Skip to navigation – Site map

The Hierarchical Model and H. L. A. Hart's Concept of Law

Massimo La Torre
p. 141–161
Le modèle hiérarchique et le Concept de droit de Hart


Law is traditionally related to the practice of command and hierarchy. It seems that a legal rule should immediately establish a relation between a superior and an inferior. This hierarchical and authoritharian view might however be challenged once the phenomenology of the rule is considered from the internal point of view, that is, from the stance of those that can be said to “use” rather than to “suffer” the rules themselves. A practice oriented approach could in this way open up a more liberal, and also somehow less parochial and ideological, road for legal theory. This is – it is argued in the paper – the programme, or better, the promise we can find in Herbert Hart’s main work, The Concept of Law. The article tries to render this promise more transparent while, nonetheless, not eschewing the blind sides of its narrative and argumentative strategy.

Top of page

Full text

1To conceive the law as a hierarchical model means first of all putting the fact of command at the heart of legal experience, that is of a power of an inferior over a superior. There is, moreover, a second possibility: that which views the law as the outcome of a social fracture between a class empowered with the authority of making or saying the law and a class whose destiny is much more humble and is that of only obeying or being the addressees of the law thus created. Finally, a further alternative is to conceive of law as a structure of layers, of entities, of rules usually, hierarchically disposed, so that one layer draws the reason of its existence or legitimacy or validity from a layer superior to the one considered. These three possibilities are not always present in one single legal theory, though they are not mutually incompatible.

2In this paper my task will be to tackle the issue of the hierarchical model in the law from the perspective of H. L. A. Hart's legal philosophy. Herbert Hart – one should not forget it – has become and still is one of the most influential jurisprudents of our times and his views imbue several of the current doctrines of legal practice and legal reasoning. Mine will thus be a presentation and discussion focussing on Hart's concept of law and its relationship to the hierarchical model. It is also intended to be a more general assessment on the fruitfulness of such an approach with special attention to international law.

3Now, before dealing specifically with our subject, two points are important to bear in mind, in order to understand the sense and relevance of the discussion around the concept of law which is the specific function of jurisprudence and legal philosophy. First of all, a discussion about the concept of law is not just a question of theory, or a mere philosophical disquisition without practical effects. The law is not an empirical object, a physical or natural thing. It is rather a human practice ruled and made possible through norms, principles, values, attitudes, ideas. In short, the law is what we believe is the law and what we practice as the law.

4This means that the law is its concept or, rather, that the law is the concept of law which we adopt and follow in our conduct. Therefore, what we accept as the concept of law has a strong, even a dramatic impact on the "hard" law, the law in action, the practice of law. Since the law is strictly, logically, connected with our society's concept of law, the law itself is a matter of discussion and controversy. Law is an essentially contested concept. But the discussion and the controversy about the law and its object are just the domain of jurisprudence and legal philosophy. The adoption of one or the other concept of law has immediate repercussions on the facts of law.

5The other point I would like to preliminarily stress is the following. Law is an ambiguous concept and an ambiguous practice. It is imbued with ambiguity. It is on the one side related to violence, force, authority, power, effectiveness, facticity, efficacy, hierarchy. On the other side, however, it is strongly connected with justice and morality, that is with autonomy in some sense. Sometimes we meet, sometimes we face law in the form of a penalty, a prison, a scaffold, electric chair: law here is very much equivalent to coercion and violence.

6But law is not just coercion and violence. There is, internal to law, a claim to correctness and justice. Law is there not just in its own right and on its own behalf; we have law not on law's own behalf, but – we might plausibly say – on behalf of a fair scheme of social and human relations. There is – so to say – an utopian element in the law, an ideal state of social practices, which is its sense and its justification. Law therefore is not just force, but it is rather a legitimate force. It is founded on legitimacy and nonetheless needs efficacy.

7Now, the history of the concept and the practice of law is the history of the ways, the strategies undertaken to deal with this ambiguity and tension, the history also of their solutions. For instance, if you think of the image ordinary people have of lawyers you will find yourself once more confronted with this ambiguity, with such duplicity. On the one side, a lawyer is seen as a pettyfogger, a rascal obsessed with emptying his clients' pockets, someone whose only real job is obscuring the laws' interpretation, distorting them, for his own personal private interests or in the best case for the private, particular interest of his client. This is the kind of lawyer which inspires Shakespeare's cry "Let's kill all the lawyers" in Henry the Sixth, Part Two. To illustrate this view, please, allow me to read an excerpt from Mary Wollstonecraft book A Short Residence in Sweden, Norway, and Denmark:

My head turned round, my heart grew sick, as I regarded visages deformed by vice; and listened to accounts of chicanery that were continually embroiling the ignorant. These locusts will probably diminish, as the people become more enlightened /.../ The profession of law renders a set of men still shrewder and more selfish than the rest; and it is these men, whose wits have been sharpened by knavery, who here undermine morality, confonding right and wrong.

8But the lawyer, the advocate, is not only Mary Wollstonecraft's pettyfogger "deformed by vice"; he is also the defendant of the unjustly indicted, the champion of human rights. Absolute kings, dictators and totalitarian States deeply dislike lawyers and advocates, often considered by them to be subversive agents. Others, on the contrary, see lawyers as a specially ethical role. Let us listen to what one great legal philosopher, the American Lon L. Fuller, thinks of the legal profession:

  • 1 Winston (2001: 275–276)

The best definition I ever heard of a lawyer was that given by the young daughter of a friend of mine. A neighbor's child asked what her father did. She said:
– He's a lawyer.
– What's a lawyer?
– A lawyer is a man that helps people.1

9According to Fuller, a lawyers' activity is not fully instrumental; it is not necessarily or conceptually oriented to success and the client's interests. It is rather functional to the good administration of justice.

  • 2 Winston (2001: 313)

In arguing cases before courts, he [the lawyer] will see his job, not as one of mere persuasion, or of a facile manipulation of legal doctrine, but as one of conveying to the court that full understanding of the case which will enable it to reach a wise and informed decision.2

10The two opposed views mentioned of what a lawyer properly is are based on opposed views about the concept of law. For one view, the law is something mainly addressed to the "bad man"; for the other view its main figure is in Hart's words "the puzzled man", a human being who asks "what ought I to do?", and looks for criteria of right conduct.

11Another field where the mentioned ambiguity and the tension intrinsic to law is fully and dramatically deployed is the international relations arena, that is international law and legality. Is there an international law, which will consequently be binding on its subjects, States, and their municipal law? For instance, John Bolton, Sub-secretary of State under the G. W. Bush administration denies that there is such a thing, while the Secretary General of United Nations, Kofi Annan, seems to defend the opposite stance. But the stance here respectively adopted is a matter of jurisprudence and of what idea of law we hold.

12Please, note that the same problem that we have with international law may also be encountered with constitutional law. That is: "Is constitutional law binding on government and officials?"; which is a thorny question, especially once we conceive of the law as a hierarchical chain of command. Once again our answer to such question will depend on our concept of law and on our jurisprudential position.

13Now, let us go back to Hart, whose doctrine today is my main subject of interest. In order to better understand Hart's theoretical enterprise, this should be seen – I believe – as a reaction or a response to four main doctrines of the concept of law. These are: (i) imperativism; (ii) realism; (iii) formalism (very well embodied in Hans Kelsen's so-called "pure" theory of law); (iv) natural law.

14(i) Imperativism is the view according to which the law is a set of commands backed, supported by the threat of a sanction, an evil, and issued by a political superior holding full sovereignty. You might remember that this is the doctrine defended by Jeremy Bentham and by John Austin, and by the whole tradition of the British – so called – analytical jurisprudence. In such a view, the practice of law is shaped in radically hierarchical terms, though not as a hierarchy of rules. In fact, imperativism has many problems in admitting the category of a rule as something distinct from the concrete prescription by superiors and from habit by inferiors.

15(ii) Realism is a complex compound of diverse doctrines. There are at least – as is well known – both an American and a Scandinavian Realism, the two holding different views. However, Hart intends to respond to the challenge mainly coming from the American realists, although his discussion of legal obligation and the interpretation of this in terms not of feeling compulsed, but just as the application of a rule is probably directed against Hägerstom's and Olivecrona's writings.

16As is well known, the fundamental thesis defended by the American brand of realism is the following. The law is nothing but a prediction about the future conduct of judges. Law is not equivalent to rules, but rather to judicial decisions, or – in another formulation – law is the the "law jobs". Law is not law in books, but law in action. According to this view, legal practice revolves around the figure of a judge, and, in a sense, turns out to be not so much an intrinsically hierarchical experience, since the judiciary is here seen as a diffuse and somehow informal social structure. In no case here, however, could we speak of law as hierarchy of rules in logical terms, since life, action, not logic or formalities run the practice of law. There might be a hierarchy of rules in the books; in front of the judge, however, such logicalities are of very little help.

17(iii) Legal formalism insists that law is a set of rules. Rules here are seen as semantical entities, propositions, Urteile – as Hans Kelsen labels them in German, following a Kantian terminology. According to Kelsen – as is well known – a rule has the form of a hypothetical judgment, "if A, then B", where A is the contrary conduct to the one promoted and desired, and B is the prescription of a sanction. According to Kelsen, moreover, rules are structured following a strict hierarchical order. The law for him is a sort of pyramid made up of different layers, each at a different level, where the rules at the lower level derive their validity from the rules at the higher level. At the top of the pyramid there is a special rule, the Grundnorm in German, the ground rule, which is the master rule giving validity to the whole system. We are all well acquainted with this doctrine.

18Kelsen's main problem is avoiding the fallacy of infinite regress which is a frequent risk in legal formalism. Legal formalism, a variant of legal positivism, conceives of two fundamental sorts of rules: (a) one kind ordering, forbidding, or permitting a conduct; and (b) another kind prescribing a sanction for the case that the first rule is not complied with. In such an approach, however, the legal character of the rules derives from the prescription of a sanction. But if this is the case, a rule ordering a sanction needs to be legal to be backed through a further rule prescribing a sanction, and this one again will be legal only if there is an additional rule backing the former through a sanction; and thus ad infinitum. In order to avoid this regressus ad infinitum, Kelsen conceives the rule as made up of two parts, one hypothetically assuming a state of afairs, and the other part which is the prescription of the sanction. Here there are no longer two rules: one prescribing a conduct, and another ordering a sanction, but there is only one rule, which prescrives a sanction for the case of a given state of affairs taking place. For Kelsen then, legal validity is not the outcome of a rule prescribing a sanction but of a rule given at a higher hierarchical level in the legal pyramidal structure.

19However, here again there is the risk of a infinite regress, since we are asked to seek an ever higher rule, a rule superior to the one which we consider legally valid. This is the reason why for Kelsen the Grundnorm, the master rule, is not a real positive rule. The Grundnorm is rather an epistemological assumption, a logical trascendental assumption in Kantian terms, whose task is to make it possible for lawyers to conceive positive rules as legally valid. The Grundnorm closes up the legal system at its summit, by rendering the search for a further higher rule pointless. All this is well known, and it is the typical model which we think of when we articulate the idea of law as a hierarchy of rules.

20(iv) Finally, we have natural law, as a fourth dominant jurisprudential approach. Natural law has a very long history, and it has been "cooked" – so to say – according to very different recipes. However, we can find a lowest common denominator and this is the following. According to natural law theories, the legal validity of a rule is mainly connected with its substantive contents and its morality, and such a morality can be ascertained through a mere cognitive exercise. Said in different terms, for natural law the "ought" of law can and should be inferred from the "is" of a moral world or substance or principle accessible to human knowledge. This means that, for natural law, law is (a) either coincident or equivalent with morality or (b) is strongly conceptually connected with the former. Law and morality are not distinct domains. They are intrinsically connected. Such an approach is usually not so much interested in the question of what law "is", but rather in what law "ought" to be. Or, said differently, natural law is above all a theory of the justification of law, being usually much less interested in offering an ontology of law. Its obsession is that the law comply with specific substantive moral requirements; what the structure or the institution of law is, or how this is articulated, is to natural lawyers a matter of expediency and, therefore, is open to multiple and flexible alternatives. However, traditionally natural law is related to imperativism, in the sense that while natural law gives the superior substantial criteria for the content of rules, these same rules are very often thought of as commands coming from a soveriegn obliged to follow the prescriptions of natural law. We can thus plausibly conclude that natural law usually defends a hierarchical view of the law, both in the sense that the fundamental piece of legal practice is command, and in the sense that law is hierarchically ordered so that positive law lies in an inferior position and is subject to natural law.

  • 3 Hart (1994: 26).
  • 4 Hart (1994: 26).
  • 5 Hart (1994: 26).
  • 6 Hart (1994: 42).
  • 7 Hart (1994: 26).

21Now, Herbert Hart is critical towards all the four mentioned approaches. (i) Against imperativism, to John Austin's analytical jurisprudence, that is the theory that views law as a set of orders backed by the threat of a sanction and issued by a political superior, he raises four main objections. (a) “Surely – he remarks first – not all laws order people to do or not to do things”.3 For instance, there are laws, such as the laws of making wills or on contracts or on marriages, which only offer facilities or confer powers to private individuals. (b) “Surely – and this is his second objection – not all laws are enacted nor are they all the expression of someone's desire” like general orders.4 For instance, custom, although a subordinate source of law in most modern legal systems, cannot be conceptualised according to the model of command. A command is personal, custom is impersonal; command is an intentional act, custom is not. A command is a deliberate, intended action; a custom is not a deliberate act: it is rather unintended. (c) “Surely – third objection – laws, even when they are statutes deliberately made, need not be orders given only to others”.5 For instance, “legislation /.../ may perfectly well have /.../ a self-binding force. There is nothing essentially other-regarding about it”.6 There is a strong analogy between a promise – by which we bind ourselves to a particular conduct – and a piece of legislation – which applies generally to all citizens (legislators included). (d) “Finally – fourth objection –, must enacted laws to be laws really express any legislator's actual desires, intentions, or wishes?” Since statutes, for instance, are enacted by complex collective bodies, should we ascribe to such bodies a capacity for emotions, desires, intentions? “Would an enactment duly passed not be law if /.../ those who voted for it did not know what it meant?”7

22So far I have said that Hart's concept of law can be better understood as a reaction to four jurisprudential models: (i) imperativism; (ii) legal formalism; (iii) legal realism; and (iv) natural law. I have already explained what these four models amount to and their relation to a possible hierarchical model of legal practice. And I have also presented Hart's line of attack against imperativism developed along four main objections.

23Now, I shall go on by illustrating Hart's response to the other three models, that is to legal formalism, to legal realism, and to natural law. In order to do so, I would like to introduce what are, in my view, the fundamental tenets of Hart's legal philosophy. These in my opinion are the following six: (i) normativism, that is, the centrality of the notion of norm or rule in the law; (ii) the centrality of an internal point of view with respect to rules and legal practice; (iii) a distinction between primary and secondary rules, and the conceptual and practical priority of the latter over the former; (iv) the so-called social thesis and the rule of recognition; (v) a moderate rule-skepticism as far as judicial reasoning is concerned; (vi) a non-cognitivistic metaethics.

24(i) Allow me to explain these items one by one. Let us start with normativism. To understand this we should refer to an example given by Hart in chapter four of his Concept of Law. Let's try – he proposes – to reconstruct a society conceptually along the model proposed by imperativism. Let us imagine a community where there is a subject issuing orders to the population, and where the population has the habit of obeying his orders. He himself does not have any similar habit of obedience to any other subject or person. Hart calls this subject issuing orders and receiving habitual obedience, Rex.

25Now, one sad day Rex dies and his place is taken by Rex the Second. Rex II starts, like his predecessor, issuing orders and claiming that these are the law of that community. However, since Rex II's government is too recent, we cannot say that there is already in that society a habit of obedience to his orders. And since such habit is a condition for having law according to the imperativistic model, we should conclude that in Rex II's reign, at least in its beginnings, there is no law. Such conclusion however is hardly tenable, if we want an account of law as a permanent situation.

26Should we give up the claim of law as a more or less permament state, we would deprive law of one of its main functions, which is that of ordering and stabilizing a society. Hence, the simple model of a society directed by Rex, the imperativistic model, which is in a sense highly hierarchical, is doomed to failure. To explain stability and permanence in law we should introduce a new concept: that of rule or norm. We thus need to found the stability of the legal system not so much on habits as rather on rules. But: What is the difference between a habit on the one side and a rule on the other? There are at least three distinguishing discriminating features.

  • 8 Hart (1994: 55). Italics are the author's.

27(1) “For the group to have a habit it is enough that their behaviour in fact converges. Deviation from the regular course need not be a matter for any form of criticism”.8 In the case of rules, deviation on the contrary is a matter of criticism.

  • 9 Hart (1994: 55). Italics are in the text.

28(2) “Where there are /.../ rules, not only is /.../ criticism in fact made but deviation from the standard is generally accepted as a good reason for making it”.9 Rules are good arguments in a discourse of justification; habits are good arguments only in a discourse of (causal) explanation. Rules are reasons for action; habits are causes or motives of a conduct. While a reason offers a reflective grounds for a course of action; and a cause or a motive does not.

  • 10 Hart (1994: 56).

(3) The third feature distinguishing social rules from habits /.../ is a feature which /.../ we shall call the internal aspect of rules. When a habit is general in a social group, this generality is merely a fact about the observable behaviour of most of the group. In order that there should be such a habit no members of the group need in any way think of the general behaviour, or even know that the behaviour in question is general; still less need they strive to teach or intend to maintain it. It is enough that each for his part behaves in the way that others in fact do.10

29For a rule to exist, on the contrary, there must be some people in the group at least who intend to follow it and and that consider the content of the rule as a general standard of conduct.

  • 11 Hart (1994: 57).

30Rules have an internal aspect which we do not find in habits. This internal aspect however should not be explained simply in terms of feelings, especially of feelings of compulsion (as it is done, for instance, by the Scandinavian realist school). “There is no contradiction in saying that people accept certain rules but experience no such feelings of compulsion”.11

  • 12 Hart (1994: 57).

What is necessary is that there should be a critical reflective attitude to certain patterns of behaviour as a common standard, and that this should dispaly itself in criticism (including self-criticism), demands for conformity, and in acknowledgments that such criticism and demands are justified.12

31The notion of rules or norms, normativism thus, according to Hart serves as a conceptual alternative both to commands (imperativism) and habits (realism). Rules are alternative to commands, in order to explain the generality, the impersonality and the self-binding character of laws and legal instruments. Rules, moreover, are alternative to habits, in order to explain the reason-giving, the reflective character of laws and their internal character.

32(ii) After "normativism" the second fundamental tenet of Hart's approach is the so-called internal point of view. This has to do with the internal aspect of rules which I have just mentioned. To approach the law, to conceive a conduct as related to law as legal, we should adopt an internal point of view. This means that we should first, and above all, consider the internal aspect of a rule. A rule is the sum of, on the one side, a regularity of conducts, a convergency of behaviour and, on the other, the idea that that particular piece of conduct is obligatory, that it constitutes a standard of behaviour. The difference between a rule and a mere habit is just that a mere habit is given only through a convergency of behaviour and nothing else.

33We can, for instance, ascertain that many people when driving a car on a hot summer day tend to use air conditioning. This, however, is a regularity, a mere habit, but not a rule. If by driving in a hot climate I do not use a car with air conditioning, I might be considered perhaps a trifle eccentric, or – who knows – stingy. But nobody would dare to accuse me of doing the wrong thing, of acting wrongly. I might in such cases be criticised because of not being prudent, or because of not acting or living in a comfortable way, but not because of being unjust or not right, or of breaking a rule. As a matter of fact, it is the possibility of the latter criticism, the criticism of being wrong, that signals the presence of a (legal) rule.

34The legal point of view therefore cannot just refer to external regularities, as it is for instance recommended by legal realists in interpreting the law as the prediction of conducts, or to a mere fact unaccessible to rules as a sheer act of decision. The legal point of view needs focussing on the internal aspect, needs ascertaining the existence of critical attitudes towards forms of behaviour in order to be able to identify a piece of legal, normative conduct. This approach is called by Hart the "internal point of view" and through such an approach Hart attacks and defeats the realist school of jurisprudence in all its variants.

35(iii) After normativism and the internal point of view, the third fundamental tenet of Hart's jurisprudence is the distinction between primary and secondary rules. We have seen when dealing with Hart's criticism of imperativism, that he points out and stresses the multiplicity of kinds of legal rules. Whereas imperativism follows a reductionist strategy, trying to compress the great variety of laws in one single form: a command or order backed by a sanction and supported by a habit of obedience, Hart's approach is open to diversity within the legal domain.

  • 13 Hart (1994: 80).
  • 14 Hart (1994: 80).

36We have seen that he points out, for instance, the existence and the relevance of rules which do not order anything but rather make it possible for citizens to enjoy certain facilities or exercise given powers and competences. In particular, Hart singles out two fundamental kinds of rules: (a) rules which prescribe a given conduct; (b) rules which ascribe power (both to officials and to private individuals). According to Hart, any attempt to see the rules which ascribe power as parts, fragments, of the rules which prescribe a conduct ends by offering a distorted picture of law and of its functions. One theory which proposes to consider rules ascribing powers as fragments of rules prescribing conducts (imposing duties) is the one defended by Hans Kelsen – the main representative of legal formalism. “Devices – writes Hart –, such as that of treating power-conferring rules as mere fragments of rules imposing duties, or treating all rules as directed to officials, distort the ways in which these are spoken of, thought of, and actually used in social life”.13 “This – adds Hart – had no better claim to our assent than the the theory that all the rules of a game are "really" directions to the empire and the scorer”.14

37Hart calls the rules imposing duties "primary rules", and the rules ascribing power "secondary rules". Only very simple and primitive legal systems can only deal with primary rules. Once we introduce devices to solve disputes or to introduce new rules into the legal system, we can have such devices only if we assume the validity of rules which ascribe law-giving and adjudicating power. Hence, Hart believes that there are three fundamental types of secondary rules, that is, of rules ascribing powers. these are: (i) rules of adjudication; (ii) rules of change; (iii) rules of recognition. (i) Rules of adjudication give judges the power of applying the law and of solving disputes. (ii) Rules of change give citizens and officials the power of modifying the prior state of the law by introducing new rules, new laws. They ascribe, therefore, legislative powers. (iii) A special case is that of the rule of recognition. This plays a role comparable to Hans Kelsen's master rule, the Grundnorm, which he put at the top of the pyramid representing the legal system.

38Through the rule of recognition we recognise which rules belong to the given legal order and accordingly we give such rules full legal validity. In this sense Hart's rule of recognition plays the same role which Kelsen attributes to his Grundnorm. There are however two major differences with respect to Kelsen's legal formalism: (a) The rule of recognition does not necessarily place the recognized rules in hierarchical order. For sure the rule of recognition is epistemologically prior to the subsequent rules recognised through it: these rules on their part nonetheless should not be deployed according to a hierarchical order so that the one derives its validity from the other. It is possible that one kind of rule, for instance statutes, legislation, could derogate other kinds of rules, for instance customary or common law. However, this does not imply that common law draws it legal character and validity from legislation according to Hart. Now, common law being in the foundational basis for judicial rulings the Anglo-Saxon systems, one might infer from such a remark that judicial power and rulings need not in Hart's view of the legal practice be subject to the general rule posed by statutes and legislation.

39(b) Another difference with respect to Kelsen's "pure" theory is that the rule of recognition is not a fiction, or a mere presupposition. It is, in contrast, a real, positive, effective rule. The latter point also marks the distance from some sort of realism: in fact the rule of recognition by Hart is seen as "law", and not just as "source of law" which can only be precipitated into "real" law through (judicial) interpretation.

40(iv) Now, we land into the fourth fundamental tenet of Hart's theory. This is the so-called Social Thesis. The rule of recognition according to him is a social practice, a social fact – it is not just a formal rule (as it is the case in Kelsen's doctrine). The validity of all law and other sources of law is therefore based on this practice, on such social facts. This is also the main argument for defending legal positivism on the part of Hart and his later followers, for instance those such as Professor Joseph Raz. By the social thesis, law is not considered an ideal phenomenon (as it is the case in natural law doctrines), but it is eminently seen as an empirical, positive domain. In particular, from this perspective laws, legal rules, draw their validity from an empirical social basic source.

41(v) Intertwined with the social thesis is the idea that the law is not conceptually connected with morality. As is well known, Hart defends a strong non-cognitivistic metaethics. Metaethics – there should not be need to recall it – is the study of the logical and epistemological status of ethical or moral statements. Within metaethics there are two opposed approaches: on the one side, cognitivism, according to which ethical statements are a product of cognition, of a cognitive attitude, and can therefore be objective more or less in the same sense as empirical statements. So that for cognitivism to say that: "George W. Bush is an American President", and to say that: "George W. Bush's war against Iraq is unjust", have the same or a similar epistemological status, that is, more or less the same claim to objectivity. At the end of the day, cognitivism in metaethics tends to become moral realism, an ontology according to which there are moral entities in the world.

42Opposed to cognitivism in metaethics there is non-cognitivism, which lays much stress on the logical difference between an is-statement ("they are bombing Bagdad" – for instance) and an ought-statement ("they should not bomb Bagdad', or "their bombing of Bagdad is wrong"). Hart adopts – we know it very well – a strong non-cognitivist view. He argues: (a) first, that is-statements are functionally deeply different from ought-statements; (b) second, that we cannot logically derive an ought-statement from an is-statement. For instance, from the fact that something is the case we could not derive that the same ought to be the case. From the fact that American soldiers torture Iraqi prisoners, we are not authorised to deduce that torturing Iraqis is the right thing to do for American soldiers or for anybody else.

43Such a non-cognitivist attitude is considered a sharp weapon by which natural law can be mortally wounded. In fact, natural law believes that (legal) rules are derivable from special or natural entities or states, so that there is a fixed hierarchy of rules and that the master rule of such a hierarchy has to be translated in terms of a statement about the world. Rules hierarchies here are static, or – if you like – somehow predetermined and substantive. Non-cognitivism, in contrast, makes it possible to conceive of rules relationships as a dynamic, procedural process.

44Summing up, I have considered four main doctrines of law faced by Hart: imperativism, realism, Kelsen's pure theory, and natural law. I have then discussed the five main tenets of Hart's approach: normativism, the internal point of view, the distinction between primary and secondary rules, the social thesis, and non-cognitivism. Each of these tenets is a tool used to defeat the four jurisprudential positions just mentioned. Normativism is a theoretical device used both against imperativism and realism; the internal point of view and the idea of obligation as subjection to a rule is a weapon against realism; the distinction between primary and secondary rules is used both against John Austin's analytical jurisprudence and against Kelsen's legal formalism; the same holds for the social thesis which is especially opposed to legal formalism; while non-cognitivistic metaethics is a formidable line of attack against natural law doctrines. But so far I have skipped the fifth main tenet of Hart's approach, a moderate rule-scepticism.

45According to Hart, rules have a semantical meaning. This meaning consists of a hard core where what is meant by the rule is clear, precise and unambiguous. If a case deals with this core, the judicial decision will find in the rule a distinct standard. There are, however, hard cases where we have to tackle the periphery of a rule' meaning. Here there is no determinacy, but a penumbra, that is, vagueness and ambiguity. If, for instance, in a park we are confronted with the rule "vehicles are forbidden", it is clear enough that such a rule prohibits people to drive cars within the park's precincts. But what about bycicles or wheelchairs – which doubtless are vehicles? What about introducing an American tank into the park to build a monument celebrating the liberation of Baghdad in April 2003?

46In such hard cases, Hart argues that rules are no longer able to direct the judge's decision. Judicial discretion takes here the upper hand over the formal rule. By this argument Hart inflicts a hard blow to legal formalism believing that judges do not really decide but only apply laws; though the blow is eschewed by Kelsen's "pure" theory which, on the contrary, defends the law-making power of judicial officials as far as the particular ruling is concerned. On the other hand, Hart thus strikes a compromise with legal realism which affirms that the concrete law-giving power are judges, and that these do strongly decide each case, not only in hard cases. However, by reintroducing from the window a relevant law-making power for judges previously expulsed from the main door, Hart is confronted with an unpleasant alternative. In hard cases, the law is no longer a matter of primary rule but only a question of secondary rule, namely of the rule of adjudication.

  • 15 Hart (1994: 81).

47The rule of adjudication is a parisitic one – as is said by Hart himself;15 it refers in order to give shape to the adjudication to another kind of (more substantive) rules. Now, the penumbra according to Hart is an infirmity which plagues exclusively primary rules which in such cases are fully derogated by secondary rules. These, nonetheless, are empty as far as the hard matter of the case is concerned. But how could the judge decide without substantive criteria. How can the parasite do without its exploited object?

  • 16 Hart (1994: 148).

48The alternative is then to state: either that there are strong objective moral criteria, but we shall thus fall into a kind of natural law approach; or that the judge can decide fully arbitrarily, according to the "tel est mon plaisir" formula, but we shall thus be taken back to legal realism in its most radical variant. In both cases the rule of recognition, now reduced to the mere indication of who is the judicial organ, will not be able to give us a standard of validity. One could argue that if the judge is empowered by the rule of recognition everything will be all right. However, even in this case, we will not have a standard or rule of validity, since such a standard should be able to hold also for the judge (if it is a standard, that is a self-binding prescription for those who use it). It should thus to able direct judges' conduct, that is, to limit or bind such conduct. But a rule just empowering the judges whatever they would like to rule cannot unfortunately play this role. On the other hand – as it is acknowledged by Hart himself – “the distinction between the uncertainty of a particular rule, and the uncertainty of the criterion used in identifying it as a rule of the system, is not itself, in all cases, a clear one”.16

49Let us start once again with the four main doctrines which Hart is opposed to: imperativism, realism, Kelsen's legal formalism, natural law. I would now like to see first how these four doctrine behave – so to say – in the domain of international law.

50(i) Imperativism – as we know already – is the view according to which the law is nothing but a set of commands or orders backed by the threat of a sanction and issued by a subject holding a supreme political power, that is, sovereignty. In this view two are the distinguishing features of law: that it is backed by force, by sanctions; that it is the outcome of a sovereign power.

51Now, as far as international law is concerned, such view leads to quite pessimistic conclusions. Since in international relations it seems that there are no sanctions nor only one sovereign power, the conclusion drawn by the imperativistic scholar is that there is no such thing as international law. In this respect, let me quote a few phrases by John Austin, the leading theorist of the analytical jurisprudence school:

  • 17 Austin (1954: 201).

International law, or the law obtaining between nations, regards the conduct of sovereigns considered as related to one another. And hence it inevitably follows, that the law obtaining between nations is not positive law: for every positive law is set by a given sovereign to a person or persons in state of subjection to its author. As I have already intimated, the law obtaining between nations is law (improperly so called) set by general opinion. The duties which it imposes are enforced by moral sanctions: by fear on the part of nations, or by fear on the part of sovereigns, of provoking general hostility, and incurring its probable evils in case they shall violate maxims generally received and respected.17

52According to Austin, therefore, international law is not "really" law, but only "positive morality". It is interesting to remember that for the British jurisprudent the same holds as far as constitutional law is concerned. Constitutional law too, in his view, is not "really" law, but only "positive morality". A disquieting conclusion indeed.

  • 18 See his Europa und Amerika : Olivecrona 1943.
  • 19 Ross (1966: 262).

53(ii) Legal realism is not usually much concerned with international law. However, insofar as legal realism is based on political realism, it tends to deny or diminish the legal status of international law. Political realism is the view of international relations according to which these are ruled only by governments or State interest, and such interests are fundamentally derived from a will of absolute power. So that for political realism (international) the law can only be a device for acquiring or maintaining State power, and never for submitting such power to an equitable rule of law. It is, in short, condemned to sway back and forth between – to use a more recent, fortunate expression – apology and utopia. To this regard we can here recall the view of two prominent Scandinavian realists: a Swede, Karl Olivecrona, and a Dane, Alf Ross. For the former, international law was something behind which lurks Carl Schmitt's Grossraumordnung, imperial hegemony.18 For the latter “the law between states (that is, international law) belongs to a category different from that of national law and lacks the latter's capacity to canalize intersts and aspirations”.19

54(iii) Much more complex and refined is Hans Kelsen's attitude. As a matter of fact, Kelsen was one of the greatest international lawyers of his time and has published extensively on the subject. I would like to remember his gigantic treatise on the United Nations of nearly a thousand pages. Kelsen was also heavily involved in the legal preliminaries to the Nuremberg trials whose actual proceedings, however, he criticised. Following Kelsen there are no conceptual difficulties in considering international law as a complete form of law. We might remember that according to him rules are legal, are law, only if they deal on sanctions. Now, one of the arguments against the legal character of international law is that there are no proper international law sanctions. Kelsen objects to such view. International law – he says – provides for a specific form of sanctions. This is war. Please, note that war here is legal, in so far as it is the reaction to a break of international law. Now, an important implication of such view is that a pre-emptive or anticipatory war, that is, a military intervention prior to an open violation of international law, would be unlawful.

55In the conceptualisation of the relationships between international law and municipal law there are two great alternatives: (a) dualism and (b) monism. Dualism purports that international law and municipal law are both law, but that they are not related the one with the other. Each law is valid in its own domain, and there are no relations of supremacy in one sense or the other between the two. This is, for instance, the stance taken by the Italian Constitutional Court, when dealing with the alleged supremacy of European Community law. Monism by contrast affirms that one of two kinds of law prevails over the other, in the sense that the validity of the one is derived or inferred from the validity of the other. As is well known, there are two kind of monism: statist monism, affirming that international law is just a derivation of municipal law, or – as the German philosopher Hegel said – "external municipal law"; international monism on the contrary believes that national law is somehow subordinate to international law. The latter position, international monism, is – we know – the one taken by Kelsen.

56You will remember that for Kelsen the law is a pyramid with different layers and a master rule on the top, the so-called Grundnorm. The lowest legal layer is given by judicial decisions; a higher level is the one of legislation authorizing judicial discretion; a still higher level is the constitution setting the tules for law-making. Now, higher than the constitutional level in the legal order pyramid is the international law layer. The constitution – according to what Kelsen says at least in some of his works – draws its validity from international law rules, in particular from the master rule of international law which is given by the so-called principle of effectiveness. The principle of effectiveness prescribes one to consider an effective State as a subject of international law, that is a fully valid legal order.

57Now, a constitution can be operative only if it is applied to an already valid legal order – but such preliminary validity is offered by international law. Kelsen's doctrine, therefore, not only recognizes the fully legal character of international law, but in addition makes the cornerstone of the entire positive legal order of such a law.

58(iv) Natural law theories are diverse and do not take a uniform attitude towards international law. For some theories, international relations are very like the state of nature where human beings live before entering in a civil state ruled by law. In the natural state subjects strive, fight, for survival and are ruled only by one principle: the preservation of life and, therefore, self-defence. Such is – as is well known – the view held by Thomas Hobbes or Baruch Spinoza. Jean-Jacques Rousseau is not very far from assuming a similar stance.

59There are, however, other natural law doctrines – for instance the one defended by Vitoria or Grotius – which assume a law valid for the entire human species regulating as well the relationships between peoples and States. For this second kind of natural law doctrines, then, there could not be any real doubt about the binding, fully legal character of international law.

60Now, what is Hart's position? At first Hart seems to assume an anti-sceptical attitude, by refusing the dogma of sovereignty.

  • 20 Hart (1994: 220).

One of the most persistent sources of perplexity about the obligatory character of international law – he writes in chapter ten of The Concept of Law – has been the difficulty felt in accepting or explaining the fact that a state which is sovereign may also be "bound" by, or have an obligation under, international law. This form of scepticism is, in a sense, more extreme than the objection that international law is not binding because it lacks sanctions.20

61Such a form of scepticism – Hart believes – should be rejected. In fact

  • 21 Hart (1994: 223–224).

belief in the necessary existence of the legally unlimited sovereign prejudges a question which we can only answer when we examine the actual rules. The question for municipal law is: what is the extent of the supreme legislative authority recognized in this system? For international law it is: what is he maximum area of autonomy which the rules allow to states?21

62That is, we cannot proceed in a prejudicial way by first posing a given notion of sovereignty and only afterwards trying to measure a law through this standard. We should rather focus directly on the phenomenon which we would like to assess without prejudices of any sort.

63Hart then is quite critical of those theories of international law which explain this through the notion of States' self-limitation. Such was, for instance, the doctrine defended by the Austrian scholar Georg Jellinek, actually one of Kelsen's teachers. Against the idea of international law as a product of States' self-limitation Hart presents three arguments. (a) Self-limitation presupposes a strong notion of sovereignty which is prejudicial to a concrete account of international law dynamics.

  • 22 Hart (1994: 224).

These theories – he writes – fail completely to explain how it is known that states 'can' only be bound by self-imposed obligations, or why this view of their sovereignty should be accepted, in advance of any examination of the actual character of international law.22

64(b) Self-limitation presupposes rules already in force.

  • 23 Hart (1994: 225). Italics are in the text.

In order that words /.../ should in certain circumstances function as a promise, agreement or treaty, and so give rise to obligations and confer rights which others may claim, rules must already exist providing that a state is bound to do whatever it undertakes by appropriate words to do.23

  • 24 Hart (1994: 225).
  • 25 Hart (1994: 226).

65(c) The practice of international law is at variance with the self-limitation doctrine. “Thirdly, there are the facts”.24 These only “can show whether this view is correct or not”.25

66As a matter of fact, in international law there are cases where a State is bound independently from any consent of the party bound. These are at least two: (i) “The case of a new state” – which is bound by the general obligation of international law; (ii) 

  • 26 Hart (1994: 226).

The state acquiring territory or some other change, which brings with it, for the first time, the incidence of obligations under rules which previously it had no opportunity either to observe or break, and to which it had no occasion to give or withhold consent.26

67This is, for instance, the case of a State acquiring access to the sea and becoming subject to the law of the sea provisions.

68Hart in a further move singles out a set of features of international law which makes this different from morality. Hart thus intends to object against both natural law doctrines and especially against John Austin's view of international law in terms of "positive morality".

69(a) A first point is the difference between legal and moral sanctions. Morality's sanction is appeal to the individual conscience, leading to feelings of guilt or shame. The same does not apply to international law.

70(b) The second point is the moral indifference of international law.

  • 27 Hart (1994: 229).

A more important ground of distinction is the following. The rules of intyernational law, like those of municipal law, are quite indifferent. A rule may exist because it is convenient or necessary to have some clear fixed rule about the subjects with which it is concerned, but not because any moral importance is attached to the particular rule.27

  • 28 Hart (1994: 230).

71(c) A third point is that while international law is based more or less on will and change, morality is not. While “there is nothing in the nature of international law which is inconsistent with the idea that the rules might be subject to legislative change”, “the very idea of change by human legislative fiat is repugnant to the idea of morality”.28 Thus morality – accordingly to Hart –, differently from law, cannot be just matter of will or decision; a view actually somehow at variance with his proclaimed metaethical non-cognitivism. (d) Nonetheless, the fourth point seems to centre around a further declaration of a non-cognitivistic espistemology. There is – he says – no moral obligation from which we could simply infer international law. Moral obligation does not lay at the root of international law.

72Following Hart's arguments up to this point it would seem that he recommends a non-sceptic view towards international law – which in other terms means that international law "really" is law and that it is or can be fully binding on States. However, Hart's conclusions do not take this direction.

73Hart's strategy is to assess whether international law is a complex "system", consisting of both primary and secondary rules, and not just a mere "set" of only primary rules. There is no permanent compulsory jurisdiction in international law – points out Hart. Therefore, we cannot affirm that there is anything close to rules of adjudication concerning international law. There is not any permanent law-making body either. Accordingly there cannot be any rule of change. Finally, since we do not find any well established body of fundamental international legal rules, we should also conclude that there is no international rule of recognition.

74The presence, however, of secondary rules, in addition to primary rules, is a necessary condition for a developed legal system. A mature system of law – says Hart – is one structured along a functional hierarchy of rules. Otherwise, we will have to deal with under-developed, primitive legal phenomena, which do not even deserve to be qualified as "systems". They are only (static) "sets" of rules, without any clear coordination and hierarchy and with a very law degree of determinacy. Such (primary and secondary) rules to be valid are only to be accepted without necessity of recurring to a reflexive standard such as the rule of recognition.

  • 29 Hart (1994: 235). Italics are in the text.

In the simpler form of society we must wait and see whether a rule gets accepted as a rule or not; in a system with a basic rule of recognition we can say before a rule is actually made, that it will be valid if it conforms to the requirements of the rule of recognition.29

  • 30 Hart (1994: 108–109).

75We could however ask whether it is true that acceptance of a rule is practicable without a previous recognition of the rule itself. We could further wonder whether such problem is solved once it is transposed at the level of the "rule of recognition", which – according to Hart – can only be used but cannot be identified from the internal point of view, so that that the existence of the rule of recognition can only be a question for the external point of view.30 We could also object that a mere acceptance of (primary) rules as it would be the case of primitive legal orders is not very far from a mere convergence of conducts, thus transforming a normative practice into a simple habit – a move – as we know – actually criticised by Hart himself.

76Be that as it may, what is important for us today is that Hart, by denying that international law comprises secondary rules, condemns it to be considered a "primitive" system of law, something – I would like to stress – which is not so different from what John Austin had called "positive morality", a form of "law not properly so called" – as he used to say.

  • 31 Hart (1994: 236).
  • 32 Hart (1994: 236). Italics are mine.

77There is no basic rule providing general criteria of validity for the rules of international law” – says Hart trenchantly.31 There are however “the rules providing for the binding force of treaties”, but these are judged to “constitute not a system but a set of rules”.32 Which means that for him, as for Austin, international law is not "really" law; actually a sad and somewhat disturbing last word. Is then John Bolton, U. S. Sub-Secretary of State for arms control eventually right?

78In a Post-script to his book The Concept of Law, written more than twenty years after its first publication, Hart tries to reply to Ronald Dworkin's formidable attack against his philosophy of law. In particular, Hart tries to include in the legal domain, in addition to rules, principles as well. He thus affirms that the rule of recognition can consist of or contain moral principles. I cannot enter here into the subtle controversy connected with this point. I would only draw your attention to the fact that, once the rule of recognition consists of principles, one of Hart's main objections against the full legality of international law should be dropped, that is, that there are no rules building up an international rule of recognition. This now however – once one accepts that the rule of recognition can consist also of principles and not merely of rules) – could be offered by (moral) principles. As a matter of fact, the supreme rule of validity, actually the rule of recognition, of international law is said by doctrine to be given by two basic principles: consuetudo est servanda (one ought to follow customary rules); and pacta sunt servanda (agreements are to be followed). Then, there is perhaps still hope for international law.

79Let me conclude. Traditionally the law is conceived as a hierarchical experience. This mainly happens in two ways. (i) First, the law is seen as the matter of orders and sheer obedience or the matter of overwhelming power and unjustified allegiance, backed (motivated) by force and violence. (ii) Second, the law is conceived as an experience reserved as articulate reasoning only to a special class, an elite – we might say – of society members, "officials". (iii) There is however a further, third way of thinking the law in hierarchical terms: this is to construct it as a system built up along different levels of rules related the one to the other as a superior to the inferior.

  • 33 Hart (1994: 155).

80Now, Hart strongly objects to the first way. “Too much that is characteristic of law is distorted – he says – by the effort to explain” it in the simple terms of a threat, or of obedience.33 Nonetheless, Hart's idea that a mature system of law cannot be explained but in terms of "secondary rules" and that such rules establish an agency endowed with a final say about the meaning of law refers at the end of the day to some kind of power or "discretion" which is not so far way from the traditional notion of command.

  • 34 Hart (1994: 114–115).

81As far as the second way is concerned, by saying that the rule of recognition holds only in so far as officials, not the generality of citizens are concerned.34 Hart makes legal obligation only a matter for a special body of agents. He thus reintroduces a sharp hierarchy between officials as superiors and citizens as inferiors.

82The third and final way in which he intends to reassert and reformulate it is through the distinction between primary and secondary rules and the notion of a rule of recognition. Whether he does so successfully or not I have tried to question in this paper.

Top of page


John AUSTIN, 1954: The Province of Jurisprudence Determined. With an introduction by H. L. A. Hart. London: Weidenfeld & Nicolson.

Herbert L.A. HART, 1994: The Concept of Law. 2nd edition, with a Postscript ed. by Penelope A. Bulloch and Joseph Raz, Oxford: Clarendon Press.

Karl OLIVECRONA, 1943: Europa und Amerika. Berlin: Junker und Dünnhaupt.

Alf ROSS, 1966: The United Nations: Peace and Progress. Totowa, N. J.: Bedminster Press.

Kenneth I. Winston, 2001: The Principles of Social Order : Selected Essays of Lon L. Fuller. Oxford: Hart Publishing.

Top of page


1 Winston (2001: 275–276)

2 Winston (2001: 313)

3 Hart (1994: 26).

4 Hart (1994: 26).

5 Hart (1994: 26).

6 Hart (1994: 42).

7 Hart (1994: 26).

8 Hart (1994: 55). Italics are the author's.

9 Hart (1994: 55). Italics are in the text.

10 Hart (1994: 56).

11 Hart (1994: 57).

12 Hart (1994: 57).

13 Hart (1994: 80).

14 Hart (1994: 80).

15 Hart (1994: 81).

16 Hart (1994: 148).

17 Austin (1954: 201).

18 See his Europa und Amerika : Olivecrona 1943.

19 Ross (1966: 262).

20 Hart (1994: 220).

21 Hart (1994: 223–224).

22 Hart (1994: 224).

23 Hart (1994: 225). Italics are in the text.

24 Hart (1994: 225).

25 Hart (1994: 226).

26 Hart (1994: 226).

27 Hart (1994: 229).

28 Hart (1994: 230).

29 Hart (1994: 235). Italics are in the text.

30 Hart (1994: 108–109).

31 Hart (1994: 236).

32 Hart (1994: 236). Italics are mine.

33 Hart (1994: 155).

34 Hart (1994: 114–115).

Top of page


Bibliographical reference

Massimo La Torre, « The Hierarchical Model and H. L. A. Hart's Concept of Law », Revus, 21 | 2013, 141–161.

Electronic reference

Massimo La Torre, « The Hierarchical Model and H. L. A. Hart's Concept of Law », Revus [Online], 21 | 2013, Online since 25 February 2014, connection on 17 January 2018. URL : ; DOI : 10.4000/revus.2746

Top of page

About the author

Massimo La Torre

Professor in Philosophy of Law at the Law School of Magna Graecia University in Catanzaro (Italy) and at the Hull University Law School (United Kingdom).

   Massimo La Torre
Dipartimento di Giurisprundeza
Università degli Studi Magna Graecia di Catanzaro
Viale Europa, Loc. Germaneto
88100 Caranzaro


Top of page


All rights reserved

Top of page
  • OpenEdition Journals