Law and reason: beyond impartiality

Law and Reason: Beyond Impartiality

I     Legislation

I have tried to demonstrate that both Dworkin and Hart were devising theories of law seeking to portray proper legal interpretation as necessarily impartial. The following conception of law has no such ambition; in fact, it presents a picture of legal interpretation where some legal questions must be answered by the use of preferences. Partiality, according to this understanding, is sometimes unavoidable in perfectly valid legal determinations. But the thrust of this conception is not critical: it is an exploration of the constraints which guide and constrict legal interpretation – even though these constraints do not eliminate partiality. Nevertheless, the concern over partiality remains, of course, looming above: to speak of the constraints of legal interpretation is to speak of the curbing of discretion and its potential for partiality; and the degree to which these constraints preclude or permit partial decision-making remains an important measure in evaluating their significance.

The Principle of Public Reason

I would like to go back to Dworkin’s puzzling question regarding checkerboard statutes:

We all feel, of course, that these suggestions are unacceptable; the question is why, and whether this has any significance to legal practice. As we saw, Dworkin believes the explanation lies in our demand that all legal requirements derive from one ‘scheme of justice’ or one ‘set of moral principles’.2 Having seen that this proposal gets nowhere, we should reopen the question of what distinguishes checkerboard statutes (for example, a statute allowing abortions to women born in odd but not even years) from noncheckerboard statutes (for example, a statute making abortions criminal for pregnant women but not for women impregnated by rape); for with his checkerboard statutes Dworkin has touched on a point that is crucial for the understanding of modern law.

Now as we saw above, it seems that the essential difference between the two statutes lies in the fact that in the checkerboard statute (the one permitting abortions to women born in odd years but not to women born in even ones) we do not see any reason for treating these two classes differently, while in the case of the regulation prohibiting abortions to all women but those impregnated by rape we do recognise reasons as to why these two classes may be treated differently (the total lack of responsibility for the pregnancy, the degree to which the newborn is unwanted, and so on). We do not have to endorse this reason in order to distinguish this statute from the checkerboard statute: we may think, for instance, that rape does not justify an abortion. Still, we see that this statute has some reason for the distinction it makes. (I shall return to this important point below.) Thus it appears that we expect to identify a reason as to why a statute treats two seemingly similar categories differently, and with checkerboard statutes there seems to be no such reason (no reason as to why to treat differently women born in odd years and women born in even years in regulating abortions, no reason for treating manufacturers of washing machines differently from car manufacturers in product liability regulation, no reason for treating discrimination on buses differently from discrimination in restaurants). Checkerboard statutes prescribe one legal requirement to one class and a second legal requirement to another while the difference between the two classes does not explain that difference in treatment. This expectation is commonly expressed in the idea of ‘equality before the law’: the similarly situated are to be treated similarly. This is not merely a formal adjudicative principle (where some positive law defines the similarly situated); it is a substantive legislative principle as well. A criminal statute which applies to those whose last names begin with the letters A to N but not to those whose last names begin with O to Z would violate our idea of equality before the law; for we can see no reason as to why an act is criminal if committed by those whose names begin with the first letters of the alphabet, but legal if committed by those whose names begin with the last. We expect statutes to have a publicly recognisable reason for the distinctions that they make.3

The relation between the category and the treatment

In raising the question of checkerboard statutes Dworkin in fact hypothesises a reason for the distinctions these statutes make: these distinctions are the result of legislative compromises – of the wish to accommodate seemingly valid but conflicting opinions (for instance, one opinion calling for the freedom to abort and another calling for its prohibition). Now why can’t this explanation furnish the required reason for the different treatments accorded by these statutes? It cannot because the different treatments it is supposed to explain have nothing to do with the features of the distinguished categories: we expect that any difference in treatment be a function of the dissimilarity between the differently treated classes (or ‘have something to do with’ the dissimilarity, or ‘be related to’ it, or ‘be connected with’ it, or any such number of formulations). When a statute forbids abortion to all women but those impregnated by rape, the prescribed difference in treatment has something to do with the distinction between being impregnated so to speak ‘normally’ and being impregnated by rape. Similarly, when the prohibition of abortion is held applicable to women born in odd but not in even years, the prescribed difference in treatment must have something to do with the difference between being born in an odd year and being born in an even one. We expect to discern a justification as to what made the legislature draw the specific distinctions it did. Dworkin’s hypothesised reason does not explain why these specific categories (women born in even years and women born in odd years) are treated the way that they are; in fact, under Dworkin’s hypothesis there is no such reason – the two categories are, presumably, arbitrarily chosen. If you ask me to explain why I treat white chickens more kindly then brown chickens and I respond ‘because I have mixed feelings about chickens’ I haven’t explained the very distinction I make between chickens that are white and chickens that are brown. The situation here is similar: what we seek a reason for is why the statute calls for a treatment of this particular category that is different from the treatment of that particular one. The response that ‘this is a compromise between pro- and anti-abortion legislators’ supplies no such reason.

Why do we demand that the reason for a statutory distinction be derived from the characteristics of the distinguished classes? This is a fundamental limitation on power that is imposed by the form of a legal order; and it is aimed both at guaranteeing the rationality of legal regulations and at making the reasons for legal regulations publicly available: a legal system is essentially a system of regulation grounded in publicly recognised reasons.

Arbitrary Distinctions?

Now these claims could be proved wrong if there were perfectly acceptable laws having arbitrary distinctions. Dworkin, whose understanding of checkerboard statutes is different from the one proposed here (but for whom arbitrary distinctions would have also proved a problem), concedes that some laws do draw arbitrary distinctions – ‘we do accept arbitrary distinctions about some matters: zoning, for example’ – but he adds that arbitrary distinctions are not acceptable where issues of justice are concerned.4 But I think there is little reason for this concession: allowing that many laws exhibit some inevitable measure of arbitrariness because of their line-drawing (for example, when the driving age is set at 17 rather than at 17 and one month), zoning ordinances exhibit just such arbitrariness and not more: given the need to draw some line between one geographic area and another, the arbitrariness inherent in zoning schemes is no greater than the one in a law setting a minimum age for driving. To put things in the terms I employed above, the different treatment accorded to different zones is not unrelated to the dissimilarities between these zones. Zoning is a well planned activity taking into account the various differences between different geographical areas (the number of residents, original character of the zone, its centrality, and so on), and it is simply wrong to think that no justification, having to do with the difference between the zones, can explain the different treatments accorded to them. We would certainly consider unacceptable any zoning scheme that delineated its sectors arbitrarily – one dividing urban space by drawing random geometrical patterns, for example.

Take another statute drawing a seemingly arbitrary distinction: a law aimed at reducing pollution and traffic by allowing vehicles with odd registration numbers to be operated only Sunday through Wednesday, and those with even numbers to operate Wednesday through Saturday. Isn’t this an arbitrary (though perfectly sensible) statutory line-drawing? The answer, once again, is that it is not: an arbitrary distinction carves up two categories whose difference cannot account for the different treatments they receive; but the distinction between odd and even registration numbers does account for the difference in treatment. The distinction between cars with odd and even registration numbers is a distinction between two groups having the same number of cars, having a roughly similar geographic distribution, where the ratio of family cars or trucks is roughly the same, and so on. This distinction is a proxy for two well defined (and presumably similar) categories; and the treatment prescribed for these two essentially similar pools is also essentially similar (the difference merely being the days of the week in which they may operate). The distinction between even and odd registration numbers is, in this respect, not an arbitrary distinction: it is a distinction which captures two well defined categories of vehicles. If it turned out that it did not – for example, if it emerged that all family cars have odd registration numbers, or that all trucks have even ones – then that statute would be as objectionable as any other checkerboard statute. There is no good reason to concede that the law draws arbitrary distinctions: it does not.5

Recharacterising the Problem with Checkerboard Statutes

Now the problem with checkerboard statutes can be better perceived if we stop talking about ‘equal treatment’ and start talking about the statutory requirement itself. Here is what I mean: the problem with Dworkin’s checkerboard abortion statute is not merely the lack of a reason for treating women born in even and those born in odd years unequally; it is the lack of a reason for prohibiting abortions to the category of women born in even years. We seek a reason, related to the features of the category singled out by the statute, as to why that category is treated as the statute demands.6 We may have a reason as to why to prohibit abortion to all women; but we have no reason as to why to prohibit it to women born in even years: we seek a reason as to why the specifically defined operative category is treated the way that it is. The fact that there may be a reason for mandating that all stores be closed on Sundays does not mean that there is a reason for mandating that all shoe stores be closed. If our operative category is ‘males 18–21 years of age with a physical disability’ then the reason we seek is not only as to why treat males the way the statute requires, nor only why treat males ‘18–21 years of age’ that way; we also want a reason for specifying the physical disability.7

In short, one need not look upon the problem of checkerboard statutes as the lack of equal treatment for two similarly situated groups: we are offended by checkerboard statutes because we identify no recognisable reason, having to do with the features of these statutes’ operative category (women born in even years, manufacturers of washing machines, and so on), as to why the state treats that category the way the statute demands. The problem is one of explanation – not, strictly speaking, of equality.8

Unreasonable Reasons

Now what counts as a proper explanation – that is to say, as a ‘recognisable reason’? As I noted above, and as Dworkin also recognises, we need not endorse a reason in order that it satisfy our demand: we may recognise a reason as sufficient even if we do not accept it as a good one. (Dworkin says that a law prohibiting abortions which exempts pregnancies caused by rape would not be regarded as a checkerboard statute even by those who think that rape does not justify an abortion.9) The perspective from which one judges whether a statute is or is not supported by a public reason is not restricted to one’s own understanding, but encompasses opinions that may very well contradict one’s own.10 This perspective, however, has its obvious limits: if we are looking for reasons, then particularly bad reasons (or ‘unreasonable reasons’, if I may use this awkward but adequate expression) cannot do.11 We recognise some opinions incompatible with our own as valid opinions (saying to ourselves: ‘well, this is certainly not my opinion on the matter, but I can accept it as a legitimate one’); but there are opinions which simply do not appear to us valid.

For example, many people think that there is a relation between the position of the planets at the time of one’s birth and one’s psychological makeup. This opinion may be popular, but it clashes with some pretty solid ideas many of us have, and a statute based upon it will strike many of us as lacking a reasonable explanation (imagine a statute granting certain employment preferences for those born between 11 March and 13 May). A statute confining Jews or Blacks to certain geographical areas similarly lacks a reasonable explanation for its treatment of its operative category – although some nasty characters are sure to recognise one. Such opinions are thought to conflict with some of our more solid moral beliefs, and they can no longer be accepted as ‘reasons’. Thus, what at one point may have appeared a perfectly acceptable statute would have been transformed, by changes in moral attitudes, into a checkerboard statute – an offence to our legal sensibilities and our understanding of what can and what cannot constitute a law. In short, we see some positions as too unmistakably false, or as conflicting with ideas we think too well defended, and we will not accept these as sufficient explanations.

Now whether an opinion is or is not reasonable is often a highly controversial determination; and the crucial question is what makes an opinion reasonable or unreasonable. It seems to me that the answer to this question is not encouraging for legal (and political) theory: the distinction between reasonable and unreasonable opinions may hinge on nothing more fundamental than whether an opinion is held among a certain elite. If one grows up in a society where ‘pro-life’ opinions are common among prominent or respected people, then one may properly conclude that this is a reasonable (if wrong) opinion. This may be all that there is to this reasonable/unreasonable business: the unrelated reasonableness of the people holding the opinion. Yet this distinction between reasonable and unreasonable opinions is of crucial importance: not only does it often influence our private actions, it is a distinction which marks decisive public lines in our increasingly pluralistic societies – and, as we shall see, it is a distinction upon which legal determinations are often based.

This important observation, and its implications for the difficulty of impartiality, will be discussed below. For now it is sufficient to realise that we expect all statutory requirements to have recognisable reasons, having to do with the features of their operative category, as to why they treat that category as they do; and that these reasons need be, well, reasonable. This means that a strict-liability statute applying to cars but not to washing machines, if legislated, would be legislated on the premise that there is a reason having to do with the difference between cars and washing machines which explains this difference in treatment. Pointing to this fact may not provide us with an uncontroversial criterion for evaluating whether such a statute is or is not acceptable: we may very well disagree about the reasonableness of that reason. But it will begin to show us how we perceive and shape our laws and our legal practice.

Is All this Significant?

Now it may be objected that all this is perfectly trivial: of course we expect our laws to be supported by reasonable reasons! But so what? We also expect them to be just, and intelligent, and efficient, and what have you; but all this is completely obvious and also completely uninteresting. There is nothing in these expectations to tell us anything insightful about the law.

This declaration misses much of what I said above: we do not merely expect statutes to be supported by reasonable reasons, we also expect these reasons to be publicly available, and to demonstrate a relation between the features of the operative category and the treatment to which a category is subjected. The attempt to analogise these expectations to the expectation that statutes be just or intelligent cannot work: we want any rule of conduct to be just or intelligent; but we want it to have a public reason related to the features of its operative category only when it appears in the form of a law. We do not think, for example, that the rules laid down by parents to their children (‘never open this drawer’), or the rules army commanders lay down to their soldiers (‘only those whose last names begin with the letters A to N get sniper training’), or the rules that religions lay down to their faithful (‘you shall not shave your beard’), need have publicly available reasons having to do with their operative categories. We do, however, think this in the case of laws. The young Iranian who talks to VS Naipaul in Beyond Belief is expressing sentiments different from those of a religious devout, as he himself acknowledges, when he complains about a religious rule, saying: ‘To me, the rules about beards have no logic. They don’t say why. They just say “Do it”.’12

In other words, we are not dealing with some general virtue (justice, intelligence, efficiency) with contingent links to our legal regulations: we are dealing with a principle that pertains specifically to our laws. Perfectly worthy regulations may be unacceptable as legal regulations because of this principle. This is what Dworkin’s hypothesised checkerboard statutes show: there is nothing wrong with a rule imposing strict liability on cars but not on washing machines (people cautious about the impact of strict liability may reasonably conclude that this is a good option) – and yet this is an unacceptable statute. Here is another example: consider a regulation rewarding war veterans by granting them an exemption from parking regulations. There seems to be nothing immoral or unreasonable about this regulation (legislators and foreign diplomats have such an exemption). Yet any statute rewarding war veterans must do so in a way that is related to their status: if there is no relation between being a war veteran and an exemption from parking regulations, then such a regulation would be unacceptable as a statute.

Now a ruler who is unconstrained by the principles of a legal system may not care for such a demand: Saddam Hussein may nonchalantly have decided to reward his war veterans by allowing them to park anywhere they felt like parking. But our legislators operate differently: they legislate laws with public justifications having something to do with the features of their operative categories. So here is one insight we can derive from our principle of public reason: the process of designing laws is guided by principles which impose limitations on the content of our laws. The institution of a legal system places substantive limits on the exercise of power. The law’s limitation on legislative power is not restricted to the frequently repeated idea that legislators are bound by their own laws: the law contains limitations that limit the regulative options themselves.

There are more practical insights to be gained here – insights about the proper standards of legal interpretation. Understanding that legislators craft laws with the principle of public reason in mind (whether consciously or not) has important implications to our understanding of statutes’ purposes – itself often a decisive factor in their application. The section entitled ‘Legislative intention’ will deal with this issue. Furthermore, as we shall also see below, legal arguments and legal determinations also preserve this principle of public reason; and this means that understanding this principle may allow us to better comprehend what makes a convincing legal argument or a convincing legal determination. But before addressing all these important points, I would like to draw some parallels between what has been said so far and the work of Lon Fuller.

The Knowability of the Law

It is a matter of common understanding that one of the requirements of the rule of law is that the law be known. But the claims advanced in the preceding pages allow us to understand this requirement in a new light: the knowability we require of the law is not merely that its subjects be advised of the standards of conduct required of them (that the law be ‘promulgated’), nor even that these standards be clear and intelligible; it is that they also know the justification for the treatment prescribed by the law. This gives a new meaning to the opposition between the rule of law and arbitrary government: this opposition runs deeper than the idea that arbitrary government does not work by way of publicly available standards. It is an opposition between standards of conduct that are themselves arbitrary and standards supported by recognised public justifications.

One legal scholar who touched on the idea of a public reason requirement was Lon Fuller. In his famous exploration of the ‘inner morality’ of the law, Fuller lists eight attributes we demand of our legal norms: we demand that they be general, promulgated, not retroactive, clear, non-contradictory, not requiring the impossible, that they be constant over time, and that there be a congruence between official action and declared rule.13 This last requirement is described, in Fuller’s little tale about a king who violates these requirements one after the other, as the lack of any ‘discernible relation between [the king’s judicial] judgments and the code they purported to apply’.14 This depicts the requirement as one of congruence between legislative command and adjudicative execution; and it therefore obligates Fuller to get into a discussion about proper legal interpretation (that is, about what adjudicative execution is about). When he does, the idea of public reason is shown hidden in his very assumptions.

Fuller says that the ‘best short answer’ to the question of the principles of legal interpretation was given in 1584 by the Barons of the Exchequer when they met to consider the difficult case of one Heydon:

It was resolved by them, that for the sure and true interpretation of all statutes in general (be they penal or beneficial, restrictive or enlarging of the common law) four things are to be discerned and considered:

1st What was the common law before the making of the Act.

2nd What was the mischief and defect for which the common law did not provide.15

3rd What remedy the Parliament hath resolved and appointed to cure the disease of the common wealth.

4th The true reason of the remedy; and then the office of all the judges is always to make such construction as shall suppress the mischief, and advance the remedy.16

Fuller then adds that ‘the central truth of the Resolution in Heydon’s case [is] that to understand a law you must understand “the disease of the common wealth” it was appointed to cure’.17 He later comments: ‘Some obscurity concerning the mischief sought to be remedied by a statute can be tolerated. But if this obscurity exceeds a certain crucial point, then no virtuosity in draftsmanship nor skill in interpretation can make a meaningful thing of a statute afflicted with it.’18

That ‘mischief’ is none other than the public reason of which we spoke. In elaborating on his idea Fuller makes use of an analogy where an inventor of household devices dies, leaving behind him an unfinished sketch of an invention. His son is bequeathed the task of completing that work. Fuller then argues that the son must discover what Fuller calls ‘the intention of the design’: the son ‘would look to the diagram itself to see what purpose was to be served by the invention and what general principle or principles underlie the projected design’.19 To explicate the analogy, the legal interpreter, looking at a statute, identifies the principles underlying its design and then uses these principles in determining what that statute requires. The reason underlying a law – the ‘intention of its design’ – must be in principle discernible from looking at it; and where it is not discernible, that statute is, as far as the law is concerned, hardly a ‘meaningful thing’.20

II     Legal Interpretation

Once we note that all statutes possess publicly available reasons for their treatment of their operative category, it naturally follows that these reasons play a role in our interpretation of these statutes. (After all, don’t we demand of statutes that they have public reasons for the very fact that their application vis-à-vis particular cases be publicly reasoned?) That role is easy to see in cases where the conditions stipulated for the operation of a statute are present, and yet the statute would be held inapplicable because its public reason fails in the case at hand. This happens, for example, when the features associated or correlated with a class, and for which the class is treated as the statute demands, are absent from an instance of that class. In Fuller’s famous example, a monument consisting of a World War II truck is to be placed in a park where a legal rule prohibits the entry of vehicles.21 Here the statute would not prohibit the entry of the truck because its public reason is simply inapplicable to the case. A similar though more difficult case was presented before the US Supreme Court in Kaiser Aluminum & Chemical Corporation v Weber, a case we encountered in Chapter 2.22 The Court was called to decide whether a private corporation implementing an affirmative action policy where black employees received priority for training was in violation of Title VII of the Civil Rights Act which prohibits discrimination ‘because of race’.23 A white employee denied a place in the training program sued his employer, but the court held that these policies did not violate the statute. Underlying this decision was the determination that certain features associated with the practices prohibited by the statute, and for which these practices are prohibited, were lacking in the case at hand: the case, in other words, did not fit the statute’s public reason.

The case of the ambulance entering the park on an emergency call presents another category of cases where the public reason of the statute obviously dictates its application: here the public reason is