Pure Theory of Law
Pure Theory of Law (German: Reine Rechtslehre) is a book by legal theorist Hans Kelsen, first published in 1934 and in a greatly expanded second edition (effectively a different book) in 1960. The second edition appeared in English translation in 1967, as Pure Theory of Law, the first edition in English translation in 1992, as Introduction to the Problems of Legal Theory. The theory proposed in this book has probably been the most influential theory of law produced during the 20th century. It is, at the least, one of the high points of modernist legal theory.
Double 'Purity' of Legal Science 
Kelsen attempted to reconstruct 'legal science (Rechtswissenschaft)' as a science of 'norms (Normen)', on analogy with Immanuel Kant's conception of natural science. This type of legal science would be 'pure' in two senses.
Firstly, it would be 'pure' in a sense analogous to that in Kant's Metaphysical Foundations of Natural Science. It would set out a priori a 'pure part' of legal science, consisting of a framework of fundamental concepts. Then, in an empirical part of legal science, this framework would be applied to empirical (sociological, historical and so on) material so as to understand that material as 'law'.
The science, in both of its parts, would also be 'pure' in a second sense, in being solely descriptive - excluding from the science any element of evaluation. The core subject matter of legal science, for Kelsen, is legal norms. He defines a 'norm' as 'the meaning of an act of will'. Thus, for Kelsen a norm does not necessarily have any element of generality - hence not all norms are rules. Indeed, as the meaning of an act of will, a norm is not intrinsically rational; departing from Kant both back toward David Hume as well as in the direction of contemporary philosophical positivism, Kelsen denies the existence of practical reason.
Legal science, as Kelsen would like it to be, has to describe a legal norm without either evaluating it or adopting it as an evaluation. This can be done by distinguishing rigorously between the social statement that is the norm itself and the legal-scientific restatement of it that is, or is contained in, a description of the norm. Kelsen terms the norm itself a 'legal norm (Rechtsnorm)' and its descriptive restatement a 'legal proposition (Rechtssatz)'. Thus, when a jurist (legal scientist) states that the law of a certain country contains the norm 'theft is to be punished with imprisonment', this legal proposition is not a comment on whether anybody in that country has a moral or other extra-legal obligation to impose, accept or evade such a penalty. In reading Kelsen, however, one has to bear in mind that, when he refers to a 'norm', he is usually speaking elliptically: he is uttering a legal proposition which stands for a legal norm, in much the same way that a picture of a dog shows the dog but does not bark.
This kind of 'purity' already excludes ideas of natural law, in which description and evaluation are intimately connected. That leaves as the subject matter of legal science only positive law. Kelsen is a 'legal positivist' both in that he excludes evaluation from legal science, through a strict separation of 'is' and 'ought', and in that he considers only positive law to be law. Although he never made much sense of the idea of 'custom', he included all public international law as 'law'.
At the same time, although it is not easy to reconcile with his rejection of practical reason, Kelsen was emphatic that politicians should support and make use of legal science. And legal scientists should assist them, for instance in law reform. Although if, when doing so, they go beyond the role of mere technician and recommend policy-based choices between possible norms, they will no longer be acting as legal scientists. For Kelsen, 'is' and 'ought' occupy separate worlds as a matter of logic, but not in practice. Their logical separation should be incorporated into practice.
Kelsen insisted that, if social scientists wished to deal with law, they would have to take their concept of law from legal science - that is, from the Pure Theory of Law. Kelsen made this point particularly against Max Weber, whose philosophical orientation and legal background were similar to his own. And, indeed, although Weber had said much about what is socially called 'law', he had not developed a clear concept of it himself.
'Legal Orders' 
For Kelsen as for other central European contemporaries, norms occur not singly but in sets, termed 'orders'. The ordering principle of an order of moral norms - and of an order of natural law, if one could exist - would be logical, as deduction. From the general norm 'do not kill other human beings', it follows deductively that A must not kill any other human being. Kelsen calls this a 'static' order. An order of positive law, he maintains, is not, or not centrally, like that. Although it forms a hierarchy, it is 'dynamic', in that its ordering principle is authorisation. Each relatively 'higher' norm authorises someone (an individual or an organ, primarily of the state) to create further and relatively 'lower' norms. In this is found the specific characteristic of positive law, hence of all law, that it regulates its own creation. Even though positive legal norms do commonly contain elements of generality, those elements are not central to the relation between a higher and a lower norm. This is a relation of legal 'validity' (not to be confused with logical validity), which is that the creation of the lower norm has been authorised by the higher norm. Since this moment of creation always involves extraneous considerations, Kelsen does not need to supplement his model with a concept of 'discretion'.
What has to be consistent in a dynamic order of norms is only the process of authorising. This process can (and perhaps always does) result in multiple and mutually uneven strands of hierarchy, such as legislative, executive and judicial strands. In each strand, at every point in the process of norm-creation, many types of extra-legal factor may be taken into account - moral, political, economic and others. Kelsen notes that, in this respect, the Pure Theory of Law has an affinity with American 'legal realism'.
At the same time, the 'hierarchy' model does not readily fit a common-law system. It takes as its paradigm a single constitutional document under which codes and statutes are enacted, with both administration and adjudication in subordinate places. However, at least outside the USA, common-law systems are moving in the direction of that paradigm. The 'hierarchy' model also fits public law more readily than private law. It is difficult to be convinced by Kelsen's contention that, in private transactions, the parties are exercising legal authorisation.
'Basic Norm (Grundnorm)' 
Kelsen assumes, however, that the scientific representation of a positive-legal order, as a hierarchy of legal propositions, must have a guarantor of unity. This guarantor cannot be other than a component of the representation, hence a legal proposition. Being a legal proposition, it counts as a representation of an actual norm. So Kelsen calls it, elliptically, a 'basic norm (Grundnorm)'. A basic norm is 'presupposed' in legal science for each order of positive law, to make it possible to understand that material as an order of positive law. This norm is simply that 'the historically first constitution is to be obeyed'. That constitution may have become established by custom or by revolution: the jurist does not evaluate the circumstances.
Kelsen was always aware that a 'constitution' (Verfassung) might or might not be a single document. In some legal orders, indeed, it can be difficult to draw a line between 'constitutional' and other norms. In 1964, Kelsen relativised the idea of 'constitution', holding that each relatively higher legal norm is a 'constitution' in relation to any norm created under its authority. This need not, however, wholly supersede a narrower sense of 'constitution', which identifies those relatively 'high' legal norms that can be changed only through a specially onerous process such as a large parliamentary majority or a popular referendum.
To presuppose a basic norm for a particular positive-legal order does not involve any moral or other extra-legal evaluation of the basic norm's requirement of obedience. The jurist simply supposes that such a requirement has been stated. The jurist does so, irrespective of whether any such norm can be empirically discovered. At this point, serious difficulties emerge - with which Kelsen wrestled for much of his long life.
One part of the problem lay in his starting point, the question 'Why should the law be obeyed?'. He gives the idea of a basic norm two roles. In one role, it is the guarantor of unity in legal-scientific representation of a positive-legal order. In another role, it is an answer, for that legal order, to the obedience question - put in the more clearly descriptive form, 'Why is the law binding?'. One might think, rather, that to suppose for a particular positive-legal order simply that the historically first constitution (identified on this occasion) 'is to be obeyed' is instead to assume that a reason for obedience has already been established outside the legal order and beyond the ken of legal science. In that case, legal science would be seriously inadequate. On the other hand, if the reason for the bindingness of a legal norm is intrinsic to its 'legal' character, that reason might have to be sought in the 'legal' form of the individual norm and not in any kind of further guarantee.
Another part of the problem lay in the requirement that the 'basic' factor be, or count as, a norm. Early on, Kelsen may have supposed that there could be such a real norm. Later, however, the 'basic norm' was relegated to being only a 'hypothesis'. In a philosophical light, nonetheless, it had the status - in Kantian terms - of a 'transcendental-logical presupposition', a logical condition for the possibility of a certain kind of knowledge.
Eventually, Kelsen accepted the criticism that the concept of a basic norm involves an infinite regress. If the basic norm of a particular positive-legal order is an authorising norm, one can ask what is the source of the authority of the basic norm itself. If 'is' and 'ought' are to remain entirely separated, that can only be a still more 'basic' norm; and then the same question could be asked again, and so on infinitely. At this point, Kelsen took a further step back, accepting that the concept of a basic norm is doubly contradictory: it is self-contradictory in that it involves an infinite regress; in addition, since no such norm actually exists, to presuppose the existence of such a norm contradicts reality. The concept is a full 'fiction' in the sense to be found in Hans Vaihinger's 'philosophy of "as if"'. But then not much remains of it, although Kelsen evidently considered that what remained was vital.
And there Kelsen ended, committed to an epistemology that was so radically relativistic as to challenge the very idea of authorisation. Nonetheless, if Kelsen does not provide an effective answer to the obedience question, he has posed that question - and many others - with exceptional sophistication.
Metaphysics and Persons 
Ideas of natural law are also excluded by Kelsen's rejection of 'metaphysics' in a strong sense, supposing something supra-human. He similarly rejects duplication in ideas of 'the state' as a supra-human entity - rather than a mere, fictional personification of the national legal order. His main target, in the 1930s, was fascist theories of state and law, such as that of Carl Schmitt. But, for Kelsen, it is also an illusion to think that, in a Rechtsstaat (roughly, 'rule of law state'), 'the state' is wholly subject to law: from a 'legal point of view', the state and the positive-legal order are the same. It is likewise an illusion to think of a 'legal person', whether relating in fact to a number of real people or to a single real individual, as anything other than a fictional personification of a bundle of norms that are rights and duties pertaining to such people. This is simply so, however, only in the 'pure part' of legal science. Kelsen is perfectly well aware that, in the empirical part, 'the state' will turn out to be a heavily armed real apparatus. One can make a similar extrapolation for the reality of legal persons, where corporate legal persons would turn out to be organisations and individual legal persons to have gender, class, ethnicity and so on.
Law and Power 
Descriptively, the Pure Theory of Law has always been about power. A positive-legal order can be identified when there are norms that regulate their own creation through a process of authorisation and, in addition, the order itself is 'by and large effective'. Kelsen does not closely specify 'by and large', but he does attribute effectiveness to coercion. In this respect, there is some affinity with John Austin's 'command' theory of law. But Kelsen's approach is quite different from Austin's. Austin finds the defining characteristic of law on the plane of the signifier, in imperative utterances. Kelsen has almost no interest in the signifier but focuses on the signified: a 'norm' is an utterance that, whatever its grammatical form, has the meaning 'ought (Sollen)'.
Austin and Kelsen also differ in that, for Kelsen, legal norms are addressed primarily to officials. Officials are directed to apply sanctions to individuals when the individuals' behaviour does not conform to a pattern specified in the norm. It is anticipated that individuals, in choosing how they will behave, will take into account the possibility that an official will apply the sanction. In that way, norms and the orders to which they belong can be effective. However, most if not all jurists now accept H.L.A. Hart's point in The Concept of Law, though directed principally against Austin, that not all legal norms are coercive - some, and some of the most important, are facilitative. The Pure Theory of Law can accommodate this by accepting that what matters most is not whether particular legal norms are coercive but whether, by containing coercive norms, the legal order as a whole is coercive - which Hart does not deny.
International and National Law 
Identification of law as characteristically coercive is more difficult with public international law, but Kelsen finds sufficient sanctions.
As to the relation between public international law and national law, however, his position changed dramatically. In the first edition of Reine Rechtslehre, 1934, like many liberal contemporaries he was an internationalist monist: he looked to public international law as global law, subsuming all national or state legal orders. This had the happy consequence that, in its contemporary application, the universality of the Pure Theory (and Kelsen insisted that only universal truths could be 'scientific') would coincide with that of predominantly global law.
By the end of World War II, however, he gave up on this aspiration as unrealistic and, with evident reluctance, accepted a sort of optional monism. He found equally tenable the view that the public international legal order is supremely valid and the view that each national or state legal order is - for itself, in a kind of 'solipsism' - supremely valid. The story of the 'basic norm' might be read as a battle to come to terms with the 'solipsist' option. Kelsen needed not to adopt such solipsism but to provide an account of it.
Toward a General Theory of Norms 
From the mid-1960s until his death in 1973, Kelsen branched out into general theory of norms. In a longer view, it was a return to his youthful passion for philosophy. His notes on this were published posthumously in 1979 as Allgemeine Theorie der Normen and appeared in English translation in 1990 as General Theory of Norms. This work, however, neither offers a completed general theory of norms nor goes far toward locating the Pure Theory of Law within a general theory of norms.
- Both editions will be included in forthcoming volumes in the Hans Kelsen Werke. A fuller and more accurate translation of the second edition is also planned. The current translation, in omitting many footnotes, obscures the extent to which the Pure Theory of Law is both philosophically grounded and responsive to earlier theories of law.
- To suppose, as is still sometimes done, that Kelsen's title Reine Rechtslehre should be translated into English not as 'Pure Theory of Law' but as 'Theory of Pure Law' is therefore wrong, conceptually as well as grammatically. Grammatically, Reine Rechtslehre translates exactly as 'Pure Theory of Law' - 'Theory of Pure Law' would have been Lehre des reinen Rechts. This theory of law is a pure theory; it is not a theory of pure law.
- Translators into English have not always served Kelsen well on this. One translation renders Rechtssatz as 'rule of law in a descriptive sense'. Yet one can have sympathy for the translators; though Kelsen's reasoning is rigorous, his style cannot be described as limpid.
- The hierarchy does not have to be, and probably will not be, even-sided, like a pyramid - although that image has been used by one of Kelsen's French translators. Reine Rechtslehre, 2nd ed sec. 35, refers to Stufenbau, which the French translation renders as pyramide while the English translation has 'hierarchical structure'. 'Steps and stairs' has also been suggested, but it might be best not to attempt an image.
- It is a 'basis (Grund)' in a logical sense. Once again, it might be best not to attempt an image.
Selected Works by Kelsen 
- Reine Rechtslehre, Vienna 1934; 2nd ed 1960.
- Introduction to the Problems of Legal Theory (1934; Litschewski Paulson and Paulson trans.), Oxford 1992; the translators have adopted the subtitle of the first edition, Einleitung in die rechtswissenschaftliche Problematik, in order to avoid confusion with the English translation of the second edition.
- Law and Peace in International Relations, Cambridge (Mass.) 1942, Union (N.J.) 1997.
- Peace Through Law, Chapel Hill 1944, Union (N.J.) 2000.
- Pure Theory of Law (1960; Knight trans.), Berkeley 1967, Union (N.J.) 2002.
- Théorie pure du droit (1960; Eisenmann trans.), Paris 1962.
- General Theory of Law and State (German original not published; Wedberg trans.), 1945, New York 1961, Clark (N.J.) 2007.
- What is Justice?, Berkeley 1957.
- 'The Function of a Constitution' (1964; Stewart trans.) in Richard Tur and William Twining (eds), Essays on Kelsen, Oxford 1986; also in 5th and later editions of Lloyd's Introduction to Jurisprudence, London (currently 8th ed 2008).
- Essays in Legal and Moral Philosophy (Weinberger sel., Heath trans.), Dordrecht 1973.
- Allgemeine Theorie der Normen (ed. Ringhofer and Walter), Vienna 1979.
- General Theory of Norms (1979; Hartney trans.), Oxford 1990.
Selected Works in English on Kelsen 
- Uta Bindreiter, Why Grundnorm? A Treatise on the Implications of Kelsen's Doctrine, The Hague 2002.
- David Dyzenhaus, Legality and Legitimacy: Carl Schmitt, Hans Kelsen and Hermann Heller in Weimar, Oxford 1977.
- William Ebenstein, The Pure Theory of Law, 1945; New York 1969.
- Ronald Moore, Legal Norms and Legal Science: a Critical Study of Hans Kelsen's Pure Theory of Law, Honolulu 1978.
- Stanley L. Paulson and Bonnie Litschewski Paulson (eds), Normativity and Norms: Critical Perspectives on Kelsenian Themes, Oxford 1998.
- Iain Stewart, 'The Critical Legal Science of Hans Kelsen' (1990) 17 Journal of Law and Society 273-308.
- Richard Tur and William Twining (eds), Essays on Kelsen, Oxford 1986.
- Lars Vinx, Hans Kelsen's Pure Theory of Law, Oxford 2007.
'The Pure Theory of Law' , Stanford Encyclopedia of Philosophy.