Hostname: page-component-54dcc4c588-2bdfx Total loading time: 0 Render date: 2025-09-30T03:14:46.121Z Has data issue: false hasContentIssue false

Can We Legally Revise the Highest Legal Rule?

Published online by Cambridge University Press:  30 September 2025

Manish Oza*
Affiliation:
https://ror.org/02grkyz14 University of Western Ontario , London, Ontario, Canada
*
Rights & Permissions [Opens in a new window]

Abstract

This paper is about whether we can revise the highest rule in a legal system using legal reasoning. HLA Hart and Alf Ross argued that we can’t: either the highest rule is unchangeable, or it can only be changed in a revolution—a merely causal change, not a legal one. I argue, drawing on an idea from Hartry Field about logic and epistemology, that we can. The emergence of the conclusion, within a legal system, that the highest rule should be replaced by something else can provide a legal reason for the change. In making the change, we act for reasons internal to the legal system. This allows us to make sense of the distinction, recognized by legal subjects, between a revolution and a legal change of sovereignty.

Information

Type
Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press

In the funhouse mirror-room you can’t see yourself go on forever, because no matter how you stand, your head gets in the way.

(John Barth, Lost in the Funhouse)

I. Introduction

For a long time, Canadian law said that Canada had been legally unoccupied before its discovery by colonial powers, and that this was why Canadian law was valid in Canada.Footnote 1 This changed in 2014, when the Supreme Court of Canada rejected these propositions.Footnote 2 Several years later, a trial court pointed out the problem:

If the doctrines of discovery and terra nullius are indeed “legally invalid” or simply inapplicable in Canadian law, what then is the legal justification validating the assertion of Crown sovereignty over Indigenous peoples and Indigenous lands?Footnote 3

The court did not answer this question. Instead, it gestured at the fact that Canadian laws are practically entrenched, and added: “It is these same laws which provide legitimacy to this Court.”Footnote 4 I think the court meant that if it rejected Canadian sovereignty, it would undermine its own ability to make holdings at all. It would no longer be speaking as a court.

There’s a strong intuition that this is right: we cannot, while reasoning within a particular legal order, revise the basis of that legal order. I want to explore this intuition—not as it occurs in the Thomas and Saik’uz case specifically, or even in relation to Canadian sovereignty, but rather as a general claim about legal reasoning and legal orders. If the intuition is right, then the foundations of a legal order can never be revised by law. If they are changed, then that change is merely causal—in other words, a revolution.

This seems to have been the view of Hans Kelsen, Alf Ross, and HLA Hart. I think it is wrong. Following some historical and conceptual preliminaries, I will set out an argument in favor of the intuition, drawing on a parallel argument by Hartry Field about logic. Next, I explain why the argument should be rejected, suggesting that a legal order can undermine itself in a way that points to a new one. I conclude with two objections: one that denies that a legal order has “foundations,” and another that questions the point of using law to revise those foundations.

A. The Foundations of a Legal Order: Historical Context

Whether the foundations of a legal order can be revised by law has implications for an old issue in legal and political thought: whether a community is entitled to bring about foundational changes in its legal order, and if so, how. While this paper isn’t meant as a contribution to historical scholarship, a sense of this context will help in understanding the broader stakes. On one view, a legal order rests on a person or group of people who have authority and exercise it to create legal rules. On another view, a legal order rests on legal rules, since no person can have authority without it being conferred on them by a rule.Footnote 5

On the first view, even the most foundational legal rule can be changed if the people who created it so choose. For example, John Locke, who held that a government’s authority derived from the people who came together to form it, argued that under certain conditions the government is “dissolved” and the people have the right to establish a new one.Footnote 6 Jean-Jacques Rousseau, who held that legal authority derived from the general will, maintained that any law can be changed by the people collectively: “there neither is nor can be any kind of fundamental law binding on the body of the people.”Footnote 7 On this view, the question whether the revision is done by law is less important, because the authority at the foundation of the legal order is not constituted or constrained by a legal rule.Footnote 8 In this paper, I will set this view aside.

On the second view, legal authority is always conferred by a legal rule, so it matters whether the most foundational rule can be changed by law. The closest approach to this question before the 20th century is in the work of Immanuel Kant, who argued that a right to revolution is impossible:

The reason a people has a duty to put up with even what is held to be an unbearable abuse of supreme authority is that its resistance to the highest legislator can never be regarded as other than contrary to law, and indeed as abolishing the entire legal constitution. For a people to be authorized to resist, there would have to be a public law permitting it to resist, that is, the highest legislation would have to contain a provision that it is not the highestFootnote 9

For Kant, the fact that the people might choose change is not enough on its own; the question is whether this change is authorized by law. And there seems to be a conceptual obstacle to a legal rule authorizing its own revision.

Kant’s argument has had a mixed reception. Commentators accept that “there may indeed be something contradictory about the thought of a constitution’s containing a provision authorizing the suspension of its authority.”Footnote 10 But this concern is often dismissed as legalistic, focusing narrowly on the constitution’s role in positive law and neglecting its role in political action.Footnote 11 This dismissal has, in turn, been criticized for overlooking the deep normative grounds of Kant’s legalism.Footnote 12 My argument tackles the conceptual obstacle directly. I aim to show that it is possible for the highest law to provide that it is not the highest—not by authorizing resistance, but by authorizing its own revision.

Indeed, the need for this possibility seems to be implicit in Kant’s discussion of constitutional change. Kant recognized that it must be possible to change an existing constitution to bring about a more rightful state—say, turning a monarchy into a republic. He argued that change should take place “only through reform by the sovereign itself, but not by the people, and therefore not by revolution.Footnote 13 Revolution is “not change … but dissolution”; instead, the sovereign should make the reform “gradually and continually,” using existing legal forms.Footnote 14 If this is right, then it matters whether such a thing is possible. I aim to show that it is possible by presenting an argument against it and then explaining why the argument should be rejected.

B. Legal Reasoning and the Legal Point of View: Conceptual Background

Next, let me explain some of the concepts I’m working with. The argument I’m going to discuss is about legal reasoning. By “legal reasoning,” I don’t mean just any kind of reasoning that has the law as its subject-matter. Rather, I mean reasoning from “the legal point of view.”Footnote 15 We take up the legal point of view when we make statements about what the law requires or permits (or about more elaborate normative effects, such as conferring rights or immunities): for example, “A is obliged to pay damages to B.”

Those who take up the legal point of view need not personally believe that they should, or even have reason to, obey the law. For example, as Kelsen observes, an anarchist might make arguments about what the law does and does not require. When someone like this takes up the legal point of view, they express themselves in “detached” statements: statements which “state the law as a valid normative system,” but “do so from a point of view … to which they are not committed.”Footnote 16

We engage in legal reasoning when we offer grounds for legal propositions from the legal point of view. A clear example is when a court issues reasons for holding one way or another. Lawyers, law professors, and law students also engage in legal reasoning when they argue about, say, whether a certain action is prohibited under the Criminal Code or whether a certain statute is constitutional.

Legal reasoning is subject to constraints that ordinary reasoning is not. If we ask, in ordinary reasoning, whether A should pay damages to B, there are a wide variety of considerations we might bring to bear on the question. We might ask whether it would contribute to human flourishing or to total welfare. Reasonable people may disagree about these considerations and what weight they should have. In legal reasoning, by contrast, the question reduces to whether there is a valid rule, not overridden by any other valid rule, requiring A to pay damages to B and whether the interactions between A and B meet the conditions of the rule; if so, then, from the legal point of view, A should pay damages to B.

One reason why we might want to use legal reasoning to settle disputes is that we can agree on whether there is a valid rule and whether its conditions are met without agreeing on the answers to questions about human flourishing, total welfare, and so on.

C. The Argument

Here it is:

  1. 1 Legal reasoning presupposes a highest legal rule.

  2. 2 Legal reasoning can only support revising a rule by appeal to a higher rule.

  3. 3 Therefore, legal reasoning cannot support revising the highest legal rule.

I will discuss each premise in turn before explaining why the argument should be rejected.

Formally, the argument looks just like one that Hartry Field discussed in his 2008 Locke Lectures on rationally revising logic.Footnote 17 But the substance of the argument can be motivated in a law-specific way, by reflecting on what we are committed to when we engage in legal reasoning—and indeed, the main parts of the argument were developed decades before Field’s discussion, by writers like Kelsen, Ross, and Hart—as I’ll now try to show.

II. First Premise: Legal Reasoning Presupposes a Highest Legal Rule

A. Highest Legal Rule

For any legal rule, there must be a legal explanation of why it is a legal rule. The demand for such an explanation is a “standing possibility” in legal reasoning.Footnote 18 A legal explanation must appeal to a higher rule, such that the lower rule is valid because it meets some condition set out in the higher rule. For example, a court may simply apply a provision of the Criminal Code, but there is always the possibility of showing that the Criminal Code is valid by appeal to some rule conferring on Parliament the jurisdiction to pass such a law.

So, for any legal rule, there has to be a higher rule that explains why it is valid. This doesn’t yet show that there has to be a highest rule. As is often observed, it seems to be consistent with three possibilities:

  1. (i) An infinite series of rules, each one higher than the last

  2. (ii) A circular structure (e.g. A > B, B > C, C > D, D > A)

  3. (iii) A highest rule

I don’t think (i) is a plausible option.Footnote 19 The assertion that a legal order contains a certain rule should be grounded in the practices of participants. While it sometimes makes sense to posit that a legal order contains a rule that has never been used in legal reasoning, such positings should nevertheless be grounded in the rules that have been used—for example, when we say that a general principle is part of the law because it helps to make sense of several more specific rules. An infinite series of never-used rules is likely to violate this condition.

What about the circular structure? Mark Walters has proposed something like it in a discussion of cases about Canadian sovereignty (2016, 47):

Interpretive theories of law, in which the resolution of difficult legal points is thought to involve an interpretive oscillation between specific rules of law and the underlying principles of political morality that they presuppose in the search of a reflective equilibrium between rules and principles that is normatively coherent and compelling, cannot draw boundary lines that mark off some but not other foundational principles as beyond the interpretive exercise. … The sovereign foundation of Canada, in other words, is an interpretive concept open to ordinary legal analysis just like any other legal concept.Footnote 20

Walters suggests that, if we accept an interpretive or “circular” theory of law, then all rules are in principle open to legal question.Footnote 21

For now, I’m going to exclude the circular theory by stipulation. What follows can be taken as an attempt to set up and solve the problem of legal revision on the supposition that there is such a thing as the highest legal rule. This is not as narrow a project as it may sound. While the claim that there is a highest rule has most often been developed in positivist theories of law, it is consistent with a non-positivist theory, that is, a theory on which the validity of a legal rule depends on its merits and not only on its sources.Footnote 22 One only has to add that any rule, to be legally valid, must satisfy certain normative constraints. That is a recognizably nonpositivist claim, both historically and in contemporary work, and adding it would not undermine either the problem I set up or its solution.Footnote 23 Thus, even on the supposition that there is a highest rule, my account should be of interest to some non-positivists.

Moreover, I don’t mean to concede that interpretivism avoids the problem of revising the highest rule. I will return to this issue at the end of the discussion, after setting out my own account. I’ll suggest that, while the problem is less central for an interpretivist, it arises in a more subtle way, and the solution I propose should be of interest to an interpretivist too.

Let’s return to the argument. If every legal rule needs a legal explanation that appeals to some higher rule, and there can’t be an infinite series or a circle of rules, then there has to be a highest rule. How do we relate to that rule when we engage in legal reasoning?

B. Presupposing

The idea of a highest rule is similar to Kelsen’s basic norm and Hart’s rule of recognition. Kelsen argued that, while legal reasoning can determine the validity of any norm other than the basic norm, it has to presuppose the basic norm.Footnote 24 For Hart, the rule of recognition is grounded in the practices of legal officials; its existence is a “question of fact.”Footnote 25 However, in interpreting Hart’s remarks on this issue, it’s important to maintain our focus on the legal point of view. I don’t take Hart to suggest that, from the legal point of view, our practices validate the rule.Footnote 26 A court doesn’t say, for example, “the constitution is valid because officials generally treat it as such.” Rather, no question of validity arises with respect to the rule of recognition.Footnote 27 In Salmond’s words, “It is the law because it is the law, and for no other reason that it is possible for the law itself to take notice of.”Footnote 28 So Hart and Kelsen agree that in legal reasoning, we treat the basic norm as binding, whether or not we think it is binding.

This is analogous to Robert Stalnaker’s notion of presupposition. Presupposing is about proceeding on the basis of certain claims—treating them as correct—not about actually believing them. As Stalnaker puts it, “Presupposing is … not a mental attitude like believing, but is rather a linguistic disposition—a disposition to behave in one’s use of language as if one had certain beliefs, or were making certain assumptions.”Footnote 29

This attitude to the highest rule makes it possible to take up the legal point of view without being committed to it, like Kelsen’s anarchist: all you need to do is treat the right claims as correct in your reasoning. In cases where someone takes up the legal point of view without being committed to it, legal reasoning gains a similar character to reasoning within a fiction.Footnote 30

I mean this literally. We reason about what is required by the law in much the way we reason about what is true according to, say, Beowulf. Footnote 31 As with Beowulf, the content of the law is not limited to what is expressly set out in its texts; we can extend it by incorporating our knowledge and drawing inferences. It’s hard to set out precise rules governing this activity, but for present purposes, one limitation is especially important: we don’t incorporate or infer content that would spoil the story.Footnote 32 This manifests as a prohibition on asking “silly questions.” In Walton’s words:

The generation of fictional truths is sometimes blocked (if not merely deemphasized) just, or primarily, because they make trouble — because they would render the fictional world uncomfortably paradoxical. But usually more than charity is involved. Recognition of a question’s silliness, and so of one reason for disallowing or deemphasizing fictional truths that give rise to it, may depend on awareness of various demands to which the artist must respond.Footnote 33

For example, it would be silly to ask why Beowulf speaks in verse.

A similar phenomenon occurs in legal reasoning. If we’re presupposing the highest rule in our legal reasoning, then certain questions—questions about the foundations of the state, whose answers might threaten the highest rule—may be disallowed or deemphasized, in awareness of demands to which the court must respond.

III. Second Premise: Legal Reasoning Can Only Support Revising a Rule by Appeal to a Higher Rule

Standard examples suggest that, when we revise a legal rule in legal reasoning, we do so by appeal to a higher rule. A court may, for example, invalidate a statutory provision by appeal to the constitution. In doing so, the court appeals to a higher rule: the constitution overrides an ordinary statute. Or, instead, a court may revise a common-law rule—say, that intangible property cannot ground an action in conversion; in this case, the court is implicitly relying on the rule that courts are empowered to revise the common law.

The claim is that this is the only way of revising a rule in legal reasoning. Are there any alternatives? The most plausible one, I think, is that we might revise a rule by appeal to itself. For example, a constitution might contain an amending formula which applies to all provisions of the constitution, including itself. If this formula is used to amend itself, it looks like a rule is being revised by appeal to itself.

An argument from Alf Ross aims to show that this is impossible. Ross argues that “from a norm of competence, no norm contrary to it can be derived.”Footnote 34 The claim is that when we rely on a rule R1 to revise a rule R2, the revision of R2 is valid only if R1 is itself valid. For example, suppose a court strikes down a provision of the Criminal Code by appealing to an interpretation of the Charter. Footnote 35 If a higher court concludes that the interpretation of the Charter was wrong, then the striking down of the provision is also wrong. So the use of a legal rule to revise itself is self-undermining. Suppose we use R1, the highest rule, to revise R1. As a result, R1 is no longer law. But then the revision of R1 no longer has a legal basis. When we ask why R1 is no longer law, we cannot explain this by appeal to R1.

Of course, anyone who accepts this argument has to make some sense of cases like the amending formula mentioned earlier. In response to Ross, Hart suggests that the effect of a rule which provides for its own amendment is “to prescribe the use of a certain procedure until that procedure is replaced by a new procedure introduced by the old procedure.”Footnote 36 And Ross offers a similar solution. For Ross, if R1 provides for its own amendment, then R1 is not in fact the highest rule. The highest rule is something like:

Follow R1 until R1 is used to revise itself, in which case follow the revised norm until it is used to revise itself, and so on.

However, when we put it this way, we can see that this rule is not revisable by legal reasoning.Footnote 37 So it remains true that the highest rule cannot be revised by legal reasoning; apparent counterexamples just show that we had misidentified the highest rule.

If legal reasoning has to presuppose a highest rule and can only support revising a rule by appeal to a higher rule, then legal reasoning cannot support revising the highest rule. This seems to get us to the intuition we started with: that we cannot, while reasoning within a particular legal order, revise the basis of that legal order.

As Ross points out, this doesn’t mean the highest rule is immune to change. It can be changed—just not by law.

For just as one legal custom may be supplanted by another legal custom, so one basic norm may be supplanted by another. But the transition in both cases is not the outcome of a legal process; it is a fact, the sociopsychological fact that the community now accepts another custom as law or another basic norm as the corner-stone of its legal order.Footnote 38

Dramatic examples of such a transition are provided by cases where the highest rule of a legal order is replaced in a revolution.Footnote 39 In these cases, the change doesn’t receive any legal explanation; rather, a revolution happens, the state apparatus begins enforcing the new rules, and people start following those rules.

One such case, State v Dosso, concerned the 1958 coup by the first President of Pakistan, Iskander Ali Mirza. Mirza dissolved the legislatures and issued a declaration abrogating the constitution, while validating most laws previously in force. The defendants in Dosso sought to invoke their rights under the constitution, requiring the court to determine whether it was still valid. The court wrote:

It sometimes happens … that a Constitution and the national legal order under it is disrupted by an abrupt political change not within the contemplation of the Constitution. Any such change is called a revolution, and its legal effect is not only the destruction of the existing Constitution but also the validity of the national legal order. … If the attempt to break the Constitution fails those who sponsor or organise it are judged by the existing Constitution as guilty of the crime of treason. But if the revolution is victorious in the sense that the persons assuming power under the change can successfully require the inhabitants of the country to conform to the new regime, then the revolution itself becomes a law-creating fact because thereafter its own legality is judged not by reference to the annulled Constitution but by reference to its own success.Footnote 40

The court concluded that the revolution had been successful and thus that the old constitution had been abrogated. If the argument we’re considering is right, this is the only way to revise the highest rule in a legal order.

IV. Why the Argument Should Be Rejected

While the argument is a neat one, I don’t think it works. To see why, it’s helpful to start with an analogy from epistemology. By setting up a problem parallel to our legal one, we gain some conceptual resources which offer a solution in the legal case as well.

A. Revising a Highest Epistemic Norm: Field’s Account

Suppose that any thinker has some highest rational norm that they follow in forming and revising their beliefs—for short, a highest epistemic norm. Such a norm would be a rule or policy mandating that the thinker form or revise beliefs under certain circumstances. An act of forming or revising a belief is epistemically rational (I will simply say “rational”) when it follows the relevant epistemic norm. To say that an epistemic norm is the “highest” is to suppose that a thinker has a hierarchy of such norms, with lower norms deriving their force from higher ones. This is not, of course, the only way one might conceive of a rational thinker, but setting things up in this way parallels the way we have been conceiving of the structure of a legal order.

On this kind of picture, it might seem impossible to rationally reject the highest norm. As Field explains, it seems that rationally rejecting the highest norm would have to involve that norm itself, and

How could any norms—or any ones that an agent could follow—tell us to revise themselves? Wouldn’t following those norms require not following them? This seems incoherent, or at least to make following the rule impossible.Footnote 41

Put in our earlier terms, a norm can be rationally revised by appeal to a higher norm, but a norm cannot be rationally revised by appeal to itself. It seems to follow that the highest epistemic norm cannot be rationally revised. It might, of course, be revised, but only in a causal way—as a result of an arbitrary decision, or being hit on the head.

Field argues, however, that this line of thought is mistaken. Even if a norm can’t dictate its own revision, this

doesn’t rule out that by following a highest level norm N, we could be rationally led to conclude that we shouldn’t be following N, and should instead be following an alternative N*. … It only tells us that N wouldn’t then dictate switching from N to N*. That doesn’t mean that we wouldn’t make the switch—only that in making it, we wouldn’t be following N. It also doesn’t mean that we wouldn’t be rational in making the switch—only that were we to rationally make it, the rationality would not be explained by its being in accordance with norm N. Footnote 42

Field wants to distinguish between two different steps in revising a norm N. The first is concluding that one shouldn’t follow N (but should follow some other N* instead). The second is actually switching from N to N*.

At the first step, there’s no obstacle to using N to conclude that one should not follow N but rather N*. For example, suppose that classical logic is my highest epistemic norm.Footnote 43 (I’ll be working with this example in some detail, because it can be developed in a fairly realistic way, but I should note that nothing in what follows turns on the use of logic as the example. One can substitute any highest epistemic norm.) I might reason using classical logic, and conclude that classical logic is flawed, and that its flaws are not shared by, say, the “Strong Kleene” logic K3. Suppose that, using classical reasoning, I come to believe that some sentences are neither true nor false, and that classical logic fails to capture this, while K3 does. I conclude that I should replace classical logic with K3. My reasoning, if fully spelled out, might be classically valid.

At the second step, Field concedes that N cannot validate the switch from N to N*; the rationality of switching from classical logic to K3 cannot be explained by classical logic. However, Field wants to maintain that the switch can, nonetheless, be rational. How? I suggest the following: the switch is rational because the old norm undermined itself in a way that pointed to the new norm.Footnote 44 This fact is itself a reason for switching to the new norm. It is not, itself, an instance of the old norm, but is a reason grounded in a conclusion drawn under the old norm. The switch is rational because it is responsive to this reason.

It’s tempting to suggest a different interpretation: classical logic was simply not my highest norm. Rather, my highest norm was something like “follow classical logic unless it leads to a problem, in which case revise it and follow the revised logic instead.” This would be the analogue of Ross’s suggestion about law. On this proposed interpretation, my highest norm remains unchanged.

I think Field should reply as follows. The attribution of an epistemic norm to a thinker is supposed to have psychological reality: an attribution is correct only if that norm really does play a role in guiding the thinker’s reasoning. And the proposed norm lacks psychological reality. Before the switch, what played a role in guiding my reasoning was classical logic, not a conditional commitment to classical logic unless certain conditions were met. Before the switch, I treated classical logic as an exceptionless set of rules; I would have rejected the suggestion that it was only valid under certain conditions. The only thing that supports attributing to me the alternative norm is the fact that I switched, together with an insistence that every change be underwritten by a higher norm. But this insistence is what Field is calling into question.

Field’s discussion suggests that a revision of the highest epistemic norm can be epistemically rational—and, importantly, can be rational precisely because it responds to a reason generated by that norm itself, rather than by appealing to some external sense of rationality. These features of Field’s discussion make it particularly fruitful in thinking about the parallel legal problem.

B. How to Revise the Highest Legal Rule

Let’s return to the highest legal rule. So far, we’ve seen two accounts of what happens when the highest rule is apparently revised by appeal to itself. First, that this is a merely causal change, legally indistinguishable from a revolution.Footnote 45 Second, that the true highest rule is the one Ross proposed, and that it remains unrevised. Neither account leaves room for a genuinely legal revision of the highest rule.

The first proposal seems problematic. Participants in legal orders recognize the difference between using the highest rule to revise itself and a revolution. We should attempt to make sense of this difference before concluding that they are just mistaken. It’s true, of course, that people sometimes gesture at the efficacy of the legal order in explaining why they follow its rules, and that they might do this in the context of a legal change. In the Thomas and Saik’uz case, having raised the question of why Canadian law is valid, the trial judge goes on to point out that Canadian law is practically entrenched. More dramatically, we can imagine a judge pointing out that Canadian law is backed up by the systematic threat of state violence—it has the police and the military behind it—such that questioning its validity seems academic. But none of this seems to be legal reasoning. In fact, a judge would be very reluctant to say that Canadian law is valid because it is backed up by the threat of state violence. They would prefer to say that it has the police and the military behind it because it is the law. By contrast, in the context of a revolution, it is very natural to say that the new rules are law because they are being enforced. So it does seem like participants in a legal order recognize the difference between a revolution and a legal change.

The second proposal—that the highest rule is “follow the Canadian constitution unless it is revised, in which case follow the revised rule”—allows us to say that the highest rule is never revised. However, this comes at a cost. As I suggested earlier, the claim that the law contains a given legal rule is correct only if that rule is appropriately related to the practice of legal reasoning. The claim that the highest rule of Canadian law is the Canadian constitution is correct in this sense. Legal reasoners actually appeal to this rule in their reasoning. By contrast, the alternative rule is only supported by the fact that it solves a puzzle; it would not be identified by legal reasoners as the highest rule that they use. The cost of Ross’ approach, then, is to detach the highest rule from the practice of legal reasoning. This is confirmed by the fact that Ross’ solution could also be used to “legalize” a revolutionary transition.Footnote 46 We just have to claim that the highest rule of the legal order is “follow the constitution unless there is a revolution, in which case follow the revolutionary government’s orders.”

Field’s discussion suggests a third approach. To start, let’s distinguish two stages in revising our highest rule. First, there is the stage of concluding (while presupposing the highest rule) that that rule is not valid. For example, the court in Thomas and Saik’uz might have reasoned:

  1. 1 Canadian law is valid only if Canada was legally unoccupied when discovered by colonial powers.

  2. 2 Canada was not legally unoccupied when discovered by colonial powers.

  3. 3 Therefore, Canadian law is invalid.

The reasoning here presupposes the highest rule of Canadian law. This is how the court is able to assert the premises, which derive from Canadian court decisions. However, in reasoning within this presupposition, the court concludes that Canadian law is invalid.

This reasoning seems unobjectionable. There’s nothing impossible about using Canadian legal reasoning to conclude that Canadian law is invalid. If a legal order includes rules about the conditions of its own validity, then it may also include the conclusion that those conditions are not met. (In terms of our earlier comparison to reasoning within a fiction, such a legal order would be like a work of metafiction: a work whose fictionality is part of its fictional content.) The difficulty comes at the second step: can the court, reasoning legally, replace the highest rule with a different rule?

I think it can. To explain how, I want to distinguish two thoughts: first, that Canadian law can dictate its own replacement; second, that Canadian law can legally ground its own replacement. Canadian law cannot dictate its own replacement. Even the reasoning sketched above doesn’t show that it can. As soon as we abandon the highest rule of Canadian law, we also lose our basis for the premises, and so for the conclusion.

However, Canadian law can legally ground its own replacement. The emergence of the conclusion, within Canadian law, that Canadian law is invalid can provide a legal basis for replacing Canadian law with something else. In making the switch, the court is responding to a legal reason that emerged internally to Canadian law, rather than externally to it. Borrowing from Raz, one can say that the court is “making new law,” but is doing so “on the ground that this is justified and required by existing law.”Footnote 47 This is why such a switch would be different from a revolutionary change, where the change is only explicable from outside the legal point of view. In short, a change of the highest legal rule can be legal—and legal precisely because it responds to a reason generated by that rule itself, not by appeal to some external sense of legality.

This approach to the problem can also solve our initial puzzle about an amending formula which applies to all provisions of the constitution, including itself. What happens when this amending formula is used to amend itself—for example, if the amending formula requires a two-thirds majority of the legislature to amend the constitution, and if a two-thirds majority votes to change this to a three-quarters majority? In this case, the original amending formula supports the conclusion that the amending formula has changed. This fact provides a reason, internal to the legal order, for recognizing the new amending formula as valid: in my terms, the formula grounds, though it does not dictate, its own replacement. Importantly, such a reason only arises when the change was done in accordance with the original amending formula; if not—if, for example, it violates some condition in that formula—then we would have an “unconstitutional amendment,” which cannot be implemented legally.Footnote 48

This solution involves a change in a legal rule that is not itself validated by a legal rule. On one view of legal reasoning, this might be impossible: legal reasoning may be thought to be essentially rule-governed. In my view, by contrast, legal reasoning can involve responding to legal reasons, even where this is not validated by a legal rule. Still, I’m not suggesting that courts can appeal in an unconstrained way to whatever reasons bear on the issue before them. Rather, in cases like these, they’re responding to a specific legal reason that is generated by the legal rules—specifically, by the fact that those rules undermine themselves in a way that points to a new rule.

You might worry about why this legal reason would be normative. If the normativity of the legal reason rests on its being generated by the prior highest rule, then it seems to fail as soon as the prior highest rule is no longer binding. If the normativity of the legal reason derives from something external to the prior highest rule, then we seem to be back with a merely causal change. (After all, a revolution can also be justified by such reasons.) I think there is a third option: the legal reason is normative because the prior highest rule undermined itself in a way that pointed to the new rule. In particular, the prior highest rule leads to the conclusion that it is not the highest rule, but some other rule is. Such self-undermining is a fact to which it is rational to legally respond.

Does this show that the legal normativity in my account arises not from the legal rules but from some external source? I don’t think so. It’s true that rationality plays a role in recognizing that the highest rule has been used to undermine itself and point to a new rule. However, rationality plays this role in all legal reasoning: we use our ordinary understanding of how to reason, and our ordinary understanding of linguistic meaning, to derive legal conclusions from legal premises. This is true even on a positivist account, and it does not show that the rules of logic and language are therefore part of the content of the law; rather, as Raz said of the rules of voluntary associations, they are rules which “courts have an obligation to apply” although they “do not become part of the legal system.”Footnote 49 Even though rationality plays a role here, there is a crucial difference between revising the highest rule on the basis of a reason that arises under it and revising it because, for some external reason, one takes the change to be good or obligatory.

C. A Worked Example: Patriation

Before moving on to consider examples, I want to apply my account to a real example rather than the hypothetical ones we’ve considered so far. This is the “patriation” of the Canadian constitution from the United Kingdom in 1982. Like the constitutionally granted independence of Australia and New Zealand, the patriation of the Canadian constitution generated extensive theoretical debate, including the view that patriation was a revocable exercise of British sovereignty and the view that it was, legally speaking, a revolution.Footnote 50 I hope to show that my account can help make sense of what happened. I should flag that this example is more complex than our earlier ones. Patriation involved not only a transition from one highest rule to another, but also a fission of one legal system into two; moreover, on some views, international law also played a role in validating the transition. I am going to bracket these issues in my discussion, focusing solely on the change of highest rules in Canadian law.

In 1982, the Canadian constitution was transferred from being a British statute, revisable by the UK Parliament, to being Canadian law, subject to revision only in accordance with its own amending formula. In the Canada Act 1982,Footnote 51 the UK Parliament provided:

  1. (1) The Constitution Act, 1982 … is hereby enacted for and shall have the force of law in Canada and shall come into force as provided in that Act.

  2. (2) No Act of the Parliament of the United Kingdom passed after the Constitution Act, 1982 comes into force shall extend to Canada as part of its law.

The Constitution Act, 1982 and the Constitution Act, 1867, together with the Constitution Acts, are two of the main parts of Canada’s constitution.

During the patriation process, the province of Quebec brought a legal challenge, arguing that the process violated a convention that required its consent to patriation. By the time the case reached the Supreme Court of Canada, the patriation process had been completed. The court held: “The Constitution Act, 1982 is now in force. Its legality is neither challenged nor assailable. It contains a new procedure for amending the Constitution of Canada which entirely replaces the old one in its legal as well as in its conventional aspects.”Footnote 52 Quebec’s legal challenge was, therefore, moot.

During the patriation process, the highest rule of Canadian law changed. Before patriation, the rule that what the UK Parliament enacts is law was higher than any rule of Canadian law and was the basis for the validity of the Constitution Acts. Post-patriation, the Canadian constitution was the highest rule in Canadian law. Its legality was, as the court said, no longer assailable. Now, if we ask why the Canadian constitution is the highest rule, we can’t straightforwardly appeal to an enactment by the UK Parliament: the UK Parliament has no legislative power in Canada. But the change was not a merely causal one. The explanation for the change is not only that people in Canada would reject an enactment of the UK Parliament that purported to legislate for them. This would be true in a revolution as well, but no enactment by the UK Parliament would be involved. Unlike a revolution, patriation is explicable from the legal point of view. The switch from a British highest rule to a Canadian highest rule itself responded to a reason that arose under UK law.

A court, reasoning after such a switch of highest rules, can still access the legal reason for the switch and offer it as a justification to subjects. To do so, the court has to engage in suppositional reasoning: reasoning in which we suppose that p, even though we don’t take p to be true. If a court supposes the old highest rule to be valid, and can infer from this that the old rule is invalid and the new rule valid instead, then this can provide a legal reason for accepting the new highest rule.

For example, a Canadian court might reason:

  1. 1 Suppose that what the UK Parliament enacts is law in Canada.

  2. 2 The UK Parliament has enacted the Canada Act 1982, s 2.

  3. 3 Therefore, what the UK Parliament enacts is not law in Canada.

This line of suppositional reasoning presents a legal reason for concluding that the old highest rule can no longer be followed. Suppositional reasoning can also point towards the new highest rule:

  1. 1 Suppose that what the UK Parliament enacts is law in Canada.

  2. 2 The UK Parliament has enacted the Canada Act 1982, s 1.

  3. 3 Therefore, the Constitution Act, 1982 is law in Canada.

Contrast this with, say, a Pakistani court reasoning after the 1958 coup. If the court reasons under the supposition that the pre-coup constitution is the highest rule, it will have no reason to conclude that the pre-coup constitution is not the highest rule. Nor does it have any reason to conclude that the post-coup constitution is the highest rule. No explanation for the change is available from the legal point of view.

This solution makes sense of the difference between a revolutionary transition and a legal one, and it does so without being entirely detached from the practice of legal reasoning. Reasoning like the above could conceivably be attributed to a court.

V. Objection: Not Change but Continuity

I’m now going to turn to two major objections to my treatment of these issues. The first objection is that the puzzle I’ve been considering isn’t a problem for law; it’s a problem for one positivist picture of law. This picture sees law as a hierarchy of authority, and it is natural that in such a picture whoever or whatever is at the top of the hierarchy will have an exceptional and problematic status. This manifests in the fact that the highest rule seems to be beyond legal revision. If we can free ourselves of this positivist picture of law, the objection goes, we may no longer have a puzzle about the highest rule.

I don’t think it’s quite right to tie the problem to positivism: as I noted earlier, it is possible for a nonpositivist to hold that there is a highest legal rule. Rather, the problem is tied to the claim that there is a highest legal rule that determines the content of the law. This claim is rejected by interpretivists, who hold that the content of the law is determined by its best interpretation. In Ronald Dworkin’s version, legal reasoning involves ineliminable appeal to moral principles which both fit and justify the legal materials.Footnote 53 Dworkin argues that there is no rule of recognition which can “provide a test for identifying principles” and that principles cannot be incorporated into the rule of recognition without reducing it to triviality: thus, “if we treat principles as law we must reject the positivists’ first tenet, that the law of a community is distinguished from other social standards by some test in the form of a master rule.”Footnote 54 This is the tradition Walters draws on when he writes that a circular theory “cannot draw boundary lines that mark off some but not other foundational principles as beyond the interpretive exercise.”Footnote 55

One response to this challenge would be to concede that the puzzle of revising the highest rule is no puzzle for an interpretivist. After all, a solution to the puzzle would still be valuable, both for our understanding of those theories which are committed to a highest rule, and for weighing up the strengths and weaknesses of different theories. However, I think this would be too quick. While the puzzle is less central in an interpretivist account, it still occurs, albeit in a more subtle way.

To begin, I’ll draw on a parallel with epistemology to explain why, for an interpretivist, the idea of a legal revision of the highest rule is less central. Following this, I’ll explain how it still gets at something important.

In “Two Dogmas of Empiricism,” WVO Quine proposed a holistic epistemology in which no statement is immune from revision:

The totality of our so-called knowledge or beliefs, from the most casual matters of geography and history to the profoundest laws of atomic physics or even of pure mathematics and logic, is a man-made fabric which impinges on experience only along the edges…. Any statement can be held true come what may, if we make drastic enough adjustments elsewhere in the system. Even a statement very close to the periphery can be held true in the face of recalcitrant experience by pleading hallucination or by amending certain statements of the kind called logical laws. Conversely, by the same token, no statement is immune to revision.Footnote 56

Quine’s target was the idea that some statements—perhaps statements about the meanings of words, or statements of pure mathematics and logic—were immune to rational revision. In response, Quine suggested that our commitments form a web, holding each other in place through their rational relations. Some statements are central in the web, while others are at the periphery, but all are revisable; it’s just that revising a statement in the center will require further revisions throughout the web. Rational relations run in every direction: statements at the periphery are supported by those in the center, but statements in the center are also supported by those at the periphery.

Interpretivism conceives of a legal order in much the same way. Legal rules form a web, holding each other in place through their rational relations. At the periphery of the web, we find specific rules applicable in specific contexts, interpretations of particular words in particular statutes, and conclusions about fault, liability, and remedies in individual cases. Moving inwards, we find the organizing principles of different domains of law, and perhaps at the center, principles that ramify throughout the legal order as a whole. Nothing in the web is “beyond the interpretive exercise”; in other words, nothing is immune to revision.Footnote 57 But revising some fundamental rule might require far-reaching changes elsewhere. Importantly, rational relations run in every direction: a specific rule might follow from a more general one, but the general rule is also justified by the fact that it makes sense of many specific rules.

Looked at in this holistic way, the striking thing about both coups and constitutional amendments is how little changes. Typically, most legal rules, including most primary rules in Hart’s sense, remain intact. What changes is the “official story,” set out in secondary rules, about why those rules are valid.Footnote 58 This was true both of the patriation of the Canadian constitution and of the 1958 Pakistan coup: an everyday contractual dispute or criminal prosecution could proceed, for the most part, just as before.

Moreover, on the holistic picture, the new “highest rule” may be justified by the fact that it can stand in rational relations to many other existing legal rules. Not only did the newly patriated Canadian constitution and the post-coup Pakistani constitution leave existing rules intact, but they also validated those rules by reauthorizing them. On the hierarchical picture, the highest rule can confer validity on other rules, but nothing confers validity on it; on the holistic picture, any rule, including a constitutional one, can gain validity by helping to make sense of other rules.

For Quine, not every revision of our commitments changes the meanings of the words involved. If we revise one statement and hold the rest mostly fixed, the meanings of the words used remain unchanged. By contrast, if we have to revise our commitments on a wide range of issues, at a certain point, we might say that the meanings of the words used have changed. On this way of thinking, whether a word means the same thing before and after a revision is a matter of degree. Similarly, interpretivism makes it natural to say that not every legal change is a revolution. If a new rule leaves existing rules largely intact and stands in rational relations with them, there has been no revolution—even if the change had no legal basis. If a new rule leads to far-reaching revisions throughout the legal order, then it is a revolution, even if the change was the result of legal reasoning. The difference between a revolution and a legal change will be a matter of degree.

The holistic picture gets a lot right. The distinction between a disruptive change and one that maintains legal continuity cross-cuts the distinction between changes that have, and changes that lack, a legal basis.Footnote 59 It’s true that, if we look at legal systems before and after a change of the highest rule, we often see massive continuity, as the newly introduced legal rule takes up rational relations with much of the existing law. This is why, for an interpretivist, the issue of whether a revision of some fundamental rule is legal seems less central. The revision, even if it is not legal, may make little difference to the process of interpretation which determines the content of the law.

To see what is left out by the interpretivist account, let me return once more to Quine. In response to Field’s puzzle about changing our highest epistemic norm, a Quinean might argue that Field’s focus is in the wrong place. After all, it is possible to revise any commitment, even a statement at the center of the web of belief, while maintaining rational continuity. But—we might point out—this possibility doesn’t yet tell us whether a given revision is rational. Some revisions are rational, and others are not; to know whether a given revision is rational, we need to look at why it happened. If I revise my commitment to classical logic because I was hit on the head, then the change is not rational, even if my overall set of beliefs is more coherent than before. Field’s account aims to explain how a change of the highest epistemic norm can be rational in this sense.

Similarly, in a legal order, it matters whether a change has a legal basis, and so it matters whether the highest rule in a legal order can be revised legally. Either the act of changing the law is driven by legal reasoning or it is not. The fact that, the day after the change, we can still make sense of the legal order doesn’t speak to whether the change itself was a legal one. What my account shows is that, even when we’re talking about the highest rule, it is possible for the change to be a legal one. This matters to those participants in the legal order who are committed to changing the law only in accordance with law.

David Dyzenhaus has recently distinguished “dynamic” and “static” theories of law: a dynamic theory “includes the dynamic process of legal change within the scope of philosophy of law while a static theory consigns change to some extra-legal space.”Footnote 60 In other words, a dynamic theory aims to explain how legal change is itself governed by law. Within a dynamic theory, it matters a great deal whether a given change can be effected in accordance with law, or only by a revolution, even if after the fact the revolutionary change is justified by its coherence with existing law.

What, then, is the interpretivist account of the types of legal changes I’ve been concerned with in this paper? Let’s grant to the interpretivist that these are not changes to the highest rule, because there is no such thing. Still, there are changes to a legal rule. Prior to 1982, the UK Parliament could legislate for Canada, and after 1982 it could not; it follows that a rule of Canadian law changed. The question is, did this change take place through law, or not? It is true that continuity was largely maintained, but I don’t think an interpretivist should (or would) hold that any change which maintains continuity is therefore a lawful one. After all, the 1958 Pakistan coup also maintained continuity, but it was not a lawful change.

A better interpretivist account might focus on the moral principles which underlie the law. If the Canadian legal order pre-1982 contained principles about the value of democracy and self-determination, and if patriation brought the positive law into better alignment with those principles, perhaps that shows that the change was lawful. By contrast, if the 1958 Pakistan coup brought the legal order into discord with its own principles, perhaps that makes it unlawful.

While this account has some appeal, and I won’t purport to comprehensively respond to it here, let me say why I think it is incomplete. Interpretivism fits well with the style of reasoning of courts in common law systems. If such a court had to deal with a challenge to a fundamental legal rule on common law grounds, it seems reasonable to say that the court’s decision would be lawful to the extent that it accorded with the principles underlying the legal order. However, no modern legal system is purely common law; all provide for change to legal rules by way of legislative action—either ordinary legislation or processes of constitutional amendment. This is what happens when the UK Parliament acts to limit its own legislative authority, or when an amending formula is used to amend itself. It is in these cases that the interpretivist account falls short: unlike a court, a legislature can revise the law in a way that creates massive discontinuity with earlier law, as long as the law is within the legislature’s jurisdiction and consistent with constitutional constraints. When a change to a fundamental legal rule is made by legislation, it is wrong to say that the change is lawful to the extent that it accords with the principles underlying the legal order (except in the trivial sense that giving effect to legislation is such a principle). A law might maintain continuity of principle while being legally invalid, or the opposite. The account I have proposed explains when, and why, a change to a fundamental legal rule can be legally valid when carried out by appeal to that very rule itself.

VI. Objection: Legal Reasoning and Legalism

This may raise a second objection. Suppose I’ve shown that the highest rule of a legal order can be revised in legal reasoning. Even so, such a revision will have a practical effect only if the revised legal order is effective—if people tend to follow its rules. This will depend on the historical, social, and political context. The successful patriation of the Canadian constitution, for example, cannot be understood solely from the legal point of view, but also has to be understood externally. As a result, you might wonder why it matters if the change was a legal one or just a political one: isn’t the crucial fact that the change happened and was effective?

Judith Shklar characterized legalism as “the ethical attitude that holds moral conduct to be a matter of rule following, and moral relationships to consist of duties and rights determined by rules.”Footnote 61 In her study of the topic, Shklar raised the question to what extent legalism is appropriate to topics such as international politics and war crimes. Similarly, we might wonder whether legalism is appropriate to a topic like the validity of Canadian law vis-à-vis Indigenous law, or the effect of a revolution.

It’s important not to stack the deck in favor of legalism here: the alternative to resolving a dispute legally need not be deciding it by a contest of power. A dispute might be resolved in a reasoned way, though the reasons in question are not legal reasons. A collective political debate about what the highest rule should be, followed by a collective decision to follow that rule (in other words, a constitutional convention), is not a mere contest of power.

Earlier, I suggested that we care about the legal point of view because of its neutrality: it allows us to settle disputes by appeal to which rules are valid rather than by resolving major normative questions. This allows us to live together despite disagreeing on the answers to those questions. However, this justification is usually offered in the context of ordinary interpersonal disputes. Does it extend to disputes about the highest rule of the legal order?

To answer this question, I want to distinguish two kinds of legalism. A rigid form of legalism says that the only reasons relevant to whether to change a legal rule are specifically legal reasons. A more open-minded legalism says that all of the reasons—legal and nonlegal—are relevant to whether the change should take place, but in some contexts, participants in the legal order are only permitted to consider legal reasons. As a result, if a change is going to happen, it should happen in a way that is legally valid, such that it can be understood even if we only consider legal reasons.

I don’t think rigid legalism is plausible. As John Gardner puts it in discussing a similar view in private law:

The reasons which judges qua judges may rely upon are not necessarily, or even imaginably, the only reasons in favour of and against their decisions. And like all other activities, private law decision-making is rationally answerable to all the various reasons which militate in favour of and against its decisions, and not only those to which the decision-maker herself is permitted to have motivational access.Footnote 62

The fact that judges are—justifiably—restricted to legal reasons does not show that other reasons are irrelevant to evaluating the decisions they make. Think of our examples: the validity of Canadian law vis-à-vis Indigenous law can’t only be evaluated by looking at the Doctrine of Discovery or rules of British imperial law. The effect of a radical political change shouldn’t depend only on whether it satisfies a constitutional amending formula.

The more open-minded kind of legalism, however, is much more plausible. All of the reasons are relevant to whether the change should take place, but if it is going to take place, it should be done in a way that is legally valid, such that the change can be understood even in contexts where we only look at legal reasons. For example, the reasons why Indigenous law should be part of the foundations of Canadian law are not simply those set out in Thomas and Saik’uz. Footnote 63 But if Indigenous law were recognized by Canadian courts through an argument like the one in Thomas and Saik’uz, that would be legally valid under Canadian law, such that the change could be understood even if we only looked at legal reasons. Similarly, the reasons for patriating the Canadian constitution from the UK were large and complex, but given that it was done through an act of the UK Parliament, we can recognize the validity of the change even if we only look at legal reasons.

This can legitimize the change even for those who disagree on where the balance of other reasons lies, if they take valid law to be authoritative.Footnote 64 Someone who disagrees that Canada should recognize Indigenous law may still accept that the change was justified by a legal argument and is therefore binding. Someone who disagrees that the Canadian constitution should have been patriated may still accept that, under the Canada Act, 1982, patriation was legally valid. A legal change of the highest rule has the prospect of bringing along people who would reject a merely causal change because they disagree on where the balance of reasons lies.

This fact, that legal reasoning can justify a conclusion independently of the broader balance of reasons, is not an unequivocal good. It allows courts and other legal institutions to do bad things—even things that they recognize to be bad—merely because they are supported by legal reasons.Footnote 65 My claim is only that, where the change in question is a good one, implementing it in a legally valid way can legitimize it across deep disagreement. What I have suggested in this paper is that this holds for the highest legal rule too.

VII. Conclusion

I have argued that the highest rule in a legal order can be changed by law, not only by revolution, just as a thinker’s highest epistemic norm can be changed by reasoning. I’ve also argued that, even on an understanding of legal order on which there is no highest rule, my account can explain what makes certain changes of fundamental rules lawful. Finally, I offered some tentative comments on why this conclusion might matter: in a community committed to changing the law through law, revising the highest rule legally can help the new rule gain acceptance even among those who disagree on the merits.

Acknowledgments

I benefited from presenting this work at the Legal Reasoning and Adjudication Workshop at the University of Wisconsin-Madison, the Canadian Section of the International Association for Philosophy of Law and Social Philosophy Webinar, the University of Western Ontario Faculty of Law, and the American Philosophical Association (Central) Annual Meeting. Special thanks for comments and questions to Emad Atiq, Nina Varsava, Adrian Liu, Michael Giudice, Jorge Fabra Zamora, Kevin Toh, Brian Leiter, Bill Watson, Catherine Rioux, Alma Diamond, Joanna Langille, Zoë Sinel, Francesco Ducci, Ryan Liss, Alan Miller, Rory Gillis, Andy Yu and two anonymous referees.

Competing interests

The author declares none.

References

1 Calder v British Columbia, [1973] SCC 313, citing Johnson v M’Intosh, 21 US (8 Wheat) 543 (1823).

2 Tsilhqot’in Nation v British Columbia, 2014 SCC 44 at para 69. For discussion see John Borrows, The Durability of Terra Nullius: Tsilhqot’in Nation v British Columbia, 48(3) UBC L Rev 701, 703 (2015).

3 Thomas and Saik’uz First Nation v Rio Tinto Alcan Inc, 2022 BCSC 15 at para 194.

4 Thomas at para 202.

5 Scott Shapiro, Legality 42 (2011).

6 John Locke, Two Treatises of Government, Book II, ss 88-89, 149-150, 220-222 (1690).

7 Jean-Jacques Rousseau, The Social Contract, I.7; cf III.18.

8 This authority is sometimes described as a “constituent power”: Emmanuel-Joseph Sieyès, What is the Third Estate?, in Oliver W Lembke & Florian Weber, Emmanuel-Joseph Sieyès: The Essential Political Writings (2014). For criticism, see David Dyzenhaus, Constitutionalism in an Old Key: Legality and Constituent Power, 1(2) Global Constitutionalism 229 (2012) and Sergio Verdugo, Is It Time to Abandon the Theory of Constituent Power?, 21(1) ICON 14 (2023).

9 Immanuel Kant, Mary J Gregor trans, Practical Philosophy Ak. 6:320 (1996), emphasis added. Kant makes a similar argument at Ak. 8:303 in commenting on the British constitution: “that the constitution should contain a law for such a case authorizing the overthrow of the existing constitution, from which all particular laws proceed … is an obvious contradiction.”

10 Katrin Flikschuh, Reason, Right, and Revolution: Kant and Locke, 36(4) Phil. & Pub. Affs. 375, 381 (2008); cf. Thomas E Hill Jr, A Kantian Perspective on Political Violence, 1 J Ethics 105, 115 (1997).

11 Lewis W Beck, Kant and the Right of Revolution, 32(3) J. Hist. Ideas 411, 414 (1971); Hill, supra note 10, at 115.

12 Flikschuh, supra note 10, at 396.

13 Kant, supra note 9, at Ak. 5:322.

14 Id., at Ak. 6:340.

15 Joseph Raz, Legal Validity, 63(3) Archiv für Rechts- und Sozialphilosophie 339 (1977); Scott Shapiro, Legality 186 (2011).

16 Joseph Raz, The Purity of the Pure Theory, 35(138) Revue Internationale de Philosophie 441, 453 (1981).

17 Hartry Field, Logic, Normativity, and Rational Revisability (2008), https://www.philosophy.ox.ac.uk/john-locke-lectures#collapse5149811.

18 HLA Hart, The Concept of Law 104 (1961).

19 However, in epistemology and metaphysics, where parallel problems arise, this option has been defended: John Turri & Peter D Klein, eds, Ad Infinitum: New Essays on Epistemological Infinitism (2014); Allison Aitken, No Unity, No Problem: Madhyamaka Metaphysical Indefinitism, 21(31) Philosophers Imprint 1 (2021).

20 Mark Walters, “Looking for a Knot in the Bulrush”: Reflections on Law, Sovereignty, and Aboriginal Rights, in Patrick Macklem & Douglas Sanderson, eds, From Recognition to Reconciliation: Essays on the Constitutional Entrenchment of Aboriginal and Treaty Rights 35, 47 (2016).

21 Mark Walters, Toward the Unity of Constitutional Value—Or, How to Capture a Pluralistic Hedgehog, 63(2) McGill L.J. 419.

22 John Gardner, Law as a Leap of Faith 21 (2012).

23 See Thomas Aquinas, Summa Theologica, I-II, q. 90, a. 2; q. 95, a. 2 (1274); Emad Atiq, Contemporary Non-Positivism (2025).

24 Hans Kelsen, The Concept of the Legal Order, 27(1) Am. J. Juris. 64, 83 (1982).

25 Hart, supra note 18, at 106, 245.

26 See Kevin Toh, Law, Morality, Art, the Works, in Luka Burazin, Kenneth Einar Himma, and Corrado Roversi, eds, Law as an Artifact 62, 70 (2018); David Dyzenhaus, The Long Arc of Legality: Hobbes, Kelsen, Hart 298 (2021).

27 Hart, supra note 18, at 105.

28 John Salmond; Glanville Williams, ed, Salmond on Jurisprudence 155 (10th ed 1947).

29 Robert Stalnaker, Pragmatic Presuppositions, in Context and Content 47, 51–52 (1974).

30 Adam Perry, The Internal Aspect of Social Rules, 35(2) Oxford J. Legal Stud. 283, 299 (2015); Robert Mullins, Presupposing Legal Authority, 42(2) Oxford J. Legal Stud. 411, 433 (2022).

31 Kendall Walton, Mimesis as Make-Believe (1990); Toh, supra note 26.

32 Richard Woodward, Truth in Fiction, 6(3) Philosophy Compass 158, 161–163 (2011); J Alexander Bareis, Fictional Truth, Principles of Generation, and Interpretation. Or: Why It Is Fictionally True That Tony Soprano Was Shot Dead, in J Alexander Bareis & Lene Nordrum, eds, How to Make Believe. The Fictional Truths of the Representational Arts 165, 171 (2015).

33 Walton, supra note 31, at 183.

34 Alf Ross, On Self-Reference and a Puzzle in Constitutional Law, 78(309) Mind 1, 21 (1969). Ross also makes an argument about self-reference: id., at 4. The argument is that a rule that provides for its own revision would say something like “this law can be amended in accordance with the following conditions,” and that this self-reference makes the rule meaningless. Hart objected that a law might well refer to itself as well as other rules, for example “all laws, including this one, can be amended in accordance with the following conditions”: HLA Hart, Self-Referring Laws, in Essays on Bentham 175 (1983). See also Joseph Raz, Professor A. Ross and Some Legal Puzzles, 81(323) Mind 415, 419 (1972) and Norbert Hoerster, On Alf Ross’ Alleged Puzzle in Constitutional Law, 81(323) Mind 422 (1972).

35 Canadian Charter of Rights and Freedoms, Part I of the Constitution Act, 1982, being Schedule B to the Canada Act 1982 (UK), 1982, c 11.

36 Hart, supra note 34, at 186.

37 Ross, supra note 34, at 24.

38 Id., at 4.

39 State v Dosso, PLD 1958 SC 533; Uganda v Commissioner of Prisons, ex parte Matovu, [1966] EA 514; and see also Madzimbamuto v Lardner-Burke, [1968] UKPC 18. For discussion, see Vidya Kumar, International Law, Kelsen and the Aberrant Revolution: Excavating the Politics and Practices of Revolutionary Legality in Rhodesia and Beyond, in Nikolas M Rajkovic, Tanja E Aalberts & Thomas Gammeltoft-Hansen, eds, The Power of Legality: Practices of International Law and their Politics (2016).

40 Dosso, at 538–539.

41 Hartry Field, Lecture 6: The Revisability Puzzle Revisited, 8 (May 28, 2008) (https://media.philosophy.ox.ac.uk/assets/pdf_file/0013/2182/Lecture_6.pdf). Given that Field’s Locke Lectures have not been published, I am quoting from the handouts. I’m not claiming that Field would currently accept these claims; rather, I’m only presenting the view expressed in the lectures and handouts.

42 Field, supra note 41, at 9.

43 There are various questions about how a formal logic might be elaborated normatively, but I will bracket these complexities here. See John MacFarlane, In What Sense (If Any) Is Logic Normative for Thought? (2004) (http://johnmacfarlane.net/normativity_of_logic.pdf); Florian Steinberger, Three Ways in Which Logic Might be Normative, 116(1) Philosophy 5 (2019).

44 Field is not very helpful on this point. Given his noncognitivism, he says there is no fact of the matter about why the switch is rational: calling it rational is just expressing approval: Field, supra note 41 at 20; but see Hartry Field, Epistemology from an Evaluativist Perspective, 18 Philosophers Imprint 1, 20 (2018).

45 HWR Wade, The Basis of Legal Sovereignty, 13(2) Cambridge L.J. 172 (1955).

46 Virgílio Afonso da Silva, A Fossilised Constitution?, 17(4) Ratio Juris 454, 467 (2004).

47 Joseph Raz, The Morality of Law 172 (1986); cf Joseph Raz, The Inner Logic of the Law, in Joseph Raz, Ethics in the Public Domain (1994).

48 Yanif Roznai, Unconstitutional Constitutional Amendments: The Migration and Success of a Constitutional Idea, 61 Am. J. Compar. L.657 (2013).

49 Joseph Raz, The Identity of Legal Systems, 59 Calif. L. Rev. 795, 811 (1971); cf Joseph Raz, Between Authority and Interpretation: On the Theory of Law and Practical Reason 203–204 (2009).

50 For a detailed survey of these debates, see Peter Oliver, The Constitution of Independence: The Development of Constitutional Theory in Australia, Canada, and New Zealand (2005).

51 1982 c. 11, s 2

52 Re: Objection by Quebec to a Resolution to amend the Constitution, [1982] 2 SCR 793 at 806.

53 Ronald Dworkin, Laws Empire 255 (1986).

54 Ronald Dworkin, Taking Rights Seriously 43–44 (1977).

55 Walters, supra note 20, at 47.

56 WVO Quine, Two Dogmas of Empiricism, 60(1) Phil. Rev. 20, 39 (1951).

57 Walters, supra note 20, at 47.

58 On the official story, see William Baude & Stephen E Sachs, The Official Story of the Law, 43(1) Oxford J. Legal Stud. 178, 179 (2023). For related discussion of how a population “might not even be aware that they have just lived through a revolution,” see Gardner, supra note 22, at 284.

59 By “continuity,” I mean that the content of the law by and large remains intact before and after a change. I am not referring to the metaphysical question of whether and in virtue of what a legal system is the same legal system at two points in time: Joseph Raz, The Concept of a Legal System 187 (1980). However, continuity in this metaphysical sense is also connected to the discussion in this paper. Some theorists take metaphysical continuity to depend on a system’s legal norms, while others take it to depend on social practices or institutions: Li-Kung Chen, The Continuity of a Legal System, 86(2) Mod. L. Rev. 307, 365. The cases of fundamental change which I’m dealing with in this paper have been taken to be problematic for norm-centred accounts: Chen, supra, at 373–374. I don’t mean to take any position on the metaphysical question here, but my account offers resources to norm-centered accounts for addressing this problem.

60 Dyzenhaus, supra note 26 at 20.

61 Judith Shklar, Legalism 1 (1964).

62 John Gardner, The Purity and Priority of Private Law, 46(3) U. Toronto L.J. 459, 464 (1996).

63 The argument from Thomas and Saik’uz is about a problem internal to Canadian law which suggests that Canadian law should make room for Indigenous legal orders. But those legal orders have their own points of view on their authority and its relation to Canadian law. A full treatment of the issue, then, would have to ask not only what Canadian law says about it but also what Indigenous law says about it. The latter is not a question I am in a position to answer; for relevant discussion, see John Borrows, Wampum at Niagara: The Royal Proclamation, Canadian Legal History, and Self-Government, in Michael Asch, ed, Aboriginal and Treaty Rights in Canada: Essays on Law, Equality and Respect for Difference (1997) and Aaron Mills/Waabishki Ma’iingan, What Is a Treaty? On Contract and Mutual Aid, in John Borrows & Michael Coyle, eds, The Right Relationship: Reimagining the Implementation of Historic Treaties (2017). For reasons suggested in the text, I think it is still worth knowing what arguments can be made from within Canadian law.

64 Fernanda Pirie, The Anthropology of Law 134 (2013).

65 This is why certain authoritarians favour using law to implement their decisions: Kim Lane Scheppele, Autocratic Legalism, 85 U. Chi. L. Rev. 545, 563 (2018).