Introduction

In The Concept of a Legal System Joseph Raz states:

By enacting a statute, making regulations, etc., the authorities create only part of a norm, the other parts of which may have been created at other times, perhaps even hundreds of years before, and often by other bodies. (…) By enacting a constitution, making a statute or a regulation, etc., the legislator creates not only a part of one norm but a part of many norms, usually of a very great number of norms. (Raz 1980, p. 70)

In the light of these facts, Raz poses unresolved Benthamian questions about the individuation of laws, viz. ‘Wherein consists the identity and completeness of a law?’, ‘What is a law?’, and ‘What are the parts of a law?’ (Raz 1972, p. 825). Contrary to those commentators who question the theoretical or practical usefulness of resolving these issues (Bix 2004, p. 98), I contend that their resolution is crucial to understanding the nature of legal interpretation. Unfortunately, jurisprudence has so far not been theoretically equipped for this task.

I first present a theory that can solve the problem of individuation. This combines the work of the Polish legal theorist Maciej Zieliński, and the Dutch linguist Teun Van Dijk, both of whom have advanced theories that are considerably more elaborate and comprehensive than Bentham’s insights. The crucial issue in solving the conundrum of individuation lies in understanding why and how those who interpret the law move from the level of a legal provision (as expressed by a clause in a legal text) to the level of a legal rule, and then apply it. I contend that this transition has to be understood in order to clarify the nature of legal interpretation as a process of deriving practical information from a complex discourse produced by the lawmaker and determining its applicability to a particular case.

I next apply the Zieliński/Van Dijk model to explain the nature of the ‘interpretation-construction’ distinction (‘ICD’). This is another theoretical model of how legal language affects judicial decisions (Solum 2010, p. 103). The ICD conceptualizes the process of decoding a rule from the provisions of a legal text, which might be better understood as deriving a semantic macrostructure from the lawmaker’s discourse. When seen from the perspective of the Zieliński/Van Dijk model, the ICD can be further theorized and the processes involved in transitioning from the level of interpretation to the level of construction can be better understood.

An Old Conundrum

The problem of individuation has been discussed by some of the greatest legal minds in history, including Bentham (1970), Kelsen (1911), Raz (1980, 1972b) and Dworkin (1972). Succinctly put, it involves individuating the elements of ‘the law’. The individuation of laws has mostly been treated as an ontological question, viz. how to determine the smallest ‘building block’ of the law. A case in point is Bentham’s analysis (1970, p. 12), as it answers typical ontological questions, such as ‘What is a law?’ or ‘What are the parts of a law?’ Raz (1972b, p. 825) similarly understands the problem of individuation as an ontological one, viz. ‘What is to count as one complete law?’, and emphasizes that it is a problem of theoretical classification. According to Raz, this is the preserve of legal philosophers; practically oriented lawyers are more concerned with the content of the law than the formal individuation of laws (ibid.).

Dworkin (1972) regarded the ontological approaches of Bentham and Raz as being beside the point. The problem of individuation, in his view, is expositional, not ontological, in nature. Lawyers sometimes need to group legal provisions into larger wholes for educational purposes, e.g. when compiling textbooks. Dworkin demonstrates the insignificance of the individuation problem in such cases with the example of someone who had just read a geological book being asked not what information it contained, but the number of propositions of fact and the theoretical basis for counting them (Dworkin 1972, p. 886).

Paradoxically, Dworkin’s example inadvertently reveals why individuation is crucial to law. Just as no one is expected to count the number of sentences a geology book contains, no one pays a lawyer to be apprised of the number of provisions in a statute. We pay lawyers to explain the possible consequences of a long and involved legal text for our particular situation. This requires being able to identify and interpret the applicable provisions. In other words, the Benthamian question of what constitutes a complete law can be understood as a question of how to order and individuate complex legal content so as to identify all of the components of the lawmaker’s communication applicable to the case at hand. This is where the problem of individuation becomes vital, not for ontological or expositional purposes, but for the fundamental purpose of applying the law.

The application approach to individuation is already present in Bentham (Tusseau 2007), who analysed it to determine the will of the lawmaker on a particular issue—a clearly practical problem that had to be solved before any legal dispute could be settled. This paper similarly treats individuation as a crucial practical issue confronting every lawyer when applying the law to a given case. From this perspective, the problem of individuation can be restated as having to understand the lawmaker’s vast and complex discourse. This discourse covers a multiplicity of issues and caters for several categories of fact situations. Selecting the relevant statutes and provisions, assembling them into a coherent whole (while resolving any vagueness, ambiguity and contradictions) and applying this to the case at hand are, and always have been, eminently practical considerations. Any proposals to neglect this matter of problematics should therefore be opposed.

In what follows, I discuss the problem of individuation, drawing on Zieliński’s theory of legal interpretation and Van Dijk’s theory of macrostructures. These contemporary theories are an insightful continuation of Bentham’s practical approach to the individuation of laws, and enable us to understand the kind of intellectual and cognitive activities involved in construing a legal text for the purpose of applying the law. First, however, I will briefly discuss the linguistic problem that underlies the discussion on the individuation of laws, viz. the problem of syntactic atomism.

The Individuation of Laws and the Problem of Syntactic Atomism

In his discussion on how particular legal provisions contribute to ‘a complete law’, Bentham (1970) was one of the first to abandon the ‘syntactic atomism’ approach, where ‘clauses’, ‘provisions’ and ‘rules’ are considered independent, isolated elements, in analysing legal texts. He realized that legal language should be analysed as a complex discourse involving interrelated, not isolated, utterances.Footnote 1

The atomistic approach did not allow the philosophy of language to be used to appropriately analyse large sets of sentences, i.e. discourses. Traditionally, individual, isolated statements were analysed, with dogged persistence, within the framework of logic, speech act theory, the theory of conversational implicatures and other linguistic theories. The break from syntactic atomism came only recently with the dynamic concept of meaning, especially H. Kamp’s discourse representation theory (Kamp 1981).

The syntactic atomism approach was transferred to jurisprudence. This is evident in the interminable discussion on the most important sentence in the history of jurisprudence, viz. ‘No vehicles in the park’ (Hart 1958). This discussion is all the more remarkable given that no legal system comprising rules that can be construed in complete isolation from each other has ever been identified.

Obviously, an analysis based on syntactic atomism does not take into account the variety of statements contained in legal texts. In addition to statements identifying the addressee, detailing the applicable circumstances and mandating or proscribing specified conduct, legal texts contain many auxiliary statements, e.g. provisions that supplement or modify others, and therefore have to be read in conjunction with them. These include legal definitions, metalinguistic statements (related to the relationship between sentences in the same text), preambles (statements that outline the objectives of the text) and constitutional principles which can be directly applied by the courts.Footnote 2 How does the presence of such statements in a legal text influence the understanding of ‘rules’? Are they rules in their own right? Or do they merely supplement other rules? If they supplement other rules, then how should lawyers relate the statements in a legal text to each other? The lack of an explicit response to these questions means that our understanding of the rules contained in a legal text appears to be barely intuitive and certainly under-theorized.Footnote 3

The considerations of the Polish legal theorist M. Zieliński supplement Bentham’s analyses in this respect. The next section highlights the advantages of Zieliński’s concept, which is much more detailed than Bentham’s. I simultaneously point out that its theoretical potential has not been fully utilized, the reason being that this concept considers the legal rule to be the utterance and not the meaning of the utterance. The removal of this shortcoming by applying T. Van Dijk’s concept of macrostructures makes Zieliński’s concept a fully adequate tool for describing legal interpretation.

Zieliński’s Concept of Legal Interpretation

In his 1972 book Interpretation as the Process of Decoding Legal Texts (written in Polish), Zieliński presents a ‘derivational theory of interpretation’.Footnote 4 This theory distinguishes between two levels of legal language: the language of the provisions of the law (i.e. the sentences in the text) and the language of rules (Zieliński 1972, p. 8; Ziembiński 1960). Legal interpretation involves transforming first-level utterances into a second-level utterance (a legal rule), which is then applied. In other words, understanding a legal text and applying it to the case at hand requires a transition from the provisions of the law to an applicable legal rule, which is an utterance with different semantic, syntactic and pragmatic characteristics.

This first-level language is a code that invariably has to be ‘disassembled’ (ibid., p. 15), i.e. several utterances of the code (provisions of the law) have to be extracted, and then reassembled and translated into the language of a single legal rule. Breaking down the rule encoded in the provisions of the law means that several provisions can comprise the single rule that Zieliński terms the ‘norm of conduct’. The interpreter’s work involves moving from the ambiguous language of the provisions of the law to the language of rules, and choosing one of the possible meanings (ibid., p. 33). The rules derived in this decoding process are unequivocal utterances mandating or prohibiting specific conduct.Footnote 5

A typical example used by Zieliński involves deriving a legal rule on homicide by interpreting the Polish Penal Code (Article 148, para. 1): ‘Anyone who kills a person is liable to imprisonment for a minimum term of eight years, imprisonment for 25 years or imprisonment for life’. This sentence does not constitute the full statement of the lawmaker on homicide, as it neither includes nor refers to any other provisions on homicide, including those on justifications or excuse (e.g. self-defence), types of qualified homicide (e.g. an after-birth infanticide by the mother or mercy-killing, which are more leniently punished), or even mens rea. For instance, Article 148 makes no mention of killing in self-defence. Only by reading Article 148 para. 1 together with Article 25 para. 1 (‘Anyone who, out of necessary self-defence, repels a direct illegal attack on any legally protected interest is not deemed to have committed an offence’) and other relevant provisions, can the full legal rule on homicide be formulated.

Moving from first-level to second-level language involves several steps that Zieliński describes in detail (2010). The operation starts with the selection of the main provision on which the rule is based (base/central provision). The base provisions contain the core of the rule. They most frequently mandate or prohibit certain behaviour in specific circumstances on the part of identified class(es) of legal persons (addressee(s)). Other relevant provisions are identified later. These modifying/supplementary provisions can influence the form of order/prohibition contained in the base/central provision, e.g. by reducing or increasing the range of entities to which it applies or by extending or narrowing the extent to which it can be applied.

Zieliński’s distinction between base/central and supplementary/modifying provisions explains e.g. how legal definitions and provisions containing defined terms are combined in legal interpretation, as well as provisions regarding crimes and legal excuses and justifications for them, or general provisions and exceptions to them. All these are examples that a legal rule arises from the combination of several statements and is therefore the result of a mini-discourse and not a single self-sufficient statement.Footnote 6

After selecting and compiling the relevant provisions, the interpreter creates a ‘rule-like utterance’ by combining the base provision with all relevant modifying and supplementing provisions. The next step is to examine the linguistic clarity of the utterance. Vagueness and ambiguity are removed, e.g. by using dictionaries or legal definitions, and/or the internal and the external context of the legal act (i.e. in which section of the act the base provision is located, how the location of the provision influences or determines its meaning, and relevant case law), etc.

Once a precise rule-like utterance is produced, its compatibility with the purpose of the lawmaker and other general standards is examined. If incompatible, it is modified in order to achieve a version that is optimal in terms of clarity and compatibility with both the purposes of the lawmaker and legal and constitutional principles. The outcome of the interpretative process is a legal rule in the form of an utterance that can be applied to the case at hand.

The final shape of the legal rule can be the result of a combination of statements derived from a variety of legal texts, enacted at different times by different lawmakers, and located in different places in the hierarchy of the law. This legal rule is treated as the complete communication of the lawmaker on the given subject. Again, Zieliński’s concept of a ‘complete’ legal rule resembles Bentham’s concept of a law that is complete ‘in point of expression’, i.e. ‘when all the legal material relating to a particular act of the legislator’s will can be comprehended in the form of an individual law’ (James 1973, p. 372).

The approach proposed by Zieliński is reflected in legal practice, which also distinguishes between a base/central provision and supplementary/modifying provisions. When formulating the legal basis of a court judgment or an administrative decision and when formulating allegations of unconstitutionality and indicating constitutional patterns of control in the practice of the constitutional judiciary, the phrase ‘Article X in connection with Article Y’ is frequently used to show the relationship between several provisions of the law. When using this phrase, lawyers indicate that the legal basis is not a single provision, but rather a legal rule formulated from several provisions that have been decoded and reassembled.

Zieliński’s is a valuable theory of legal interpretation. It is, however, not free of imperfections, two of which particularly stand out. The first is that Zieliński’s legal norm is, following Bentham, always structured as a duty-imposing rule. As such, it does not readily accommodate Hart’s postulate of the variety of laws (Hart 1961). Unfortunately, there is no space here to discuss this problem in detail. However, there is no theoretical obstacle to applying the concept of ‘decoding’ to the provisions from which power-conferring rules (e.g. rules of change and adjudication) are derived. It suffices to assume that depending on the objectives of the interpreter, the syntactic form of the legal rule decoded from the legal text can be ‘Z should do S’ (a duty-imposing rule) or ‘Y is treated as X’ (a power-conferring rule), where ‘Y’ denotes a series of actions performed by an agent, and ‘X’ denotes the legal consequences thereof (e.g. a valid court judgment).

The other flaw is more pertinent to the topic discussed in this paper. Zieliński argues that legal rules, like legal provisions, are utterances, not the meanings of the utterances (Zieliński 1992, p. 105). As is shown below, this assumption is untenable. While provisions of the law are real-life utterances, produced in concrete spatio-temporal contexts, a legal rule is a meaning obtained by combining these utterances—not a sign composed of other signs, but the meaning of an ‘assemblage of signs’ (Bentham 1970, p. 1). Let us see why the ontological status of legal rule is important for understanding the nature of legal interpretation.

Legal Rule: An Utterance or a Combined Meaning of Several Utterances?

The assumption that a legal rule is an utterance is based on the intuitive conviction that the lawmaker regulates the legal situation of the addressees of the law through directive statements addressed to them, and that these are similar to the messages that they send each other in the course of normal communication. Another reason for making this assumption is that utterances are compositional (Szabó 2008): more complex utterances can be built from smaller constituent components, which is obviously desirable from the point of view of the concept of interpretation, which involves the reconstruction of a more complex utterance (a legal rule) from simpler utterances (provisions of the law).

Treating a legal rule as an utterance is susceptible to criticism for syntactic and pragmalinguistic reasons. The syntactic problem is a controversy regarding the actual structure of the rule and involves reasonable doubt as to whether a legal rule, as an utterance, is the full ‘statement’ of the lawmaker, i.e. whether it encompasses all the components expressed in the given area of the regulation. A similar problem can be found in Bentham, who argues that ‘a complete law’ comprises ‘all the legal material relating to a particular act of the legislator’s’ (James 1973, p. 372). This might imply that the criterion of the completeness of a law is not the unity of the subject of a particular law, but the unity of an act of expressing the law by the lawmaker.

However, when applying the law, it is not the full statement of the lawmaker that is decoded from the legal text, but only those components relevant to the facts under consideration. Suppose a provision of criminal law on causing a land, water or air traffic disaster is to be applied.Footnote 7 Only the normative pattern that is relevant to the facts under examination will be reconstructed, e.g. in the case of a land traffic disaster, there is no need, either theoretically or practically, to recreate the full normative pattern, which would include clauses on sea and air traffic disasters (Morawski 2010, p. 17). The other problem is the pragmalinguistic problem, i.e. identifying the person that ‘utters the legal rule’. The natural candidate is the lawmaker. This person is, however, not very well suited to this role, if only because the constituent components from which the rule is to be decoded have been created by various lawmakers (parliaments of different terms of office, the authors of primary and secondary legislation, domestic and supranational lawmakers, etc.). There is therefore no single person who created the whole utterance. One response to this problem is to assume that the legal rule is an utterance ‘even if it is only conceived’ (Zieliński 2010, p. 16) and need not therefore be actually stated. Another is to assume that the rules are uttered not by several real lawmakers but by a single ‘ideal’ or ‘rational’ lawmaker, i.e. to integrate all the real lawmakers into a single virtual lawmaker. However, in the pragmalinguistic sense, the ‘abstract’ or ‘theoretical’ context of an utterance is an oxymoron. The pragmalinguistic context is always the spatio-temporal context in which specific language is used by a specific person.

Yet another way of dealing with the pragmalinguistic problem is to assume that the interpreter is the one who ‘utters’ the rule. This is difficult to accept, because it contradicts the authority of the lawmaker as the entity from whom the legal rule originates and who thereby governs the behaviour of its addressees. If the interpreter of the rule is to be its author, he replaces the lawmaker as the author of the law, which is fatal to the lawmaker’s authority and to the idea that the interpreter is bound by his decisions.Footnote 8

In addition to the syntactic and pragmalinguistic problem, Zieliński’s concept of rule as being an utterance has a semantic problem. There is a noticeable contradiction in considering a legal rule to be an utterance while expecting it to be unequivocal. The assumptions underlying M. Zieliński’s concept allow for the following reasoning, which leads to a contradiction when a legal rule is assumed to be an utterance:

P1. An utterance is a sign.

P2. A sign requires interpretation.

P3. A legal rule is unequivocal.

P4. A legal rule does not require interpretation (it is the product of an interpretation).

K1. A rule is not an utterance.

Premises P1 and P2 are not controversial. Premise P1 is predicated on the conviction that a linguistic action produces an utterance, i.e. an artefact intended to achieve a communicative result, viz. a conventional sign. Premise P2 is an element of a broader omnia sunt interpretanda assumption that is characteristic of Zieliński’s concept of interpretation. This assumption implies that any linguistic action to which meaning can be assigned is subject to interpretation. It is also consistent with assumptions underlying semiotic theory.Footnote 9

As for premises P3 and P4, in order to resolve a given matter when applying the law, the interpretation must result in an unequivocal rule. If, at the end of the interpretation process, the person applying the law concludes that the rule/utterance derived from the legal text still lacks clarity, but nevertheless applies it to a specific case, he would in fact be continuing the interpretation process instead of closing it.Footnote 10 This situation could be presented as follows:

  1. a)

    At the end of the interpretation process, X states that a fragment of text T is an utterance N that can be assigned at least two meanings—Z1 or Z2;

  2. b)

    While applying the law, X acknowledges that N applies to the facts of the case, which is a referent of the Z2 meaning of utterance N;

  3. c)

    This means that X has disambiguated utterance N by acknowledging that it refers to the facts covered by its Z2 meaning, and not by its Z1 meaning.

It therefore appears that conclusion K1 is correct: a legal rule cannot be an utterance (sign). As will be presented below, a legal rule is the meaning of an utterance or rather a meaning of a combination of utterances (signs), i.e. provisions of the law, and as such can be unequivocal.

The Structure of the Meaning vs. the Structure of an Utterance

As we have seen, the reasons why legal rules are perceived as utterances include the conviction that rules qua utterances have a syntactic structure, and can therefore be combinations of simpler utterances (legal provisions). But does this mean that the meaning of the utterance cannot have a structure and cannot be built from simpler components?

What is required is a theory which assumes that a complex structure of meaning can be built from simpler semantic components, as only then could the meaning of several legal provisions be ‘reassembled’. These simpler semantic components would be the meanings of individual provisions of the law, while the complex semantic structure built from the meanings of the individual provisions would be the legal rule. Discourse representation theory (Kamp 1981; Van Eijck and Kamp 1997) fits the bill and has been applied to texts as T. Van Dijk’s theory of macrostructures (Van Dijk 1980).Footnote 11

Van Dijk’s concept is an alternative approach which preserves Zieliński’s intuition regarding the need to put order to the plethora of provisions into a coherent whole. This concept describes the phenomenon of breaking down larger semantic structures into smaller structures that have to be combined to reconstruct this larger structure and it does so in a surprisingly similar way to the derivational concept offered by Zieliński. The difference is that Van Dijk’s concept focuses on the structure of the meaning rather than the structure of the utterance.Footnote 12

Van Dijk’s approach is based on the intuitive conviction that our cognitive activities (e.g. perception, understanding discourse and interaction) are based on the fundamental dichotomy between part and whole, between ‘global’ and ’local’ structures, and between cognitive units interpreted as ‘wholes’ and assemblies of ‘parts’ or ‘elements’ (Van Dijk 1980, p. 4).Footnote 13 For instance, when interpreting a discourse, we combine several semantic elements into a semantic whole by using connecting components such as theme, topic, gist, upshot or point (p. 5). In effect, ‘words and sentences are seen as parts of the discourse, and the theme or topic is seen as a property of the whole’ (ibid.). Van Dijk calls this overall structure, into which individual smaller components are assembled in the process of understanding discourse, ‘global meaning’, ‘global structure’, or simply a macrostructure, i.e. a conceptual structure that organizes detailed cognitive material into a larger whole (ibid.). He emphasizes that this macrostructure ‘is a legitimate part of a linguistic semantics of discourse and not merely a component of a psychological model of discourse processing’ (p. 26).

The relationship between words/sentences and a macrostructure in Van Dijk corresponds to Zieliński’s distinction between the provisions of the law and legal rules. This relationship is presented as a relationship between semantic microstructures (e.g. the meanings of words, phrases, clauses and simple actions) and macrostructures (‘global semantic information only relative to the microstructures of discourse, cognition, and interaction’) (p. 13).

The constituent components that enable us to understand the discourse are micro-speech acts that make up a certain sequence (p. 6). That sequence of speech acts does not, however, become a new, full speech act which, in the pragmalinguistic sense, someone ‘happens’ to ‘perform’. According to Van Dijk, a sequence of mini-speech acts remains a sequence and is understood as a whole, not spoken as a whole. The semantic macrostructure that is created on the basis of the sequence of speech acts is therefore not the speech act itself, but a semantic structure.

A legal rule is a semantic structure of this kind, not an utterance. To be treated as an utterance, it would have to be a new, separate speech act. It would have to be a specific spatio-temporal event and would have to have a single ‘performer’. However, a legal rule is a concatenation of micro-speech acts (legal provisions), performed by several entities. In other words, provisions are spoken or written at specific points in time and space (as Van Dijk emphasizes, micro-acts are of an empirical nature), but nobody ‘utters’ the legal rule as a meta-locutionary act, i.e. a linguistic event taking place in a specific place and at a specific time, because the macrostructure is theoretical (p. 9). In effect, Van Dijk’s theory does not allow for the assumption that a legal rule is an utterance, but enables Zieliński’s basic idea of two levels of legal language to be maintained.

Not only is Zieliński’s distinction between two levels of legal language directly reflected in Van Dijk’s concept, but so is ‘derivation’. As he argues, ‘a macrostructure may be derived or inferred from microstructures’ (p. 15), and it is possible to define rules, operations and transformations between both levels (ibid.). Van Dijk emphasizes that

macrostructures have an essential semantic function. They define higher-level or global meaning derived from lower-level meanings [emphasis added]. This process of derivation may involve the construction of new meaning (i.e., meaning that is not a property of the individual constitutive parts). Hence, as their crucial function, macrostructures allow additional ways of comprehension for complex information. (p. 15)

Therefore, meaning can have structure: larger semantic structures can be created from smaller semantic structures. In the case of legal interpretation, this means reconstructing the meaning of individual provisions of the law and then creating a semantic macrostructure from these meanings, constituting a legal rule. In this concept, a sequence of sentences (discourse) can have a meaning that differs from the sum of the meanings of its component sentences. This is, after all, a distinguishing feature of deriving a legal rule from a set of legal provisions.

Another of Van Dijk’s distinctions, viz. the one between simple and complex facts is also of interest to legal philosophers. ‘Facts’ are ‘chunks of represented reality’ (p. 20), situations represented by the propositions (assertions) that we make or interpret. They may represent ‘certain aspects of “what the case is” in a certain situation: the individuals involved, their properties and their relationships’ (pp. 20–21). A given utterance that takes the form of a sentence represents a situation, event, or action (p. 21), while its structure encompasses a representation of the individual elements of reality.

The complexity of the assertions we make determines the complexity of the facts they represent. The facts may be simple (‘Mary being ill’ or ‘Peter calling his dog’) or more complex (‘The happy boy kissing the blonde girl on her cheek in the back seat of his car’).Footnote 14 Greater complexity may be due to e.g. more participants in a given fact or more ‘modifiers’ (sentences that change or supplement the representation of a fact in an assertion, the equivalents of Zieliński’s modifying provisions). Van Dijk maintains that we tend to express simple facts with single sentences. However, using more sentences to describe a given fact increases its complexity. This also creates connections between facts (situations), e.g. conditional relationships (‘Fact Y is contingent on Fact X’) and the ‘nesting’ of facts. All these relations between facts represented in language are visualized when we interpret the sentences we read or heard (pp. 20–21).

The language and the visualization of the facts in the mind enables them to be combined, separated and broken down into smaller components and reassembled. When analysing the example of the kissing scene given above, Van Dijk argues that the interpreter can easily isolate the fact that the boy was happy or that the girl was blonde or that both the boy and the girl were sitting in the back seat of a car (pp. 20–21). The isolation can be represented by different clauses or sentences in a discourse and, at the same time, by a different speech act (viz. two or more assertions) (ibid.).

Like Zieliński, Van Dijk refers to the possibility of breaking down larger elements and combining smaller elements into a larger whole. In Zieliński’s case, the rules are broken down, whereas in Van Dijk’s case, complex facts represented in language are broken down. The provisions from which the rule is decoded are combined into a larger whole within the framework of the derivational concept of interpretation, whereas in the case of Van Dijk, simple facts are combined into a larger whole of complex facts. The parallel between Zieliński and Van Dijk is no accident. In both cases, we are dealing with the interpretation of language, which takes place by reading sentences and creating a more complex, larger whole on their basis.

Van Dijk’s theory sheds light on the function of the legal rule qua macrostructure, i.e. ensuring that the discourse (the set of legal provisions) is coherent. Without macrostructures, the interpreter would perceive the connections between the pieces of micro-information, but would not be able to arrange them into a larger whole that acknowledges the meaning and function of the discourse, or distinguish one discourse/one action sequence from another (p. 14). However, a discourse is not only coherent at the local level (e.g. at the level of pairwise connections between sentences). Macrostructures at the global level are what make the coherence of the discourse explicit (p. 10).

Reconstructing a legal rule from provisions of the law also ensures that the lawmaker’s discourse is coherent. Reconstruction eliminates the contradictions that can arise between micro-acts, especially if—as in the law—micro-acts (provisions) are introduced into the discourse by different persons at different times. This is precisely why lawyers have a whole range of instruments at their disposal, ranging from conflict rules of the lex specialis derogat legi generali type, to the general requirement to avoid conclusions ad absurdum.Footnote 15

Back to the Problem of Individuation

When Zieliński and Van Dijk’s approaches are merged, legal interpretation can be understood as being based on:

  1. a)

    finding individual utterances in the text which relate to the ‘theme’ (e.g. ‘murder’) and which are relevant to the facts of the case at hand;

  2. b)

    reading these utterances, and decoding their meaning by creating a mental representation of simple facts;

  3. c)

    combining, separating and modifying individual mental representations of simple facts into a larger whole, i.e. into a representation of a complex fact (a macrostructure—a legal rule).

The parallels between Zieliński’s and Van Dijk’s concepts also apply to the internal structure of the larger whole obtained through interpretation. In the case of a legal rule, these are the addressee, the circumstances and the mandate or prohibition. In the case of the semantic macrostructure, these are Agent, Patient, Goal, Object, Beneficiary, Instrument, State, Event, Action and Process (ibid., p. 22). Within this integrated framework, the legal rule is the structured meaning of the set of provisions of the law contained in the text. As such, a legal rule is always a concrete mental representation decoded from the text in a given place at a given time, while its internal abstract structure is a construct of the legal doctrine.Footnote 16 Van Dijk’s concept also offers a solution to the syntactic problem of whether the lawyer reconstructs the whole of the legal rule from the provisions, or only that part relevant to the situation under consideration. Zieliński assumed that the lawmaker’s utterance, by definition, required the reconstruction of one holistic legal rule encoded in the provisions of the law. To assume otherwise would be tantamount to accepting that there are many different rules encoded in the same provisions of the law.

Within the framework proposed by Van Dijk, nothing prevents a mental representation being reconstructed from the legal text that only refers to those facts whose compliance with the law is being assessed.Footnote 17 There is therefore a direct relationship between the facts being assessed and the scope of the legal rule that a lawyer has to reconstruct from the text. The facts of the case trigger the intellectual process of interpretation. It is as if the judge is being asked what to do about a certain fact situation and is required to find the answer in the lawmaker’s discourse, i.e. in the legal text, by selecting those utterances that refer to this type of fact situation. If the facts in question apply to the administrative permit to build a nuclear power plant, it is clear that the scope of the normative pattern to be reconstructed from the legal text is determined by the scope of the facts, and the interpreter selects the provisions of the law that are relevant to the facts under review. The Atomic Law and the Construction Law would definitely need to be referenced, as would certain provisions of the Administrative Procedures Code. However, the Act on Social Welfare would be of no relevance whatsoever.

The principle of individuation proposed by the Zieliński/Van Dijk model therefore differs from the principles proposed to date, including those put forward by Bentham, Kelsen, Raz and Dworkin. Individuation is not an ontological or expositional process, and is not only discussed by legal philosophers who wish to identify the smallest building blocks of a legal system for the sake of human knowledge. The principle of individuation proposed by the Zieliński/Van Dijk model is consummately pragmatic: a law has to be individuated to decide a case. A semantic macrostructure that encapsulates the meaning of that part of the lawmaker’s discourse that is relevant to the facts of the case is derived ad hoc. Even if this macrostructure, which we term a ‘legal rule’, is derived from the same provisions of the law, it may differ from case to case, if the material facts differ.

The Zieliński/Van Dijk model is open to several objections. The first results from its assumption that the legal rule may vary depending on the material facts. It may seem counterintuitive to those used to legal rules being constant and stable. The traditional concept of a legal rule understood as an utterance perceives it to be a permanent and uniform structure. Moreover, it is not natural to assume that the lawmaker would say something different depending on the fact situation assessed by the lawyer. However, if a legal rule is treated as a macrostructure derived from the provisions of the law, considered to be elements of the discourse conducted by lawmakers, then it need not be permanent or unchangeable. This is because it is not an utterance of a real or imaginary person, but rather a tool of thought, which orders the complex and chaotic communication of real lawmakers. This structure makes it possible to extract the meaning required to assess given facts from the legal text—it is the Benthamian ‘intellectual whole’ that should be distinguished from the ‘physical’ (legal text) (Bentham 1960, p. 429). Despite the malleability of the legal rule as a structure/cognitive tool, the discourse in the form of provisions of the law, which are actual messages from real lawmakers, remains unchanged.

The second objection is that the Zieliński/Van Dijk model may lead to excessive subjectivism in legal interpretation. After all, if a legal rule has to be interpreted each time it is applied, and if its form depends on the facts to which it is applied, there is a risk of instability. This issue has been addressed extensively in the theory of mental representation. The key anti-subjectivist argument is that the processes involved in producing, comparing and manipulating mental representations all have ‘proper functions’ (Millikan 2017, p. 223), i.e. functions that have been selected on account of their demonstrated usefulness in a long series of analogous situations in which the organisms that produced the representations were involved (a ‘lineage’—Millikan 2005, p. 38). Because of the similarity and repetitiveness of the stimuli that triggered the mental representations, and the requirement that the mental representations must be fairly stable in order to perform their proper function, the whole process leads to the functions of the representations being ‘stabilising functions’ (Millikan 1984, pp. 77–78). This stabilizing capacity of the process by which mental representations are produced and used minimizes the potential for divergence. The risk of subjectivism is therefore insignificant.

A third objection to the Zieliński/Van Dijk model may be that, for all its sophisticated language and theoretical references, what it proposes is quite simple and obvious—we perceive complex phenomena as consisting of less complex ones, and our language reflects this obvious relationship between the whole and its parts. The two prongs of this objection may be countered by observing that: (i) the simplicity of a theory has no bearing on its validity; and (ii) if the Zieliński/Van Dijk model can address issues that have emphatically not been obvious for many years (at least not to the long list of legal philosophers struggling with individuation), then the charge of obviousness fails. Another answer is that legal philosophy needs a theory, albeit an intuitive one, that provides a useful terminology for analysing how legal language works and how it is interpreted by its addressees. The Zieliński/Van Dijk model links the discussions of legal philosophers with those of discourse theory—a discipline that offers significant insights into human communication and can therefore enrich the toolbox of legal philosophy.

Interpretation/Construction Distinction and the Zieliński/Van Dijk Model of Legal Interpretation

The two-tier nature of legal communication, which distinguishes a provision of the law from a legal rule, resembles the ‘interpretation/construction’ distinction adopted in the American theory of legal interpretation (Whittington 1999; Solum 2010).Footnote 18 The interpretation component of the ICD ‘recognizes or discovers the linguistic meaning of an authoritative legal text’ (Solum 2010, p. 100), i.e. the recreation of the semantic content of a fragment of a legal text. The construction component of the ICD, per contra, entails extracting and/or assembling the legal content, which can differ from the semantic content. A text is given legal effect when a court translates the semantic content of the text into its corresponding legal content, and then applies it to a particular case (Solum 2008, p. 67 et seq.).

Similarly to the Zieliński/Van Dijk model, the ICD distinguishes between two levels of legal language—the language of the provisions and their meaning in the form of legal rules. In the former case, the interpreter has to determine the semantic content of the provisions of the law, whereas in the second, he has to establish a legal rule that embodies the legal effect of a legal text, e.g. by defining the obligation(s) of the addressee. In this approach, ‘interpretation’ can be seen as the process of understanding individual provisions of the law and ‘construction’ as the set of processes by which the semantic macrostructure is derived from them.

The Zieliński/Van Dijk model can be useful in supporting the ICD and in giving it a more advanced methodological shape. The theoretical strength of the model is that moving from interpretation to construction is transparently defined as a process of decoding macrostructures (legal rules) from microstructures (legal provisions). The ICD assumes that this transition takes place, but is silent as to the cognitive process by which it does so. Therefore, the Zieliński/Van Dijk model provides an explanation for the relationship between interpretation and construction; one that clearly shows the linguistic foundations of the transition from the level of provisions to that of rules.

In particular, the model allows for a more theoretically advanced presentation of ‘construction’. This process extends beyond the determination of the meaning of individual provisions primarily because of the need to combine the content of individual provisions into a single semantic macrostructure. This combination is a complex process in which the lawyer also uses information other than information on the semantic content of the individual provisions. Construction frequently involves modifying the semantic content of individual provisions of the law and even eliminating these provisions (e.g. by applying the lex superior or lex specialis rules). Construction enables contradictions to be avoided and gaps to be filled. It interprets law systemically on the basis of arguments ‘from the overall structure of the text’ (Solum 2010, p. 107).

The transition from interpretation to construction is the crux of every lawyer’s work. This is where the vastness and complexity of an abstract legal text is transformed into a practical legal rule that can be used to determine whether a particular state of affairs has been deemed impermissible by the lawmaker. A variety of factors is involved in this transition. First, the nature of the state of affairs to which the law is to be applied determines the provisions to be interpreted. The state of affairs of ‘murder’ and the state of affairs of ‘killing in self-defence’ call for different combinations of the provisions of the law. As different states of affairs require different combinations of provisions, the provisions within a particular combination are transformed into different macrostructures. Even if meanings can be attributed to the abstract legal provisions taken in isolation, applying them requires that the relevant ones be merged into a whole that represents the meaning of the mini-discourse. The meaning of the whole is not simply the sum of the meanings of its parts.

Second, the moment of applying the law is when the vagueness and ambiguity of legal language have to be removed. As shown above, an ambiguous or vague provision cannot be applied; it has to be made precise. Application, understood as establishing that term X refers to state of affairs Y, removes any ambiguity and vagueness, because there is no longer any doubt as to the applicability of term X in the given situation. One way to remove ambiguity and vagueness is to compare the current state of affairs with the previous states of affairs to which a particular term has been applied (Recanati 2004, pp. 150–151), and to apply it to the facts in a way that ‘use has approved’ (Endicott 2020). When making the transition from interpretation to construction, the interpreter therefore has to take into account the previous practice of applying the same terms. In case law, this practice involves examining previous court decisions and studying doctrinal discussions on the proper applicability of the terms being interpreted. Consequently, the transition between interpretation and construction allows for the semantic content of the legal provisions to be infused with additional elements taken from the precedent and the legal literature.

The transition from interpretation to construction, as seen within the framework of the Zieliński/Van Dijk model, shows how the semantic content of the legal provisions is enriched by both syntactic and pragmatic factors. Interpreting the provisions of the law establishes their semantic meaning. Construction enables a more complex structure to be built out of these semantic building blocks and their meaning to be made more adequate to the case to be decided.

Contrary to the claims of the ICD critics (Schauer 2019), the construction stage, as seen from the perspective of the Zieliński/Van Dijk model, is an element of the internal perspective on legal decision-making; one firmly based on the authoritative text of law. As the macrostructure derived within the construction stage is a meaning of the lawmaker’s textual discourse, it cannot be treated as an element of the external perspective, associated with legal realism, in which construction is perceived not as a result of having consulted the relevant texts, but as an ‘all things considered’ judicial decision about a desired result of the decision-making process (Schauer 2019, p. 6).

Both opponents and proponents of the ICD believe that it is based on the idea that ‘the meaning of language within law must be distinguished from what the law ought to do with that language on a particular occasion’ (Schauer 2019, p. 19; Solum 2010, p. 95). The Zieliński/Van Dijk model shows that this is not necessarily the case: the ICD proves that the meaning of the most locally applicable legal provision has to be distinguished from the meaning of the legal text—understood as the aggregate of all relevant legal documents that are binding in a particular jurisdiction at a particular time—which is a representation of the entirety of the lawmaker’s discourse.

The Zieliński/Van Dijk model also helps us understand the dispute as to whether the communicative content of legal language is sufficient for legal interpretation (Greenberg 2010). The answer given by the Zieliński/Van Dijk model is as follows: the critics of the Standard Picture may be correct in saying that the semantic content of the legal provisions is insufficient for the purposes of legal interpretation. It has to be supplemented through syntactic and pragmatic operations, in particular by selecting the relevant provisions, interpreting them as a whole once a macrostructure has been created, and removing any ambiguity and vagueness from the terms used in the provisions. This last process may involve reviewing the instances of these terms as used in previous judicial decisions.

On the other hand, however, the criticism of the Standard Picture is not fully justified. Contrary to what the critics claim, a sufficiently sophisticated linguistic approach based on discourse representation theory can fully explain the workings of legal interpretation and application. For example, the Zieliński/Van Dijk model allows arguments seemingly taken from outside the authoritative text of the law, e.g. previous court decisions and doctrinal disputes, as nevertheless linguistic and therefore able to assist in decoding the semantic macrostructure from the authoritative text of the law.Footnote 19 As such, the Zieliński/Van Dijk model, with its two-level approach to legal communication, is a theoretically advanced description of the construction process and a holistic and historical approach to the legal text as the entirety of the lawmaker’s discourse. It is therefore a useful tool for both practitioners and philosophers of law.