The legal positivism of HLA Hart

The Legal Positivism of HLA Hart

I     Hart’s Thesis

The theories of law with which this book is concerned attempt to explain what makes a legal claim or a legal proposition legally valid: they offer a description of how it is determined that the law requires X, or that it does not require Y. Those theories that aim at establishing the impartiality of legal determinations seek to exclude the use of preferences from that decision-making process: establishing what the law requires, they say, is a decision-making process which doesn’t take preferences into account. HLA Hart’s theory of law is such a theory. It is also a theory that is surprisingly commonsensical. According to Hart, the law is a collection of rules whose status as legal rules is a consequence of some official action (they are, for the most part, declared by a legislature or by courts); and the application of these legal rules involves a determination of whether a case falls inside or outside the conventional meaning of the words appearing in them. Section I of this chapter will elaborate and explain these claims, and will also examine how they seek to solve the problem of judicial impartiality. Sections II and III will criticise Hart’s claims, as well as his solution to the impartiality problem. Section IV will briefly survey the state of legal positivism after Hart.

Identifying the Authoritative Legal Rules

Hart’s theory of law, appearing in its most complete formulation in The Concept of Law, articulates roughly two stages for determining what the law requires (two stages of legal interpretation): the first involves the identification of legal rules, the second the identification of the requirements of these rules vis-à-vis a particular case.1 Together these two processes establish what the law says – or whether it says anything – on any given matter.

According to Hart, all mature legal systems have a test for the identification of legal rules, a test which all legal practitioners know and agree upon. Hart calls this test the ‘rule of recognition’. Only rules which satisfy that test are recognised as legal rules (rather than moral rules, or social rules, or what have you).2 The ‘rule of recognition’ specifies certain ‘sources of law’ where legal rules are to be found: these sources include the collection of statutes passed by parliament, judicial precedents, the decisions of administrative agencies, and other such authorities.3 If a rule is found in those specified sources, it is a legal rule; if it ain’t there – it ain’t legal. As there are various legal sources in a complex, modern legal system, these sources are ranked for relative supremacy within the ‘rule of recognition’ (for example, the ‘rule of recognition’ may declare that a legislated statute is superior to a judicial precedent): this guarantees that it will be possible to identify one authoritative legal rule in cases where rules taken from different sources conflict. The ‘rule of recognition’ is, therefore, a rather complex theoretical entity: it contains references to various legal sources, as well as a ranking of these legal sources in order of supremacy. It may take quite a few pages to put this ‘rule of recognition’ on paper. Yet, however complex or lengthy, the ‘rule of recognition’ must be shared by the legal practitioners of the legal system to which it belongs: they must agree as to what is the governing ‘rule of recognition’. The legal practitioners must share the test for the identification of legal rules.

According to Hart, it is the very fact that practitioners share the ‘rule of recognition’ that constitutes that rule: the ‘rule of recognition’, says Hart, exists by virtue of being accepted as a normative standard and followed. What does Hart mean by that? He means that the ‘rule of recognition’ exists as the test for valid legal rules because people accept it as a test and abide by it: it is, in other words, a conventional standard. Here is an analogy: the meaning of a word is what it is by virtue of its being accepted as such – that is, by virtue of the fact that we all agree that this is its meaning. The meaning of a word is also a conventional standard; it is constituted by conventional agreement. (Indeed it has been suggested that Hart’s idea that the ‘rule of recognition’ is a conventional standard has linguistic conventions as its very model.4) Thus, just as the correct meaning of words (or the correct test for using them) is supposedly constituted by conventional agreement among speakers, so is the ‘rule of recognition’ (or the correct test for the legal validity of rules) constituted by agreement among those who are practising law.

Now Hart concedes that the ‘rule of recognition’ is not explicitly articulated, and certainly not in its entirety: people are not in the habit of citing a ‘rule of recognition’ when they identify legal rules.5 Nevertheless, says Hart, we know the rule exists, and we know what it requires, because this is ‘shown in the way in which particular rules are identified, either by courts or other officials or private persons or their advisors’.6 One can detect the ‘rule of recognition’ by observing the way in which people (private individuals, lawyers, judges, state officials) identify legal rules: people look for legal rules in certain ‘sources of law’ (the pronouncements of the legislator, the decisions of courts, the promulgations of administrative agencies), and in doing so they follow the ‘rule of recognition’ – thereby making it known. What’s more, deducing the ‘rule of recognition’ from the actions of these people does not involve any complicated evaluative or controversial reasoning: according to Hart, this is a straightforward factual observation as to where people go when they look for legal rules.7

So the ‘rule of recognition’ – which can be identified by observing the actions of legal practitioners – points practitioners to the ‘sources of law’ where legal rules can be found. Sometimes the source will not present the practitioner with a ‘ready-to-use’ rule: Hart thinks that judicial decisions, for example, may lay down a precedent without explicitly articulating its governing rule.8 In such cases, the person seeking the legal rule must extract the rule from the precedent before her (she must grasp the rule laid down by the precedent). Yet, ultimately, all rules of law are linguistically formulated rules using general classifications: Hart says that ‘all rules involve recognising or classifying particular cases as instances of general terms’; rules of law ‘refer to classes of persons, and to classes of acts, thing, circumstance; and [the law’s] successful operation … depends on a widely diffused capacity to recognise particular acts, things, and circumstances as instances of the general classification which the law makes’.9 In short, the law consists of rules, found in certain recognised sources identified by the ‘rule of recognition’; and these rules divide the world into classes of ‘acts, things, or circumstances’ using general words (for instance, the class of ‘assaults’, the class of ‘contractual agreements’, or the class of being ‘unemployed’), and then prescribe the treatment that is to be accorded each of these classes.

Applying Legal Rules

Once the legal rule is identified, establishing what the law requires on any given case is a rather straightforward exercise in linguistic classification: according to Hart, the legal interpreter applies the law by examining whether the case (or the hypothetical instance) falls inside or outside the conventional meaning of the classificatory term of the relevant legal rule. For example, if a rule reads ‘no vehicles in the park’, then whether the rule is or is not violated by a certain conduct depends on whether that conduct involves a ‘vehicle’ and a ‘park’ according to the conventional meanings of these terms.

Now general terms have cases of clear and certain applications – ‘General terms would be useless to us as a medium of communication unless there were such familiar, generally unchallenged cases’10 – as well as cases of uncertain applications: take any general term, says Hart, and you will find that it has some applications regarding which it is unclear whether it applies or not. The term ‘silk’, for example, has unmistakable applicability to a certain substance secreted by silkworms; but imagine that we discover that the substance somewhat changes its chemical constitution when exposed to certain pollutants. This would present us with an uncertain case of applicability: the altered material is a borderline case falling in the ‘penumbra’, or the ‘open texture’, of the term ‘silk’. (‘Open-textured’ and ‘penumbral’ were the adjectives Hart used to denote uncertain cases of application.11) What, for Hart, distinguishes core applications from borderline applications? Core applications are those characterised by agreement: the ‘plain case’ is one where ‘there is general agreement in judgment as to the applicability of the classifying term’ – whereas a borderline case of applicability occurs where ‘no firm convention or general agreement dictates its use’.12 In this latter case, the determination of applicability is not resolved by the meaning of the general term: the question whether the altered material is ‘silk’ or not ‘silk’ cannot be answered by consulting the meaning of that term, because meaning is presumably a function of convention, and there is no convention governing the matter. We may summon various arguments in deciding whether the altered material falls inside or outside the term ‘silk’ – and some arguments would be better, some worse; but no position would be ‘correct’ or ‘incorrect’ as far as the meaning of our term is concerned. Hart says that ‘if in such cases doubts are to be resolved, something in the nature of a choice between open alternatives must be made by whoever is to resolve them’.13 (‘Open alternatives’ here signifies that neither option is uniquely correct.) So once we encounter a borderline case, we must make a choice regarding the applicability of the term: the meaning of the term does not settle the issue. But in ‘core’ cases there is no such choice: applicability is determined by linguistic conventions which leave no room for any ‘open alternatives’.

This means, for Hart, that legal rules follow a similar pattern – since their applicability is a function of the applicability of the general terms appearing in them. Hart says that ‘in case of everything which we are prepared to call a rule, it is possible to distinguish clear central cases, where it certainly applies, and others where there are reasons for both asserting and denying that it applies. Nothing can eliminate this duality of a core of certainty and a penumbra of doubt when we are engaged in bringing particular situations under general rules’.14 There are core cases for the application of legal rules – for example, those involving core instances of ‘vehicles’ and ‘parks’; and there are cases where the rule’s application is unclear – for example, when it is unclear whether the relevant object is a ‘vehicle’ (say, the object is a motorised wheelchair). When such cases occur in the application of legal rules, we shall have to go beyond what the law says: we shall have to appeal to considerations external to the law in deciding the applicability of the rule, and no answer will be correct or incorrect as far as the law or existing legal requirements are concerned.

It is, of course, an uninteresting truism that there are cases where the law is clear (‘core’ cases) and cases where it is unclear (‘penumbral’ cases); but Hart is offering an explanation of what makes legal cases clear or unclear. According to Hart, the core and penumbra of the law track the core and penumbra of the general classificatory terms appearing in legal rules. In other words, whether a case is a clear case for the law is a function of whether it is a clear case of applicability for the general terms appearing in the relevant legal rule.15 Where the applicability of the general terms is uncertain, the law is uncertain. Indeed when judges decide the applicability of a rule in a borderline case they ‘have settled a question as to the meaning, for the purposes of this rule, of a general word’.16

So legal rules have core cases of applicability when a case is a core case for the application of their general terms; and they have borderline cases of applicability when a case is a borderline case for the applicability of their general terms. Since what the law requires is simply a function of what the conventional meanings of the general classificatory terms require, when the applicability of these terms is unclear, the legal interpreter must go ‘outside the law’ in order to determine the applicability of the relevant legal rule. In such cases the law has no clear answer on the matter, and it is no help to keep staring at the law. If there is a decision to be made, the legal interpreter will have to go somewhere else in order to make it. At this point, if the case is in court, the judge must move from the task of legal interpretation (from establishing what the law is) to the task of judicial legislation (to establishing what the law should be): from the task of identifying the law to the task of writing it.

Nevertheless, according to Hart such cases are a minority existing only on the ‘margins of rules’:17 given the communicative function of linguistic terms, the questionable cases – those regarding which it is indeterminate whether they fall within or without a term – are few. Hart says that ‘failure to do justice to [the phenomenon of open texture] will always provoke exaggeration’;18 in reality ‘the life of the law consists to a very large extent in the guidance both of officials and private individuals by determinate rules which … do not require from them a fresh judgment from case to case’.19 Just as our everyday use of language encounters few questionable borderline cases, so does the law – whose applicability is a function of the applicability of our natural language – mostly handle clear and unambiguous cases.

Hart’s Theory and Impartiality

It is easy to see how Hart’s theory establishes the impartiality of proper legal interpretation. To begin with, legal rules are identified by the ‘rule of recognition’, and everybody agrees what that rule is: it is a conventional standard. People identify legal rules by following the same procedure that everybody else follows. Identifying legal rules is therefore a decision-making process having no place for considerations of preference. Preferences are by definition controversial; conventions are by definition a matter of consensus. If identifying legal rules is a matter of following a conventionally accepted standard, then there is no danger of partiality in the process of identifying legal rules. Moreover, in most cases there is no danger of partiality in the application of these rules either: their applications are, too, grounded in conventions – in linguistic conventions regarding the correct use of general terms. Whether a case falls within or without a general classificatory term is, for the most part, a determination with no place for considerations of preference.

Admittedly, there are cases where no convention governs the applicability or inapplicability of a general term (no convention on the use of the term ‘vehicle’ establishes whether a motorised wheelchair is or is not a ‘vehicle’). When such instances occur, no convention establishes the correct legal determination, and the legal interpreter must resort to arguments about the pros and cons of finding the term, and thereby the rule, applicable. Here, partiality may rear its head: preferences may certainly influence some of these determinations. Yet, as we saw, these cases are a small minority – they are the exception, rather than the rule.20 Moreover, these are not cases of ascertaining what the law requires; these are cases of ascertaining what the law should require – meaning that preferences may be used only in cases where the law does not resolve the issue in question. There are gaps in the law. Cases may arise that nobody has contemplated; the legislature may do a bad job or neglect its duties, or may even intentionally leave some areas unregulated: in such cases legal interpreters must go beyond the law. But what else can they do? Legal cases must be resolved, and refusing to hear a case merely amounts to handing the victory to defendants. Nevertheless, to repeat, legal interpretation (establishing what the law says – not what it should say) is an activity that is necessarily impartial; and the majority of cases involve legal interpretation. Determining legal rights and duties is, for the most part, a straightforward business revolving around the use of readily available conventional standards, so that in the majority of cases partiality is not even a possibility for the conscientious legal interpreter.

II     Hart and the Problem of ‘Essentially Contested Concepts’

There are two principal problems with Hart’s picture of legal interpretation. The first concerns what have come to be known as ‘essentially contested concepts’. (My discussion will use the words ‘term’ and ‘concept’ interchangeably.) Many linguistic terms do not sit well with Hart’s analysis of the core and the penumbra: they do not have large cores of certain applications upon which everybody agrees. Instead, their applications are very commonly disputed, and, as far as the speaker is concerned, correct applicability may have little to do with whether everybody else does or does not agree. This brings into question the sense in which the application of rules is a function of conventional agreement, and it threatens to sap much explanatory power out of Hart’s theory. The second problem, which I will address in section III, concerns a different but more fundamental difficulty: viz, practitioners’ habitual disregard for the ‘boundaries’ set by the linguistic meaning of the terms appearing in legal rules.

On Essentially Contested Concepts and Their Containment

Hart’s understanding of linguistic terms having large cores of conventional agreement and marginal penumbras of disagreement does not sit well with a whole range of rather common terms.21 They have been named, among other things, ‘essentially contested concepts’ – concepts like ‘justice’, ‘democracy’, ‘art’, or ‘good faith’, regarding which speakers have different, and often competing, understandings.22 Essentially contested concepts do have applications upon which everybody agrees, but these applications are not as common: where essentially contested concepts are concerned, there is no good reason to think – as Hart does – that the penumbral cases (those cases where people disagree about applications) are a marginal minority. It would appear, instead, that each time a contested concept or expression is involved, disagreement is the rule rather than the exception. Moreover, essentially contested concepts (such as ‘reasonableness’, ‘recklessness’, ‘excessive force’, ‘unfair dealing’, ‘coercion’, and ‘excessive burden’) figure as classificatory general terms in innumerable legal rules, and those exceptional cases that do generate consensus are not likely to arise in the judicial context. In short, Hart’s claim that the ‘life of the law’ consists of clear and unambiguous cases, and that ‘judicial legislation’ is the exception rather than the rule, is simply wrong as far as essentially contested concepts are concerned; and there are an awful lot of essentially contested concepts in the law.23

Hart is not oblivious to this obvious point; he explicitly acknowledges the problem when discussing what he calls ‘very general’ classificatory terms whose applicability is usually controversial: ‘Of course, even with very general standards there will be plain indisputable examples of what does, or does not, satisfy them. Some extreme cases of what is, or is not, a “fair rate” … will always be identifiable ab initio … But these are only the extremes of a range of different factors and are not likely to be met in practice.’24 This, says Hart, ‘entails a relative indeterminacy … and a need for further official choice’.25 However, Hart thinks that this is not a serious problem for his theory, because he thinks that such legal standards are purged – rendered determinate – by subsequent court decisions (the further ‘official choice’ he mentions): where a standard like ‘reasonableness’ appears in a legal rule, individuals ‘are required to conform to a variable standard before it has been officially defined, and they may learn from a court only ex post facto when they have violated what, in terms of specific actions or forbearances, is the standard required of them. Where the decisions of the court on such matters are regarded as precedents, this specification of the variable standard is very like the exercise of delegated rule-making power by an administrative body …’.26 In other words, such legal standards are a mere delegation of authority for the courts to lay down the real legal rule containing the real legal standard.

Putting aside the question of systems where court decisions are not regarded as precedents, Hart’s claim remains highly questionable. One way of reading his argument is that courts replace vague and indeterminate standards (or ‘essentially contested’ terms) with other more determinate ones – that is, with standards with substantial cores of conventional agreement. Although there can be little doubt that courts applying standards like ‘reasonableness’, ‘due care’ or ‘unfair competition’ give these more determinate contents – indeed this they must do if only by recounting the facts of the case – it is doubtful that courts replace