John Austin and Constructing Theories of Law

Abstract

One of the standard criticisms of John Austin’s work is that his portrayal of law, as essentially the command of a sovereign to its subjects, does not fit well with the way law is practiced in many or most contemporary legal systems, or the way that it is perceived by lawyers, judges, and citizens who are participants in those systems. However, contemporary legal theories also seem to fail to fit current practice in significant ways. The question is one of the power of the insights a theory brings, and whether it is worth whatever “distortions” it requires. For one generation, the insights of Austin’s (or Kelsen’s or Raz’s) theory might seem central, and the deviations trivial, while for a later generation, the insights might seem small or hard to accept, while the deviations seem fatal. Theory construction, especially where the theory is not anchored by falsifiable predictions, is often more a matter of persuasiveness, rather than a matter of truth. And if John Austin’s theory seems less sustainable than it once did, that may say as much about us, and what concerns us, as it does about his theory.

An earlier version of this paper was presented at the University College London Conference, “John Austin 150th Anniversary” and a different version of portions of the paper was published in (2011) 24 The Canadian Journal of Law and Jurisprudence 431–440. I am grateful for the comments of Andrew Halpin, the other participants at the University College London Conference, and Brian Tamanaha.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1.1 Introduction

One of the standard criticisms of John Austin’s work is that his portrayal of law, as essentially the command of a sovereign to its subjects,¹ does not fit well with the way law is practiced in many or most contemporary legal systems or the way that it is perceived by lawyers, judges, and citizens who are participants in those systems. The argument continues: that since the theory “fails to fit the facts,” Austin’s theory must be rejected in favour of later theories that have better fit.

This seems like a standard move in theory construction. Where the objective is to describe or explain some practice, any conflict between the theory and the practice being described counts strongly against the proposed theory, and we should search for an alternative theory that fits the practice better.

The importance to jurisprudential theory-construction of fidelity to participants’ understanding has been reinforced by the move in English-language legal theory towards a hermeneutic approach to legal theory (as in the “internal point of view” introduced by Herbert L. A. Hart,² and accepted by theorists as far apart methodologically as John Finnis and Joseph Raz³). In very rough terms, this approach argued (or, at times, merely assumed) that theories of law would be better to the extent that they accounted for the perspective of those citizens who viewed the law as giving them reasons for action. This approach to legal theory, in turn, reflects the general “hermeneutic” or “Verstehen” approach to the social sciences: a view that knowledge of social institutions is distinctly different from knowledge in the physical sciences, and that a primary focus of theorizing is and should be awareness of the motivations and purposes of participants, emphasizing participants’ understanding, not merely their behaviour.

For many influential modern approaches to the nature of law, including Joseph Raz’s exclusive legal positivism and Ronald Dworkin’s interpretivism, while they criticize the lack of fit of theories like Austin’s, those theories themselves unapologetically offer characterizations of legal practice that deviate in significant ways from the way most people practice or perceive law. Thus, at least at first glance, it appears that many contemporary legal theorists wish to have it both ways: they use the deviations from conventional understandings as grounds for dismissing some theories by other scholars, but forgive or overlook comparable deviations in their own theories.

This chapter will begin to explore what general principles can be learned, or developed, regarding when or to what extent deviation from the way law is practiced and perceived is appropriate in a theory of the nature of law. Additionally, the chapter will also consider whether, in light of the proper approach to fit and mistake in theory-construction, Austin’s theory of law might be a more viable alternative than is conventionally assumed.

1.2 Deviations and Mistakes

Joseph Raz writes:

John Austin thought that, necessarily, the legal institutions of every legal system are not subject to – that is, do not recognize – the jurisdiction of legal institutions outside their system over them. (…) Kelsen believed that necessarily constitutional continuity is both necessary and sufficient for the identity of a legal system. We know that both claims are false. The countries of the European Union recognize, and for a time the independent countries of the British Union recognize, the jurisdiction of outside legal institutions over them, thus refuting Austin’s theory. And the law of most countries provides counterexamples to Kelsen’s claim. I mention these examples not to illustrate that legal philosophers can make mistakes, but to point to the susceptibility of philosophy to the winds of time. So far as I know, Austin’s and Kelsen’s failures were not made good. That is, no successful alternative explanations were offered.⁴

In a sense, this sort of criticism of Austin, and of Kelsen, is, within the jurisprudential literature,⁵ perfectly common-place. Both theorists are presented as having interesting theories, but ultimately ones that are deeply flawed. In recent years, if scholars and students are familiar with Austin’s work at all, it tends to be through H. L. A. Hart’s use of Austin’s work as a stepping stone to his own approach: the way Hart used purported weaknesses in Austin’s command theory to justify Hart’s own quite different form of legal positivism.⁶

Hart offered a series of criticisms of Austin’s theory that are often now taken as proven accusations, with little attention given to potential defences of the theory. Hart’s criticisms included: (1) that, contrary to Austin’s theory, law contains much greater variety than is presented by a theory that equates law (only) with commands; (2) that Austin’s theory cannot distinguish a legitimate legal system from the rule of gangsters or terrorists; (3) that theories that equate law with the command of a sovereign cannot account for the legal status of custom, and may also have trouble accounting for judicial legislation; and (4) that many communities do not have anything that would count as a “sovereign” in the sense used by Austin, a person or institution that has no limits or constraints. In fact, Austin noted many of these objections in his own works, and offered responses,⁷ but these responses (some inevitably more substantial than others) have been largely forgotten in the rush to place Austin in his role as “the sincere but limited theorist whose faults were corrected by later and wiser writers.”

This is not the place to give any final reckoning to the individual criticisms of Austin’s work, but to consider the general sort of criticism raised. In particular, what I find intriguing about Raz’s quoted criticism of Austin and Kelsen, is that the author of the criticism himself offers claims about law that other theorists and observers might similarly characterize as subject to “counter-examples,” or as simply “mistaken” or “false.”

Raz has famously argued for what others have labelled “exclusive legal positivism,” a view that holds that moral evaluation can never play a role in determining what the law is (though it can play a role in determining what the law should be). When critics argue that there are clear contrary examples – moral standards in constitutional provisions or the use of moral reasoning in determining the content of common law legal norms – Raz denies that these are in fact refutations, or even counter-examples, to his theory. In the face of purported counter-examples, Raz notes that the judges’ characterizations of what they are doing in opinions are often the result merely of conventions of presentation, or slightly mis-leading labels used so as not to provoke those naively attached to certain preconceptions (e.g., that judges do not legislate).⁸

However, one would think that comparable arguments could be offered on behalf of Austin (and Kelsen, for that matter): arguing that whatever lack of fit there appears to be between their theories and current practices and perceptions would be removed or minimized by careful re-characterizations. Yet, for some reason, that move is rarely made by those commentators who are (too) quick to dismiss these theories.

1.3 Hart and Errors

And it is not just the rejected legal theorists of prior eras who must face accusations of lack of fit between theory and practice. Such claims reach even more established theorists.

The usual narrative of analytical jurisprudence, at least as given in most English and American university courses and in countless books and articles, is that John Austin has the merit of being the first, or one of the first, legal positivists, but that his theory was deeply flawed, flaws pointed out most clearly by H. L. A. Hart, whose own work set the standard for modern theories of law. However, though Hart’s work is treated, in this narrative, as significantly superior to Austin’s, and as the groundwork of all of merit in what has come since, there are occasional references to possible mistakes.

Some of the alleged errors are not relevant for our purposes, e.g., because they relate to propositions that are tangential to Hart’s theory of law; they can be abandoned without affecting the basic structure and basic claims of the theory (e.g., regarding the tenability of Hart’s practice theory of rules⁹). However, other claimed errors in Hart’s work cannot be so easily shrugged off: e.g., as to whether legal systems should be equated with the union of primary and secondary rules, whether every legal system has one (and only one) rule of recognition, and whether law is mostly a matter of rules.¹⁰

As regards the union of primary and secondary rules, Simon Roberts argued that this criterion for the designation “legal system” (or “non-primitive legal system”) improperly excluded many communities with more informal dispute-resolution and norm-creation systems.¹¹ To which, one might answer, on Hart’s behalf, that the mere fact that under some definition of law, or set of criteria for law, not all communities would be said to have law (or to have law in its fullest sense) is not, by itself, a reason to reject that definition or set of criteria.¹²

More central, perhaps, are two other criticisms. Under Hart’s approach, all legal systems (at least all sophisticated legal systems) have a rule of recognition, which sets the criteria by which one determines which norms are part of that legal system. The rule of recognition is the highest (or, to change the metaphor, the most basic) norm in the chain of justification and authorization within the legal system. And, more implied than either asserted or argued for, each legal system has only one such rule of recognition. Raz has argued that there is no reason to assume that this will in fact be the case; that legal systems could well have two (or more) rules of recognition.¹³ And Dworkin has argued forcefully that legal systems have principles as well as rules, legal standards that cannot be correlated with the sort of content-neutral “pedigree” criteria associated with a Hartian rule of recognition.¹⁴ These claims of error cannot be brushed aside as easily as Roberts’, and the arguments for and against have created a substantial literature, to which this article cannot do justice.¹⁵

1.4 Trade-Offs

One point I hoped to make by this too-quick tour of major legal theorists and their critics is that accusations that theories deviate from practices and perceptions are widespread, and by no means the end of the discussion. Perhaps, it would be better if a theory matched perfectly participants’ perceptions of a practice, but it is acceptable if it does not, as long as there is some benefit one gets in return. More to the point, perhaps, a perfect match between theories on one side, and practices and perceptions on the other, is not to be expected.

Theories are models: efforts to “boil down” complicated reality, and the variety of experience over time and across societies, to claims regarding what is “essential” amid the details and the differences. One could even argue that the problem is with theories of law that work too hard to account for nuance (e.g., accounting for all the different kinds of legal rules, etc.,) that they lose the basic insight about law’s nature. They are like maps that are large and detailed, almost as big as the area they purport to describe, creating realistic portraits of the area, but doing so at such a large size that they are no longer functional, and can no longer serve their intended function of helping us to find quickly the best route from one place to another.¹⁶

Theories and models involve, by their nature, trade-offs. The power or insight of the theory is to be weighed against the simplification, distortion, or mis-characterization involved in reducing the complexity of life to a simple picture. In economic modelling, it is sometimes argued that any distortions of human behaviour presented in the model are compensated for by the value of the model in predicting human behaviour. There is debate regarding whether in fact economic models are successful in predicting behaviour,¹⁷ but that is, for our purposes, beside the point. What is relevant is that prediction of events is a (relatively) objective matter, a marker most of us can agree upon as a valuable counter-weight to the cost of any distortion within the model.

However, within jurisprudence there are additional problems. How can one discuss the meta-theoretical trade-offs in theories of law if there is no consensus regarding either intermediate or ultimate values? One must first know what one is aiming for and what would count as success before one can even think about costs and benefits in relation to theory construction. What is it that we are doing, or trying to do, when we theorize about (the nature of) law?

This is a basic question for legal philosophers – as Nigel Simmonds put it in discussing the challenge facing Herbert L. A. Hart and those who came after him: “once essentialism (…) was avoided as an option, it became hard to see how an investigation of law’s nature could be anything other than an empirical matter.”¹⁸ Is there something philosophical to be said about law, that goes beyond mere historical and sociological investigations? But certainly Kelsen and Hart, and Finnis, Dworkin, and Raz – and likely Austin as well – thought of themselves as doing something different than empirical investigation.

What is the benefit we seek from a successful theory of law? Raz speaks of the ultimate objective of legal theory as explaining part of our community’s self-understanding.¹⁹ For Ronald Dworkin, it is an interpretive process that reworks existing practices in their best moral or political light.²⁰ For Liam Murphy, it is selecting or constructing the theory whose belief by society would have the best consequences.²¹ For Sean Coyle, it is part of an exploration of the role of law in realizing the good.²² For John Finnis, similarly, the objective of legal theory is, or should be, about asking the “why” question: “Why have law?” How does law fit within the moral requirement to seek the common good?²³ If we do not know what the objective of theorizing is, or if we cannot agree on what it should be, it will be difficult even to begin the discussion of when a theory’s lack of fit is justified by its achievements.

Ronald Dworkin’s concept of constructive interpretation gives an example of how trade-offs might be understood in theorizing. According to Dworkin, an interpretation (here of a social practice, though for Dworkin the claim is generalized to all interpretation) must meet some minimal level of fit with the practice being interpreted; otherwise it would not even qualify as an interpretation. Beyond that, one would either choose the best interpretation that also had the minimal level of fit (according to an early version of the theory²⁴) or choose the theory that had the best combination of fit and value (according to later versions²⁵).

A different question arises when it is not a straight trade-off, but rather a weighted choice. When Hart urges us to take into account the internal point of view, his argument is that this perspective is more central and (therefore) more important than the perspective of those who do not perceive the law as giving them reasons for action.²⁶ It is because this is a richer, better, fuller, or more central explanation that the theory should be built around it, rather than on a different basis, even if that other basis might have a better overall fit with perceptions and practices.²⁷

It is in the nature of trade-offs, that the greater the insight one believes that the theorist (Austin or Raz or Kelsen or Dworkin) has offered, the greater the deviation from participant perception (“lack of fit” or “mistakes”) that one will condone in the details of the theory. Even granting this much, the problem is that the existence and quality of an “insight” often seems to vary from one reader (observer) to another, and also over time. Thus, while in one era a theory’s mistakes might seem trivial relative to the insight offered, in another era that same lack of fit might seem fatal. Thus, to many, and perhaps to most, Austin’s theory looks untenable now, and, for these same observers, it may not be easy to understand how Austin’s theory could ever have been as dominant as it was. While for many of these same contemporary commentators, any lack of fit exhibited by, say, Joseph Raz’s exclusive legal positivism is worth carrying for the insights that theory offers about the connections between law, rules, reasons for action, and authority. How much deviation from practice one believes a theory can carry will inevitably be a matter of judgement.

1.5 Not (Quite) Trade-Offs

Perhaps we move too fast to be speaking of trade-offs for theories of law. Some theorists argue that there is no need to speak of trade-offs, because the theories in question in fact do not suffer from any lack of fit. Rather, the practices and perceptions that purport to differ from the theories are in fact untenable. For example, under a Razian analysis, judges may think that because they are applying moral-sounding constitutional provisions, they are declaring a pre-existing legal status, rather than making new law, when they invalidate a statute. However, Raz would say that this cannot be, for it is contrary to matters essential to the nature of law.²⁸ Similarly, under a Dworkinian analysis, a judge may think that she is declaring the legislators’ intentions for some statute, intentions that are purely matters of fact, but Dworkin would insist that this simply misunderstands what legislative intentions are or could be.²⁹

A final example, further afield, comes from the Scandinavian legal realists (e.g., Alf Ross, Karl Olivecrona, and Vilhelm Lundstedt), who criticized the normative language (e.g., “right” and “duty”) used in law.³⁰ The Scandinavian realists believed that concepts like “legal right” and “legal duty” were phrases without a reference, and could be explained only in terms of subjective psychological feelings of power or bindingness, or the residue of ancient beliefs about magical powers. These theorists did not doubt that citizens and legal officials referred to “legal right” and “legal duty” as though they were objects that somehow existed in the world, but in that, the Scandinavian realists argued, the citizens and officials were simply deceiving themselves.

A different alternative to a “trade-off” analysis would be that theorizing should be understood in terms similar to Willard V. O. Quine’s “web of beliefs.” Under this analysis, we have inter-connected views, that hang or fall together, and facts that do not initially seem to fit into our beliefs may require adjustments in aspects of the interconnected propositions, but that almost any such fact can be accommodated, albeit at times with some uncomfortable stretching in those beliefs.³¹

Theorists who do not entirely deny that there are mistakes or deviations in comparing their theories to actual practices and perceptions may instead discount the importance of the deviations. These discounting arguments come in certain common forms. First, there is the argument that the way certain judges, lawyers or citizens speak about the law does not reflect their actual views about the law, but instead reflects only certain conventions of presentation. This argument is often used in response to the observation that judges frequently speak about “finding” or “discovering” existing law (rather than creating new law) even when the outcome seems far different than prior decisions and other settled law.³² Second is the argument that judges and lawyers may characterize their actions in a certain way to respond to the political pressures and misunderstandings by naïve citizens (e.g., who do not want to think of their unelected judges as making new law, or making “political” judgements in interpreting and applying the law); according to this argument, these judges and lawyers do not believe the characterizations they report. Third (though this is seen far less often than the other two) is the claim that the judges, lawyers, and some citizens as well, are simply deceiving themselves. When the great English common law judges and commentators of the medieval and renaissance periods claimed that judges merely discover existing law, is it possible that at some level even these sophisticated and worldly observers actually believed that? Perhaps some of them did, and perhaps they did because it helped them to avoid facing unpleasant political and legal issues.

1.6 Is Law Distinctive?

In discussing legal theories, past and present, in this work, I have spoken in abstract terms regarding the process of building theoretical models, and the trade-offs within a theory. One issue left unconsidered is whether law, and theorizing about law, might be different in important ways from other theorizing about social practices, distinctive in ways that affects our thinking about models and trade-offs.

One difference that might be worth noting that is law has a role (at least an arguable role) in our practical reasoning – reasoning about what we ought to do and how we ought to live – that most other social practices do not have or claim to have. This aspect of law has been particularly emphasized by natural law theorists, but is accepted, to different degrees and in various ways, by many other theorists as well. As John Finnis has argued, law has a “double life”: it is simultaneously a social/historical fact and a normative system.³³ Law as a social-historical fact is constituted by the actions of officials within a particular legal system from its beginning to the present. There are propositions about law which are primarily summaries of what decisions legislatures and judges, and perhaps also administrative agencies and executive/enforcement officials, have made over time. Such claims are made by social scientists and other academics, as well as by legal practitioners and judges.

Often, when claims are made about the law, there is some ambiguity regarding whether the claims are descriptive/historical, regarding what actions were actually taken by officials in the past, or whether there is some element of modifying, re-characterizing, or reforming the rules to make the current (or future) cases better. And when theories are offered of areas of law, the detailed case outcomes are built into generalizations in ways that reflect a conscious bias towards making the overall picture more just or at least more coherent. This is sometimes described as “rational reconstruction.”³⁴

The way that law has an aspect of social practice and an aspect of practical reasoning certainly complicates any effort to theorize about the nature of law. And it may make a difference on what counts as a cost or a trade-off in theorizing about law. However, it is not clear that law’s distinctive nature changes the general meta-theoretical question about how one balances insight and distortion: the comparison of costs and benefits appears to remain comparable with what occurs with theories in other areas.

1.7 A Different View of Austin

I want here to take a brief break from general discussions of theorizing and lack of fit to return to Austin, and consider what arguments might be raised on his behalf.

The argument for Austin might go as follows (and no claim is being made that this argument can be found in Austin’s own works, or even that he would have approved of it had it been brought to his attention). Austin’s theory simplifies, and therefore distorts, but the simplification is a necessary cost for an important objective: uncovering a basic insight about law.

Regarding the problem of theoretical objectives, discussed earlier, one might note first that there are significant doubts regarding what Austin saw himself as offering in this theory. At one or two points in his lectures on jurisprudence – but (to my knowledge) not much more often, in the course of over 1,000 pages of text – Austin describes his work as offering a “science” of law.³⁵ This may parallel the continental European theorists he had read, and later continental theorists like Kelsen, who saw their conceptual analyses as part of a “science” of law. At the same time, modern commentators find that Austin’s discussions could be as easily interpreted as description (this is what is true of all known legal systems) as conceptual (this is what is necessarily true of any legal system).³⁶ A conceptual objective would make it easier to speak of “insights” that justify any lack of fit.

There are different (but related) ways of characterizing the insight(s) about law that can be drawn from Austin’s command theory. First, that law is essentially about power.³⁷ Second, that law is best understood (and best practiced) as a top-down institution, with norms imposed by the government on its citizens, rather than as a bottom-up institution (as both the classical commentators on the English common law and the continental historical jurisprudence theorists would have it). Third, that every legal system has some entity whose power is effectively unconstrained.³⁸

From the perspective of some of these perspectives, it is a benefit, not a drawback, that Austin’s theory does not incorporate the perspective of citizens who view the law as creating reasons for action. From this Austinian approach, it is the Hartian legal positivist approach that is mistaken, in its apparent willingness to join certain strains of natural law theory in focusing too much on how law can or should create (moral) reasons for action.³⁹

Of course, one response to a revised Austinian theory would be in much the same tune as prior criticisms: that this is a theory built on a poor fit with actual practices and perceptions (what might less delicately called “mistakes”), and thus cannot claim to have uncovered insights, only distortions. Under the Dworkinian analytical structure mentioned above, the argument is that the theory’s fit with the practice is too poor to even qualify as an “interpretation.” (Perhaps under a coherentist view, like Quine’s web of beliefs or Thomas Kuhn’s discussion of “paradigms,”⁴⁰ it is the claim that certain facts are so hard to incorporate into the existing analytical or conceptual structure that the whole structure must be rejected and replaced).

To some extent this is the argument that is still going on in legal theory, relating not only to Austin’s work, but more generally regarding the role of coercion in the nature of law.⁴¹ Those who believe that coercion is central to law’s nature think that theories of law that omit or discount coercion are missing something basic. Theorists on the other side of the issue make comparable criticisms, asserting that it is the coercion-centred theories that are missing something essential.⁴²

1.8 Conclusion

Of course, there are no bright-line rules for determining when a theory of some practice is tenable and when it is not, and when an existing way of understanding a practice needs to give way in the face of purportedly recalcitrant facts. (And this is not merely because we are dealing here with social practices rather than the physical sciences; a similar lack of bright lines applies also to when one Kuhnian paradigm within science must give way to another⁴³).

To some extent, the success or failure of a theory becomes a matter of perception and a matter of judgement among the consumers of theory. Legal theories – like all other ideas – arise in response to the intellectual questions and practical concerns of the time in which they arise.⁴⁴ They may yet adapt or be re-characterized in ways that make them seem responsive to the questions and concerns of another period, but inevitably there will come a time when a theory that once seemed powerful and important begins to seem instead quaint and without use to a new generation of thinkers. And such changes in perception likely occur also at the level of components of theory, and components of theory-construction. For one generation, the insights of Austin’s (or Kelsen’s or Raz’s) theory might seem central, and the deviations trivial, while for a later generation, the insights might seem small or hard to accept, while the deviations seem fatal.

Theories of the nature of law are relatively “unmoored,” lacking, on one hand, the constraint of prediction of events; and, on the other hand, any agreed purpose. It should thus not be surprising that there is significant disagreement among theorists. There is disagreement about how to characterize certain of the facts on the ground, but even where agreement can be found on that, disagreement remains, as reasonable people can choose differently when faced with theories that make different choices about what is important, what counts as “insight,” and how much of participants’ perceptions or “common sense views” one can or should throw overboard in the name of theory-building.

The theorist has resources available when faced with apparent deviations between a theory and people’s practices and perceptions. It can be argued that apparent deviations just reflect conventions of presentation, deceptions, or self-deceptions. Alternatively, it can be argued that any characterization of the relevant practices other than the one offered by the theory is unsupportable. Beyond that, a theorist’s claim in the face of recalcitrant data will be some variation of the trade-off metaphor: that the cost involved in deviating from the practices and perceptions is worth accepting in light of the insights discovered and displayed by the theory.

Theory construction, especially where the theory is not anchored by falsifiable predictions, is often more a matter of persuasiveness, rather than a matter of truth. And if John Austin’s theory seems less sustainable than it once did, that may say as much about us, and what concerns us, as it does about his theory.