Making Interpretation More Explicit: International Law and Pragmatics

In: Nordic Journal of International Law
View More View Less
  • 1 Switzerland
  • 2 Switzerland

Building on earlier research, the present article integrates linguistics and pragmatics into the study of interpretation of treaties in international law. This approach aims to make the reasoning of interpreting agents and their appeals to interpretive canons more explicit and transparent. This is consequently demonstrated with a number of practical examples in which the process of legal interpretation and its accommodation of the mentioned norms of interpretation can be adequately described and modelled. At the same time, it is shown that legal language possesses certain particularities, but nevertheless ultimately follows the basic pragmatic rules of communication. Nonetheless, linguistics and pragmatics can only provide an ultimately descriptive account of interpretation, so that evaluative judgements on the normative questions of how to respect the norms of international law still need to be made and – from a normative perspective – the rules of international law are not replaced or abrogated.


Building on earlier research, the present article integrates linguistics and pragmatics into the study of interpretation of treaties in international law. This approach aims to make the reasoning of interpreting agents and their appeals to interpretive canons more explicit and transparent. This is consequently demonstrated with a number of practical examples in which the process of legal interpretation and its accommodation of the mentioned norms of interpretation can be adequately described and modelled. At the same time, it is shown that legal language possesses certain particularities, but nevertheless ultimately follows the basic pragmatic rules of communication. Nonetheless, linguistics and pragmatics can only provide an ultimately descriptive account of interpretation, so that evaluative judgements on the normative questions of how to respect the norms of international law still need to be made and – from a normative perspective – the rules of international law are not replaced or abrogated.

1 Introduction

In this interdisciplinary article, a legal scholar and a linguist specialising in pragmatics aim to combine the scientific strands of linguistics and international law. We make a case for integrating modern linguistics and pragmatics into the study of legal interpretation and outline the benefits such an approach could offer. In a previous article, we highlighted that international legal scholarship is increasingly focused on the interpretation of international law, which has engendered a rich set of interdisciplinary approaches.3 Intuitively, one might think that the field of law and language might be a good candidate for exploring the link between international law, interpretation and linguistics because law is language in the sense that it relies on language for its use.4 Despite the increasing recognition that “linguistics has a role to play in understanding legal interpretation[,] there is little evidence that legal scholars are developing the skills to make use of linguistic skills”, which stands in surprising contrast to their embracing the fundamentals of economics.5

Given this surprising desideratum, in our previous article we presented a comprehensive programmatic overview of our joint approach to modelling legal interpretation. In this article, we focus more closely on our posited parallel between the scientific domains of pragmatics and the interpretation of treaties6 in international law and how well they map onto – or apply to – each other. For reasons of simplicity, we henceforth simply speak of interpretation meaning the interpretation of treaties in international law. While linguistics and pragmatics may be applied to any field of law, we confine our analysis to the debate on treaty interpretation in international law. This is not only because the pragmatic perspective is notably lacking from the latter,7 but also because it is a promising field for research. In particular, interpretation in international law is of interest to a broad international audience of lawyers and scholars who are all faced with the same problems based on the same norms in one common legal order.

Our parallel is based on the fact that pragmatics, which “studies utterance interpretation in context”,8 has evolved from the so-called “waste-basket” of linguistics to become a fully-fledged discipline.9 The same should be true of legal interpretation and its study in international law, which should no longer be considered merely a legal afterthought – as if meaning were “all in the text” of a treaty.10 To make this parallel more explicit, a caveat is needed at this point: the practice or process of interpretation needs to be distinguished from the respective corresponding scientific domains. While pragmatics has become a domain of its own, legal interpretation is not a separate domain, but a process based on a number of norms which form a central part of the respective fields of law, e.g. international law. In other words, what may intuitively prevent an exact mapping of the two fields of research appears to be the domain structure within law. However, despite the lack of a proper “domain”, legal interpretation has become an object of study of its own.

In this article, we zoom in on central tenets of a modern cognitive pragmatic theory, namely Relevance Theory,11 and examine the extent to which this theory of meaning, communication and utterance interpretation can be applied to – or maps onto – legal interpretation or, conversely, the extent to which the legal domain represents a special case that falls outside of the scope of any general pragmatic theory.12 We test the validity of our posited parallel with the help of practical examples of interpretation processes in the field of international law.

In the context of this sample analysis, we take up linguist Carston’s argument that interpreting agents in the legal domain appear to appeal to interpretive canons to support their favoured interpretation, yet they do so mostly implicitly.13 Her observation lends further support to our previous finding that interpreting agents do not always make their reasoning in using the norms of interpretation of international law fully transparent or explicit.14 We thus test the hypothesis that a relevance-theoretic pragmatic analysis can help to make legal reasoning more explicit. This analysis may, in turn, help to build a stronger scientific foundation for the debate within international law on interpretation15 and thus to provide “some better grounded explicit guidance in (at least some) cases of disputed interpretation”.16

Further to our posited parallel and the above-mentioned question whether and to what extent meaning can be said to be in the text, there appear to be similarities in the debate between textualist and non-textualist approaches to legal interpretation and the debate between literalist (or minimalist) and contextualist views in semantics.17 Due to its limited scope, this article only briefly touches upon these debates and the fact that, similarly to the criticism of the textualist notion of the primacy of the text, there are “semantics/pragmatics border wars”18 relating to the question of what role context plays in interpretation.19

We begin this article with a brief overview of the relevance-theoretic account of pragmatic utterance interpretation and explore the boundaries between semantics and pragmatics along with the role of context. Then we discuss the particularities of legal language and assess how well Relevance Theory may be applicable to this special case of communication. After briefly recalling the norms of interpretation in international law, we undertake an analysis of practical examples to test whether a pragmatic theory can offer an opportunity to adequately describe and model the process of legal interpretation, accommodating the mentioned norms of interpretation. Finally, we examine the potential benefits and limits of an integration of modern linguistic and pragmatic theory into the study of interpretation in international law.

2 Semantics, Pragmatics and Relevance Theory

Recent writing in international law arguably does not fully take into account current research in the fields of linguistics and pragmatics. It may thus be useful at this point to provide a brief introduction to pragmatics, the cognitive turn and the emergence of a complex relationship between the fields of semantics and pragmatics. For the sake of simplicity, we mainly use the terms of utterer and addressee. The main tenets apply regardless of whether communication takes place in oral or written form. In the same way, these tenets apply whether communication is simultaneous or deferred.20

2.1 The Development of Pragmatics

Depending on one’s scientific standpoint, pragmatics may be defined as belonging to different fields. The discipline of pragmatics has its origins in the philosophy of language, in particular the work of Paul Grice, who made a crucial distinction between what an utterer says and what they mean by saying it.21 The discipline of pragmatics is also regarded as a sub-discipline of linguistics, where it is distinguished, among others, from semantics: “while semantics is concerned with the meaning encoded in the formal components of a language (its words and syntax), pragmatics focuses on the way those forms with those meanings are used in particular contexts to express or communicate a range of different concepts or thoughts”.22 Furthermore, Grice’s work paved the way for pragmatics to become situated within cognitive science, which intends to explain the way the human mind works.23 In this context, greater emphasis is put “on the actual processes of interpretation, the constraints on those processes due to limits on addressees’ cognitive resources, and the mental representations that an addressee forms of a speaker’s intended meaning”.24 The goal of pragmatic theory is thus to account for this intended meaning or, in other words, speaker meaning, and to model how “hearers/readers bridge the gap between linguistically encoded meaning” and communicatively intended speaker meaning.25

To explain this development of pragmatics, we must clarify the difference between two perspectives on verbal communication. The first perspective is that communication is encoded, directly or indirectly, in language (a code model). The second perspective leads back to Grice’s work. Grice postulated that successful communication does not depend on exact linguistic encoding and decoding. Instead, such success depends on the addressee’s ability to recognise the utterer’s intention. Grice introduced the distinction between a sentence and an utterance.26 In linguistics and pragmatics, sentences and their meaning are traditionally viewed as context-independent and as referring to information or meaning associated with that sentence according to the underlying linguistic system.27 Utterances and their meaning (also called speaker meaning,28 see above) are regarded as context-dependent and as referring to the information associated with that utterance according to the intentions of the utterer.29 In Grice’s view, a speaker or writer who utters intends the addressee to recognise the intention behind the utterance, and this recognition is what produces an effect on the addressee,30 such as understanding.31

Grice thus laid the foundations for an alternative to the code model of communication, i.e. an inferential model, in which “a communicator provides evidence of [their] intention to convey a certain meaning, which is inferred by the audience on the basis of the evidence provided”.32 This does not mean that the code model of communication is replaced by a solely inferential model. Verbal comprehension involves an element of decoding because an utterance is a linguistically-coded piece of evidence. However, the linguistic meaning thus decoded is “just one of the inputs to a non-demonstrative inference process which yields an interpretation of the speaker’s meaning”.33 In other words, the correct interpretation is a joint product of linguistic information, contextual premises and deductive processes.34

2.2 Relevance Theory

Sperber and Wilson built on a number of elements from Grice’s work when they made their seminal contribution to pragmatics in the form of Relevance Theory. As an inferential approach to pragmatics, Relevance Theory follows Grice’s postulate that the recovery of speaker meaning is an intentional process. In the relevance-theoretic account, all forms of communication involve a two-fold intentional (ostensive-inferential) process.35 The utterer must explicitly or overtly show a communicative intention (ostension) to communicate a particular piece of information to the addressee, which the addressee then has to infer (inference).36

Sperber and Wilson also accept Grice’s claim that utterances “automatically create expectations which guide the hearer toward the speaker’s meaning”.37 However, Relevance theorists criticise that – as Gricean approaches would suggest – these expectations should be based on Grice’s Cooperative Principle and the conversational maxims of Quality (truthfulness), Quantity (informativeness), Relation (relevance) and Manner (clarity).38 Sperber and Wilson emphasise that interlocutors may not cooperate (or may even be self-interested, deceptive or incompetent), and still successfully communicate.39 They give special importance to the content of only one of Grice’s maxims, the notion of Relation (relevance) in the sense that “the expectations of relevance raised by an utterance are precise and predictable enough to guide the hearer toward the speaker’s meaning”.40

In Relevance Theory, the technically defined notion of relevance is a potential property of any input to any perceptual or cognitive process.41 Utterances raise expectations of relevance not because of communicative conventions such as the Cooperative Principle and maxims, but because “the search for relevance [regarding those stimuli and representations] is a basic feature of human cognition, which communicators may exploit”.42 Human cognition is entirely constrained by the principle of relevance, also termed the First (or Cognitive) Principle of Relevance43: human cognition is oriented towards maximising relevance.44 This property of human cognitive systems is grounded in “their evolved orientation towards achieving as many improvements to their representational contents as possible”.45 The First Principle of Relevance entails the Second (or Communicative) Principle of Relevance:46 every utterance presumes its own optimal relevance.47 This presumption of optimal relevance “unpacks as an implicit guarantee that the speaker/author has been as relevant as possible (to the addressee) within the parameters of [their] abilities and preferences”.48

Given this technical notion of relevance, any external stimulus or internal representation (e.g. sights, sounds, thoughts, memories or conclusions of inferences) may be relevant or matter to a given individual at a given time49 to the extent they yield varying cognitive effects. The most important type of cognitive effect is a contextual implication, i.e. an implication or conclusion which follows “(inferentially) from new input in combination with a context of existing assumptions, but not from either alone”.50 While yielding cognitive effects is a prerequisite for relevance, relevance is not to be measured in absolute terms, but rather comparative ones.

Sperber and Wilson point out that there are more or less implicit – as well as nonverbal – forms of communication.51 They argue that implicit communication is much vaguer than explicit statements, and that this vagueness is often intentional.52 Relevance Theory can accommodate these cases because the utterer’s informative intention is to modify “not the thoughts [directly] but the cognitive environment of the audience”.53 In contrast to accounts that follow a strict code model of communication, Relevance Theory claims that “thoughts do not travel” from one brain to another54 by means of encoding and decoding. Instead, thoughts must be attributed and inferred on the basis of the interlocutors’ respective cognitive environments. A cognitive environment is a set of assumptions that are “manifest to an individual; that is, assumptions that are entertained as true or inferable” in the inferential process of verbal communication.55 The interpretive or contextualising process of formulating, confirming or infirming hypotheses strengthens, weakens or suppresses old assumptions and generates new ones altogether.56 The process as a whole is driven by the search for relevance. The more similar assumptions two people share, the greater the overlap between their cognitive environments,57 and the more likely the search for relevance will lead to successful communication.58 Verbal communication is a special kind of communication in the sense that it raises specific expectations about the utterer’s intention due to its input in the form of utterances.59 Utterances encode logical forms (conceptual representations, however incomplete or fragmentary) which the utterer has manifestly chosen to provide as input to the addressee’s inferential comprehension process. As a result, verbal communication can achieve a higher degree of explicitness than non-verbal communication.60

The comparative notion of relevance is further characterised in terms of cost-benefit and can be schematised as cognitive effects and processing effort.61 The more cognitive effects or pieces of information are available to the addressee, the greater the relevance of the utterance, and vice versa.62 Information is relevant if it has at least one positive cognitive effect in a given context – if it adds, modifies or deletes information.63 A “positive cognitive effect is a worthwhile difference to the individual’s representation of the world: a true conclusion, for example”,64 such as establishing the correct meaning of a treaty provision in international law. False conclusions are not worth having because they are cognitive effects, but not positive ones.65 Based on its cognitive and communicative principles along with the overtness – or explicitness – of the utterer’s communicative and informative intentions that accompany an utterance (see above), Relevance Theory suggests that cognitive effects and cognitive processing can be naturally balanced internally66 by means of a specific comprehension procedure, which is automatically applied in the online processing of verbal utterances:

Relevance-theoretic comprehension procedure

a. Follow a path of least effort in computing cognitive effects (contextual implications):

Test interpretive hypotheses (disambiguations, reference resolutions, lexical adjustments, implicatures, etc.) in order of accessibility.

b. Stop when your expectations of relevance are satisfied.67

To explain how inferential processes lead to the correct interpretation of utterances, the relevance-theoretic account of comprehension draws on cognitive science, specifically the Theory of Mind. Relevance Theory is a form of cognitive pragmatics in which utterers are mind-readers.68 The ability to read minds is “neither random nor the result of social or linguistic conventions”, but is produced by what the cognitive philosopher Daniel Dennett called the intentional stance – the idea that humans have an innate ability to attribute mental states, such as intentions and beliefs, to others.69 Relevance Theory thus combines the study of utterance interpretation in context with elements of cognitive theory.70 The intentional stance, though not infallible,71 allows an individual to explain and predict the behaviour of others and to predict and manipulate their mental states,72 based on two premises: first, other individuals are rational agents; second, other individuals are endowed with beliefs, desires and other mental states.73

Based on these premises, the intentional stance allows Relevance Theory to replace the nigh-unattainable mutual knowledge requirement of a code model with the weaker notion of mutual manifestness.74 Mutual knowledge could only be achieved if an utterer and an addressee could know that only shared assumptions are used in the communication process. For this purpose, the addressee must, however, know – and not merely assume – that the utterer knows that the addressee holds an assumption, which again the addressee must know and so forth ad infinitum. Mutual knowledge is thus impossible to achieve in the practice of verbal communication.75 In turn, it suffices that a fact be perceptible or inferable at a given moment for this fact to be manifest to an individual. Since manifestness depends on the cognitive abilities and the physical and cognitive environments of an individual at a given moment,76 an individual may attribute a similar manifestness to their interlocutor (i.e. assume that there is a good chance that the cognitive environments of the two interlocutors overlap). As a result, mutual manifestness is possible in communication. Two restrictions apply, which show why the intentional stance is fallible: “people never share their total cognitive environments. Moreover, to say that two people share a cognitive environment does not imply that they make the same assumptions: merely that they are capable of doing so.”77

Inferential processes in verbal communication are thus interpretive processes of formulating and confirming or infirming context-based hypotheses. Context, in turn, is partly determined by the linguistic meaning of utterances,78 and partly by the information the addressee can retrieve at a given moment from short-term or working memory – for instance, the addressee may process sensory information, or the preceding utterances or long-term memory triggered by the semantic content of utterances.79 Thus, when an addressee follows the path of least effort and arrives at an interpretation that satisfies their expectations of relevance and there is no contrary evidence, then this is the most plausible hypothesis about the utterer’s meaning. Given the fact that comprehension is a non-demonstrative inferential process, this hypothesis may well be false; but it is the best a rational addressee can do.80

2.3 The Process of Utterance Interpretation and the Notions of Explicature and Implicature

At this point, we can now describe the process by which utterances are interpreted, including the notions of explicature and implicature, which we examine with the help of practical examples in section 5. In short, Relevance Theory holds that meaning is linguistically underdetermined and that it must be contextually enriched or adjusted81 in a variety of ways by an addressee in order for them to infer the utterer’s meaning.82 In other words, linguistic evidence constrains but does not determine utterance comprehension or interpretation.83 This is because linguistic evidence or, in other words, explicit content in utterances does not usually express or encode complete propositions (i.e. representations of states of affairs that may be evaluated by the addressee as to their truth or falsity)84, which distinguishes this view from the Gricean conceptions of “what is said” by an utterance.85

While presumably possible in principle, utterers do not usually aspire to full explicitness because it would not normally be economical and/or compatible with their preferences.86 To put the relevance-theoretic position more precisely, linguistic underdeterminacy is “not just a matter of processing convenience (saving of speaker or hearer effort) but is, in fact, an essential property of natural language sentences, which do not encode full propositions but merely schemas for the construction of (truth-evaluable) propositional forms”.87 While there is not space here to present truth-conditional semantics in detail, note that “the truth-conditions of a sentence are the conditions under which it would be true – that is what the world would have to be like in order for [that particular] sentence to be true. … On the other hand, the truth value of a sentence in some particular world is a specification of whether the sentence is in fact true in that world.”88 To give an example, we cannot say whether a propositional form such as “The defendant is guilty” is true or false, or what state of affairs in the world it represents, until we know the answers to a number of questions: we need to know who the defendant refers to, at what time or under which circumstances the referent is guilty, whether the utterer is expressing a belief of their own or attributing it to someone else, etc. In other words, the addressee must not only contextually infer what the utterer meant by what they said, but also what is actually said.89

The addressee thus decodes linguistic meaning and follows a path of least effort to “enrich it at the explicit level and complement it at the implicit level until the resulting interpretation meets [their] expectation of relevance”.90 The different sub-tasks involved can be schematised as follows:

Sub-tasks in the overall comprehension process

  1. a.Constructing an appropriate hypothesis about explicit content (Explicatures) via decoding, disambiguation, reference resolution, and other pragmatic enrichment processes.
  2. b.Constructing an appropriate hypothesis about the intended contextual assumptions (Implicated Premises).
  3. c.Constructing an appropriate hypothesis about the intended contextual implications (Implicated Conclusions).91

It is important to note that this schematic outline is considerably oversimplified as explicatures and implicatures, i.e. implicit premises and conclusions, are arrived at by a process of mutual parallel adjustment, in which hypotheses about both are considered in order of accessibility:92

These sub-tasks should not be thought of as sequentially ordered: the hearer does not FIRST decode the logical form, THEN construct an explicature and select an appropriate context, and THEN derive a range of implicated conclusions. Comprehension is an on-line process, and hypotheses about explicatures, implicated premises, and implicated conclusions are developed in parallel against a background of expectations which may be revised or elaborated as the utterance unfolds.93

The need for explicatures, i.e. the pragmatic enrichment at the explicit level of an uttered sentence, can be illustrated with the help of an example of reference resolution. Begin with the utterance of the sentence: “She carried it in her hand”. The addressee must pragmatically enrich the uttered sentence to obtain its full meaning or proposition.94 Pragmatic enrichment means that the addressee has to draw on their cognitive and/or situational context to determine who carried what in whose hand, or, in other words, what the uttered sentence explicitly says, because pronouns can potentially refer to anyone or anything, i.e. their referents cannot simply be decoded.95 Only after the enrichment does the linguistically provided template, here a sentence, typically become fully propositional and thus truth-evaluable.96 Explicatures may therefore be viewed as so-called “pragmatic intrusions” into the truth-condition of utterances.97 Traditionally, semantics and pragmatics are regarded as separate realms: semantics focuses on elements that influence the truth-condition of utterances; pragmatics is concerned with elements unrelated to truth,98 i.e. with implicatures – what the utterer meant by what they said.99 One marked difference between Relevance Theory and Gricean accounts is that Relevance Theory does not restrict pragmatic inferential processes that trigger implicatures to what is meant or implicated, but also relies on those processes to determine “what is said”. The content of explicatures thus goes well beyond Grice’s “what is said”, requiring pragmatic contributions for its recovery.100 In the relevance-theoretic approach, the traditional level of “what is said” has no psychologically significant role to play in interpretation, as this level has been shown not to be rich enough to form the basis for further inferences.101

In contrast to explicatures, implicatures are situated at the implicit level of the utterance. They are contextual inferences that addressees draw from the pragmatically-enriched explicit meaning or evidence and lead to cognitive effects in the addressees’ cognitive environments,102 such as contextual implications. If Peter says “It is four o’clock”, Mary might derive the implicated conclusion, depending on the context, that they will be late for a class that starts at four o’clock or that they still have time for a coffee before an appointment at five o’clock.103 In each context, Mary will have constructed the implicated premise that they do have a class at four o’clock or an appointment at five o’clock, or even that drinking coffee will help them to stay focused on the upcoming tasks, etc. The proposition “It is four o’clock” may be either true or false. Implicatures, in contrast, fall on a range from stronger to weaker.104 How strongly a proposition is implicated also correlates to how essential it is for the addressee to derive the respective proposition in order to satisfy the expected relevance raised by the utterance.105 This also means that an utterer may not always be fully committed to the truth of implicatures,106 but instead may have more or less confidence in them and communicate them more or less powerfully.107

It thus appears that the borders between semantics and pragmatics are not clear-cut. For instance, these borders appear to be closely entwined with the “explicit/implicit distinction”, which Relevance Theory draws differently from the Gricean “what is said” and “what is meant”.108 Also, we have seen that many factors affect truth-condition, which would normally relegate them to semantics, but they may also draw on context, which place them within pragmatics.109 This fact blurs a traditional distinction between a context-dependent pragmatics vs. a context-independent semantics.110 The afore-mentioned semantics/pragmatics border wars thus refer to the debate between literalist (or minimalist) and contextualist views in semantics: “literalists” assume that semantics is basically autonomous with little or minimal “pragmatic intrusion”; “contextualists” adopt the basic outlines of the relevance-theoretic view of pragmatic contribution at every level. However, “no one denies that contextual facts and pragmatic reasoning are needed at the near side of what is said [i.e. the explicit content level]”.111 This debate, which reflects the debate surrounding the primacy of the text in international law, leads back to the question as to how far our posited parallel extends or the extent to which the legal domain presents a special case of language and communication that does not readily fall within the reach of the relevance-theoretic account.

3 Assessing the Particularity of Legal Language

As there appears to be a certain familiarity of legal scholars with early, classical or more traditional models of pragmatics,112 an objection raised by legal scholars such as Marmor is that “it is unlikely that Gricean maxims apply to legal language because … ‘the enactment of a law is not a cooperative exchange of information’ but rather a form a strategic behaviour”.113 While we agree with Carston that it would be foolish to deny that legal language has some special properties, such as its strategic nature,114 we have already seen that cooperation is not essential for communication, which Relevance Theory accounts for based on the Second Principle of Relevance, i.e. the optimal relevance of an ostensive stimulus.115 On top of that, we further agree with Carston that there seems to be a somewhat attenuated notion of cooperative, purposeful activity that might cover all kinds of linguistic communication, including legal language: “the producer of the language wants to get a certain meaning across to an audience and the audience wants to grasp that meaning”.116

Another objection commonly levelled against the claim that pragmatic theory could apply to legal interpretation is based on the fact that an international legal utterer is not a single person, but part of a group or chain of authors whose intentions may differ: “for most such legal texts, the documents were drafted, approved, or ratified by large groups … It is not clear how, if at all, the differing individual intentions can be ‘summed up’ into a collective intention for the legal text”.117 Also, international legal communication is often temporally and spatially deferred, though it aims to regulate hypothetical future cases.118 This type of deferred communication implies that the respective authors/utterers may address multitudes in varying contexts. The easy way out of this dilemma is, supposedly, to abstract from inferable intentions and to conclude that all meaning is in the text (“‘word meaning’, ‘plain meaning’”)119 and can be decoded from the text. This conclusion is, however, cognitively untenable. Given the way language works and the extensive role pragmatics plays in verbal comprehension, it should be very rare that legislative content means exactly what its linguistic evidence or explicit content says.120

The concept of cognitive environments, Theory of Mind and the intentional stance arguably offer potential solutions to the collective intention problem. International legal utterers and addressees are normally international lawyers or are, at least, very familiar with international law, and they normally assume a mutually manifest cognitive environment, albeit not always explicitly. The given addressee may assume that the existence of a potential multitude of addressees was manifest to the utterer. The intentional stance tells both utterer and addressee that they are rational and that it is unlikely that communication should be geared towards failure.121 Since communication is ostensive-inferential, utterers and addressees should be able to infer the intended meaning implied in the treaty text from the linguistic evidence. Due to the functioning of relevance, there may have been implicit weaker or less relevant assumptions just below the level of conscious attention – even if there was no conscious consideration of a problem by (all) utterers122 – because in order to have a representation of a set of assumptions “it is not necessary to have a representation of each assumption in the set. Any individuating description may do.”123

On this basis, one may reasonably claim that that the group or chain of authors of a given legal text must have had at least some degree of overlapping or collective/joint intention(s). Otherwise, the following question arises: if the signatories to a treaty find that they are in disagreement about its correct interpretation, how can one of the divergent interpretations be judged correct when all parties signed up in good faith to a text that they each took to be concrete evidence of their own intention? Can they be said to have unwittingly signed different treaties?124 One can find reflections of attempts to answer this question in the legal domain. In Scalia’s words, interpreting agents “look for a sort of ‘objectified’ intent – the intent that a reasonable person would gather from the text of the law”,125 which is in line with the attribution of (mutually) manifest assumptions in Relevance Theory.

Lastly and most importantly, we address the objection that pragmatics, as some critics suppose, can only model ordinary, non-specialised communication in the form of conversation. However, there has been no convincing effort in philosophy of language or related fields to show that legal interpretation is something entirely other.126 In particular, no one has demonstrated that legal interpretation can be seen as communication without cognition. The cognitive tendency to maximise relevance, as described by the Cognitive Principle of Relevance Theory, is a universal human characteristic according to which interlocutors may “predict and manipulate the mental states of others”.127 It is, therefore, active in all forms of communication, including discourse in international law.

Texts in international law are like other specialised forms of communication in that they use everyday as well as technical language.128 Most visibly, these texts are particular in their uses of archaisms, long-winded and complex sentences and generally peculiar style.129 Arguably, such stylistic elements may be found to differing degrees in any genre and text or discourse type. Even the technical terms legal utterers use are just conceptual terms that refer to conceptual content. The expert utterer (the international lawyer) should, as a result of their professional experience, have a mental representation of such content as well as related assumptions. Their subject-specific experience should also have provided them with mental representations of, and the ability to make assumptions about, the formal properties of treaty texts in international law.

Their assumptions also include value judgements or second-order assumptions. Some may find problematic the assertion that assumptions are neither true nor false, but are measured in terms of their strength, i.e. as “more or less strong or well-founded”.130 But the idea of relative strength is arguably not problematic per se. The cognitive environment of an international lawyer, in which their mental representations and assumptions reside, also contains assumptions about acceptable and unacceptable inferred conclusions. If mentally represented content of the norms of interpretation is part of the cognitive environment, this content is readily available and applied, just as “can you pass the salt?” is understood as a request and not a question about one’s ability to pass the salt.131 The fact that such usage is not simply decodable, but has to be processed for use before it can become mentally manifest is made clear by the fact that laypersons, who are not (proficient) speakers of the “language” of international law, may not be able to infer such implicatures, despite semantic evidence.132

The notion of value judgements and the related assumptions about acceptable and unacceptable inferred conclusions lead to the only phenomena that appear to be truly particular to interpretation in international law. First, the most important way in which law differs from linguistics and pragmatics is that the former is not descriptive, but normative:

[L]egal texts … have a prescriptive or normative content …, and that content has to carry over to times, places, and conditions distant from the time/place/conditions of origin. This combination of properties might well make for certain tendencies in the linguistic formulation of the documents. We might expect a much higher degree of verbal explicitness than is typical (or necessary) in face-to-face speech, due to a less rich or clearly shared context in which to interpret the language.133

While more explicit formulations may be found to differing degrees in any form of written/deferred communication, the normative nature of the law is closely linked with a second true particularity of the legal language: its many explicit rules or conventions of interpretation.134 However, in line with our posited parallel, Carston highlights a similarity between pragmatic principles/maxims and legal norms, which are both explicitly formulated:135 “as a matter of interpretive practice, at least some of the [norms] used by judges and other interpreters of the law are instantiations of Gricean maxims”.136 Just like the explicitly formulated Gricean maxims, norms on the interpretation of international treaties might be surface manifestations of a general foundational principle underlying all pragmatic or contextual inferences, i.e. the cognitively grounded principle of relevance in the relevance-theoretic account.137

After this short overview, one may conclude that the peculiarities of legal language do not a priori appear to be incompatible with the relevance-theoretic description of ordinary utterance interpretation, perhaps apart from “certain provisos and/or modifications”138 concerning the integration of explicit rules or norms into the relevance-theoretic account. Before we put our posited parallel to the test, we will, therefore, take a quick look at the content of the rules on the interpretation of treaties in international law.

4 The Norms of International Law on the Interpretation of Treaties

If it is accepted that interpretation in law is not some extraordinary form of communication that has nothing to do with “ordinary” communication, linguistics and pragmatics can help us to better understand the process of interpretation with their concepts and notions. In essence, legal interpretation is thus a process of communication in which lawyers use their cognitive environment and background knowledge to draw inferences that are to some extent different from those drawn by a layperson. This is due to the fact that a number of norms have become part of their cognitive environment and make them give more weight to some inferences and less to others, thus virtually excluding some interpretations that would be considered “unsound“ or “not rational“ from a purely legal-positivistic point of view.

The norms of interpretation contained in the Vienna Convention on the Law of Treaties (vclt)139 can be understood as such norms, and it is in this light that they together with other norms of interpretation are to be set out in a short overview at this point. Article 31(1) lays down that 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”. This norm and a number of subsequent norms enshrined in the Convention establish a number of instructions as to how international law requires a law-applying agent to interpret treaty norms. However, there are also further instructions in international law to be found outside the framework of the Vienna Convention.

Based on the doctrine,140 a number of elements can thus be identified.141

  1. As a starting point, the norms applicable to interpretation in international law are subject to development in time, so that it is important to keep in mind that legal developments may make some principles lose force and render others more “relevant”. As an example, current doctrine no longer adheres to early views that the meaning of treaties can be so “clear” that they only need to be “applied” and not “interpreted”.142
  2. There are claims that the norms applicable to interpretation undergo some modifications with certain kinds of treaties, the typical example cited being human rights treaties.143 Others, however, argue that the somewhat modified approach to interpretation is based on the very same norms of interpretation being applied somewhat differently because of the treaty at issue.144
  3. Article 31 vclt contains three elements set out in more detail in its three paragraphs. Put simply, these elements are the text, its context, and the object and purpose of the treaty in question. As to the relationship between these three elements, the general consensus is that there is no hierarchy and that the three elements are to be examined in “a logical progression”.145
  4. The textual element consists in looking for the “ordinary meaning” of a term,146 although it is generally admitted that there typically is no such thing as one single ordinary meaning.147 Interpreting agents thus often use dictionaries or specialist books to find a definition.148 In the case of generic terms, a dynamic or evolutionary approach to interpretation is often employed because the intent of the parties in using such a term is assumed to have been to allow that such a term be given meaning in light of the circumstances prevailing at the time of interpretation.149
  5. The context is defined in Article 31(2) vclt and, generally speaking, requires an interpreter to take into account the treaty as a whole with the preamble and the protocols, its systematic structure,150 certain agreements relating to the treaty and certain instruments made in connection with the treaty.151 Furthermore, according to Article 31(3) vclt certain subsequent agreements and practices152 as well as relevant rules of international law applicable between the parties are to be taken into account “together with the context”.153 Under the element of context, an interpreting agent can, for example, compare whether the same term is used elsewhere in the treaty or whether a related treaty contains analogous wording.154
  6. As the third element, Article 31 vclt requires the interpreter to take into account the object and purpose of a treaty, which commentators see as an expression of the principle of effectiveness. 155 The interpreting agent must assess what the aims of the treaty are, e.g. by looking at the title, the preamble or special clauses on the objectives of the treaty,156 and attempt to promote them during the interpretation of the treaty’s terms.157
  7. Article 32 vclt states that recourse may be had to supplementary means of interpretation “including” the preparatory work of the treaty or the circumstances of the treaty’s conclusion, but only if the interpretation process based on Article 31 vclt leaves the meaning of a term “ambiguous or obscure” or leads to a “manifestly absurd or unreasonable” result.158 It thus establishes a relationship between its elements and those of Article 31 vclt.
  8. Certain norms of interpretation do not so much describe elements that interpreters should focus on but rather the relationship between the elements of interpretation, e.g. that inferences should be drawn from the subsequent practice to a treaty rather than from its preparatory work.159
  9. The norms of interpretation contained in the Vienna Convention are generally seen as not being exclusive, so that there can exist other norms of interpretation in international law.160
  10. A principle not contained in the Convention is the principle of restrictive interpretation, also called in dubio mitius, according to which treaties are to be interpreted in favour of state sovereignty. However, despite its relevance in earlier international legal jurisprudence161 its existence as a rule of customary international law and relevance is disputed.162
  11. Another principle that is often only classified as part of interpretation in light of the object and purpose of a treaty163 is that of effectiveness (ut res magis valeat quam pereat). According to the principle, treaty provisions are to be interpreted so that every part of the text can be attributed meaning and that they receive their fullest effect and weight.164
  12. Aust lists a number of further, unwritten principles that are derived from domestic (often Roman) law and that can help as “aids” to discover ordinary meaning.165 For example, express mention of a circumstance can exclude others (expressio unius est exclusio alterius); or the use of special words followed by general words means that the general words are limited by the genus/class of the special words (ejusdem generis).

Based on this overview of the norms of interpretation in international law, we can now return to the relevance-theoretic account and pull the threads together. Practical examples can be used to show how implicature and explicature are at work during the process of interpretation; at the same time, the actual operation of various norms of interpretation in the interpretive process becomes more transparent, notably without their content being affected.

5 Applying Relevance Theory to Legal Interpretation

We now present three practical examples of interpretive processes in the context of international law, which serve to illustrate the sub-tasks in the overall relevance-theoretic interpretation process and which can be categorised along the lines of explicature and implicature (see section 2). In this framework, we also examine the claim that interpreting agents do not always make their reasoning in using the norms of interpretation of international law fully transparent or explicit166 in order to test whether a relevance-theoretic pragmatic analysis can help to make legal reasoning more explicit.

5.1 Explicature: Prisoners of the Sea

Our first example makes a case for the construction of explicatures, albeit not explicitly referred to as such, by interpreting agents in legal interpretation. Remember that Relevance Theory holds that language is not a code and that cognitive economy ensures that language is never fully explicit, and even linguistically underdetermined. This means that even the explicit content of the linguistic evidence in a treaty (i.e. what is explicitly expressed by an utterance, or rather by the author of that utterance),167 e.g. the explicit content of conceptual terms, must be pragmatically inferred. According to Relevance Theory, a conceptual term does not encode a fully-fledged concept. Rather, the encoded lexical meaning of a term only provides the basis for pragmatic enrichment processes, which lead to the more specific or broader concepts a term can be used to express.168

More precisely, in Relevance Theory most words encode concepts. These concepts are, however, atomic or mere schemas. They are thought of as addresses in memory which are associated with a lexical, a logical and an encyclopaedic entry. These entries constitute pointers to the related linguistic information, mental representations, assumptions, the applicable inference rules, etc. These pointers indicate the conceptual space on the basis of which an expression or conceptual term will get its actual reading in a particular utterance situation169 following the relevance-theoretic comprehension procedure (see section 2). The relevance-theoretic account may thus provide an explanation and a model of why “all too often, what the ‘ordinary meaning’ is in particular instances is an issue” in legal interpretation.170 According to Relevance Theory, this issue arises because even the explicit content of a word or term is dependent on context or, in other words, the construction of an explicature.

Aust mentions an interesting example of an explicature or, as he calls it, an “implied term” in a treaty.171 Article 22 of the Third Geneva Convention172 states that prisoners of war “may be interned only in premises located on land”. However, during the Falklands conflict in 1982, there was not enough accommodation in the territory for the large number of Argentine prisoners of war because tents had been lost when a British ship had been sunk by the opponent. After consultations between the United Kingdom and the International Committee of the Red Cross, the prisoners were kept on merchant ships and warships until their repatriation. Aust suggests that in this situation a term could be read into Article 22 that a party that cannot comply with that norm due to reasons beyond its control may hold prisoners of war on ships if that is preferable to leaving them on land without adequate protection from the harsh conditions. His conclusion is that “[g]ood interpretation is often no more than the application of common sense”.173 International lawyers might certainly argue over whether Aust’s proposed solution is dogmatically tenable, with some finding it a pragmatic proposal, others arguing that the norms of interpretation in international law oppose such undogmatic pragmatism. Our point is a different one: namely, to demonstrate that Aust’s reasoning can be made more explicit and transparent using the tools of linguistics and pragmatics, in particular the concept of explicature. As a consequence, even a sceptical appraisal in international law can engage in more depth with Aust’s approach based on the understanding gained through linguistics and pragmatics.

Aust’s conclusion illustrates our and Carston’s observation that interpreting agents do not always make their reasoning in using the norms of interpretation fully transparent or explicit. Relying on the relevance-theoretic framework, one could describe Aust’s reasoning and the related interpretive process as a broadening or loosening of the linguistic content encoded in “land”.174 While the concept of “land” encodes conceptual meaning (e.g. the encyclopaedic entry of the concept pointing to or making accessible the assumption that land may border on water, but is different from it), this meaning must be decoded and pragmatically enriched in interpretation, which here results in an ad hoc concept. The pragmatic enrichment processes which lead to the construction of ad hoc concepts are affected by a variety of factors, including “context, accessibility of encyclopaedic assumptions, and considerations of relevance”.175 Ad hoc concept formation is thus one of the processes that lead to an explicature. In ad hoc concept formation, the conceptual content of a given lexical item is pragmatically adjusted to express conceptual content which is narrower or looser.176 This process is to be distinguished from the resolution of lexical ambiguity (homonymy) where a lexical form maps onto unrelated meanings, such as in “bank” (e.g. the riverside vs. the financial institution). Nor does an ad hoc concept necessarily imply any non-literal use.177

Aust thus modulates the linguistic content of “land” based on context: first, the immediate utterance context, i.e. “premises located on land”, in which the “premises” make available or inferable the assumption of firm ground with housing structures on it, and second, textual context, i.e. assumptions taken from the wider context of the convention, which make available the information that the firm ground with housing structures on it is meant to serve as shelter/protection from the elements (which may be viewed as a first step of narrowing the conceptual meaning of “land” in this context). Since the semantic content of “ships” makes available sufficiently similar information, the linguistic content of “land” can be loosened or extended to include “land” that is adjacent to it, even if this “land” is technically located on top of water. Because of Aust’s loosening (following the initial narrowing) of the linguistic content of “land”, the term does no longer apply to land which lacks accommodation, making the holding of prisoners of war on ships (the only remaining option) compliant with the norm.

Interpreting agents of international law may argue that the construction of an explicature appears to be at odds with looking for the “ordinary meaning” of a term, which appears to be understood as the “literal”178 meaning of a term. One may reply: first, the explicature takes into account the text, context and purpose of the convention in logical progression. Second, the notion of literal meaning leads back to the debate between literalist (or minimalist) and contextualist views in semantics and its similarity to the criticism of the textualist notion of the primacy of the text, relating to the question of what role context plays in interpretation. Relevance Theory is situated on the contextualist end of the spectrum. In the relevance-theoretic framework, the particular tension between literal vs. non-literal meaning (or the Gricean “what is said” and “what is meant”) does not arise due to the assumed linguistic underdeterminacy. Context-immune decoding of the linguistic evidence in an utterance results in a minimal proposition which is normally uninformative, irrelevant, trivially true or even patently false, and must therefore be pragmatically/contextually enriched before it can be deemed to be anything close to “what is said” by a utterer in the process of utterance comprehension.179 In other words, the minimal proposition must be enriched at the explicit level until the resulting interpretation meets the addressee’s expectation of relevance. Against the backdrop of the debate between literalist and contextualist views in semantics, it does not appear clear whether in legal interpretation the ordinary meaning of a word or phrase is to be equated with the encoded meaning or with the usual, most familiar interpretation.180 Despite the semantics/pragmatics border wars, it appears, however, safe to say that the notion of context-independence vs. context-dependence is, in fact, gradual in the sense that there are not one but several concepts of what constitutes context.181 Literal meaning is, in turn, not necessarily prior to non-literal meaning and both literal and non-literal meaning is in some way context-dependent.182

Ad hoc concept formation happens because, according to Relevance Theory, it would be cognitively uneconomical to store all this highly specific or very broad conceptual content in memory,183 which can be viewed as our mental lexicon.184 Similarly, based on considerations of economy the dictionaries to which interpreting agents often turn in search of a definition will not normally provide a very large number of contextually enriched word meanings or definitions. On the other hand, interpreting agents proceed to look for ordinary meaning on the basis of an explicitly formulated norm, or, in other words, they follow an explicitly formulated convention. In this way, legal norms are like Gricean maxims. They appear to be “rules of thumb” which license the interpreting agent to infer or enrich semantic content in a certain stereotypical way,185 but these rules do not make the underlying cognitive principles transparent or explicit.

The norms of interpretation in international law are not throughout hierarchically organised, as has been shown. Like between the Gricean maxims, there is no hierarchy between the three elements set out in Article 31 vclt, the text, its context, and the object and purpose of the treaty in question. Aust may therefore move from ordinary meaning to also employing purpose and effectiveness as guiding rules for his pragmatic inference process, here an explicature. Disagreements about the correctness of Aust’s interpretation may arise, not only because of the linguistic underdeterminacy of the explicit content and the defeasibility of pragmatic inference processes,186 but also, and more importantly, for the reason that the norms are not hierarchically organised, but complementary, and are therefore “potentially in conflict with each other and give opposite results in some cases”, similarly to Gricean maxims.187 This is also recognised at least by some in international law. Some scholars insist that the norms of interpretation give precedence to the “textual” or “literal” approach and argue that interpreters should only look at the object and purpose of a treaty to confirm what they have already found by merely looking at the text.188 The contrary and arguably at least equally widespread view, however, is that the ordinary meaning is not an element of treaty interpretation “to be taken separately”, that all the elements like the context and the object and purpose of a treaty are to be taken “in conjunction” and that the ordinary meaning, if at all, can be no more than a “very fleeting starting point”.189 The correct interpretation is thus ultimately a question of who has more or the most plausible or convincing arguments – or, in relevance-theoretic terms, who is more or the most relevant. As a descriptive theory, Relevance Theory thus appears to be able to help to make Aust’s underlying reasoning and his norm-guided inference processes more transparent or explicit, which may be helpful in case his interpretation is disputed, for instance in an attempt to show that he may not have correctly followed the norms in enriching the linguistic evidence.

5.2 Implicated Premise: The International Law Exam

Having analysed an example of pragmatic contribution to explicit utterance content, we now look at two examples of pragmatic contribution to implicit content, i.e. implicature. Unlike our first example, our second example does not make a case for the construction of an appropriate hypothesis about explicit content, but about intended contextual assumptions, called implicated premises. Drawing inferences about the (appropriate) context can substantially change what an addressee understands in an utterance whose semantic elements themselves may not change at all. In the relevance-theoretic framework, this is possible because contextual assumptions are not encoded in the linguistic form of a given utterance. Rather, the linguistic form provides access to contextual assumptions, thus contributing to the construction of a hypothesis about the utterer’s meaning.

One of the authors recently posed an exam question to his first-year international law students: “How does one determine which state is authorised to exercise diplomatic protection for a particular legal person?” This seemingly straightforward question on a classical topic of international law190 prompted most students to explain what was expected of them. They typically showed the various existing theories to prove a link between a legal person and a state, commented on the International Court of Justice’s view expressed in the Barcelona Traction case191 and some went as far as to express their own views of what should be the dominant theory. However, one student took a different approach. They also explained the various theories, but then concluded briefly: “But in Switzerland, the control theory is the predominant and thus applicable theory.” After some reflection, the author felt compelled to give the student full points for their answer,192 although the answer given was considerably shorter than those of their peers. The student had, however, read the exact same exam question. The only difference was that the student inferred that they were answering this question in an international law exam at a Swiss university, so that the question was about how to identify the relevant state according to Swiss law. The author could not think of a valid reason why this inference should not be valid and thus decided that there was no reason to punish the student for simply providing an excellent example of how inference works.

In this example, we are not so much interested in the student’s answer itself (“But in Switzerland, the control theory is the predominant and thus applicable theory”), which is an inferred conclusion from the exam question. Rather, we are interested in the reasoning that led them to make reference to “Switzerland”. Neither the term “Switzerland” nor the phrase “but in Switzerland” can be said to be a development of the encoded logical form of the utterance, i.e. the enriched explicit content of the exam question (“How does one determine which state is authorised to exercise diplomatic protection for a particular legal person?”): for instance, neither “which state” nor “state” nor “one” can be narrowed or loosened to explicitly express “Switzerland” in this utterance context.

At the same time, there was no linguistic element in the text precluding the student’s inference of the contextual assumption that in an international law exam at a Swiss university Swiss law must be considered. This contextual assumption served as an implicit premise for the student’s conclusion and answer to the exam question: the student followed the path of least effort, tested an easily accessible hypothesis (based on the situational context, the meaning of the terms in the text and the object and purpose of the exam) about intended contextual assumptions and, since there was nothing to cancel or prohibit this implicit premise, was licensed to stop as their expectation of relevance was satisfied. As “Switzerland” appeared to be a relevant implicated premise to the student, they felt licensed to base their conclusions, and thus their answer, on it.

Note that in our example, the student was not bound by the norms on the interpretation of treaties. These rules did, therefore, not influence which inferences the student felt allowed to draw. However, an important influence was, of course, the student’s cognitive environment, which was in this case marked by knowledge of international law (e.g. what the dominant theory in Switzerland is concerning the identification of the state entitled to exercise diplomatic protection for a legal person). A layperson would thus in all likelihood not have drawn the same implicated premise, not knowing about the different possible theories of diplomatic protection or the dominant theory in international law, or that situational context may play a role in the sense of the existence of a relevant theory applicable in Switzerland according to Swiss law.

5.3 Implicated Conclusion: The Day the Law Stood Still

In our third example, we focus again on pragmatic contribution to implicit content, i.e. implicature. This time, however, we focus on the construction of an appropriate hypothesis about the intended contextual implications, called implicated conclusions, and the extent to which implicated conclusions can be influenced by the norms on the interpretation of treaties. In other words, we take a closer look at the cognitive environment of international lawyers, i.e. their manifest assumptions (those assumptions that are entertained as true or inferable) about acceptable and unacceptable inferred conclusions that may be represented in their cognitive environment. In legal communication, many explicit rules that lawyers have internalised by studying and practicing law form part of this background knowledge. The interpreting agent’s background knowledge thus plays a role in determining the acceptability or weight of inferences. This can lead to situations where particular inferences can effectively play hardly any role at all. An example taken from an actual legal situation that has arisen under European Union capital markets law provides an illustration to what extent such background knowledge can effectively “exclude” inferences. Note that the problem may fall under eu law, so that one may argue that the norms of interpretation differ; the problem that arises, arguably, would however have raised identical legal issues for all practical purposes if it had been an international treaty and not an eu legal act that required a change in the domestic law.

On 3 July 2016, a new regulation on market abuse entered into force, fully harmonising certain elements of the law such as the prohibition of insider trading.193 To bring its law into compliance with the new directly applicable norms of eu law, Germany decided to abrogate its national rules on the same subject. The amending law,194 however, incorrectly195 entered into force already on 2 July 2016,196 leading to a situation in which for the whole of 2 July 2016, there was no prohibition of certain acts like insider trading in Germany. This curious result was clearly neither the intention of the legislator nor of anyone else.197 A non-legally trained person would perhaps consider to infer from the obvious intent of all the parties involved in drafting the law that what had been meant or intended was that the old norms should be abrogated only with the end of 2 July 2016, so that there was no lacuna until the entry into force of the eu regulation on 3 July 2016. However, any lawyer worth their salt will know with certainty that this is an implicated conclusion that must not be drawn.

The text of the law speaks of the entry into force on 2 July 2016. Fundamental legal principles do not allow for an extension of the temporal scope of application of a law linked with criminal liability by means of interpretation. Similarly, the prohibition of retroactive effects of criminal law norms also makes it impossible to interpret the law as meaning that the eu norms should have started to be applicable already on 2 July 2016. The construction of an appropriate hypothesis about the intended implicated conclusions may have provided a solution more in line with the intents of given parties; however, explicit rules which are manifest in legal background knowledge force the legally knowledgeable interpreter to refuse from granting certain inferences the relevant weight to change their interpretation of the legal rule in question. The example shows that not all inferences may actually be drawn, and that because of the specificity of background knowledge in legal communication, there are clear limits to what can be inferred in the context of a given legal norm.198

The norms of interpretation in international law thus are very much like pragmatic maxims in the sense that they attempt to put prevailing constraints on both utterer and addressee with the aim of facilitating successful communication and interpretation.199 In the legal domain, explicitly formulated constraining norms appear to be grounded in the normative content of the law. It seems that the relevance-theoretic account can accommodate this particularity of legal interpretation: the more explicit and detailed a norm is, such as in the case of the prohibition of the extension of the temporal scope of application (or of retroactive effects) of a law linked with criminal liability, the more easily accessible it will be and, therefore, the more easily or straightforwardly it will satisfy expectations of relevance.

As hinted at in our first example, disagreements may arise because the relevant legal norms are, for the most part, not (explicitly and precisely) graded or hierarchically organised,200 which is, however, again grounded in the normative content of the law and the fact that the content of legal texts has to carry over to distant times, places and conditions.201 While, similarly to Gricean maxims, norms potentially conflict with each other,202 they can still be accommodated by Relevance Theory, as it grounds pragmatic expectations in a single underlying principle of communicative relevance.203 It appears that in legal communication, the same comprehension procedure applies as in ordinary, non-specialised communication, with the proviso that norms, which are manifest in the form of contextual assumptions, are implicit premises which may impact the strength and acceptability of inferences or even “exclude” them, especially those in the form of explicatures (the explicature of the element of “2 July 2016” thus being “from exactly 00:00 cet to exactly 24:00 cet on 2 July 2016”) and implicated conclusions about the utterer’s meaning. Interpreting agents may not stop, but must construct hypotheses guided or constrained by norms, e.g. that for the whole of 2 July 2016, there was no prohibition of certain acts like insider trading in Germany.

While Relevance Theory may thus make such reasoning more transparent or explicit in an inferentially sound way, there are, however, limits to what pragmatics can contribute to interpretation in international law. During the process of interpretation, it will always be necessary to make evaluative judgements on the normative question of how to respect the norms in international law.204 These choices cannot be “replaced” or prescribed by examining the process of interpretation from the perspective of Relevance Theory, as it is fundamentally descriptive and not normative.205

6 Conclusion

This article has made a case for interdisciplinary research, and, more specifically, for integrating linguistics and, in particular, pragmatics into the study of interpretation in international law. A number of publications have focused on interpretation in international law from a classical legal-dogmatic perspective or from the angle of methods like law and economics. There is, however, still rather little research using the tools of linguistics and pragmatics. The examples in this paper have shown how the notions and concepts of linguistics and pragmatics can be made fruitful for the study of interpretation in international law. Of course, pragmatic theory cannot do everything; it cannot take the final interpretive decision itself due to its descriptive nature. But based on concepts such as the intentional stance, cognitive effect/processing effort, linguistic underdeterminacy and pragmatic enrichment, cognitive pragmatic Relevance Theory appears to be able to model, and thus to make transparent and explicit, the mental processing of the decoded linguistic evidence, the available contextual assumptions and the inferred conclusions, taking into account contextual elements such as the cognitive environments of utterers and addressees.

We have also shown that legal language has some special properties, which can nevertheless be mapped onto the relevance-theoretic model of meaning and utterance interpretation in context, which thus appears to confirm our posited parallel. Legal interpretation is like ordinary, non-specialised communication in that it appears to require a kind of intentionalism to enable it to consider intentions that are not textually expressed and to make considerations about the appropriate context for interpretation,206 e.g. contextual assumptions about norms in the form of implicated premises. However, these premises have special properties in the context of legal interpretation in that they guide and constrain the interpretive process in a normative way, which may even overrule hypotheses about the utterer’s intention.

International lawyers might argue that, in practice, one can actually “do” international law without worrying about the theoretical scaffolding that upholds the use of language. Based on the intuition that communication is successful in most cases, international lawyers may hold that it suffices to apply the legal norms of interpretation in order to “discover” the correct interpretation of a treaty and to be able to “justify” this correct interpretation. However, this lack of reflection hardly satisfies the standards of a scientific discipline, which international law claims to be. Even if it is accepted that we may intuitively reach a similar outcome in most cases, linguistics and pragmatics may help us to understand and model how a text like a treaty is processed by the human mind as it is interpreted. Our aim to provide a description of the “discovery” process is warranted because, as Linderfalk argues, the very idea of having norms of interpretation presupposes providing an answer to the question of how addressees manage to understand utterances.207

Further (empirical) investigation is needed to support our posited parallel between interpretation in pragmatics and international law and to ultimately enable us, with the help of pragmatic theory, to provide cognitively grounded reasons for favouring a particular legal interpretation in the case of an explicit interpretive dispute. However, our effort to help to enable courts and other interpreting agents to apply norms as objectively as possible208 will not eliminate disagreement, as language is underdetermined and pragmatic inferences are defeasible.209 In other words, linguistic underdeterminacy potentially leaves room for more than one interpretive “discovery” process – similarly to the “underdeterminacy” of the hierarchy of the norms of interpretation in international law. As to the “justification” of the discovery process, the underlying goal or motivation of pragmatic Relevance Theory is described as follows: “It seems that human cognition is aimed at improving the individual’s knowledge of the world.”210 This underlying motivation is compatible with the overarching goal of international law’s prescriptions on interpretation, i.e. to achieve a correct interpretation in compliance with the norms of interpretation, as such a correct interpretation adds to the interpreting agent’s world knowledge. However, pragmatic Relevance Theory does not establish rules that ought to lead to a single correct interpretation because the technical notion of relevance is not normative. Therefore, this notion cannot provide a normative “justification” of an alleged correct interpretation. What linguistics and pragmatics can provide are their own criteria, such as the cognitive effort involved in determining a given interpretation, for favouring a particular interpretation. Such criteria can be of use to international lawyers confronted with problems within their own system of norms, e.g. to determine the ordinary meaning of a treaty term. In this context, higher processing effort, while normally representing a cognitively grounded reason against a particular interpretation, may be legally acceptable if an adequately strong legal justification can be procured.

It is worth noting that interpretation in international law may, at least theoretically, reside on the more conscious side of reasoning and interpretation: while Relevance Theory is an account of “automatic online processes of interpretation which occur (for the most part) at an unconscious ‘subpersonal’ level, … the explicit marshalling of (alleged) contextual evidence and the reconstructions of lines of reasoning that go on in the disputed interpretations … are obviously conscious, highly reflective, ‘person’ level activities”.211 In contrast to sub-attentively processed phenomena, which may or may not come to an individual’s conscious attention,212 the explicitness of the norms literally incites conscious effort, thinking and evaluation. This greater effort is, however, licensed by the greater real-world effect or impact which the implicated premises and conclusions may have. In other words, as more recent insights from pragmatic theory move us towards a more cognitively grounded theoretical architecture, we are arguably also licensed to move away from simplistic thinking about and understanding of communication, interpretation and intentionality in international law and towards a better understanding of the mental processes behind legal interpretive decisions.




The authors thank Ulf Linderfalk, Moshe Hirsch, Cosmina Hodoroaga and Hasmik Jivanyan for helpful comments on earlier versions of this article and Stefan Reitemeyer and an anonymous student for providing examples used in this article. Research for this article was generously supported by the Jean Nordmann Foundation. All errors remain within our responsibility.


J. Smolka and B. Pirker, ‘International Law and Pragmatics. An Account of Interpretation in International Law’, 5 International Journal of Language & Law (2016) pp. 15 et sequ.; see also J. Dunoff and M. Pollack (eds.), Interdisciplinary Perspectives on International Law and International Relations: The State of the Art (Cambridge University Press, Cambridge, 2013), passim. See also for an example of a more legal theory-based approach: Maks Del Mar, ‘Metaphor in International Law: Language, Imagination and Normative Inquiry’, 86.2 Nordic Journal of Int. Law (2017) xx–xx., this issue.


K. McAuliffe, ‘Language and Law in the European Union: The Multilingual Jurisprudence of the ecj’, in P. Tiersma and L. Solan (eds.), The Oxford Handbook of Language and Law (Oxford University Press, Oxford, 2012) p. 200.


M. Freeman and F. Smith, ‘Law and Language: An Introduction’, in M. Freeman and F. Smith (eds.), Law and Language - Current Legal Issues Volume 15 (Oxford University Press, Oxford, 2013) p. 5. In our mentioned previous article, we described the fact that even those international legal scholars who do engage with linguistics have not yet fully exploited the potential of pragmatics for the study of interpretation in international law, Smolka and Pirker, supra note 3, p. 14 et sequ. See for an alternative linguistic approach to international law based on corpus linguistics Kjaer’s contribution in this special issue.


We deliberately focus on the interpretation of treaties as it is a well-confined area of debate with widely agreed upon rules. Of course, one could also discuss interpretation in other contexts, see e.g. on the interpretation of customary law. P. Merkouris, Article 31(3)(c) vclt and the Principle of Systemic Integration - Normative Shadows in Plato’s Cave (Brill, Leiden, 2015) pp. 263 et sequ.


Smolka and Pirker, supra note 3, p. 14 et sequ.


D. Sperber and G. Origgi, ‘A pragmatic perspective on the evolution of language’, in D. ­Wilson and D. Sperber (eds.), Meaning and Relevance (Cambridge University Press, Cambridge, 2012) p. 331.


J. Mey, ‘How to do good things with words: A social pragmatics for survival’, 4:2 Pragmatics (1993) p. 247, quoting Yehoshua Bar-Hillel.


A. Reboul and J. Moeschler, Pragmatique du discours. De l’interprétation de l’énoncé à l’interprétation du discours (Armand Colin, Paris, 1998) p. 17.


R. Carston, ‘Legal Texts and Canons of Construction: A View from Current Pragmatic Theory’, in M. Freeman and F. Smith (eds.), Law and Language (Oxford University Press, Oxford, 2013) p. 8.




Ibid., p. 22.


Smolka and Pirker, supra note 3, pp. 3–4.


Ibid., p. 2.


Carston, supra note 11, p. 30.


Ibid., p. 8.


L. Horn, ‘The border wars: A neo-Gricean perspective’, in K. von Heusinger and K. Turner (eds.), Where Semantics Meets Pragmatics (Elsevier, Amsterdam, 2006) p. 21.


See Stanford Encyclopedia of Philosophy, Pragmatics, section 3.4, , visited on 1 December 2016, quoted hereinafter as “Stanford Encyclopedia”. See also Smolka and Pirker, supra note 3, pp. 4 et sequ.


See next section.


Carston, supra note 11, p. 9. For a brief history of pragmatics, see Smolka and Pirker, supra note 3, pp. 4 et sequ.


Carston, supra note 11, p. 9, emphasis added.


A. Reboul and J. Moeschler, La pragmatique aujourd’hui: une nouvelle science de la communication (Editions du Seuil, Paris, 1998) p. 59.


Carston, supra note 11, p. 9.




Reboul and Moeschler, supra note 23, p. 50.


D. Blakemore, Understanding Utterances: Introduction to Pragmatics (Blackwell, Oxford, 1992) pp. 3–10.


D. Sperber and D. Wilson, Relevance. Communication and Cognition (Blackwell, Oxford, 1995) p. 21.


Blakemore, supra note 27, p. 5–6.


See Stanford Encyclopedia, supra note 19.


D. Wilson and D. Sperber, ‘Relevance Theory’, in L. Horn and G. Ward (eds.), The Handbook of Pragmatics (Blackwell, Malden, 2006) p. 611.


Ibid., p. 607, emphasis added.




J. Moeschler, ‘Pragmatics, Propositional and Non-Propositional Effects. Can a Theory of Utterance Interpretation Account for Emotions in Verbal Communication?’, 48:3 Social Science Information (2009) p. 452; see also below. In our previous paper, we adopted the term “by-product” from Moeschler (Smolka and Pirker, supra note 3, pp. 9 and 21). We suggest replacing it by the term “joint product” because the term “by-product” might be misread as “derivative or supplementary interpretive product or result”, whereas it is essential in arriving at a correct interpretation.


Reboul and Moeschler, supra note 23, p. 72.


S. Zufferey and J. Moeschler, Initiation à l’étude du sens (Sciences Humaines Éditions, ­Auxerre, 2012) p. 108.


Wilson and Sperber, supra note 31, p. 607.




Ibid., pp. 611–613.


Ibid., p. 607.


Carston, supra note 11, p. 27.


Wilson and Sperber, supra note 31, p. 608.


Sperber and Wilson, supra note 28, p. 261.


Carston, supra note 11, p. 28.


Ibid., p. 27.


Sperber and Wilson, supra note 28, p. 261.


Moeschler, supra note 34, p. 455, quoting Wilson and Sperber, emphasis added.


Carston, supra note 11, p. 31. For a discussion on whether Relevance Theory might present a reductionist view of human cognitive systems and whether this view might be complemented by looking to emotion-oriented psychology to examine whether there might be a more psychologically realistic description of the motivational, affective and cognitive factors behind improving representational contents rather than this being “simply a matter of making the most efficient use of the available processing resources [… which] is something we would all want to do, given a choice” (Wilson and Sperber, supra note 31, p. 610, emphasis added), see Smolka and Pirker, supra note 3, pp. 30 et sequ. (section 5.5); see also J. Smolka, Cognitivism on the Couch: Revisiting Auchlin’s Experiential Approach to Pragmatic Discourse Analysis (Lehmanns, Berlin, 2014). This discussion might help to better explain how, in Relevance Theory, interpretation rests on a choice which involves the preferences of the utterer. Similarly, this discussion may help to better model the motivational force behind those preferences.


Wilson and Sperber, supra note 31, p. 608.


Carston, supra note 11, pp. 27–28.


Sperber and Wilson, supra note 28, pp. 59–60.


Ibid., p. 56.


Ibid., p. 58.


Ibid., p. 1.


Moeschler, supra note 34, p. 456, emphasis added.




Sperber and Wilson, supra note 28, p. 41.


Ibid., p. 44.


Carston, supra note 11, p. 28.


Wilson and Sperber, supra note 31, p. 614.


D. Wilson, ‘Relevance Theory and Lexical Pragmatics’, 15:2 Italian Journal of Linguistics (2003) p. 282.


Zufferey and Moeschler, supra note 36, p. 108.


Moeschler, supra note 34, p. 454. Sperber and Wilson highlight that relevance is comparative rather than quantitative: “even when absolute measures exist …, we generally have access to more intuitive methods of assessment which are in some sense more basic. I[t] therefore seems preferable to treat effort and effect … as non-representational dimensions of mental processes …; and when they are mentally represented, it is in the form of intuitive comparative judgements rather than absolute numerical ones”, Wilson and Sperber, supra note 31, p. 610, emphasis omitted.


Wilson and Sperber, supra note 31, p. 608.


ibid., p. 608.


Moeschler, supra note 34, p. 455.


Carston, supra note 11, p. 28, emphasis omitted.


Moeschler, supra note 34, p. 453, citing S. Baron-Cohen, Mindblindness: An Essay on ­Autism and Theory of Mind (mit Press, Cambridge, 1995) as a representative of this approach.


J. Moeschler, ‘Is Pragmatics of Discourse Possible’, in A. Capone (ed.) Perspectives on Language Use and Pragmatics. A Volume in Memory of Sorin Stati (Lincom Europa, ­Munich, 2010) p. 223, emphasis added, citing D. Dennett, The Intentional Stance (mit Press, ­Cambridge, 1987), as the scholar who coined this concept.


J. Moeschler and A. Auchlin, Introduction à la linguistique contemporaine (Armand Colin, Paris, 2009) p. 178.


Reboul and Moeschler, supra note 10, p. 48; see also below.


Wilson and Sperber, supra note 31, pp. 610 and 623.


Reboul and Moeschler, supra note 10, p. 47.


Sperber and Wilson, supra note 28, pp. 18 and 41.


Ibid., pp. 17–21.


Ibid., p. 39.


Ibid., p. 41.


Reboul and Moeschler, supra note 10, pp. 49–50.


Moeschler and Auchlin, supra note 68, p. 180.


Wilson and Sperber, supra note 31, pp. 614.


Cf. Carston, supra note 11, pp. 12–13.


Wilson and Sperber, supra note 31, p. 613.


Carston, supra note 11, p. 13.


Blakemore, supra note 27, pp. 6–7.


R. Carston, ‘Relevance Theory and the Saying/Implicating Distinction’, in L. Horn and G. Ward (eds.), The Handbook of Pragmatics (Blackwell, Malden, 2006) p. 636.


Cf. Smolka and Pirker, supra note 3, p. 28.


Carston, supra note 82, p. 654.


B. Birner, Introduction to Pragmatics (Wiley-Blackwell, Chichester, 2013) p. 16, emphasis omitted.


Cf. B. Clark, Relevance Theory (Cambridge University Press, Cambridge, 2013) pp. 306–308.


Wilson and Sperber, supra note 31, pp. 613, emphasis omitted and added. The addressee’s expectation of relevance corresponds to a hypothesis about the utterer’s meaning which satisfies the presumption of relevance conveyed by the utterance (ibid., p. 615).


Ibid., p. 615.


Ibid., p. 617.


Ibid., p. 615.


J. Moeschler and A. Reboul, Dictionnaire encyclopédique de pragmatique (Seuil, Paris, 1994) pp. 124–125.


The example illustrates pragmatic enrichment with the help of procedural meaning. The distinction between conceptual and procedural meaning is another element not typically covered by legal scholars. In the statement “She carried it in her hand”, she, it and her are not descriptive, as would be the case with conceptual meaning. Procedural meaning is, for instance, contained in personal pronouns and conjunctions, which guide the addressee’s interpretation. In the above statement, personal pronouns instruct the addressee to determine their referent. This is because she and it do not correspond to definite concepts, but mark an unoccupied place where a concept might go (Sperber and Wilson, supra note 28, pp. 72–73). International lawyers often neglect procedural meaning in favour of conceptual meaning. This is why, to keep matters simple, we also continue to focus on conceptual meaning.


Carston, supra note 82, p. 633.


Zufferey and Moeschler, supra note 36, p. 176.


Ibid., p. 180.


Carston, supra note 82, p. 633.


Ibid., p. 636.


K. Börjesson, The Semantics-Pragmatics Controversy (De Gruyter, Berlin/Boston, 2014) p. 120.


Moeschler and Auchlin, supra note 68, p. 179.


Zufferey and Moeschler, supra note 36, p. 122.


Sperber and Wilson, supra note 28, p. 199.


Cf. Wilson and Sperber, supra note 31, pp. 620–621.


U. Linderfalk, ‘The Functionality of Conceptual Terms in International Law and International Legal Discourse’, 6:2 European Journal of Legal Studies (2013/2014) p. 33.


Sperber and Wilson, supra note 28, p. 199.


Carston, supra note 82, p. 633.


Zufferey and Moeschler, supra note 36, p. 180.


There is “an ambiguity in the word ‘semantics’ that too often goes unnoticed”, which, in turn, makes the context-dependence/independence distinction even less clear-cut (K. von Heusinger and K. Turner, ‘(By way of an) Introduction: A first dialogue on the semantics-pragmatics interface’, in K. von Heusinger and K. Turner (eds.), Where Semantics Meets Pragmatics (Elsevier, Amsterdam, 2006) p. 2): semantics can refer both to context-independent words or sentences, but also to “what is said” in utterances (ibid., p. 1, quoting Stalnaker). In Relevance Theory, which draws on Grice’s theory, this distinction translates into “a conception of semantics as actually two separate, though related systems that interact with a pragmatics system” (Börjesson, supra note 101, pp. 302–303): “whereas linguistic semantics deals with the lexical meanings of linguistic expression and their composition, it is in real semantics that propositions generally are semantically interpreted” (ibid., p. 298).


See Stanford Encyclopedia, supra note 19, section 3; emphasis added.


Cf. Smolka and Pirker, supra note 3, p. 2.


Carston, supra note 11, pp. 16–17, quoting A. Marmor, ‘The Pragmatics of Legal Language’, 21:4 Ratio Juris (2008).


Carston, supra note 11, p. 17; see also below.


Wilson and Sperber, supra note 31, pp. 611–613.


Carston, supra note 11, p. 17. Carston highlights that cooperation and strategy, or competition, are not necessarily mutually exclusive: “legal theorist … Hart discusses joint practices of a rule-governed sort in which the participants, although competing with one another, must cooperate in following the rules that constitute the practice or otherwise the practice could not be maintained. … See H. L. A. Hart, ‘Are there any natural rights?’ (1995) 64(2) The Philosophical Review 175” (ibid.).


B. Bix, ‘Legal Interpretation and the Philosophy of Language’, in P. Tiersma and L. Solan (eds.), The Oxford Handbook of Language and Law (Oxford University Press, Oxford, 2012) p. 152.


U. Linderfalk, On the Interpretation of Treaties. The Modern International Law as Expressed in the 1969 Vienna Convention on the Law of Treaties (Springer, Dordrecht, 2007), supra note 27, p. 96.


Bix, supra note 109, p. 153.


Carston, supra note 11, p. 20.


The notion of the intentional stance leads us back to the afore-mentioned argument that legal communication is strategic: optimal relevance does not mean that utterers will make their utterances “as easy as possible to understand” (Wilson and Sperber, supra note 31, pp. 613–614) per se. Rather, since utterers cannot normally want communication to fail in theoretically possible, though unforeseeable, future contexts, they will strategically make the formulation of their utterances as relevant to these contexts as possible, e.g. potentially vague as to facilitate the temporally deferred interpretive process. The notion of least effort is to be understood against the backdrop of the background knowledge of the expert (legal) utterer, who can be assumed to have sophisticated expectations of ­relevance and is, therefore, able to deal with mistakes and bad faith (ibid., p. 625).


R. Benett, ‘Constitutional Interpretation’, in L. Solan and P. Tiersma (eds.), The Oxford Handbook of Language and Law (Oxford University Press, Oxford, 2012) p. 117.


Sperber and Wilson, supra note 28, p. 58, emphasis added; see also Smolka and Pirker, supra note 3, pp. 24–25. When producing an utterance, an utterer will often have several intentions. To return to our example from section 2, when Peter says “It is four o’clock” to Mary, the addressee should recognise that, depending on the context, they will be late for a class that starts at four o’clock, or that they still have time for a coffee before an appointment at five o’clock. The utterance may come with further intentions: that the addressee should think that the utterer cares about her and does, therefore, not want either of them to be late, or that the utterer is a good friend who loves spending time with the addressee, etc. Since the first of these intentions is normally much more overt and accessible, it is plausibly a correct interpretation which satisfies the interpreter’s expectations of relevance, and she is therefore licensed to stop interpreting (cf. Carston, supra note 11, p. 24). The distinction between strong and weak implicatures thus helps Relevance Theory to account for multiple intentions. An utterance can achieve relevance in a variety of ways: “Some utterances (e.g. technical instructions) achieve relevance by conveying a few strong implicatures. Others achieve relevance by weakly suggesting a wide array of possible implications, each of which is a weak implicature”, e.g. in poetic uses of language (Wilson and Sperber, supra note 31, p. 621).


Carston, supra note 11, p. 25.


Ibid., quoting Dworkin’s comment in A. Scalia, A Matter of Interpretation. Federal Courts and the Law - An Essay (Princeton University Press, Princeton, 1998) p. 17.


Bix, supra note 109, p. 155.


Wilson and Sperber, supra note 31, p. 610.


Linderfalk, supra note 27, p. 67.


H. Mattila, Comparative Legal Linguistics (Ashgate, Farnham, 2013) pp. 72, 119–127. For example, English language legal discourse is said to include the following characteristics: specialised lexis, where words acquire meaning via definition or institutionalised use; Latin and archaic expressions, especially Norman French; doublets and triplets; complex prepositional and nominal phrases; the use of directive shall; peculiarities of style, e.g. self-referentiality, formality, verbosity (I. Witczak-Plisiecka, ‘Speech actions in legal contexts’, in M. Sbisà and K. Turner (eds.), Pragmatics of Speech Actions (De Gruyter Mouton, Berlin, 2013) p. 183). Carston expects “at certain points in a legal text [in any language], a judicious use of linguistic indeterminacy and/or vagueness, even beyond that which is inherent to language itself, where law-makers have recognized that they cannot envisage the full range of possible future situations that a law might need to be applied to. They may deliberately choose linguistic expressions which will leave open to future judicial decision whether or not the law is to be applied in particular cases. See T. Endicott, Vagueness in Law (Oxford: Oxford University Press, 2000) and papers in A. Marmor and S. ­Soames (eds), Philosophical Foundations of Language in the Law (Oxford: Oxford University Press, 2011)” (Carston, supra note 11, p. 19). In contrast to other peculiar stylistic features in the legal domain, such as long-winded sentences, whose historically developed existence may not be necessary for comprehension but can be justified by their processing ease and accessibility to the expert legal utterer who is familiar with these features, indeterminacy and vagueness may be more closely related to the normative nature of legal texts (see below).


Linderfalk, supra note 27, pp. 43–44.


Reboul and Moeschler, supra note 23, p. 55.


The idea of a mutual cognitive environment must not be overextended in this context since the intentional stance is fallible. The usefulness of the concept is best illustrated when the same rules of law are interpreted by different kinds of lawyers. When lawyers have divergent legal cognitive environments, they may argue over the correct interpretation. This becomes clear when we examine the example of the interaction between two legal regimes: human rights law and international humanitarian law. Although the semantic elements of the applicable norms do not vary, the legal cognitive environment of the human rights lawyer and the international humanitarian lawyer are not the same, and these divergences have become visible in the case law, for instance in the interpretation of different norms prohibiting the deportation of civilian populations in situations of war, see Smolka and Pirker, supra note 3, pp. 26 et sequ. for a discussion of this example.


Carston, supra note 11, p. 19.


Linderfalk, supra note 27, p. 48.


Carston, supra note 11, p. 19.


Ibid., p. 17.


Ibid., pp. 16, 18, 27 and 33.


Ibid., p. 32.


1155 unts 331, 8 ilm 679, entered into force Jan. 27, 1980.


We mainly base ourselves on a number of current “classical” accounts of the Vienna Convention and the rules on interpretation in international legal doctrine, without claiming to be able to treat the vast literature on the topic exhaustively. Also, our focus lies on Articles 31 and 32 vclt as the central, although certainly not exclusively relevant norms.


Again, we opt to present certain “prominent” rules to demonstrate the usefulness of our approach rather than to aim for exhaustiveness at this point. We thus leave aside e.g. questions of treaties concluded in several languages (see also on the role of English in international law Tomuschat’s contribution in this special issue).


A. D. McNair, The Law of Treaties (Clarendon Press, Oxford, 1961) p. 365 note 1. See for the current approach e.g. R. Gardiner, ‘The Vienna Convention Rules on Treaty Interpretation’, in D. B. Hollis (ed.) The Oxford Guide to Treaties (Oxford University Press, Oxford, 2012) p. 478.


R. Gardiner, Treaty Interpretation (Oxford University Press, Oxford, 2010) p. 21.


O. Dörr, ‘Article 31’, in O. Dörr and K. Schmalenbach (eds.), Vienna Convention on the Law of Treaties - A Commentary (Springer, Berlin, 2012) p. 536 para 29; B. Çalı, ‘Specialized Rules of Treaty Interpretation: Human Rights’, in D. B. Hollis (ed.) The Oxford Guide to Treaties (Oxford University Press, Oxford, 2012) pp. 546–547.


A. Aust, Modern Treaty Law and Practice (Cambridge University Press, Cambridge, 2013) p. 208. Others speak similarly of a “single combined operation“ (Dörr, supra note 136, p. ­541 para 39) or a “process of progressive encirclement“, see Aguas del Tunari v Bolivia, 21 October 2005, icsid Case No. arb/02/03, Award, para 91, <>, visited on 1 December 2016.


Aust, supra note 137, p. 209.


Gardiner, supra note 134, p. 480.


Gardiner, supra note 135, p. 164.


Dörr, supra note 136, p. 534 para 25, and p. 535 para 27 with special reference to the European Convention on Human Rights.


Ibid., p. 543 para 44.


See concisely Aust, supra note 137, pp. 210–212.


See comprehensively G. Nolte (ed.), Treaties and Subsequent Practice (Oxford University Press, Oxford, 2013), passim.


See on these three elements Gardiner, supra note 134, pp. 482–487.


See with examples Dörr, supra note 136, pp. 544–545 para 48 et sequ.


Gardiner, supra note 135, p. 189.


Dörr, supra note 136, p. 546 para 56.


Gardiner, supra note 135, p. 190.


See on the lengthy debate on the adequate role of preparatory work Gardiner, supra note 134, pp. 487–490. See on the difficulties in actual practice Aust, supra note 137, p. 218.


U. Linderfalk, ‘Is Treaty Interpretation an Art or a Science? International Law and Rational Decision Making’, 26:1 European Journal of International Law (2015) pp. 173–174, speaks of “first-order” and “second-order assumptions”.


Dörr, supra note 136, p. 538 para 33.


See as the locus classicus ss ‘Wimbledon’, 17 August 1923, pcij, Judgment, pcij Series A No 1, 1923, p. 24, <>, visited on 1 December 2016.


Dispute Regarding Navigational and Related Rights (Costa Rica v Nicaragua), 13 July 2009, icj, Judgment, icj Reports 2009, p. 213, para 48. Sceptical with regard to human rights treaties Gardiner, supra note 135, p. 349, <>, visited on 1 December 2016.


Iron Rhine (‘Ijzeren Rhin’) Railway Arbitration (Belgium v Netherlands), 24 May 2005, 27 riaa 35, para 49, XXVII/35–125.pdf>, visited on 1 December 2016.


Gardiner, supra note 134, p. 496; Dörr, supra note 136, p. 539 para 35.


Aust, supra note 137, pp. 220–221 (emphasis omitted).


Smolka and Pirker, supra note 3, pp. 3–4; Carston, supra note 11, p. 22.


Börjesson, supra note 101, p. 114.


Ibid., pp. 68–69.


Ibid., p. 67–68. Note that there is considerable debate in linguistics and pragmatics on the precise nature of encoded word meaning (Carston, supra note 11, p. 12; for an overview see Börjesson, supra note 101).


Carston, supra note 11, p. 22.


Aust, supra note 137, pp. 221–222.


75 unts 3 (No. 972).


Aust, supra note 137, p. 222.


Note that cases of broadening appear to be less widely agreed on than those of narrowing (Carston, supra note 11, p. 12).


Wilson and Sperber, supra note 31, p. 618.


Börjesson, supra note 101, pp. 117–118.


Carston, supra note 11, pp. 11–12.


Aust, supra note 137, pp. 206.


Carston, supra note 82, pp. 633–634, 639 and 649–650. See also section 2; see furthermore Stanford Encyclopedia, supra note 19, section 3.4.


Carston, supra note 11, p. 19.


Börjesson, supra note 101, pp. 3 et sequ.


Cf. ibid., pp. 27–28, 42 and 69.


Ibid., pp. 117–118.


For an overview of the debate between minimalist and contextualist approaches on how much semantic material is coded in the mental lexicon, see ibid., pp. 49 et sequ.


Carston, supra note 11, p. 19.


Ibid., p. 33.


Ibid., p. 15.


Aust, supra note 137, p. 209.


Gardiner, supra note 135, pp. 161–162.


See e.g. for one authoritative treatment of the issue of diplomatic protection C. Amerasinghe, Diplomatic Protection (Oxford University Press, Oxford, 2008).


Barcelona Traction, Light & Power Co. (Belgium vs. Spain), 5 February 1970, icj, ­Judgment, icj Reports 1970, p. 3, p. 42, <>, visited on 1 ­December 2016.


To be eligible for diplomatic protection by Switzerland, a company must indeed be effectively controlled by Swiss nationals or companies, see Note de la Direction du droit international public of 20 August 2002, addressed to the Political Division ii of the Département fédéral des affaires étrangères, reprinted in 13:4 Revue Suisse de Droit International et Européen (2003) pp. 454–455.


Article 8 Regulation (eu) No. 596/2014 of the European Parliament and of the Council of 16 April 2014 on market abuse, Official Journal L 173, 12 June 2014 pp. 1–61.


Erstes Finanzmarktnovellierungsgesetz (hereinafter “1. FiMaNoG”), Bundesgesetzblatt vom 30.6.2016, Jahrgang 2016 Teil I Nr. 31 pp. 1514 et sequ.


See on the exact details of this error during the legislative process C. Rothenfußer and C. Jäger, ‘Generalamnestie im Kapitalmarktrecht durch das Erste Finanzmarktnovellie­rungsgesetz’, 69:37 Neue Juristische Wochenschrift (2016) pp. 2689–2690.


Article 17 (1) 1. FiMaNoG.


The surrounding discussion continuously emphasised that the law should enter into force „as provided for by the relevant eu legislation“, see Rothenfußer and Jäger, supra note 195, p. 2690.


Contrary to what scholars who give central importance to the possible influence of social practices on interpretation in law sometimes claim, see e.g. A. Bianchi, ‘Law, Time, and Change: The Self-Regulatory Function of Subsequent Practice’, in G. Nolte (ed.) Treaties and Subsequent Practice (Oxford University Press, Oxford, 2013) pp. 136–137.


Carston, supra note 11, p. 24.


This fact illustrates the arbitrary or opportunistic nature of this type of constraining or guidance: in the worst case, “you [may only] stop when you run out of time, money, energy”, J. Polich, ‘The Ambiguity of Plain Meaning: Smith v. United States and the New Textualism’, 68:1 Southern California Law Review (1994) p. 288.


Carston, supra note 11, p. 19.


Ibid., p. 15.


Ibid., p. 27.


Ibid., p. 21, quoting Endicott, ‘Law and Language’, from Stanford Encyclopedia of Philosophy, , visited on 1 December 2016.


However, our cognitive abilities allow us to mentally represent information, including assumptions and evaluative judgements related to norms as well as to ethical, social or political issues that may play a role in legal interpretation, and Relevance Theory considers this information to be part of cognitive context (cf. Smolka and Pirker, supra note 3, pp. 29 et sequ.).


Carston, supra note 11, p. 8.


Linderfalk, supra note 118, pp. 33–34.


Carston, supra note 11, p. 33, quoting Polich.




Sperber and Wilson, supra note 28, p. 47.


Carston, supra note 11, p. 30.


Sperber and Wilson, supra note 28, p. 151.

If the inline PDF is not rendering correctly, you can download the PDF file here.