Taking the legal perspective seriously

Abstract

1. Adam Perry (2023) rejects a perspectivalist understanding of legal obligations. That A is legally obligated to φ, under perspectivalism, can be translated to:

(1) According to law, A has a moral obligation to φ.

Perry says that perspectivalism cannot account for mixed arguments. By this he refers to arguments that involve both legal and non-legal claims. He offers the following example:

(2a) All and only persons over 18 have a legal right to vote.

(2b) A has a legal right to vote.

(2c) So A is over 18.

(2a) and (2b) are legal claims, while (2c) is not. So this is a mixed argument in Perry’s sense. The argument appears valid. But suppose we were to translate the argument along perspectivalist lines:

(3a) According to law, all and only persons over 18 have a moral right to vote.

(3b) According to law, A has a moral right to vote.

(3c) So A is over 18.

Now the argument is plainly invalid.

As Perry (2023: 720) recognizes, there is a simple change that preserves its validity. Suppose we were to prefix (3c) with ‘according to law’, the perspectivalist operator. Then we would get:

(3a) According to law, all and only persons over 18 have a moral right to vote.

(3b) According to law, A has a moral right to vote.

(3c*) So according to law, A is over 18.¹

To infer (3c*) from (3a) and (3b), however, presumes the legal perspective is closed under implication. This requires the addition of a legal closure principle. Here is how Perry formulates the principle:

(Legal Closure) If p is entailed by q, r, s, …, and according to law each of q, r, s, …, then according to law p.

But this leads to absurd conclusions. Consider:

(4a) According to law, A is a competent adult.

(4b) If A is a competent adult, then A is morally obligated to keep her promises.

(4c) So, given Legal Closure, A is morally obligated to keep her promises according to law.

(4d) So, given perspectivalism, A is legally obligated to keep her promises.

As Perry concludes, (4d) could easily be false even if (4a) and (4b) are true. The problem generalizes. Given Legal Closure, far too many things will be true according to law.

We agree. Nothing about this, however, should lead someone to reject perspectivalism. For perspectivalism can explain mixed arguments without Legal Closure. A closure principle is necessary, but not one which leads to the problem Perry identifies. Take this alternative:

(Legal Closure*) If according to law [Q therefore P], and according to law Q, then according to law P.

The only important difference from Legal Closure is that it requires the conditional [Q therefore P] be endorsed by the legal perspective.² This precludes the inference Perry finds problematic. Given Legal Closure*, (4c) no longer follows from (4a) and (4b). This is because the conditional in (4b) is not prefixed with ‘according to law’. Thus we cannot attribute the consequent to law (even if the antecedent is true according to law).

For the argument to follow, we must make the following modification:

(4a) According to law, A is a competent adult.

(4b*) According to law, if A is a competent adult, then A is morally obligated to keep her promises.

(4c) So, given Legal Closure*, A is morally obligated to keep her promises according to law.

(4d) So, given perspectivalism, A is legally obligated to keep her promises.

Now the conclusion ceases to be problematic. In many jurisdictions, breaking a promise (on its own) does not violate a legal obligation. But this is because (4b*) does not obtain. In those jurisdictions, the law does not take a position, one way or another, on whether competent adults are under a moral obligation to keep their promises.

Further, Legal Closure* allows the perspectivalist to make sense of mixed arguments. Return to this mixed argument:

(3a) According to law, all and only persons over 18 have a moral right to vote.

(3b) According to law, A has a moral right to vote.

(3c*) So according to law, A is over 18.

Under Legal Closure*, if according to law Q therefore P, and according to law Q, then according to law P. Let P be ‘A is over 18’. Let Q be ‘A has a moral right to vote’. Given (3a), [Q therefore P] is true according to law. So, given Legal Closure* and (3b), A is over 18 according to law. It follows that perspectivalism, with the help of Legal Closure*, avoids Perry’s objection.

2. You may worry that Legal Closure* is ad hoc. Why think – if the legal perspective is closed under implication (as it must be if it is to account for mixed arguments) – it is closed along the lines of Legal Closure* and not Legal Closure? Because when we close a perspective, we should use only those premisses that are true according to that perspective.

Take a novel (N) in which the main character (Sherlock) lives at 221B Baker Street. Something akin to Perry’s closure principle would license the following inference:

(5a) According to N, Sherlock lives at 221B Baker Street.

(5b) If Sherlock lives at 221B Baker Street, he lives in a bank.

(5c) So according to N, Sherlock lives in a bank.

But this cannot be right. For (5c) could easily be false. After all, in the novel 221B might be a rooming house, not a bank. That, as Lewis (1983: 262) points out, is precisely what the Sherlock fiction says.³ (5a) is true only because it is ‘taken as prefixed’ (Lewis 1983: 264). Conversely (5b) is true only because it is ‘taken as unprefixed; for in the stories there was no bank there but rather a rooming house’ (Lewis 1983: 264). So (5c), a prefixed conclusion, simply cannot follow from these premisses. Indeed, ‘nothing follows: neither the unprefixed conclusion that Holmes lived in a bank nor the prefixed conclusion that in the stories he lived in a bank’ (Lewis 1983: 264).

Perry was therefore right to reject Legal Closure. But he misses the diagnosis. It is not wrong because the legal perspective is not closed. Rather it goes awry by trying to close a perspective using claims (here, conditionals) which are not true according to that perspective. The truth of (5b) is irrelevant to whether the consequent is true according to the fiction. And so, for the argument to be valid, (5b) must be replaced with:

(5b*) According to N, if Sherlock lives at 221B Baker Street, then Sherlock Holmes lives in a bank.

Even though (5b*) is false, the argument is now valid. All the premisses must be prefixed for a prefixed conclusion to follow. This lesson, as applied to law, leads to Legal Closure*.

3. If the legal perspective is analogous to a novel, then Legal Closure* is clearly a well-motivated way to close the legal perspective. However, many treat the legal perspective as a moral perspective (e.g. Raz (1990: 175–77), Coleman (2007: 596–97), Shapiro (2011: 279–80) and Gardner (2012: 133)). Those who do may be inclined to reject Legal Closure*. To see why, consider the following moral argument:

(6a) According to utilitarianism, one should φ if and only if φing would maximize utility.

(6b) φing would maximize utility.

(6c) So according to utilitarianism, one should φ.

This is familiar. We regularly seem to reason from a general normative claim of what is true according to a moral theory, and some fact about the world, to specific normative claims that are true according to that moral theory. One might suppose any closure principle applicable to moral perspectives must preserve the validity of this argument. Yet a moral version of Legal Closure* renders the argument invalid.

(Moral Closure*) If according to Moral-Theory-β [Q therefore P], and according to β [Q], then according to β [P].

Let β be utilitarianism, P that one should φ, and Q that φing would maximize utility. Moral Closure* does not permit inferring (6c) from (6a) and (6b). While [Q therefore P] is true according to utilitarianism, that φing would maximize utility is not prefixed. Without the prefix, however, the inference is not licensed under Moral Closure*. We have now arrived at the objection:

(7a) From (6a) and (6b) one can infer (6c); the argument is valid.

(7b) If Moral Closure* is correct, (6c) cannot be inferred from (6a) and (6b), rendering the argument invalid.

(7c) So Moral Closure* is incorrect.

(7d) Because the legal perspective is a moral perspective, if Moral Closure* is incorrect, then Legal Closure* is incorrect.

(7e) So Legal Closure* is incorrect.

We first take aim at (7a). Then we address (7d). Because they are either false or unsupported, this objection fails.

Here is an initial worry concerning (7a). Denying the inference to (6c) is putatively costly because it deprives us of a common form of moral argumentation. But how common is this argument form?

In ordinary discourse concerning what is right or wrong, we should not understand our claims to be ‘prefixed’. This is because, if prefixed, our ordinary moral claims could turn out true even if we were morally mistaken. But on those occasions we would typically take ourselves to have uttered a falsehood.

The academic setting is a possible exception. There, we sometimes see Kantians consider what utilitarianism would be committed to. Such claims might be understood as prefixed when they are embedded in arguments. But those arguments need not have the form of (6a)–(6c). Rather:

(8a) According to utilitarianism, one should φ if and only if φing would maximize utility.

(8b) φing would maximize utility.

(8c) If the claim made by utilitarianism in (8a) is true, then one should φ.

Suppose the Kantian wishes to leverage their conclusion (8c) by way of objection. Utilitarianism is false because we should not actually φ. Either (6c) or (8c) allows the Kantian to make this manoeuvre. What this shows is that, in this context, (6a)–(6c) and (8a)–(8c) are functionally equivalent. Given this, it is not costly to deny (7a). Common moral arguments can proceed even if Moral Closure* renders (6a)–(6c) invalid. For it preserves the validity of (8a)–(8c).

Here is a second worry for (7a). While Moral Closure* does not permit inferring (6c) from (6a) and (6b) as written, it does permit the inference if (6b) is replaced with:

(6b*) According to utilitarianism, φing would maximize utility.

As Lewis (1983: 262) explains, we often drop the prefixed operator ‘by way of abbreviation leaving us with what sounds like the original sentence … but differs from it in sense’. Once the prefixed operator is made explicit, Moral Closure* permits the inference. The argument is now valid. This further undermines support for (7a). The intuitive validity of inferring (6c) from (6a) and (6b) could rest on the tendency to implicitly read (6b) as (6b*).

Those inclined to object could raise the following point. Replacing (6b) with (6b*) permits (6c) to be inferred if (6b*) is true. But perhaps (6b*) cannot or need not be true. This is because in (6b*) utilitarianism makes a descriptive claim and moral perspectives cannot, or do not, make descriptive claims.

However, if utilitarianism cannot make descriptive claims, there would also be good reason to think utilitarianism cannot make specific normative claims. ‘P should φ’ is a specific normative claim. How utilitarianism could endorse that claim without being committed to the truth of a descriptive claim is unclear. Plausibly, to say ‘Jim should φ’ is to commit to the existence of Jim.⁴ It would be highly irregular to say ‘it is true that Jim should φ, but I lack a belief, one way or another, as to whether Jim exists’. And so, for it to be true according to utilitarianism that Jim should φ, it must also be true according to utilitarianism that ‘there is a Jim’. This raises a dilemma. Either utilitarianism can make descriptive claims (allowing for specific normative claims) or it cannot make specific normative claims (because it cannot make descriptive claims). On the first horn, (6b*) is licit. Then the argument is valid. On the second horn, (6c) must be false. Then the argument fails irrespective of Moral Closure*.

Perhaps you are sceptical of our efforts to undermine (7a). Nonetheless, the objection fails for a different reason. Recall the line of argument. It starts with the thought that law – and therefore the legal perspective – is a kind of moral perspective. Then it holds that it is impossible for moral perspectives to make descriptive claims. This is meant to be a problem for our position because Legal Closure* requires the law to make such descriptive claims. So far, we have sought to show why (7a) should be rejected. But there is another problem with this line of argument. Once we assume moral perspectives cannot make descriptive claims, (7d) becomes exposed to a serious vulnerability. If law can make descriptive claims, there is a clear disanalogy between the legal and moral perspective. Law can make descriptive claims while morality cannot. The truth of Legal Closure* would then be unsurprising even if Moral Closure* were false.

And law can surely make descriptive claims.⁵ We know this because what the law says about descriptive facts often diverges from what is actually true. Legal fictions are a good example. Sometimes the law contains irrebuttable factual presumptions. On those occasions the law presumes the truth of certain facts despite overwhelming evidence to the contrary. We could make the same point, however, with more prosaic instances of judicial error. Judges regularly make findings of fact. Those findings can be mistaken. Such a mistake, in turn, may lead judges to recognize legal rights which did not (prior to their ruling) exist. Suppose the law were to confer a right to damages to those carelessly injured by others. The judge falsely finds the plaintiff to be injured. This leads the judge to enter judgement in favour of the plaintiff. Doing so was a mistake, but one traceable to an erroneous finding of fact.

If legal perspectives can make both moral and descriptive claims, while moral perspectives can make only moral claims, then we would have good reason to reject (7d) even if (7c) were true.⁶ Hence the objection is stuck. Either Legal Closure* works even if law is a moral perspective (because moral perspectives can make descriptive claims) or the law takes the perspective of something other than morality. Regardless, Legal Closure* prevails and Perry’s objection thus fails.

4. Perry makes the following two-step objection to perspectivalism: (i) to account for mixed arguments, the legal perspective must be closed under implication, but (ii) doing so commits the perspectivalist to saying far too many things are true according to law. We agree with (i). But here we have shown why (ii) – and therefore Perry’s objection – fails. Each premiss of an argument must be true from a perspective if a conclusion from that perspective is to follow. This allows the legal perspective to be closed without too many things being true according to law.

At the same time, the soundness of Legal Closure* remains an open question. There may yet be reasons, which we have not addressed, to reject it. Those who go down this road, however, will need to reject perspectivalism. This is an important lesson we can draw from Perry’s argument, even if his core objection fails. We can now better see the terms under which a perspectivalist approach to law must proceed.

To be clear, none of our arguments are meant as a full-throated defence of perspectivalism. For our part, one of us rejects it while the other finds it a promising way to cash out the content of the law. Which of us is correct is a question for another occasion.⁷

References

Footnotes