Hart's Concept of Law Lecture Two 2001


(a) Rules which pertain to different activities are different;

(b) Rules which originate in different ways are different.

(c) Rules may differ in their relation to conduct.

ad (a) This is a Wittgensteinian view. If the meaning of a word is in its use and if rules are used to define or constitute "activities" or "language games" then every "language game" or "form of life" has its own rules with their unique use and hence different meanings. But the argument is hardly persuasive. The rules could all be of the same logical type though performing different functions. It does not follow that the rules (or laws) that relate to marriage and those that relate to commerce are necessarily of different types simply because they serve different social purposes or sustain very different "activities".

ad (b) No one who holds that rules are of the same logical type is committed to the belief that they originate in the same way. It may be possible to regard rules of law as of the same logical form whether they originate in custom, from courts or from the legislature and one question for General Jurisprudence is precisely whether the variety of legal phenomena can be "reduced" to one logical form. It may or may not be the case that statutory rules and common law rules are logically distinct and cannot be assimilated to one model. But that is a big question within jurisprudence on which theorists differ. For example, Dworkin takes the view [at p 313 of Law's Empire ] that "much the same techniques of interpretation" are appropriate to reading statutes as to deciding common law cases whereas Hart, though he does acknowledge that the differences can be exaggerated, nonetheless regards the difference as significant [pp 121-124] The important point of this discussion is whether or not it is possible to subsume both common law and statutory law under one concept of law even though they clearly do originate in different ways.

ad (c) Quite what Hart means is unclear. Unless we have an idea of how rules relate to conduct we cannot suppose different relations between rules and conduct. Something along the following lines may reconstruct Hart's position: The relation between rules and behaviour is normative. The behaviour is what ought to be according to the rule. But this "ought" may be too coarse or general a term adequately to capture the relationship between rule and behaviour because a rule might command behaviour or merely authorise it. Such a fundamental difference between "must" and "may", that is between normative functions or what I call "normative modalities" should, it might well be thought, be clearly marked by a terminological distinction rather than fudged by being merged into one general term such as "ought". The suggestion here is that "ought" should be further divided into different normative functions or modalities such as commanding, i.e. that which one "must" do and authorising - that which one "may" do. Any analysis of law which does not make such a distinction as for example Austin's and allegedly Kelsen's might well be thought to be inadequate and distorting. Kelsen gave his "ought" four separate meanings: command, authporise, permit, and (added later) derogate.

Now, such a view would be extremely interesting and important. But though it is possible to read it into Hart, there is no corroborating evidence that this is what he meant. And at this stage there is no justification for the view that there are rules of different types. Yet this view is central to Hart's thesis. And "rules" are unlikely to be helpful because, according to Hart, they are complex, puzzling, confusing, and baffling.. As he writes, "Some jurists indeed find the notion utterly mysterious" [p 9].

Hart then presents what is to him a crucial distinction between rules and habits. A mere convergence of behaviour has to be distinguished from the idea that something ought to be done. Hart is distinguishing what is done "as a rule" from what is done in accordance with or because of an existing rule. It is a feature of the latter to bring in normative vocabulary such as "must", "should", "ought", etc. Though Hart tend to cloud the issue in niceties of linguistic analysis the point is important. There is a significant distinction between description and evaluation; between fact and value. Value judgments expressly of impliedly invoke standards or "norms"; they are or invoke proposition about what ought to be. Such value judgment or norms may or may not be prescriptive as well as normative, that is to say they may prescribe conduct under specified conditions. Clearly moral norms and legal norms are prescriptive in this sense whereas aesthetic judgment about the beauty of a sunset or the merits of a painting are not obviously if at all prescriptive.

According to Hart regularities of social conduct are not all the product of existing rules. Thus to point to a particular regularity is not of itself to imply that such behaviour ought to be. Rules are not evident from mere observation of regularities: 19. "Even after watching a thousand games [of chess] it would still be possible to believe that it is against the rules to open with a rook's pawn". (Alf Ross, On Law and Justice, London, 1958, p.15). Hart suggests that a difference between habits and rules is that conduct deviating from the latter will probably meet with hostile reactions and in the case of legal rules punishment imposed by officials. Breach of habits, however, if one may even put it like that, does not lead to such hostile reaction or sanction. This distinction, which characterises much jurisprudence, places emphasis on the predictive element of social rules but Hart says - and rightly - that this overlooks the way that judges and officials within the system look upon the rules. They do not treat the rules as predictions of their own behaviour. Rather they regard the rule as a justification or a reason for their behaviour. The predictive element is irrelevant from the point of view of "our friend the bad man" : 20. "Take the fundamental question, What constitutes the law? You will find some text writers telling you that it is something different from what is decided by the courts of Massachusetts or England, that it is a system of reason, that it is a deduction from principles of ethics or admitted axioms or what not, which may or may not coincide with the decisions. But if we take the view of our friend, the bad man, we shall find that he does not care two straws for the axioms or deductions, but he does want to know what Massachusetts or English courts are likely to do in fact. I am much of his mind. The prophecies of what the courts will do in fact and nothing more pretentious are what I mean by law". (Oliver Wendell Holmes, Collected Papers, New York, 1920, p.173).

Hart is beginning here to set up the distinction between the "internal" and the "external" aspect of rules. Internally rules provide a standard for official behaviour. The predictive view-point is external and thus cannot adequately account for the way in which rules account for the conduct of officials. The point is familiar enough to Kelsen scholars. Kelsen talks of "the immanent meaning of law". This, says, Kelsen, is a normative meaning; an ought proposition. Law, for Kelsen, is a system of norms stipulating when coercion may, or can, or must be applied. Law, then, is a set of directives by which official behaviour is regulated. The legal norms (or "standards" in Hart's terminology) determine what officials ought to do. And, of course, no amount of predictive utterances, no amount of sociological description can recapture the internal normative aspect of law. It is because of this that Kelsen insists upon a strict separation of law as a system of norms regulating human conduct and actual human behaviour; a radical separation of normative propositions and statements of fact; of prescription and description; of normative jurisprudence and descriptive sociology.

But Hart does not develop the internal/external distinction at this stage. Rather he seek to resist the strict, predictive view of the Scandinavian Realists which denies that law has or can have any such immanent meaning and he calls for " a further elucidation of the distinction between social rules and mere convergences of behaviour" [p 11], a distinction which he thinks is crucial to an understanding of law. For Hart, then, the internal aspect of law is central to an understanding of law.

Further, Hart draws attention to the rule scepticism of American Realists who have challenged the view that law consists wholly or even primarily of rules, on the basis that judges do not apply rules:-

(a) in cases of first impression;

(b) in cases where rules conflict; and

(c) generally they always interpret the law, and are therefore not merely applying but also making law.

But such rule sceptics claim too much. From the proposition that judges sometimes do not apply legal rules it cannot be deduced that judges never follow rules, which is what the extreme rule-sceptic, and the later Critical Legal Studies movement, have claimed. Hart counters the extreme sceptic by arguing that legal rules exhibit a central core of indisputable meaning and in many cases this is so clear that one could not imagine a dispute as to the meaning of the rule breaking out. Hart returns to this challenge in Chap VII where he deals with the issues in much more detail.

Finally in Chap I there is a note on definition, the gravamen of which is that the popular method of definition per genus et differentiam is inappropriate in jurisprudence because we lack a genus or family of things to which the concept of "law" belongs. The significant claim from the point of view of criticism is again hidden away amid linguistic philosophising but Hart here clearly denies that "law" can be defined per genus et differentiam because there is no obvious genus. The most obvious is "rules of behaviour" but 21. "... the concept of a rule, as we have seen, is as perplexing as that of law itself so that definitions of law that start by identifying law as a species of rule usually advance our understanding of law no further". (Hart, op.cit., p.15). Savour this passage. Think about it. Hart says that definition per genus et differentiam is not always helpful because it requires that we know and understand the genus; otherwise we are explaining obscurum per obscurius This method won't do for law: there is no obvious genus. The most likely is "rules". But that is too vague. Therefore it will not do. Yet, as we will see, Hart proceeds at length, about 200 pages in all, to extol the virtues of an analysis of law in terms of the union of rules of different types. This is peculiar. If "rules" is unclear then a fortiori "rules of different types" must be unclear.

However Hart claims not to be defining but merely elucidating, that is, he is not laying down necessary and sufficient conditions. On that basis, perhaps, the vagueness of the concepts "rule" is not so important. Yet this hardly seems adequately to explain the self-refuting nature of Hart's views. Perhaps he will clarify the notion of a rule sufficiently to permit of its place as a central or "key" concept. We must look out for this but you may agree that, by the end of the Concept, hart has not sufficiently clarified the notion of a rule to allow the central role to the "union of rules of different types" which he claims for it. For, as you may know, this "union of different kinds of rules" is, Hart claims, the "new key" to the science of jurisprudence.

Austin, John (1790-1859). John Austin is rightly celebrated as one of the founding fathers of English jurisprudence for all that he was very much the pupil and Jeremy Bentham (1748-1832) the master as later research has so convincingly revealed. Austin was born into a good family and married well. He had joined the army when sixteen years old, but sold his commission after about five years and took up law studies with a view to a career at the Bar and he was called to the Inner Temple in 1818. Although his name appears on the list as an equity practitioner until 1829, he had given up all thoughts of practice as early as 1825, success having eluded him. When the University of London (now University College) was founded in 1826 he was appointed its professor of jurisprudence and he set about preparing a course of lectures with characteristic and painstaking thoroughness over the following two years. He found writing difficult, imposing upon himself high standards of exactitude and clarity and his life was riven with nervous illness, depression and self-doubt.

The first lectures in 1828 drew a sizeable and distinguished audience, including John Stuart Mill, but Austin did not attract new students, his audiences dwindled virtually to vanishing point, and he resigned his chair in 1832. Of her husband, Sarah Austin wrote that he lived 22 "a life of unbroken disappointment and failure" (Austin 1968: xvi) though it does appear that throughout his life he enjoyed the unstinted support of a loving wife and loyal friends. Indeed, it was chiefly through Sarah's efforts that an expanded version of his introductory lectures was published as The Province of Jurisprudence Determined (including an Outline of the whole lecture course) in the Autumn of that year and although little notice was taken of it at the time it is now regarded as a classic of English jurisprudence even though it was intended by Austin to be merely a preparatory work. A second edition was published in 1861 due again to the indefatigable Sarah, and an essay on "The Uses of Jurisprudence" with which Austin had commenced his lectures was eventually published in 1863. Sarah also published two volumes of Austin's lectures in 1863 and a later edition of these Lectures in Jurisprudence, edited by Robert Campbell, was published in 1885.

Austin also delivered his lectures by invitation at the Inner Temple in 1834, but again there was no interest and he failed to attract an audience. Through the good offices of his influential and loyal friends he was appointed by Lord Brougham to the Criminal Law Commission but he was soon disillusioned and resigned. He was subsequently appointed to a commission to enquire into the government and administration of Malta and in this Austin was uncharacteristically successful and the resultant documents are highly praised. He even earned some money from the exercise but otherwise it is unlikely that he earned even one hundred pounds from lecturing and writing. In the latter years of his life he spent much time in Germany and France, supported by Sarah's resources and earnings and finally settled in Weybridge, in England where he died in Sarah's words 23 "in obscure and honest poverty" (Macdonell 1885) in 1859. Apart from the Province, Austin, alway self-critical and dissatisfied with his work, published little although an important essay on "Centralization" appeared in Volume 85 of the Edinburgh Review in 1847 and a trenchant pamphlet, entitled A Plea for the Constitution was published in 1859.

Austin's immortality rests on two related but distinct contributions: first, his account of the nature of jurisprudence itself and, secondly, his command theory of law. It is less Austin's own peculiar view of nature of General Jurisprudence as 24"... the science concerned with the exposition of the principles, notions and distinctions which are common to systems of law" (Austin 1968:367), than it is the important distinction that he drew between general and particular (or national) jurisprudence and his belief that such a generalised approach to the study of law was both possible and beneficial, that secures Austin's immortality. As Hart rightly observed (Austin 1968: xv), an emphasis on "principles" of law common to all mature systems does tend rather too much towards the very natural law thinking that otherwise Austin was at considerable pains to denounce. However, an emphasis on "notions and distinctions" common to all mature systems inspired 25 "a line of English analytical jurists [including] Amos, Clark, Markby, Hearne, Holland, and Salmond" (Austin 1968: xvii) and in America one can point to Gray's The Nature and Sources of Law (1909) and Hohfeld's Fundamental Legal Conceptions (1923) as works in the Austinian analytic tradition. (Morison 1982: 160-167).

Thus, though he did not live to know it, Austin founded a distinctive and enduring style or school of thinking about law, namely Analytical Jurisprudence:26 "The analysis of legal concepts is what jurisprudence meant for the student in the days of my youth. In fact it meant Austin. He was a religion ..." (Buckland 1949: 2). And whilst there may have been a tendency in the past amongst the practitioners of this style of thinking about law to regard the analysis of concepts as an end in itself, its real significance is as groundwork for the critical evaluation of beliefs and, as such, it is an important adjunct to all jurisprudential thought, not only in the positivist, but also in the naturalist, rationalist, or realist traditions:27 "... the duty of the scientific inquirer is to distinguish the meanings of an important word from one another, to select the meaning appropriate to his own purposes, and consistently to employ the word during his investigations in this sense and no other" and "... one inestimable service of the Analytical School to jurisprudence and morals [is] that it furnishes them with a rigidly consistent terminology" (Maine [1874]1914: 374 and 369)

Austin's command theory of law provides a second claim to immortality. The theory was neither original nor complex. It can be seen in the words of Thasymachus in Plato's Republic. Essentially for Austin, 28 "... the term command [an order backed by a sanction for non-compliance]... is the key to the science of jurisprudence" (Austin 1968: 13) and "... every positive law, or every law strictly so-called, is a direct or circuitous command of a [legally illimitable] monarch or sovereign number in the character of a political superior..." (Austin 1968: 134). Although this was the dominant theory of law in England for a century there is a great deal wrong with it.

First, it does not fit much of what is usually recognised as "law". Indeed, it fits very little other than, perhaps, the English criminal statute. The law of property, trusts, succession, and contract can be brought within the model only by the implausible device of treating nullity as a sanction but 29 "nullity cannot ... be assimilated to a punishment attached to a rule as an inducement to abstain from the activities which the rule forbids" (Hart 1961:34) Austin, himself, excluded international law and constitutional law. Administrative law, too, is hard to fit into the model. Worse still, the model can only accommodate the common law on the basis of the unsatisfactory fiction that what the sovereign permits, the sovereign commands and in any event the common law is, perhaps, better understood as unwritten, customary law rather than circuitous sovereign commands: 30 "The notion that the common law consists of rules which are the product of a series of acts of legislation (mostly untraceable) by judges (most of whose names are forgotten) cannot be made to work, if taken seriously, because common law rules enjoy whatever status they possess not because of the circumstances of their origin, but because of their continued reception" (Simpson 1986:14) Secondly, the model fits uneasily if at all into a 31 "charter society" (Wallachow 1994: 142-165), that is one where ordinary legal rules are susceptible of override if inconsistent with the provisions of a charter or bill of rights and that calls into question both the sovereign legislature's immunity from legal limitation and the command theory's status as a contribution to general jurisprudence. Thirdly, the central notion of "sovereignty", albeit influential on the great Dicey's constitutional theory, is misconceived in that it is descriptively inaccurate and normatively unappealing (see Maine [1874] 1914, Ch XII). Austin, himself, had some difficult in locating a sovereign even in the legal system of his own time and place, and federal systems, especially if combined with a bill of rights, as in the United States of America, cannot without great ingenuity, if at all, be accommodated to the model. Nor is it at all obvious that a legal system that concentrates all law making in one supreme individual or institution is normatively more attractive than one in which the legislative power is divided and decentralized, even if, as Austin (and Bentham) stipulated, the sovereign ought to makes laws only in accordance with the principle of utility, "the greatest happiness of the greatest number". The stronger that constraint upon the sovereign legislature, the less normatively unappealing the theory. However, the principle of utility is indeterminate at least as to the relevant tract of future time and as to the relevant population and therefore a very wide range of legislative products can be thought consistent with the principle. Accordingly compliance with the principle of utility may be a far lesser constraint than Austin (and Bentham) contemplated.

With so many demerits it is remarkable that the command theory endured. Part of the reason is the certainty payoff and the obvious appeal of the theory to those of an executive cast of mind. In the case of Bentham, the command theory also tied very comfortably into his codification project which, had it ever been implemented, would have abolished the common law and reduced the role of the judge (as it was under the French Civil Code of 1804) to that of "... the mouth which pronounces the words of the law", in Montesquieu's oft repeated phrase. But Austin took a very different view of the common law:32 "That part of the law of every country which was made by judges has been far better made than that part which consists of statutes enacted by the legislative. Notwithstanding my great admiration for Mr. Bentham, I cannot but think that, instead of blaming judges for having legislated. He should blame then for the timid, narrow, and piecemeal manner in which they have legislated, and for legislating under cover of vague and indeterminate phrases ..." (Austin 1968: 191). Given such tensions within the theory that he had adopted it is perhaps unsurprising that, 33 "dissatisfied with all he did, Austin wrote little ... [and] could not even be induced to prepare a second edition of his work on jurisprudence" (Macdonell 1885). Yet it has seemed to many that there was much to be learned even from Austin's mistakes and it is fitting to give the final word to Hart, whose Concept of Law (1961) marked the official end of the "Age of Austin" in jurisprudence: 34 "... buried in the idea of command there are ... elements which are crucial to the understanding of law" (Hart 1982: 268).

Chapter II: Austin's Theory of Law. Hart draws attention to "the simple imperative theory" in order that we may learn from its errors. This is introduced with a consideration of the "variety of imperatives" where, in rapid succession hart distinguishes requests, pleas, warnings, orders, the giving of an order, and coercing, but he says we need not bother with these linguistic niceties. He bemoans the absence of a general word to cover all these different types of utterances - though perhaps "prescription", "directive" or even "act of will" would serve.

The chief point of his investigation is simply made. Hart declares his intention to use the phrase "orders backed by threats" or "coercive orders" to refer to orders like the gunman's which are supported only by threats; and "obey" or "obedience" to refer to compliance with such orders. The point is to strip"orders" and "obedience" of any normative implications. As Hart intends to use these terms there is no element of authority. The simple gunman situation is the model whereby Hart explains his intended meaning . That (for Hart) is the kind of situation in which these two words are correctly used. Thus the simple imperative theory of law is the "gunman situation writ large".

Hart does not let the occasion pass without criticising Austin's central concept - "the key to the science of jurisprudence" on the basis that the type of situation characterised by Austin is not one in which we naturally speak of commands: 35. "... a command is primarily an appeal not to fear but to respect for authority". (Hart, op.cit., p.20). But Austin defined "command" for his own purposes as follows: 36. "A command, then, is a signification of desire. But a command is distinguished from other significations of desire by this peculiarity: that the party to whom it is directed is liable to evil from the other, in case he comply not with the desire". (John Austin, op.cit., p.14). Austin therefore left us in no doubt as to what he meant and we do not have to consider whether or not the type of social situation to which he refers was "naturally" [ p 20] one to which the word "command" was appropriate. Therefore Hart's criticism is beside the point. More important is the question of "authority" Hart admits that an analysis of this concept is no easier than that of "law" itself and that we cannot profitably use the concept of authority as the basis of our account of law. So it may be that Austin did not do so badly to avoid it. Further, it may be that Austin, like the Scandinavian Realists and some neo-Marxists, did not so much "ignore" [p 20] the distinction between power and authority, but rather rejected it !

Hart insists, rightly, that the simple imperative model has to be augmented if it is to reproduce the characteristics of the modern legal system. To get from the gunman situation to an Austinian position four features not present in the gunman situation must be added:-

1. Generality; 2. Enduring nature of law; 3. A habit of obedience; and 4. Sovereignty

ad 1. In the operation of the law face-to-face encounters between officials and citizens are rare; rather the law utilises the technique of what Hart now calls "general forms of directions" [p 21] which do not name and are not addressed to particular individuals; and do not indicate a particular act to be done. The "criminal statute" [p 21] is doubly general; i.e. general as regards type of conduct; and general as regards classes or categories of persons. Official individuated face-to-face directives are rare and in any event derive from "general directives" [p 21] Notice the various synonyms for "rules" that Hart is now deploying. He can't very well use the word "rule" because only 6 pages earlier he affirmed that the concept of a rule is too vague and puzzling to explain law. But the point here is that generality must be added to the simple imperative theory in order that it reproduce the character of law.

Clearly Hart is not criticising Austin's theory because Austin was well aware of the distinction between general commands on the one hand and occasional or particular commands on the other. Rather Hart is here building up the Austinian position step by careful step from an artificially simplified imperative model. Austin is criticised for speaking of laws being "addressed to the community at large" [Province p 22] because "addressed" suggests a face-to-face situation which does not obtain and was not actually intended by Austin. "Ordering", for Hart, does entail actually addressing individuals, attracting their attention, or taking steps to attract their attention. "Making law" for people does not. Austin, says Hart, in speaking of laws being addressed to people meant only that these are the people to whom the law applies, though there need have been no publication or actual communication. Thus Hart suggests that Austin's term "addressed" blurs the distinction between two questions: (i) To whom does the law apply? And (ii) To whom has it been published?

ad 2. A further element which needs to be added to te gunman situation is that of the endurance of laws. In the gunman situation the gunman has temporary or transient ascendency over the bank teller. Thee is no other relationship of superiority or inferiority. His orders lack the enduring quality of law. We need therefore to add a general belief that the threat will be executed not only on the first promulgation of the order but continuously thereafter until the order is withdrawn or cancelled. This continuing belief changes "orders" into "standing orders". [p 23]

ad 3 This general belief is to be distinguished from a third feature with which the gunman situation is to be fortified; that of a general habit of obedience; i.e. that most of the orders are more often obeyed than disobeyed. Though this is a vague notion it is an important one; it is what distinguishes law from the merely transitory supremacy of the gunman.

ad 4 There is another characteristic which needs to be added. The legal system of a modern state is characterised by a certain kind of supremacy within its territory and an independence of other systems which is not in our simple model. Bear in mind that Hart published this in 1961, some years prior to Britain joining the common market and the enactment of the European Community Act, 1972. However, Hart's concluding paragraph is a concise summary of the Austinian position which he has built up out of the artificially simplified gunman model... 37. "... there must, wherever there is a legal system, be some person or body of persons issuing general orders backed by threats which are generally obeyed, and it must be generally believed that these threats are likely to be implemented in the event of disobedience. This person or body must be internally supreme and externally independent. If, following Austin, we call such a supreme and independent person or body of persons the sovereign, the laws of any country will be the general orders backed by threats which are issued either by the sovereign or subordinate in obedience to the sovereign". (Hart, op.cit., p.52).

So we have built up a situation which parallels that of the criminal statute in a modern state. But, says Hart, there are "types of law" [p 24] which do not seem to fit:- 38 "The concept of general orders backed by threats given by one generally obeyed, which we have constructed by successive additions to the simple situation of the gunman case, plainly approximates closer to a penal statute enacted by the legislature of a modern state than to any other variety of law. For there are types of law which seem prima facie very unlike such penal statutes, and we shall have later to consider the claim that these other varieties of law also, in spite of appearances to the contrary, are really just complicated or disguised versions of this same form". (Hart, op.cit., p.24). Now Hart is not here referring to the "laws which are not imperative" admitted by Austin, but to what he calls power-conferring laws, laws which create facilities, e.g. contracts, wills, etc He says that we shall have to consider claims which present power-conferring rules as disguised versions of "general orders backed by threats" [p 24] - the allusion is to Bentham's "fragmentation" and Kelsen's idea of "a dependent norm", of which more later.

Hart's programme in this Chapter has been to present the simple imperative model of law and by analysing the gunman situation build up the traditional Austinian view of law as strongly as possible, in order, subsequently to show that even at its best, it is not adequate for explaining the law of "modern" systems, even though Austin, himself, had expressly limited his account to "mature systems".


[BRIDGE TO NEXT LECTURE: I will turn next to a brief resume of Austin's Command Theory of Law. It may be simply stated in Austin's own words: 39. "Every positive law ... is set, directly or circuitously, by a sovereign individual or body, to a member or members of the independent political society wherein its author is supreme ... The party or parties to whom a law is set, or the party or parties on whom a duty is laid, are necessarily obnoxious to the sanction which enforces the law and the duty ... It is set by a party armed with might, to a party or parties to whom that might can reach". (John Austin, op.cit., p.350)].