• 1. Introduction

  When he taught at Harvard University in 1957, H.L.A. Hart surprised his American students by comparing law to baseball (Lacey 2004, p. 185). John Searle is also fond of using games as examples. In his book on The Construction of Social Reality he talks at length about American football, for instance. But Searle’s most favourite example is chess (1995, p. 28, 66; 2010, p. 10). In his article ‘How Law is like Chess’, (2006a) the legal philosopher Andrei Marmor claims that indeed law and chess are in important respects comparable social institutions. From this comparison he builds up an argument that is meant to convince us that legal science and philosophy can and should be just as value free as science generally is, or at least generally aspires to be.
  Marmor is not the only one who believes this. In The Concept of Law and in the Postscript (published posthumously in the second edition of the book in 1992) Hart explicitly argues that his philosophy of law is meant to be descriptive (or conceptual) and morally neutral (Hart 2012, p. 240; Cp. Postema 2011, p. 341, 271; Marmor 2006b, p. 695; Lacey 2006, p. 954, 955). This is a heavily contested claim, but the legal positivism that Hart founded is developed further by what has come to be known as ‘institutionalism’, building on both Hart’s work and on John Searle’s speech act theory. If the arguments are convincing, this offers an interesting view on the practice and method of legal doctrinal research. In this article I propose to explore some of the pivotal arguments and examine their consequences for the method of legal science.
  The fact/value distinction appears to be at the core of Hart’s positivism. For the sake of clarity Hart considered it very important, before analysing and evaluating the law, to first distinguish what the object of legal research and philosophy is exactly. Indeed, according to Hart, this separation between fact and value (which informs the crucial distinction between law and morality) makes a critical evaluation possible in the first place (1958). The analysis of the concept of law is therefore meant to offer an exposition (rather than a strict definition¹x‘[The purpose of the book] is not to provide a definition of law, in the sense of a rule by reference to which the correctness of the use of the word can be tested; it is to advance legal theory by providing an improved analysis of the distinctive structure of a municipal legal system and a better understanding of the resemblances and differences between law, coercion, and morality, as types of social phenomena. The set of elements identified […] are treated as the central elements in the concept of law and of prime importance in its elucidation’ (2012, p. 17; Cp. Postema 2011, 265).) of the law as a distinct object of inquiry. Hart thought that this was not an empirical (psychological or sociological) matter but a conceptual one, which has important implications for the method of legal research. Before we can do empirical research we need to first elucidate the normative concepts involved by means of a thorough analysis.²x‘Hart’s reference to “descriptive sociology” was a gesture to the Austinian precept that an understanding of linguistic usage illuminates the world. […] But Hart’s own final statement of his position goes further yet. In one of his notebooks of the late 1980s, while accepting that he had been mistaken in making such a broad claim, he argued that a better formulation would have been that the book provided the ‘normative concepts required for a descriptive sociology’ (Lacey 2006, 949).
  In short, Hart argues that the object of study should first be determined and only then should judgments be made with regard to its merit (Marmor 2006b, p. 691; Lacey 2006, p. 956, 961). In his ‘Legal Positivism: Still Descriptive and Morally Neutral’ (2006b), Marmor literally quotes Hart in the title and defends his idea that it is indeed methodologically possible to describe without somehow evaluating (endorsing or criticizing) the object under scrutiny against those who are critical of this claim to objectivity and neutrality. Some are critical with regard to the claim to such objectivity in the exact sciences: is value-free science truly possible? But in the social or cultural sciences we may have even more reason to be sceptical. How can we ever do justice to human social and cultural phenomena if we do not truly and deeply understand, and engage with, the norms that structure them and the human needs, interests and values that guide them? Many have argued that a social and cultural phenomenon such as law is inevitably a matter of interpretation, not of description.
  In ‘The Practice and Discourse of Legal Scholarship’ (1988) Edward Rubin goes to the heart of the matter when he argues that an academic discipline such as law studies a particular social practice and a practice is not a body of objective information. Method is therefore not a matter of logic (that is, not a matter of ‘methodology’ in the sense of the ‘logic’ of scientific ‘method’) but of judgment. Rubin explains that
  

  ‘The defining characteristic of a practice, in contrast to methodology, is that its determinations of validity ultimately must be made on the basis of judgment, and perhaps intuition, rather than according to fixed rules. Being based on judgment, a practice is necessarily embedded in a culture and a language. As a result, the conclusions that any scholarly discipline produces are bounded by a cultural horizon that is finite, although not necessarily unchanging.’ (Rubin 1988, p. 841)

  
  H.L.A. Hart would not agree. Indeed, law as an institutionalized discipline is a practice, but – and possibly this distinguishes law and legal science from many other practices – the determination of the validity of a legal claim is not a matter of judgment but indeed generally made on the basis of rules. The constitution prescribes the rules and procedures for valid lawmaking, and such laws prescribe the rules and procedures for lower-level legal rules and standards to be valid law. The whole structure is derived from a general rule that is applied in legal practices with regard to what counts as a valid legal source. To be a bit more precise, Hart famously argued that both the officials of the law and legal subjects generally distinguish legal rules from other rules by means of a rule of recognition. This rule guides us to the valid sources of law and the valid legal rules and standards these contain. Of course, there is a need for judgment in hard cases, but often law is a matter of applying rules.
  Of course, you need to know what a social practice such as law is for, what purpose it has for the participants, if you truly want to understand the practice. Ronald Dworkin argues, for instance, that you do not know the law if you do not see the point of it all – the moral and political purposes that are part and parcel of this particular practice (Dworkin 1986, 2011). Such political and moral purposes determine, according to Dworkin, the judgment of what counts as law and what not. This implies that you do not only need to know the point, you have to make its purpose integral to your interpretation and therefore endorse or internalize the political morality underlying the law to truly understand your subject and to come up with a convincing interpretation. Dworkin argues that there is no alternative to this internal point of view and the engagement it requires.
  Andrei Marmor takes sides with Hart on this issue and argues that with regard to the matter of validity there is no such dependence on moral or political standards: ‘Non-positivists assert, while positivists deny, that moral considerations form an essential part of the conditions of legal validity, and conceptually so’ (Marmor 2006b, p. 687-688). This issue is important because science in general prescribes the obeisance of the distinction between facts and values, partly because our valuations tend to get in the way when we try to get the facts right. Science should be as value free as possible, Max Weber famously argued, science demands unprejudiced observation and analysis, because our prejudices with regard to the value of the phenomena under investigation tend to make us blind to the facts. Nicola Lacey argues that H.L.A. Hart was indeed influenced by Weber’s work (Lacey 2004, p. 229-231).
  I hope to clarify this discussion by changing the subject to a game. What can we learn from the chess analogy? Cannot we describe a game of chess without at the same time endorsing, valuing or criticizing this particular phenomenon? We need to know the rules of the game, to be sure, and we need to know what general and particular purposes and human values are involved; but this does not seem to make it impossible to stay neutral with regard to both the purposes of games in general and of the game of chess in particular. But does this also hold for the study of the law? Is it possible to stay at a critical distance from our subject of inquiry here, and still be able to describe, analyse and understand the law in a relatively detached way? The chess analogy suggests that indeed ­objective ­knowledge of the law is possible and that the method cannot be other than description and analysis, before we evaluate our findings by means of whatever evaluative perspective you choose.
  To put these claims in perspective I propose to first have a brief look at the philosophical method Hart was introduced to in Oxford by J.L. Austin. Subsequently, I will discuss the institutionalist conception of the law as a (at least partially autonomous) structure of regulative and constitutive rules or conventions. The chess analogy will lead us to the idea of legal science as, in Hart’s words, ‘still descriptive and morally neutral’.

• 2. Ordinary Language Philosophy

  In The Concept of Law (2012) H.L.A. Hart argued that at least two distinct characteristics distinguish law from other subjects of scientific investigation. For one thing, law is a normative phenomenon. Legal rules are prescriptive and therefore not some fact that is there for us merely to observe and explain in terms of causal relations. Legal rules are there to observe in the sense of to obey. The law confers duties on both citizens and the legal officials that administer the law. If you do not see this normative aspect of law, you do not see the law at all.
  However, not only is the law a set of norms that confer duties, it also consists of rules that confer particular competences or powers to particular individuals. In the Dutch legal order, for instance, according to the Grondwet the government and the parliament together have the power to proclaim the law. From one moment to t7he other, the statement ‘this is a valid law in this legal order’ has become a true statement merely by the legislature stating this fact. Such a declaration is called a ‘performative utterance’, and the law is full of such utterances. ‘I hereby pronounce you man and wife’, to give but one well-known example (Searle 1969; Cp. Culler 2000). Thus, the law entrusts many particular individuals and institutions, and not only those of the state, with the power to make a particular statement true by merely uttering it.
  The notion of ‘performative utterance’ was devised by J.L. Austin, who became, together with Gilbert Ryle, the most influential philosopher of what has come to be known as ‘ordinary language philosophy’. H.L.A. Hart was a close colleague of Austin in Oxford (they taught courses together), and arguably Hart’s Concept of Law would not have been written if Austin had not introduced Hart to his analysis of the concept of ‘rule’, a concept that became pivotal in Hart’s jurisprudence (Lacy 2004, p. 134). It is helpful to explore this approach to philosophy a bit further by contrasting it with the logical positivism it tried to supplant.
  Interestingly, Bertrand Russell and most members of the Vienna Circle who grounded logical positivism had a background in the exact sciences. But as the centre of gravity of British analytical philosophy gradually shifted from Russell’s Cambridge to Oxford, it shifted to a generation of philosophers that generally had a humanistic academic training (Schwartz 2012, p. 120; Lacey 2004, p. 137; Postema 2011, p. 264, 265). The Oxford philosophers J.L. Austin and Gilbert Ryle argued that the earlier generation had been too impressed with modern ­mathematics, logic and the natural science generally to appreciate the sophistication and subtle precision of ordinary language. The logical positivists were unimpressed by ordinary language because it could not match the exactness of mathematics and logic and was therefore generally vague and often even nonsensical (Cp. Ayer 1990).
  But Ryle and Austin turned the tables and came to think of logical positivist ­philosophy as crude and imprecise: in an effort to discover the ‘logic of science’, as Rudolf Carnap characterized philosophy (in: Rorty 1992, p. 54), logical positivists forced language into the straitjacket of a highly abstract and technical formal ­language. They thus impoverish the language to such a degree that they were blinded to the subtle and rich structure of the ordinary, natural language we have at our disposal to talk about the world. We should instead take ordinary language at face value and, as J.L. Austin argued, ‘avoid at all costs oversimplification, which one might be tempted to call the occupational disease of philosophers if it were not their occupation’ (1962, p. 38).³xOxford philosopher Peter Strawson: ‘… ordinary people employing the ordinary conceptual resources of mankind have at their disposal not a crude, rough-and-ready instrument, but an enormously sophisticated instrument, for thinking; and it’s an immensely and inexhaustibly interesting task to trace the various connections between the concepts which people handle in their ordinary life with no particular difficulty’ (in: Magee 1971, p. 136, 137; Cp. Lacey 2006).
  Austin was interested in the language of the law in a way logical positivists never were, and this marked his relationship with the former barrister H.L.A. Hart. Hart recognized that the phenomenon of performative utterances was indeed pivotal in law. He furthermore recognized that Austin’s way of doing philosophy was similar, or at least related to the way lawyers analyse legal rules and concepts. The language of the law could be thought of as just as subtle and rich in structure as ordinary language generally. Hart realized that
  

  ‘… his detailed knowledge of the subtle texture of legal reasoning, provided him with a fund of examples ripe for philosophical analysis. […] He began to realize that legal usage provided example of conceptualization and distinction-­making which could be put to use in philosophy.’ (Lacey 2004, p. 144)

  
  Indeed, the influence went both ways. The analysis of legal concepts could be highly instructive for philosophers, just as philosophical analysis could take legal science to a higher level. Nicola Lacey argues that, on the one hand, Hart’s expertise contributed significantly to the development of Austin’s ‘speech act’ theory (2004, p. 145) and that, on the other hand, Austin’s analysis of the meaning of the concept of ‘rule’ as it is ordinarily used proved pivotal for Hart’s analysis of the concept of law.⁴xThe notion of a rule is, of course, not only important for lawyers and moralists, but in fact fundamental to language in general. Speaking a language is following rules. For this reason Wittgenstein famously devotes a substantial part of his Philosophical Investigations (1953) to all kinds of rule-following, and his analysis has proven to be of great interest for philosophers and linguists generally.
  How did Austin go about his philosophical exploration of the concept of ‘rule’? He knew that rules are of many kinds and that therefore in order to analyse the meaning, and hence the use, of this word ‘rule’ thoroughly, all sorts of actual rules should be investigated: rules of grammar, rules of particular games or sports, rules of evidence, etc. His close colleague Geoffrey Warnock recounted how Austin apportioned among his group of collaborators particular kinds of rules to analyse, thus systematically dividing the big problem into limited tasks, the better to cover much ground (Magee 1971, p. 96).
  H.L.A. Hart was not part of this particular group at the time, but he seems to have greatly profited from this exercise. The Concept of Law probably would not have been written without J.L. Austin’s work in general and his work on the notion of the rule in particular (Lacey 2004, p. 156). Hart realized that the influential, positivistic analyses of law by Jeremy Bentham and John Austin were seriously flawed because they mistook the notion of a legal rule for the notion of a command, which is quite a different sort of thing. As opposed to Bentham and Austin, Hart came to think of the notion of a rule as basic to a modern legal order: the law is indeed (maybe among other things; as stated previously, Hart was not looking for one strict definition) a system of rules, and this shift in emphasis greatly changes the way we look at the subject.
  John Austin’s idea of law as ‘the command of the sovereign backed by threats’ had, apart from being clear and concise, the merit of being an attempt to make possible a scientific characterization of law, in the sense that the concept of law was clearly delimited and reducible to observable facts. This definition not only straitjacketed what we mean by the word ‘law’ but also presented the law from an explicitly external point of view. With regard to the natural world this is, of course, the only point of view that is available to us. But with regard to the social world we miss something important if we keep looking at it from the outside, John Searle argues, ‘because the description of the overt behaviour of people […] misses the underlying structure that makes the behaviour possible’ (1995, p. 5).
  Indeed, as Hart writes, such an external, strictly scientific point of view loses sight of an important feature of a (legal) rule: its normative character.
  

  ‘If […] the observer really keeps austerely to this extreme external point of view […] his description of their life cannot be in terms of rules at all, and so not in the terms of the rule-dependent notions of obligation or duty. Instead, it will be in terms of observable regularities of conduct, predictions, probabilities, and signs. […] What the external point of view, which limits itself to the observable regularities of behaviour, cannot reproduce is the way in which the rules function as rules in the lives of those who normally are the majority of society. […] For them the violation of a rule is not merely a basis for the prediction that a hostile reaction will follow but a reason for hostility.’ (2012, p. 89, 90)

  
  Interesting from a methodological point of view is that in his introduction to The Concept of Law, Hart notes that this difference between the external and the internal points of view ‘shows itself linguistically’. He distinguishes here between mere convergent or conformist behaviour and truly following a rule, and argues that
  

  ‘In describing the latter we may, though we need not, make use of certain words which would be misleading if we meant only to assert the former. These are the words ‘must’, ‘should’, and ‘ought to’, which in spite of differences share certain common functions in indicating the presence of a rule requiring certain conduct.’ (2012, p. 9)

  
  The law does not generally feature in the lives of the legal subjects as some external force to be reckoned with, but as a system of rules that they experience as a source of obligations (2012, p. 82). In other words, although the members of a legal community sometimes have to be forced to comply if they choose to disrespect the law, they generally choose to comply voluntarily on the basis that they feel obliged to do so. They do so when they acknowledge that some rule is actually the law and when they see law as a source of norms, not merely of ‘commands backed by threats’.
  If the law did not have this normative aspect, the resources needed to order a human society would be beyond measure; no orderly society could exist without this sense of obligation. But H.L.A. Hart and J.L. Austin seem to have learned this, and this is important to note here, not by means of empirical research (which employs a more or less external point of view), but by means of the analysis of ordinary language, in this case of how we use the word ‘rule’. The existence of this normative realm ‘shows itself linguistically’, Hart says, and by analysing its specific character and meaning we learned something about the law that we would not learn from the external, scientific point of view.⁵xWe should also note that in the normative domain, generally, norms need not be practised in order to be valid. Indeed, we know of many perfectly valid moral and legal rules that are often not followed. To express this categorical difference we distinguish facts from norms. Facts have an empirical reality and norms have validity. Somehow, somewhere, facts and norms are aspects of one reality, of course; Hart argues that the rule of recognition is where factuality and validity meet.

• 3. The Separation Thesis

  We should strictly distinguish legal phenomena from other normative phenomena such as morality or rules of etiquette, Hart argues, and he uses the same method here. What we refer to when we use the word ‘law’ and ‘morality’ is different in kind, and any competent user of the English language should be able to tell the difference. Moreover, not only do legal scientists actually make this distinction, legal officials and professionals working in legal practices and institutions, and many lay people, actually apply a particular rule to distinguish legal norms from other social norms in society. I refer to Hart’s ‘rule of recognition’,⁶xHart introduces the rule of recognition thus: ‘The simplest form of the remedy for the uncertainty of the regime of primary rules is the introduction of what we shall call a “rule of recognition”. This will specify some feature or features possession of which by a suggested rule is taken as a conclusive affirmative indication that it is a rule of the group to be supported by the social pressure it exerts’ (2012, p. 94). ‘The rule of recognition is neither itself a rule of law nor a transcendental presupposition; rather it is a social rule accepted and practiced by judges and other legal officials’ (Postema 2011, p. 285, 286). of course, and ‘its existence is shown in the way in which particular rules are identified, either by courts or other officials or private persons or their advisors’ (2012, p. 98).
  But note that this is actually an empirical claim. Although legal scientists and philosophers of law may often not be too interested in the actual behaviour of human beings in society when they study law, there is one important exception to this rule. Hart explains,
  

  ‘[A] rule of recognition is unlike other rules of the system. The assertion that it exists can only be an external statement of fact. For whereas a subordinate rule of a system may be valid and in that sense ‘exist’ even if it is generally disregarded, the rule of recognition exists only as a complex, but normally concordant, practice of the courts, officials, and private persons in identifying the law by reference to certain criteria. Its existence is a matter of fact.’ (2012, p. 110)

  
  Hart argues (in short) that the existence of a legal order presupposes the ‘separation between law and morals’ and that in a modern society such as ours we apply a strictly formal rule of recognition to determine whether a norm is a legal norm or not (Leiter 1998, p. 534, 535). This rule of recognition is not everywhere the same – the Dutch use a different rule from the Americans, for instance – but Hart claimed that indeed the existence of such a rule is characteristic of a modern legal order. And to repeat, this is an empirical claim: a particular and significant part of society actually uses this rule of recognition to decide what the law demands. In particular, legal officials and professionals will do so.
  In legal philosophy, this thesis has come to be called ‘legal conventionalism’ as the employment of a rule of recognition is thought of as a matter of convention (Hart 2012; Marmor 1998; Postema 2011, p. 342). What does this mean for Hart’s version of the so-called separation thesis, the thesis that law and morality are conceptually distinct? Hart does not use the term, but in effect he argues that a legal system has a different structure. Hart depicts the law as a system of two kinds of rules, ‘primary rules’ and ‘secondary rules’, and he argues that particularly the intricate structure of these secondary, power-conferring rules coordinating the set of primary, duty-conferring rules (conferring specific duties on the legal subjects) distinguishes law from other normative phenomena such as morality. In Hart’s own words:
  

  ‘Under rules of the one type, which may well be considered the basic or primary type, human beings are required to do or abstain from certain actions, whether they wish to or not. Rules of the other type are in a sense parasitic upon or secondary to the first; for they provide that human beings may by doing or saying certain things introduce new rules of the primary type, extinguish or modify old ones, or in various ways determine their incidence or control their operations. Rules of the first type impose duties; rules of the second type confer powers, public or private. Rules of the first type concern actions involving physical movement or changes; rules of the second type provide for operations which lead not merely to physical movement or change, but the creation or variation of duties or obligations.’ (2012, p. 81)

  
  Every legal rule is, qua legal rule, valid on the basis of some power-conferring secondary legal rule, which is itself valid on the basis of some higher power-conferring legal rule. The rule of recognition functions as the ultimate ‘foundation’ or ‘source’ of the hierarchical structure of rules that is thus created. Indeed, the fact that the whole structure of rules is founded in this rule of recognition makes the law into a systematic whole, makes it a system of rules ‘… for the rules are now not just a discrete unconnected set but are, in a simple way, unified’⁷x‘One distinguishing mark of constitutive conventions […] is that [they] typically come in systems. There must be a system of conventional rules to establish a social practice, that is, some cluster of rules intertwined in a more or less complex structure. This is a feature that stems from the constitutive function of such conventions, from the fact that they constitute a practice’ (Marmor 1998, p. 522; Cp. Searle 1995, p. 28). (Hart 2012, p. 95).
  Moral systems may consist in primary rules and thereby confer duties on individuals within a particular society, and they may also be relatively unified in the sense of being systematically and hierarchically ordered, but there is no explicit and consistent set of secondary rules making clear, for instance, what exactly should be done in the case of violation of the rule, who exactly is competent to decide whether a violation has taken place, and how exactly we should react to such a violation. Indeed, according to Hart, supplementing the primary rules with secondary rules regulating such matters can be thought of – conceptually, not historically – as the decisive step from the pre-legal (moral) world into a modern legal world like ours (2012, p. 94; Cp. Marmor 2002, p. 106).
  Note that a power-conferring rule authorizes a particular person or institution to particular performative utterances. We already referred to the state official who is empowered to declare two persons man and wife, a competence that is usually explicitly conferred on officials in a particular legal statute. Thus, the official is empowered by a legal statute; but what makes this statute valid? The validity of this statute depends on the power of the formal lawgiver to make valid ­statutory laws, and to change the law, a power that is conferred on this institution by the constitution. Treaties are, likewise, valid law because the legislative power is authorized by the constitution to declare them such. The constitution furthermore empowers judges to decide legal conflicts on the basis of the law and the constitution is itself valid because it was declared by the formal constitutional lawgiver, often a qualified majority in parliament. This chain of validity is hier­archically ordered, but the buck has to stop somewhere. Hart argues that the rule (or convention) of recognition is ultimately a particular practice people regard as obligatory, and this is where the chain is ultimately founded; this conventional social fact is the fountainhead of law as we know it.

• 4. Regulative and Constitutive Rules

  John Searle offers an account of how we collectively create such a legal order that has proved to be highly influential among Hart’s followers. Searle’s The Construction of Social Reality (1995) seems to fit Hart’s positivism like a glove because the distinction between duty-conferring primary rules and power-conferring secondary rules coincides (at least largely) with Searle’s distinction between ‘regulative rules’ and ‘constitutive rules’ (Cp. Weissbourd & Mertz 1985, p. 627, 628). This has opened up an interesting perspective for understanding Hart’s analysis of the concept of law, and many have since explored this idea. The result is known as ‘institutionalism’, pioneered by Neil MacCormick and Ota Weinberger in their groundbreaking An Institutional Theory of Law (1986). I will, however, focus on the work of Andrei Marmor here, whose version ‘Legal Conventionalism’ also builds on the work of Hart and John Searle. Let me try and give a short outline of law as a system of regulative and constitutive rules.
  Regulative rules regulate a previously existing practice, while constitutive rules are at the very foundation of the practice involved in the sense that they form its structure. In Searle’s own words: ‘constitutive rules do not merely regulate, they create or define new forms of behaviour’ (1969, p. 33, 1999, p. 123; Cp. Marmor 2006a, p. 350). The convention prescribing that we drive on the right side of the road, for instance, solves a coordination problem in the sense that it is not important on which side in particular we drive but all the more that we all drive on the same side of the road (Cp. Lewis 2002). This convention is thereby an example of a regulatory rule that is premised on the need to behave in a predictable and therefore regular way, not on the content of the particular rule. Many legal rules are of this nature, of course.
  Marmor, however, argues that, although conventions are indeed often the relatively arbitrary solution to coordination problems, there are also conventions in society where there was no previous coordination problem to solve. Such ‘constitutive conventions’, as he calls them, make a particular practice possible in the first place.⁸x‘Conventions evolve in response to a wide variety of social needs, and the need for coordination is just one of them. For example, conventions constitute various ways in which we express respect to one another, both linguistically and otherwise, conventions constitute artistic genres, games, and numerous social institutions; social conventions often regulate interpersonal relationships in such settings as a workplace, or a party, or an academic gathering, and such. In short, conventions serve numerous functions by constituting ways in which people interact with each other and engage in socially valuable activities’ (Marmor 2006a, p. 357). The practice has a point, of course, and the particular human needs and values it serves are part of the meaning of the constitutive rules involved, but there is no previous practice that is now regulated. Moreover, as these rules are constitutive of the practice, the practice can be thought of as (at least partly) autonomous: ‘its autonomy […] consists in the fact that the conventions constituting the practice are radically underdetermined by those general values and human concerns that they instantiate’ (Marmor 1998, 520-523).
  The game of chess provides us with the classic example. Most, if not all, of the rules of this game are constitutive in the sense that without such rules chess would not exist – illustrated by the fact that when we change the rules of this game we start wondering whether we are still playing chess or some other game. Playing games in general satisfies the human need or desire for playful, competitive social interaction, which is, at the bottom line, also the point of playing chess.⁹x‘The rules constituting chess are constitutive conventions. They constitute what chess is and how to play the game. In part, they constitute the point of playing the game and some of the specific values associated with it. But all this is possible only against the background of a deeper layer of shared understandings about what competitive games are. In order to play chess as a competitive game of a certain kind, players must follow a fairly elaborate set of norms, or deep conventions, determining the concept of games and the essential point(s) of engaging in such a practice. Chess, as a game of a particular kind, is only an instantiation of a more general human activity that we call “playing a competitive game.” […] Deep conventions typically come to our attention when we face some deviant behavior or some doubts about the practice we are engaged in. […] The rules of recognition of the kind Hart had in mind are surface conventions. They determine what counts as law in a particular legal system or in a particular community. These surface conventions of recognition are instantiations of deep conventions about what law in our particular culture is. Roughly, the idea is that the rules of recognition relate to the deep conventions of law like particular games relate to deep conventions about what it is to play a competitive game’ (Marmor 2006a, pp. 364, 366, 369; Cp. Searle 1995 on ‘the Background’). However, our desire for competitive interaction could be satisfied in many different ways, and Marmor argues that this need radically underdetermines what the exact rules and particular goals, and thereby the particular character of the game, will be. In other words, we know only very little about chess if we only know the human need or values it serves. Importantly, exactly this fact makes the practice of playing chess autonomous from other games. When we know the rules, we recognize a game of chess when we see it; we are not likely to confuse it with a game of football, for instance. Chess has, although it serves human needs and values that are also served by other practices, a distinct character of its own.
  Indeed, legal institutionalist theory is likewise premised on the idea of the distinct character, and thereby the autonomy of legal institutions (Ruiter 1998; ­MacCormick & Weinberger 1986). The human needs and values that are served by the law also underdetermine the specific character the law will have in a ­particular society. Marmor argues, however, that social practices such as law are never completely autonomous (not even the game of chess) because they are all related to some general human concerns, and such general concerns inform many different practices that are thereby related (Marmor 1998, 523, 530). Here lies a possible weakness of institutionalism. Is the autonomy of the law as a social practice not partly the product of a rationalization, or even idealization, of the practice by legal science and philosophy?¹⁰xWe should, however, note that in many continental ‘civil’ law systems the law is largely the product of legal science. This is the kind of science that can make its claims to what is law true when judges or the legislature make the results of legal scientific analysis and rationalization into law. If the practice is largely a matter of rule-following we might say no; if, by contrast, the practice is largely a matter of judgment, the idea of the systematic autonomy of law is just another scientific formalist dream.
  According to the institutionalists, however, the autonomy of the practice, however partial, as the result of the fact that it is a structure of constitutive conventions or rules, has important epistemological and methodological consequences. In Speech Acts (1969) John Searle argued that many of these constitutive rules are essentially analytic in the sense that they give definitions and confer powers but do not describe independently existing facts.¹¹xQuine (1953) famously argued that there is no such thing as purely analytic (‘grounded in meanings independently of matters of fact’) as opposed to purely synthetic truths that are grounded in fact. Grice and Strawson defended the distinction (1956) by pointing to the particular use that is made of it in philosophical circles. I assume here that Searle’s use of the term ‘analytic’ is well founded and consistent. Such rules specify, for instance, what counts as a valid move or what counts as winning the game but do not have any empirical content. Thus, constitutive rules can be thought of as analytical statements, in the sense that they totally depend for the truth or falsity on the definition. In other words, constitutive rules are true by definition:
  

  ‘Regulative rules regulate a pre-existing activity, an activity whose existence is logically independent of the rules. Constitutive rules constitute (and also regulate) an activity the existence of which is logically dependent on the rules. Regulative rules characteristically take the form of or can be paraphrased as imperatives. […] If our paradigms of rules are imperative regulative rules, such non-imperative constitutive rules are likely to strike us as extremely curious and hardly as rules at all. Notice that they are almost tautological in character, for what the ‘rule’ seems to offer is part of a definition of ‘checkmate’ or ‘touchdown’. That, for example, a checkmate in chess is achieved in such and such a way can appear now as a rule, now as an analytic truth based on the meaning of ‘checkmate in chess’. That such rules for checkmate can be construed as analytic is a clue to the fact that the rule in question is a constitutive one.’ (p. 34)

  
  Driving on the right side of the road is like chess in that it is also a rule-based activity. But in this case we can change the rules: we might drive on the left side of the road, for instance, without structurally or essentially changing the practice of driving on the road. We are therefore dealing with a regulative rule here and not with a constitutive rule. But we should also note something else about the difference between driving and chess. You can understand road travelling without knowing the particular rule to keep on the right side of the road. By contrast, you cannot understand chess, you cannot know what chess is, without knowing the particular rules of this game. Indeed, even if you know that human beings like playful competitive interaction, without knowledge of these rules what you see when you watch two people play is completely incomprehensible.
  Let us now return to what this example may teach us about the law. Surely law consists of regulatory rules. The rule to keep on the right side of the road is actually a legal rule on the European continent, and so is the rule that forbids citizens to kill another human being. This latter rule is regulatory in nature. The fact that human beings kill other human beings is a fact of life, whether such a legal rule forbidding it exists or not. In order to prevent such violent and socially disruptive behaviour, the law prohibits it, and threatens to punish the violation of this rule. Criminal law thus has (at least) an ordering function in society, and this function is largely to prevent violence and to retaliate if necessary. In the words of John Searle,
  

  ‘The whole point of the criminal law is regulative, not constitutive. The point is to forbid, for example, certain antecedently existing forms of behaviour such as killing. But to make the regulations work, there must be sanctions, and that requires the imposition of a new status on the person who violates the law. […] Thus the regulative ‘Thou shalt not kill’ generates the appropriate constitutive ‘Killing, under certain circumstances, counts as murder, and murder counts as a crime punishable by death or imprisonment.’’ (1995, p. 50)

  
  What constitutive rules essentially do is specify the particular conditions under which a particular ‘status function’ is imposed on some person or object.¹²xSearle (2010): ‘The distinctive feature of human social reality, the way in which it differs from other forms of animal reality known to me, is that humans have the capacity to impose functions on objects and people where the objects and the people cannot perform the functions solely in virtue of their physical structure. The performance of the function requires that there be a collectively recognized status that the person of object has, and it is only in virtue of that status that the person or object can perform the function in question. Examples are pretty much everywhere: a piece of private property, the president of the United States, a twenty dollar bill, and a professor in a university are all people or objects that are able to perform certain functions in virtue of the fact that they have a collectively recognized status that enables them to perform those functions in a way they could not do without the collective recognition of the status’ (p. 7). Many human activities and practices that are regulated by means of legal provisions existed before these regulations came into force and will remain existent if these regulations are changed or abolished. Killing other human beings is regrettably among them. The judge is empowered to qualify a particular action as a criminal act and to subsequently declare the suspect guilty as charged by means of a performative utterance. Searle explains that a very simple formula is involved here: ‘X counts as Y in context C’ (Searle 1995, p. 43, 2010, p. 10). In the context of a criminal proceeding the premeditated and intentional killing of another human being counts as murder, and the judge is empowered to decide what punishment is due.
  What is distinctive of the law, then, as opposed to other normative systems in society, is that in addition to the primary or regulative legal rules it also consists of a supplementary set of rules that specifies, as clearly and distinctly as possible, the conditions under which a person can be lawfully qualified as a murder suspect and thus arrested. Such rules also regulate the legal procedure according to which the suspect’s guilt should be established and when, and by whom, the murderer can at the end of such a procedure be convicted to a long-term imprisonment.
  To be sure, the law here satisfies the human need for the predictability of the state’s reaction to certain behaviour and serves the good or value of protecting individuals against the arbitrary use of state power. This may be true, but this particular human need radically underdetermines what shape the rules will take in a particular legal order. In other words, we cannot say we know the law until we know all the specifics, the particular constitutive and regulatory rules involved.¹³xMarmor argues that this explains the conventional nature of the rule of recognition and the rules it identifies as legal rules. The legal order is ‘radically underdetermined’ by the general purposes and values it serves, and the legal order can therefore take different shapes and colors. ‘First, it is crucial to realize that the needs, functions, or values that such conventions respond to radically underdetermine the content of the rules that constitute the relevant social practice. Our need or desire to play some competitive intellectual board game does not determine the content of the rules of chess; the creative needs or desires to stage dramatic performances do not determine the particular theatrical genres that have emerged over time in various cultures; the reasons for showing respect to people under various circumstances do not determine the practices of etiquette that we have’ (2001b, p. 360; Cp. Postema 2011, p. 524). When you change these rules significantly, criminal law as a social phenomenon undoubtedly changes beyond recognition. Indeed, if you change or abolish enough of these rules we do not even recognize it as law anymore.
  Thus, law is a complex whole of primary, duty-conferring rules, on the one hand, and secondary, power-conferring or constitutive rules, on the other. Constitutive rules or conventions do not merely regulate but also constitute legal practices and institutions. The complex and intricate set of these constitutive rules gives the law its specific structure or character; and we can therefore say that Hart was right when he said that in modern societies a legal system is structurally different from other normative phenomena. If we want to know what the law demands of us and what this means exactly, we turn to this particular system of primary and secondary legal rules. This system is also the prime subject matter of what is traditionally known as doctrinal legal research.

• 5. Language and Institutional Facts

  Let us now explore the implications of what we have learned. To begin with, let us change our perspective and try and look scientifically at a game of chess; that is, from a strictly external point of view (while keeping the example of law in the back of our mind). What do we see exactly when we closely watch two people play? Of course, we can discern patterns in the movements of the two players. From these patterns we might induce that particular physical laws are at work. But notice that these physical laws are not truly perceptible. David Hume made us aware of the fact that we only see patterns; we do not actually see the laws that explain the regularities. Seeing the physical laws behind or underlying these patterns or correlations is a matter of interpretation. We, so to speak, ‘add’ this information as the result of previous experience (Hume argued, in fact, that the fact that we see causality is the product of custom) or – and there is scientific evidence for this – as the result of our particular biological make-up (Kahneman 2011, p. 76).
  What can never be truly perceived by someone taking the external point of view is that these patterns are the result of not merely certain physical laws at work but also the players following certain rules. This was, of course, a major point for Hart. Seeing people abide by rules is certainly a matter of interpretation: because we know the rules of the game of chess we understand the reasons for the moves the players make, and we see how these moves contribute to the ultimate goal of defeating the other by checkmating the king. Like causality, we do not truly see these rules; seeing the behaviour as the result of following rules is also a matter of interpretation that requires particular knowledge of and experience with the rules in question. Hart explicitly argued that it requires an internal, hermeneutical point of view.¹⁴x‘For the understanding of this the methodology of the empirical sciences is useless; what is needed is a “hermeneutic” method which involves portraying rule-governed behavior as it appears to its participants, who see it as conforming or failing to conform to certain shared standards’ (Hart 1983, 13; Cp. MacCormick & Weinberger 1986, 15).
  One might therefore argue that the physicist and the analyst of social institutions such as a game of chess work on different levels of description. The physicist analyses the patterns in order to formulate a theory regarding the physical laws that explain patterns in the behaviour of natural objects. The analyst of the game of chess, by contrast, analyses the patterns in the behaviour of the two players and proposes an understanding in the sense of an interpretation of this behaviour within the context of the general rules and purposes of playing a game of chess. Without any foreknowledge of this particular context we do not see the game; and it is actually a defining feature of the physicist’s habitus, and its external perspective, to abstract away from such social meanings to focus on the underlying physical structure. Indeed, with regard to the natural world we have only an external perspective, while with regard to our social world an external perspective does not truly exist because from such a point of view it totally vanishes.
  There are a number of interesting things we should now note about the rules of chess. First, although these rules are formulated in natural language terms, they are not so vague and indeterminate as we might expect. (Bertrand Russell would agree that the rules of chess are, for one thing, much more clear and distinct than much of the traditional philosophy he so vehemently criticized is.) Indeed, when in doubt, you can refer to a rule book. The rules of chess are now accepted all over the world as they are collected, unified and ‘codified’ in a rule book that is officially recognized by the World Chess Federation. Chess players, and especially chess officials and referees worldwide, thus use a (conventional) rule of recognition to identify the rules of the game. Many even know these rules by heart. In other words, chess is a highly institutionalized practice.
  Second, unlike traffic rules that were referred to, the rules of chess do not regulate some preexisting practice. As already mentioned, without these rules the game of chess would not exist at all, and changing one of those rules immediately raises the question whether we are still playing chess or some other (new) game. Many games do have rules that can be thought of as arbitrary in the sense that we might change them without changing the character of the game beyond recognition. Hart often refers to such regulatory rules of a game, but he also recognizes the existence of rules ‘which veto certain types of conduct under penalty and secondary rules which specify what must be done to score or to win’ (1961, p. 9). The latter will most likely be essential in the sense that changing them will literally be a game changer; they are thus constitutive rules or conventions.
  This implies that the game cannot be adequately described and explained by means of a set of rules, terms, concepts and distinctions other than those that are used in the practice under scrutiny. We can describe and explain natural phenomena with the help of mathematical equations (which are not constitutive of the natural order), but we cannot, for instance, describe legal institutions and practices in non-legal terms without seriously distorting them. Searle argued that constitutive rules are analytic, and changing the jargon will therefore be a game changer in the sense that we will not recognize it to be about law anymore. Legal scholars therefore ‘employ the same legal terms, and treat them with comparable levels of respect or disdain. This shared conceptual framework between legal scholars and their subject matter may be referred to as the unity of discourse’ (Rubin 1988, p. 1859-1860).
  Importantly, and this is the third point, the game of chess consists in a system of rules and is therefore completely normative in nature. Searle argues that the whole of our social reality is indeed normative in this sense. The basic structure of our social reality, and that of the law as a distinct part of it, is thus deontological. Searle argues that the whole structure is based on a systematic whole of collectively recognized rights and duties; that is, on a complex system of social competences and obligations that are conferred on persons and things (used by persons) with the help of constitutive rules (2010, p. 80-84). Hart’s analysis of law as a system of rules seems congruent with Searle’s theory in that, as I have argued, in both cases it is rules all the way down.
  We should now note one last thing about the game of chess: without a language to formulate and communicate the rules of chess this game would not exist. ­Playing a game of chess is therefore completely language dependent in a way that the existence of the natural world around us is not – or so Searle argues (1995, p. 60-66). This is also true of our social reality generally, and of the law in particular. But the fact that this practice is language dependent does not mean that the chess analyst, or anyone who knows the rules and purpose of this game, cannot make objective statements about, for instance, the famous World Championship match between Boris Spassky and Bobby Fisher in 1972. Searle argues that, within a particular practice such as chess, the statement that Fisher won this game by 12½–8½, and thus became the new World Champion is a true statement. These are thus perfectly objective facts, he says, albeit facts of a particular kind: ‘institutional facts’.
  

  ‘Some facts exist independently of any human institution. I call these brute facts. But some facts require human institutions in order to exist at all. An example of a brute fact is that the Earth is 93 million miles from the sun, and an example of an institutional fact is that Barack Obama is president of the United States. Institutional facts are typically objective facts, but oddly enough, they are only facts by human agreement or acceptance. Such facts require institutions for their existence. […] And what exactly is a human institution? […] An institution is a system of constitutive rules, and such a system automatically creates the possibility of institutional facts.’ (2010, p. 10)

  
  This means that, although the rules of chess are normative – they prescribe what you should and should not do – playing a game of chess is in itself a fact. Being language dependent, such a fact is, of course, not a ‘brute fact’ of nature like Mount Everest is but what Searle dubbed an ‘institutional fact’ (Searle 2010, p. 10). Such institutional facts are the product of collectively ascribing or imposing a particular function and status on particular objects, persons and behaviour. When you participate in such an institution you can objectively describe, analyse and explain what people do, and exactly what it is that they are trying to achieve. Indeed, Searle argues that although institutional facts are a matter of interpretation and therefore ‘ontologically subjective’, the fact that people collectively recognize the status functions and meanings of the particular moves in the practice makes institutional facts ‘typically objective facts’. Thus, although they are ontologically subjective, they are ‘epistemologically objective’ (1995, p. 8).
  Hart’s argument has been construed by institutionalists as saying that this also goes for law. Law is just as language dependent as chess: without language, law does not exist. But from within legal practices and institutions we can make epistemologically objective statements about the law, just as we can about chess. The whole complex of secondary rules in a legal order can be thought of as constitutive of that order; it is, in other words, what characterizes the legal order as such and thereby distinguishes it from other social phenomena such as morality. We can thus link Hart’s analytical jurisprudence to Searle’s social constructivist philosophy: Hart’s secondary rules are (at least largely) constitutive rules in Searle’s sense.
  There is one last important implication I want to mention here. This concept of law as an institutional fact makes the idea of objective legal science feasible. We can study the law as a system of regulative and constitutive rules or conventions in a relatively detached way; that is, we can describe, analyse and understand the law without actively endorsing, or committing to, the normative claims such laws make or the values and interests they serve (Hart 1983, p. 14-15; Postema 2011, p. 289). Indeed, legal positivism is ‘Still Descriptive and Morally Neutral’, Marmor writes, and we can objectively describe and analyse the law before we evaluate it according to, for instance, a particular conception of justice or morality. Thus, a legal scientist can, and should, meet the criterion of providing a value-free analysis of his subject matter.

• 6. Conclusion

  What did the chess analogy teach us? According to the legal positivists, we can think of law as a practice that is, at least, partially autonomous and built out of constitutive rules or constitutive conventions. Law and chess are both heavily institutionalized practices, and that makes the comparison the more interesting. We all learned what chess is by learning what the exact objective of the game is and what the rules are, and subsequently by playing the game according to its rules, of course. When we are in doubt about the rules, we can refer to an authoritative institutional source to find them. Chess players and legal subjects alike apply some kind of rule or convention to distinguish the rules of their ‘game’ from the rules of other practices, institutionalized or not.
  Of course, chess is much more autonomous than law is, but the analogy helps us understand how the law can be partially autonomous in the first place. Furthermore, just as we can distinguish chess from other games (although they serve the same fundamental function), we can see how law has its own distinct structure as compared with other norm systems in society. Marmor argued that law, like chess, also has a certain independence from the human needs and values that it serves, because these underlying purposes do not determine the exact rules and standards of the law as it stands. Thus, to know the law we turn to the sources of law as they are identified by the rule of recognition; there is no necessary connection with rules and standards of other normative phenomena in society such as morality.
  Another interesting finding is that the game of chess, or a particular game for that matter, cannot truly be described in terms other than its own. Likewise, law is a structure of norms that can be described only from an internal perspective and in its own terms; an external perspective and a foreign jargon will distort the subject under inquiry because it will change it beyond recognition. Hart distrusted the reductionist tendencies in the social sciences of his day and therefore felt much more at home in the Oxford school that practised ‘ordinary language philosophy’ and therefore took (legal) discourse at face value (Postema 2011, p. 265). To be sure, there are many empirical questions we can ask with regard to law that require a different method of inquiry, but the analogy with chess opens our eyes to the fact that law is generally to be studied and analysed from within, on its own account, and that we should try and appreciate the sophistication and the often subtle precision of legal language as the historical product of dealing with specific social and political problems.
  Moreover, not only is this structure of a normative nature but it also claims authority in the sense that we are all supposed to observe the law in the sense of obeying it. The fact that the law claims authority in itself makes it important that we can distinguish legal norms from other norms in society. However, not only citizens who want to know what the law demands of them will want to be able to distinguish law from morality, for instance; but also legal research will start by delineating their subject of inquiry from other subjects. Before evaluating the subject under scrutiny they will first have to describe the law under scrutiny, in the sense that they will have to give a fair and unprejudiced – ‘detached’ – account of what their object of inquiry is, before criticizing (or praising) the law as they found it.

References

• Austin, J.L. (1957). A Plea for Excuses: The Presidential Address. Proceedings of the Aristotelian Society, New Series, 57, 1-30.

• Austin, J.L. (1962). How to Do Things with Words. Oxford, UK: Clarendon Press.

• Ayer, A.J. (1990). Language, Truth and Logic. London: Penguin.

• Carnap, R. (1992). On the Character of Philosophic Problems. In R. Rorty (Ed.), The Linguistic Turn: Essays in Philosophical Method. Chicago: The University of Chicago Press.

• Culler, J. (2000). Philosophy and Literature: The Fortunes of the Performative. Poetics Today, 21(3), 503-519.

• Dworkin, R. (1986). Law’s Empire. Cambridge Mass.: Harvard University Press.

• Dworkin, R. (2011). Justice for Hedgehogs. Cambridge Mass.: Harvard University Press.

• Grice, H.P., & P.F. Strawson (1956). In Defense of a Dogma. The Philosophical Review, 65(2), 141-158.

• Hart, H.L.A. (1958). Positivism and the Separation of Law and Morals. Harvard Law Review, 71(4), 593-629.

• Hart, H.L.A. (1983). Essays in Jurisprudence and Philosophy. Oxford: Clarendon.

• Hart, H.L.A. (2012). The Concept of Law. Oxford UK: Oxford University Press.

• Kahneman, D. (2011). Thinking Fast and Slow. London: Penguin.

• Kenny, A. (2007). Philosophy in the Modern World: A New History of Western Philosophy. Oxford UK: Clarendon Press.

• Lacey, N. (2004). A Life of H.L.A. Hart: The Nightmare and the Noble Dream. Oxford: Oxford University Press.

• Lacey, N. (2006). Analytical Jurisprudence Versus Descriptive Sociology Revisited. Texas Law Review, 84, 945-982.

• Lewis, D. (2002). Convention: A Philosophical Study. Oxford: Blackwell.

• Leiter, B., (1998). Realism, Hard Positivism, and Conceptual Analysis. Legal Theory, (4), 533-547.

• MacCormick, N., & O. Weinberger (1986). An Institutional Theory of Law: New Approaches to Legal Positivism. Dordrecht: D. Reidel Publishing Company.

• Magee, B. (1971). Modern British Philosophy. London: Secker & Warburg.

• Marmor, A. (1998). Legal Conventionalism. Legal Theory, (4), 509-531.

• Marmor, A. (2002). Exclusive Legal Positivism. In J. Coleman, & S.J. Shapiro (Eds.), The Oxford Handbook of Jurisprudence and Philosophy of Law (pp. 104-124). Oxford: Oxford University Press.

• Marmor, A., (2006a). How Law Is Like Chess. Legal Theory, 12(4), 347-371.

• Marmor, A., (2006b). Legal Positivism: Still Descriptive and Morally Neutral. Oxford Journal of Legal Studies, 26(4), 683-704.

• Postema, G.J. (2011). Legal Philosophy in the Twentieth Century: The Common Law World. Dordrecht: Springer.

• Quine, W.V.O. (1953). Two Dogma’s of Empiricism. In From a Logical Point of View (pp. 20-46). Cambridge Mass.: Harvard University Press.

• Rorty, R. (1992). The Linguistic Turn: Essays in Philosophical Method. Chicago & London: The University of Chicago Press.

• Rubin, E.L. (1988). The Practice and Discourse of Legal Scholarship. Michigan Law Review, 86(8), 1835-1905.

• Ruiter, D.W. (1998). Structuring Legal Institutions. Law and Philosophy, (17), 215-232.

• Searle, J.R. (1969). Speech Acts. Cambridge UK: Cambridge University Press.

• Searle, J.R. (1995). The Construction of Social Reality. New York: The Free Press.

• Searle, J.R. (1999). Mind, Language and Society, Philosophy in the Real World. London: Weidenfeld & Nicolson.

• Searle, J.R. (2010). Making the Social World, The Structure of Human Civilization. Oxford UK: Oxford University Press.

• Schwartz, S.P. (2012). A Brief History of Analytic Philosophy: From Russell to Rawls. Chichester: Wiley-Blackwell.

• Weissbourd, B., & Mertz, E. (1985). Rule-Centrism versus Legal Creativity: The Skewing of Legal Ideology through Language. Law and Society Review, 19(4), 623-659.

Noten

• 1 ‘[The purpose of the book] is not to provide a definition of law, in the sense of a rule by reference to which the correctness of the use of the word can be tested; it is to advance legal theory by providing an improved analysis of the distinctive structure of a municipal legal system and a better understanding of the resemblances and differences between law, coercion, and morality, as types of social phenomena. The set of elements identified […] are treated as the central elements in the concept of law and of prime importance in its elucidation’ (2012, p. 17; Cp. Postema 2011, 265).

• 2 ‘Hart’s reference to “descriptive sociology” was a gesture to the Austinian precept that an understanding of linguistic usage illuminates the world. […] But Hart’s own final statement of his position goes further yet. In one of his notebooks of the late 1980s, while accepting that he had been mistaken in making such a broad claim, he argued that a better formulation would have been that the book provided the ‘normative concepts required for a descriptive sociology’ (Lacey 2006, 949).

• 3 Oxford philosopher Peter Strawson: ‘… ordinary people employing the ordinary conceptual resources of mankind have at their disposal not a crude, rough-and-ready instrument, but an enormously sophisticated instrument, for thinking; and it’s an immensely and inexhaustibly interesting task to trace the various connections between the concepts which people handle in their ordinary life with no particular difficulty’ (in: Magee 1971, p. 136, 137; Cp. Lacey 2006).

• 4 The notion of a rule is, of course, not only important for lawyers and moralists, but in fact fundamental to language in general. Speaking a language is following rules. For this reason Wittgenstein famously devotes a substantial part of his Philosophical Investigations (1953) to all kinds of rule-following, and his analysis has proven to be of great interest for philosophers and linguists generally.

• 5 We should also note that in the normative domain, generally, norms need not be practised in order to be valid. Indeed, we know of many perfectly valid moral and legal rules that are often not followed. To express this categorical difference we distinguish facts from norms. Facts have an empirical reality and norms have validity. Somehow, somewhere, facts and norms are aspects of one reality, of course; Hart argues that the rule of recognition is where factuality and validity meet.

• 6 Hart introduces the rule of recognition thus: ‘The simplest form of the remedy for the uncertainty of the regime of primary rules is the introduction of what we shall call a “rule of recognition”. This will specify some feature or features possession of which by a suggested rule is taken as a conclusive affirmative indication that it is a rule of the group to be supported by the social pressure it exerts’ (2012, p. 94). ‘The rule of recognition is neither itself a rule of law nor a transcendental presupposition; rather it is a social rule accepted and practiced by judges and other legal officials’ (Postema 2011, p. 285, 286).

• 7 ‘One distinguishing mark of constitutive conventions […] is that [they] typically come in systems. There must be a system of conventional rules to establish a social practice, that is, some cluster of rules intertwined in a more or less complex structure. This is a feature that stems from the constitutive function of such conventions, from the fact that they constitute a practice’ (Marmor 1998, p. 522; Cp. Searle 1995, p. 28).

• 8 ‘Conventions evolve in response to a wide variety of social needs, and the need for coordination is just one of them. For example, conventions constitute various ways in which we express respect to one another, both linguistically and otherwise, conventions constitute artistic genres, games, and numerous social institutions; social conventions often regulate interpersonal relationships in such settings as a workplace, or a party, or an academic gathering, and such. In short, conventions serve numerous functions by constituting ways in which people interact with each other and engage in socially valuable activities’ (Marmor 2006a, p. 357).

• 9 ‘The rules constituting chess are constitutive conventions. They constitute what chess is and how to play the game. In part, they constitute the point of playing the game and some of the specific values associated with it. But all this is possible only against the background of a deeper layer of shared understandings about what competitive games are. In order to play chess as a competitive game of a certain kind, players must follow a fairly elaborate set of norms, or deep conventions, determining the concept of games and the essential point(s) of engaging in such a practice. Chess, as a game of a particular kind, is only an instantiation of a more general human activity that we call “playing a competitive game.” […] Deep conventions typically come to our attention when we face some deviant behavior or some doubts about the practice we are engaged in. […] The rules of recognition of the kind Hart had in mind are surface conventions. They determine what counts as law in a particular legal system or in a particular community. These surface conventions of recognition are instantiations of deep conventions about what law in our particular culture is. Roughly, the idea is that the rules of recognition relate to the deep conventions of law like particular games relate to deep conventions about what it is to play a competitive game’ (Marmor 2006a, pp. 364, 366, 369; Cp. Searle 1995 on ‘the Background’).

• 10 We should, however, note that in many continental ‘civil’ law systems the law is largely the product of legal science. This is the kind of science that can make its claims to what is law true when judges or the legislature make the results of legal scientific analysis and rationalization into law.

• 11 Quine (1953) famously argued that there is no such thing as purely analytic (‘grounded in meanings independently of matters of fact’) as opposed to purely synthetic truths that are grounded in fact. Grice and Strawson defended the distinction (1956) by pointing to the particular use that is made of it in philosophical circles. I assume here that Searle’s use of the term ‘analytic’ is well founded and consistent.

• 12 Searle (2010): ‘The distinctive feature of human social reality, the way in which it differs from other forms of animal reality known to me, is that humans have the capacity to impose functions on objects and people where the objects and the people cannot perform the functions solely in virtue of their physical structure. The performance of the function requires that there be a collectively recognized status that the person of object has, and it is only in virtue of that status that the person or object can perform the function in question. Examples are pretty much everywhere: a piece of private property, the president of the United States, a twenty dollar bill, and a professor in a university are all people or objects that are able to perform certain functions in virtue of the fact that they have a collectively recognized status that enables them to perform those functions in a way they could not do without the collective recognition of the status’ (p. 7).

• 13 Marmor argues that this explains the conventional nature of the rule of recognition and the rules it identifies as legal rules. The legal order is ‘radically underdetermined’ by the general purposes and values it serves, and the legal order can therefore take different shapes and colors. ‘First, it is crucial to realize that the needs, functions, or values that such conventions respond to radically underdetermine the content of the rules that constitute the relevant social practice. Our need or desire to play some competitive intellectual board game does not determine the content of the rules of chess; the creative needs or desires to stage dramatic performances do not determine the particular theatrical genres that have emerged over time in various cultures; the reasons for showing respect to people under various circumstances do not determine the practices of etiquette that we have’ (2001b, p. 360; Cp. Postema 2011, p. 524).

• 14 ‘For the understanding of this the methodology of the empirical sciences is useless; what is needed is a “hermeneutic” method which involves portraying rule-governed behavior as it appears to its participants, who see it as conforming or failing to conform to certain shared standards’ (Hart 1983, 13; Cp. MacCormick & Weinberger 1986, 15).