I. Introduction
Take the following claim:
X has a moral right to P.
Now replace “moral” with “legal,” such that
X has a legal right to P.
This brings us to a question that attracts the interest of general jurisprudence. How do they differ?
Clearly, they must in some respects. X can have a right to P without having a legal right to P. Moral rights exist in virtue of promises made among friends. They are not all, or even mostly, legal rights. Those who break a promise to meet for coffee do not necessarily commit a legal wrong.
At the same time, many have felt the force of a powerful intuition that the two are, in an important respect, about the same sort of thing. When we say X has a moral right to P, we mean X is entitled to P in a sense that fundamentally matters. To say X has a legal right to P is to invoke that very same sense of importance.
Some square this circle by saying “legal” refers to a species of the moral. There are several variants of this view, commonly known as legal antipositivism. In one, legal rights are specified as only those moral rights that arise in the right way.Footnote 1 In another, legal rights are specified as only those moral rights enforceable in court.Footnote 2 These variants of antipositivism share the following features. First, although all legal rights are moral rights, not all moral rights are legal rights. Second, legal rights straightforwardly correspond to moral significance. Legal rights are a kind of moral right. They are important in precisely the same way all moral rights are.
A common concern is that this inadequately separates the legal from the moral. Antipositivists say all legal rights are moral rights, even as they accept that not all moral rights are legal rights. Many find this unacceptable. Not all legal rights, they insist, are moral rights. On this view, we can have a legal right to P despite lacking any moral entitlement to it.
The mystery is how to reconcile this with the intuition that claims to legal rights are about moral rights. How can legal rights potentially diverge from the content of moral rights in this radical fashion, yet still concern a fundamental moral entitlement? The leading answer is to understand statements about legal rights in terms of a prefixed proposition. Known as legal perspectivalism, it translates “X has a legal right to P” as
According to law X has a moral right to P.
This is a claim about which rights exist from the legal perspective. So legal rights are best understood as claims “by law” regarding the content of moral rights.Footnote 3
Many assume perspectivalism is compatible with a popular way of thinking about law. It goes:
Hartian Orthodoxy. Some norm N is part of the law only if, and in virtue of, N being ultimately recognized or validated by the rule(s) of recognition.Footnote 4
For Hart, the rule of recognition is a social rule to which legal officials adopt a critical attitude. By this, he means these officials use the rule as a standard by which to evaluate the conduct of themselves and others. In this way, the rule of recognition is a rule which legal officials accept and practice.
Hartian Orthodoxy is committed to the following two claims. First, what norms are part of the law (and thus “legal”) is coextensive with what the rule of recognition ultimately validates.Footnote 5 Only the rule of recognition, at the most fundamental level, determines which norms are part of the set of legal norms. Call this Exhaustive Determination. Second, a certain social practice of legal officials exclusively determines the content of the rule of recognition. The rule’s content is wholly contingent and dependent on that practice. There is no content that the rule must have, simply given the kind of thing law is.Footnote 6 Call this Socially Given.
Here we show that perspectivalism is incompatible with Hartian Orthodoxy. If perspectivalism is true, then either Exhaustive Determination or Socially Given is false. There is now a choice. Those attracted to the advantages of perspectivalism will need to jettison Hartian Orthodoxy. Conversely, those who wish to insist on Hartian Orthodoxy will need to make do without perspectivalism.
We arrive at this position through a close consideration of Adam Perry’s recent objection to perspectivalism.Footnote 7 As Perry shows, naïve perspectivalism cannot explain the validity of certain legal arguments. We agree this poses a serious problem. Their validity is an explanatory desideratum that perspectivalism must accommodate. But our conclusions diverge. Perry says no version of perspectivalism can survive this desideratum. We disagree. Here, however, we show that once perspectivalism is revised in accordance with the desideratum, it becomes incompatible with Hartian Orthodoxy. Those committed to Hartian Orthodoxy can therefore read our argument as a different way to reject perspectivalism. Alternatively, for those committed to perspectivalism, our argument calls for a reconsideration of Hartian Orthodoxy.
The paper proceeds as follows. To start, we lay out the desideratum that motivates the rest of our arguments and illustrate how perspectivalism can accommodate it (§2). Then we show that perspectivalism, suitably revised to accommodate the desideratum, is incompatible with Hartian Orthodoxy (§3). After clarifying the scope of our conclusion (§4), we consider various objections (§§5–6). A brief conclusion follows (§7).
II. The Desideratum
Perry rejects perspectivalism. He does so because it cannot account for the validity of certain inferences we can make using legal propositions. Specifically, it cannot account for inferences that mix legal and nonlegal claims. He offers the following example:
(1a) and (1b) are legal claims, while (1c) is not. So this is a mixed argument in Perry’s sense. It also appears valid. But suppose we were to translate the argument along perspectival lines.
Now the argument is invalid. We cannot infer from some conditional being true in f, and the antecedent of that conditional being true in f, that the consequent of the conditional is in fact true.Footnote 8 Call this form of argument an export argument, for it uses two legal claims to infer a nonlegal claim.
We will return later to why Perry thinks perspectivalism cannot account for the validity of this argument. But let’s start with a point of agreement: perspectivalism needs an answer. It is a theory of the meaning of legal terms. A desideratum for any such theory is its ability to explain plausibly valid arguments, like (1a)–(1c), without rendering them irreparably invalid.
The burden is therefore on perspectivalism to show how we can validly draw inferences in mixed arguments. The task is to either (i) render the arguments valid, or (ii) explain why they are invalid despite, at first glance, appearing valid. Perry says perspectivalism can do neither. We show otherwise. But it requires perspectivalism to depart from Hartian Orthodoxy.
How, then, can perspectivalism explain mixed arguments? One approach starts by adding the prefix “according to law” to (2c). The export argument, suitably revised, would go:
This alone cannot render the argument valid. For that, perspectivalism would need to close the legal perspective in accordance with some inference rule.Footnote 9 We show in previous work the existence of such a rule.Footnote 10
Conditional Legal Closure (CLC). If according to law [q therefore p], and according to law q, then according to law p.
Now, (2c*) follows. Let p be “Jane is over 18.” Let q be “Jane has a moral right to vote.” Given (2a), it is true that according to law q, therefore p. So, given Conditional Legal Closure and (2a)–(2b), we can conclude that, according to law, Jane is over 18.Footnote 11
III. The Incompatibility
So far, we have shown that, for perspectivalism to accommodate the validity of mixed arguments, it must endorse a closure principle. Now we turn to how this renders perspectivalism incompatible with Hartian Orthodoxy.
To be clear, we have yet to show any incompatibility. For the addition of a closure principle, with respect to export arguments, it cannot touch Hartian Orthodoxy. Why? Take once more our formulation of Hartian Orthodoxy, which we restate for convenience.
Some norm N is part of the law only if, and in virtue of, N being ultimately recognized or validated by the rule of recognition.
This is primarily concerned with when some norms are part of the law. Not, crucially, what descriptive facts are true according to law.
Hence the lack of tension. What the closure principles make true in export arguments are not legal norms. Rather, they make true in law certain descriptive facts. Understood this way, perspectivalism does not yet pose a threat to Hartian Orthodoxy. But the threat isn’t too far away.
Recall the desideratum requiring perspectivalism to have the ability to validate mixed arguments. For now, we focused solely on one kind of mixed argument, namely export arguments. Once we use a closure principle to validate another kind of mixed argument, however, the incompatibility between perspectivalism and Hartian Orthodoxy comes to the surface.
This, too, is a mixed argument since it has both legal and nonlegal claims. But this time, the argument uses a legal claim and a nonlegal claim to arrive at a legal conclusion. Call this an import argument.
Such arguments are prolific in legal practice. We often infer from some general legal norm (3a), coupled with some facts about the world (3b), to arrive at a specific legal conclusion (3c). Indeed, this mode of inference is common even outside legal practice. Suppose a witness sees her neighbor perform some action, perhaps taking property belonging to another without authorization. Even those who lack legal training often have a general grasp that something along these lines violates a legal norm against theft. Now, say the witness were to draw the inference that her neighbor committed theft. The problem is that, absent a closure principle, perspectivalism lacks a straightforward way to validate this sort of inference.
This argument, translated along perspectival lines, is invalid.Footnote 12 Moreover, it seems part of the desideratum we initially set out to vindicate. How could a closure principle validate it?
The first step is the same as the export argument. There (2c) had to be prefixed and turned into (2c*). Here (4b) needs to be prefixed and turned into (4b*), such that it reads “according to law, Jane is over the age of 18.**Footnote 13 Once the swap is made, Conditional Legal Closure validates the inference. Unlike export arguments, the conclusion of an import argument is a legal norm. Footnote 14 It is a moral norm true according to law. It is therefore the sort of thing with which Hartian Orthodoxy is concerned.
What, precisely, is the tension between import arguments (so validated) and Hartian Orthodoxy? Consider the function of closure principles. They seem to make P part of the law in virtue of how it relates to other statements which are true according to law. Consider:
It looks like (4c) is part of the law. Thus, it falls under Exhaustive Determination. But why is (4c) part of the law? Because of a closure principle, which makes it true in law. This presents a dilemma for perspectivalism when it comes to Hartian Orthodoxy. There are two options.
First, you may think the closure principle is not part of the rule of recognition. If so, whether N is part of the law is not only determined by the rule of recognition. It is also determined by the inference rule under which the legal perspective is closed. This straightforwardly rejects Exhaustive Determination.
Second, you may think the closure principle is part of the rule of recognition. This will mean we must reject Socially Given. For the truth of the closure principle itself, as opposed to its implications, is not dependent on the practice and acceptance of legal officials. It is simply a metaphysical truth that a closure principle applies.
To be sure, which statements are true according to law will fundamentally be sensitive to social practice. What is insensitive is there being a closure principle in virtue of which the consequent is true according to law if social practice is such that: (i) the conditional is true according to law, and (ii) the antecedent is true according to law.
Could the closure principles be entirely subject to social practice? If so, it would be possible for a legal system to lack a closure principle. That is, for a closure principle not to apply to the legal perspective of that jurisdiction. But then the following import argument, relativized to that jurisdiction, would be invalid:
We find this unacceptable.Footnote 15 There is a strong sense that (3a)–(3c) and its relativized variant (3a*)–(3c*) are valid. The desideratum with which we are concerned just is the need to account for how these arguments (or something easily confused with them) could be valid, and therefore necessarily valid, under perspectivalism.
So perspectivalism faces a dilemma. It needs a closure principle to accommodate mixed arguments. But if it includes one, the conclusion of import arguments is (ultimately) true in virtue of something apart from the rule of recognition. If instead it incorporates a closure principle into the rule of recognition to preserve Exhaustive Determination, then the ensuing rule will not be entirely Socially Given. It follows that one can endorse or Hartian Orthodoxy, but not both.
IV. Norms and Claims
A preliminary issue with our argument pertains to its scope. We frame the target as the conjunction of legal perspectivalism and Hartian Orthodoxy. But you may worry that the same issue arises for any version of Hartian Orthodoxy, irrespective of the addition of legal perspectivalism. To see why, take our original import argument:
We assume (3a)–(3c) is valid, and that therefore (3c) is true. But if (3c) is true, then something must explain how (3c) is part of the law when (3a)–(3b) are true. This explanatory challenge arises prior to the translation of (3a)–(3c) along perspectival lines, which we earlier presented as (4a)–(4c).
A plausible answer is that (3c) follows logically, via modus ponens, from (3a)–(3b). But if (3c) is a statement of a legal norm, this poses a problem for Hartian Orthodoxy quite apart from perspectivalism. The dilemma arises regardless: either (i) modus ponens is, independently, among the fundamental explainers of legal validity, or (ii) it corresponds to a feature which any rule of recognition must include. Either way, we have seemingly arrived at an objection to Hartian Orthodoxy simpliciter.
As it turns out, a complication arises if our argument is not restricted to variants of Hartian Orthodoxy which accept legal perspectivalism. So we are only prepared to defend the more limited conclusion that the conjunction of the two is false, while leaving open the truth of Hartian Orthodoxy on its own. To see the complication, consider what a variant of Hartian Orthodoxy shorn of perspectivalism might look like. It could not hold that legal norms are claims about what exists according to law. They would instead likely be nonpropositional prescriptions, directives, commands, plans, or so on. It follows that, if (3a) is understood as simply being the legal norm, those who subscribe to this family of views must reject the validity of (3a)–(3c). After all, validity is a relation that obtains between propositions, and (3a), on the hypothesis we are considering, is not a proposition.
Still, you may wonder whether we could understand (3a) as a proposition along Hartian Orthodox—but not perspectival—lines. One natural way to do so is to construe (3a) as asserting a covariance between the property of “being over 18” and “having a legal right to vote.” That is, (3a), on this interpretation, would mean: “there are no cases where [X is over 18] and [X lacks a legal right to vote].” Say we understood (3a) in this way. Then the argument would read:
But now we no longer have reason to think there is a logical axiom making, or describing a property of the thing that is making, (3c) part of the law. This is because (3c) is not made true, such that it is part of the law, because it is entailed by (3a*)–(3b) via a logical axiom. If anything, the explanatory arrow runs the other way. The truth of (3b)–(3c) takes us to the truth of (3a*). After all, (3a*) asserts that, for any X, a certain set of conjunctions will hold. As such, it being true that “Jane is over 18” and “Jane has a legal right to vote” is logically prior to the more general assertion in (3a*). For if the conjunction that [Jane is over 18 and Jane lacks a legal right to vote] were true, then (3a*) would be false. On this view, (3a) is understood as an exceptionless generalization along the lines of (3a*), and the truth of such generalizations is logically posterior to the falsity of a certain set of conjunctions. This means no closure principle is required. So understood, the view is not vulnerable to our argument.
By contrast, on a perspectivalist understanding of Hartian Orthodoxy, it is the truth of (3a)–(3b) which explains the truth of (3c). This is why, when coupled with a closure principle, (3c) is necessarily true when (3a)–(3b) are true. The principle is therefore a crucial part of the story for (3c) being part of the law. This renders the view vulnerable to our argument.
So far, we assumed a nonperspectivalist Hartian can interpret (3a) as (3a*). Say you disagree. Perhaps (3a) should instead be understood as a proposition about a legal norm, the content of which is “all and only persons over 18 have a legal right to vote.” Then the argument would proceed as follows:
This argument, as phrased, is invalid. To make it valid, a bridge principle of the following form is required:
If there is a legal norm that “N only if P,” then “N” only if P.
Here, N refers to a normative legal entity, such as a legal right or duty. The inclusion of this principle secures the validity of (3a†)–(3c). Further, the truth of (3a†) and (3b) explains the truth of (3c). At the same time, it seems to render the nonperspectival Hartian vulnerable to our objection. For now, the truth of the bridge principle appears to be an essential part of the explanation for why Jane has a legal right to vote. After all, without it, (3a†) and (3b) could not render true that claim that Jane has a legal right to vote. The bridge is either applicable in virtue of the rule of recognition or it is not. The former seems inconsistent with Socially Given, the latter with Exhaustive Determination. Is this version of nonperspectival Hartian Orthodoxy, therefore, vulnerable to our argument?
A preliminary point is that, while we think the reinterpretation of (3a) as (3a†) could work, those who instead opt for (3a*) can straightforwardly explain the validity of the original argument without a bridge premise. By contrast, perspectivalists must resort to a closure principle to accommodate its validity.Footnote 16
More substantively, nonperspectivalist Hartians who reinterpret (3a) as (3a†) can endorse a conception of legal norms which avoids the dilemma we set out. Specifically, it is possible to understand the bridge premise such that its truth is part of the nature of legal norms. As such, the explanation for why the bridge premise is true need not rely on a feature the rule of recognition must possess, nor on something beyond the rule of recognition. All that is necessary is to understand something about the nature of legal norms themselves, which seems less problematic for Hartian Orthodoxy.
What is this account of legal norms? Take moral norms first. Clearly, we want to say the moral norm [something is bad if it causes unwarranted suffering] explains, in part, why it is bad to torture a baby for fun. David Enoch offers a possible way to do this.Footnote 17 He says the norm makes it the case that the property of “suffering” (when paired with, say, the property “unwarranted”) grounds the normative property “bad.” The moral norm makes it the case that a certain set of descriptive properties grounds a certain normative property. That is, the moral norm metaphysically gives descriptive properties the power to ground moral properties. On this view, we do not need any inference rules to bridge the gap between there being a moral norm that “X is bad if X bears property-P” and a given token of X actually being bad if that token bears P. The latter follows from the former simply from the nature of a moral norm.
As Enoch suggests, we could apply the same view to legal norms.Footnote 18 The legal norm [it is prohibited to drive, on road R, over 80 km/h] makes it the case that a token of the act type “driving” bears the property “being prohibited” in virtue of bearing the properties “being on road R” and “being over 80 km/h.” If so, we do not need a bridge principle to conclude that Jill has violated the law if she drove 90 km/h on road R. Given the general legal norm, the property of being legally prohibited obtains in virtue of the descriptive properties.
The nonperspectivalist Hartian can therefore explain the bridge principle if we assume that, in this respect, the nature of legal norms is analogous to moral norms. Once we understand legal norms as conferring powers on descriptive properties to ground legal properties, there can exist no gap between there being a legal norm “q if p” and the truth of the proposition “q if p.” The existence of the norm entails the existence of the exceptionless generalization, from which we can then straightforwardly validate the argument. No separate inference rule or bridge principle is required.
Crucially, on this view, there is no need to invoke anything outside the rule of recognition as a fundamental determinant of legal validity. The proximate explanation of why Jane being over 18 makes it the case that Jane has a legal right to vote is the legal norm giving Jane being 18 this grounding power. This is consistent with a straightforward Hartian explanation of why the legal norm exists, namely, the legislature having enacted a statute. That this legislature is a source of law, in turn, will be given by the rule of recognition.
At first glance, the view we just sketched strikes us as a viable way for nonperspectival variants of Hartian Orthodoxy to go. If so, they are not vulnerable to our argument. This, in addition to the alternative construal of (3a) as (3a*), explains why we restrict the target of our objection to the conjunction of legal perspectivalism and Hartian Orthodoxy.
V. Restricting the prefix
Our argument, once formalized, is as follows:
-
(I) If legal perspectivalism is true, then a closure principle must be true.
-
(II) If a closure principle must be true, then either Exhaustive Determination or Socially Given must be false.
-
(III) If either Exhaustive Determination or Socially Given is false, then Hartian Orthodoxy is false.
-
(IV) So if perspectivalism is true, then Hartian Orthodoxy is false.
The previous section explains why (I) is framed in terms of legal perspectivalism. In this section, we address a preliminary worry with (I). Then in the next section, we respond to some other objections to our argument.
Grant for a moment that the closure principle we propose—Conditional Legal Closure—is in serious tension with Hartian Orthodoxy. Still, you may wonder whether perspectivalism really needs it to accommodate the desideratum. Here we address an initially attractive way to do without the principles: to limit the coverage of the prefix “according to law” such that it doesn’t take scope over the whole conditional.
How? We can understand “all and only persons over 18 have a legal right to vote” as asserting:
X has a legal right to vote if X is over 18.
The core claim of perspectivalism is that “legal” serves to indicate that the expressed proposition features a prefix. Previously, we assumed the prefix “according to law” has a wide scope. That is, it takes scope over the entirety of the statement in which it is embedded. The perspectivalist reading would then go:
According to law X has a right to vote if X is over 18.
But, even if this were the right understanding of the prefix for simple statements of legal rights, you may reject it for more complex statements like the one we gave above. You could, instead, think the prefix has a narrow scope. That is, it takes scope only over the consequent of the conditional, rather than the whole conditional. Thus:
If X is over 18, then [according to law X has a right to vote].
Understood this way, the mixed arguments are valid without a closure principle.Footnote 19
It now looks like this strategy can accommodate the desideratum without any apparent tension with Hartian Orthodoxy. (Later we will show some nonapparent ways in which the two are in tension.) Things are not so easy, however. This approach must confront a serious hurdle: explaining how the whole conditional—that is, “all and only persons over 18 have a right to vote”—is part of the law. As Perry explains,
The law is not that persons have a right to vote; it is that certain persons have a right to vote, namely, those over 18. What is true from the legal perspective is at least the content of general legal norms—that is, the law.Footnote 20
In other words, under perspectivalism, the content of the law includes p only when p is true according to law. So if it is part of the law that “all and only persons over 18 have a right to vote,” the prefix must apply to the whole conditional. Limiting the scope of the prefix, therefore, robs Peter to pay Paul. It accommodates the desideratum, but at the cost of distorting legal content.
You may think this response is too quick. After all, is it not open to perspectivalism to deny this notion of legal parthood? The thought is that certain propositions are part of the law even if they are not true according to law. How so? The perspectivalist could adopt the following disjunctive view:
p is part of the law if and only if either (i) p is a normative claim which is true in law, or (ii) p determines whether normative claims are true in law.
If so, “all and only persons over 18 have a legal right to vote” would be part of the law even if the prefix only takes scope over the consequent. For “if and only if X is over 18, then [according to law A has a right to vote]” determines whether it is true in law that “X has a right to vote” (a normative statement).
Nonetheless, this response fails for two reasons. The first is that the prefix can surely take scope over the whole conditional on at least some occasions. But then the challenge of accounting for mixed arguments reemerges. The second is that this response is entirely ad hoc. For the perspectivalist, it may be helpful to think of “the legal perspective” by analogy to the world of a fictional story. Things are thus and so according to the novel, or “in the world of the novel,” even if they aren’t that way in the real world. Now ask yourself: what is “part of” that novel (or the world of the novel)? Intuitively, only those things that are true according to the novel. The novel does not include whatever rules determine whether some claim is true according to the novel. Consider by way of analogy:
If J.R.R. Tolkien wrote p in the official version of The Lord of the Rings trilogy, then according to The Lord of the Rings p.
Clearly, the conditional is about The Lord of the Rings. But it doesn’t seem part of Tolkien’s fiction. Why wouldn’t the same be true for law? Limiting the scope of the conditional redefines when something is part of a perspective in ways unique to law. The second response is that the prefix can surely, on at least some occasions, take scope over the whole conditional. But then the challenge of accounting for mixed arguments reemerges.
You could simply insist these arguments, as formulated, are invalid. But there are a couple of problems with this. For one, it means “all and only persons over 18 have a legal right to vote” is ambiguous between: (i) “according to law all and only persons over 18 have a moral right to vote” and (ii) “all and only persons over 18 [according to law have a moral right to vote].” But the argument is only valid under (ii). So the validity of the argument is itself ambiguous. This is odd, since the initial argument, prior to being translated along perspectival terms, seems unambiguously valid.
For another—and this is the more serious issue—we must confront the relationship between (i) and (ii). Can they diverge in their truth conditions? If they can, a problem arises. Consider again:
As a matter of legal practice, we cannot infer (4c) from (4b) if (4a) were false. If it is not the law that all those over 18 have a right to vote, then lawyers cannot reason from someone being over 18 to conclude they have a right to vote. Yet if (i) and (ii) can diverge in their truth conditions, this inference would be possible. For it could be true that:
Now we can infer (4c) given the truth of (4a*)–(4b), even if (4a) is false. This cannot be right. So if the truth value of (4a) and (4a*) can diverge, this strategy renders the wrong verdicts about which inferences we can make.
One response is to try and ensure some connection between (4a) and (4a*) holds such that their truth values cannot diverge. But how? There would need to be some link between when (4a) is true and when (4a*) is also true.
We think this is possible. But the best way of forging this connection leads us to Conditional Legal Closure, and as such reintroduces the tension between perspectivalism and Hartian Orthodoxy.
The simplest possible link would be:
But this cannot account for legal fictions. Such fictions arise when the law stipulates the truth of some descriptive proposition even when it is false. It can therefore be true that “according to law p,” even though p is false (where p is some descriptive proposition). Here’s one example. Take the conviction of a defendant who is actually innocent. Upon conviction, it becomes the case that “according to law D φed,” where φ picks out the conduct which forms the basis of the criminal offence. And this can be true according to law even if D did not actually φ.Footnote 21
As such, where there is a legal fiction, “according to law q” can be true despite “according to law q if p” being true and p being false. For there could be a legal fiction that p is true, such that p is true according to law despite being false. The naïve link we outlined as (i) cannot make sense of this. The lesson: we cannot guarantee correspondence between (4a) and (4a*).Footnote 22
Those who wish to defend perspectivalism without the help of closure principles could try and make do with a weaker link between (4a) and (4a*). That weaker connection is:
This is an improvement since it does not preclude legal fictions. But note that under (ii) we cannot infer (4c) unless (4b) is taken as prefixed. We have seen this before. Recall:
Only now is the argument valid given (ii). As it happens, this is because (ii) is a derivation of Conditional Legal Closure.Footnote 23 Since (ii) is a notional variant of our preferred closure principle, this attempt to restrict the scope of the prefix cannot undermine our argument.
To sum up, we should reject this strategy because it involves an ad hoc notion of legal parthood. And, even if you were willing to ignore this, it does not support a genuine alternative to our closure principle. But suppose we set all this aside. Our argument goes through regardless. Even with the scope of its prefix restricted, perspectivalism is still incompatible with Hartian Orthodoxy.
An initial incompatibility arises because this strategy is committed to the rule of recognition being part of the law. This is because the rule determines whether normative claims are true in law. Under (ii), then, it is itself a legal rule. This involves some cost since it renders perspectivalism incompatible with Exhaustive Determination. If the rule of recognition is part of the law, then under Exhaustive Determination it must have validated itself. This, however, seems false. The rule of recognition need not meet the same criteria of validity it uses to determine the validity of other legal norms. This notion of legal parthood therefore makes the rule, under perspectivalism, a counterexample to Exhaustive Determination.Footnote 24
Nonetheless, we won’t press this point further since there is a response that we do not find too costly. That is, carve out an exception to Exhaustive Determination for the rule of recognition. For Hartians, the rule of recognition is plausibly special in many respects. One of those ways could be its being an exception to Exhaustive Determination.
There is, however, a more serious tension between perspectivalism (once its prefix is restricted) and Hartian Orthodoxy. The problem is (4a*), which says if A is over 18 then according to law, A has a moral right to vote. The rule determines whether (4c) obtains, and thus whether Jane has a legal right to vote. This is contrary to Exhaustive Determination. For there is now a rule, apart from the rule of recognition, which is an ultimate determinant of legal validity.
There are two ways to avoid this conclusion. First, (4a*) could be part of the rule of recognition. Second, it could be the case that (4a*) is in some way traceable to the rule of recognition, such that the rule of recognition still ultimately exhaustively determines the content of the law.
Both options, however, violate Socially Given. If (4a*) is part of the rule of recognition, then one of its components would be independent of official practice. This problematic component would hold that NC (the consequent of a general legal norm) is part of the law if AN (the antecedent of a general legal norm).
If, instead, this rule (that NC is part of the law if and only if AN) is validated by the rule of recognition rather than being among its components, then there must be some criterion which corresponds to the validation. That criterion would then necessarily be part of the rule. Either way, Socially Given is violated.
VI. False exits
Just to restate, here’s the formalization of our argument.
-
(I) If legal perspectivalism is true, then a closure principle must be true.
-
(II) If a closure principle must be true, then either Exhaustive Determination or Socially Given must be false.
-
(III) If either Exhaustive Determination or Socially Given is false, then Hartian Orthodoxy is false.
-
(IV) So if perspectivalism is true, then Hartian Orthodoxy is false.
The previous section rejected an objection which sought to undermine (I). Here we consider an additional objection to (I), as well as objections to (II) and (III).
A. Rejecting (I)
Under (I), for perspectivalism to be true a closure principle must hold. This expresses a necessity relation. For any jurisdiction, at any time, a closure principle cannot help but be true. Why? Due to the desideratum concerning the validity of certain mixed arguments. To account for this desideratum, perspectivalism must have the tools—that is, a closure principle—to address this need.
But you could object. Perhaps you think the desideratum, put this way, is too strong. There is a difference between arguments such as (3a)–(3c) being valid once suitably relativized to a particular jurisdiction, and such arguments being valid respecting every possible jurisdiction. For some jurisdiction, such as the United Kingdom, a version of (3a)–(3c) relativized to it may be valid. At the same time, there could be a legal system—or so the objection goes—where (3a)–(3b) are true yet (3c) is not. That is, it might be possible to have a legal system whose perspective is not closed in accordance with Conditional Legal Closure. What separates a system with a closure principle from one that is not? Simple: the rule of recognition in the former, but not the latter, either incorporates or is accurately described by the closure principle.
Notably, this objection does not reject the desideratum wholesale. After all, the arguments (suitably relativized) are valid if there is a closure principle. And in jurisdictions like ours, where we can safely presume a closure principle exists, arguments like (3a)–(3c) are valid. The objection merely insists that there are some possible legal systems which are not like ours in this respect. If so, the presence of a closure principle is consistent with both Exhaustive Determination and Socially Given. As such, the objection raises an important worry with our argument. But we do not think the objection succeeds, for three reasons.
First, the objection does not preserve validity respecting all suitable relativizations of (3a)–(3c). But we should not accept that there is some jurisdiction when, at some point in time, a conclusion which would follow from CLC is not true.
To see why, consider other domains in which people have proposed to understand the relevant discourse as including a fictional or perspectival operator. For instance:
This argument seems undeniably valid. Now suppose we were to translate the argument along perspectival lines.
Now it, like (2a)–(2c*), is invalid absent a closure principle.Footnote 25 For (6a)–(6c) to be valid, F would need to be closed by a suitable closure principle. Crucially, this principle must apply to any fiction. Otherwise, we would need to accept the existence of arguments involving fictional terms such as (5a)–(5c), which, prior to their prefixed translation, seem undeniably valid, yet are no longer valid posttranslation. This would be a serious black mark against prefixed accounts of the meaning of fictional discourse.Footnote 26
A possible worry is that we are being insufficiently imaginative regarding fictions. A particular fiction, perhaps, could deny modus ponens or adopt a paraconsistent logic. Then an argument along the lines of (5a)–(5c), which seems undeniably valid, would become invalid after the prefixed translation. Why can’t the same be true for law?
We think it can. That law (or some fiction) need not be closed in accordance with classical logic does not, however, mean it need not be closed. The legal perspective could instead be closed in accordance with whatever logic it itself endorses. For instance, to use a nonlegal example, consider the worldview of Graham Priest.Footnote 27 He thinks there can be true contradictions. If we closed his worldview using classical logic, this would mean everything is true according to his worldview. But Priest is not committed to trivialism. If his worldview is wrong, then it is wrong for some other reason. We would avoid this issue by closing his worldview with the paraconsistent logic that he himself endorses.
Now return to the legal perspective. If it were closed in accordance with the axioms of classical logic, then it would adhere to:
Entailment Legal Closure (ELC). If p is entailed by q, r, s …, and according to law each of q, r, s …, then according to law p.
The objection—namely, that the law may not follow classical logic—would have force if we were committed to ELC. For, on ELC, the only way to avoid a conclusion stemming from an otherwise valid argument (that is, an argument that is valid prior to any prefixed translation) would be to suspend or do away with a closure principle. But we do not rely on ELC. Rather, we used the following in §2:
Conditional Legal Closure (CLC). If according to law [q therefore p], and according to law q, then according to law p.
CLC, crucially, differs from ELC precisely because it requires the legal perspective to affirm the inference rules (including the logical axioms) used to reach the conclusion of an argument from its premises. It follows that CLC could be necessarily true of legal systems, while it is also true that a particular legal system—given the social practice of its officials—carves out some exception from modus ponens or some other axiom of classical logic.
The possibility of perspectives adhering to a nonclassical logic does not, therefore, undermine the need for a closure principle. Still, you might think the law isn’t closed in line with CLC. To see why, return to the following import argument:
Now say a legal official denies Jane the right to vote. This brings up a possible worry with (I). While (4a) and (4b*) are still true after the official’s decision, (4c) is not. We therefore have an apparent failure to adhere to the closure principle. If correct, this would show that, contrary to (I), the applicability of a closure principle is socially contingent.
But this mischaracterizes the situation. To start with, the decision is most naturally understood as a legal mistake. That is, the official has now changed the law via a mistaken decision. Why, though, did the refusal to apply the norm to Jane change the law? Our view offers an answer. It was, prior to the decision, true that Jane had a legal right to vote. The official mistakenly failed to recognize. Of course, the official’s failure could then change the law. But this is consistent with its being a mistake. If one wishes to avail themselves of this explanation of the denial of Jane’s legal right to vote, however, then they need to explain how it was true that “according to law Jane is entitled to vote” prior to the mistake. This, we argue, requires a closure principle be applicable to the jurisdiction.
Now turn to the sense in which, after the decision, Jane lacks a legal right to vote. How is this compatible with the legal perspective adhering to a closure principle? The answer is that the official’s conduct is best understood as making one of the premises in the argument false. Here are two possibilities. First, the decision could be understood as having amended the general norm, such that there is an exception with respect of Jane when it comes to (4a). Second, it could have introduced a legal fiction with respect of Jane’s age such that (4b*) is no longer true.
These explanations are consistent with both ELC and CLC. But suppose you reject them. That is, you think it possible for (4c) to be false even when (4a)–(4b*) are true. We accept this is inconsistent with ELC. But, as we will now show, there are two additional ways, consistent only with CLC, to understand what is going on.
First, the official’s decision could have carved out an exception to modus ponens from the legal perspective. In short, the decision makes it the case that, going forward, valid inferences from these particular premises will depart from classical logic. Second, it could be the case after the decision that Jane both has and lacks a legal entitlement to vote. In other words, the decision has made a contradiction true according to law. This is unacceptable under ELC because, given the principle of explosion, all propositions would now be true according to law. But if this jurisdiction rejects the principle of explosion—as CLC allows—then no such problem arises.
As such, neither nonclassical logics nor official mistakes are a threat to CLC. Now let’s turn to a more fundamental point. There are good reasons to believe that a closure principle must be true of a legal system. For it is hard to see how the truth of a closure principle could be socially given. If it were socially given, then we should be able to conceive of a legal system that systematically does not adhere to a closure principle. But we do not think such a system is conceivable.
To be clear, we accept that the soundness of import arguments is often, if not always, contingent on the practice of legal officials. This is because whether its premises are true—such as whether the law claims Jane is over 18, or that all and only those over 18 have a moral right to vote—is socially contingent. Additionally, whether the relevant inference rule—like modus ponens—is true according to law may also be socially contingent. What is not contingent, however, is the need for a legal perspective to endorse an inference rule according to which a conclusion follows from premises it also endorses.
To see why, start with the edge case. Can a legal perspective never endorse a conclusion because it follows from premises it does endorse? No. Such a perspective would be radically open in an inconceivable manner. Whenever an official must decide on the existence of a legal right in a new case, there would simply be no relevant legal norms which predate the dispute. Absent a closure principle, legal norms could only arise via explicit contribution, that is, directly through some exercise of a legal power.Footnote 28 The class of cases in which judges apply existing law, rather than create new norms, would then be vanishingly small. The objection therefore, relies on the law being open in a way that few have sought to defend.Footnote 29
Say you accept it is impossible for a system to never endorse the consequences of what it does endorse. Still, you could wonder whether it is tenable for a legal system to only sometimes do so. Could a given system’s rule of recognition block the operation of a closure principle on certain occasions? Even if we were to concede that this is possible, the suggestion no longer poses a threat to (I). What (I) needs is for a legal system to necessarily contain a closure principle, the content of which is not contingent on social practice. It does not require that the closure principle apply to all parts of the law. It is enough that any legal system must contain a closure principle which is (i) not socially contingent, and (ii) at least on some occasions, fundamentally determines the law. And, insofar as it is true of any legal system that there will be some norms which owe (in part) their validity ultimately to the closure principle (even if these norms are only part of a restricted domain), then (i) and (ii) will hold.
B. Rejecting (II)
Now consider two objections to (II). Recall:
The first objection seeks to show that a closure principle somehow depends on the rule of recognition without being part of the rule itself. The second objection seeks to show that, even if a closure principle is needed to make sense of legal claims, the principle need not determine the content of the law.
The first objection takes us to familiar territory. It consists of two steps. To start, deny that the closure principle is part of the rule of recognition. Then try to trace the principle to the rule of recognition. This came up previously when trying to restrict the scope of the prefix. We shall show this objection to (II) fails for the same reasons as earlier.
Consider how the objection might get off the ground. Our formulation of Hartian Orthodoxy invokes the notion of ultimacy. Under Exhaustive Determination, some norm N is part of the law in virtue of ultimately being validated by the rule of recognition. This is consistent with many legal norms not being directly validated by the rule of recognition.
Suppose the United Kingdom’s rule of recognition is that whatever the Westminster Parliament enacts is law. Many UK laws, however, are not enacted by Parliament. Take Scotland, for instance. Clearly, the enactments of the devolved Scottish Parliament can still be valid. This is possible because, under the rule of recognition, the Westminster Parliament is the ultimate source of law. But then it could enact a power-conferring law which grants the Scottish Parliament the power to enact certain laws for Scotland. On this picture, the devolved laws of Scotland are ultimately traceable to the UK rule of recognition, but they are not the direct products of the Westminster Parliament.
Hartian Orthodoxy can explain the legal validity of Scottish law. Now comes the objection: what stops us from explaining the operation of closure principles along similar lines? Perhaps the necessary closure principle is the proximate or direct source of the validity of certain mixed arguments. Still, it may be tempting to think we can deny (II) so far as the closure principle is itself validated by the rule of recognition.
We should reject this thought. And for the same reasons as before. Either the rule of recognition necessarily or contingently validates the closure principle. If necessarily, then Socially Given is false. The rule of recognition can only necessarily recognize the closure principle if (i) there is some criteria picked out by the rule which suffice to make it part of the law, and (ii) that criteria is present in any rule of recognition the legal practice could have, no matter the social practice of legal officials. This is incompatible with Socially Given. If, however, the validation is merely contingent, then we are back to our arguments for the necessity of a closure principle.
There is, however, a nearby objection which avoids the need to trace. Exhaustive Determination simply requires the rule of recognition to be the only ultimate determinate of whether some norm is part of the law. Up to now, we argued the closure principle “makes” the conclusion true and is therefore another determinate. Suppose you deny this. You could, instead, think the closure principle simply captures—that is, describes—what the rule of recognition already makes the case.
This view has serious motivation. Take the logical inference rules. On one view, the laws of logic do not make a conclusion true when it follows from the premises.Footnote 30 Instead, they simply describe what follows from what. If ABC is true of the world, then the logic rules tell us that XYZ will also be true of the world.
Can the same be true of closure principles in law? Then they could accurately describe the rule of recognition’s activity without threatening Exhaustive Determination. We think so. But this runs afoul of Socially Given for reasons we already addressed. Either the rule of recognition’s activity only contingently covaries with what the closure principle licenses, or it necessarily does so. If the former, we are right back to rejecting (I). If the latter, then presumably this is because the rule of recognition picks out some criteria which obtain in a manner necessarily coextensive with the closure principle. For that to be true, these criteria must be necessarily part of the rule. Hence, this route requires the rejection of Socially Given.
Or does it? Our current hypothesis is:
If the truth of a closure principle depends on the social practice of legal officials, then we cannot account for the necessary validity of mixed arguments.
You could deny this. That is, you could insist the closure principle entirely depends on official practice yet also insist the mixed arguments it validates are necessarily valid. This supposes something can be both dependent and necessary. How? Since, although Z depends on Y, Z is necessary because Y is necessary. Suppose, for instance, the natural numbers necessarily exist. The set of all natural numbers depends on the existence of those numbers. But if the natural numbers necessarily exist, then (plausibly) their set also necessarily exists.
The essence of Socially Given is dependency, not contingency. Legal practice is what determines the content of the rule of recognition. The rule only has the content it does because it is what officials accept. Now, if what officials accept is contingent, then the rule’s content would also be contingent. Generally, those who defend Hartian Orthodoxy will, we assume, wish to endorse this contingency.
But perhaps closure principles, or the criteria which covaries with those principles, are an exception. It is not contingent that officials accept a closure principle. The thought might be that, if officials were to reject the principle, their practice would no longer be legal in character. Still, on this view, the rule of recognition possesses whatever content it does in virtue of social practice. As such, the objection denies (II). It says a closure principle can necessarily hold in a manner consistent with Socially Given. We should reject this, however, for a couple reasons.
For one, to be plausible, the objection must explain why officials necessarily accept a closure principle, but not other criteria specified by the rule of recognition. Absent an explanation, the maneuver looks ad hoc.
For another, if legal officials must necessarily accept a closure principle, then this would render the legal perspective fundamentally distinct from other perspectives. The closure principle, which applies to fictions, worldviews, or moral theories, does not depend on our practices in this way. It is not merely that some mixed arguments respecting these perspectives are necessarily valid. Rather, these closure principles obtain given the kind of thing these perspectives are.Footnote 31 So, perspectivalists who defend Hartian Orthodoxy would be forced to exempt the legal perspective from what is true of other perspectives.
Now turn to the second objection to (II). Recall how the import argument upon which we rely, once translated along perspectival lines, proceeds:
Our argument builds on the fact that (4c) is true, where it is true, partly in virtue of a closure principle which renders certain propositions true according to law due to how they relate to other propositions also true according to law. Since (4c) is a legal norm, if it is made true by a closure principle not traceable to a rule of recognition, then Exhaustive Determination is false. If it does trace to the rule of recognition, then Socially Given is false. Either way, legal perspectivalism is inconsistent with Hartian Orthodoxy.
But what if we reject (4c) being a legal norm? Suppose instead (4c) is understood as a legal claim (or proposition) instead, such that it does not imply the existence of a norm.Footnote 32 If this works, then the truth of (4c) would no longer pose a threat to Exhaustive Determination; it would not require the existence of a legal norm—that is, a right to vote which Jane possesses—whose inclusion as part of the law must be accounted for. The only norm is that “all and only persons over 18 have a right to vote.” That norm, in turn, might make true the proposition that “according to law Jane has a right to vote.” But (4c) being true need not imply the existence of any new norms in law. On this view, a closure principle may describe which legal claims follow from other legal and nonlegal claims. It says nothing, however, about the content of the law—that is, legal norms—which are the purview of the rule of recognition. Put another way, the objection is that a closure principle and the rule of recognition do two strictly separate things. That the closure principle works in its domain (claims) does nothing to threaten either the ultimacy of the rule of recognition in its domain (norms) or the social contingency of its criteria of validity.
We do not think this objection succeeds. Before we explain why, let’s address a preliminary issue. Some who endorse legal perspectivalism do, it seems, conceive of legal norms as prefixed propositions.Footnote 33 On this view, the legal norm “all and only persons have a legal right to vote” simply is the proposition “according to law, all and only persons have a moral right to vote.” The norm is valid when this proposition is true. If so, there is no sharp distinction between legal claims and legal norms. It follows that our argument straightforwardly applies to this variant of perspectivalism.
Additionally, such a variant would otherwise come with certain advantages. Ruling it out is therefore costly to some extent. One advantage is that it allows us to treat the term “legal” as doing the same thing in both talk of legal norms and talk of legal rights. In both cases, it functions to indicate that the statement involves a wide scope prefix. To say “there is a legal norm that N” is equivalent to “according to law, there is a moral norm that N.” And to say “S has a legal duty to φ” is equivalent to “according to law, S has a moral duty to φ.” All else equal, it would be better for our legal semantics to be consistent with respect to how the term “legal” functions. If perspectivalist theories were to conceive of “legal norms” as something other than prefixed propositions, then this consistency is lost.
Another advantage relates to a particular concern with positivism, namely that it proliferates normativity or otherwise adds a set of novel normative entities to our ontology.Footnote 34 If we accept that the law consists of legal norms, and that those norms are entities (just not robustly normative ones), then it appears we have done just that. By contrast, if legal norms just are prefixed propositions about what is true according to law, then this metaphysical worry does not arise.Footnote 35
Setting this aside, there is a more fundamental problem with this objection. The link between the truth of legal propositions and the existence of legal norms is, in some respects, stronger than the objection supposes. The objection does not merely require a distinction between legal claims and legal norms. It also requires that, when a legal claim such as (4c) is true, it is not true in virtue of there being some unique legal norm (such as Jane’s legal right to vote). Perhaps this is plausible enough respecting a claim like (4c), which we might describe as a “particular” or “specific” legal claim. Crucially, however, we can construct import arguments whose conclusion is not a particular legal claim. For them, the assumption that no new legal norm is at play is far less plausible. Consider:
Of the premises, (7a) reflects a common view of the UK’s constitutional arrangements and (7b) is a descriptive claim about whether, according to law, Parliament has said the content of N.Footnote 36 Now it seems obvious (7c) should follow. But once formalized, it is apparent that (7a)–(7c) is an import argument. Moreover, the structure of the argument generalizes. Whenever body-B has the power to make law by declaration, there will be a conditional:
According to law if B says N, then N.
So long as N can include another power-conferring rule, it is possible for B to say:
If C says N, then N.
Then, given an appropriate closure principle, it follows:
According to law if C orders N, then N.
Crucially, irrespective of whether (7c) is a legal norm, it is plainly not a particular legal claim. So we are inclined to say (7c), and legal claims like it, are true in virtue of there being a new legal norm, namely the power the Scottish legislature now possesses.
It follows that the objection, to the extent it seeks to reconcile Exhaustive Determination with the presence of a closure principle, fails. For (7c) is true partly in virtue of a closure principle, and the truth of (7c) implies (or requires, or entails) the existence of a corresponding legal norm. As such, there is a norm whose membership in the law needs to be explained, and a closure principle must feature in that explanation.
Could those inclined to the objection deny the existence of a new legal norm corresponding to (7c)? The thought is that (7c) is simply the best interpretation of what the norm in (7a) requires. On this hypothesis, given the facts, “adhering” to (7a)—or, alternatively, the rule of recognition licensing (7a)—simply involves (7c). If so, there would be no new norm in (8c) for which we must separately account.Footnote 37
We do not think this works. A minor point is that it is unclear what “adhering” to (7a) might amount to. There are no duties in (7a), which is instead a power-conferring rule. More seriously, if this maneuver is to work, then any proposition we must infer from another legal proposition will not be true in virtue of a new norm. After all, if we must infer a certain legal proposition, then—given perspectivalism—we will need a closure principle. And if that proposition being true implies there is a new legal norm on the scene, we will need to account for that new norm. This would then generate another candidate counterexample to Exhaustive Determination.
As such, the objection relies on the rule of recognition being the only norm (or, if there are multiple rules of recognition, norms) in a given legal system. The other “norms” would need to be replaced with true propositions representing what “would” be true—normatively speaking—if the rule of recognition were robustly normative. But this is unacceptable. There are more norms than just the rule of recognition—or, indeed, more norms than just the rule of recognition plus those norms the rule directly validates. We should therefore think that there is a new norm which makes (7c) true. And that is all we need for our argument to work.
C. Rejecting (III)
So much for objections to (II). What about (III)? Again:
The objection would need to argue that Exhaustive Determination or Socially Given are not essential to Hartian Orthodoxy. Or, put differently, that they could be abandoned while still maintaining the core spirit of Hartian Orthodoxy. Here, we take for granted that this is impossible for Socially Given. This still leaves open Exhaustive Determination. But as we now show, it too must be kept.
A possible approach is to restrict Exhaustive Determination to general norms. So:
Exhaustive Determination (General). Whether some general norm NG is part of the law ultimately depends on NG being validated by the rule of recognition.
If so, Hartian Orthodoxy only requires norms like “all and only persons have a legal right to vote” to be validated by the rule of recognition. But it takes no position as to how, if at all, specific norms—like “Jane has a legal right to vote”—become part of the law.
This objection fails, however, because we can construct import arguments which have a general legal norm as their conclusion. Recall:
As such, perspectivalism is also incompatible with Exhaustive Determination (General). The only way out is to merge the closure principle into the rule of recognition, but that would render perspectivalism inconsistent with Socially Given (since an aspect of the rule of recognition would exist necessarily).
Exhaustive Determination (General) is not, however, the only possible refinement. Consider:
Exhaustive Determination (Noninferential). Whether some norm N is noninferentially part of the law ultimately depends on whether N is picked out by the rule of recognition.
This position distinguishes two kinds of legal claims: (i) the inferential, and (ii) the noninferential. Inferential claims are those which depend for their truth on being properly inferrable from other claims. Conversely, the truth of noninferential claims is not so dependent. They are true without the need to be inferable from other claims.
The divide between (i) and (ii) typically corresponds to the divide between general and specific norms. For general norms are often noninferential. But not always, as we have just shown. Equally, specific legal norms are not necessarily inferential. For instance, a judge could authoritatively declare that Jane has a legal right to vote. By doing so, the judge can, perhaps, noninferentially make it the case that, according to law, Jane has a moral right to vote.
This brings us to the final objection to (III). On Exhaustive Determination (Noninferential), the rule of recognition need only to exclusively determine the ultimate validity of noninferential claims (and their associated norms). Conversely, closure principles only apply to inferential claims (and their associated norms). So the domains of rules of recognition and closure principles never overlap.Footnote 38 As such, closure principles are unable to undermine Exhaustive Determination. This preserves the ultimate validating role of the rule of recognition within its (noninferential) domain.
Let’s take stock. What we now have is a refinement of Hartian Orthodoxy solely to address our incompatibility claim. Absent some independent motivation for this maneuver, this seems suspect.
Worse, it leads to the worrying conclusion that Hartian Orthodoxy does not apply to a substantial part of the law. Suitably revised, it holds:
Hartian Orthodoxy*. Some noninferential norm N is part of the law only if, and in virtue of, N being ultimately recognized or validated by the rule of recognition.
Earlier, we showed that inferential claims include not just many specific legal norms (like “Jane has a legal right to vote”) but also some general legal norms (like “if the Scottish Parliament says N, then N**). That Hartian Orthodoxy does not seek to explain them would be startling. It would restrict the position to a vanishingly small aspect of the legal domain. What was supposed to be the key to understanding legal validity would only apply to norms whose legal membership does not rely on the inferential relation.
This is not the only problem with Hartian Orthodoxy*. Strictly speaking, it is consistent with inferential norms being necessarily determined, in part, by morality. How? Although our closure principles do not involve any moralized elements, other possible closure principles do. For instance, one could endorse a closure principle whereby P only follows if P is minimally morally correct. That is consistent with Hartian Orthodoxy* to the extent it only applies to the noninferential part of the legal domain.
This would radically weaken legal positivism. Suppose you took the view, which some attribute to Ronald Dworkin, that there are two steps to a complete picture of legal content.Footnote 39 First, some aspect of legal content consists entirely of social facts. Second, another aspect of legal content consists of principles inferred from the legal rules at the first step (as a matter of fit) in light of background moral principles which render the law more morally attractive (through a process of justification). This picture, whether defensible or not, is unrecognizable as legal positivism. For Hartian Orthodoxy to somehow be rendered consistent with this two-step moralized account of law would amount, in effect, to its destruction. The lesson we should draw is that Hartian Orthodoxy applies to more than just noninferential norms.
But if so, we can see no tenable alternative to Exhaustive Determination, which allows Hartian Orthodoxy to be consistent with perspectivalism. This addresses the objection to (III). And so, from (I)–(III), we can conclude,
VII. Conclusion
We began with the premise that perspectivalism must accommodate the validity of certain mixed arguments. Then we showed how it could use a closure principle to do so.
This took us to our central argument: that perspectivalism, so revised, is inconsistent with Hartian Orthodoxy, a widely held position within legal positivism. It holds that a norm is part of the law only if, and in virtue of, it being ultimately traceable to the criteria identified by a social rule among legal officials. If we are right, either perspectivalism or Hartian Orthodoxy must go.
It is possible for legal positivists to jettison perspectivalism, but not without cost. For perspectivalism is a powerful way to explain the sense in which law is “about” the robustly normative world. Absent it, positivists will need another way to explain, or explain away, this resonance between law and morality.
Acknowledgements
Our thanks to Tom Adams, Alma Diamond, Hasan Dindjer, James Edwards, Ezequiel Monti, Adam Perry, Andreas Vassiliou, and Dan Ward.
Open access
