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.
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.
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.