Law as authoritative fiction
Law and fiction as closed prefixed contexts
different points of view in an argument may result in a fallacy. So here is Lewis’s famous example:
Sherlock Holmes, we are told in the Arthur Conan Doyle mysteries, lived at 221B Baker Street in London. Let us assume, therefore, that there is some sense in which (1) is true:
(1) Sherlock Holmes lived at 221B Baker Street, London
It turns out, however, that the building at 221B Baker Street in London at the time was a bank. Let us therefore assume that at the relevant times, (2) is true:
(2) The building at 22IB Baker Street, London, is a bank.
The inference from (1) and (2) would seem to be:
(3) Sherlock Holmes lived in a bank.
But, of course, (3) is clearly false. What has gone wrong here? Lewis tells us that we made the mistake of moving from a prefixed to an un-prefixed context. (1) is true only if it is prefixed by an operator such as “In the fiction F ...”; whereas (2) is true only if taken as un-prefixed (in the real world, as it were). Thus, unless (2) is prefixed by the same operator “in fiction F ...”, you cannot conclude that (3) is true in the fiction; and because (1) is true only if it is prefixed, you cannot conclude that (3) is true in an un-prefixed sense. Surely, this is quite right, and I will refer to this problem as the Lewis fallacy.
Now, the Lewis fallacy requires some refinement, even in the case of fiction. Consider the following inference:
- (4) Sherlock Holmes lived in London.
- (5) London is a city in the United Kingdom
- (6) Sherlock Holmes lived in the United Kingdom.
We have the same structure here as in (1) to (3), but a very' different result. (4) is clearly prefixed by' “In the fiction F...”, whereas (5) seems to be un-prefixed; it is just a fact in the real world that London is in the UK. But there is no fallacy here. Any' sensible reader of the Conan Doyle mysteries would have assumed, and rightly so, that Sherlock Holmes’s escapades take place in the UK. And, crucially, this would be the case even if the United Kingdom (or England, or Great Britain) is never explicitly mentioned in the text. So what is it that makes the inference of (4) to (6) valid, as opposed to (1) to (3) which is not?
The solution has to be this: Although (5) seems to be an un-prefixed proposition, in the context of this inference it is not; (5) is incorporated into the fiction by' implication. The assumption here is that fictions typically incorporate by' implication an indefinite, though limited, number of facts (or factual assumptions) about the world, at least those that are relevant, and can be assumed to be salient and well known to potential readers. A reasonable reader of the Sherlock Holmes mysteries can be expected to know that London is a city in the UK, and thus, even if the text does not mention this explicitly, it can be regarded as incorporating it by implication. (Assuming, of course, that there is nothing in the text to suggest otherwise.) Thus we avoid the Lewis fallacy; the entire inference of (4) to (6) should be regarded as contained within the prefixed context. Similarly, even if the mysteries never refer to Sherlock Holmes’s nose, we can assume that he had one (and only one), in virtue of the stated fictional fact that he is a man. These kind of unmentioned facts can be said to be incorporated in the fiction by implication and thus, for inferential purposes, they can be regarded as prefixed statements. The conditions and limits of implicitly incorporated assumptions in fiction are problematic, of course, and sometimes indeterminate.
As we mentioned above, there are prefixes which create a Lewis-ty'pe fallacy, and others which don’t. Let me call them closed and open prefixes, respectively. Open prefixes are such that they can occur in arguments with un prefixed statements to yield valid conclusions. For example, prefixes such as “According to the laws of nature (or, according to science; or according to the laws of thermodynamics) ...”; or “According to modal logic ...”. So what is it about closed prefixes that they create the Lewis fallacy'? One suggestion might be to look at the semantics of the prefix. It is probably' implicit in the semantics of scientific prefixes that they range over un-prefixed statements to yield valid conclusions. In other words, it is probably part of what it means to prefix a statement by “according to science” that the statement following the prefix is meant to apply unconditionally.7 Whereas it is part of the meaning of a prefix such as “according to fiction F...” that it ties the truth value of the statement to be contained within a world demarcated by the prefix -that is, the world of fiction F. This is probably true, but it may not be enough. Still, you may' wonder, what makes it the case that some prefixes are closed? The suggestion that I will endorse and try' to support throughout this chapter is that some prefixes are such that they designate a constitutive relation to the truthvalues of the statements prefixed by them. A statement is true in a fiction, if it is, because the fiction states it. The saying so makes it true, so to speak. It doesn’t mean that there are no limits here; even fiction cannot make it the case that X is true in the fiction if X is something that we cannot possibly imagine or otherwise entertain in your minds (such as a “square circle” or even, perhaps, “a delightful torture of an innocent child”.) What those limits are and how to account for them is an interesting question, but we need not try' to resolve it here.8
In short, closed prefixes are those in which a constitutive relation obtains between certain essential features of the world/context designated by the prefix, and the truths of the statements expressed in that world/context. Now' you might wonder what other prefixes are closed, besides fiction, that is? Law, I will argue shortly, is a closed prefix, but there are many others. Structured games, for example, constitute closed prefixed contexts. When the umpire in a soccer game orders a player “out”, it is true, in the game of soccer, that the player has to leave the field. The game is a closed prefixed context. It is only according to the rules and conventions of the game that if the umpire orders the player “out”, then the player, in the game, from the game’s perspective, as it were, is to leave. The rules and conventions of the game constitute the context that determines certain statements about that context to be true or false.9
Why is the law, however, a closed prefix? A good starting point is to see that the truth value of any statement about the content of a legal premise depends on the specific jurisdictional prefix. Take any statement expressing some legal premise, such as (7) “A ought (legally) to ip in C”, and you will immediately realize that whether (7) is true or false depends on the specific prefix. According to the law in Si at t, (7) can be true, and according to the law in S2 at t it may not be, and the same goes for the time variant. It may have been true until last week but no longer is. Some legal philosophers may doubt that this is necessarily the case. Perhaps there are some universal truths about legal contents that are not prefix-variant. What I have in mind here are ven' general principles, such as the idea that wrongs need to be rectified, criminals need to be punished, or agreements need to be kept, and things like that. There is a sense in which it might be true that all legal systems must instantiate some such principles, and therefore there might be a sense in which such abstract and general precepts are universally legal, as it were, un-prefixed. I doubt that there is a serious sense in which such principles are legal, but it doesn’t matter much for our purposes. After all, actual laws can deviate from such principles; not all wrongs are legally rectified and not all promises are binding contracts, etc. So there might be some general principles that law purports, perhaps even necessarily, to instantiate; but it is the instantiation that legally matters, not the guiding ideals.
Be this as it may, suffice it for our purposes if you agree that, by and large, the truth value of legal statements is prefix-dependent. But still, you might think that this is not enough to make the legal context a closed prefix. After all, we don’t seem to get the Lewis fallacy by mixing legal statements with un-prefixed statements of non-legal facts. On the contrary', almost every legal syllogism seems to do just that, reach a legal conclusion from statements about the law and statements about non-legal facts. Consider this example:
- (8) According to the law in S at t, anyone who is an X ought to
- (9) John is an X (in S at t)
- (10) Therefore, John ought, legally, to ip in C.
Clearly (8) is a prefixed statement and it would seem that (9) is not; it is just a fact in the world, so to speak, that John happens to have feature X in the relevant time and circumstances. And yet, despite the fact that we seem to mix a prefixed premise (8) with an un-prefixed premise (9), we get a valid conclusion in (10). But this is an illusion. The legal conclusion in (10) follows only if (9) is incorporated into the legal prefix by implication, presupposition, or stipulation. It has to be the case that according to the law in S at t, John is an X. Factual premises must be incorporated, explicitly or implicitly, into the legal framework to make legal syllogisms valid.10
Consider a simple case. Suppose it is the law (here and now) that accepting an offer under specified conditions amounts to the formation of a legally valid contract. Suppose you offer to sell me your car for a certain amount of money, and I reply by saying “Yes, I accept the offer”. May we conclude that we just formed a legally valid contract? Not so fast. It depends on whether our relevant expressions in the particular context count as an offer and an acceptance in the eyes of the law; there has to be an explicit or implicit legal finding that they do. And notice that these kind of findings can always be legally challenged and debated. Events in the world need to be internalized into a legal framework; they need to be legally recognized as relevant or established facts, to form premises in legal arguments. Without such incorporation to the prefixed context, we do get the Lewis fallacy.
Lawyers often talk about something similar but in a way that is potentially confusing. Lawyers acknowledge the “finding of fact(s)” as an essential aspect of any legal argument, but the expression is ambiguous. In one sense it can mean exactly what I suggest here - that the relevant facts need to be incorporated into the legal framework - but in another, perhaps more common sense, it can simply mean the authoritative ascertainment of what really happened. In litigation contexts, therefore, the finding of facts is a two-stage affair: courts may need to ascertain what really happened, whether A really said that X or whether B really intended to kill his victim, things like that; and then when the event in the world has been found, authoritatively, to have occurred, there is a finding, as it were, though often implicit, that the way things happened counts as the legally incorporated fact or event. It may be worth keeping in mind that, quite generally, the assumptions about what counts as a legally relevant fact inevitably guides the findings of fact in the factual sense. After all, events and actions in the world allow for an indefinite number of descriptions and levels of specification. Which ones we pick out to ascertain for purposes of formulating premises in a legal argument depends on what counts, legally speaking, as the relevant kind of fact. Therefore, when we include a factual assumption in a legal argument, that is, when we rely on a certain non-legal fact as a premise in the argument, we thereby implicitly incorporate it into the legal prefix. By stating that this fact matters, legally, we make that fact, as it were, part of the legal framework; we implicitly assume that it is the kind of fact that counts, legally, as the relevant kind.
To put this in a formal way, any legal syllogism involving premises about legal content and non-legal facts, must conform to this form:
- (1) According to the law in S at t, if X then legal result Y
- (2) Ex [an event in the world of X type]
- (3) Ex legally counts as X in S at t.
- (4) Therefore according to S at t, X
- (5) Therefore [legal result] Y in S at t.
Unlike in fiction, however, it would seem that there are no antecedent conditions attached to the kinds of facts law can incorporate into its prefixed context. In the fictional context we would think that factual assumptions can be incorporated within a given fictional story only under certain conditions, such as common knowledge of the relevant facts, relative salience, relevance to the story, lack of contrary' indication in the text, etc. In the legal case, however, it seems that there is no room for such constraints; in principle, any non-legal fact or event in the world can be incorporated into the legal context if the law applies to it. And the law can apply to almost anything. Now this, I think, is true but only up to a point or, to be more precise, with an important caveat. The law does not necessarily incorporate into its prefixed context every true factual assumption that follows, logically and empirically, from facts it relies upon, explicitly or implicitly. In other words, non-legal facts are not automatically incorporated into the legal context even if they are true and potentially relevant; incorporation of facts to the prefixed world of a legal context often requires authoritative finding, and sometimes the actual authoritative finding defies empirical truth, and even logical consistency.
Suppose, for example, that the law in S at t is -
(11) “Anyone in possession of an object with feature X ought to
Now suppose that people would normally assume that objects of type A, B, and C are objects with feature X; so they would assume that in context Z, anyone who possesses A, B, or C ought to
- [1] Saying so makes it so 2 At least part of what explains the fact that law, like fiction, is a closed prefixed context stems from the fact that in both cases, of law and of fiction, the domain
