II. LAWS, COMMANDS, AND ORDERS

1. VARIETIES OF IMPERATIVES

The clearest and the most thorough attempt to analyse the concept of law in terms of the apparently simple elements of commands and habits, was that made by Austin in the Province of Jurisprudence Determined. In this and the next two chapters we shall state and criticize a position which is, in substance, the same as Austin’s doctrine but probably diverges from it at certain points. For our principal concern is not with Austin but with the credentials of a certain type of theory which has perennial attractions whatever its defects may be. So we have not hesitated where Austin’s meaning is doubtful or where his views seem inconsistent to ignore this and to state a clear and consistent position. Moreover, where Austin merely gives hints as to ways in which criticisms might be met, we have developed these (in part along the lines followed by later theorists such as Kelsen) in order to secure that the doctrine we shall consider and criticize is stated in its strongest form.

In many different situations in social life one person may express a wish that another person should do or abstain from doing something. When this wish is expressed not merely as a piece of interesting information or deliberate self-revelation but with the intention that the person addressed should conform to the wish expressed, it is customary in English and many other languages, though not necessary, to use a special linguistic form called the imperative mood, ‘Go home!’ ‘Come here!’ ‘Stop!’ ‘Do not kill him!’ The social situations in which we thus address others in imperative form are extremely diverse; yet they include some recurrent main types, the importance of which is marked by certain familiar classifications. ‘Pass the salt, please’, is usually a mere request, since normally it is addressed by the speaker to one who is able to render him a service, and there is no suggestion either of any great urgency or any hint of what may follow on failure to comply. ‘Do not (p. 19) kill me’, would normally be uttered as a plea where the speaker is at the mercy of the person addressed or in a predicament from which the latter has the power to release him. ‘Don’t move’, on the other hand, may be a warning if the speaker knows of some impending danger to the person addressed (a snake in the grass) which his keeping still may avert.

The varieties of social situation in which use is characteristically, though not invariably, made of imperative forms of language are not only numerous but shade into each other; and terms like ‘plea’, ‘request’, or ‘warning’, serve only to make a few rough discriminations. The most important of these situations is one to which the word ‘imperative’ seems specially appropriate. It is that illustrated by the case of the gunman who says to the bank clerk, ‘Hand over the money or I will shoot.’ Its distinctive feature which leads us to speak of the gunman ordering not merely asking, still less pleading with the clerk to hand over the money, is that, to secure compliance with his expressed wishes, the speaker threatens to do something which a normal man would regard as harmful or unpleasant, and renders keeping the money a substantially less eligible course of conduct for the clerk. If the gunman succeeds, we would describe him as having coerced the clerk, and the clerk as in that sense being in the gunman’s power. Many nice linguistic questions may arise over such cases: we might properly say that the gunman ordered the clerk to hand over the money and the clerk obeyed, but it would be somewhat misleading to say that the gunman gave an order to the clerk to hand it over, since this rather military-sounding phrase suggests some right or authority to give orders not present in our case. It would, however, be quite natural to say that the gunman gave an order to his henchman to guard the door.

We need not here concern ourselves with these subtleties. Although a suggestion of authority and deference to authority may often attach to the words ‘order’ and ‘obedience’, we shall use the expressions ‘orders backed by threats’ and ‘coercive orders’ to refer to orders which, like the gunman’s, are supported only by threats, and we shall use the words ‘obedience’ and ‘obey’ to include compliance with such orders. It is, however, important to notice, if only because of the great influence on jurists of Austin’s definition of the notion of a (p. 20) command, that the simple situation, where threats of harm and nothing else is used to force obedience, is not the situation where we naturally speak of ‘commands’. This word, which is not very common outside military contexts, carries with it very strong implications that there is a relatively stable hierarchical organization of men, such as an army or a body of disciples in which the commander occupies a position of pre-eminence. Typically it is the general (not the sergeant) who is the commander and gives commands, though other forms of special pre-eminence are spoken of in these terms, as when Christ in the New Testament is said to command his disciples. More important—for this is a crucial distinction between different forms of ‘imperative’—is the point that it need not be the case, where a command is given, that there should be a latent threat of harm in the event of disobedience. To command is characteristically to exercise authority over men, not power to inflict harm, and though it may be combined with threats of harm a command is primarily an appeal not to fear but to respect for authority.

It is obvious that the idea of a command with its very strong connection with authority is much closer to that of law than our gunman’s order backed by threats, though the latter is an instance of what Austin, ignoring the distinctions noticed in the last paragraph, misleadingly calls a command. A command is, however, too close to law for our purpose; for the element of authority involved in law has always been one of the obstacles in the path of any easy explanation of what law is. We cannot therefore profitably use, in the elucidation of law, the notion of a command which also involves it. Indeed it is a virtue of Austin’s analysis, whatever its defects, that the elements of the gunman situation are, unlike the element of authority, not themselves obscure or in need of much explanation; and hence we shall follow Austin in an attempt to build up from it the idea of law. We shall not, however, hope, as Austin did, for success, but rather to learn from our failure.

2. LAW AS COERCIVE ORDERS

Even in a complex large society, like that of a modern state, there are occasions when an official, face to face with an individual, orders him to do something. A policeman orders (p. 21) a particular motorist to stop or a particular beggar to move on. But these simple situations are not, and could not be, the standard way in which law functions, if only because no society could support the number of officials necessary to secure that every member of the society was officially and separately informed of every act which he was required to do. Instead such particularized forms of control are either exceptional or are ancillary accompaniments or reinforcements of general forms of directions which do not name, and are not addressed to, particular individuals, and do not indicate a particular act to be done. Hence the standard form even of a criminal statute (which of all the varieties of law has the closest resemblance to an order backed by threats) is general in two ways; it indicates a general type of conduct and applies to a general class of persons who are expected to see that it applies to them and to comply with it. Official individuated face-to-face directions here have a secondary place: if the primary general directions are not obeyed by a particular individual, officials may draw his attention to them and demand compliance, as a tax inspector does, or the disobedience may be officially identified and recorded and the threatened punishment imposed by a court.

Legal control is therefore primarily, though not exclusively, control by directions which are in this double sense general. This is the first feature which we must add to the simple model of the gunman if it is to reproduce for us the characteristics of law. The range of persons affected and the manner in which the range is indicated may vary with different legal systems and even different laws. In a modern state it is normally understood that, in the absence of special indications widening or narrowing the class, its general laws extend to all persons within its territorial boundaries. In canon law there is a similar understanding that normally all the members of the church are within the range of its law except when a narrower class is indicated. In all cases the range of application of a law is a question of interpretation of the particular law aided by such general understandings. It is here worth noticing that though jurists, Austin among them, sometimes speak of laws being addressed¹ to classes of persons this is misleading in (p. 22) suggesting a parallel to the face-to-face situation which really does not exist and is not intended by those who use this expression. Ordering people to do things is a form of communication and does entail actually ‘addressing’ them, i.e. attracting their attention or taking steps to attract it, but making laws for people does not. Thus the gunman by one and the same utterance, ‘Hand over those notes’, expresses his wish that the clerk should do something and actually addresses the clerk, i.e. he does what is normally sufficient to bring this expression to the clerk’s attention. If he did not do the latter but merely said the same words in an empty room, he would not have addressed the clerk at all and would not have ordered him to do anything: we might describe the situation as one where the gunman merely said the words, ‘Hand over those notes’. In this respect making laws differs from ordering people to do things, and we must allow for this difference in using this simple idea as a model for law. It may indeed be desirable that laws should as soon as may be after they are made, be brought to the attention of those to whom they apply. The legislator’s purpose in making laws would be defeated unless this were generally done, and legal systems often provide, by special rules concerning promulgation, that this shall be done. But laws may be complete as laws before this is done, and even if it is not done at all. In the absence of special rules to the contrary, laws are validly made even if those affected are left to find out for themselves what laws have been made and who are affected thereby. What is usually intended by those who speak of laws being ‘addressed’ to certain persons, is that these are the persons to whom the particular law applies, i.e. whom it requires to behave in certain ways. If we use the word ‘addressed’ here we may both fail to notice an important difference between the making of a law and giving a face-to-face order, and we may confuse the two distinct questions: ‘To whom does the law apply?’ and ‘To whom has it been published?’

Besides the introduction of the feature of generality a more fundamental change must be made in the gunman situation if we are to have a plausible model of the situation where there is law. It is true there is a sense in which the gunman has an ascendancy or superiority over the bank clerk; it lies (p. 23) in his temporary ability to make a threat, which might well be sufficient to make the bank clerk do the particular thing he is told to do. There is no other form of relationship of superiority and inferiority between the two men except this short-lived coercive one. But for the gunman’s purposes this may be enough; for the simple face-to-face order ‘Hand over those notes or I’ll shoot’ dies with the occasion. The gunman does not issue to the bank clerk (though he may to his gang of followers) standing orders to be followed time after time by classes of persons. Yet laws pre-eminently have this ‘standing’ or persistent characteristic. Hence if we are to use the notion of orders backed by threats as explaining what laws are, we must endeavour to reproduce this enduring character which laws have.

We must therefore suppose that there is a general belief on the part of those to whom the general orders apply that disobedience is likely to be followed by the execution of the threat not only on the first promulgation of the order, but continuously until the order is withdrawn or cancelled. This continuing belief in the consequences of disobedience may be said to keep the original orders alive or ‘standing’, though as we shall see later there is difficulty in analysing the persistent quality of laws in these simple terms. Of course the concurrence of many factors which could not be reproduced in the gunman situation may, in fact, be required if such a general belief in the continuing likelihood of the execution of the threat is to exist: it may be that the power to carry out threats attached to such standing orders affecting large numbers of persons could only in fact exist, and would only be thought to exist, if it was known that some considerable number of the population were prepared both themselves to obey voluntarily, i.e. independently of fear of the threat, and to co-operate in the execution of the threats on those who disobeyed.

Whatever the basis of this general belief in the likelihood of the execution of the threats, we must distinguish from it a further necessary feature which we must add to the gunman situation if it is to approximate to the settled situation in which there is law. We must suppose that, whatever the motive, most of the orders are more often obeyed than disobeyed by most of those affected. We shall call this here, following Austin, (p. 24) ‘a general habit of obedience’ and note, with him, that like many other aspects of law it is an essentially vague or imprecise notion. The question how many people must obey how many such general orders, and for how long, if there is to be law, no more admits of definite answers than the question how few hairs must a man have to be bald. Yet in this fact of general obedience lies a crucial distinction between laws and the original simple case of the gunman’s order. Mere temporary ascendancy of one person over another is naturally thought of as the polar opposite of law, with its relatively enduring and settled character, and, indeed, in most legal systems to exercise such short-term coercive power as the gunman has would constitute a criminal offence. It remains indeed to be seen whether this simple, though admittedly vague, notion of general habitual obedience to general orders backed by threats is really enough to reproduce the settled character and continuity which legal systems possess.

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. But if we are to reproduce the features of even a penal statute in our constructed model of general orders generally obeyed, something more must be said about the person who gives the orders. The legal system of a modern state is characterized by a certain kind of supremacy within its territory and independence of other systems which we have not yet reproduced in our simple model. These two notions are not as simple as they may appear, but what, on a common-sense view (which may not prove adequate) is essential to them, may be expressed as follows. English law, French law, and the law of any modern country regulates the conduct of populations inhabiting territories with fairly well-defined geographical limits. Within the territory of each country there may be many different persons or bodies of (p. 25) persons giving general orders backed by threats and receiving habitual obedience. But we should distinguish some of these persons or bodies (e.g. the LCC or a minister exercising what we term powers of delegated legislation) as subordinate lawmakers in contrast to the Queen in Parliament who is supreme. We can express this relationship in the simple terminology of habits by saying that whereas the Queen in Parliament in making laws obeys no one habitually, the subordinate lawmakers keep within limits statutorily prescribed and so may be said in making law to be agents of the Queen in Parliament. If they did not do so we should not have one system of law in England but a plurality of systems; whereas in fact just because the Queen in Parliament is supreme in relation to all within the territory in this sense and the other bodies are not, we have in England a single system in which we can distinguish a hierarchy of supreme and subordinate elements.

The same negative characterization of the Queen in Parliament, as not habitually obeying the orders of others, roughly defines the notion of independence which we use in speaking of the separate legal systems of different countries. The supreme legislature of the Soviet Union is not in the habit of obeying the Queen in Parliament, and whatever the latter enacted about Soviet affairs (though it would constitute part of the law of England) would not form part of the law of the USSR. It would do so only if the Queen in Parliament were habitually obeyed by the legislature of the USSR.

On this simple account of the matter, which we shall later have to examine critically, there must, wherever there is a legal system, be some persons 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 subordinates in obedience to the sovereign.

CHAPTER II NOTES

Page 18. The varieties of imperatives. The classification of imperatives as ‘orders’, ‘pleas’, ‘comments’, &c., which depends upon many circumstances such as the social situation and relationships of the parties and their intentions as to the use of force, is as yet a virtually untapped subject of inquiry. Most philosophical discussion of imperatives is concerned either with (1) the relationships between imperative and indicative or descriptive language and the possibilities of reducing the former to the latter (see Bohnert, ‘The Semiotic Status of Commands’, 12 Philosophy of Science (1945)), or (2) the question whether any, and if so what, deductive relationships exist between imperatives (see Hare, ‘Imperative Sentences’, 58 Mind (1949), also The Language of Morals (1952); Hofstadter and McKinsey, ‘The Logic of Imperatives’, 6 Philosophy of Science (1939); Hall, What is Value (1952), chap. 6; and Ross, ‘Imperatives and Logic’, 11 Philosophy of Science (1944)). Study of these logical questions is important; but there is also great need for a discrimination of the varieties of imperatives by reference to contextual social situations. To ask in what standard sorts of situation would the use of sentences in the grammatical imperative mood be normally classed as ‘orders’, ‘pleas’, ‘requests’, ‘commands’, ‘directions’, ‘instructions’, &c., is a method of discovering not merely facts about language, but the similarities and differences, recognized in language, between various social situations and relationships. The appreciation of these is of great importance for the study of law, morals, and sociology.

Page 18. Imperatives as the expressions of the wish that others should act or abstain from action. In characterizing in this way the standard use of the imperative mood in language, care must be taken to distinguish the case where the speaker simply reveals that he wishes another to act in a certain way, as a piece of information about himself, from the case where he speaks with the intention that the other shall be moved thereby to act as the speaker wishes. The indicative not the imperative mood would normally be appropriate in the former case (see on this distinction Hägerström, Inquiries into the Nature of Law and Morals, chap. 3, s. 4, pp. 116–26). But though it is necessary, it is not sufficient to characterize the standard use of the imperative mood that the speaker’s purpose in speaking is that the other should act in the way he wishes; for it is also necessary that the speaker should intend the person addressed to recognize that this is his purpose in speaking and to be influenced thereby to act as the speaker desires. For this complication (which is neglected in the text) see Grice, ‘Meaning’, 66 Philosophical Review (1957) and Hart, ‘Signs and Words’, 2 The Philosophical Quarterly (1952).

Page 19. The gunman situation, orders and obedience. One of the difficulties to be faced in the analysis of the general notion of an ‘imperative’ is that no word exists for what is common to orders, commands, requests, and many other varieties, i.e. the expression of intention that another should or should not do some action; similarly no single word exists for the performance of, or abstention from, such action. All the natural expressions (such as ‘orders’, ‘demands’, ‘obedience’, ‘compliance’) are coloured by the special features of the different situations in which they are normally used. Even the most colourless of these, viz. ‘telling to’ suggests some ascendancy of one party over the other. For the purpose of describing the gunman situation we have chosen the expressions ‘orders’ and ‘obedience’ since it would be perfectly natural to say of the gunman that he ordered the clerk to hand over the money, and that the clerk obeyed. It is true that the abstract nouns ‘orders’ and ‘obedience’ would not naturally be used to describe this situation, since some suggestion of authority attaches to the former and the latter is often considered a virtue. But in expounding and criticizing the theory of law as coercive orders we have used the nouns ‘orders’ and ‘obedience’ as well as the verbs ‘order’ and ‘obey’ without these implications of authority or propriety. This is a matter of convenience and does not prejudge any issue. Both Bentham (in Fragment of Government, chap. i, note to para. 12) and Austin (The Province, p. 14) use the word ‘obedience’ in this way. Bentham was aware of all the difficulties mentioned here (see Of Laws in General, 298 n.a.).

Page 20. Law as coercive orders: relation to Austin‘s doctrine. The simple model of law as coercive orders constructed in Section 2 of this chapter differs from Austin’s doctrine in The Province in the following respects.

(a) Terminology. The phrases ‘order backed by threats’ and ‘coercive orders’ are used instead of ‘command’ for the reasons given in the text.

(b) Generality of laws. Austin (op. cit., p. 19) distinguishes between ‘laws’ and ‘particular commands’ and asserts that a command is a law or rule if it ‘obliges generally to acts or forbearances of a class’. On this view a command would be a law even if it were ‘addressed’ by the sovereign to a single individual so long as it required him to do or abstain from a class or kind of action and not merely a single act or a set of different actions specified individually. In the model of a legal system constructed in the text the orders are general, both in the sense that they apply to classes of individuals and refer to classes of acts.

(c) Fear and obligation. Austin occasionally suggests that a person is bound or obliged only if he actually fears the sanction (op. cit., pp. 15 and 24, and The Lectures, Lecture 22 (5th edn.), p. 444, ‘The party is bound or obliged to do or forbear because he is obnoxious to the evil and because he fears the evil’). His main doctrine, however, seems to be that it is enough that there is the ‘smallest chance of incurring the slightest evil’ whether the person bound fears it or not (The Province, p. 16). In the model of law as coercive orders we have stipulated only that there should be a general belief that disobedience is likely to be followed by the threatened evil.

(d) Power and legal obligation. Similarly, in his analysis of command and obligation, Austin at first suggests that the author of the command must actually possess the power (be ‘able and willing’) to inflict the eventual evil; but he later weakens this requirement to the smallest chance of the smallest evil (op. cit., pp. 14, 16). See on these ambiguities in Austin’s definitions of command and obligation Hart, ‘Legal and Moral Obligation’, in Melden, Essays in Moral Philosophy (1958), and Chapter V, s. 2.

(e) Exceptions. Austin treats declaratory laws, permissive laws (e.g. repealing enactments), and imperfect laws as exceptions to his general definition of law in terms of command (op. cit., pp. 25–9). This has been disregarded in the text of this chapter.

(f) The legislature as sovereign. Austin held that in a democracy the electorate, and not their representatives in the legislature, constitute or form part of the sovereign body, though in the United Kingdom the only use made by the electorate of its sovereignty is to appoint its representatives, and to delegate to them the rest of their sovereign powers. Though he claimed that ‘accurately speaking’ this is the true position, he permitted himself to speak (as all constitutional writers do) of Parliament as possessing the sovereignty (op. cit., Lecture VI, pp. 228–35). In the text of this chapter a legislature such as Parliament is identified with the sovereign; but see Chapter IV, s. 4, for a detailed scrutiny of this aspect of Austin’s doctrine.

(g) Refinements and qualifications of Austin’s doctrine. In later chapters of this book certain ideas which have been used in defending Austin’s theory against criticisms are considered in detail, though they are not reproduced in the model constructed in this chapter. These ideas were introduced by Austin himself though, in some cases, only in a sketchy or inchoate form, anticipating doctrines of later writers such as Kelsen. They include the notion of a ‘tacit’ command (see Chapter III, s. 3, above, p. 45, and Chapter IV, s. 2, above, p. 64); nullity as a sanction (Chapter III, s. 1); the doctrine that the ‘real’ law is a rule addressed to officials requiring them to apply sanctions (Chapter III, s. 1); the electorate as an extraordinary sovereign legislature (Chapter IV, s. 4); the unity and continuity of the sovereign body (Chapter IV, s. 4, p. 76). In any assessment of Austin attention should be paid to W. L. Morison, ‘Some Myth about Positivism’, 68 Yale Law Journal, 1958, which corrects serious misunderstandings of earlier writers on Austin. See also A. Agnelli, John Austin alle origini del positi o giuridico (1959), chap. 5.

CHAPTER II 3rd ed. NOTES

Page 18. Austin’s theory. Hart makes it clear that he is discussing a simplified reconstruction of Austin’s theory (itself a simplified version of Bentham’s theory). For Austin’s own views see W. L. Morison, John Austin (Stanford University Press, 1982); and W. E. Rumble, The Thought of John Austin: Jurisprudence, Colonial Reform, and the British Constitution (Athlone Press, 1985). For a treatment of English legal thought in the period see Michael Lobban, The Common Law and English Jurisprudence 1760–1850 (Oxford University Press, 1991).

Pages 18–20. Law as imperatives. See also Neil MacCormick, ‘Legal Obligation and the Imperative Fallacy’, in A. W. B. Simpson ed., Oxford Essays in Jurisprudence, 2nd series (Oxford University Press, 1973). For the suggestion that something of an imperatival theory survives even in Hart, see G. J. Postema, ‘Law as Command: The Model of Command in Modern Jurisprudence’ (2001) 11 Philosophical Issues 470. For attempts to save aspects of an imperatival theory see Matthew H. Kramer, In Defense of Legal Positivism: Law Without Trimmings (Oxford University Press, 1999) 83–7; and Robert Ladenson, ‘In Defense of a Hobbesian Conception of Law’ (1980) 9 Philosophy and Public Affairs 134.

Pages 20–2. The generality of law. For thicker and thinner interpretations of generality compare Friedrich Hayek, Law, Legislation, and Liberty, Vol. I (University of Chicago Press, 1973), chap. 2; Lon L. Fuller, The Morality of Law (rev. edn., Yale University Press, 1969), 46 ff.; and Timothy Endicott, ‘The Generality of Law’ in Luís Duarte d’Almeida, Andrea Dolcetti, and James Edwards eds., Reading The Concept of Law (Hart Publishing, 2013).

Pages 24–5. Supremacy and Independence. See also Joseph Raz, The Concept of a Legal System (2nd edn., Oxford University Press, 1980) chap. 1.

FOOTNOTES CHAPTER II

1. ‘Addressed to the community at large’, Austin, above, p. 1 n. 4 at p. 22. ↩