Introduction

By contrasting the external and internal perspectives one can adopt towards a legal system, H.L.A. Hart in The Concept of Law brings to light the conceptual distinction, commonly adopted by legal theorists today, between external legal statements and internal legal statements. This distinction is succinctly captured in Kevin Toh’s remark that the former are ‘statements about individual laws or legal systems’ uttered by an observer of the legal system while the latter are ‘statements of law’ uttered by an adherent, typically an official, of the legal system (emphases in original).Footnote 1 ‘What are legal officials doing when they utter internal legal statements?’ is a question that has attracted much attention in analytical jurisprudence, presumably because as Jules Coleman points out, our answer to this question has important implications for whether we regard law as having legitimate authority.Footnote 2

Hart, as one of the most important figures in Anglo-American jurisprudence, has also tackled this question. However, jurisprudes today differ in their interpretations of Hart’s answer: drawing upon the philosophy of language and metaethics to reconstruct Hart’s view, they disagree as to whether Hart should be interpreted as an expressivist or quasi-expressivist. In this paper I propose a third reconstruction, under which Hart adopted an inferentialist analysis of the semantics of internal legal statements. In executing this reconstruction, I focus on Hart’s inaugural lecture, and utilize the theoretical apparatus of Robert Brandom, a leading advocate of inferential role semantics of our time, to bring Hart’s methodological pragmatist—and more specifically inferentialist—insight in that lecture to the fore.

A preliminary point to note from the outset: in their expositions of Hart’s analysis of internal legal statements, jurisprudes use ‘internal legal statements’ to refer to two types of statements of law: first, statements of legal validity and secondly, what Toh calls ‘statements that enunciate and apply rules’.Footnote 3 This usage follows Hart’s own practice of calling both types of statements ‘legal statements’Footnote 4 or ‘internal statements of law’.Footnote 5 The crucial difference between these two types of statements is this: in uttering the first type of statements legal officials are applying the rule of recognition (R) of a legal system to determine whether a particular rule (L) is valid according to R, while in uttering the second type of statements they are applying a legal rule L, whose validity according to R is presupposed, to a particular case. The first type of legal statements hence typically takes the form ‘it is the law that … (a statement of L)’. The second type of legal statements, on the other hand, concerns not legal validity but the legal rights and duties of the parties to a particular case or the legal relations between those parties. Hart’s own examples of statements of the second type include, e.g., ‘A has a right to be paid £10 by B’ and ‘A & Company, Ltd. have a contract with B’.Footnote 6

Although jurisprudes have discussed both types of internal legal statements in their expressivist and quasi-expressivist reconstructions of Hart’s analysis, the inferentialist reconstruction I propose in this paper is based on Hart’s analysis of the second type of internal legal statements. It has this basis because throughout his career Hart focused almost exclusively on the second type of internal legal statements as the target of his philosophical analysis. The importance of this preliminary clarification will become clear when we come to the question whether Hart retained an inferentialist analysis of legal statements in The Concept of Law, where Hart seems to have offered a different analysis for each type of internal legal statements (henceforth I use the expression ‘legal statements’ to refer specifically to ‘internal legal statements’).

Expressivist and Quasi-expressivist Interpretations of Hart’s Analysis of Legal Statements

As a major proponent of an expressivist interpretation of Hart’s analysis of legal statements, Toh maintains that according to Hart, ‘in uttering an internal legal statement, a speaker expresses his acceptance of norms that make up the legal system’.Footnote 7 Importantly, this interpretation is presented as an analysis of the semantics (i.e., the meaning) of legal statements and not merely of their pragmatics (i.e., ‘the characteristic purposes for which those statements are articulated’).Footnote 8 The result is that under Toh’s interpretation, Hart’s analysis of legal statements is a non-cognitivist account whereby those statements lack truth value by virtue of being merely an expression of the utterer’s mental state.Footnote 9 The problem with this interpretation of Hart as an expressivist qua non-cognitivist is that it only fits Hart’s earliest published work in jurisprudence—his 1949 essay ‘The Ascription of Responsibility and Rights’Footnote 10—which Hart intentionally chose to exclude from his 1968 collection of essays.Footnote 11 As we will see, in as early as 1953, when Hart delivered his inaugural lecture (published as ‘Definition and Theory in Jurisprudence’ in Hart’s 1983 collection of essays), he had abandoned a non-cognitivist account of the meaning of legal statements in favour of a cognitivist yet non-descriptivist account.

Matthew Kramer has recently provided a detailed criticism of Toh’s interpretation of Hart as an expressivist in regard to the semantics of legal statements. Drawing a clear distinction between semantics and pragmatics, Kramer maintains both a negative thesis and a positive thesis in his criticism. The negative thesis rejects the characterization of Hart as an expressivist regarding the semantics of legal statements. According to Kramer, Hart is not an expressivist who analyses the meaning of a legal statement only in terms of ‘the function of the statement in giving voice to some non-cognitive attitude(s)’.Footnote 12 The positive thesis is that Hart can indeed be properly called an expressivist, but only regarding the pragmatics of legal statements.Footnote 13 Kramer thus reminds us that ‘we should be chary of assuming that [Hart’s] expressivism was focused on the semantics of legal pronouncements rather than solely on their pragmatics’.Footnote 14 In this paper I argue that although Kramer is right to reject the interpretation of Hart as an expressivist qua non-cognitivist, he is overly hasty to conclude that Hart’s expressivism was focused solely on the pragmatics of legal statements, for there remains an unexplored possibility of interpreting Hart as an inferentialist regarding the semantics of legal statements. Before exploring that possibility, let us first turn to the quasi-expressivist interpretation of Hart’s analysis.

The quasi-expressivist account, proposed by Stephen Finlay and David Plunkett,Footnote 15 takes Kramer’s reminder seriously. Duly observing the distinction between semantics and pragmatics, Finlay and Plunkett attribute to Hart an analysis that assigns an expressivist (and non-cognitivist) role to the pragmatics of legal statements and a descriptivist role to their semantics. According to this analysis, in regard to semantics, ‘a statement of the form ‘it is the law that L (in X)’ semantically expresses the proposition that L is a rule […] satisfying the criteria of the rule of recognition R of legal system X’.Footnote 16 Meanwhile, in regard to pragmatics, such a statement expresses the utterer’s non-cognitive attitudes or prescriptions, i.e., the utterer’s own attitude of ‘being motivated to act in accordance with L’Footnote 17 or the utterer’s prescription for the audience to accept R and derivatively to obey L.Footnote 18

The quasi-expressivist interpretation fits well with the view Hart somewhat cursorily expressed in the introduction to his 1983 collection of essays.Footnote 19 In that introduction, Hart indicates that he was right to reject non-cognitivism but wrong to espouse non-descriptivism in regard to legal statements, i.e., he was wrong to deny that legal statements were descriptive of certain facts.Footnote 20 The root of this mistake, according to Hart, is his conflation of the semantics and pragmatics of legal statements:

I fail to allow for the important distinction between the relatively constant meaning or sense of a sentence fixed by the conventions of language and the varying ‘force’ or way in which it is put forward by the writer or speaker on different occasions […] Neglect of this distinction […] vitiates part of my account […] of the meaning of statements of legal rights or statements about corporations.Footnote 21

Having recognised this conflation, Hart reverts to descriptivism regarding the semantics of legal statements while retaining expressivism regarding their pragmatics:

It was just wrong to say that such statements are conclusions of inferences from legal rules, for such sentences have the same meaning on different occasions of use whether or not the speaker or writer puts them forward as inferences which he has drawn. If he does put such a statement forward as an inference, that is the force of the utterance on that occasion, not part of the meaning of the sentence (emphasis in original).Footnote 22

As such, it might appear that Finlay and Plunkett have made a strong case for interpreting Hart as a quasi-expressivist. However, as I will argue in the following sections, Hart’s most elaborated analysis of legal statements—the analysis that he offered in his inaugural lecture and partially retained in The Concept of Law—is not quasi-expressivist as it rejects the clear distinction between semantics and pragmatics, a distinction that motivates the quasi-expressivist analysis. Instead, embodying a kind of methodological pragmatism, Hart’s analysis is an inferentialist one that is superior to both its expressivist and quasi-expressivist counterparts.

Hart as a Methodological Pragmatist

The Thesis of Methodological Pragmatism

There has been much talk about pragmatism in legal theory in recent decades, and pragmatism as a philosophical doctrine is not to be confused with pragmatism as a theory of adjudication.Footnote 23 The kind of methodological pragmatism I attribute to Hart in this paper is a specifically philosophical thesis about the meaning of language. It is the thesis that we cannot grasp the semantics (i.e., the meaning) of a single term or a statement without first grasping its pragmatics. The most famous articulation of this thesis is arguably due to the later Wittgenstein. As Brandom points out, ‘under the banner ‘Don’t look to the meaning, look to the use,’ Wittgenstein […] radicalizes the pragmatist critique of semantics’Footnote 24 and he achieves such radicalization by ‘[p]ointing out, to begin with, that one cannot assume that uses of singular terms have the job of picking out objects, nor that declarative sentences are in the business of stating facts’.Footnote 25 The crux of methodological pragmatism is hence a kind of prioritization of pragmatics over semantics, and this is captured by Brandom’s formulation of methodological pragmatism as ‘the view that the point of introducing a notion of semantic content or meaning (and hence the source of the criteria of adequacy of resulting theory) is to explain or at least codify central properties of their pragmatic use’ (emphasis in original).Footnote 26 As we will see, this is an apt characterization of the methodology of elucidation that Hart adopted in his inaugural lecture.

Replacing Definition with Elucidation

Although The Concept of Law is generally regarded as Hart’s greatest legacy, his analysis on the meaning of internal legal statements in that book is relatively brisk. To get a clearer understanding of Hart’s view, we need to examine the methodological lesson Hart was preaching in his inaugural lecture in 1953. In that lecture, Hart starts with noting the ‘great ambiguity’ in questions like ‘What is law?’, ‘What is a right?’, and ‘What is a state?’Footnote 27 Hart thinks these questions are ambiguous because they ‘may be used to demand a definition or the cause or the purpose or the justification or the origin of a legal or political institution’.Footnote 28 He proceeds to observe that if we try to disambiguate these questions by rephrasing them as explicit requests for the meaning of certain words, we seem to be trivializing these questions and we are still left puzzled even if we know how to use those words correctly.Footnote 29

Why are we still puzzled? Hart identifies two anomalies in relation to legal concepts that generate persistent puzzlement: first, we apply legal concepts to a wide range of cases, yet we also instinctively think that there must be some underlying principle that accounts for why we lump these diverse cases under the same concept, and we feel the urge to elucidate that underlying principle.Footnote 30 Secondly, such elucidation cannot be accomplished by the usual method of definition because legal concepts ‘do not have the straightforward connection with counterparts in the world of fact which most ordinary words have and to which we appeal in our definition of ordinary words’.Footnote 31 According to Hart, these anomalies motivate questions that take the form of quests for definitions, and such quests in turn give rise to ‘vast and irreconcilable theories’ that are detached from the actual workings of a legal system.Footnote 32

Dissatisfied with this ‘incubus of theory’,Footnote 33 Hart proposes a new methodology that replaces the quest for definition with a quest for elucidation. Drawing inspiration from Bentham’s advice that to explain legal concepts we must shift our attention from single concept-words (e.g., ‘a legal right’) to sentences where these concept-words appear,Footnote 34 Hart urges us to shift our focus from these concepts taken alone to statements applying those concepts, and then try to ‘[specify] the conditions under which such statements are true and the manner in which they are used’.Footnote 35 In the introduction to his 1983 collection of essays, Hart reaffirms and summarises the methodology of elucidation he advocated in his inaugural lecture as follows:

what was needed was a close examination of the way in which statements e.g. of legal rights or of the duties of a limited company relate to the world in conjunction with legal rules, and the important first step to take was to ask under what conditions statements of this kind have a truth value and are true.Footnote 36

However, specification of truth conditions for the relevant legal statement is only the first step in the elucidatory project, to be followed by ‘showing how [the statement] is used in drawing a conclusion from the rules in a particular case’.Footnote 37

Methodological Pragmatism in Hart’s Elucidatory Project

Having glimpsed into Hart’s general methodological outlook, we are now in a good position to explore the elements of methodological pragmatism in Hart’s preferred methodology of elucidation. First, Hart emphasises that what motivates traditional legal theorists to contrive vast and irreconcilable theories for legal concept-words is ‘the absence of some counterpart to ‘correspond’ to these words’,Footnote 38 and the common mistake of traditional legal theorists is to try to find some ‘unobvious counterpart’ while ignoring that ‘the primary function of [legal concepts] is not to stand for or describe anything but a distinct function’.Footnote 39 These remarks neatly align with Brandom’s observation that methodological pragmatism is partly captured by the Wittgensteinian caveat that ‘one cannot assume that uses of singular terms have the job of picking out objects’.Footnote 40 They provide a prima facie case for reading Hart as implicitly adopting a methodological pragmatist view according to which we can only understand the meaning of legal concept-words by attending to their pragmatic use, not by identifying their semantic referents.

Secondly, there is an interesting resemblance between Hart’s view that we should shift our focus from individual concept-words to sentences where the concept-words appear and the views of pivotal methodological pragmatists. For example, in rejecting the correspondence theory of truth, Richard Rorty has urged us to shift our attention from individual sentences to vocabularies or language games as a whole:

When the notion of ‘description of the world’ is moved from the level of criterion-governed sentences within language games to language games as a whole […] the idea that the world decides which descriptions are true can no longer be given a clear sense. It becomes hard to think that that vocabulary is somehow already out there in the world, waiting for us to discover it […] [F]or example, that the fact that Newton’s vocabulary lets us predict the world more easily than Aristotle’s does not mean that the world speaks Newtonian.Footnote 41

Although it might seem that Hart’s suggestion only shifts the unit of meaning from words to statements and hence falls one step short of Rorty’s view, the pivotal step in Hart’s method of elucidation is not merely to situate the legal concept-words in legal statements that apply those concept-words, but to situate those legal statements in the broader context where they play their functional role. This is indicated by Hart’s retrospective reflection that the methodology he advocated stems from

a conviction that longstanding philosophical perplexities could often be resolved not by the deployment of some general theory but by sensitive piecemeal discrimination and characterization of the different ways, some reflecting different forms of human life, in which human language is used (emphasis added).Footnote 42

The purpose of Hart’s method of elucidation is hence to explicate the implicit context within which legal statements fulfill their functional role, and it is the ability to fulfill that functional role that endows those legal statements with meanings.

A third dimension in which Hart’s elucidation exemplifies methodological pragmatism is a potential connection between Hart’s method of elucidation and inferential role semantics (hereinafter ‘IRS’), a connection that we will further explore in following sections. Notably, IRS has received some application in jurisprudence by legal theorists who see it as the proper theory to analyse the meaning of legal concepts. For example, Coleman maintains that ‘the inferential roles our concepts play reveal the holistic (or semi-holistic) web of relations in which they stand to one another, and it is this web that determines a concept’s content’.Footnote 43 In a similar vein, Benjamin Zipursky has also relied on IRS to develop what he calls ‘pragmatic conceptualism’, according to which ‘our concepts […] have content only because of the web of commitments in which they are enmeshed, and […] those commitments are themselves defined only in relation to a community’s set of practices’.Footnote 44 However, neither Coleman nor Zipursky has recognised that such an inferentialist account can be traced back to Hart’s inaugural lecture once we reconstruct it using the apparatus of IRS. Drawing on the works of Brandom—arguably the most prominent inferentialist today—I undertake this reconstruction after a brief overview of Brandom’s inferentialist framework.

Brandom’s Inferentialist Framework

Truth Conditional Semantics vs. Inferential Role Semantics

As a semantic theory that aims to explain how the sentences we utter come to have meaning, Brandom’s IRS should be understood against the backdrop of the traditional truth-conditional semantics, whereby the meaning of a sentence is explained in terms of the conditions under which it would be true.Footnote 45 Brandom’s IRS differs from the traditional approach in three crucial respects.

First, truth-conditional semantics takes the notions of truth and reference as explanatorily primitive such that a sentence’s meaning can be couched in its truth conditions, which are in turn explained in terms of the relation of reference between the single terms or predicates and the object or property to which they refer. The notion of truth can then be used to explain a good inference from one sentence to another as an inference that preserves truth.Footnote 46 In contrast, Brandom’s IRS takes the notion of inferential role as explanatorily primitive. According to Brandom, his account ‘starts with a practical distinction between good and bad inferences, understood as a distinction between appropriate and inappropriate doings, and goes on to understand talk about truth as talk about what is preserved by the good moves’ (emphasis in original).Footnote 47

Secondly, the strategy of taking inferential role as explanatorily primitive mandates a top-down approach that takes sentences—instead of single terms or predicates—as the starting point in explaining sentence meanings, in contrast to the bottom-up approach of truth-conditional semantics. This is not surprising given Brandom’s conviction that ‘to be propositionally contentful is to be able to play the basic inferential roles of both premise and conclusion in inferences’,Footnote 48 and the fact that only sentences, not subsentential expressions like single terms or predicates, can directly enter into inferential relations (Brandom’s account of how subsentential expressions can play indirect inferential roles will have to be omitted here due to limited space).

Thirdly, truth-conditional semantic theories are atomistic while Brandom’s IRS is holistic. The holism also follows naturally from the strategy of taking inferences as explanatorily primitive, for if the meaning of a sentence consists in the inferential role it plays within a network of inferences, one cannot understand the sentence without understanding what other sentences it follows from and what other sentences it entails. The result is that for IRS theorists, ‘no sentence plays its inferential role all on its own. For what a claim is evidence for or against and what is evidence for or against it depends on what collateral commitments are available to serve as auxiliary hypotheses in extracting inferential consequences’.Footnote 49

Commitments, Entitlements, and Incompatibility

Driven by the methodological pragmatist conviction that semantics must answer to pragmatics, Brandom’s IRS takes the strategy ‘to identify the particular structure [social practices] must exhibit in order to qualify as specifically linguistic practices’ (emphasis in original).Footnote 50 The particular structure Brandom identifies is the ‘inferential game of making claims and giving and asking for reasons’ (emphasis in original).Footnote 51 Under this structure, to have a certain concept (and hence to understand what a concept-word means) is to ‘have practical mastery over the inferences [the concept] is involved in—to know, in the practical sense of being able to distinguish, what follows from the applicability of a concept, and what it follows from’ (emphasis in original).Footnote 52

Commitments and entitlements are normative statuses interlocutors acquire in the game of giving and asking for reasons. Brandom uses the notion of commitment to capture what is more colloquially called a belief. A speaker can either acknowledge a commitment herself or attribute a commitment to others. When she acknowledges a commitment to a certain statement, she also becomes committed to other statements that follow from the original statement she acknowledges herself to be committed to. For example, it follows from the speaker’s commitment to ‘This sweater is red’ that she is also committed to ‘The color of this sweater has a wavelength longer than that of the color blue’. Consequential commitments like this can be attributed to the speaker without being acknowledged by her.Footnote 53

The normative status of entitlement is used by Brandom to capture what is more colloquially called justification for beliefs. According to Brandom, in acknowledging a commitment to a certain statement one also incurs a ‘liability to demands for justification—that is, demonstration of entitlement’.Footnote 54 For many (though not all) statements—Brandom uses examples such as ‘I have ten fingers’Footnote 55—when the speaker acknowledges a commitment to that statement, she is by default entitled to that commitment (i.e., others will see her as prima facie justified in making that statement).Footnote 56 In asserting a statement (and assuming that there is no successful challenge to that statement) the speaker acquires an entitlement to that statement, and she also acquires entitlements that follow from the original entitlement. Just like consequential commitments, consequential entitlements that are attributed to the speaker may or may not be acknowledged by her.Footnote 57

Incompatibility is a relation between commitments and entitlements. As Brandom indicates, ‘two claims are incompatible with each other if commitment to one precludes entitlement to the other’.Footnote 58 An obvious example is that a commitment to the statement ‘This sweater is red’ precludes the entitlement to the statement ‘This sweater is green’, making the two claims incompatible. If this incompatibility is acknowledged by the speaker or attributed to her, she acquires an incompatibility commitment, i.e., a commitment to the relation of incompatibility between these two claims.Footnote 59 A speaker may or may not acknowledge incompatibility commitments that are attributed to her, just as is the case with consequential commitments and consequential entitlements.

Deontic Scorekeeping: UIAs and DICs

It is important to introduce the key notions of commitments, entitlements, and incompatibility because these are the notions in which the pragmatic force of a sentence is couched, and the sentence’s semantic content is in turn couched in its pragmatic force. In a conversation, the interlocutors grasp the pragmatics of a sentence through a kind of deontic scorekeeping: once a sentence is uttered, it makes a difference ‘to what commitments and entitlements are attributed and undertaken by various interlocutors’,Footnote 60 thereby changing the deontic score of each interlocutor. An interlocutor’s grasp of the pragmatics of a sentence depends on her ability to keep track of such change to the deontic scores of herself and others. More specifically,

understanding or grasping the significance of a speech act requires being able to tell in terms of such scores when it would be appropriate (circumstances of application) and how it would transform the score characterizing the stage at which it is performed into the score obtaining at the next stage of the conversation of which it is a part (consequences of application).Footnote 61

This is why the game for which the interlocutors keep deontic score is an inferential game: the connections interlocutors draw between the original sentence to other sentences are inferential connections. By drawing such connections, the interlocutors are specifying the original sentence’s inferential role in terms of both the sentence’s upstream inferential antecedents (hereinafter ‘UIAs’) and downstream inferential consequences (hereinafter ‘DICs’).Footnote 62 According to Brandom, it is our ability to engage in this game that distinguishes us as linguistic creatures capable of producing meaningful expressions from non-linguistic creatures lacking this ability. Parrots can be trained to reliably utter the noise ‘That’s red’ whenever they are presented with red things, but they do not thereby possess the concept of red.Footnote 63 To know what the concept-word ‘red’ means is to have the capacity of keeping score in the inferential game of giving and asking for reasons, the capacity to tell ‘what else one would be committing oneself to by applying the concept, what would entitle one to do so, and what would preclude such entitlement’.Footnote 64

The Norm-Governedness of Meaning as Inferential Role

Let us finish our brief outline of Brandom’s IRS by highlighting a key feature that will be pertinent to our interpretation of Hart as an inferentialist. According to Brandom, the meaning of a sentence is conferred by the sentence’s inferential role in the network of material inferences, which are generally non-monotonic and hence can be defeated by adding an additional collateral premise.Footnote 65 People often have different views on which inferences are materially good because they are committed to different collateral premises. However, under Brandom’s IRS meaning is not relative to individual interlocutors. The material inferences that confer propositional contents on expressions are all normative because those inferences ‘implicitly contain norms concerning how it is correct to use expressions, under what circumstances it is appropriate to perform various speech acts, and what the appropriate consequences of such performances are’ (emphases in original).Footnote 66 For example, when a speaker asserts and hence acknowledges her commitment to ‘This sweater is red’ yet fails to acknowledge a commitment to ‘The color of this sweater has a wavelength longer than that of the color blue’, we can say that her failure is a mistake because it is a consequential commitment that she ought to have acknowledged. Deontic scorekeeping is therefore norm-governed in the sense that it can be done correctly or incorrectly: what constitutes the differences in score, which confer the propositional content of an assertion, is objective to the extent that any single interlocuter can be mistaken about it. As we will see later, this norm-governedness distinguishes inferentialism from expressivism and provides the inferentialist with the same advantages that quasi-expressivism has over expressivism.

Hart’s Inferentialist Analysis of Legal Statements

Having introduced Brandom’s inferentialist apparatus, let us proceed to explore the viability of reading Hart as a forerunner of applying an inferentialist account of meaning to analyzing legal statements. Compared with a truth-conditional approach, the inferentialist approach aligns better with Hart’s observations that legal concepts ‘do not have the straightforward connection with counterparts in the world of fact which most ordinary words have’Footnote 67 and that ‘the primary function of [legal concepts] is not to stand for or describe anything but a distinct function’.Footnote 68 This is because as we have seen above, truth-conditional semantics is fundamentally representationalist for it explains the meaning of subsentential expressions by virtue of their referents. In contrast, Hart’s central insight is anti-representationalist in pointing out that we cannot understand the meaning of legal statements without considering the form of life in which those statements play their functional role.

This anti-representationalist insight opens up the space to reconstruct Hart’s inaugural lecture as an inferentialist analysis of legal statements, for as we will see, according to Hart, the functional role played by legal statements is precisely to express the conclusion of an inference made according to legal rules. Thus interpreted, Hart’s pragmatist account of the meanings of legal statements would come close to endorsing—in regard to legal discourse—the kind of IRS deployed by Brandom. The congeniality of Brandom’s IRS to Hart’s method of elucidation is reinforced by the fact that in his later career Hart uses the expression ‘inference’ as an equivalent to what he originally called a ‘conclusion of law’.Footnote 69

Hart’s Truth Conditions Reinterpreted

To carry out an inferentialist reconstruction of Hart’s inaugural lecture, we first need to address the following questions: Hart apparently subscribed to truth-conditional semantics in trying to specify the truth conditions for legal statements. This is illustrated by Hart’s contention that to carry out the first step of his two-fold elucidatory project, we must specify ‘conditions under which the statement […] is true’ or ‘conditions necessary for the truth of a sentence’ (emphases added).Footnote 70 How could Hart subscribe to truth conditional semantics and be an inferentialist at the same time? The answer to this question, as I will argue, is that what Hart takes as truth conditions for legal statements can be interpreted as a combination of what Brandom calls UIAs and DICs of those statements. The argument will have to be fleshed out in a few steps.

To begin with, we should note that in his inaugural lecture Hart carefully distanced himself from the thought that the meanings of legal statements can be solely explained in terms of their truth conditions without specifying ‘the manner in which [such statements] are used’.Footnote 71 In the following passage Hart explicitly points out that the specification of truth conditions is not sufficient for us to understand the meanings of legal statements:

If we take a very simple legal statement like ‘Smith has made a contract with Y’, we must distinguish the meaning of this conclusion of law from two things: from (1) a statement of the facts required for its truth […] and also from (2) the statement of the legal consequences of it being true […] [I]t seems as if there is something intermediate between the facts, which make the conclusion of law true, and the legal consequences […] [A legal statement] is an utterance the function of which is to draw a conclusion from a specific rule under which, in circumstances such as these, consequences of this sort arise, and we should obviously neglect something vital in its meaning if […] we said it meant the facts alone or the consequences alone or even the combination of these two.Footnote 72

This passage demonstrates that for Hart, the specification of truth conditions for legal statements is only part of the elucidatory project, and those truth conditions are only relevant in so far as they indicate the circumstances under which legal rules can be appropriately applied. The meanings of legal statements consist not in those circumstances, but in the statement’s function to express what follows from the application of those legal rules under those circumstances. It is hence the second step of ‘showing how [the statement] is used in drawing a conclusion from the rules in a particular case’,Footnote 73 rather than the first step of specifying the statement’s truth conditions, that plays the pivotal role in Hart’s elucidatory project.

One might point out that even if Hart regarded truth conditions as insufficient for conferring meanings on legal statements, he apparently thought they were a necessary component to their meanings, alongside the legal statements’ inferential roles. As such, Hart can only be interpreted as an advocate of what Brandom calls weak inferentialism—whereby the articulation of a sentence’s inferential role is necessary but not sufficient for constituting the sentence’s meaning—instead of Brandom’s strong inferentialism, whereby the articulation of inferential role is not only necessary but also sufficient for meaning constitution.Footnote 74 This might be the most obvious interpretation of Hart’s approach, but it is not the only interpretation available. For one thing, Hart conceives of truth conditions as ‘facts required for [a legal statement’s] truth’.Footnote 75 This seems to presuppose a correspondence theory of truth, a theory that is undermined by well-known problems relating to both the notion of facts and the purported relation of correspondence.Footnote 76 These problems might have convinced Hart to adopt a deflationary theory of truth, which Kramer thinks ‘Hart would very likely have been attracted to’,Footnote 77 and in that case Hart would have found the appeal to truth conditions as a necessary component of a theory of meaning both unilluminating and unviable.

The appeal to truth conditions would be unilluminating because from a deflationist perspective, the truth conditions for legal statements are specified using Tarskian T-sentences, an example of which would be ‘‘James has made a contract with Lily’ is true iff James has made a contract with Lily’. This truth condition by itself does not tell us anything illuminating about what it is for someone to make a contract. A Davidsonian truth-conditional semanticist would use truth conditions like this to provide a translation between the object language and the meta-language (though its being a translation is not obvious when the object language and the meta-language are the same). This translational project only works on the presumption that subsentential expressions have referents that we can find out through empirical observations conducted in a process like Davidsonian radical interpretation. However, the central insight Hart tried to convey in his inaugural lecture is precisely that many legal concepts do not have a referent and the assumption that they do causes philosophical bewilderment. Moreover, combining a deflationary theory of truth and a truth-conditional theory of meaning would arguably be unviable. As Brandom observes, ‘anyone who holds to a deflationary theory of truth is precluded from explaining propositional contents in terms of truth conditions’.Footnote 78 The incompatibility between a deflationary theory of truth and a truth-conditional theory of meaning has also been noted by philosophers like Paul Horwich, Hartry Field, and Huw Price.Footnote 79 Admittedly, whether Hart would endorse deflationism about truth remains ultimately speculative, and whether deflationism about truth is fundamentally incompatible with truth-conditional semantics remains a controversial issue.Footnote 80 Therefore, I am not arguing that Hart’s elucidatory project must be interpreted as a kind of strong inferentialism, but only that it is congenial to such an interpretation. It is hence possible for us to keep Hart’s non-descriptivist analysis intact by understanding Hart’s truth conditions as a legal statement’s UIAs, i.e., statements ‘that can serve as premises from which entitlement to the original [statement] can be inherited’.Footnote 81 This is in effect replacing Hart’s truth conditions with assertibility conditions, for UIAs are just the set of sufficient conditions under which one is justified in asserting the statement.Footnote 82 They are the reasons that the speaker has to give in order to vindicate her entitlement to the original statement when confronted with a non-gratuitous challenge in the game of giving and asking for reasons. Thus interpreted, Hart’s approach will embody a kind of strong inferentialism whereby a legal statement’s inferential role is not only necessary but also sufficient for conferring its meaning.

Why Hart was Not a Verificationist

To say that Hart’s truth conditions can be interpreted as assertibility conditions seems to imply that Hart can be interpreted as a verificationist in regard to the meaning of legal statements. After all, the crucial difference between a verificationist account of meaning and Brandom’s inferentialist account is that the former focuses exclusively on a statement’s assertibility conditions (i.e., UIAs), while the latter also incorporates the statement’s DICs.Footnote 83 Did Hart ignore legal statements’ DICs by characterizing legal statements as mere ‘conclusions of law and the tail ends of legal calculations’?Footnote 84 The answer is negative, for the truth conditions identified by Hart comprised both UIAs and DICs.

Hart’s observation of DICs is perhaps most conspicuous in the truth conditions he specifies for the statement ‘Nusquamia owes you £1,000’. The second of those conditions is that pursuant to the rules of the fictional state Nusquamia’s legal system, certain consequences follow from someone’s being entitled to receive a sum of money from the fictional state.Footnote 85 Hart also attends to DICs in his more general remark that we can capture the meaning of a legal statement only by ‘reproduc[ing] the distinctive manner in which the […] statement is used to draw a conclusion from a specific but unstated rule under which such a consequence follows on such conditions’ (emphasis added).Footnote 86 The fact that certain consequences are prescribed by a legal rule to follow from certain circumstances was treated by Hart as a truth condition (now reinterpreted as an assertibility condition) that, along with other assertibility conditions (e.g., the existence of a legal system), justifies one’s assertion of the legal statement. This treatment is problematic as it seems to be putting the cart before the horse. We can see, e.g., that the consequence of A being liable to pay damages to B when A has breached its contract with B does not justify the statement ‘A has a contract with B’. It is the statement that justifies the consequences rather than the other way around. Hart’s formulation hence conflates both UIAs and DICs of a legal statement by lumping them together as truth conditions. But despite this infelicity, Hart’s formulation reveals that he is not a verificationist and that his account for the meanings of legal statements is perfectly intelligible as an inferentialist account.

Something Intermediate Between the UIAs and DICs?

So far, we have seen that once we reinterpret Hart’s truth conditions as a combination of Brandom’s UIAs and DICs, Hart’s analysis of legal statements aligns neatly with Brandom’s IRS. But an attentive reader might point out that this cannot be right. Hart maintained in his inaugural lecture that ‘we should obviously neglect something vital in [the meaning of a legal statement] if […] we said it meant the facts alone or the consequences alone or even the combination of these two’ (emphasis added).Footnote 87 If Hart were an inferentialist, wouldn’t he have thought that the meanings of legal statements consist precisely in the constellation of those circumstances (UIAs) and consequences (DICs)?

To answer this question, we need to see what Hart was driving at in his contention that ‘it seems as if there is something intermediate between the facts, which make the conclusion of law true, and the legal consequences’.Footnote 88 If we interpret facts as a legal statement’s UIAs and consequences as its DICs, then what is that something intermediate between them? Hart did not provide a clear answer, but he kept emphasizing the same point, i.e., the function of the legal statement is to ‘draw a conclusion from a specific rule under which, in circumstances such as these, consequences of this sort arise’.Footnote 89 We can infer from this remark (and Hart’s similar remarks in the inaugural lecture) that what mediates between the UIAs and DICs is nothing but the legal statement itself, and what enables the legal statement to play this mediating role is just its function to inferentially connect the UIAs to the corresponding DICs. In other words, the mediating role is just the legal statement’s inferential role, the role it plays in virtue of its position in an inferential structure as something that follows from the UIAs and entails the DICs.

The legal statement indicates certain background assumptions that from the speaker’s perspective warrant an inference from the UIAs to the DICs. Without these background assumptions, there will not be a materially good inference from the UIAs to the DICs. For example, the inference from ‘the defendant has intentionally killed another human being’ to ‘the defendant has committed a heinous crime’ will not be materially good without the statement ‘the defendant has committed murder’ mediating between them. This is because it is only under certain assumptions—implicitly encapsulated in the statement ‘the defendant has committed murder’—that one who has committed intentional killing has committed a heinous crime, or indeed a crime at all. Such assumptions would include, e.g., that the killer was not conducting proportionate self-defence, that the killer is not an executioner in a jurisdiction where death penalty is legal, and that the killer is not a doctor who has administered euthanasia to a patient in a jurisdiction where euthanasia has been legalised.

The fact that a legal statement plays this mediating role in such an inferential structure would seem like a platitude for an inferentialist because this structure is built-in at the very foundation of IRS: the UIAs are UIAs to the statement and the DICs are DICs of the statement. The meaning of the statement cannot be explained as a combination of UIAs and DICs if we just treat the UIAs and DICs as certain circumstances and consequences without putting them into the inferentialist structure. The importance of this inferentialist structure—whereby the circumstances and the legal consequences are inferentially connected with each other through the legal statement mediating between them—is what Hart was trying to reveal by pointing out that there is something intermediate between the facts and the consequences. However, without the apparatus of IRS available to him, he had to express it in a somewhat awkward and confusing way, e.g., by characterizing a legal statement as ‘the tail-end of a simple legal calculation’ or as ‘a conclusion of law’.Footnote 90 Once we interpret Hart’s elucidatory project through an inferentialist lens, we should see that Hart’s observation adds nothing extra to the inferentialist framework because it is already part and parcel of that framework itself. Hence, Hart’s remark about there being something intermediate between the circumstances and the consequences does not undermine but rather strengthen the interpretation of Hart as an inferentialist.

Did Hart Retain an Inferentialist Analysis in The Concept of Law?

At this point, it should be clear that Hart had jettisoned a non-cognitivist analysis of legal statements in his 1953 inaugural lecture, as demonstrated by his method of elucidation whose first step is to specify the conditions under which legal statements are true. The endeavour to specify truth conditions of legal statements presupposes the truth-aptness of those statements and hence the falsity of non-cognitivism. In place of the non-cognitivist and ascriptivist account he endorsed in 1949, Hart offered a cognitivist, non-descriptivist, and arguably inferentialist analysis of legal statements in his inaugural lecture. Did Hart retain this analysis in his landmark work The Concept of Law? The answer is complicated.

Finlay and Plunkett interpret Hart’s analysis in The Concept of Law as a quasi-expressivist analysis.Footnote 91 To substantiate this interpretation, they cite a passage in The Concept of Law where Hart says ‘[w]e can indeed simply say that a statement that a particular rule is valid means that it satisfies all the criteria provided by the rule of recognition’.Footnote 92 Finlay and Plunkett see this as a clear indication that in The Concept of Law Hart had adopted a descriptivist view in regard to the semantics of legal statements, statements describing the fact that a certain rule satisfies the rule of recognition of a legal system. However, as clarified in the introduction of this paper, statements of legal validity are only one type of legal statements and—as demonstrated by the foregoing exposition—not the type that Hart focused on when analyzing the semantics of legal statements. Even if Finlay and Plunkett are right about Hart’s descriptivist analysis of statements of legal validity, there is countervailing evidence that Hart retained an inferentialist analysis of the second type of legal statements in The Concept of Law.

In that seminal work, Hart argues that the gunman situation fails to illuminate the meaning of the statement of the form ‘one has an obligation to φ’, whose meaning can only be explained when we put the statement in the following social situation: first, there exists a social rule ‘making certain types of behaviour a standard’Footnote 93 and secondly, ‘the distinctive function of such statement is to apply such a general rule to a particular person by calling attention to the fact that his case falls under it’ (emphasis added).Footnote 94 This analysis demonstrates that Hart was still committed to the methodological pragmatist view that we can only understand the semantics of a statement by first grasping its pragmatics, i.e., by situating the statement within the discourse where it plays its functional role.

Admittedly, compared with what Hart said in his inaugural lecture, the inferentialist tenor of that functional role is no longer apparent, for Hart now avoids characterizing the function of the statement as drawing an inference (what he originally called a ‘conclusion of law’),Footnote 95 but instead states it more abstractly as applying a general rule to a particular case. Nonetheless, the pragmatist outlook and general structure of Hart’s analysis remain the same and there is a strong case for believing that in The Concept of Law Hart had not jettisoned the non-descriptivist, inferentialist analysis regarding the second type of legal statements. Hart’s rather ambiguous phrase ‘calling attention to the fact’ is especially interesting, for Hart did not use the phrase ‘describing the fact’, which would have made his descriptivist stance explicit had he adopted one. Instead, by using this more ambiguous phrase, Hart makes it possible to treat the fact that one’s case falls under a rule as an assertibility condition—i.e., a UIA—of the statement ‘one has an obligation to φ’. Therefore, compared with a straightforward quasi-expressivist interpretation of Hart’s analysis in The Concept of Law, a more balanced view is that Hart partially retained the methodological pragmatist and inferentialist analysis he adopted in the inaugural lecture.

Hart’s Regrettable Retraction

In the introduction to his 1983 collection of essays, Hart makes it abundantly clear that he is retracting the non-descriptivist analysis of legal statements in his inaugural lecture and reverting to what Finlay and Plunkett call the quasi-expressivist view, which is descriptivist in regard to the semantics of legal statements and expressivist in regard to their pragmatics.Footnote 96 This retraction of what could have become a great insight for analytic jurisprudence is regrettable. In hastily conceding that legal statements have the same meaning regardless of their occasions of use, Hart slides back into a descriptivist picture, which tries to ‘explain what a legal sentence means […] independently from the peculiar speech context within a linguistic practice and the propositional attitudes of the participants’.Footnote 97 This is a picture that Hart’s inaugural lecture could help us leave behind as far as legal statements are concerned. It is also a picture that curiously neglects some of the most important messages of both J.L. Austin and the later Wittgenstein, given that Hart regarded them as the first and second most important figures who had influenced his own philosophical development.Footnote 98 With the benefit of hindsight, we can see that the inferentialist analysis Hart implicitly endorsed in his inaugural lecture enjoys the same advantages possessed by the quasi-expressivist analysis over its expressivist counterpart. We therefore have strong reasons to believe that had Hart been aware of these advantages, he would not have so easily disowned the methodological pragmatism and non-descriptivism he originally espoused.

Explaining Internal-External Disagreements

In their campaign for quasi-expressivism, Finlay and Plunkett point out that as an analysis of legal statements, quasi-expressivism enjoys two advantages over expressivism: first, quasi-expressivism is able to account for ‘disagreements about law between one speaker making an internal legal statement, and another speaker making an external legal statement concerning the same legal system’.Footnote 99 According to Finlay and Plunkett, expressivists will have a difficult time explaining such disagreements for the following reasons: compared with internal legal statements, external legal statements are much more resistant to an expressivist treatment, i.e., it is much harder to view them as expressions of the speakers’ attitudes or mental states.Footnote 100 If A is making a description by uttering an external legal statement and B is expressing her mental state by uttering an internal legal statement, it is unclear how there can be a genuine disagreement between them, but intuitively we think that there can be such disagreements, so expressivists flounder.Footnote 101

Quasi-expressivists have no problem explaining such disagreements, because for them both internal and external legal statements are ‘descriptive claims about the same subject: whether a particular first-order rule satisfies the criteria of a particular rule of recognition’.Footnote 102 This is certainly correct, but we need not espouse descriptivism in order to account for such disagreements, which are perfectly intelligible within an inferentialist framework. Recall that under Brandom’s IRS, the meaning of a sentence—as encapsulated in its inferential role—is objective to the extent that any single interlocuter can be mistaken about it. This is because the correct meaning of a legal statement is governed by a norm that any individual interlocutor can fail to grasp. The objectivity of inferential role makes it possible for an inferentialist to explain the disagreement between A and B as a genuine disagreement about what norm is or was actually governing the correct use of the statement. Just like quasi-expressivists, inferentialists can offer this explanation under a unified semantic theory without the need to devise ‘a separate account of the semantics of external statements of law’.Footnote 103

Avoiding the Frege-Geach Problem

The second virtue of quasi-expressivism highlighted by Finlay and Plunkett is that while expressivists have a hard time dealing with the Frege-Geach problem, quasi-expressivists have the advantage of not falling prey to this problem in the first place. We know that Hart retracted the non-descriptivist analysis of legal statements and adopted a quasi-expressivist stance because he was persuaded by Peter Geach’s objection to expressivism.Footnote 104 However, by endorsing inferentialism, one can reject expressivism without reverting to descriptivism. Given Hart’s generally pragmatist and anti-metaphysical outlook, this is presumably what he would have done had he realised that just like quasi-expressivists, inferentialists can also avoid Geach’s challenge.

Geach’s objection shows that an expressivist account that explains the semantic content of statements in terms of their pragmatic use will not work in force-stripping contexts: an expressivist will struggle to explain, e.g., how the free-standing statement ‘torture is wrong’ can retain its meaning (explained as an expression of the speaker’s attitude against torture) when embedded in the conditional ‘if torture is wrong, then soldiers should not torture their prisoners’ so as to avoid the fallacy of equivocation. The quasi-expressivist analysis of legal statements avoids this problem by ‘[identifying] the semantic content of a simple declarative sentence about the law as an ordinary descriptive proposition with an ordinary kind of truth value’.Footnote 105 Inferentialists, however, can avoid the same problem without committing to descriptivism. The key to finding a non-expressivist and non-descriptivist way around the Frege-Geach problem, as succinctly put by Matthew Chrisman, is to ‘appeal to an account of how rule-governed linguistic role constitutes meaning, which comports with the compositional systematicity of meaning’.Footnote 106 Brandom’s IRS offers exactly such an account.

As has been noted above, the meaning of a legal statement consists in its inferential role within the network of inferences that connect the appropriate UIAs and DICs. That inferential role is governed by a normative rule, and in case of legal statements, by a legal rule. The legal rule prescribes an objectively correct answer to what the legal statement means because it specifies the inferences one ought to acknowledge as materially good given one’s commitment to the legal statement. Because the legal statement is governed by the same legal rule even when it is uttered in different contexts, it will have the same inferential role and therefore a constant meaning whether it is free-standing or embedded. Because the meaning of the legal statement consists in its objective inferential role as prescribed under a legal rule instead of the utterer’s subjective mental state of approval or disapproval, the threat of the Frege-Geach problem evaporates and we can retain an analysis of legal statements that avoids the pitfalls of expressivism yet retain the insights of non-descriptivism.Footnote 107

Accounting for the Objectivity of Legal Statements

At this juncture, one might point out that even if inferentialism offers the same advantages over expressivism as quasi-expressivism does, it suffers from the defect of not being able to account for the objectivity of legal statements. According to Finlay and Plunkett, a distinctive advantage of the quasi-expressivist analysis is that it ‘accommodates the appearance that legal statements aim to describe objective, attitude-independent facts about law’.Footnote 108 As we have seen, inferentialism is also able to accommodate the objectivity of legal statements at least to the extent that any individual interlocutor can be mistaken about the statement’s inferential role and hence about its meaning. It is a more difficult question, however, as to whether inferentialism can vindicate a stronger notion of objectivity, which not only accommodates individual mistakes, but also accounts for the possibility of community-wide mistakes, i.e., the possibility for legal statements to have what Kramer calls ‘strong observational mind-independence’Footnote 109 or what Coleman calls ‘modest objectivity’.Footnote 110 The topic of legal objectivity has attracted ample attention in jurisprudential literature and is too complex to be tackled here. A few brief remarks will suffice for the purpose of this paper.

Whether legal statements indeed have strong observational mind-independence is a controversial issue.Footnote 111 Assuming that legal statements do have strong observational mind-independence, it is possible for an inferentialist to account for this feature of legal discourse without treating legal statements as descriptions of attitude-independent legal facts. Indeed, Brandom himself has proffered such an account by appealing to a Hegelian model of legal concept determination, which aims to explain how officials’ past practices of legal concept application can institute rules that normatively govern how they ought to apply those concepts in the future.Footnote 112 A more radical pragmatist might draw inspiration from Price’s account of our concept of objective truth as convenient friction,Footnote 113 an account that would allow us to vindicate the objectivity of legal statements by explaining the practical function served by our concept of legal objectivity. Such an account would likely involve hypotheses about what behavioral changes the belief in modest objectivity of legal discourse would cause and what benefits those behavioral changes would bring to the functioning of a legal system. A good example is provided by Coleman and Leiter in the following passage:

Modest objectivity is a normative conception of objectivity in the sense that it provides a criterion for assessing whether adjudication is legitimate or justifiable [...] Modest objectivity provides criteria by which to criticize actual adjudicatory practice as falling short of objectivity, e.g., for its lack of impartiality, complete information, imaginative empathy, logic, etc (emphasis in original).Footnote 114

Unsurprisingly, proponents of the descriptivist account of legal statements will have an easier time accounting for legal objectivity: legal statements are objective because they describe attitude-independent legal facts. This ease, however, comes at a cost.

Not all legal statements purporting to describe legal facts will succeed in doing so. Moreover, whether a legal statement accurately describes the legal fact it purports to describe is a question that cannot be answered within the legal discourse itself. This is what makes the legal facts attitude-independent and thereby endows the legal statements with strong observational mind-independence. However, by separating the domain of legal discourse (where legal statements are uttered) and the domain of legal facts (where the real meanings of those statements belong), one postulates for legal statements a test of truth that is—to borrow Hilary Putnam’s expression—‘radically non-epistemic’.Footnote 115 Proponents of the descriptivist account will then face the same formidable challenges encountered by proponents of the correspondence theory of truth. By rejecting the representationalist view towards legal statements—a view that the descriptivist analysis implicitly holds on to—inferentialists can vindicate the objectivity of legal statements without the burden of defending the correspondence theory of truth, which many jurisprudes have found implausible when applied to the domain of legal discourse.Footnote 116

Conclusion

Drawing on Brandom’s framework of IRS, I have tried to reconstruct Hart’s analysis of legal statements as an inferentialist analysis. Based on a close reading of Hart’s inaugural lecture, this reconstruction is made possible by reinterpreting what Hart viewed as truth-conditions of a legal statement as a combination of the statement’s UIAs and DICs. I argue that in comparison with the expressivist and quasi-expressivist reconstructions, the inferentialist reconstruction more accurately captures Hart’s cognitivist yet non-descriptivist view towards the semantics of legal statements, a view that was partially retained by him in The Concept of Law. Although Hart eventually retracted the inferentialist analysis and reverted to descriptivism (thereby becoming a quasi-expressivist) in his later career, it is a regrettable retraction that does not fit well with his generally pragmatist and anti-metaphysical outlook. We might as well wonder whether Hart would have made the same decision had he realised that inferentialism not only offers the same advantages that quasi-expressivism has over expressivism, but also has the potential to vindicate the objectivity of legal statements while remaining both non-descriptivist and anti-representationalist.