Elephants and Flashlights: On Martin on Raz

Elefantes y linternas: sobre Martin sobre Raz

André L. S. Coelho
National Faculty of Law - Federal University of Rio de Janeiro, Brasil

Elephants and Flashlights: On Martin on Raz

Isonomía. Revista de Teoría y Filosofía del Derecho, no. 55, 2021, pp. 168 -176

Received: 23 August 2021

Accepted: 02 September 2021

Abstract: The paper responds to some of Margaret Martin’s criticisms of Joseph Raz’s legal theory in her book Judging Positivism, especially the alleged change in Raz’s account of adjudication and its potential negative impact on Raz’s legal theory. To that end, I treat Raz’s theory in particular, and analytic legal theories in general, as providing a picture of what law would be like if it were fully determined by certain aspects rather than revealing essential aspects of what law is. In this sense, legal theories are compared, as in the famous Indian parable, to blind men trying to describe what an elephant is, a comparison that is later completed by the idea of the periodic rotation of a lantern, which illuminates each time a different aspect of the animal.

Keywords: Joseph Raz, Margaret Martin, legal exclusive positivism, adjudication.

Resumen: El ensayo responde a algunas de las críticas de Margaret Martin a la teoría jurídica de Joseph Raz, en su libro Judging Positivism, especialmente la afirmación de que Raz cambió su teoría de la adjudicación y que este cambio produjo un impacto negativo en su teoría del derecho. Con ese fin, trato la teoría de Raz en particular, y las teorías jurídicas analíticas en general, como si, en lugar de revelar aspectos esenciales de lo que el derecho es, proporcionaran una imagen de cómo el derecho sería si estuviera totalmente determinado por ciertos aspectos. En este sentido, las teorías jurídicas se comparan, como en la famosa parábola india, con ciegos que intentan describir qué es un elefante, comparación que luego se completa con la idea de la rotación periódica de una linterna, que ilumina cada vez un diferente aspecto del animal.

Palabras clave: Joseph Raz, Margaret Martin, positivismo jurídico excluyente, decisión judicial.

I. Introduction: Elephants and Flashlights

Hart once made this famous observation about the “seemingly paradoxical utterances” produced by some legal philosophers of the past about their concepts of law:

For, understood in their context, such statements are both illuminating and puzzling: they are more like great exaggerations of some truths about law unduly neglected, than cool definitions. They throw a light which makes us see much in law that lay hidden; but the light is so bright that it blinds us to the remainder and so leaves us still without a clear view of the whole (Hart, 1994, 2).

I find this quote fascinating, most likely for different reasons from the ones Hart had in mind. I believe this metaphor of the bright light that both reveals something hidden and blinds us to the remainder is not merely a compensatory explanation of some eccentric legal theories and their outlandish views, but rather an apt portrayal of every legal theory, no matter how sensible it sets out to be.

I see legal theories as the blind men of the Indian parable, conceptualizing what an elephant is by having each of them touching a different part of its giant, polymorphous body. Legal positivists touch the leg and claim that the animal is a pillar like a tree-trunk; natural lawyers place their hands upon the elephant’s side and claim it is a wall; legal realists rub the ears and claim it is a fan; interpretivists grab the trunk and claim that the elephant is a long, thick serpent, etc. None of them is totally wrong, but the right account, that would only emerge from their combined views, is impossible to arise through conceptual debate alone, since each view would seem incompatible with the other ones, leading to the inevitable conclusion that only one of them can be true.

Helpful as far as visual aids go, the elephant parable is limited to a synchronic dimension. Legal theories take different paths not only by focusing on different aspects of law (synchronic diversity), but also by reacting to each other over time (diachronic diversity). Later theories emphasize aspects neglected by earlier ones. This diachronic dimension must be added to the mix so that we can see the problem more comprehensively. Combining Hart’s quote with the elephant parable, I would then say that whenever one theory directs the flashlight beam towards the leg, the next one cannot but illuminate the side, and the next one the trunk, the next the tusks, and so on.

I am of course aware that this is not how the proponents of each legal theory see them. They propose their views on the law as complete, thereby excluding all other ones. But if the elephant parable offers any warning, it is that we must not expect the blind men to recognize the limits of their accounts. After all, they were there. They aid hands on the elephant. Thus, they are fully confident they know the whole truth. However, this should not stop us, we who are but blind men listening to the reports of blind men, from being more cautious and interpreting their accounts as simply offering pieces of the truth.

That is the standpoint from which I see the many merits of Raz’s theory of law. His ideas in Practical Reason and Norms (Raz, 2002) about authority, exclusionary reasons, prescriptive and non-prescriptive rules, institutional systems, and the legal point of view are, in my particular reading, better appreciated as a reaction to interpretive and morally charged theories of law and adjudication. Whereas these theories point the flashlight beam to hard cases and the indeterminate content of rules, emphasizing implicit values, personal judgments, and normative disagreements, Raz moved the luminescence back to the idea of obligation, calling attention to what it means to be under the authority of law.

Quite understandably, that was not the way Professor Margaret Martin chose to critically address Raz’s legal theory, or rather, the array of legal theories he defended throughout his prolific career. Her book, Judging Positivism (Martin, 2014), is certainly the most thorough examination of Raz’s main works and ideas and contains some of the best developed criticisms directed against them. The way she chose to address Raz’s theory is not, as I would view it, as a good theory of the elephant’s leg, but instead as a bad theory of the elephant, pointing to the many shortcomings and contradictions that come from depicting the whole animal from touching one of his legs. She also did not take his theory as a shift of the flashlight, that is as a welcome reminder of some truths of law neglected by earlier theories, but instead as a failed attempt at grasping its whole truth.

The contrast between Professor Martin’s preferred way of critically addressing a legal theory and mine is likely to be the result of different conceptions of the functions and merits of theories in general. In what follows, I will try my best to comment on some of Martin’s critical theses in the context of her own main concerns. But it is very likely that my lack of both hope and interest in a complete and unified theory of law impacts the way I am more inclined to read her contributions. That is why I chose to begin speaking about elephants and flashlights, a theme that will make a comeback in my arguments a little further down in the text.

From the many interesting topics Martin raised about Raz’s work, I chose two, belonging to chapter 3 of her book, on which to comment. Both are related to the distinction between a theory of law and a theory of adjudication. Martin argued that (1) from Practical Reason and Norms to Ethics in the Public Domain (Raz, 1995), Raz changed his account of adjudication from a positivist to a realist stance; and (2) this change negatively impacted his theory of law, especially the sources thesis, which became more apparent in his account of rights, where he took a position that, according to Martin, resembles more Holmes’ views rather than Hart’s.

It is worth saying that my comments on these points do not depend on my agreement with Raz’s theses. Much on the contrary, I find Raz’s discretionary model of adjudication unappealing and one of the least interesting aspects of his theory. However, as I do think the distinction between a theory of law and a theory of adjudication is defensible, I do not think the reason why Raz’s theory of adjudication should be considered problematic stems from any negative impact it might have on his theory of law. In fact, I sustain that both theories are independent.

Before proceeding, I would like to pay a brief look at what some descriptive theories in analytic philosophy truly are about.

II. Leg-based Theories of Elephants

Whenever we say that analytic legal theories are descriptive, what are they supposed to be descriptive of ? Law is the shortest, but also the vaguest of answers. We certainly do not mean each legal system with the very norms that happen to be and their contents. What we mean instead is one of two things: some traits or social practices common to all or most legal systems we are culturally and historically exposed to. Validity, integrity, authority, etc., are traits. Creating, identifying, interpreting, obeying, and applying rules and principles are social practices. These are the real objects of description in so-called descriptive legal theories.

Since they are the artifacts of philosophers, and not of social scientists, such descriptions are conceptual, not empirical. They proceed by proposing concepts that, arguably, parallel the everyday intuitions of those familiar with the law. Those intuitions are the theorists’ own, as long as they think that such intuitions are widespread and shared with every relevant person. But no matter how much a legal theory claims to grasp the thinking of the ordinary citizen, it never consults her. Instead, it relies on conceptual intuitions of armchair thinking about what is thought, known, or believed by the ordinary citizen – sometimes not without much stretching, picturing ordinary citizens too much like philosophers. Armchair conceptual intuitions are the basic building blocks of every analytical legal theory, and once we understand that, their shortcomings become at the same time less surprising and more forgivable.

This brings us back to the elephant parable. Sketching a legal theory from a powerful intuition is like pointing the flashlight at one part of the elephant and taking that part as the key to the whole. We, who are aware of what a whole elephant really looks like, know that there is no key to the whole, because the elephant is giant and polymorphous, made of different heterogeneous parts. Treating any of them as key to all other parts is committing oneself to a big mistake from the start. But that is basically what analytical legal theories are, that is, a bunch of leg-based theories of elephants. That is the nature of the descriptive enterprise.

In Raz’s case, a description of the entirety of law based on the concept of authority is the casebook example of a leg-based theory of the elephant. Raz claims that he is describing how social practices really work, because he is convinced that the powerful conceptual intuition he is starting from is the key to those social practices as a whole. But since there is no key to the whole, what he obtains is not so much a description of how social practices of law truly are, but rather a description of how those social practices would be if his central concept were the key to the whole. That is, if social practices were completely determined by the concept of authority. From all the possible worlds of law, his description of it corresponds to what law is in the authority-based world – a world that, although conceivable and maybe even attainable, is not our world. He is describing how the elephant would look like, if it were entirely made of legs – a possible animal, just not anything like an elephant at all (this is not particular to Raz in any case, since Hart’s convention-based, Dworkin’s integrity-based, and Finnis’s flourishing-based conceptions of law are also leg-based theories of elephants).

But it is important to distinguish a leg-based theory of elephants from a normative theory of elephants. For it is one thing to have touched the leg and leaped to the conclusion that elephants are entirely made of legs, and another to think that elephants are not entirely made of legs but maybe they should be to become better elephants. Raz’s authority-based theory of how law works is a theory of the first kind. Whenever he drops his descriptive hat and sports his normative one, he reveals himself much more of a supporter of individual freedom, solidary communities, moral legitimacy, and non-institutionally constrained moral judgment. So, whenever we refer to his authority-based theory of law, or to how ordinary citizens behave in his authority-based world, we are neither talking of real, concrete social practices nor of normative, idealized ones. We are talking of highly stylized practices that are real in the authority-based world, which means that, compared to the real, concrete practices of our world, they capture at most the aspects of these practices that are authority-based. This will prove to be important later on in our considerations.

III. The Change of Raz’s Account on Adjudication

In her book, Martin stated that Raz changed his view about adjudication from a fact-based, positivist account in Practical Reason and Norms to a value-based, realist account in Ethics in the Public Domain. According to Martin, in the earlier book Raz maintained that judges were obligated to apply the law, whereas in the later book he changed his mind and said that judges were moral reasoners that might or might not apply the law.

I believe Martin would be right if Raz had defended a normative position about adjudication in Practical Reason and Norms. However, I do not think he actually stated that judges were morally obligated to apply the law (as identified by the sources thesis). The most normatively charged passage is this:

The second important consequence of the difference between institutionalized systems and systems of absolute discretion is that the former contain, indeed consist of, norms which the courts are bound to apply regardless of their view of their merit. A more accurate formulation would be that institutionalized systems consist of norms which the primary organs are bound to apply and not at liberty to disregard whenever they find their application undesirable, all things considered. (Raz, 2002, 139)

But, despite the ambiguity of language, this passage is a descriptive one. It states the consequences of taking the concept of an institutionalized system seriously, that is, of organizing the entirety of the application of its norms according to its conceptual difference regarding a system of absolute discretion.

Raz is not talking here about how law should work or how judges should decide cases, but rather of how the concept of law works, or how law would work if every aspect of its working was guided by the necessary traits of the concept it belongs to. If this explanation is true, then, the other reference more suitable to be interpreted as normative has to do with the legal point of view.

In the last chapter of Practical Reason and Norms Raz offered his doctrine of the legal point of view as a solution to the problem of the normativity of law, that is, of why legal rules are to be considered norms. He explained the normativity of law based neither on reasons to act nor on generalized beliefs. He explained it as a conditional normativity, dependent on assuming what he called the legal point of view.

The legal point of view (of system S), we could say, consists of the norms of S and any other reasons on which the norm subjects of S are required by the norms of S to act. The ideal lawabiding citizen is the man who acts from the legal point of view. He does not merely conform to law. He follows legal norms and legally recognized norms as norms and accepts them also as exclusionary reasons for disregarding those conflicting reasons which they exclude (Raz, 2002, 171).

The legal point of view is the perspective of the agent that treats the entire body of legal norms as legitimate and binding. No one is obligated to assume the legal point of view (for that would incur in the same problems of basing the normativity of law on actual reasons to act), but once you do, you come to see every prescription of law as a sufficient reason to act upon its content. Law, therefore, possesses normativity only for those that see it from the legal point of view.

The impression that judges are obligated to assume the legal point of view might emerge from passages like this:

It is not necessary for a legal system to be in force that its norm subjects are ideal law-abiding citizens or that they should be so (i.e. that legal norms are morally valid). But it is necessary that its judges, when acting as judges, should on the whole be acting according to the legal point of view. This entails that the courts must regard ordinary citizens as required to be ideal law-abiding citizens and judge them accordingly (Raz, 2002, 171).

But there is a difference between a requirement for a legal system to be in force and a moral obligation. Judges are not morally obligated to apply the law. There are no set of reasons capable of showing that, for a judge, applying the law is always, all things considered, the best moral choice. But, for a legal system to be in force, it is necessary that the judges assume the legal point of view and, accordingly, treat the citizens as if the latter assumed the same point of view. The following example helps to realize the difference:

If I go with a vegetarian friend to a dinner party I may say to him, ‘You should not eat this dish. It contains meat.’ Not being a vegetarian, I do not believe that the fact that the dish contains meat is a reason against eating it. I do not, therefore, believe that my friend has a reason to refrain from eating it, nor am I stating that he has. I am merely informing him what ought to be done from the point of view of a vegetarian. Of course, the same sentence can be used by a fellow vegetarian to state what ought to be done. But this is not what I am saying, as my friend who understands the situation will know (Raz, 2002, 175-6).

So, saying that judges as judges should apply the law is much like saying that vegetarians as vegetarians should not eat a dish that contains meat. In both cases, the obligation in question is conditional to assuming a certain point of view. But there is no definitive reason why someone should assume that point of view. Those statements would still make perfect sense even if the one making them believed the judge should not apply the law and the vegetarian should not refrain from eating meat.

In the case of a legal rule that, applied to a certain case, would give rise to great injustice, it would make sense to say both that the judge as judge, assuming the legal point of view, should apply that rule nevertheless; and that the judge, as a moral reasoner, should not, all things considered, apply that rule. The normativity of law is conditional to assuming the legal point of view, and for a legal system to be in force it is necessary that most judges assume this point of view most times. This does not mean, however, that, from time to time, judges cannot refrain from applying the law; and it certainly does not mean that a normative theory of adjudication cannot recommend doing so.

IV. The Negative Impact on Raz’s Theory of Law

There is, to be fair, no theory of adjudication in Practical Reason and Norms, merely a vague reference to the obligation of judges as judges, once they assume the legal point of view, to apply the law. When something is said about applying norms regardless of their merits, it is in the context of describing the consequences of the institutionalized system model. The term formalism is not used. There is not even an explanation of whether the idea of applying the law entails giving preference to certain methods (textualism, intentionalism, etc.) instead of others. So, the defense of a highly discretionary theory of adjudication in Ethics in the Public Domain is not so much a change as it is the first real formulation of a theory of adjudication in Raz’s work.

Martin argues that this new theory of adjudication negatively impacts Raz’s theory of law, especially the sources thesis. Here too I take the route of disagreement, for I think there is no actual connection between both things.

In Ethics in the Public Domain, Raz distinguishes between reasoning about the law and reasoning according to the law. Reasoning about the law tries to determine the content of what the positive law says about a certain issue. It is part of the judicial business, but it is far from being all of it. Judges, Raz explains, are moral reasoners. They do not simply decide on the issue of what the law says. They use a wide range of reasons in order to reach the solution that, all things considered, is right. That is reasoning according to the law. Now legal considerations are the starting point, but not the finishing line. Judges may use non-prescribed criteria and non-written principles to extend, diminish, modify, or replace the content of the law. Which means that how judges decide cases is not predictable from what positive law says.

Martin argues that in no other point of the theory the negative impact shines more brightly than in Raz’s theory of rights. According to her explanation, in Practical Reason and Norms rights are made dependent on legal rules, given the sources thesis, and are necessary to guide social action, since they tell the addressees what they can and cannot do. The way the addressees determine which rights they have is by looking backwards at legal rules, and not forward at judicial decisions. Given the formalist account of adjudication, they expect judicial decisions to conform to legal rules and use the latter as their source of information about their rights. The sources thesis, then, has a role not only in the identification of law but also in the identification of rights, contributing therefore for social coordination.

But, Martin explained, once the formalist theory of adjudication in Practical Reason and Norms was replaced with a realist one in Ethics in the Public Domain, the way the norm subjects determine what rights they have would have to shift from looking backwards at legal rules to looking forward at judicial decisions. Since judicial decisions might or might not apply the law, forming expectations based on legal rules would be naïve and deceptive. Guessing or predicting judicial decisions would become the new method of identifying rights, which Martin claims, is closer to Holmes’ idea of prophecy than to Hart’s idea of rules.

But again, this seems to me like a confusion between the descriptive and the normative dimensions of the theory. As long as the description of the law is authority-based, legal rules, identifiable via the sources thesis, would still be the main source of information about rights, if the practice of identifying rights is described as it would be in an authority-based world. If, on the other side, judges are morally obligated to decide as moral reasoners, withdrawing from legal rules whenever a different decision showed itself to be, all things considered, morally better, this (normatively ideal) display of discretion would only affect the (descriptively factual) social practice of identifying rights if judicial discretion became (descriptively) such a dominant legal trait that social practices of law had to be re-conceived as they would be in a discretion-based world, and not in the authority-based one. Real individuals would perhaps learn and change; but the norm subjects of the authority-based world are not real individuals.

Of course, other answers would be possible. If, for example, judicial decisions were discretionary, often withdrawing from legal rules, but most individuals did not notice that and thought they were applying positive law, nothing would change in the identification of rights. If judicial decisions were discretionary, often withdrawing from legal rules, and most individuals noticed that, but most rights were left untouched by those decisions, nothing would change either. If judicial decisions were discretionary but did not withdraw from legal rules all that often, no change would come too. And if discretionary decisions were right, but most judges did not follow that, relying instead on the application of legal rules, Raz’s normative model, to the extent that it would have no impact over real-life decision making, nothing would change either in the social practice of identifying rights. Only if Raz’s normative model were adopted, if it changed most judicial decisions, if those decisions withdrew from legal rules very often, if those decisions changed most rights, if the norm subjects realized that and if guessing or predicting was, comparatively, more successful than relying on legal rules – only then, this line of argumentation suggests, it would change the behavior of the norm subjects in the identification of their rights (I would not pursue this line of argumentation because it treats the social practices legal theory talks about too much like real social practices, instead of conceptually conceived ones: it speaks of legal theories as elephant-based, and not leg-based, as I think they truly are).

V. Conclusion: Raz as a Flashlight Wielder

In this last section, I would like to advance a provisional conclusion about how my elephant and flashlight approach would see Raz’s legal theory.

First, Raz’s theory has the merit of taking one trait of law, namely authority, and conceive of the whole of law as it would be in an authority-based world, that is, as if law were completely determined by authority, and authority alone. Since the authority-based world is not our world, the social practices the theory describes are not ours either, but they help us see what our real-world practices might have that is indeed authority-based. There’s something in this approach that is resemblant of Platonic models and Neo-Kantian constructs, but I am fairly convinced that it applies to the product of analytic legal theories, even if in defiance of their methodological self-understanding.

Most people would say that, once the authority-based world is not our world, Raz’s theory is not successful as a descriptive theory. Some would confound its idealization in the descriptive sense with its idealization in the normative sense. And I would disagree with both groups. Descriptive idealizations are the best we can achieve by using conceptual methods (empirical methods have similar problems, but that is another issue for another time.) And since conceptual methods are prone to the “key to the whole” approach, there can be no elephant-based theory of elephants: only leg-based, side-based, ear-based, trunk-based, and tusk-based ones, each teaching us something valuable about the elephant, but at the same time failing miserably in giving us a picture of the whole animal that is free from distortion and generalization.

However, the flashlight part must come right after, to emphasize the importance of Raz’s theory in reminding us of aspects of law neglected by other theories. Moreover, depending on the moment of legal history one is situated at, a certain part of the elephant might become the most important to be beamed at with the flashlight. Of course, at the level of valuable descriptive idealizations, Hart’s convention-based, Dworkin’s integrity-based, and Finnis’s flourishing-based worlds of law might be just as informative and interesting. But in a quadrant of Latin American legal history so wrecked by judicial discretion and lawfare, where the constitution, the rule of law, and legality itself, have been neglected and harmed (not without a good deal of blame of those that glorified judicial interpretation and creativity), paying due attention to the authority leg of our elephant is perhaps not the worst idea after all.


Hart, H.L.A., 1994: The Concept of Law. Oxford: Clarendon Press.

Martin, Margaret, 2014: Judging Positivism. Oxford: Hart Publishing.

Raz, Joseph, 1995: Ethics in the Public Domain: Essays in the Morality of Law and Politics. Oxford: Clarendon Press.

Raz, Joseph, 2002: Practical Reason and Norms, third ed. New York: Oxford University Press.