KEVIN TOH’S EXPRESSIVIST READING OF H. L. HART, OR HOW NOT TO RESPOND TO

recognition Abstract: This paper criticises Kevin Toh’s expressivist reconstruction of H. L. A. Hart’s semantics of legal statements on  A first draft of this paper was presented at the Colloquium Expressivism in Contemporary Legal Philosophy, as part of the IV Congresso Internacional de Direito Constitucional e Filosofia Política, held at the State University of São Paulo, in November 2019. I thank the organizers, Thomas Bustamante and Thiago Decat, for the invitation, and the participants, in particular Kevin Toh, for valuable feedback. This work was supported by the National Council for Scientific and Technological Development


INTRODUCTION
In Hart's "Expressivism and His Benthamite Project", Kevin Toh attributes an expressivist and noncognitivist semantics of internal legal statements to H. L. A. Hart 1 , according to which speakers express their acceptance of legal norms in uttering internal legal statements, i.e. statements of law ascertainment or law application. This expressivist semantics depends on there being a noncontingent connection between normative utterances and speakers' reasons for action (Toh 2005, p. 79): acceptance of a norm as a reason to act is constitutive of the meaning of normative statements.
In this paper, I argue against such a reconstruction of Hart's semantics of legal statements on the grounds that (1) it blurs the distinction between legal statements and statements of value, and (2) it allows for municipal law without a rule of recognition. Section 1 makes the case for (1), and Section 2 argues for (2). Section 3 appeals to the distinction between detached and committed statements to argue that the expressivist component of Hart's analysis of legal statements is to be understood as part of a pragmatic account -rather than a semantic account -of committed statements. Section 4 presents an alternative reading of Hart's semantics of legal statements.

REPLYING TO DWORKIN
According to Toh (2005, p. 88), Hart's analysis of the meaning of internal legal statements has two prongs. First, speakers express their acceptance of the norm that they consider to be the rule of recognition -i.e. the norm that identifies legal norms -of their legal system. Second, speakers presuppose that such a norm is generally accepted and complied with in their community. Here, Toh has in mind what he calls "full-blooded acceptances [...] involved in making the paradigmatic internal legal statements that committed internal legal statements are" (Toh 2005, p. 90).
What follows from this semantic account of internal legal statements is an important correction to Hart's analysis of the rule of recognition. In Chapter 6 of The Concept of Law (1961, p. 109), Hart argues that there cannot be an internal legal statement of the content of the rule of recognition because this would be a statement to the effect that the rule of recognition is legally valid, while legal validity is always validity in conformity with the criteria that constitute the rule of recognition. Toh cannot accommodate the preclusion of internal legal statements that explicitly state the content of the rule of recognition in his reading of Hart's legal semantics since he argues that, for Hart, an internal legal statement actually expresses speakers' acceptance of the rule of recognition of their community's legal system. This amounts to a major rupture with Hart's views, as becomes clear as Toh's analysis progresses.
Toh claims that his reading of Hart's semantics of internal legal statements is capable of providing a compelling reply to Ronald Dworkin (1972, pp. 46-80;1986, ch. 1), who criticised Hart's legal philosophy for allegedly being unable to account for genuine legal disagreement, what Dworkin calls "theoretical disagreements". Since expressivism claims that in uttering internal legal statements speakers are expressing their acceptance of a legal norm while attempting to influence others' legal opinions, Toh (2005, p. 113) considers this analysis to be "specifically designed to explain normative disagreements".
Furthermore, Toh (2005, p. 113) is keen to dismiss the second prong of his reconstruction of Hart's analysis of internal legal statements. According to Toh, a speaker can have a disposition to accept a legal norm even if he or she does not believe in the pragmatic presupposition that makes up the second prong of this analysis of internal legal statements. This being so, it follows from Toh's twopronged reconstruction that legal officials can "disagree about the content of the rule of recognition of their legal system" (Toh 2005, p. 114). This is why expressivism is supposed to provide a compelling Hartian reply to Dworkin. It is worth quoting the passage in which Toh depicts a legal debate between two speakers on the rule of recognition in full: [A] speaker who believes that R1 is the rule of recognition of his community, and another speaker who believes that R2 is the rule of recognition of the same community, can have a genuine legal disagreement by uttering roughly the following two statements, respectively: 'Let us act according to a norm that is a part of a system of norms with R1 on top and other secondary norms in the middle tiers!', and 'Let us act according to a norm that is a part of a system of norms with R2 on top and other secondary norms in the middle tiers!' Clearly, such two speakers share a normative meaning despite their disagreement about the content of their community's rule of recognition. (Toh 2005, p. 114) Such a conversation is perfectly plausible, and it depicts meaningful normative disagreement. Nevertheless, Hart would not identify a discussion like this as a properly legal one. In order to understand why, let us consider the kinds of arguments that could be used by the speakers in the quotation.
Each speaker urges the other to act in accordance with a norm -perhaps even the same norm. This is not impossible, since different rules of recognition can identify the same norm as legally valid. Crucially, however, what is at issue here is not the legal validity of the norm defended by each speaker according to the criteria provided by the corresponding rule of recognition. The divergence concerns the rule to be used to identify norms: R1 or R2. In other words, the point at issue is these rival sets of criteria. Thus, the core of my first objection against Toh is that such a debate is a normative discussion of the value or desirability of a certain legal system rather than a normative discussion that is internal to a legal system.
As Hart (1961, p. 109) observes, the value or desirability of a rule of recognition is of course open to debate. Hart (1961, p. 107) even provides examples of these questions: "Does it [a rule of recognition] produce more good than evil? Are there prudential reasons for supporting it? Is there a moral obligation to do so?" Nowhere does Hart argue that these questions are meaningless. He is merely claiming that there is a relevant difference between these questions and questions about the validity of a norm in accordance with intra-systemic criteria, a difference on which we can draw the distinction between legal statements and statements of value: These [the questions quoted above] are plainly very important questions; but, equally plainly, when we ask them about the rule of recognition, we are no longer attempting to answer the same kind of question about it as those which we answered about other rules with its aid. (Hart 1961, p. 107) And Hart goes on: when we move from the statement that a particular enactment is valid, to the statement that the rule of recognition of the system is an excellent one and the system based on it is one worthy of support, we have moved from a statement of legal validity to a statement of value. (Hart 1961, p. 108) This being so, by introducing the acceptance of a rule of recognition into the meaning of legal statements, the analysis of internal legal statements that Toh attributes to Hart blurs the distinction between law and morality 2 (or values in general) 3 , as well as the distinction between jurisprudence and moral philosophy. Therefore, while a strategy of this sort would clearly satisfy Dworkin, 4 it could never satisfy Hart. In the next section, I argue that the loss of the distinction between legal statements and statements of value is not the only baffling consequence of the attribution of an expressivist semantics to Hart's legal theory.

LAW WITHOUT SECONDARY RULES?
The quotation in the last section in which Toh (2005, p. 114) depicts a purported legal debate about the rule of recognition clearly entails the dismissal of the rule of recognition itself. After all, there is no rule of recognition in Hart's sense if there is no jointly accepted rule of recognition. Toh endorses this implication 5 , but he does not consider it a reason to give up on his expressivist reading of Hart. As Toh writes: [H]is discussion of international law in chapter 10 of The Concept of Law clearly shows that Hart is willing to attribute, if not a legal system, then at least what could be called a 'legal regime' to a community without settled secondary norms. (Toh 2005, p. 115) This is a clever move, even though the sequence of this assertion is hard to follow. We are told that, if it is to be distinguished from a legal system, a legal regime obtains in a community "insofar as the members of that community regulate their conduct according to a set of norms the applicability of which they treat as amenable to authoritative determination by an appeal to an ultimate norm of legal validity" (Toh 2005, p. 116). Does this mean that Toh believes that Hart would attribute a legal regime -to be distinguished from a debate on values -to a community the members of which act and assess others' actions according to different ground norms? Again, according to Hart's philosophy, would such a community have positive law or a normative debate about the best law to have?
The next paragraph in Toh's paper does not make things any easier to understand. We are told that, "in the absence of a jointly accepted rule of recognition", Hart could still attribute a legal system to a community provided that "its members aim at the discovery and maintenance of such a jointly accepted ultimate norm of legal validity" (Toh 2005, p. 116;my italics). But what is there to be discovered in such a community? If there is no factual agreement about the ultimate norm of legal validity, it is meaningless to speak of a rule of recognition to be discovered unless Toh has now turned Hart's purported expressivism into a form of moral realism, or at least into a Dworkinian kind of moral objectivism. And this seems to be the case according to what Toh says in the last lines of the same paragraph: "even if the members disagree about the content of the rule of recognition, if they act as if there is a right answer to the question of what the true rule of recognition is, it seems that we can attribute to their community a legal system" (Toh 2005, p. 116;my italics). In other words, Toh's reconstruction of Hart's theory not only turns legal statements into statements of value but also turns empirical or social agreement on an actual rule of recognition into a matter of correctly answering a question whose exact nature is hard to specify. As far as I can see, this question could only be one that concerns the value of a rule of recognition. There is no other way to make sense of this paragraph, which depicts members of a community disagreeing about the content of a rule, the existence of which, as a matter of social fact, is disproved by that very disagreement.
Be that as it may, we still need to view Toh's appeal to Chapter 10 of The Concept of Law -the chapter on international law -as evidence that Hart allows for the possibility of the attribution of a legal regime (if not a legal system) to a community the members of which do not jointly accept a rule of recognition. Indeed, as is clear from his discussion of international law, Hart does not consider the rule of recognition or any secondary rule to be part of the definition of law. Toh is right to call attention to this. Nonetheless, the rule of recognition (like other secondary rules) is much more important to Hart's elucidation of law than Toh suggests. Not by chance, the title of Chapter 5 of The Concept of Law is Law as the Union of Primary and Secondary Rules.
Beyond its title, Chapter 5 makes clear that a "simple social structure of primary rules" (Hart 1961, p. 92) among individuals does not count as a legal regime at all. Indeed, Hart refers to the introduction of secondary rules -among them the rule of recognition -as "a step from the pre-legal into the legal world" (Hart 1961, p. 94;my italics). A few lines further, Hart refers to law -and, again, not merely to a legal system -as what may "most illuminatingly be characterised as a union of primary rules of obligation with such secondary rules" (Hart 1961, p. 94).
Thus, from the fact that Hart elucidates the law for a community of states without the concept of a rule of recognition in Chapter 10, it does not follow that he is willing to endorse the attribution of a legal regime to a community of individuals devoid of a rule of recognition. The indispensability of a rule of recognition to municipal law is due to the particularities of its factual background, as Hart (1961, p. 219) argues in Chapter 10. The point is that international law is devoid of centralised sanctions because of its particular factual background. This is why it can also be devoid of secondary rules such as the rule of recognition. As Hart (1961, p. 98) claims earlier, in Chapter 5: "secondary rules provide the centralised official 'sanctions' of the system".
Therefore, contrary to what Toh's expressivist reading of Hart leads us to believe, 1) municipal law is treated by Hart as the central case of law; and 2) only international law -that is, law without centralised sanctions -can dispense with a rule of recognition.
If my remarks in this and the previous section are sound, we have sufficient reason to dismiss the possibility of an expressivist semantics as a reading of Hart's analysis of internal legal statements on the basis of its unacceptable consequences. If Dworkin's claim (see for instance 1986, p. 66) that interpretation has two dimensions is correct -i.e. the dimension of fit and the dimension of justification -then it would seem that because Toh tries so hard to justify Hart's point of view in light of his own vision of what is required for good jurisprudence, his interpretation simply does not fit Hart's own jurisprudence. With this said, we still need to understand why Toh ascribes such an alien legal semantics to Hart and what a better alternative on this point might look like.

A PRAGMATIC APPROACH TO DETACHED AND COMMITTED STATEMENTS
As noted above, Toh's reconstruction of Hart's semantics of internal legal statements is based on committed internal legal statements, since these are paradigmatic internal legal statements (Toh 2005, p. 90). Committed internal legal statements are to be contrasted with detached legal statements. In The Concept of Law, Hart had only distinguished between internal legal statements -as statements that manifest speakers' acceptance of a normand external legal statements -as statements that state or predict regular behaviour. Nonetheless, this distinction, as Hart was willing to admit later, is insufficient. This being so, Hart came to accept the Razian distinction between committed and detached statements (see for instance Raz 1981, p. 305), as we will see below.
Even though (as Toh acknowledges) committed internal statements are paradigmatic cases of legal statements, detached statements are an essential addition to the analysis of legal statements insofar as they allow for an account of lawyers' statements "describing the contents of a legal system (whether it be their own or an alien system) whose rules they themselves in no way endorse or accept as standards of behaviour" (Hart 1983, p. 14; my italics; see also Hart 1966, p. 154). This report, Hart (1983, p. 14) adds, is given in "normative form"; thus detached statements are not the old external statements from The Concept of Law.
Why does this new distinction matter to an assessment of Toh's reconstruction of Hart's analysis of internal legal statements? My claim is that Toh mistakes the pragmatic element that distinguishes committed legal statements from detached legal statements -norm acceptance -for the pivotal element that constitutes the meaning of internal statements. In other words, I am claiming that the distinction between committed and detached statements is pragmatic, not semantic. Therefore, expressions of acceptance -as central as they are to Hart's analysis of legal statements -are not a matter of semantics in his theory of law.
Recently, Matthew H. Kramer (2018) has also argued, against Toh, that Hart's expressivism is better characterised as pragmatic, not semantic. With this noted, addressing the question of how the law gives reasons for action, David Enoch has constructed an argument that better supports my claim against Toh. The relevant passage from Enoch's paper is worth quoting in full: [I]t is quite possible that the normative flavor of internal, committed legal statements is not a part of their semantic content, but rather a part of their pragmatic features. Indeed, that this is so is strongly suggested by the thought that detached and non-detached legal statements behave logically as if there is no semantic difference between them: When a committed official says 'R is legally valid' and a noncommitted outsider says 'R is legally invalid', they seem to be genuinely disagreeing with each other, indeed contradicting each other, not talking past each other; furthermore, the outsider can report the official's view by saying 'Official said that R is valid' -and it's not clear how this can be so if there is a genuine difference in the semantic content of such locutions when uttered by the official and the outsider. (Enoch 2011, p. 23) Turning back to the thoughts that underlie Hart's acceptance of the distinction between committed and detached statements, we can understand legal statements uttered by judges to justify punishment for breaches of law as committed statements while understanding legal statements uttered by lawyers who are counselling clients on the verge of being punished for breaches of law as detached statements. Based on the passage from Enoch quoted above, we can see that if different semantics were to be applied to the meaning of these statements depending on who is uttering them, uncommitted lawyers would not be very useful to their clients since, as Enoch would say, judges and lawyers would be talking past each other. Hence, if judges and lawyers are not talking past each other, and if lawyers are not necessarily expressing acceptance of a norm when counselling their clients, acceptance is not part of the meaning of committed legal statements but an aspect of its pragmatics.
But has Hart himself not been guilty of taking norm acceptance to be a component of the meaning of internal legal statements in The Concept of Law? I do not think so, and I will explain why in the next section.

HART'S WEAK CONCEPTION OF NORMATIVITY
In Chapter 6 of The Concept of Law, Hart (1961, p. 105) considers "how the judge's own statement that a particular rule is valid functions in judicial decision". There, Hart (1961, p. 105) says that a judge's statement that "a rule is valid is an internal statement recognising that the rule satisfies the tests for identifying what is to count as law in his court". Hart then adds (1961, p. 105) that this statement constitutes "part of the reason for his decision". My suggestion is that the best way to rationally reconstruct this passage in light of the distinction between detached and committed statements (which Hart would later accept) is to view the recognition that a rule satisfies the test for identifying what counts as law as the meaning shared by those statements, and to view the function or use of the rule (as part of judges' reasons for their decision) as the pragmatic element that identifies committed statements in particular (or, in the language of The Concept of Law, internal statements).
The core Hartian semantic thesis at issue here is that a provision of reasons for action or decision is not contained in the meaning of internal statements. This semantic thesis makes Hart's conception of legal normativity a weak conception that is fully reducible in accordance with intrasystemic criteria of validity, which was already thoroughly clear in The Concept of Law: "We can indeed simply say that the statement that a particular rule is valid means that it satisfies all the criteria provided by the rule of recognition" (Hart 1961, p. 103;my italics). This is why, as we saw above (Section 1), there is no internal legal statement regarding the content of the rule of recognition.
To say that the judgment that a law exists is not about reasons for action and does not imply anything about reasons for action just in virtue of its meaning is to deny the first of three considerations that Toh claims to motivate expressivism: [E]xpressivists are struck by the fact that, at least for some normative concepts, the content of an assessment using one of those concepts implies, or the act of making the assessment requires, that any agent within the scope of that assessment -which can include the agent making the assessment -possesses a reason or motive to act according to the assessment… (Toh 2005, p. 79) Toh believes that Hart embraces this consideration when he says, at the beginning of The Concept of Law, that "the judge, in punishing, takes the rule as his guide and the breach of the rule as his reason and justification for punishing the offender" (Hart 1961, p. 11). According to Toh (2005, p. 82), Hart is here appealing to a "non contingent connection between a normative assessment, on the one hand, and assessing [a] person's possession of reason or motive, on the other, to reject the prediction theory [of law]". In addition, Toh (2005, p. 82, n. 14) argues that Hart calls the prediction theory out on its being unable to "explain why judges see themselves as having reasons to impose sanctions".
My take on this passage is quite different. Certainly, Hart is asserting that judges use legal rules as reasons for their decisions. But this is a description of what judges do as judges, a description made in order to point out that the prediction theory of law -in reducing legal statements to means of predicting judicial behaviour -is unable to account for such a phenomenon. Hart is not concerned with why judges treat the law as a source of reasons to impose sanctions. Judges are committed officials precisely because doing soi.e. treating the law as a source of reasons to impose sanctions -is their official duty. Speakers endowed with different roles -like the lawyers discussed above (Section 3), or even judges in different contexts of speech -can understand and utter the same legal statements that judges utter when justifying judicial decisions without considering such statements reasons for action. Therefore, the connection between the internal statement that a law exists and the consideration that it provides reasons for action is contingent: it depends on social roles and subjective motives.
I am glad to admit, however, that Toh (2005, p. 83) is (almost) right when he says that, according to Hart, "where a person makes a judgment that a law exists, he considers some action nonoptional or obligatory". Indeed, at least a certain type of law cannot be internally stated by speakers unless they consider some actions legally obligatory. Yet this is not to say that the mere recognition of laws that establish legal obligations entails acceptance of reasons for action for the subject of this recognition or for anybody else. For Hart, this is why the meaning of a legal obligation differs from the meaning of a moral obligation.
In Hart's view, judges can impose legal obligations in a way that assumes that offenders understand their legal obligations without presupposing that offenders should accept that there are reasons for action in conformity with such obligations. Judges have their own subjective motives for committing themselves to their legal systems and for using legal statements as reasons for their own actions. Nonetheless, as noted above, those who suffer the imposition of sanctions in accordance with legal statements are capable of considering those statements valid legal statements without accepting that anyone (even a judge) has reasons for action that are based on them. This last point is the core of my reading of Hart's conception of legal normativity, a conception that Hart (1966, p. 153) himself contrasts with Joseph Raz's cognitive "account of normativity in terms of reasons for action". What Hart means by the term 'cognitive' in the context of his debate with Raz is an account of statements of law ascertainment and law application that explains legal normativity objectively in terms of reasons for action. Accordingly, an expressivist non-cognitive account of the same statements explains legal normativity subjectively, in terms of acceptance of reasons for action. Neither of these accounts -the Razian (cognitive) or the expressivist (noncognitive) -describes Hart's own non-cognitive position, which simply disconnects normativity from reasons for action. This is Hart's clearest description of his account: [A]s the etymology of 'duty' and indeed 'ought' suggests, such statements refer to actions which are due from or owed by the subjects having the duty, in the sense that they may be properly demanded or exacted from them. On this footing, to say that an individual has a legal obligation to act in a certain way is to say that such action may be properly demanded or extracted from him according to legal rules or principles regulating such demands for action. (Hart 1966, p. 159;my italics) This meaning of legal duties or obligations allows judges to accept statements of law ascertainment and law application as reasons for their official actions while avoiding the necessity of appealing to moral motives (Hart 1966, p. 160;1982, p. 264). After all, since the meaning of these statements has nothing to do with reasons for action, a judge can meaningfully impose legal duties or obligations on offenders in a "technically confined way" (Hart 1982, p. 266), i.e. without presupposing that offenders will or should see such duties or obligations as reasons for action for them or for anybody else.

CONCLUSION
Toh's rational reconstruction of Hart's analysis of legal statements is insightful and thought-provoking, and this is why it has been so influential. Nevertheless, Toh's reading of Hart results in the displacement of the concept of a rule of recognition from centre stage in Hart's theory of law. Moreover (and related to this), his interpretation is unable to preserve the distinction between legal statements and statements of value. I have argued that these implications are sufficient to motivate the pursuit of a different reading of Hart's semantics of legal statements. The clue to this alternative reading is the distinction between detached and committed legal statements. While committed statements do express speakers' acceptance of norms, detached statements describe norm application and norm ascertainment in normative terms without expressing any kind of norm acceptance. Expressing norm acceptance and describing norm application or norm existence are different ways of using the same statement. The meaning shared by detached and committed statements of law application and law ascertainment is the conformity of actions and decisions to rules and the conformity of these rules to applicable criteria of validity. Therefore, legal normativity has a merely technical or intra-systemic meaning in Hart's legal theory. This is why the rule of recognition -as the source of legal criteria of validity -takes centre stage in Hart's semantics of legal statements, and this is why the distinction between legal statements and statements of value must be preserved from a Hartian point of view.