoning" (Odermatt 2015, p. 122). This contribution is, however, only concerned with the application of the rules as they are enshrined in the VCLT-a sociological analysis of how court composition or functioning influence the application of the VCLT, or even lead to an occasional divergence or departure from it, would be beyond its scope.
In the introduction, the relating communication was described as being twofold: between the legislator or international lawmaker and the court or judge, and between the judge and a legal audience consisting of the legal parties, other lawyers and judges, and the legal community in general. However, international law has no legislature (only state and international organizations creating, e.g., treaties, i.e., the international lawmaker) and there are fewer checks and balances on international courts than on domestic ones (Tumonis 2012, p. 132). This distinction is much less applicable in the context of European Union law with its considerably higher degree of institutionalization.
A judge's interpretation of a given treaty justifies the way he or she applies it in a given case, and he or she, therefore, has to justify this interpretation by appropriate (or convincing or persuasive) arguments (MacCormick and Summers 1991a, p. 511). In legal justification, a judge must also show that the interpretation/ruling/decision is consistent with certain legal norms, such as treaty provisions, and is supported by certain legal values and principles that have preferably been authoritatively stated by courts or tribunals (cf. for national law e.g., MacCormick 2005).
In the field of international law, authoritative rules for treaty interpretation have been provided by the 1969 VCLT in its now widely accepted Articles 31 and 32 (Klingler et al. 2019, p. xxv). According to Article 31(1), a treaty "shall be interpreted 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". The three main elements are the text, its context, and the object and purpose of the treaty in question. The title of Article 31 emphasizes that the process is a unity-a single, closely integrated rule (Yearbook of the ILC 1966 II, pp. 219-20). 5 This "general rule" refers to a holistic approach and does not establish any hierarchical or chronological order (Dörr 2012, p. 541).
The textual element consists in looking for the "ordinary meaning" of a term (Aust 2013, p. 209), although it is generally admitted that there typically is no such thing as one single ordinary meaning (Gardiner 2010, p. 480). Interpreting agents often use dictionaries or specialist books to find a definition (ibid., p. 164). The contextual element is defined in Article 31(2) and, to put it very briefly, requires an interpreter to take into account the treaty as a whole with the preamble and the protocols, its systematic structure (Dörr 2012, pp. 534, 535), certain agreements relating to the treaty, and certain instruments made in connection with the treaty (Aust 2013, pp. 210-12). According to the object and purpose of a treaty, the interpreting agent must assess the aims of the treaty, for example, by looking at the title, the preamble, or special clauses on the objectives of the treaty (Dörr 2012, p. 546) and attempt to promote them during the interpretation of the treaty's terms (Gardiner 2010, p. 190). Article 32 states that if the interpretation process based on Article 31 leaves the meaning of a term "ambiguous or obscure" or leads to a "manifestly absurd or unreasonable" result, recourse may be had to supplementary means of interpretation "including" the preparatory work of the treaty or the circumstances of the treaty's conclusion. Article 32 thus establishes a relationship between its elements and those of Article 31, that is, a hierarchy in which Article 32 is subordinate to Article 31.
Articles 31 to 33 VCLT (Article 33 relates to treaties authenticated in several languages, which will not be discussed due to lack of space) are considered to reflect the general rules of customary international law. This means that international courts and tribunals apply them even when the parties in dispute have not ratified the VCLT. The rules also apply to any written texts establishing international legal norms, for example, resolutions of international organizations or other soft law instruments (Pellet 2019, pp. 5-6). The elements of these articles are binding, even though they leave a wide margin of appreciation to the interpreter (ibid., pp. 6-7).
The VCLT makes no reference to canons of interpretation, as they are referred to in other fields of legal interpretation such as statutory or contractual interpretation and which were relied on in treaty interpretation prior to the VCLT's adoption. While the VCLT gives priority to the rules enshrined in its Articles 31 to 33, the canons may still be referred to and provide guidance (Klingler et al. 2019, xxv). However, international tribunals rely on them without resorting to the underlying domestic origins (Waibel 2019, p. 29). Examples of canons are expressio unius est exclusio alterius according to which the express mention of a term or terms implies the exclusion of others (Klingler 2019, p. 74), and ejusdem generis according to which a general word is to be interpreted by reference to surrounding specific words (Baetens 2019, p. 145).
Although canons are not expressly mentioned, it can be argued that some of them may be implicitly encompassed or referred to, especially by Article 31(3)(c), according to which the interpreter shall take into account "any relevant rules of international law applicable", and by Article 32, according to which "supplementary means of interpretation" may be used (Pellet 2019, pp. 7-8). It must, however, be noted that while canons may subsidiarily assist the interpretation based on the provisions of the VCLT, they cannot be applied to reverse or undermine such interpretation (Dörr 2012, p. 242). Canons are thus in a lower normative category than-or hierarchically subordinate to-rules (Pellet 2019, p. 10). The canons which should be considered predominant cannot be decided in general, but only in context (Yearbook of the ILC 2006 II Part Two, p. 178), 6 i.e., on a case-by-case basis.
There is thus a hierarchy between rules of interpretation in international law. But is this particular hierarchy reflected by the hierarchy between legal rules of interpretation, as modeled by PD, one of the most influential representatives of argumentation theory (Oswald 2007, p. 179)? 7 PD considers legal argumentation as a specific institutionalized form of argumentation, and legal procedure as a specific institutionalized form of a critical argumentative discussion in which the parties and the judge use arguments to resolve a dispute or difference of opinion about the application of a particular legal rule in a specific case (Feteris 2017, p. 221;Feteris and Kloosterhuis 2009, p. 322). PD highlights that in adjudication, it is not the parties themselves who terminate their dispute but the judge as a third party who will make a reasoned decision on who is wrong and who is right according to a set of rules. The parties, in turn, readjust their roles from trying to persuade each other to trying to convince the judge (Van Eemeren and Houtlosser 2006, p. 384). The judge anticipates the critical reactions which may be put forward by institutional antagonists: the party who may want to appeal the decision, the judge in an appeal procedure (Feteris 2017, p. 219)-which, in international law, is quite rare-or the legal community as a whole.
Judges have discretion as to how to apply the law. They, therefore, normally account for the decision in the form of a justification-although, unlike PD claims (cf. ibid., p. 224), they are not obliged to do so. PD describes the decision as a speech act in the form of an assertive declaration, and the justification as a complex speech act of argumentation in which the acceptability of the decision is defended on the basis of the decisions about the facts and the applicable law (ibid., p. 224). The fact that judges are not obliged to justify decisions explains why, as PD notes, judges' reasons are not always explicit, clear, and well ordered, and they may not give an account of all considerations underlying the decision which would be necessary for a complete justification (e.g., because they consider it obvious), or may adduce arguments obiter dicta that are superfluous to the justification of their decision (Feteris and Kloosterhuis 2009, p. 308). This fact can also be attributed to the strategic nature of legal argumentation, or what PD calls strategic maneuvering. In the presentation of the justification for their decision, judges often try to present their decision as a self-evident result of the application of the law to the facts of the case. However, this application is often less self-evident than it is presented to be (ibid., p. 326). Judges are thus trying to reconcile certain standards of reasonableness with their own rhetorical goals, that is, with resolving the dispute in their favor or, more precisely, according to their own points of view (Van Eemeren and Houtlosser 1999, p. 481) about the law as it ought to be (Kloosterhuis 2009).
PD reconstructs the judge's often implicit justification via prototypical argumentative patterns (Feteris 2017, p. 225 ff.), that is, specific types of legal argument schemes that are characteristic for the resolution of different types of legal disputes in different fields of law (ibid., p. 352). PD introduces a hierarchy of rules of interpretation which are mainly based, among others, on MacCormick and Summers (1991b), who inferred a "universalist" (Feteris 2017, p. 12) preferred order for the use of various forms of interpretative arguments from the practices of legal justification in nine countries (Summers 1991b, p. 3) with comparable legal cultures (on the similarities between common and civil law see Summers and Taruffo 1991, pp. 508-9;Müller 2000). The hierarchy of legal arguments adopted by PD on this basis is as follows. First, judges are to look for grammatical or linguistic arguments referring to the meaning of the words and expressions used in the rule. If such an argument offers no acceptable solution, a judge may look for systematic arguments referring to the position of the rule in the legal system and its relation to other rules (i.e., context). If systematic arguments do not offer an acceptable solution, a judge can look for teleological-evaluative arguments which refer to the goals (or purposes) and values the rule is intended to realize (Feteris 2016, p. 63; see also Feteris 2017, pp. 12-14;MacCormick and Summers 1991a, p. 531). This hierarchy clearly runs counter to the aforementioned hierarchy in international law in which those three types of arguments, that is, the text, its context, and the object and purpose, form a single, closely integrated rule without any hierarchical or chronological order. While some authors speak of a "logical progression" by which the three elements are to be examined (Aust 2013, p. 208), starting with the text, this does not imply any hierarchy.
It is sometimes noted in PD that the mentioned hierarchy pertains to the interpretation of national laws, for example, that this hierarchy in which the linguistic interpretation rule has the highest position exists "in many legal systems" of common or civil law tradition (Feteris 2017, p. 231); however, this is not always made explicit. For instance, regarding EU law and international law, it is only mentioned that the VCLT leaves courts "a certain latitude in interpreting and applying the law" (Feteris 2017, p. 14), yet the absence of a hierarchy in the VCLT between what PD calls linguistic, systematic, and teleological-evaluative arguments is not discussed. In addition, it is mentioned that the ECJ has developed its own interpretative culture in which the "traditional" rules of interpretation, such as linguistic/grammatical/textual interpretation, do not suffice and, for this reason, "supranational" rules of interpretation are used, such as autonomous and consensus interpretation (ibid., pp. 9-10, 15). However, the latter rules are arguably not "specific" to EU law but exist under other names in other fields of law. It can thus be argued that international law is not taken into consideration in this "universalist" hierarchy. It can also be argued that international law is, in fact, an "important" (ibid., p. 14) field of law that has become so pervasive that it cannot and should not be ignored from the perspective of national law. Moreover, a truly universalist hierarchy of legal rules of interpretation should not content itself with taking into account differences between common and civil law-or limit itself to the interpretational practices of "higher courts" in these traditions (Summers 1991b, p. 2)-and thus take a Western-centric viewpoint, excluding other legal traditions. This critique does not, however, imply that PD's prototypical argumentative patterns which reflect the aforementioned hierarchy cannot be used in international law. It may well be the case that such a hierarchy arises in a given case; the present critique just means that one cannot in the abstract assume such a universalist hierarchy.
This lack of consideration of the rules of interpretation of international law may also be due to the fact that a certain focus on statutory interpretation exists in research on the interface between pragmatics/linguistics and legal interpretation/law (e.g., Busse 2017; Capone and Poggi 2016;Carston 2013;Giltrow and Stein 2017;Poggi 2011Poggi , 2013Skoczeń 2013Skoczeń , 2016Skoczeń , 2019Slocum 2017), although this focus is not as pronounced (see, e.g., Kjaer and Lam 2022). Due to its high degree of institutionalization, EU law may be an exception to this focus (e.g., Solan 2009; Tiersma and Solan 2012; Vogel 2019); it must, however, be noted again that the VCLT applies not to intra-EU law, but only to treaties the EU concludes with third parties.

Can Rules of Interpretation Be Translated into (Neo-)Gricean Maxims?
In Section 2, it was highlighted that interpretation in law can be characterized as essentially argumentative. Since arguments are essentially instances of language use, argumentation and pragmatics are two related fields (Walton et al. 2021, p. 7) and research on this joint interface with the law seems worth pursuing. This section discusses elements of (neo-)Gricean pragmatics but does not primarily look at how utterances are comprehended. Instead, it looks at how a certain understanding of utterance meaning is justified (cf. ibid., p. 10). As discussed, legal arguments which justify an interpretation can be classified into different types. The question in this section is: Can these justificatory argument types be translated into (neo-)Gricean maxims or principles/heuristics?
The idea that such a translation is possible appears to have emerged from Llewellyn's famous critique of canons of construction (i.e., U.S. rules of interpretation, which this contribution focuses on as an illustrative example of statutory interpretation) in relation to the interpretation of statutes (Llewellyn 1950, p. 401). Llewellyn noted that there were typically two opposing maxims on almost every issue; canons were thus not determinate and could be used by judges to justify outcomes that were driven by motivations other than neutral application of the law, namely their own political, cultural, or moral preferences (Murphy 2019, p. 15).
Llewellyn extracted a list of 24 canons from U.S. cases and paired them with 24 canons in opposition. He noted that the canons are a conventional vocabulary of argument (1950, p. 401); that is, shorthand for expressing a complex logical claim (Murphy 2019, pp. 16-17). A case outcome could simply depend on the canon a judge selected from each pairing (ibid., pp. [15][16]. Murphy (2019, pp. 16-17) showed that this exercise can be similarly undertaken for rules, canons, and principles of international law, for instance, opposing ordinary or "plain meaning" and "intentions". 8 Due to Llewellyn's critique of the nature and operation of the canons, U.S. academics and, to a certain extent, U.S. courts have moved toward resolving textual ambiguities within statutes by other means, namely the legislature's intent or purpose on the basis of, in large part, a statute's legislative history (Murphy 2019, pp. 17-18). It must be stressed, however, that the U.S. interpretation in light of the statutory purpose is normally only accepted if ordinary (or technical) meaning arguments cannot be generated due to ambiguity or vagueness (Summers 1991a, p. 441). In other words, favoring strong evidence of legislative intent/purpose over credible ordinary meaning to the contrary is a maxim the U.S. Supreme Court seldom applies or even allows (ibid., p. 439). As indicated in Section 2 and explicitly stated by the Court, this is due to the fact that "general statements of overall purpose contained in legislative reports cannot defeat the specific and clear wording of the statute" (St. Martin Lutheran Church v. South Dakota, 451 U.S. 772 [1981], p. 786). 9 By contrast, the rules in international law may be understood as mediating without favoritism (Yearbook of the ILC 1966 II, pp. 219-20) 10 among the three approaches to treaty interpretation (text, context, and purpose) (Murphy 2019, p. 19)-while the preparatory work of a treaty, which roughly corresponds to a statute's legislative history, is hierarchically subordinate to these three approaches.
Taking up Llewellyn's critique, Miller stressed that the reason such rules of interpretation may stand in tension with each other is simply a matter of competing inferences drawn from the evidence. The evidence may be ambiguous but that does not mean that the inferences themselves are useless (Miller 1990(Miller , p. 1202. Miller showed that insights from Grice's system of maxims-consisting of the cooperative principle which, in turn, generates a set of maxims classed under quantity, quality, relation, and manner and their respective submaxims (Grice 1989)-for interpreting language in conversational settings, i.e., Grice's theory of implicatures, "appear[s] remarkably similar in form to many of the leading maxims of statutory interpretation" (Miller 1990(Miller , p. 1182. Miller also stressed that while there is no perfect equivalence of maxims of interpretation across legal systems, there is sufficient continuity to suggest that the maxims reflect some relative universal principles for interpreting statutes (ibid., p. 1182). Like Gricean maxims, legal maxims in many cases will not yield a single unambiguous result since competing inferences are involved (ibid., p. 1202).
Take the following examples. Expressio unius can be characterized as an instance of the third Gricean submaxim of manner, that is, "be brief". In a list of grouped terms followed by a general term, this particular submaxim would be flouted (i.e., intentionally or blatantly not observed) by the legislature if it intended to convey, for example based on context, that the general term should be understood in its broadest sense because this would make the other words in the list superfluous. Besides the maxim of manner, expressio unius corresponds to the first submaxim of quantity, that is, "make your contribution sufficiently informative" (ibid., pp. 1195-96, 1226-27). Its opposite canon, ejusdem generis, corresponds to the second submaxim of quantity, that is, "do not make your contribution excessively informative" (ibid., pp. 1199-200, 1226). If cats, dogs, and other animals are barred from a public park, can a mounted policewoman enter the park? While plain-or ordinarymeaning would bar her because her horse is an animal, this would also flout the second submaxim of quantity, for if all fauna were covered, "cats" and "dogs" would provide more information than necessary. Compliance with the maxim is re-established if the reference to cats and dogs modifies the term "animals" to only like animals (ibid., p. 1200). Similar to expressio unius, ordinary meaning corresponds to the first submaxim of quantity, as well as to the first and second submaxims of manner, that is, "avoid obscurity" and "avoid ambiguity" (ibid., pp. 1226-27). Plain meaning in U.S. law also refers, albeit inexactly, to a pragmatic process of weighing competing considerations: "the clarity of the statutory language, its consistency with the underlying legislative purposes and whether the costs of resort to extrinsic aids to interpretation (such as legislative history) are likely to outweigh whatever benefits might be realized from such an enterprise" (ibid., p. 1224). Finally, purpose corresponds to the maxim of relation, that is, "do not say anything irrelevant" (ibid., p. 1227) to the purpose of a statute (or treaty).
Several other authors have stated that rules of interpretation (with some provisos or modifications) can also be considered as similar to or instantiations of neo-Griceanespecially Hornian (Horn 1984) and Levinsonian (Levinson 2000)-principles (e.g., Carston 2013; Skoczeń 2019; Slocum 2016); many of them show that Horn's Q-and R-principles are "clearly mirrored by the legal canons expressio unius est exclusio alterius and eiusdem generis, respectively" (Poggi 2020(Poggi , p. 1201. According to Horn's (1984Horn's ( , 1995 account, the principle of sufficiency (Q), "Make your contribution sufficient; say as much as you can (given R)" combines Grice's first quantity submaxim with his first two manner submaxims, while the principle of least effort (R), "Make your contribution necessary; say no more than you must (given Q)" combines Grice's second quantity submaxim with his relation maxim and his third and fourth manner submaxims.
Carston noted that the ordinary meaning rule not only bears considerable similarity to Grice's manner maxim, but also to the neo-Gricean I-principle (which is similar to Horn's Rprinciple) and its heuristic, which licenses hearers to interpret unmarked linguistic expressions in an ordinary stereotypical way (Carston 2013, p. 19;Levinson 2000, p. 37), as well as to the Q-principle and its corresponding heuristic, "What isn't said, isn't the case" (Carston 2013, pp. 13-14;Levinson 2000, p. 31). 11 Carston (2013, p. 16) illustrated that the correct application of the two maxims is highly context sensitive. For instance, in Smith v. United States (508 U.S. 223 [1993]) 12 the neo-Gricean principles lead to exactly contradictory results regarding the ordinary or plain meaning of "use of a firearm": The Q-principle leads to the interpretation of using a firearm as a weapon. By contrast, the I-principle leads to the interpretation that any use of a weapon falls within the scope of the provision (Carston 2013, pp. 21-23; see also discussion in Skoczeń 2019, p. 114). The contradictory result illustrates that while rules of interpretation are general rules for the use of language, they also make use of general terms which themselves require interpretation (Hart 2012). Carston concluded that neo-Gricean principles may be "somewhat superficial manifestations of some other deeper or more general principle underlying all pragmatic inferences", that is, the relevance-theoretic principle of "relevance" (Carston 2013, p. 16; see also Section 5).
Macagno et al., in turn, built on Levinson's (Macagno et al. 2018, pp. 69, 72) and Atlas' neo-Gricean approach in order to develop "a framework of linguistic interpretation within the structure of an inference to best explanation" among legal arguments for or against a given interpretation (Macagno et al. 2018, pp. 69-70) which may be presented when different possible interpretations are available and doubts arise (ibid., p. 74). 13 In legal theory, the term "interpretation" is used in two ways. In a broad sense, it includes both the process and the result of determining the actual meaning of a legal source (see, e.g., Tarello 1980). In a more restricted sense, it only concerns cases in which reasonable doubt or a conflict is raised concerning the prima facie understanding of the meaning of a text. According to Macagno et al. (2012, pp. 64-65), the prima facie meaning is arrived at by attributing a default context or rule to the source statement, that is, an unchallenged presumptive meaning which is very similar to how ordinary language interpretation is modeled by Levinson (and by Horn). The notion of presumption can be linked to the idea that we have preferred or default-or presumptive-pragmatic interpretations or inferences (Levinson 2000, pp. xiii, 1, 372). 14 Since (neo-)Gricean conversational heuristics are defeasible, that is, defeated by stronger assumptions concerning the goal of the cooperative activity, they can be used to capture the presumptions which guide the process of legal interpretation, as well as to calculate and support a given interpretation in a legal interpretive dispute (Macagno et al. 2018, pp. 86-87).
Macagno et al. focused on hard cases in which there was a difference of opinion about the correct interpretation of the rule. Unlike in easy cases, a linguistic argument cannot function as a decisive argument because there are different views with respect to the exact meaning of the rule. The reason a linguistic argument may suffice as a justification in an easy case is that according to the "universalist" preferred order of rules of statutory interpretation mentioned in Section 2, reference to the clear intention of the legislator as it appears from the wordings of the law must be taken as the starting point for the application of the law (unlike in international law). In hard cases, other sources are necessary to establish the intention of the legislator (Feteris 2009, p. 3). In Macagno et al.'s model, the best interpretation is one that best fits both the shared background presumption in the context and the communicative intention attributable to the speaker in the light of "what he has said" (Atlas and Levinson 1981, p. 42). The best interpretation is thus less controversial or less subject to defeat by conflicting propositions in the common ground (Macagno et al. 2018, p. 83). As it is less controversial, the best interpretation is the more informative one and thus corresponds to the pragmatic principle of informativeness (Atlas and Levinson 1981, pp. 40-41) which, in turn, closely corresponds to Levinson's I-heuristic (Birner 2013, p. 83) or, in other words, to the R-principle (cf. Walton et al. 2021, pp. 109, 141). In conclusion, justificatory legal arguments can be translated into (neo-)Gricean maxims/principles/heuristics. It seems that the neo-Gricean idea of a default context or rule may help model prima facie understanding in both easy and hard cases due to the primacy of ordinary meaning in statutory interpretation, whereas a post-Gricean/relevancetheoretic perspective (Sperber and Wilson 1995) might better reflect the fact that which rule of Article 31 VCLT is to be preferred depends on the context.

Based on Their Translation into (Neo-)Gricean Maxims, Can Rules of Interpretation Be Translated into a Hierarchy of Presumptions Which Guide Legal Interpretation?
Section 3 has shown that both neo-Gricean and legal maxims or canons on their own "provide little ground for assessing which interpretation is the best one" (Macagno et al. 2018, p. 90) in hard cases. So what, if anything (apart from the hierarchies described in Section 2) gives legal argument types (or schemes) priority over others (on a case-by-case basis)? 15 This section focuses on Macagno et al.'s framework, as it serves to order interpretive legal arguments (or canons or maxims) hierarchically (ibid., pp. 69-70) and, to this end, "translate[s]" legal arguments of interpretation into argumentation schemes (Macagno et al. 2012, p. 70). 16 Argumentation schemes are based on Toulmin's notion of warrant, that is, "general, hypothetical statements, which can act as bridges, and authorize the sort of step to which our particular argument commits us" (Toulmin 1958, p. 91), and they aim to represent the combination between a semantic principle-for example, classification, cause, consequence, authority-and a type of reasoning-for instance, deductive, inductive, or abductive reasoning. They provide abstract patterns of legal argumentation which represent types of arguments that carry probative weight for supporting or attacking a conclusion but are-in the most typical instances-defeasible (Macagno et al. 2012, p. 68). Argumentation schemes and (neo-)Gricean maxims are similar in that they can be viewed not as providing an interpretation, but rather as providing reasons to support it (Slocum 2015, pp. 203-7;Walton 2002, p. 191).
The argument schemes into which pragmatic maxims/heuristics and their corresponding legal rules of interpretation are translated are largely based on sets of interpretive arguments by MacCormick and Summers (1991b) as well as Tarello (1980), for instance, the analogy or absurdity argument (Macagno et al. 2018, p. 91 ff.;Macagno and Walton 2017, pp. 49-52). For example, plain meaning corresponds to the argument from natural meaning; purpose corresponds to the teleological argument (Macagno et al. 2018, pp. 91-94). Contextual interpretation according to the VCLT appears to correspond to the systemic argument, and, like purpose, it thus corresponds to the Gricean maxim of relation (ibid., pp. 94, 96). Like its national law counterpart-called contextual harmonization in U.S. law (Summers 1991a, pp. 413-14), and which serves as an illustrative example of statutory interpretation in this contribution,-contextual interpretation according to the VCLT may also correspond to a contrario, a simili, and the economic argument (Macagno et al. 2018, p. 98), and thus with a contrario to the first quantity submaxim and the third manner submaxim, with a simili to the second quality submaxim, and with the economic argument to the second quality submaxim and the third manner submaxim. 17 Macagno et al. stress that all interpretive arguments may be rebutted by contrary arguments based on the same maxims or rules of interpretation and supported by contrary contextual evidence (ibid., p. 98). The success of one argument over another is thus not a matter of the nature of the legal rule of interpretation itself, but of the whole argumentation brought for and against the interpretative conclusions (Walton et al. 2021, p. 151). Macagno et al. (2018, p. 86; see also, e.g., Macagno and Walton 2013) state that conversational maxims/heuristics are defeasible in the sense that they are defeated by stronger assumptions relating to the goal of the cooperative activity; therefore, Gricean maxims need to be ordered and analyzed together with other types of presumptions governing conversation, the foremost of which is the purpose of the dialogue in the sense of Grice's definition of the cooperative principle: "Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged" (Grice 1975, p. 45, emphasis added). The authors note that Grice's notion of "direction" represents a dialogical intention to which the meaning of single speech acts needs to be linked. Since the purpose of a statement of law is often retrieved by taking into account "the whole co-text of the provision in which it occurred", the authors conclude, first, that the mechanism of retrieval of the speakers' intention in the Gricean sense applies to statutory interpretation, 18 and second, that the "purpose of the law" is a presumption to which all other presumptions must be subordinated (Macagno et al. 2018, p. 106).
Purpose which is thus grounded on a meta-presumption is the first-or highest or most fundamental (ibid., p. 103)-presumption (level 0). This presumption type can be called pragmatic as it connects the illocutionary force of a speech act to its presumed intention. The second type (level 1) refers to the conventional presumptive meaning of lexical items (ibid., p. 107). The third type (level 2) concerns expectations about relations between facts or events that can be used to interpret a specific content or action (ibid., p. 107), for example, "Weapons are usually used to commit violent crimes" (ibid., p. 109). The last type (level 3) consists of contextual "rebuttable interpretive assumptions", for example, "Usually words are used with the same meaning within a statute", and co-textual information, that is, other statements of law (ibid., p. 109). The four levels are contextually ranked according to their respective possibilities of being subject to default-that is, how easily they can be rejected by relying on more contextual information (ibid., p. 108)-depending on the accessibility of information (e.g., mutual information concerning specific contexts is more accessible than encyclopedic information about events/facts) (ibid., p. 107). Levels 1 to 3 are ordered based on their degree of context dependence (ibid., p. 109).
Macagno et al. claim that arguments from purpose, that is, teleological arguments, are more complex as they presuppose linguistic, co-textual/contextual, and factual presumptions yet directly refer to overarching pragmatic presumptions (ibid., p. 101). 19 According to the authors, presumptions which determine the legal purpose are more general/basic or fundamental to the legal system (ibid., p. 97), for example, the "law needs to be understood, address relevant social problems, etc." (ibid., p. 109). However, according to MacCormick and Summers (1991a, p. 538), on whom the authors rely, it is a misunderstanding that basic values (or presumptions) are directly deployed only in arguments of the teleological/evaluative type, as the justificatory force of all argument types depends on fundamental legal-constitutional and political values; there should thus be an equilibrium between the different values expressed by different arguments. For instance, natural meaning arguments express the values of democracy, separation of powers, rule of law (ibid., p. 534), and systemic-not purposive-arguments appear to primarily express values of intelligibility of the law to citizens (ibid., p. 535).
MacCormick and Summers also stress that it is not necessary in every interpretational situation to go deeply into purposes (ibid., p. 540), and that in some legal systems, proceeding to the level of purpose or teleological argumentation is not appropriate unless considerable weight is ascribed to the purposes or values at stake (ibid., p. 531). In civil law systems, the notion of literal meaning is frequently evoked as a self-sufficient entity at the theoretical level as a guardian of the law's certainty and predictability (Skoczeń 2016, p. 627). Based on MacCormick and Summers' hierarchy discussed in Section 2, the ordinary language scheme should be taken as the default setting unless there are superior reasons to interpret the expression as fitting one of the other schemes. Macagno, Sartor, and Walton, however, consider all interpretative canons as defaultive, 20 even if claims based on some canons can only be raised under specific conditions (Walton et al. 2016, pp. 62-63;Walton et al. 2021, p. 290).
The goal of these authors is for their framework to "transcend [ . . . ] the narrower jurisdictional issues" and to thus be applicable on a "worldwide" basis (Walton et al. 2021, p. 6), that is, to be applicable to statutory interpretation in both common and civil law (Macagno and Walton 2017, p. 47). The authors' other goal is to test their framework on arguments used in "problem cases" of statutory interpretation (Walton et al. 2016, p. 52), declaring, for instance, the aforementioned Smith v. United States 21 case a "crucial case for analyzing the presumptions and levels of presumptions" underlying statutory interpretation which "can be applied to the analysis of other cases" (Macagno et al. 2018, p. 97). The issue with such problem cases seems to be that if purpose "overcomes" other presumption types, especially plain meaning (ibid., p. 104), this is due to an exception. As discussed in Section 3, there are only "few cases in which the U.S. Supreme Court makes an exception to the standard meaning of a statutory rule" in light of its purpose (Feteris 2016, p. 62), which the authors do not seem to discuss apart from noting that all mentioned rules of interpretation are defaultive-or defeasible or subject to exceptions (Sartor 2009, p. 21). The other issue with problem cases is that they are often quoted or are (in)famous precisely because the judges' interpretation/decision is perceived as controversial, unlike the authors appear to claim. For instance, according to Macagno et al., in Smith 22 the purpose of avoiding drug-related crimes and in particular any association between drug selling and weapons "easily" undercuts the linguistic presumption that "to use a firearm" means "to use a firearm as a weapon"; the latter should, therefore, be interpreted as any firearm use, as long as it is related to drug trafficking (Macagno et al. 2018, pp. 98, 109). Contrary to the authors' assertion, the Court's decision may lead to a high degree of legal uncertainty-just think of the average person reading this law and having to predict this meaning to be able to abide by it.
Another problem with purpose is that it invites the-"hotly disputed"-question of "whose purpose should govern, and what evidence should be consulted as to purpose" (MacCormick and Summers 1991a, p. 519). The disputing parties may advance contrary purposes between which a judge has to decide. Macagno et al.'s hierarchy of presumptions offers a hierarchy of competing purposes in the sense that, first, the more global/basic purposive or pragmatic presumptions which are directly related to the basic principles of the relationship between the lawmaker and the citizens are the strongest presumptions (Macagno et al. 2018, p. 111), and, second, presumptions relative to a more specific context (factual/level 2 and contextual/level 3) provide the strongest grounds for reconstructing a specific pragmatic presumption (ibid., p. 110). In Smith, 23 Macagno et al. appear to argue that the context of the statute (drug trafficking) overcomes more basic pragmatic or purposive presumptions, that is, the law cannot be unjust or absurd (ibid., p. 98) and should be understood by citizens (ibid., p. 101); or, at least, they appear to argue that the defendant's argument grounded on the former basic presumption should not count as such (ibid., p. 98). The question appears to be one of conflicting interpretations of what, for example, "unjust" means. It thus seems that the proposed presumption hierarchy does not lead to clear-cut results in controversial cases. 24 The authors' focus is on statutory interpretation. From the perspective of the present contribution, it is not clear that such a hierarchy of presumptions is applicable to international law, in which at the theoretical level, the means of interpretation of Article 31 VCLT are considered a "single combined operation" in which all elements applicable to a case are "thrown into the crucible, and their interaction [will] give the legally relevant interpretation" (Yearbook of the ILC 1966 II, pp. 219-20). 25,26 Thus, the proposed hierarchy of presumptions might well reflect what judges in international law may sometimes de facto do, but it appears difficult to reconcile with the existing normative legal categories.

An International Law Example and a Tentative Conclusion
To give an example of a hard case in international law, let us briefly consider the judges' ruling on preliminary objections in the ICJ case Ukraine v. Russian Federation (2019). 27,28 The judges had to decide whether the procedural preconditions of Article 22 of the International Convention on the Elimination of All Forms of Racial Discrimination (CERD) were met in order to decide whether the Court had jurisdiction. The Court sided with Ukraine's interpretation that the ordinary meaning of "or" in the phrase "not settled by negotiation or by the procedures expressly provided for in [CERD]" is disjunctive, not conjunctive (although from a pragmatics perspective, one might argue that inclusive interpretations are preferred or stronger in negative contexts; see Noveck et al. 2002, p. 304). The disjunctive reading was supported by the Court's interpretation in light of the context-which "does not support" a cumulative reading-and "also" of the object and purpose of CERD, that is, "to eradicate racial discrimination effectively and promptly", which would be more difficult to achieve if the procedural preconditions under Article 22 were cumulative, that is, under a conjunctive reading of "or". It may be argued, however, that this purpose may not necessarily be more quickly achieved by a court procedure than by other negotiation/conciliation procedures provided for in CERD and that one may give more weight to the ordinary meaning of the provision at issue. For instance, Russia pointed out in its preliminary objections (2018, pp. 188-89) 29 that in another international law case, the ordinary meaning of "or" under negation was interpreted as cumulative/conjunctive. Russia's interpretation of Article 22 CERD was seconded in Qatar v. United Arab Emirates by the U.A.E.'s preliminary objections (2019, p. 77 ff.) 30 based on the same interpretation by five judges' joint dissenting opinion (2011, p. 156) 31 in Georgia v. Russian Federation.
This example serves to illustrate the points made in Sections 2-4. First, while the "universalist" hierarchy of the three rules of interpretation focused on in this contribution does not in theory apply to international law, it has arisen in this case-yet no critical legal discussion rule would have been violated if it had not; neither has strategic maneuvering with linguistic argumentation derailed in this case because the standards of reasonableness and conditions for an acceptable use of such argumentation appear to be met (cf . Feteris 2009;Van Eemeren and Houtlosser 2006). Second, while it appears in this case that purpose, with the help of context, has overcome ordinary meaning, there is no such theoretical notion of "overcoming" in international law in this context, as ordinary meaning, context, and purpose are considered a single general rule. It thus seems more difficult than in national law to view presumptions which determine the legal purpose as more general/fundamental-perhaps pointing to a risk of making "superficial comparisons" or transfers of legal concepts to other areas of argumentation (Gama 2017, p. 557) or pragmatics. This contribution suggests that arguments in national and international law are-at least in theory and despite the fact that both can be translated into (neo-)Gricean maxims-not exactly the same. In other words, this contribution has assessed the merits of different pragmatic approaches to argumentation in legal interpretation, identified problems with neo-Gricean approaches in international law, and suggested a post-Gricean account as an alternative. According to RT, the drawing of inferences is always context-sensitive, guided by the principle of relevance. Which maxim-or rule of interpretation in Article 31(1) VCLT-is applied thus depends not on any hierarchy or defaults, but on its relevance (in RT's technical sense) to the case. More generally, international law should not be overlooked by argumentation theory despite its focus on statutory interpretation, as states must use the interpretative methods of international law to honor their international obligations (Ammann 2020, pp. 191-92).
To pull the strands of pragmatics, argumentation, and legal interpretation further together, one might attempt to explain the perception of legal decisions as controversial by an alternative to strategic maneuvering. Under a post-Gricean/relevance-theoretic perspective, the rhetorical persuasiveness of legal interpretations/arguments may be grounded on how far-fetched they are given the explicitly stated meaning (see Smolka and Pirker 2021). Since which rule of interpretation will prevail likely depends on the preferred outcome, which likely depends on policy (Tumonis 2012, p. 133), cultural, or moral preferences (Murphy 2019, p. 15), it may be interesting under a relevance-theoretic perspective to experimentally test how strong-or contextually relevant (Oswald 2016, pp. 25, 30)-such considerations are (see Pirker and Skoczeń 2022 Due to its focus on international law, this contribution uses terms such as "international lawmaker" and "applicable legal rules" alongside corresponding national law terms such as "legislator" and "enacted regulation" (cf. terms used in Skoczeń 2013, p. 53). 3 Legal arguments in judicial contexts can be divided into evidentiary arguments concerning the reconstruction of the facts relevant to the case and interpretive arguments concerning the applicable legal norms, that is, rules, standards, or principles, stated in authoritative texts or precedents (Canale and Tuzet 2019). This contribution focuses on the latter. 4 The relevance-theoretic comprehension procedure follows a cost-benefit logic, with processing effort as costs and cognitive effects-e.g., a change to the set of assumptions an individual entertains-as benefits. Put very simply, the greater the effects-and the smaller the effort, the greater the relevance. Comprehension, or interpretation, follows a path of least effort (Wilson 2003, p. 282 ff). It must be noted that researchers in pragmatics and law express reservations against the use of RT because of "the kind of conscious effortful scrutiny a legal text may be subjected to" (Carston 2013, p. 32); legal interpretation is, therefore, not oriented at processing that requires least effort (Skoczeń 2019, p. 62). More generally, "legal pragmatics is not interested in the kind of psychological approach proposed by RT" (ibid.) due to its conception of the notion of intention, which merits discussion in a separate contribution. For lack of space, suffice it to say that the central role of subjective mental states in the form of intentions may be difficult to map onto legal notions of intentions in the sense that, in legal interpretation, it may be preferred to substitute such an internal intention with external, public cues found in the context (ibid., p. 135). For arguments against this reservation as well as other common reservations against the applicability of pragmatic theories to legal interpretation, see Pirker and Smolka (2017). The intention rule of interpretation merits discussion (though not explicitly mentioned in Articles 31 to 33 VCLT); it is, however, left aside due to lack of space. 10 See note 5 above. 11 Carston appears to conflate Horn's and Levinson's accounts, e.g., Horn's Q-and R-principle and Levinson's Q-and I-heuristic. She notes that there are small differences in the formulation of their heuristics and other more substantial differences in outlook-Levinson believes in "default" interpretations, Horn does not-but that none of these are significant for their application to legal interpretation (Carston 2013, p. 13). Due to lack of space for discussion, this position is adopted in this contribution. 12 Smith v. United States, (508 U.S. 223 [1993]). 13 According to Allott (2010, p. 17), an inference to the best explanation is a variety of abductive inference, which, unlike deductive inference, is non-demonstrative or, in other words, uncertain, and open to revision. In abductive inference, an individual thus "reasons from a fact that is to be explained to an explanation for that fact"-and this explanation will be the one that, based on the individual's knowledge of the speaker, the conversational situation, and the world more generally (ibid., p. 94), appears to be the best, although possibly mistaken, explanation to this individual (ibid., p. 17). See note 12 above. 23 See note 12 above. 24 For this reason, Skoczeń (2019) argued that a game-theoretic approach to hard cases is more descriptively accurate. 25 See note 5 above. 26 Ordinary meaning, context, and purpose according to the VCLT cannot "be ordered in hierarchies depending on the specific legal context" (Walton et al. 2021, p. 319) because there is no hierarchy among them. For a linguistic analysis of U.S. law cases, see e.g., Solan (1993). Among other things, Solan discusses examples with "or" (ibid., pp. 45-55). It must be noted that in U.S. law there are linguistic rules covering the use of "or" (as well as "and"). The linguistic rules amount to subordinating ordinary meaning to purpose (ibid., p. 45) (i.e., an exception to the primacy of plain/ordinary meaning), although their "application is very difficult to explain in terms independent of the results achieved" (ibid., p. 55) in the sense that judges "are using linguistic principles to accomplish an agenda distinct from the[se] principles" (ibid., p. 62). These linguistic rules are applied in cases like that of the ICJ example used in this section in which "the connector (and or or) is within the scope of some logical operator, such as a negative" (ibid., p. 46), presumably because "the relationship between conjunction [and disjunction] is a potential source of confusion" (ibid., p. 49)-although the overall frequency of their application appears to be "miniscule compared to the number of times that statutes and other legal documents use conjunction and disjunction" (ibid., 45). However, these linguistic rules have no equivalent in the VCLT. More precisely, such rules fall under the category of canons of interpretation, which, in the VCLT, are hierarchically subordinate to the rules of interpretation (see Section 2). This contribution limits itself to comparing the use of the VCLT rules with congruent canons of statutory interpretation, which, based on MacCormick and Summers (1991b), have been identified by prominent pragmatic approaches to legal argumentation as forming a universalist preferred order for the use of forms of interpretative arguments in the field of law.