In 1957, H.L.A. Hart, Professor of Jurisprudence at the University of Oxford, spent a year at Harvard as a visitor and was asked to deliver the Oliver Wendell Holmes lecture. In this lecture, Hart decided to defend analytical jurisprudence and legal positivism arguing for a separation of law and morals in a climate inclined to see theoretical problems from a practical perspective and generally hostile to legal positivism. Lon Fuller, Carter Professor of General Jurisprudence at Harvard, demanded to be given a reply to the lecture to express his anti-positivist stance. Both lectures, given at the height of their careers, were published in the Harvard Law Review in 1958 and immediately became a point of reference for understanding positivism and some of the more important criticisms of positivism (for more on the historical and biographical background of the debate, see Lacey, 2004).
What does positivism claim? What was Hart willing to defend? What were Fuller's main objections, and do they really succeed? Was there really a debate between Hart and Fuller? A good starting point to organize these questions is one of the footnotes in the published text of Hart's lecture: ‘Positivism and the Separation of Law and Morals’. Here, Hart mentions what might be understood as positivism's main tenets:
It may help to identify five (there may be more) meanings of “positivism” bandied about in jurisprudence:
- the contention that laws are commands of human beings..
- the contention that there is no necessary connection between law and morals or law as it is and ought to be….
- the contention that the analysis (or study of the meaning) of legal concepts is (a) worth pursuing and (b) to be distinguished from historical inquiries into the causes or origins of laws, from sociological inquiries into the relation of law and other social phenomena, and from the criticism or appraisal of law whether in terms of morals, social aims, “functions”, or otherwise…
- the contention that a legal system is a “closed legal system” in which correct legal decisions can be deduced by logical means from predetermined legal rules without reference to social aims, policies, moral standards….and
- the contention that moral judgments cannot be established or defended, as statements of facts can, by rational argument, evidence, or proof (“noncognitivism” in ethics)… . (1958, pp 601–602 note 25)
Let us call these claims as follows: (1) the command theory; (2) the no necessary connection thesis; (3) the methodological claim; (4) the positivism as formalism charge and the problem of interpretation; and (5) the meta-ethical confusion. In light of these five claims, I will explore whether the exchange of views between Hart and Fuller in 1957 truly amounted to a debate. Sorting out this issue is the aim of this article.
Hart presents his article ‘as part of the history of an idea’ (1958, p. 594), namely, the positivist commitments of the utilitarian tradition represented in the work of Jeremy Bentham and John Austin. According to Hart, Austin and Bentham held (1) the command theory, (2) the no necessary connection thesis, and (3) the methodological claim. Hart himself was willing to defend this tradition with the exception of (1) the command theory. Regarding claims (4) and (5), he thought they misrepresented the positivist tradition, and in the article was at pains to reject the claim that (4) and (5) were part of the tradition.
I will start with the easy points and therefore will not necessarily follow the claims in order. First, (1) the command theory: This is an easy point because Hart and Fuller agree that the command theory is unacceptable as a rendering of law's nature. Explaining Hart's account of the errors of the command theory would involve developing his theory in detail (see Hart, 1961, especially Chapters 2–4). A few points will suffice for our purposes here.
The utilitarians thought that laws were commands. As Hart puts it, on the command theory, a law
is simply an expression by one person of the desire that another person should do or abstain from some action, accompanied by a threat of punishment which is likely to follow disobedience (1958, p.602).
In the 1958 exchange, and later in his seminal book ‘The Concept of Law’ (1961), Hart criticizes this idea on several grounds, but two are especially relevant: a) a desire that another person should do or abstain from some action accompanied by a threat of punishment is surely not a good account of law. Hart famously points out that when a gunman says to his victim ‘Give me your money or your life’, we have a desire and a threat of punishment and surely this is not law; and b) on Austin's view, these commands are issued by sovereign (a political superior) to whom the recipients of these commands owe habitual obedience; moreover the sovereign is not in the habit of obeying anyone superior to it. According to Hart, it was wrong to characterize the act of following law's directives as a habit of obedience. Law, like other normative phenomena, has rules, not habits, as an essential element, and if we concentrate on rules we have a more promising route to explicating law's nature. Concentrating on rules, Hart develops his notion of primary and secondary rules, and settles on the view that their union is the key to the ‘science of jurisprudence’.
Fuller explicitly acknowledges that Hart's example of the gunman situation is right and does not challenge this part of the argument (1958, pp. 638–639). However he does challenge Hart's notion of secondary rules. At one point, Hart claims that: ‘…fundamental accepted rules specifying what the legislature must do to legislate are not commands habitually obeyed, nor can they be expressed as habits of obedience to persons’ (1958, p. 603). Here, Hart is trying to highlight the importance of secondary rules that supplement primary rules of obligation, some of which establish procedures and empower certain bodies to make law (see Chapter V of The Concept of Law, 1961). Hart's point is that these are best explained in terms of the acceptance of rules, not merely habits of obedience. Fuller, in turn, focuses on the ‘acceptance’ issue Hart raises and argues that the acceptance of these rules must be due to moral reasons. Hart's account is flawed, according to Fuller, because he cannot make sense of the notion of acceptance without this being moral acceptance of the rules that make law possible. Fuller even states that the acceptance of these rules ultimately rests on a perception that they are right and necessary (1958, p.639).
This apparent disagreement may be a case of Fuller not understanding Hart's argument. The notion of acceptance of a rule is important in the context of Hart's notion of the internal point of view (1961, pp. 55–56). The internal point of view is one of the features that distinguishes rules from habits. Nevertheless, although the internal point of view acknowledges the importance of acceptance and a reflective critical attitude towards patterns of behavior, it is neutral regarding the reasons one might have for accepting rules. Hart states:
…[I]t is not even true that those who do accept the system voluntarily, must conceive of themselves as morally bound to do so,….their allegiance to the system may be based on many different considerations: calculation of long-term interest; disinterested interest in others; an unreflecting inherited or traditional attitude; or the mere wish to do as others do (1961, p.198).
There is more to discuss about the advantages and disadvantages of the notion of the internal point of view, but this is not the relevant issue here. It is clear that Fuller, in developing his notion of moral ‘acceptance’ of secondary rules, is aiming at an important connection between law and morality, a point Hart was bound to deny in any case.
So, regarding (1) the command theory, we do not have much of a debate, because both authors agree on the basic point that the command theory needed to be refuted. However, when Fuller disagrees with Hart on how to refute the theory, he misses the target in his criticism.
Let us now turn to (4), the charge of positivism as formalism and the problem of interpretation. Here, we have two important issues. The first is Hart's insistence that the utilitarian tradition and positivism generally was not attributing to judging the role of mere automated machines that use only logical deductions to reach decisions. On this formalist scenario, the judge is limited to reasoning in two steps: first, identifying the facts and the rule applicable to those facts; and second, a deduction to the pertinent legal consequence. According to Hart, this is a claim made only by formalism; it need not be accepted by positivism or analytical jurisprudence (if someone was to be blamed for this misconception, it is people like Blackstone and Montesquieu, 1958, p. 610).
While clearing the ground about what positivism entails and arguing that it does not entail a formalistic account of judicial reasoning, Hart does mention briefly a few ideas on the topic of interpretation. In this exchange, we have an introduction to what will become Chapter 7 of The Concept of Law, namely, to the idea of legal rules having a core of settled meaning and a penumbra of uncertainty. Hart introduces the now famous example of a legal rule that forbids vehicles in public parks (1958, p. 607). There is a core of settled meaning of the term ‘vehicle’, which applies, e.g., to automobiles, and also a penumbra of uncertainty regarding cases like bicycles and roller skates.
Fuller takes these ideas very seriously and dedicates much of his article to a response to them. According to Fuller, Hart advances a theory of judicial interpretation (1958, p. 662) and even worse a prescriptive theory of judicial interpretation (1958, p. 662). He thinks that Hart's ‘theory of judicial interpretation’ recommends that we interpret legal texts by analyzing single words of a statute. He responds that in interpretation, we also have to consider the context that gives rise to these problems of interpretation and take into account a more integral view of the statute and of law in general (1958, p. 662–666).
These are interesting questions Fuller raises concerning interpretation, but unfortunately, here, he also misses the point. Hart in the 1958 article and later in The Concept of Law was not trying to advance any theory of judicial interpretation, even less a prescriptive one. In this exchange with Fuller, he has a more modest aim: to be very clear that positivism was not formalism. Positivism, because of the notion of a penumbra of uncertainty, recognizes the creative aspect of interpretation and also that there is a settled core of meaning. In this sense, the important question for Hart is, to what extent are courts constrained by the law in matters of adjudication? The answer is to some extent provided by the settled core of meaning. This is a much narrower question and answer than would be delivered by a theory of judicial interpretation.
But there is a second important issue in this part of the exchange that highlights some interestingly diverse views of the two authors on the topic of interpretation. Fuller thinks that the issue of interpretation shows how morality has a more important role to play than the one envisioned by positivism. Fuller argues that the following questions are required in the process of interpretation: ‘What can this rule be for? What evil does it seek to avert? What good is it intended to promote?’ (1958, p. 665). In other words, for Fuller, we ask questions of ‘ought’ to determine what the rule ‘is’ (1958, 666). Moreover, we always consider the interpretation of a rule in light of its purpose even in easy cases. Fuller says that in easy cases, that is, those located within Hart's core of settled meaning, ‘we can see clearly enough what the rule “is aiming at in general” ’ (1958, p. 663). So interpretation for Fuller is an activity that always demands an analysis of a question of ‘ought’.
Hart has two very important objections to this line of reasoning. First, he states that even if we concede that these issues of interpretation turn on questions of what ‘ought’ to be the case, the ‘ought’ does not necessarily mean a moral ought: The ‘ought’ in question is a standard of criticism but not necessarily a moral one (1958, 612–613). Thus, we may have an analysis of a moral ought in some issues but not in every issue of interpretation. Hart mentions an example of a judge under the Nazi regime who decides a case by considering a question of ‘ought’. He thinks this is obviously not a moral ought (1958, 613–614).
Hart's second objection is to resist the idea of presenting interpretation as a ‘discovery’ of what the rule was all along (1958, p. 612). According to Hart, this is an invitation to revise our conception of what a legal rule is:
We are invited to include in the “rule” the various aims and policies in the light of which its penumbral cases are decided on the ground that these aims have, because of their importance, as much right to be called law as the core of legal rules whose meaning is settled (1958, p. 614)
He claims, however, that not all questions are questions like those of the penumbra (1958, p. 615). The idea of the penumbra only makes sense if there is a core of settled meaning: ‘even if there are borderlines, there must first be lines’ (1958, p. 614).
So, even if Fuller wrongly thinks that Hart is advancing a prescriptive theory of judicial interpretation, the exchange does give us important issues to consider regarding the nature of interpretation and the role interpretation plays in our renderings of the nature of law. As I will mention in my discussion of the next claim, here, we have a debate that up to this day is still controversial.
Let me continue with claim (5), the meta-ethical confusion. There is little further to discuss here in addition to the problem of interpretation that was already outlined above. Hart insists that it is a misconceived notion that positivism necessarily entails a view in meta-ethics, such as subjectivism, relativism or non-cognitivism, or a view about statements of facts and statements of value, or about statements of means and statements of ends. According to Hart, Fuller has a distinctive claim about statements of means and statements of ends because for Fuller, in some cases, in deciding questions of means, you must consider and clarify issues of ends (1958, p. 628 and 668). These cases are penumbral cases of instructions or rules, and how for Fuller you need an analysis of the ends or purposes of that instruction or rule to know what it means, a discussion that naturally brings us again to the topic of interpretation in the law (on ‘means’ and ‘ends’ and the Hart-Fuller debate, see Green, 2010).
So, Hart again takes issue with Fuller's claim that, to decide a question of what ‘is’ a rule, one must consider the question of ‘ought’. Hart reiterates that Fuller's way of presenting the problem of interpretation is misleading. Instead, Hart insists that there is not only one way of deciding questions of interpretation. We have several – ‘nearly equal in attraction’ (1958, p.629) – and, concerning uncertainties, it is better to be clear about them and not hide them talking as Fuller did about the ‘ought’ question and interpretation (1958, p. 628–629).
I mentioned at the end of the discussion of the last claim that there is an important debate between Hart and Fuller concerning issues of interpretation. On the one hand, Hart insists on the uncertainties inherent in the interpretative task and the need to avoid talking of discovering one way of deciding cases, even if ‘ought’ questions are taken into consideration. On the other hand, Fuller insists that the question of ‘ought’ take us to what the rule really is. One cannot avoid here bringing Ronald Dworkin into the debate. Fuller's claims constitute a preliminary analysis of Dworkin's idea that there is a right answer in legal disputes and that it is misleading to say that judges create new law or legislate. For Fuller and also for Dworkin, we discover what the law is through an analysis of an ‘ought’ question; in Dworkin's case, this is finding the value served by the practice of law (Dworkin 1977 and 1986, Stavropoulos 2003, also Summers 1984 and Vega Gomez 2011 regarding Fuller and the right answer thesis).
We are left with claims (2), the no necessary connection thesis, and (3), the methodological claim. Both claims reveal an interesting debate between Hart and Fuller. Let me begin with the methodological claim and end with the most important issue of the no necessary connection thesis.
(3) The methodological claim: In the footnote mentioned above, Hart makes the methodological point that Austin and Bentham and positivism in general argue for the analysis (or study of the meaning) of legal concepts. This task is one worth pursuing and
to be distinguished from historical inquiries into the causes or origins of laws, from sociological inquiries into the relation of law and other social phenomena, and from the criticism or appraisal of law whether in terms of morals, social aims, “functions”, or otherwise (1958, p. 602).
Actually, Hart barely touches upon the methodological claim and does not see the need to defend it. Nevertheless, for Fuller, this ‘positivist’ way of understanding how to do jurisprudence is one of the most irritating features of positivism in general. Fuller does not understand the project of jurisprudence in the absence of reflection about its purpose – without hesitation we might add moral purpose – and Hart's idea of clarity as a ‘sovereign virtue in jurisprudence’ (1958, p. 593) is not enough for him.
Fuller's methodology of jurisprudence contains two important features that although they are not thoroughly defended, nevertheless are highlighted in the exchange. These are the need to acknowledge the prescriptive component in jurisprudence and the need to attend to specific and practical problems in the law.
Fuller is committed to the need to acknowledge the prescriptive character of our theories. He claims that even if positivism states that it is just descriptive, it is in fact prescribing and owes ‘its special prescriptive powers precisely to the fact that it disclaims prescriptive intentions’ (1958, p. 632). So for Fuller, our descriptions are already in the prescriptive domain, even if we claim they are not. This methodological position helps to explain many of his criticisms of positivism; e.g., how definitions are ‘direction posts for the application of human energies’ (1958, p. 632), why according to Fuller, Hart needs to defend a definition of law that makes sense of the obligation of fidelity to law (1958, p. 635), and generally, why he thinks that the analysis of the concept of law and definitions of law elaborated on purely conceptual grounds are fruitless (1958, p. 635).
Fuller's second commitment – though it is related to the point just mentioned – is his insistence on jurisprudence as a problem-solver for specific practical dilemmas encountered by the law. At one point, while discussing the problem confronted by the drafting of a constitution for a country emerging from a violent period, he states,
What disturbs me about the school of legal positivism is that it not only refuses to deal with the problems of the sort I have just discussed, but bans them on principle from the province of legal philosophy. In its concern to assign the right labels to the things men do, this school seems to lose all interest in asking whether men are doing the right things (1958, p. 643).
Furthermore, Fuller demands a jurisprudence that answers to the problems a judge faces, for instance, those of a trial judge whose views on commercial matters diverge from those of his supreme court and therefore is caught in a dilemma. Fuller asks: What does positivism have to say to the judge? (1958, p. 646–647).
The answer positivism has to these sorts of criticisms is straightforward. Positivism does not need to provide an answer to the problems judges face or to issues raised in times of constitutional crisis, because, as Leslie Green claims, positivism is not in the business of offering advice but of trying to understand the nature of law (2008, p. 1036).
In this part of the exchange, Fuller's claims also serve as a preliminary to what Ronald Dworkin proposes in the domain of methodology in jurisprudence. Dworkin insists on the two methodological points highlighted by Fuller, i.e., the need to acknowledge the prescriptive component in jurisprudence and the need to attend to specific and practical problems in the law (2004, see also Vega Gomez, 2011 for this similarity between Dworkin and Fuller).
In one of his recent essays, Dworkin argues that despite its claim to the contrary, Hart's theory does have a theory on how judges should decide cases (2004). In Fuller, as in Dworkin, we have jurisprudence playing a role in aiding judges in their practical problems – through the recommendation to consider the ‘ought’ question – and a claim that even if positivists insist on elaborating a descriptive jurisprudence, positivism actually prescribes and has a theory of how cases should be decided.
Hence, there is a lively debate between Hart and Fuller over the methodological claim (3), which persists today between the positivists and the non-positivist approach to jurisprudence represented by Dworkin. (On the methodological dispute, additionally see Finnis 1980, Waluchow 1998, Perry 2001, Dickson 2001, Marmor 2006, among others, and on the implications of the methodological dispute between Hart and Fuller, see Lacey 2008.)
Finally, we come to the important claim (2), the no necessary connection thesis.
I emphasized earlier that this claim defended by the utilitarians is one that Hart is willing to stand by against Fuller and others. A claim that initially argues for a separation of law as it is from law as it ought to be. What does the separation amount to and what does it not imply in its original utilitarian form? The thesis states that it does not follow from the mere fact that a rule violates standards of morality that it is not a rule of law, and it does not follow from the mere fact that a rule is morally desirable that it is a rule of law. Hart mentions that this is what Bentham and Austin were anxious to assert (1958, p. 599). The reasons that served as a background for this utilitarian claim are also expressly mentioned by Hart: First, it avoids the problem of law and its authority supplanting someone's conception of what law ought to be, and second, it prevents the danger of laws not being able to be criticized because it decisively answers the question of what we should do (1958, p. 598). Both dangers are to be avoided in order to clear the way for the prescriptive task of law reform (1958, p. 597).
This is what the claim implies in its utilitarian form. And what does it not imply? The claim has never denied relations between law and morality, or as Hart said, ‘an intersection of law and morals’ (1958, p. 598). The utilitarians recognized this fact; that is, that legal systems are influenced by moral opinions and vice versa and that in some cases, morality might be brought into the law for adjudication purposes (1958 pp. 598–599).
So much for the claim in its original utilitarian form. The important issue arises when Hart tries to defend the claim from an objection, namely, that although the claim might be true with respect to each and every law of a legal system – that is, a particular law to be law need not pass any moral tests – what happens if we ask the question from the perspective of a legal system as a whole? Is there a necessary connection between law and morality if we ask this second and more important question?
After reformulating the problem as a question of a necessary connection between law and morality in this wider sense, Hart is willing to make two important concessions in favor of necessary connections between law and morality. The first is his idea of a ‘natural necessity’, one that arises from the contingent natural fact that human beings are vulnerable and need food and other goods to survive. All legal systems therefore have rules that prohibit murder, violence, theft, and rules that protect a minimum form of property. Because all legal systems have rules that overlap with moral rules, Hart concedes that there is a ‘necessary’ connection between law and morals; however, this is a ‘natural’ contingent necessity because one day, things may be different and we will no longer need these rules (1958, p.622–623). Hart's position here is an introduction to what in The Concept of Law will be called the minimum content of natural law.
A second concession derives from the characteristic nature of rules in a legal system. Rules are general in the sense that they refer to a generality of cases and a generality of persons. Through the general character of rules, there is a connection with justice and the notion of ‘treating like cases alike’. This is one important element of justice and morality that overlaps with one important element of all legal systems (1958, p. 624).
Fuller also advances a conception of an overlap between law and morality, which is similar to Hart's in some respects. It centers on characteristics pertaining to justice in the administration of the law. Fuller claims that law has an inner morality, a morality proper to law conceived of as ‘order’, a functioning order that guides the conduct of citizens. This order inherent in the law is what we call the Rule of Law, it exists when certain procedural principles obtain. Order or the Rule of Law requires the enactment of general norms, making norms understandable, limiting the use of retroactive laws, demanding publicity of the laws, avoiding contradictory norms, making norms congruent, avoiding frequent changes in the norms, and not demanding the impossible from citizens (see also Fuller 1969).
At one point in the discussion, Fuller states that even to create bad laws, you have to follow this implicit and inner morality in the law (1958, p. 645), a claim that immediately triggers the following question. If for both Hart and Fuller there is an important overlap between law and morality, which is the product of certain characteristics pertaining to justice in the administration of the law – in Hart's case, ‘generality’ and in Fuller's ‘order’ – and both accept that this morality is compatible with bad laws, then where is the disagreement? It seems that Hart and Fuller were basically arguing for the same conclusion. But the difference is crucial: Hart's concession that there is an overlap and two necessary connections between law and morality is compatible with pernicious legal systems. For Fuller, when we have an important perversion of the inner morality of law, a legal system ceases to be legal (1958, p. 660). So, in the end, Fuller's notion of an inner morality of law is not as formal as one thinks. This may have to do with Fuller's additional notion of an external morality of law, one that corresponds to the morality of the authority of law, i.e., the morality of the authority needed to make law and how according to Fuller, the internal morality and external morality ‘reciprocally influence one another’ (1958, p. 645). Unfortunately, this central element of reciprocity between the two moralities is not fully argued for by Fuller in this exchange, and there is no argument stating what role each morality plays in our evaluation of law.
The difference between the two positions is highlighted through the famous Grudge Informer Case. In Germany in 1944, a wife wanting to get rid of her husband reported to the authorities his negative remarks about Hitler and the Nazi Regime, an act punishable by law at that time. The husband was subsequently sentenced to death but never executed. After the collapse of the Nazi regime, the wife was brought to trial for intending to imprison her husband. The problem was that the wife acted ‘lawfully’ according to the statutes in force at the time she made the accusations. So, the German Courts after the Nazi regime in 1949 had the problem of either recognizing those hideous statutes as valid law or, on the other hand, declaring that hideous statutes such as these could never be considered valid law (for more implications of the case for the debate, see Dyzenhaus, 2008).
In response to this case, Hart echoes his idea that a separation between law and morals allows a clearer view of immoral laws (1958, p. 618). He claims that the best way to make sense of the problem was to recognize the statute as valid law and punish the wife under another statute, which retroactively made criminal an act that before was legal.
On the other hand, Fuller argues that the device of the inner morality of law would have been sufficient to declare these Nazi statutes invalid. For Fuller, gross violations of the inner morality help determine when a rule and/or a legal system cease to be legal. Fuller puts the issue as follows:
To me there is nothing shocking in saying that a dictatorship which clothes itself with a tinsel of legal form can so far depart from the morality of order, from the inner morality of law itself, that it ceases to be a legal system (1958, p. 660).
Therefore, as was foreseen by Hart, his two important concessions, a ‘humble minimum’ as he calls them (1958, p. 623), were bound not to satisfy natural law theory, or Fuller in this case. For Fuller, even certain procedural principles of order in the law, i.e., the inner morality of law, call for a more robust overlap between law and morality, an overlap that enables us to determine when a legal system is indeed legal or ceases to be so (see the important questions raised by Waldron regarding Hart's response to Fuller, 2008).
We have now made sense of Hart's defense of claim (2) and Fuller's response to it. I will end with some brief remarks about how contemporary positivists understand Hart's separation thesis (see also Waluchow 1998).
For some positivists, Hart never defends the separation thesis as a no necessary connection claim (Gardner 2001); for others, he denies all necessary connections with morality despite the concession that a certain overlap is ‘naturally’ necessary (Green 2008). I think the source of this confusion is in Hart himself notwithstanding the overall clarity he championed and practised in jurisprudence.
The confusion is attributable to Hart because he closes the discussion with only these two concessions; these two necessary connections between law and morality (see 1961 p. 198–207) as if accepting more necessary connections will defeat his defense of positivism and the separation claim in its original utilitarian form (a particular law to be law does not need to pass any moral tests) and modified later version (a legal system is still legal despite violations of standards of morality).
But contemporary positivists successfully claim that we can argue for additional necessary connections between law and morality and still be faithful to the separation claim. I think this is Joseph Raz's point (2003) when he makes sense of the truism that law can be good but also the source of much evil, we can make complete sense of the truism and defend it – as Hart was interested in doing while making his two necessary connections between law and morality compatible with pernicious legal systems – and accept several and some important additional necessary connections between law and morality (see also Green, 2008).