Hans Kelsen's Pure Theory of Law

Audio version created with Paper2Audio.

Listen on Paper2Audio

Chapter Four

Hans Kelsen (1881 to 1973) was a prolific and influential Austrian legal theorist, who spent the last decades of a long, productive life in the United States, having escaped Europe at the time of Hitler's rise to power. Kelsen's work is important in international law as well as jurisprudence, and he was a central figure in the drafting of the Austrian constitution after the First World War.

Over the course of five decades of jurisprudential writing, Kelsen published dozens of books and articles, with his position on various matters changing in ways both large and small. This presents a difficulty for any attempted summary of Kelsen's view, a task already complicated by the sophistication of Kelsen's theory and the unfamiliarity (to American and English audiences at least) of the philosophical traditions within which Kelsen was working, in particular, neo-Kantianism, a school of thought that attempted to apply Immanuel Kant's ideas more broadly to questions of social and ethical theory.

The picture of Kelsen's theory presented will attempt to contain the general themes that continued throughout most of his writings, while there will be little attention paid to the ways in which Kelsen's views changed. Inevitably, the portrayal will be something of a simplification relative both to the full complexity of Kelsen's theory and to its transformation over time.

The Pure Theory of Law

Kelsen referred to his theory as “reine Rechtslehre”, a “pure theory of law”. In Kelsen's words, the theory was “pure” “because it only describes the law and attempts to eliminate from the object of this description everything that is not strictly 'law'”. Moral judgments, political biases, and sociological conclusions were all to be pushed aside, as improper for a “scientific” description of the social institution of law.

Chapter 3 noted the importance of the normative aspect of law for H.

50. A. Hart's legal theory (it is central to the “internal aspect of rules”, which in turn is central to Hart's theory and how it differs from empirical theories like that of John Austin). Within Kelsen's theory, the normativity of law is, if anything, an even more central and dominating factor. Most of what is puzzling to readers of Kelsen's legal theory can be better understood if one keeps in mind the theory's focus on normativity.

There are two basic starting points for understanding Kelsen's approach to legal theory. First, normative claims—arguments for how one ought to act or for how things ought to be—can be grounded only on (justified by) other normative claims. This is the argument, usually attributed to David Hume, that one cannot derive a normative conclusion from purely factual premises: it is sometimes phrased, “one cannot derive an 'ought' from an 'is'”. In other words, a purely factual description of a situation will never be sufficient, by itself, to justify a conclusion that something ought (morally) to be done. One can only justify such a conclusion by first accepting or inserting a moral premise.

Secondly, such lines of justification must necessarily come to an end at some point. In day-to-day discussions, each (normative) argument put forward is based on (justified by) some more general or more basic argument. We tend to forget that if we look closely enough at the chain of arguments in favour of a particular position, we will eventually come to an argument that is not justified by some other argument, and the validity of this final argument can only be based on its being tacitly or explicitly accepted (accepted “on faith”, as it were).

Consider the following example. A religious person tells you that it is wrong to commit adultery. When you ask her why, she says, “because that is what is said in our sacred book.” Being in an obstinate mood, you say “So what?”, to which her response is that the sacred book is the revealed word of God. To a second “So what?”, her patient response would be that we should all do as God tells us to do. However, if at this point you ask why that is so, you are likely to get no more than a puzzled (or hostile) look. This line of argument has come to an end; either one accepts that one ought to do what God says or one does not. And there is a sense in which the foundational argument, “we ought to do what God says”, is entailed by or implied by the religious person's initial assertion that “one should not commit adultery”. (This is not to say that one could not reach the same normative conclusion using other starting points, but only that, for this person, this conclusion derives from or implies that starting point.)

Kelsen's argument was that there is a foundational argument implied (“presupposed”) by legal statements just as there is a foundational argument implied by religious statements. In more technical language, Kelsen applied a “neo-Kantian” approach to legal theory, an approach based on aspects of Kant's theory of knowledge, in particular Kant's Transcendental Argument.

The best way to understand Kelsen's project may be to think of him as asking: “What follows from the fact that some people treat legal rules as valid norms?” Like many important philosophers, Kelsen tried to show what is interesting or paradoxical about matters that seem ordinary and unremarkable to us. For Kelsen, the ordinary and unremarkable fact to be considered is that while looking at a simple collection of actions, we sometimes see those actions as normative. Whenever one looks at people putting slips of paper into a box, and sees “voting”; or looks at an organised group of people raising and lowering their hands in various sequences, and sees “the passage of valid legislation”, this translates empirical actions into normative meanings. The translation is clearer on the occasions when someone says that since those certain actions have been done (the group of people raising and lowering their hands), one now “ought” to do something (for example pay a certain tax). The border between “is” and “ought” has seemingly been crossed, and the question is: what can be learned from that?

Here we need to return to the idea of the normative chain of justification. One starts with some simple legal-normative statement: for example, “one cannot park here (it is illegal to do so) ”. If the person making this statement was asked why it was so, she would probably note that this regulation was validly promulgated by some city council, judge or administrator. If the questioner pushes further, the chain could be followed back, for example that the administrator was authorised to act in this area by an act of the legislature, and the act of the legislature was passed according to the procedures set down in the constitution. Things get slightly trickier when one gets to the constitution itself. The document might itself have been a modification of an earlier basic law, or it might have been drawn up under the authorisation of an earlier basic law. However, again, we will eventually come to a point either so foundational or so early in the society's legal history that one cannot go any further back, and no further justification can be offered.

Following the whole chain through then leads to the following implication: to assert the (normative) validity of the individual legal rule (“one cannot park on this street”) is implicitly to affirm the validity of the foundational link of the chain (for example “one ought to do whatever Parliament orders”), for the same reason that affirming an individual religious belief implicitly affirms the foundational norm of the religion (“one ought to do whatever God commands”). To put the matter differently, the affirmation of the foundational norm is “presupposed” by any express or implied affirmation of individual legal rules. This foundational norm of a legal system (“one ought to do whatever is authorised by the historically first constitution”) is what Kelsen calls the “Grundnorm” or “Basic Norm” of the legal system.

For Kelsen, analysis and justification within law thus occurs at both a “static” and a “dynamic” level. It is static in the way that legal norms are justified by more general or more basic norms. It is dynamic in the way that more basic norms authorise the (later) creation of more specific norms by legal officials. This hierarchical view of law and legal validity is sometimes known by the German, “Stufenbaulehre” (“step structure doctrine”), and was adopted by Kelsen from the work of another Austrian theorist, Adolf Julius Merkl.

Reduction and Legal Theory

Hans Kelsen believed that all legal norms could and should be understood in terms of an authorisation to an official to impose sanctions: if A (citizen) does X (wrong action), then B (an official) is authorised to impose Y (a sanction).

Thus, Kelsen would want us to translate “you shall not murder”, into the following instruction to an official: if any citizen murders, you (the official) have the authority to impose a sanction upon that person. As the instruction to the official is only an authorisation, one might wonder how Kelsen can explain the fact that officials are bound to impose sanctions—it is not usually just a matter within their discretion. Kelsen would say that where officials have an obligation to act, this only means that there is another norm, instructing a higher official to the effect: “if the lower official does not impose a sanction in this situation, you are authorised to impose a sanction on that official”—and so it would go up the hierarchy.

This is a slightly awkward formalisation of criminal laws as it stands, but its awkwardness becomes far greater when we try to put civil laws, in particular laws which confer powers, into the same form. For example, a statute authorising the formation of wills might read: if A creates a valid will (by following certain procedural and substantive requirements), and then dies, and A's executor refuses to follow the instructions of the will, then B (an official) has the authority to impose a sanction on A's executor.

Reduction is the natural tendency whenever one posits a theory or a model of behaviour. In some ways, it is the essence of the activity. To the extent that one can discuss a complex social phenomenon, like law, in terms of one or two concepts, the process of theorising seems to be a success. There is no point in a theory that merely replicates the complexity of the phenomena about us. That gives us nothing. An explanation is necessarily an attempt to explain an activity in other, simpler terms.

There is something satisfying about being able to say something like “law is basically or essentially . . .” (where the blank might be filled in by “orders backed by threats”, “authorisations to officials to impose sanctions”, or “plans”). To understand the essence of something has always been considered the core of wisdom, so we tend to welcome the opportunity, when a theorist tells us that she has “discovered” what the essence is of law (or government or property or marriage, etcetera).

On the other hand, simplification is often distortion. The more one tries to re-characterise the variety of experience as though it were homogenous, the more awkward and inaccurate the description will be. All social theorists (economists and anthropologists as well as legal and political theorists) must consider the proper balance between descriptive accuracy and explanatory power. (It is a problem that is particularly significant in understanding the limitations of the law and economics movement, which will be discussed in Ch. 18). Kelsen's theory lies towards an extreme in reduction: an attempt to reduce all laws to a particular form. However, as H. L. A. Hart pointed out when discussing John Austin's approach to law, while such reductions seem to have the benefit of simplicity, this benefit is largely a surface matter, as the likely consequences of trying to force the various legal norms into a single structure are awkwardness, poor fit and a risk of misleading the reader.

Hart versus Kelsen Perhaps because of the limited dialogue between (or overlap in) H. L. A. Hart scholars and Hans Kelsen scholars, the differences between Hart and Kelsen are frequently poorly understood. Sometimes Kelsen is seen as an imperfect stopping point between Austin's mistaken views and Hart's solutions (a position that does not stand up long under close examination). One text stated that Hart is merely Kelsen in clearer prose. Even if this is meant to be complimentary to Kelsen, it does a disservice to both theorists. This section will briefly discuss some of the things that join and separate the two writers.

There is one question that theorists who focus on the normativity of law—and Hart as well as Kelsen would fit into this category—could be said to be trying to answer: how is a legal system to be distinguished from the orders of gangsters? For Hart, this question led to an investigation of the differences in action and attitude between how we act when we are following a rule and how we act when we are being compelled to do the same action. This, in turn, led to Hart's discussion of the “internal aspect” of rules and of law, which is basic to his approach to legal theory.

Kelsen's response to the gangster/law question would be simple: those who see the actions of the people in power in a normative way, and thus presuppose the Basic Norm in dealing with official promulgations, see the officials as legitimate authorities; those who do not see the actions this way will see the people in power as gangsters or their equivalent. In a sense, Kelsen's response is comparable to Hart's: the difference between the commands of valid law and the orders of gangsters is determined by, indeed is constituted by, the attitudes of the citizen subjects.

Here, we also see how legal positivism links Hart and Kelsen: both analyse the difference between gangsters and legitimate government by focusing on the more or less “neutral” question of citizens' reactions. Hart and Kelsen's positions avoid making moral judgments. They pass by the more obvious answer to the gangster/law question, which would be quickly given by a natural law theorist: that the difference between legitimate leaders and gangsters is that the former act justly and for the common good, and the latter do not.

The differences between Hart and Kelsen are also interesting and significant. While both Hart and Kelsen emphasised the normative aspect of law in response to and criticism of more reductive/empirical approaches, their notion of the “normative” differed. Hart's view of the normative reduced it to a combination of certain types of social facts, while Kelsen resisted any reduction of “normative” to facts—that is, he denyed that statements of what one ought to do (for example legally or morally) could be translated without loss into some descriptions of facts about the world. While Hart's theory tried to track and explain actual social practices (with labelling of the work as “descriptive sociology”, and the careful distinctions, for example feeling obliged versus having an obligation, acting out of habit versus following a rule, and the different kinds of rules), Kelsen's theory tended to be more abstract—appropriate for what purported to be a “pure theory” and a neo-Kantian analysis.

The most obvious differences may be ones of methodology, which have been hinted at in passing over the course of this chapter and the previous one. Hart's analysis builds on close attention to actual practices (and how they are perceived by their participants) and linguistic usage. On the other side, Kelsen is offering a kind of logical analysis of law and of normative thinking in general.

There are also interesting similarities and differences between Kelsen's Basic Norm and Hart's rule of recognition. Both the rule of recognition and the Basic Norm rest on the idea of chains of normative validity: a particular legal norm is only valid because it has been authorised by a more general or more basic legal norm. This chain of validity must end somewhere, with a foundational norm that carries no further justification, other than its “acceptance” or its having been “presupposed”. It is again important to note the difference of approach and methodology here: Hart's theory is meant as an analytical description of actual practices, while Kelsen sought a theory purified even of sociological observation, and is best understood as a neo-Kantian transcendental deduction from the fact that we treat certain rules as legal norms.

Both the idea of a (single) rule of recognition and a (single) Basic Norm derive from assumptions that societies' legal regulations occur or are viewed as occurring in a systematic way—all the norms fitting within a consistent, hierarchical structure of justification. If one does not think that legal systems must be systematic in this way, then one could conclude that there could be more than one rule of recognition or more than one Basic Norm.

on the Nature of Norms In his last works, Kelsen became caught up in questions regarding the nature of norms. Analysis in (metaphorical) terms of one norm “justifying” or “generating” another, and inquiries regarding whether a legal system can contain norms with contradictory contents, seemed to create a confusion in Kelsen, “between a norm as a kind of sentence or sentence-meaning and as a contingent entity created and repealed by certain social events.” Arguably, this line of inquiry was what was behind many of the changes in his theory over time, as well as some of the stranger notions of the later work. The problem of the “existence”, “meaning” and “logic” of norms remains a difficult one for all theorists, though H. L. A. Hart's work (see Ch. 3) deals with those challenges much less awkwardly than Kelsen's last works do.

Suggested Further Reading

Brian H. Bix, “Rules and Normativity in Law”, in Tomasz Gizbert-Studnicki, Krzysztof Pieszka, Michał Arszkiewicz & Paweł Banas eds., Problems of Normativity, Rules and Rule Following (Springer, 2015), pp. 125–146.

Luís Duarte d'Almeida, John Gardner and Leslie Green eds., Kelsen Revisited: New Essays on the Pure Theory of Law (Hart Publishing, Oxford, 2013).

Michael Steven Green, “Hans Kelsen and the Logic of Legal Systems”, 54 Alabama Law Review 365 (2003).

Hans Kelsen, General Theory of Norms (M. Hartney trans., Clarendon Press, Oxford, 1991).

—, Introduction to the Problems of Legal Theory (B. L. Paulson and S. L. Paulson trans., Clarendon Press, Oxford, 1992).

—, Pure Theory of Law (M. Knight trans., University of California Press, California, 1967).

Andrei Marmor, “The Pure Theory of Law”, in Edward N. Zalta ed., Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/lawphil-theory/ (2016).

Stanley L. Paulson, “Continental Normativism and Its British Counterpart: How Different Are They?”, 6 Ratio Juris 227 (1993).

—, “A ‘Justified Normativity’ Thesis in Hans Kelsen’s Pure Theory of Law?: Rejoinders to Robert Alexy and Joseph Raz,” in Matthias Klatt ed., Institutionalized Reason: The Jurisprudence of Robert Alexy (Oxford University Press, Oxford, 2012), pp. 61–111.

—, “The Neo-Kantian Dimension of Kelsen’s Pure Theory of Law”, 12 Oxford Journal of Legal Studies 311 (1992).

Stanley L. Paulson and Bonnie Litschewski Paulson eds., Normativity and Norms: Critical Perspectives on Kelsenian Themes (Clarendon Press, Oxford, 1998) (includes contributions by H. L. A. Hart, Alf Ross, Carlos S. Nino, Joseph Raz, Neil MacCormick, and Georg Henrik von Wright).

Jeremy Telman ed., Hans Kelsen in America – Selective Affinities and the Mysteries of Academic Influence (Springer, Dordrecht, 2016).

Richard Tur and William Twining eds., Essays on Kelsen (Clarendon Press, Oxford, 1986) (includes contributions by Joseph Raz, Ota Weinberger, J. W. Harris, and Stanley L. Paulson).

“Dozens” is actually a bit of an understatement. By one count, Kelsen published 387 separate works (approximately 100 of which dealt exclusively with legal theory). Of those 387 works, 18 books and 121 articles were available in English as of 1991. Michael Hartney, “Appendix: Bibliography of Kelsen's Publications in English”, in Hans Kelsen, General Theory of Norms, pages 440 to 454.

2 Especially if one takes into account Kelsen's very last writings, which were unpublished during his lifetime, some of his changes in ideas and attitude were quite dramatic, at times appearing to support just the set of views he had most vigorously opposed earlier. See Stanley L. Paulson,"Kelsen's Legal Theory: The Final Round", Oxford Journal of Legal Studies at 265 to 266 (1992); Neil Duxbury,"Kelsen's Endgame", Cambridge Law Journal (2008). On the different"phases" of Kelsen's work, see Hartney,"Introduction", in Kelsen, General Theory of Norms, pages 20 to 53; Stanley L. Paulson,"Metamorphosis in Hans Kelsen's Legal Philosophy", Modern Law Review (2017).

Because Kelsen wrote much of his important work in German (and until recently was poorly served by his English translators) and because he wrote out of a different philosophical tradition (a continental tradition strongly influenced by Kant), his work has not been as central to the development of English-language legal theory as might have been warranted (in many non-English-speaking countries, Kelsen is far better known and far more influential than Hart). For the above reasons, I used Hart rather than Kelsen to introduce the topic of modern legal positivism, even though many of Kelsen's works were published long before the publication of Hart's most important works.

The last, “sceptical” phase of Kelsen's work will be discussed briefly at the end of the chapter. This chapter will not discuss Kelsen's early “critical constructivist” phase. On that topic, see, for example Stanley L. Paulson, “Four Phases in Hans Kelsen's Legal Theory? Reflections on a Periodization”, Oxford Journal of Legal Studies at 156 to 157 (1998).

Hans Kelsen, The Pure Theory of Law, p. 1; see also Stanley L. Paulson, “The Purity Thesis”, Ratio Juris (2018).

One should not over-read Kelsen's talks about a "science" of law. Here, "science" is the translation of the German "Wissenschaft", whose meaning and application generally is broader than the English "science". For example, it is usual and uncontroversial in German to use the term Wissenschaft when referring to literary theory, where in English one would not label it as a "science".

7 See Kelsen, The Pure Theory of Law, pages 4 to 10.

8 See David Hume, A Treatise of Human Nature (2nd ed., with text revised and notes by P. H. Nidditch, Clarendon Press, Oxford, 1978), Section 3.1.1, pages 469 to 470.

However, the moral premise in question can be “obvious” or something “everyone agrees with”. One should note that there are some philosophers who contest the general view that one cannot derive an “ought” proposition from an “is” proposition. That is a complicated debate; for present purposes, one need not only that Kelsen's approach to law is grounded on the more conventional view that such a derivation is not possible.

10 See Kelsen, The Pure Theory of Law, pages 193 to 195.

That the argument could be stretched a step or two further does not alter the basic analysis. For example, the religious person could say, “one ought to do what God says because He created

humanity and all the world", with the implied claim that one ought to obey the entity that created us. However, there is no particular reason why everyone must accept that normative position.

12 See, e.g. Kelsen, The Pure Theory of Law, pp. 201–205; Stanley L. Paulson, “The Neo-Kantian Dimension of Kelsen’s Pure Theory of Law”, 12 Oxford Journal of Legal Studies 311 (1992). A transcendental argument is one:

“that elucidates the condition for the possibility of some fundamental phenomenon whose existence is unchallenged or uncontroversial . . . Such an argument proceeds deductively, from a premise asserting the existence of some basic phenomenon (such as meaningful discourse . . .), to a conclusion asserting the existence of some interesting, substantive enabling conditions for that phenomenon.”

Robert Audi ed., The Cambridge Dictionary of Philosophy (2nd ed., Cambridge University Press, Cambridge, 1999), p. 925 (entry on “transcendental argument”).

13 See Paulson, “The Neo-Kantian Dimension of Kelsen’s Pure Theory of Law”, p. 324.

14 See, e.g. Hans Kelsen, Introduction to the Problems of Legal Theory (Bonnie L. Paulson and Stanley L. Paulson trans., Clarendon Press, Oxford, 1992), pp. 6–12.

There are complications for Kelsen's argument when an official acts within her area of delegated power, but acts in an unauthorised (illegal) way. See, e.g. Stanley L. Paulson, "Material and Formal Authorisation in Kelsen's Pure Theory", 39 Cambridge Law Journal 172 (1980).

See, e.g. Kelsen, Introduction to the Problems of Legal Theory, pp. 56–60. Neil MacCormick helpfully adds: “[A]lthough a basic norm is indeed presupposed in the case of any working constitution, it is not a mere presupposition. It is the content of a living custom, a convention that can be articulated as an explicit norm.” Neil MacCormick, Institutions of Law (Oxford University Press, Oxford, 2007), p. 288. Kelsen introduced his notion of the Basic Norm in 1914, adapting ideas from Immanuel Kant and Rudolf Stammler. Stanley L. Paulson, “A Great puzzle: Kelsen’s Basic Norm”, in Luís Duarte d’Almeida, John Gardner and Leslie Green eds., Kelsen Revisited: New Essays on the Pure Theory of Law (Hart Publishing, Oxford, 2013), pp. 43–61 at pp. 45–46.

17 See, e.g. Kelsen, The Pure Theory of Law, pp. 108–278.

18 See András Jakab, “Problems of the Stufenbaulehre: Kelsen’s Failure to Derive the Validity of a Norm from Another Norm”, 20 Canadian Journal of Law and Jurisprudence 35 (2007); Stanley L. Paulson, “How Merkl’s Stufenbaulehre Informs Kelsen’s Concept of Law”, 21 Revus 29 (2013).

Kelsen's actual terminology is that the official "ought" to impose the sanction, but Kelsen uses the word "ought" broadly, in a sense which is best summarised as "authorised to" rather than "should perform". See Hans Kelsen, "On the Basis of Legal Validity", 26 American Journal of Jurisprudence 178 at 178–179 n. b (1981) (Stanley L. Paulson trans.) (translator's note on Kelsen's use of "bestimmen" and "sollen").

20 See, e.g. Kelsen, Pure Theory of Law, pp. 114–130.

Kelsen is neither the first nor the last theorist to make such an attempt. See, e.g., J. W. Harris, Law and Legal Science (Clarendon Press, Oxford, 1979) (attempting to analyse all laws in terms of duties).

$ ^{22} $ Hart, The Concept of Law, pp. 27–42.

23 Jeffrey G. Murphy and Jules L. Coleman, Philosophy of Law (revised ed., Westview Press,

1990), p. 27.

One can find Hart's comments on Kelsen in H. L. A. Hart, Essays in Jurisprudence and Philosophy (Clarendon Press, Oxford, 1983), pp. 286–342. However, one often gets the impression that Hart did not entirely understand Kelsen's work, perhaps because Kelsen's starting point was so different from the Anglo-American tradition within which Hart wrote.

25 See Hart, The Concept of Law, pp. 6–7, 20–25, 82–83; Kelsen, Pure Theory of Law, p. 8.

Stanley L. Paulson, “Continental Normativism and Its British Counterpart: How Different Are They?”, 6 Ratio Juris 227 (1993).

27 Paulson, “Continental Normativism and Its British Counterpart: How Different Are They?” at 236.

$ ^{28} $ Hart, The Concept of Law, pp. 100–110.

29 See Kelsen, Introduction to the Problems of Legal Theory, p. 59.

30 See, e.g. Paulson, “The Neo-Kantian Dimension of Kelsen’s Pure Theory of Law”.

31 See Joseph Raz, The Authority of Law (2nd ed., Oxford University Press, Oxford, 2009), pp. 122–145; Joseph Raz, The Concept of a Legal System (2nd ed., Clarendon Press, Oxford, 1980), pp. 197–200.

32 Hartney, “Introduction”, pp. xlii–xliii.

For an overview of those changes and notions, see, e.g. Hartney, “Introduction”, pp. xlii–liii; Paulson, “Kelsen’s Legal Theory: The Final Round”; Duxbury, “Kelsen’s Endgame”. For example, in Kelsen’s late works: “The legal norm emerges as the meaning of an act of will, and the validity of the legal norm is ‘conditional upon the act of will of which it is the meaning’”. Paulson, “Kelsen’s Legal Theory” at 266 (footnotes omitted). For evidence that the seeds for this late transformation were present earlier in Kelsen’s work, see Stanley L. Paulson, “Metamorphosis in Hans Kelsen’s Legal Philosophy”, 80 Modern Law Review 860 (2017).