Skip to main content

What Is It to Apply the Law?

A Correction to this article was published on 09 August 2021

This article has been updated

Introduction

That courts apply the law is tritely true. To say what that involves, though, is not easy. My goal in this paper is to give a jurisprudential account of the notion; to clarify what applying the law is.

Three sets of questions present themselves at once. One concerns what I will call the direct object of law-application. When a court applies the law, what exactly is it applying? This is not a trick question. Courts are said to ‘apply’ all sorts of things: statutes, rules, principles, policies, doctrines, other courts’ views, and more. Is ‘the law’ the name of another item on this list? That sounds wrong. Or is it by applying some or any of those items that courts apply the law? And if it is, can we give a principled criterion for determining what the list contains?

A second set of questions regards the indirect object of law-application. Courts do not just apply the law; they apply it to something. To what, exactly? To cases? To facts? To the ‘facts of the particular case’ before them, as the common phrase goes? We also say that a court applies the law in a certain case; and in deciding a certain case. Are all these the same? And what precisely is it to apply the law to a case? Indeed – what is a ‘case’?

Third, there are questions about the content of law-application. Is ‘applying the law’ the name of a specific type of act? If so, what are its distinctive features? If not – what then? There are also things courts are said to do – decisions they are said to issue – by applying the law. Are these themselves acts of law-application?

There is no obvious order in which to address these issues. Any attempt to deal with the questions in one set would benefit from our having first tackled those in the others. Still, we need a strategy, and mine will be to leave the first set of questions for last. I will not explicitly discuss the direct object of law-application before the very end of the paper; and until then, for convenience, I will focus on the application of statutory and other written legal provisions – indisputable instances of law-application.

I cannot, however, begin by simply turning head-on to any of the other two sets of questions. That is because there are – or so I want to suggest – two kinds of law-application, each giving rise to its own issues of content and indirect object. So I must start by explaining what the two kinds are. I call them inferential law-application and pragmatic law-application. I discuss the former in Section II, the latter in Section III, and the relation between them in Section IV. In Section V, I look into the notion of a ‘case’; and I return in Section VI to our three opening sets of questions.

Inferential law-application

A working example

Consider this excerpt from Lord Denning’s leading judgment in The Hollandia:Footnote 1

Article III, paragraph 8 [of the Hague-Visby Rules for the international carriage of goods by sea, an international Convention given ‘the force of law’ by a UK Act of 1971] . . . is of the first importance in our present case. It says that: ‘Any clause . . . lessening such liability [i.e. the liability of a carrier for loss or damage to goods] otherwise than as provided in these Rules, shall be null and void and of no effect’.

Now apply that article III, paragraph 8, to clause 2 of this bill of lading [issued by the defendants]. In so far as clause 2 restricts the liability of the carrier to D.fl. 1,250 it is clearly null and void: because it lessens the liability of the carrier to much less than his liability under article IV, paragraph 5(a) of the Hague-Visby Rules which comes to £11,490.96.

. . .

My conclusion is that, in proceedings in the courts of the United Kingdom clause 2 of this bill of lading is null and void.

It is, I take it, uncontroversial that Denning applied Article III, paragraph 8, of the Hague-Visby Rules to clause 2 of the bill of lading before him. We would say so even if Denning had not explicitly used that language. So I will take this as a working example. It gives us a clear illustration of one sense in which we speak and think of law-application.

So what does it involve?

A mental act

One thing it does not involve is the giving of an argument of any sort. Denning is, of course, offering in that passage an argument for a certain conclusion. His conclusion is that clause 2 in the bill of lading before him is null and void as a matter of law. But giving that argument is not part of his applying Article III, paragraph 8, of the Rules to the relevant clause.

Denning’s applying the provision is a purely mental act. It is the act of reasoning towards that conclusion – that the clause is null and void – on the grounds of some link that he takes to hold between the provision and the clause in question. This does not mean that Denning’s belief in that conclusion must have been the product of his mental engagement with the relevant provision. He could have justifiably formed such a belief on grounds unconnected to Article III, paragraph 8, of the Hague-Visby Rules. But it does mean that Denning must have held the further belief that the provision and the clause were related in such a way that, given that relation, his conclusion could not be false.

So when Denning goes on to publicly give, in writing, an argument for his conclusion – when he goes on to justify that conclusion to his readers – he has already applied the provision. What he writes documents his line of reasoning. It is evidence, for us, of how he thought. But he would still have applied Article III to the relevant clause (whether or not we had been able to know it) had he failed, or refused, for whatever reason, to put forward any argument at all.

Applying a provision to some object x – in this sense in which we would say of Denning that he applied Article III to the clause in the bill of lading – is something we can do purely by thinking. It is reasoning towards a certain conclusion about x.

Application and applicability

If Denning’s applying that provision was for him to reason towards the conclusion that the clause was null and void, then, I just claimed, he must have taken these two items – provision and clause – to be linked in a certain way. In what way? What is the role in that reasoning of the provision being applied?

The answer is, I think, that applying a provision p to some object x is to reason towards a certain conclusion c about x on the grounds that (a) p applies to x, and (b) if p applies to x, then c holds as a matter of law. Denning does not explicitly use the language of applicability, but it would not have been surprising if he had. His conclusion must have been premised on Article III, paragraph 8’s applying to the particular clause before him.

Is it licit for me to explain what it is for a person to apply a provision p to a certain object x by saying that p must be taken by that person to apply to x? There is no circularity. Applying a provision to x is not the same as a provision’s applying to x. To say that Denning applied Article III, paragraph 8, to the clause is to say something about what Denning did. To say that Article III, paragraph 8, applied to that clause is to say something about how the provision related to the clause: it is to say that the provision is (or was) applicable to it. But we need, of course, to be clearer about this notion.

Here is my take on the issue. First, to say that a provision p applies to an object x is to say that there is some specific consequence (or consequences) that the very existence of p has for x as a matter of law. It is, therefore, to say that there is some claim about x that is true as a matter of law, and that could not, given p, be false.

But one can know that a provision p applies to an object x and not know exactly what the consequence is that p has for x as a matter of law. One can know, that is, that there is some claim about x that must, given p, be true as a matter of law; and yet not know exactly what that claim is. And the reverse is possible as well. One can know that if p applies to x, then a certain claim c about x will be true; but not know whether p does apply to c.

To apply p to x is to work both these matters out. It is to form one’s reasoned view about a certain consequence (or consequences) that p does have for x as a matter of law.

What Denning’s applying Article III, paragraph 8, of the Hague-Visby rules to clause 2 in the bill of lading before him comes down to is his working out that that provision had the consequence of making the clause null and void as a matter of law.

Applicability claims

I said that applying a provision to a certain object x and thereby reaching a certain conclusion about x does not have to go hand in hand with actually giving an argument for that conclusion. But we can learn something about law-application by considering how such an argument, when given, is best reconstructed.

How should we think of the structure of Denning’s argument, the argument he gives in the above-quoted passage? It might be tempting to reconstruct it along the following lines:

(1) For every x, if x is a clause in a contract of carriage lessening the carrier's liability for damage to goods otherwise than as provided in the Hague-Visby Rules, then x is null and void.

(2) Clause 2 of the bill of lading issued by the defendants is a clause in a contract of carriage lessening the carrier's liability for damage to goods otherwise than as provided in the Hague-Visby Rules.

Therefore (from (1) and (2)),

(3) Clause 2 of the bill of lading issued by the defendants is null and void.

This is an instance of what is commonly called a ‘legal syllogism’: a deductively valid argument that combines, as premises, (a) a general first-order statement of law (the ‘major premise’, as it is normally – if inaccurately – called), and (b) a statement about some aspect of the case in hand that instantiates the antecedent of that general statement of law (the ‘minor premise’); and whose conclusion, taken to be true as a matter of law, is a statement about that same aspect of the particular case.

The model of the legal syllogism – widely endorsed among legal theorists – is meant to capture the justificatory structure of law-applying decisions like Denning’s. It goes together with the view that applying the law is (to quote but a recent formulation) ‘to solve a legal dispute by subsuming an individual case under a general normative premise (a legal rule, legal standard, legal principle, legal precedent, etc.)’.Footnote 2 As I have argued elsewhere, however, the model of the legal syllogism is incapable of achieving its own goals, and should be rejected.Footnote 3 On the theory behind the model, a claim like (1) is supposed to give a statement of the ‘rule’ expressed by the provision that Denning was applying. But there is no reason to attribute to Denning a commitment to the truth of anything like the claim in (1) above. Denning does not say, and has no need to say, and would in all likelihood not want to say, that all liability-restricting clauses in contracts of carriage other than as provided in the Hague-Visby rules are null and void: all of them, regardless of context or surrounding circumstances, with no room for qualifications or riders of any kind. Nor does he say that the consequence of applying that provision must always be, or only be, the nullity of some particular liability-restricting clause in some contract of carriage.

The point, note, is not that the language in (1) sticks too closely to the language of the provision itself, whereas a suitably qualified universal would almost certainly have to depart from that text. (A ‘suitably qualified universal’ is one that – if it could be given at all, which is itself unlikely – might plausibly stand as a true statement of law.) The point, rather, is that Denning does simply not engage with any universal statement of that sort, however qualified.

And the same is true of judges everywhere. When addressing whether a provision applies to the case in hand, courts do not – and would probably not even be able to – provide watertight descriptions of the properties that would have to be satisfied by any case to which the provision applies. They cannot even be reasonably expected to offer watertight descriptions of the relevant properties of the case at hand. They do not, as Denning does not, articulate and commit themselves to statements of first-order ‘rules’ that they take the relevant provision to express, and under which they then simply subsume the case in hand.

And yet it is that provision that Denning is applying. It is by reference to it that he argues for the conclusion that the clause in the bill of lading was null and void; that the provision applies to the clause is what supports his conclusion. How then should we go about reconstructing arguments like these – like Denning’s?

One could think of different alternatives.Footnote 4 Here is what I suggest:

(1′) If Article III, paragraph 8 of the Hague-Visby Rules applies to clause 2 of the bill of lading issued by the defendants, then clause 2 of the bill of lading issued by the defendants is null and void.

(2′) Article III, paragraph 8 of the Hague-Visby Rules applies to clause 2 of the bill of lading issued by the defendants.

Therefore (from (1′) and (2′)),

(3′) Clause 2 of the bill of lading issued by the defendants is null and void.

There are two points to highlight here. The first is that premise (2′) is – as is, therefore, the antecedent of the conditional in (1′) – a second-order claim about the applicability of a certain provision to a certain object x. It is what I propose to call an applicability claim. I say it is a second-order claim because – unlike the ‘major premise’ in the legal syllogism, which purports to be a first-order statement of a ‘rule’ putatively expressed by the provision that is being applied – premise (2′) is a statement about the provision itself and how it relates to a certain object.

Second, the argument in my proposed reconstruction is wholly about the relation between the provision and a particular object. It is not an argument about clauses (plural) in bills of lading (in general); nor is it about the range of effects that Article III, paragraph 8, of the Rules would possibly have on any such clause. It is an argument about a certain clause – clause 2 in the specific bill of lading that Denning had before him – and about the provision’s effect on it. Whether or not Article III, paragraph 8, of the Rules would also have the same effect, or a different effect, or no effect at all, on other foreign law clauses in other bills of lading, or on any other object, under different circumstances – that is beside the point, and was not a part of either Denning’s argument or his reasoning towards the relevant conclusion.

Nor should the justification of the applicability claim itself – the justification of premise (2′) – be taken to involve anything analogous to the legal syllogism; to involve, that is, the articulation of a general rule about the provision’s very applicability and whose antecedent is instantiated by the clause in the bill of lading. We should not, in other words, be looking to reconstruct Denning’s argument for his applicability claim along the following lines:

(i) For every x, if x is a clause in a contract of carriage and x lessens the carrier's liability for damage to goods otherwise than as provided in the Hague-Visby Rules, then Article III, paragraph 8 of the Hague-Visby Rules applies to x.

(ii) Clause 2 of the bill of lading issued by the defendants is a clause in a contract of carriage and lessens the carrier's liability for damage to goods otherwise than as provided in the Hague-Visby Rules.

Therefore (from (i) and (ii)),

(2′) Article III, paragraph 8 of the Hague-Visby Rules applies to clause 2 of the bill of lading issued by the defendants.

How then are the relevant premises – both the applicability claim, and premise (1′), which specifies the consequence the provision is taken to have for a certain object if it does apply to it – justified in the context of arguments like Denning’s? On the basis of reasons of different kinds, and of how such reasons bear on the particular object to which the provision’s applicability is being discussed. That the exact language of Article III, paragraph 8, of the Hague-Visby Rules can be straightforwardly used to describe the clause in the bill of lading – to say that that clause did ‘lessen’ the ‘liability of a carrier’ for loss or damage ‘otherwise than as provided in the Rules’ – is, of course, one such reason; and a strong reason at that. But fit, however clear, with the authoritative language of a provision is neither necessary nor sufficient to justify a claim that the provision does apply to some relevant object – as judicial practice shows beyond doubt. The same holds for premise (1′). Here too we can describe the consequence that Denning takes the provision to have – that clause 2 is null and void – in language found in the text of the provision itself. But legal provisions can have consequences that their texts do not mention and may even sometimes contradict. It all depends on what reasons happen to bear on the case in hand – and correspondence with the text is but one of those reasons. Just think of the diverse range of considerations normally called the ‘canons’ of statutory interpretation. These are nothing more and nothing less than pointers to kinds of reasons that can support, in arguments like Denning’s, the justification both of applicability claims and of claims about the consequences a provision has for a certain object.

Law-application is, in a sense, a particularistic affair. What I mean is not that, for any conclusion correctly reached on the grounds that some provision applies to some object x, there is no true universal principle that captures all and only those aspects of x that are relevant for that conclusion – a principle that could be used to construct an inference that would fit a scheme identical to that of the legal syllogism. What I mean is that we – and courts in particular – do not (indeed, could not even be expected to be able to) engage mentally with such principles as a necessary, typical, or even minimally frequent step in either our reasoning or our arguments towards such conclusions.

Any object? Any conclusion?

I have been speaking of the application of provisions to ‘some object x’, and of law-application – in the sense discussed so far – as reasoning towards a conclusion ‘about x’ that is taken to be true as a matter of law. But are there not restrictions on the range of objects to which a provision can be applied? And should we not say instead, perhaps, that applying the law must be reasoning towards a statement of law, or at any rate a statement of the law ‘as applied’ to something?

I think not. ‘Statement of law’ is an equivocal phrase. It is sometimes used to refer to so-called ‘normative’ or ‘deontic’ statements: statements about the normative positions (duties, liberties, powers, and so on) we have under existing law. Courts do often reason towards statements of that kind – statements, for example, like ‘The defendant was under a duty to provide services to the plaintiff’ – by applying existing law. But the range of conclusions that can be reached and justified in that way is not confined to statements of normative positions (Denning’s conclusion in The Hollandia, that clause 2 in that particular bill of lading was ‘null and void’, is itself not – at least not obviously – a deontic statement), and there is therefore no principled restriction on the range of objects to which existing law can (conceptually) potentially be applied.

What does have to be the case is that the conclusion, whatever it is, is taken to be true as a matter of law. This is a notion I cannot fully analyse here. But a simple example will bring out one relevant point: that a statement about some particular object can be put forward either as true as a matter of law, or as true sans phrase. Consider the following pair of provisions, subsections 54(1) and (2) of the Race Relations Act 1976:

54. – (1) A complaint by any person (‘the complainant’) that another person (‘the respondent’) –

(a) has committed an act of discrimination against the complainant which is unlawful by virtue of Part II; or

(b) is by virtue of section 32 or 33 to be treated as having committed such an act of discrimination against the complainant,

may be presented to an industrial tribunal.

(2) Subsection (1) does not apply to a complaint under section 12(1) of an act in respect of which an appeal, or proceedings in the nature of an appeal, may be brought under any enactment, or to a complaint to which section 75(8) applies.

In Khan v. General Medical Council,Footnote 5 a doctor of Asian origin was appealing a decision of an industrial tribunal. The General Medical Council had denied him full registration as a Medical Practitioner; and after unsuccessfully appealing to a Review Board, he complained of indirect racial discrimination to the industrial tribunal, under section 54(1) of the Race Relations Act 1976. The industrial tribunal, applying section 54(2), dismissed the application; and the Employment Appeal Tribunal held in Khan that that had been the correct decision.

The conclusion warranted by the industrial tribunal’s application of section 54(2) to the applicant’s complaint was that section 54(1) does not apply to that complaint. So here is how that tribunal’s argument for that conclusion could be reconstructed:

(1) If section 54(2) of the Race Relations Act 1976 applies to the applicant’s complaint, then section 54(1) of the Race Relations Act 1976 does not apply to the applicant’s complaint.

(2) Section 54(2) of the Race Relations Act 1976 applies to the applicant’s complaint.

Therefore (from (1) and (2)),

(3) Section 54(1) of the Race Relations Act 1976 does not apply to the applicant’s complaint.

Here, both premise (2) and the conclusion are statements of the same kind. Both are applicability statements: one positive, the other negative. But in the context of this argument, there is a crucial difference between the two. The conclusion is being put forth as a claim taken to be true as a matter of existing law. It is justified, as was the conclusion in Denning’s argument, on the grounds that there is a certain provision that is part of existing law – section 54(2) – that applies to, and carries a certain consequence for, the applicant’s complaint.

Premise (2) is also being put forth as true, of course; it is part of the same argument. But it is not (or at any rate not necessarily) being put forth as true as a matter of law. Perhaps it is being put forth – as was the similar premise in Denning’s argument – on the basis of whatever substantive reasons bear on the applicability of section 54(2) to the applicant’s complaint.

A claim about some object x is put forth as true as a matter of law if it is put forth on the grounds that there is some section of existing law that applies to x, and whose applicability to x is taken to suffice for that claim to be true. But any such claim about some object x will ultimately be premised on an applicability claim that is itself put forth as true but not as a matter of law: an applicability claim such that there is no section of existing law that suffices to make it true.

Definition

Our discussion brought out four main aspects of law-application in the sense we have been discussing – a sense I propose to refer to as inferential law-application. First, to inferentially apply the law is to reason in a certain way. Second, it is to reason towards a conclusion about some object x – any object. Third, it is reasoning towards that conclusion on the grounds that a certain provision p applies, with a certain consequence, to x. And fourth, the conclusion is taken to be true as a matter of law.

Here then is a definition:

Inferential law-application (for written provisions): To inferentially apply a provision p (on its own) to some object x is to reason towards a conclusion c about x on the grounds that (a) p applies to x, and (b) if p applies to x, then c is true as a matter of law.Footnote 6

This definition is meant to capture one sense in which we – lawyers and judges, but citizens too – commonly think and speak of law-application. But as I said, inferential law-application is only one of two different kinds of law-application. Let us now consider the other.

Pragmatic law-application

Definition

As you can imagine, inferentially applying the law was not all Denning did in The Hollandia. In fact, had Denning limited himself to reasoning and arguing towards the conclusion that the clause in the bill of lading was null and void – had that been all he had done – he would have been rightly accused of having failed to apply the law to the case before him.

What else did he have to do? He had to, and did, allow the appeal:

My conclusion is that, in proceedings in the courts of the United Kingdom clause 2 of this bill of lading is null and void. This action should not be stayed. It should proceed – with the limit of liability being that prescribed by the Hague-Visby Rules.

. . .

I would allow the appeal, remove the stay imposed by the judge, and allow the action to proceed in England.Footnote 7

So there is a sense of ‘applying the law’ that refers, not to any mental act of reasoning towards some conclusion, but to the actual performance of an external, non-mental action with certain characteristics. This is what I propose to call pragmatic law-application, and here is how I think it can be defined:

Pragmatic law-application (for written provisions): To pragmatically apply a provision p is to perform an action φ such that:

(a) the agent takes φ to be an action that either (a1) she legally ought to perform, or at least (a2) she is legally permitted to perform (in the sense that it is not the case that she legally ought not to perform it);

(b) the agent takes the relevant ought-claim – that is, (a1) or (a2) – to be supported by the conclusion of the inferential application of p to some object x; and

(c) by φ-ing, the agent purports to be authoritatively settling some particular question or matter.

As the definition makes clear, pragmatic and inferential law-application are closely related. I will say more about this relation in Section IV. First, though, there are some aspects of the definition that I want to discuss.

Judicial action and judicial decision

The definition mentions the performance by an agent of a certain action. This could be understood in a wide sense, to include omissions. But in the case of courts, it is the performance of actions in the strict sense, not omissions, that is invariably at stake. That is because doing nothing is seldom, if ever, a live option for a judge. Denning’s option was not to either allow the appeal or simply do nothing. It was to either allow or refuse to allow it – by actually issuing a decision to that effect.

‘Decision’, indeed, is how the relevant action is normally referred to in legal discourse. The court in The Hollandia, we would say, decided to allow the appeal. This does not normally mean much more than ‘the court allowed the appeal’. But it conveys the thought that the court was concerned, as courts typically are, with a driving normative question: ought it to allow the appeal? At stake in pragmatic law-application is the legal justification of such decisions. A court’s decision d is legally justified in one of two scenarios: (a) if d is the decision that the court legally ought to issue in the relevant decision-making context; or, if there is no single decision the court legally ought to issue in that context, (b) if d is not a decision the court legally ought not to issue – if, that is, the court is legally permitted (in one sense of the word) to issue it.

Not all legally justified judicial decisions are law-applying decisions. But what courts legally ought to do turns in part – in large part – on existing law. If a court in deciding in a certain way does correctly apply the law, that suffices for its decision to be legally justified.

This is what we see in The Hollandia. Denning manifestly took the view that allowing the appeal was the action the court legally ought to perform; and he took this view on the basis that there was a certain provision – Article III, paragraph 8, of the Hague-Visby Rules – which applied, with a certain consequence, to the particular clause in the bill of lading. So he took the court’s decision to be legally justified by reference to that provision; he took the provision to normatively support – to provide, as it were, normative warrant to – the court’s action of allowing the appeal.

We can say, then, that to pragmatically apply a provision is to perform an action – to issue a decision – that one takes to be legally justified by reference to the provision. But ‘by reference to the provision’ is still an opaque description: the justificatory link itself, the link between provision and action, is what now needs to be clarified.

Justifying law-applying decisions

One way to think about it is again to consider the structure of the argument that a judge like Denning would give to justify such a decision, to argumentatively show the decision to be legally justified qua law-applying decision. It is not part of my proposed definition that whoever pragmatically applies a provision must also give an argument showing the corresponding action to be legally justified. A court could simply act in the way described in the definition, and it would be applying the law – even if it were impossible to tell, externally, that the action had been performed as an action of law-application. But again, courts do commonly give arguments to that effect, and we can learn something about pragmatic law-application by reflecting on the structure of such arguments; for the justificatory link between the provision being applied and the action being performed will be made apparent in the fully reconstructed inference.

How should we reconstruct Denning’s argument? My suggestion:

(1) If Article III, paragraph 8 of the Hague-Visby Rules applies to clause 2 of the bill of lading issued by the defendants, then clause 2 of the bill of lading issued by the defendants is null and void.

(2) Article III, paragraph 8 of the Hague-Visby Rules applies to clause 2 of the bill of lading issued by the defendants.

Therefore (from (1) and (2)),

(3) Clause 2 of the bill of lading issued by the defendants is null and void.

(4) If clause 2 of the bill of lading issued by the defendants is null and void, and there are no countervailing considerations, then we [i.e. the court] legally ought to allow the appeal and remove the stay on the plaintiff’s action.

(5) There are no countervailing considerations.

Therefore (from (3)-(5)),

(6) We legally ought to allow the appeal and remove the stay on the plaintiff’s action.

This is a complex deductive argument, in two steps. The first step – from (1) and (2) to (3) – is the argument that corresponds, as we saw in Section II, to the inferential application of the provision to the clause in the bill of lading. Premise (2) is the relevant applicability claim; the conclusion, a statement of the effect that the provision is taken to have on the clause. But it is the second step – from (3), (4), and (5) to (6) – that primarily concerns us now.

Premise (4) is the crucial one. What it purports to capture is that what is asserted in premise (3) – which describes the outcome of the inferential application of the provision to the clause – is taken by the court to be a reason for adopting a certain course of action. Premise (4) identifies the action – allowing the appeal – that counts, in the particular decision-making context, as the action of applying the provision; and signals, with the ‘there are no countervailing considerations’ rider, that the fact that the clause in the bill of lading is null and void supports normatively, although it does not conclusively establish, the conclusion that the court legally ought to allow the appeal.

Logically speaking, of course, the inference could be compressed. If a court is disposed to endorse premises (1) and (4) in the inference above, then it could instead directly put forward the following conditional, which those two claims entail (by hypothetical syllogism):

(1′) If Article III, paragraph 8 of the Hague-Visby Rules applies to clause 2 of the bill of lading issued by the defendants, and there are no countervailing considerations, then we legally ought to allow the appeal and remove the stay on the plaintiff’s action.

Together with (2) and (5), this claim entails the final conclusion. So the court could simply have offered the following one-step inference, rather than the more complex one I gave above:

(1′) If Article III, paragraph 8 of the Hague-Visby Rules applies to clause 2 of the bill of lading issued by the defendants, and there are no countervailing considerations, then we legally ought to allow the appeal and remove the stay on the plaintiff’s action.

(2) Article III, paragraph 8 of the Hague-Visby Rules applies to clause 2 of the bill of lading issued by the defendants.

(5) There are no countervailing considerations.

Therefore (from (1′), (2), and (5)),

(6) We legally ought to allow the appeal and remove the stay on the plaintiff’s action.

This would do, argumentatively, but only if neither (1) nor (4) in the more complex version of the argument happened to be controversial in the particular decision-making context. The court would, in any event, have to be prepared to both unpack and defend the move from the applicability claim – premise (2) – to the ought-claim in the conclusion; and the way to do it would be by specifying the substantive claim that is the conclusion of the inferential application of the provision to the clause.Footnote 8

Support

I said that in the complex argument above, premise (4) presents the fact that the clause in the bill of lading is null and void as a reason supporting the court’s decision to allow the appeal. Why merely as a reason rather than as a fact sufficient to establish that allowing the appeal is what the court legally ought to do? Was the court not legally bound to allow the appeal if indeed the clause was null and void?

No. For one, the applicability of Article III, paragraph 8 of the Hague-Visby Rules, with its effect of rendering the clause null and void, is consistent with there being other sections of positive law that might also bear on the matter – including other provisions applicable to the same clause – and that could provide reasons against allowing the appeal (or at least support the conclusion that it is not the case that the court legally ought to allow it); there can be conflicting but equally valid applicable provisions. And more generally, even in the absence of any conflicting provision, it could be that the court was legally permitted not to allow the appeal in the name of some legally relevant countervailing reason. What a court legally ought to do is never solely determined by the applicable positive law.

By a ‘reason’ I do not mean only a consideration that counts either for or against deciding in a certain way. Consider provisions like the following (from the Criminal Justice (Scotland) Act 2016):

35 Authorisation for questioning

(1) The court may authorise a constable to question a person about an offence after the person has been officially accused of committing the offence.

. . .

This provision, which confers a liberty, is one that a court that decides to authorise such questioning may be pragmatically applying. The court may correctly take its decision to be one that it is legally permitted to issue (and so one such that it is not the case that the court ought legally not to issue it); and it will take the provision to support (though, again, not conclusively) that ought-proposition. But it does not follow, and does not have to be true, that the court legally ought to authorise questioning. It could be true neither that it ought, nor that it ought not, which would suffice for the decision to be legally justified; and in any case, that provision does certainly not provide the court with any reason for deciding either way. Not all facts that support ought-propositions are reasons for or against anything; and by a ‘reason’ what I mean is a fact that supports an ought-proposition.Footnote 9

Note, by the way, that the conclusion of the inferential law-application of some provision may well be a claim directly about the same action that the court is then justified in performing if it is to pragmatically apply the provision. That is what we see in the example just given. The provision that justifies (absent countervailing considerations) the court’s decision to authorise a constable to question someone is itself a provision about courts’ decisions to authorise constables to question someone. (In The Hollandia, by contrast, the provision being applied concerned liability-restricting clauses in contracts of carriage, not whether certain appeals should be denied.) But we should still differentiate in such cases between the conclusion of the inferential application of that provision in some particular situation (e.g. ‘This court has a liberty to authorise a constable to question this person’), and the ought-claim that that conclusion supports (‘It is not the case that this court legally ought not to authorise a constable to question this person’.). The conditional ‘If Section 35(1) of the Criminal Justice (Scotland) Act 2016 applies to the present situation, then it is not the case that this court ought not to authorise a constable to question this person’ is not a tautology.

Acting in an authoritative capacity

What makes Denning’s action – his decision – an act of law-application, though, is not simply the fact that he took it to be one that he ought legally to perform. It would not be an act of law-application had it not been performed – or, at least, had Denning not purported to perform it – in a specific role (that of a judge, in his case) and as a means of authoritatively putting an end (even if only a provisional one) to a certain matter or controversy.

We can as citizens, some authors remark, engage with the law in various ways – we can follow it, comply with it, break it, avail ourselves of it, invoke it, and so on – but not quite by ‘applying’ it.Footnote 10 This is only partly true, as we now know: for we do often as mere citizens inferentially apply the law. Suppose you want to get married but, having recently obtained a divorce, wonder whether you are legally able to marry again so soon. So you read the relevant statutes and conclude that you are indeed permitted to remarry. You have applied – inferentially applied – the relevant provision or provisions to come to that conclusion.

But it is true that by then going ahead and actually getting married you will not be performing any further act of law-application; you will merely be exercising a legal power and availing yourself of the accompanying liberty.Footnote 11 Likewise, I may by inferentially applying some provision come to the conclusion that legally I have a duty to pay you a certain amount of money; but by paying you I am not ‘applying’ the law: I am merely discharging my duty.

It is not in our capacity as citizens that we pragmatically apply legal provisions. We pragmatically apply a legal provision only when, in some specific role or capacity – such as, paradigmatically, that of a judge – we exercise (or at least purport to exercise) the authority to bring some matter or controversy to a resolution of some kind. The relevant action – the decision – is meant to settle such matter, and is taken by the agent to be justified by reference to the provision.Footnote 12

How are inferential and pragmatic law-application related?

Pragmatic law-application conceptually presupposes – as my definition makes clear – inferential law-application. But there is more to say about the relations between the two kinds.

First, the agent who pragmatically applies a provision p need not be the same person who has inferentially applied it to whatever the relevant object happens to be. Typically, of course, it will be the same judge or court who will both inferentially apply a provision to some object x, drawing the relevant conclusion; and then go on to pragmatically apply that provision by performing the action that that conclusion supports. But the two operations can fall upon different entities. That is what appears to be the case, for example, when a national court of a member state of the European Union refers an interpretative question of EU Law to the Court of Justice of the European Union. The Court of Justice’s ruling may well involve the inferential application of EU Law – and will then be relied upon by the national court as it proceeds to pragmatically apply the relevant law.Footnote 13

Second, neither pragmatic nor inferential law-application is explanatorily reducible to the other. Why might someone think that, given my definitions, pragmatic law-application is reducible to inferential law-application? The thought would be that the pragmatic application of a certain provision p could be understood in terms of the inferential application of something else: of the putatively legal rule that, if p applies with a certain consequence to a certain object, then, absent countervailing considerations, the court legally ought to decide in a certain way (or at least it is not the case that legally it ought not to decide in a certain way). The pragmatic application of p would then be explainable as (a) the inferential application of this rule, simply coupled with (b) the actual performance by the court of the relevant action. In The Hollandia, for example, the court’s pragmatic application of Article III, paragraph 8 of the Hague-Visby Rules could be explained – so the thought would run – as the inferential application of the rule that, if that provision applies to the clause of the bill of lading with the consequence of making it null and void, then, if there are no countervailing considerations, the court legally ought to allow the appeal. Having reasoned towards this conclusion – seemingly meeting my definition of inferential law-application – the court simply acted accordingly. There is nothing left to account for or explain.

But this line of reasoning is misguided. It is true that in pragmatically applying a provision p, a court will have reasoned that if p applies with a certain consequence to a certain object, then, absent countervailing considerations, its decision is one that it legally ought to issue (or not one it ought not legally to issue). But conditional statements like this, if true, are not themselves true as a matter of law. If they state a rule, it is not one that is part of existing law. Rather, it would be a rule about existing law and how it bears on what the court legally ought to do; a rule about what to do if some part of existing law – provision p – applies to some given object.

Or to put things more generally: the rule (or principle, if that is what it is) that courts ought to apply the law is not itself part of the law that it directs courts to apply. This is not to say that there could not be a legal rule – part of existing law – requiring courts to apply existing law; a self-referential rule even. But we would still need to know if courts legally ought to apply existing law, including that rule: and if they ought, the rule or principle that so determined would itself not be part of existing law. So pragmatic law-application cannot be explained in terms of inferential law-application. (Not to mention, of course, that inferential law-application is based, as I argued in Section II.D, on second-order applicability claims, not first-order claims stating the content of whatever it is that is being applied.)

Nor is inferential law-application reducible to pragmatic law-application. Here, the thought would be that if inferential law-application comes down to, as I suggested, the performance of a mental act, then it could be brought under my definition of pragmatic law-application, given a suitably broad reading of ‘action’. But again, this would not do. What pragmatic law-application comes down to is the performance not just of an action, but of an action that the agent takes to be legally justified on the basis of existing law; and that is never true of the mental act of coming to some conclusion about some object on the basis of existing law. Perhaps it is true that in The Hollandia, Denning ought to have reached, as he did, the conclusion that the clause in the bill of lading was null and void as a matter of law. But the mental act of reaching this conclusion is itself not an act that stands to be justified by reference to existing law. Rather, it stands to be justified by reference to whatever substantive (and logical) considerations govern how courts should rely on existing law to reason towards conclusions about particular objects like clauses in bills of lading.

Still, the two kinds of law-application are related not just conceptually but also normatively. I mentioned in Section II.D that the justification of applicability claims (and of claims about their consequences) turns on a range of substantive considerations, to which a court should be responsive. These include the further consequences that will (or would) follow if the conclusion that is the outcome of inferentially applying that provision is indeed taken to be true as a matter of law, and acted upon. In other words, the fact that there would be a certain decision d that a court ought to issue if the relevant provision applies with a certain consequence to some object x is itself a consideration to be taken into account when determining whether or not the provision does apply to x. If d is in some way an undesirable or unwelcome or absurd decision to issue in the particular decision-making context, then that is a consideration that bears on the court’s reasoning about whether or not the relevant provision applies. Pragmatic law-application, in short, normatively feeds back into inferential law-application.

Cases

As I mentioned in Section I, courts are often said to apply the law both to and in cases. So far I have tried to avoid the language of ‘cases’, and succeeded almost always. But my proposed definitions had better be at least compatible with it. Are they?

The notion of a case is less clear than it seems. We encounter in legal discourse, I think, at least four senses of the word.

Sometimes – call this Sense 1 – what we mean by a case is just, well, nothing at all. Consider the following claim:

(A) The provision in Article III, paragraph 8, of the Hague-Visby rules applies to the case of a foreign law clause in a bill of lading.

This is something that the The Hollandia court could have said. But what is this ‘case of a foreign law clause’? It is not what the court would call ‘the case in hand’; that would make no sense. The claim above is simply, it seems to me, a needless periphrasis.Footnote 14 It is a roundabout way of saying that

(B) The provision in Article III, paragraph 8, of the Hague-Visby rules applies to foreign law clauses in bills of lading,

which is itself to be understood as the claim that the fact that a clause is a foreign clause in a bill of lading does not mean that the provision will not apply to it.

Likewise,

(C) The provision in Article III, paragraph 8, of the Hague-Visby rules applies to the case of clause 2 in the bill of lading issued by the defendants

means nothing more than

(D) The provision in Article III, paragraph 8, of the Hague-Visby rules applies to clause 2 in the bill of lading issued by the defendants.

In another sense of the word – the one in play when we speak of a judge deciding the case before her – a case is really a particular question. It is a question brought to a judge and on which she is expected to rule. It is always a normative question, and it always concerns some action by the judge or court. Should the judge enforce this contract? Convict this defendant? Allow this appeal? To ‘decide the case’ is for the court to adopt one particular course of action: the course of action that corresponds to what the judge thinks she legally ought to do when facing such a question.

But although a case in this sense of the word – Sense 2 – is a question rather than a set of facts, it is often a question raised in connection with a particular set of facts (or descriptions of facts) also brought before a judge. These are what we would normally call the ‘facts of the case’; and such facts – themselves not the ‘case’ in the sense just specified – are what we sometimes also call ‘the case’ before the judge. So in this sense – Sense 3 – of the term, a ‘case’ is a series of descriptions of particular facts, in connection with which a certain normative question (a ‘case’ in Sense 2) has been raised and brought before a judge to be decided.

In a fourth sense, the term refers more broadly to any decision-making context in which a court is asked to rule on a case in the second sense of the term. ‘In this case’, the court could have said in The Hollandia, referring to the particular decision-making context in which it found itself, ‘we decide to allow the appeal’.

(There is a fifth sense, although it is of no interest for our present purposes; it is the sense in which we speak, for example, of The Hollandia as itself a ‘case’: a ‘case’ we can refer to, quote from, analyse, criticise, and so on. Here the term refers comprehensively to a certain decision-making event with all its components, including the actual decision issued by the court, together with its record in the court’s written opinion.)

Do courts, then, apply the law to or in – or in deciding – cases in any one of these senses? Yes. In Sense 1, the term is used to refer – although redundantly – to the indirect object of inferential law-application: to that to which the law is being inferentially applied. In Sense 4, it refers to the decision-making context in which the court faces a normative question to which the ought-claim of pragmatic law-application provides an answer. And in Sense 2, it refers to this very question, the question in the decision of which the court will be pragmatically applying the law.

Conclusion

I started this paper by identifying three sets of questions about law-application: questions about what I called (a) its direct object, (b) its indirect object, and (c) its content. I left all three for the end, though for different reasons.

I postponed dealing with questions of indirect object and content because, I explained, these might receive different answers depending on the type of law-application we have in mind. What I have just said, in Section V, about the notion of a case, shows that this is indeed true of the indirect object of law-application. The indirect object of inferential law-application is a ‘case’ only in that first, empty sense of the word. It is, as we saw in Section II.E, any object x such that the applicability of some part of existing law to x will bear on whether a certain statement about x is true as a matter of law. That is what a court will inferentially apply the law to. Pragmatic law-application, by contrast, is application, not to but in a case, or in deciding a case – but in other senses of ‘case’. As to the content of each kind of law-application, that is precisely what my proposed definitions purport to characterise.

Questions about the direct object of law application, though, I postponed merely for reasons of convenience. Focusing on the application of single written provisions allowed me to rely on clear and simple examples as I developed and defended my analyses. But what can we say more generally about the direct object of law-application?

We can say that it encompasses any context-salient part of existing law: any section of positive (or indeed non-positive, depending on what your legal metaphysics allows) law as it exists at the time of application, and which can be identified as such – as part of existing law – independently of any views on its applicability and application to any object (and independently also of any views on the normative desirability of pragmatically applying it in any particular decision-making context). It could be a provision, but also, for example, a set of provisions taken together, or a precedent decision, or a custom – or whatever it is that properly counts as part of law under the correct theory of what law is.

Change history

Notes

  1. The Hollandia [1982] Q.B. 872 (at 884A-B).

  2. See José Juan Moreso and Samuele Chilovi, ‘Interpretive Arguments and the Application of the Law’ in G. Bongiovanni et al (eds.), Handbook of Legal Reasoning and Argumentation (Springer 2018) 495.

  3. For detailed discussion, and further references, see Luís Duarte d’Almeida, ‘On the Legal Syllogism’, in David Plunkett, Scott Shapiro and Kevin Toh (eds.), Dimensions of Normativity: New Essays on Metaethics and Jurisprudence (Oxford University Press 2019) 335–364.

  4. Including wondering whether such arguments might perhaps be characterised as non-deductive. Here, however, I assume that the justification of law-applying decisions can indeed be understood as a deductive inference: that too the received view; but on this specific point the received view is, I think, well supported. (This is not to say that all legal arguments are best characterised as deductive.) For references see, again, Luís Duarte d’Almeida ‘On the Legal Syllogism’, supra n 3.

  5. Khan v. General Medical Council (17 BLMR 1-10 [1993]).

  6. Both this definition and that of pragmatic law-application (shortly to be offered in Section III.A) are meant as analyses of the notions being defined. The discussion so far has sought to identify features that we, as competent users of those two concepts (though not, of course, under the names coined here for each), do take – and can now, upon reflection, recognise – the two kinds of law-application to have. The following section does the same for pragmatic law-application. The proposed definitions aim to crystallise such features in a perspicuous formulation.

  7. The Hollandia [1982] Q.B. 872 at 884G-885A.

  8. In ‘Applicability and Effectiveness of Legal Norms’, Law and Philosophy 16 (1997) 201–219, at 203–207, Pablo E. Navarro and José Juan Moreso distinguish between the ‘internal’ and the ‘external’ applicability of legal norms. The former notion is intended to capture what I called – in Section II, when discussing inferential law-application – applicability tout court; it concerns a conceptual relation between a legal norm and a certain individual case. (See also Pablo E. Navarro and Jorge L. Rodríguez, Deontic Logic and Legal Systems, Cambridge: Cambridge University Press (2014), 126–129.) ‘External’ applicability, by contrast, is a normative notion: to say that a norm N is externally applicable to a certain generic case C is, in their definition, to say that the application of N to the individual cases that are instances of C is required (or at least permitted) by some other norm. But this is not sufficiently fine-grained: their notion of external applicability – with its generic reference to the application of a norm, a concept they do not explore – glosses over the distinction between inferential and pragmatic law-application. It is the latter kind only, I think, that they have in mind.

  9. Euan MacDonald and I try to give an an account of reasons in terms of the ‘support’ relation – which focuses on the relation between the facts that constitute reasons and the truth-value of ought-claims (as opposed to the ‘favouring’ relation, which focuses on the relation between the facts that constitute reasons and what they are reasons for – actions, beliefs, etc.) – in Luís Duarte d’Almeida and Euan MacDonald, ‘Contra Tantum Reasons’ (unpublished typescript, on file with the author). For present purposes, however, it suffices to say that my use of ‘reason’ includes both ‘favourers’ and ‘supporters’ – both reasons for and reasons against a certain course of action.

  10. See, e.g., Gerald C. MacCallum, ‘On Applying Rules’ in Legislative Intent and Other Essays on Law, Politics, and Morality (edited by Marcus G. Singer and Rex Martin), The University of Wisconsin Press (1993) 64–74, at 63, 72–74; Giorgio Pino, ‘L’applicabilità delle norme giuridiche’, Diritto e Questioni Pubbliche 11 (2011) 797–871 at 802–809; Mathieu Carpentier, ‘Validity versus Applicability: A (Small) Dose of Scepticism’, Diritto e Questione Pubbliche 18 (2018) 107–132 at 109–111.

  11. Thanks to Antony Duff for discussion on this point.

  12. This is not to say that we can never as mere citizens perform actions that might be called actions of ‘law-application’ in some sense of the notion – a sense not reducible to notions such as that of complying with the law, that of following the law, that of doing as the law requires, and so on. Considerations like the ones just mentioned are not enough to settle that point. But it is not a point I aim to settle in this paper. My focus is the particular and distinctive sense in which we say of courts (and other institutional agents) that they characteristically by their actions apply the law. We think of courts as law-applying bodies; and what we mean is not that courts simply happen to – as anyone else might too – apply the law.

  13. Thanks to Maggie O’Brien for discussion here, and to Euan MacDonald for the example.

  14. Arthur Quiller-Couch has some amusing remarks on the word ‘case’ – ‘jargon’s dearest child’, he calls it – in On the Art of Writing (Cambridge: Cambridge University Press 1923) 77. Fowler was also not a fan: ‘there is perhaps no single word so freely resorted to as a trouble-saver, and consequently responsible for so much flabby writing’; see Fowler’s Modern English Usage, ed. Ernest Gowers (2nd edition, Oxford: Oxford University Press 1968) 76–77; and H.W. and F.G. Fowler, The King’s English (third edition, Oxford: Oxford University Press 1931) 15–16. In case you are wondering, Fowler does not object to ‘fan’. (He would also not have objected – in case you are also wondering – to ‘in case you are wondering’: it is one of the ‘legitimate uses’ of the word. ‘In the case of ‘, though – that is ‘the worst offender’.)

ACKNOWLEDGMENTS

Thanks to Antony Duff, Euan MacDonald, Gonçalo de Almeida Ribeiro, Hasan Dindjer, Lucas Miotto, Maggie O’Brien, audiences in Edinburgh, Florence, Glasgow, Lisbon, Milan, and Oxford, and two anonymous reviewers, for helpful comments on earlier versions. I started to think about the topic of this paper in the spring of 2017, when I was a Fernand Braudel Research Fellow at the European University Institute in Florence. I am grateful to the EUI for hosting me, and for the excellent conditions that I enjoyed during my stay.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Luís Duarte d’Almeida.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The original online version of this article was revised due to a retrospective Open Access order.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Duarte d’Almeida, L. What Is It to Apply the Law?. Law and Philos 40, 361–386 (2021). https://doi.org/10.1007/s10982-021-09405-x

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10982-021-09405-x