What Makes Law
AN INTRODUCTION TO THE PHILOSOPHY OF LAW
Liam Murphy
New York University
7
2 Morality and the Grounds of Law
Adjudication and the Grounds of Law
The main traditional dispute about the nature of law, and the one that is
the main focus of this book, is a dispute about what morality has to do
with iguring out the content of the law in force in any particular place.
On the one hand we have matters of fact, such as the meaning of what
is written in some document that is issued by some legal institution such
as a legislature. On the other hand we have moral considerations, in the
broad sense that includes not just individual right and wrong but also normative
political theory (about, say, social justice or the proper limits of
state power). Everyone thinks that matters of fact are relevant to determining
the content of current law. The main dispute about the nature of
law is a dispute about whether moral considerations are also relevant.
This is a dispute about what Dworkin (1986, 4) called the grounds of
law – of what makes legal propositions true.
This dispute is obviously distinct from the question of whether moral
judgment does or should inluence those who make law. We can take for
granted that it should and to some extent does. The question is whether
moral considerations are relevant to iguring out what the law already is.
What may not be so obvious is that the issue of the grounds of law is also
distinct from that of whether judges should appeal to moral considerations
when adjudicating disputes.Explaining this contrast is perhaps the best way
to bring the traditional debate about the grounds of law into focus.
WHAT MAKES LAW8
Any government oficial whose role it is to determine the legal rights
and duties of others requires a theory of legal decision making. In the
case of judges, we call this a theory of adjudication. Oficials from other
branches of government also make decisions about the legal rights and
duties of others, which will be important in later chapters, but judges provide
the central case and the natural place to start.
A judge’s theory of adjudication may be sketchy and perhaps only
implicitly believed, but she must have one. Decisions about people’s legal
situations obviously cannot be made without having views about which
considerations it is appropriate to take into account. It is not essential, by
contrast, that judges have a theory of the grounds of law – a theory that
would tell them whether and when moral considerations are relevant to
the question of what the law (already) is.
A record of the reasoning behind a judicial decision may cite many
factors that were thought to be relevant, such as the contents of constitutions,
statutes, prior judicial opinions, the prevailing custom in a
particular industry or place, the opinions of legal scholars, and considerations
of social welfare, justice, and fairness. A judicial opinion (in
common-law jurisdictions, at any rate) explains how and why a decision
was reached, and therefore tells us a lot about the judge’s theory of
adjudication, but it will not necessarily reveal her views about which
of the factors on which the decision was based were part of already
existing law.
If judges were called on always to announce the state of the prior law
as they found it,they would need to reveal their views,for example,about
whether considerations of fairness are part of the law or rather fall into
the category of considerations not part of law that are nonetheless legitimately
taken into account in adjudication. Such categorization requires a
theory about which kinds of factors are, in principle, relevant to iguring
out the content of the law; it requires a theory of the grounds of law. It is
typically not necessary for judges to engage in such categorization, and
so it is typically not necessary for judges to have a theory of the grounds
of law.
MORALITY AND THE GROUNDS OF LAW 9
This holds true even when common-law courts explicitly “overrule”
or “decline to follow” a precedent. Such a statement leaves open whether
the discredited precedent was, in the view of the court, formerly part
of the law, which is now changed, or instead a mistake that was never
part of the law properly understood. Both views are found in traditional
common-law thinking, but judges working within the common law need
not take a stand.
Some judges and some legal theorists believe that all normative considerations
that judges are authorized to take into account when deciding
a case are necessarily part of the existing law. We can call this the adjudicatory
or adjudicative (Perry 1987) view of law. The implication of the
adjudicatory view is that there is no interesting gap between determining
what the law is and marshaling the considerations relevant to resolving a
particular dispute before a court.1
If the adjudicatory view of law is correct,
then it is misleading to say that it is not necessary for judges to have
a theory of the grounds of law. But the adjudicatory view of law may or
may not be correct, and it is not necessary for judges to have a view about
whether it is.
The majority and the dissent in the nineteenth-century New York
case Riggs v. Palmer2
disagreed about statutory interpretation. Francis
Palmer’s will,formally valid under the relevant statutes in NewYork State,
left his estate to the person who murdered him. The majority argued for
the relevance of the fact that the legislature, had it ever considered such
a case, would never have intended to allow a murdering heir to inherit. It
also argued that statutes should be interpreted in light of “fundamental
1
There will still be a gap, on any plausible view: mathematics and logic can help determine
the outcome of a legal dispute, but no one believes that they are part of the law in
force.The terms of a contract will in part determine the outcome of a contract dispute.
It seems natural not to regard those terms as part of existing law, but some may prefer
to talk that way – to treat the contract as private legislation between the parties. But
this is not an interesting disagreement.What matters is whether the normative considerations
of fairness, justice, and the rest are, if legitimately appealed to in adjudication,
therefore part of the law in force.
2
115 N.Y. 506 (1889).
WHAT MAKES LAW10
maxims of the common law.”The dissenting judges argued for a straightforward
application of the literal meaning of the words of the relevant
statutes. Dworkin characterizes this dispute as one “about what the law
was, about what the real statute the legislators enacted really said” (1986,
20). But while the judges clearly disagreed about proper adjudication, we
simply do not know whether they all embraced the theory of the grounds
of law according to which all normative factors legitimately taken into
account in adjudication are at the same time relevant to determining the
content of the existing law.They did not say.They did not take a stand on
the nature of law because it was not necessary.
It is sometimes claimed that all, or almost all, judges hold the adjudicatory
view of law. If this were true, it would provide some support
for it. But the evidence does not support the claim. In the United States,
prominent and scholarly judges who have addressed the issue – from
Oliver Wendell Holmes Jr. to Learned Hand to Richard Posner – have
thought it obvious that judges must, on occasion and under constraint,
“legislate.”3
Benjamin Cardozo (1921) elaborated a theory for such legislative
adjudication – using what he called the “method of sociology,”
a judge should ill gaps in the existing law with recourse to community
morality. Most judges, however, neither announce their theory of law in
their opinions nor write books or articles about the judicial process, so it
is hard to know what theory of law they hold.
It is true that in recent times it has become typical for judges and aspiring
judges in the United States publicly to disavow “legislating from the
bench.” But many of these same judges would equally adamantly deny
that applying the law ever requires them to have recourse to “their own”
moral or political views.As it would precisely be inclusion of moral judgment
within the bounds of law that would give existing law the resources
3
“I recognize without hesitation that judges do and must legislate, but they can do so
only interstitially;they are conined from molar to molecular motions.”Southern Paciic
Co. v. Jensen, 244 U.S. 205, 221 (1917, Holmes, J., dissenting). Hand (1952) takes a similar
view, as does Posner (2008, 81–92).
MORALITY AND THE GROUNDS OF LAW 11
to generate an answer to all questions before a court, this makes it a little
mysterious just what theory of law these judges hold.4
Of course there is this truth in what most judges say: whatever else is
controversial in the theory of adjudication, judges ought to apply the law.
Is the upshot that even if applying the law is not all that judges must do,
they nevertheless will need a theory of law – that this is in fact the primary
thing that they do need (Dworkin 2006, 18–21)?
It is true that one way to develop a theory of adjudication is to
start with a theory of the grounds of law and only later turn to the
issues of when judges may depart from prior law (if the theory of the
grounds of law makes that ever seem necessary) and of how decisions
should be made when valid law does not settle the issue (if the theory
of the grounds of law makes that a possibility). But it is also possible
to start from the immediate practical question of how to decide a
4
It is a commonplace of jurisprudence – that no theorist or judge could really disagree
with – that a literal reading of legal sources will not in itself yield answers to all legal
questions.Though some legal systems or areas of law are more or less determinate than
others, there will always be cases where a reasonable reading of the materials could go
either way. There may or may not be a method of adjudication, such as “originalism”
or Cardozo’s method of sociology, that could allow a judge to supplement the guidance
of legal sources without reference to her own moral or political judgment. (That
seems dubious, but it is not our topic here.) But even if one were available in principle,
there is no consensus (in most jurisdictions) that any such method is mandated
by existing legal sources, read plain. It is quite appropriate, then, that Justice Antonin
Scalia of the U.S. Supreme Court defends originalism on political grounds (Scalia
1998), just as Cardozo before him defended the method of sociology. Both hold that
judges should appeal to other material when they need more guidance than the legal
sources can give them; both hold that this other material should be something other
than a judge’s own moral convictions, something more objective. But neither holds
that his preferred method of adjudication is mandated by existing law. To be a good
judge, even on these approaches, you need to reach the right moral conclusion about
how to make decisions on your own; no legal authority is directing you in this matter.
In the face of its evident absurdity, the pervasive contemporary disavowal of the need
for moral reasoning on the part of judges, at least in the United States, is something
of a political pathology that is itself worthy of investigation. For an extended discussion,
see Kennedy (1998). The current point, however, is simply that in this political
environment we would be unwise to treat what judges say about the nature of law at
face value.
WHAT MAKES LAW12
case.5
And if we start from that end, there is no need, within the theory of
adjudication, to mark the boundary between law and not law.
There have been exceptional cases where courts have declared that a
decision about the content of prior law was necessary to resolve a legal
dispute. Shortly after the reuniication of Germany, two former East
German border guards were convicted of homicide for the shooting of a
person attempting to climb the Berlin Wall.The guards appealed, in part
on the ground that the shooting was in accordance with the law of the
former German Democratic Republic: the killing was therefore justiied
by law in force at the time, and so the homicide convictions must be seen
as retroactive criminalization that violated the constitution of the now
uniied Federal Republic of Germany. The appellate court, the Criminal
Division of the German Federal Court of Justice, agreed that the convictions
would have been unconstitutional if the killings had been in accordance
with valid East German law. But there was no retroactivity, the
court held, because the legislation of the German Democratic Republic
that regulated the policing of the border was,in virtue of its grossly unjust
content (authorizing guards to shoot at people leeing over the border),
no law at all.6
In this case the court made a statement about the grounds of law. It
was moved to do so presumably because it seemed that there was no
other way it could uphold the convictions in the face of the potentially
justifying legislation. In support of its declaration that grossly unjust law
was not law at all, it cited a 1946 article by Gustav Radbruch (2006) published
in a German law journal. There was, however, a less adventurous
route to the same outcome. Reviewing a similar set of convictions, the
German Constitutional Court, while not impugning the constitutionality
of the approach of the Federal Court of Justice, showed how the
5
Here is how Posner sees it: “A judge does not reach a point in a dificult case at which
he says,‘The law has run out and now I must do some legislating.’ He knows that he has
to decide and that whatever he does decide will (within the broadest of limits) be law;
for the judge as occasional legislator is still a judge” (2008, 85).
6
BGHSt 39, 1 (1992).
MORALITY AND THE GROUNDS OF LAW 13
convictions could be upheld without taking a stand on the grounds of
law.7
The Constitutional Court upheld homicide convictions stemming
from shootings at the wall on the basis of a creative interpretation of
the constitutional ban on retroactivity. Article 103(2) of the German
Constitution holds that “An act may be punished only if it was deined
by a law as a criminal offense before the act was committed.”8
This, the
court held, did not apply to a situation where an otherwise criminal act
was justiied by a grossly unjust law.The court justiied this interpretation
with an argument about the point of having a ban on retroactive punishment
in a democratic society.Approaching the issue in this way, the court
avoided having to pronounce on the validity of the legislation governing
the activities of the border guards at the time of the shooting and so
avoided having to take a stand on whether morality was relevant to the
determination of the content of the law.
The German Constitutional Court in this case certainly engaged in
moral reasoning along the way to reaching its decision. Suppose we think
that this was entirely appropriate.That would indicate nothing about our
view on whether moral considerations are relevant to determining the
content of the law. One excellent way to characterize what legal philosophers
have been disagreeing about is precisely this question:When a conscientious
judge appropriately appeals to moral considerations to reach a
decision, has he in so doing gone beyond the mere application of existing
law and in part also made new law? One side says yes, its opponents no.
The dispute about the grounds of law is not about what judges should do,
but you could say it is about the correct description of what they do.9
Once this becomes clear, however, many start to wonder why anyone
ever thought this dispute mattered. If judges can do without a theory of
7
BVerfGE 95, 96.
8
Oficial translation available at: https://www.btg-bestellservice.de/pdf/80201000.pdf.
9
“Much of legal theory can be seen as an attempt better to understand what a court
takes from the law and what it gives to it, where the application of the law stops and
judicial discretion begins, what the boundary is between the law the court inds and the
law it creates” (Raz 1986, 1117).
WHAT MAKES LAW14
the grounds of law and that theory affects merely how we can describe
what judges do, why would anyone bother with it?
The Law in Force
The reason is that anyone who is attempting to answer a question about
the content of law must as an initial matter have a view about the grounds
of law.
Had a German law professor been asked, prior to the trials of the
former border guards, whether the intentional killings at the wall were
criminal homicides when they happened, she would not have been able
to avoid taking a stand on the grounds of law. If they were not criminal
homicides, the justifying East German law (which we can assume was
grossly unjust) was valid, and Radbruch’s account of the grounds of law
is rejected. Suppose that this is the view of our professor. We might next
ask her what she thinks about the Constitutional Court’s argument that
the convictions should stand even if the killings were not crimes at the
time. The professor may dodge the question by saying that it was a good
decision, that she “agrees” with it. That is multiply ambiguous. She may
mean, irst, that the court correctly applied German constitutional law as
it was prior to its decision. Or she may mean that the court, acting within
its legal authority and thus in accordance with law, reinterpreted and thus
changed, at least for this case, German constitutional law in a manner she
thought wise. Last, she may mean that, even though the court exceeded
its legal authority in acting as it did, this was, in the circumstances, the
morally right thing to do – taking into account, she might say, the importance
of some kind of public accountability for the crimes of the former
East German regime.After all, she might add, even though judges are, as
a matter of morality, duty bound to act in accordance with law, that duty
is not absolute, and sometimes other factors may outweigh it.
To keep things reasonably simple, we can leave aside the issue of
whether, and if so when, it is morally appropriate for a judge to make
decisions that are not in accordance with law. If we ask our professor
MORALITY AND THE GROUNDS OF LAW 15
to choose between the irst and second ways of understanding what she
said, we will force her, once again, to take on the issue of the grounds of
law.We know that she holds the justifying East German legislation valid.
If she believes that the convictions were nonetheless constitutional under
the law as it was before the court made its decision, she must believe that
the argument of political morality the Constitutional Court engaged in to
justify its decision falls within the domain of law. If the professor believes
that the convictions were not valid under prior German constitutional
law, but that the decision was nonetheless in accordance with law and
a good decision, then she believes that while the argument of political
morality did not fall within the boundary of law, it was nonetheless a
legitimate set of considerations for a judge acting according to law to
take into account, and that the court’s reasoning was sound.
Consider a very different kind of example. Whether the exclusion of
same-sex couples from the institution of marriage in New York was in
violation of the state constitution in, say, 1995 depends on the grounds
of law. Some people believe that the content of the law is determined
entirely by legal sources: statutes, constitutions, judicial opinions, and
so on, all interpreted in a fashion that never requires the independent
moral judgment of the interpreter. If we take this view, we will probably
conclude that the existing legal sources did not determinately settle the
legality of same-sex marriage in New York in 1995, and so the matter
remained open until the Court ofAppeals settled it,in the negative,in the
2006 case of Hernandez v. Robles.10
Subsequently, the Marriage Equality
Act of 2011 changed the law. Others believe that the content of the law
is determined by the best, the morally best, interpretation of the existing
legal materials. If we take this view and also believe that equal protection,
properly interpreted, implies equal participation, regardless of sex
or sexual orientation,in important social institutions such as marriage,we
will have little dificulty in reaching the conclusion that the exclusion of
same-sex couples from the institution of marriage in New York violated
10
7 N.Y.3d 338, 855 N.E.2d 1 (2006).
WHAT MAKES LAW16
the equal protection clause of the New York State Constitution in 1995.
We could also conclude that the holding to the contrary in Hernandez
was a mistake and that the legislation of 2011 recognizing same-sex marriage
was strictly speaking unnecessary.
On the one view there was no answer to the legal question before
Hernandez; after that decision, there was for ive years the clear answer
that same-sex marriage was not legally available in New York until the
New York State Legislature changed the law. On the other view, it has
for a long time been contrary to law to exclude same sex-couples from
the institution of marriage in New York. Different views of the grounds
of law, different conclusions about what the law is. And it does seem to
matter, to New York residents at any rate, what the law about marriage
in New York is.
Each of us needs to take a stance on the dispute about the grounds
of law if we want to be able to offer answers to legal questions. Far
from nothing turning on it, it looks as though, as far as the law is concerned,
everything turns on it. Of course, if it did not matter what the
law is, then the dispute about the grounds of law would not matter. This
“eliminativist” attitude to the law, radical though it is, must be taken seriously;
I discuss it in Chapter 6.
Law and Morality: A Brief History
Some insist that moral considerations are relevant to iguring out what
the law is and others that this is never so; each camp believes that the
other misunderstands the nature of law in a fundamental way. But that
neither camp is obviously mistaken has been recognized at least since
Aristotle.
In the Rhetoric, Aristotle distinguishes between the “written law,”
which is speciic to a particular place, and the unwritten “common law,”
which is the same everywhere and expresses what is just by nature
(1373b). He then advises those arguing a lawsuit before an Athenian jury
to take advantage of the fact that the law governing the dispute could be
MORALITY AND THE GROUNDS OF LAW 17
thought to be grounded either entirely in the written law, or alternatively
in the written law as corrected by the common law.Where the written law
is against you, you should argue from the common law, pointing out that
the written law does not always “perform the function of law.”Where the
written law is on your side, you should remind the jury that its job is not
to try to be “wiser than the laws” (1375a–1375b).The advice is to lean on
one or another way of understanding what kind of thing Athenian law is,
depending on which understanding better suits your case.
In language that has been familiar in the West since St. Thomas
Aquinas’s Treatise on Law,11
we could put Aristotle’s point this way.Take
for granted the fundamental distinction between the natural law, which is
a distinctive way of understanding what modern philosophers would call
universal morality, and positive or human law, which is the law that is in
force in a particular place.The question then arises of whether the human
law in a particular place is determined solely by its written or customary
legal sources or whether it is at least in part also determined by the natural
law. Aristotle sees that both possible answers to this question about
the nature of (human) law can be made to seem plausible.
Aquinas perhaps agreed; he did not take a clear stand on whether the
natural law is partly determinative of what the human law is,or instead just
an external basis for determining what it ought to be. At the start of the
Treatise on Law, Aquinas writes that all law (human law included) must
concern itself with“the happiness of the community.”Later,when writing
speciically about the human law, he notes that a law is unjust if (among
other reasons) it is contrary to the beneit of the community, or imposes
unequal burdens, or exceeds the lawmaker’s power, and then comments:
“These are acts of violence rather than laws, for as Augustine said,‘A law
that is unjust is considered no law at all’” (qu. 96). Such remarks suggest
the position that conformity with the natural law is always a criterion of
validity for the positive law. But other passages suggest the contrary view.
The passage just quoted continues: “Therefore laws of this kind do not
11
Summa Theologiae, Qu. 90–7 in Aquinas 1988.
WHAT MAKES LAW18
bind in conscience except to avoid scandal or disorder.” That suggests
that unjust human law still is human law, it is just that the reasons for
obeying are not the same as those we have for obeying human law that is
in accord with the natural law.This interpretation seems compatible with
the statement that if a human law “disagrees in some respect from the
natural law, it is no longer a law but a corruption of law.” So it is plausible
to read Aquinas as allowing unjust human law the status of human law,
but refraining from granting it the status of law in the fullest sense of
the word, which implies its conformity to the natural law and therefore
human reason (Finnis 1996, 2003; I follow Finnis’s interpretation).
The important point is that Aquinas seems relatively unconcerned
with issues of the validity and interpretation of the human law. Rather, he
was concerned with the question of what the human law ought to be and
with the nature of the obligation to obey just or unjust rules promulgated
by the state, whether they are properly called law or not.
Not until the modern period did legal theorists and philosophers with
legal expertise start to offer views on the question of whether morality
could or could not be relevant to the determination of what the
(positive) law is. Hobbes explicitly states the “command theory” version
of legal positivism: law is the command from one person to another who
is obliged to obey.12
But this is not yet positivism in the classical form it
took on with Jeremy Bentham in the late eighteenth century, because
Hobbes did not believe that the content of the law is one thing, justice or
morality another; he believed that there was no such thing as justice without
law and so could hardly have allowed that considerations of justice
are relevant to the determination of what the law is. Hobbes did hold that
the sovereign was subject to the natural law, and that positive law should
be interpreted in light of the laws of nature,but as the natural law consists
for Hobbes in precepts of rational self-interest, the issue of the relation
between positive law and morality is still not joined (see Dyzenhaus 2012
for a contrary interpretation).
12
Leviathan, ch. XXVI.
MORALITY AND THE GROUNDS OF LAW 19
Positivism proper starts with Bentham. More than that, we can say
that philosophical discussion of the nature of human law – philosophical
discussion of whether, given what kind of thing law is, morality can
be a ground of law – starts with Bentham too. It seems that none of the
early modern philosophers in the natural law tradition – Suárez, Grotius,
Pufendorf, Locke – were any more concerned than Aquinas to take a
stand on this question.13
And neither Hume nor Kant was terribly interested
either.14
Bentham’s opponents were not philosophers but theorists
of the common law – especially William Blackstone, who was the target
of Bentham’s irst published work, A Fragment on Government, in 1776
(see Lieberman 1989; Postema 1986).
Blackstone did appear to take a clear stand on the role of morality in
determining the content of the common law. When writing about precedent,
and a judge’s duty to make decisions based on the existing law
rather than his own private judgment, Blackstone commented:
Yet this rule admits of exceptions, where the former determination is
most evidently contrary to reason; much more if it be clearly contrary
to the divine law. But even in such cases the subsequent judges do not
pretend to make a new law, but to vindicate the old one from misrepresentation.
For if it be found that the former decision is manifestly
absurd or unjust, it is declared, not that such a sentence was bad law;
but that it was not law; that is, that it is not the established custom of
13
Like Aquinas, Suárez makes remarks that could be interpreted as taking a stand:“The
human lawmaker ... does not have the power to bind through unjust laws, and, therefore,
were he to command unjust things, such prescriptions would not be law, because
they have neither the force nor the validity necessary to bind.”“For a law to be genuine
law, it must ... be just and reasonable, because an unjust law is not law.” Both passages
are from De legibus, and are cited in Garzón Valdés 1998, 267. Garzón Valdés reads
these passages as implying a view that there is no such thing as an unjust human law.
They can be read that way, though Suárez in other respects, such as his embrace of the
command theory of law (Postema 2012), suggests a reading like Finnis’s reading of
Aquinas that I embrace in the text. It seems that, as for Aquinas, the important upshot
of a “law” being unjust for Suárez is that it does not “bind.”
14
Waldron (1996) plausibly argues that Kant was a legal positivist, but this has to be
extracted from the text by way of implication.
WHAT MAKES LAW20
the realm, as has been erroneously determined. And hence it is that
our lawyers are with justice so copious in their encomiums on the reason
of the common law; that they tell us, that the law is the perfection
of reason, that it always intends to conform thereto, and that what is
not reason is not law. (Commentaries, 1: 69–70)
This sounds like the view that an unjust law is an impossibility; if it is
unjust, it cannot be law. In the past century or so, many people have used
the label “natural law” for that view. But it is doubtful whether anyone
ever held such a view.We have seen that Aquinas most probably did not,
and the thought that Blackstone did is contradicted by several passages
where he rather clearly acknowledges the possibility of unjust law (see
Finnis 1967). It is more plausible to read Blackstone as advancing the
more subtle and plausible claim that determination of the common law
involves the application of reason,understood to include moral judgment,
and that on occasion this will justify treating certain (though not all) past
decisions as mistakes, and so not part of the body of decisions that are
“evidence of what is common law.”15
On this reading, Blackstone’s nonpositivism
was something rather close to Dworkin’s theory of law,according
to which the content of the law is determined via an interpretation,
guided by moral judgment, of existing legal materials.16
Bentham confronted this idea of law with a deinition of law as the
command of a sovereign supported by the threat of a sanction (Bentham
1970; see Postema 1986, chap. 9; 2011). This account of law received
its most widely read presentation in John Austin’s The Province of
Jurisprudence Determined, irst published in 1832. By the time the ifth
edition of this book was published in 1885, debate among philosophers
15
“So that the law and the opinion of the judge, are not always convertible terms, or one
and the same thing; since it sometimes may happen that the judge may mistake the law.
Upon the whole, however we may take it as a general rule, ‘that the decisions of the
courts of justice are the evidence of what is common law’” (1: 71). For discussion, see
Lieberman 1989, Part I.
16
On the afinity between traditional common-law theory and Dworkin’s views, see
Perry 1987.
MORALITY AND THE GROUNDS OF LAW 21
and lawyers about the relation between law and morality was widespread
throughout Europe, with positivism in the ascendance, at least in
England. Bentham’s insistence on the distinction between what the law
is and what it ought to be could be described, as early as 1877, and by no
less careful and fair-minded a writer than Henry Sidgwick, as “to us so
obvious a truism that it seems pedantic to state it expressly” (2000, 207).
In Germany, by contrast, the merits of positivism – usually in the guise of
“statutory positivism” – remained a matter of contentious debate in the
nineteenth century and would become politically explosive during the
Weimar period (see Dyzenhaus 1997; Paulson 2006).
In any event, not until the twentieth century were signiicant philosophical
contributions to the debate about the nature of law made,particularly
through the work of Hans Kelsen, H. L.A. Hart, Ronald Dworkin,
and Joseph Raz.
Two Pictures
The most fundamental divide within the debate over the grounds of law
is between those who believe that the foundations of the law in a particular
place are exclusively matters of fact and those who believe, to the contrary,
that the content of the law is also partly a moral matter. Different
views within each camp differ on the way factual or moral considerations
determine the content of law. These differences, though very important,
are downstream of that more fundamental disagreement.The core idea of
positivism, then, is that it is in the very nature of law, wherever it is found,
that the inal determinants of its content are facts, not norms (Coleman
2001a, 75). The opposing view has lacked an appropriate name, which is
perhaps why “natural law” has misleadingly been used.“Legal moralism”
would also be misleading, as it already has a place in a different debate
about the legitimacy of the legal enforcement of morality for its own
sake. Dworkin’s label “law as integrity” would not do because we are
looking for a name for the broader class of views of which Dworkin’s is
not the only example (though it is by far the most fully developed and
WHAT MAKES LAW22
the most important).This leaves “nonpositivism.” It may seem too broad
a label, since to deny that the law is determined just by social facts is
not, strictly, to say that it is in part determined by moral considerations
in particular. But when we limit ourselves to plausible views, there is no
third class of considerations that might be relevant to the content of law:
it is either social facts alone or social facts along with moral considerations,
especially those of political morality, so it is clear enough what
“nonpositivism” means.
In the next two chapters,I try to motivate each of these pictures of what
law is, and discuss some of the main internal issues within each camp.
23
3 Legal Positivism
A Normative System Grounded in Facts
To bring out what motivates positivism, it is best to think of it irst in its
purest and most plausible version, according to which moral reasoning
is never required or appropriate in iguring out what the law is, even if
legal sources can be interpreted as calling for it.1
Legal argument on this
view is never, in any part, about what ought to be, never involves thinking
about right and wrong, fairness or justice. Of course the language lawyers
use overlaps with that used in moral discussion – in both domains we talk
of rights, duties, obligations, wrongs, and so forth. But that is no embarrassment
to the positivist, who can accept that the law typically and ideally
relects someone or some collective’s moral conclusions about what
the law should be. Moreover, a statute or constitution or a history of decisions
might give a great deal of guidance about what the speciically legal
meaning some moral term, such as “justice,” may have acquired. What is
important is that while the content of the law may have been, and ideally
would have been, arrived at after moral deliberation, and while the law
may use words also used in moral argument, moral deliberation is not
required to identify law’s content after it has been made.
The law is what is posited, or put forward by a person or people.We
may all hope that what gets posited is good, that it matches closely with
1
This has come to be known as “hard” or “exclusive” positivism; I discuss the “soft” or
“inclusive” alternative later in this chapter.
WHAT MAKES LAW24
what the law ought, morally speaking, to be. But, insists the positivist,
it would be simply mad to look at what has been put forward as law by
people and see there, instead, what ought to have been put forward.
Suppose someone were to argue that slavery is illegal in a particular
place in part because it is a violation of people’s moral rights. The positivist
sees such an argument as like defending the claim that sexual
promiscuity causes disease by saying that promiscuity deserves to be
punished.
Sometimes positivism has been confused with the peculiar view that
it is a matter of moral indifference what the content of the law is. In fact,
the origins of positivism lie in Bentham’s enthusiasm for the reform of
law in line with what he took to be the correct account of what it ought
to be. One of his concerns was that if people believed it was in the nature
of law to be good, this would deter them from subjecting the law to moral
critique and then reform (Postema 1986, 304–5). Bentham also believed
that the sole basis for sound critique was the principle of utility; but positivism
as a theory of law takes no stand on the appropriate moral theory
for evaluation of the law.
Positivism is also compatible with the idea there might be something
meritorious about law as such.Perhaps there is no law without a legal system,
and perhaps anything that qualiies as a legal system has some merit.
Perhaps too, there are reasons for thinking that, even though law as such
does not necessarily provide reasons for action, it often does (a position
I myself defend in Chapter 7).The positivist picture of law does not offer
any kind of fully general claim about the lack of necessary connections
between morality and law (see Gardner 2001 for a clear statement of this
point). The claim it makes about the nature of law is very focused: law is
a social phenomenon in the speciic sense that its content is always ixed
by social facts, and so, in reasoning toward a view about the content of
current law, it is never appropriate to take into account considerations of
what would be better, of what ought to be.
If our thoughts about what it would be morally better for the law
to be cannot at all guide our thinking about what it is, what does guide
LEGAL POSITIVISM 25
us? The answer is legal sources: statutes, constitutions, judicial opinions,
and so on, all interpreted in a fashion that never requires the independent
moral judgment of the interpreter (a classic statement is Raz 1994,
194–221).These are the social facts that ground propositions of law. So a
positivist account of law must explain how we are to identify which texts
or practices count as legal sources, as well as an account of how they may
be interpreted in a morally neutral way. Positivist theory has traditionally
been more concerned with the irst of these.
The positivist rejects such claims as that materials count as legal
sources if it is appropriate, from the point of view of political morality,
to treat them as such. A positivist may grant that the best possible justiication
for coercing people according to certain rules is that the rules
were arrived at through a democratic process, but insists that this is not
at all relevant to the question of whether legislation is a source of law. If
legislation is a legal source, we must be able to identify it as such without
entering into questions of justiication.
Typically, a particular text counts as part of the legal order in a particular
place because another, higher-level legal source tells us so. Thus the
validity of administrative orders may be established by appeal to legislation,
and the validity of legislation by appeal to a written constitution.As
Bentham was perhaps the irst to insist, legal sources are generally not
identiied as valid one by one, but rather form a system (Bentham 2010;
see Postema 2012 for discussion). But of course the system is not selfvalidating;
if a highest-level legal norm exists that explains the validity
of all lower-level norms, the question of the validity of that highest-level
norm still arises. Put otherwise, we might be able to identify or dream up
a number of systems of rules, with their own internal structure and organizing
highest norms. The question we would need to answer is which
of those systems is in force as the legal system of a particular place. The
challenge for the positivist, then, is to tell us what matter of fact establishes
that a particular system of legal norms is in force.
For Bentham and Austin, the solution lay in a certain account of
sovereignty. Law, “strictly and properly so-called” as Austin liked to put
WHAT MAKES LAW26
it, was the command of a sovereign, backed up with a threatened sanction.
This position is positivist because commands can be identiied as
events that happen in the world, and the sovereign, the person or collective
whose commands count as law, is also identiied purely descriptively.
The sovereign is not such because of some right to rule, but because of
the existence of a habit of obedience. Therefore, to ind out the content
of the law it is necessary to igure out whose commands are habitually
obeyed and then igure out what he has commanded.
Hart’s detailed criticism of the command theory has been very inluential
(though for some reasonable doubts, see Schauer 2010a). Part of
the critique pointed out that many legal rules are only very awkwardly
understood as commands. The rules of contract law, for example, are
better understood as facilitative than imperative. Also, the idea that
all legal rules are backed with sanctions cannot be made to it much of
what is commonly regarded as law. (For example, Austin had to concede
that, according to his view, a statute repealing another statute was not
strictly a law.)
Hart’s alternative account is that the validity of legal rules is determined
by other legal rules in a hierarchical structure that ends with a set
of ultimate criteria of legal validity, identiied as such by a rule called the
rule of recognition. But this special rule is not really a rule of the system
of legal sources; rather it is a rule deinitive of the system, grounding it
from outside. To avoid the question, “But how do we know that the rule
of recognition is valid?” Hart holds that it is neither valid nor invalid, it
just is. By which he means, it is the rule of recognition that is in effect.
It is the rule of recognition that the participants in the practice of making,
but especially applying, the law in a particular place actually accept –
they treat it as appropriately governing their activities within the legal
practice.The situation is analogous to the“rule”that tells us that the rules
of chess are the rules of chess. How do we know that these are the rules of
the game, rather than those? No rule of the game of chess can answer that
question.We can only answer, well, these are the rules that are generally
accepted to be the rules of chess.
LEGAL POSITIVISM 27
It seems evident to most commentators that Hart’s account of
law is superior to Austin’s. Given that there is so much disagreement
about the nature of law, this area of agreement might seem puzzling.
We cannot say that it seems evident to everyone that Hart has more
accurately described the nature of law, because it does not seem so to
nonpositivists – from their point of view, both accounts are utterly hopeless.The
best way to characterize the relationship between the command
theory and Hart’s account is that they are competing efforts to provide
detail for the basic picture of law as social fact. If that basic picture is
taken for granted, Hart’s account is better to the extent that it its better
with our ordinary contemporary ways of talking and thinking about the
law.The great achievement of The Concept of Law is to tell a story about
how the law is determined purely by social facts that appears entirely
natural, even commonsensical, to the reader who is already attracted by
that basic picture.
There is supericially a great contrast here with the work of Hans
Kelsen.2
Kelsen’s commitment to a “presupposed” valid Grundnorm
(“basic norm”) as the inal step on the ladder of validity strikes many
people, at least in the English-speaking world, as philosophically extravagant
and obtuse. But actually the difference between Hart’s and Kelsen’s
accounts is not nearly as great as it may seem, and for a rather banal reason:Though
both wrote that they were giving accounts of legal “validity”
(“Geltung”), they meant different things by the word (Raz 1979, 150). For
Hart, the ultimate criteria of validity as set out in a rule of recognition
simply identify those rules that are “in force,” or are part of the law, in
a particular place. This leaves open, as an issue requiring separate discussion,
what reason, if any, a person might have to obey those rules. In
his rejection of Austin’s account of law, Hart pointed out that for many
people legal rules play a signiicant deliberative role; they are not simply
2
Kelsen’s views changed over time. My account, which is heavily inluenced by Raz 1979,
is based mainly on Kelsen 1967, but seems compatible with earlier and later publications.
For discussion of the development of Kelsen’s views, see Paulson 1998.
WHAT MAKES LAW28
obeyed out of habit, but are treated as providing reasons for action. He
also introduced the idea of “internal” legal statements – those that, in
making claims about the law that is in force, express the utterer’s acceptance
of the legal norms as reason giving. For example, the statement,“it
is my legal duty to pay my taxes,” is, for Hart, an internal legal statement
in this sense. But all of this is compatible with the possibility that valid
legal rules provide no one with any genuine reason for action at all; in
particular, neither the fact that a rule is legally valid nor people’s willingness
to make internal statements about that rule implies anything about
its moral force. Hart did believe that he had to provide an account of the
discourse of “legal obligation,” but that account too leaves quite open the
issue of whether I have a real reason to do what I have a legal obligation
to do. Law, in this sense, is just like the system of rules we call etiquette,
or indeed any rule-governed game; we can make sense of proper and
improper moves we can make from within the game, without taking any
kind of stand about why that is a game we should play. This aspect of
Hart’s view is discussed further in Chapter 7.
Kelsen used the term “validity” differently. For Kelsen, legal norms
purport to impose oughts, and insofar as they are valid, they do impose
oughts. There is some controversy about whether Kelsen thought that
legal oughts were in some sense a different kind of beast than moral
oughts, but such a distinction would only matter for current purposes if
he also thought that legal oughts were somehow less binding than moral
oughts, which he clearly did not. And so the reason Kelsen had to insist
that the foundation of a legal system was a presupposed basic norm
rather than some matter of fact, was that, as he saw it, validity implies
ought and you can’t get an ought from an is.As Kelsen (1928, 393–4) puts
it, “This, indeed, is the problem of the positivity of law: The law appears
as ‘ought’ and ‘is’ at the same time, while logically these two categories
are mutually exclusive.”
Nonetheless, facts are the foundation of the grounds of law for Kelsen
as well. The content of law for Kelsen is determined by the rules of the
constitution that is effective, and a constitution is effective if the norms
LEGAL POSITIVISM 29
created in accordance with it are by and large applied and obeyed. It is
just that those norms are only legally valid in Kelsen’s sense if there is
a further norm, the basic norm, which states that people ought to obey
the effective constitution.Without that grounding ought statement, there
would only be matters of fact, no oughts.That would do for a sociological
account of law, but not for a “juristic” one (Kelsen 1967, 218).
So we see that both Hart and Kelsen offer accounts of how to determine
which legal norms are in force without at the same time taking a
stand on whether they impose genuine obligations. Hart calls this an
account of legal validity, while Kelsen insists that it is not yet that. There
is no need to quibble over these words. I generally use “validity” in Hart’s
normatively neutral sense because it seems to comport with current
English usage. The issue of what kind of obligation valid legal rules can
impose is discussed in Chapter 7, and that is the right place to discuss
Kelsen’s doctrine of the basic norm (see also Marmor 2010).
The question we are addressing now is how to determine the content
of the law that is in force. And though Hart’s account of the law that is
in force is in some ways an advance on Kelsen’s, it is fundamentally the
same view. For Hart, the content of the law is determined by the ultimate
criteria of legal validity. The ultimate criteria of legal validity are
set out in the rule of recognition, which is in effect if it is actually followed
by people in the business of identifying law.The content of law for
Kelsen is determined by the rules of the constitution that is effective.The
“constitution,” in this usage, means the collection of rules regulating the
creation of legal norms that apply to people, and is in that sense roughly
equivalent to Hart’s rule of recognition, which sets out ultimate criteria
of legal validity (Kelsen 2006, 124; 1967, 222). A constitution is effective
if the norms created in accordance with it are by and large applied and
obeyed. On both accounts, then, law has the content it does because of
certain facts about the practice of people, especially legal oficials.3
3
Hart believes that unless oficials take an “internal attitude” to the rule of recognition,
a legal system will not persist. But this internal attitude is still a matter of fact.Whether
WHAT MAKES LAW30
It might seem that on these accounts of how to determine the law in
force we have not just given up on Kelsen’s thought that to say a legal
norm is valid is to say that it is genuinely binding; we have also given up
on any plausible understanding of the law as a set of norms at all. For
whether or not it imposes genuine obligations, it must be possible, if law
is to be understood as a set of norms, for there to be divergence between
what the law requires and what happens in legal practice; that seems hard
to square with the idea that the content of the law is determined by matters
of fact. But the positivist model of how to determine the content of
law – not just the model of Hart and Kelsen, but also that of Bentham
and Austin – is compatible with this minimal sense in which law is a set
of norms. It does not reject the idea that there is a truth of the matter
about what the law is in particular places and that legal oficials can be
mistaken about this. On the command model, a judge could go wrong in
misunderstanding the content of the sovereign’s command. In Hart and
Kelsen’s version, a decision might be mistaken when a particular legal
oficial decides contrary to the constitutional norms that are generally
applied, or simply fails to reason correctly within the normative system
the ultimate criteria of validity set up.
To sum up: the positivist picture of law is distinctive in inding the
grounds of the law, the ultimate sources of its content, in matters of fact.
Reducing the accounts of Hart and Kelsen to what is relevant to an
account of the grounds of law (as opposed to an account of the role law
plays in the deliberations of most legal oficials, or an account of what
kind of obligation, if any, law imposes), we extract a plausible way to ill
out this picture: the social facts that provide the grounds for the content
of law are practices that most legal oficials actually engage in most of
the time. In saying that it is plausible, I mean that for those of us who ind
the picture of law as social fact appealing in the irst place, this account
introduces no awkwardness; it is compatible with us (including those of
oficials can in fact treat the rule of recognition as providing reasons for action is discussed
further in the next section.
LEGAL POSITIVISM 31
us who live in constitutional democracies) continuing to say most of the
things we are in the habit of saying about the law.
Positivist Puzzles
If we take Hart’s formulations as our positivist model, there are some
ambiguities to unravel and revisions to be made. Hart (1994, 110) writes:
“The rule of recognition exists only as a complex, but normally concordant,
practice of the courts, oficials, and private persons in identifying
the law by reference to certain criteria. Its existence is a matter of fact.”
It is this formulation that is closest to Kelsen’s account of the eficacy of
a constitution.
Hart also tells us that the oficials of a legal system must accept the rule
of recognition.On the notion of acceptance introduced to illustrate the idea
of the internal point of view, those who accept a rule treat it as providing a
general reason for action that also applies to others and justiies criticism,
of themselves and the others, for noncompliance. Just why they treat the
rule in this way is, as we have seen, entirely open. It could be because they
believe that prudence counsels it, or morality demands it, or because, like
most other people, they just do, without having thought much about it.
The rule of recognition, Hart tells us,
if it is to exist at all, must be regarded from the internal point of view
as a public, common standard of correct judicial decision, and not as
something which each judge merely obeys for his part only.Individual
courts of the system though they may, on occasion, deviate from these
rules must, in general, be critically concerned with such deviations as
lapses from standards, which are essentially common or public.This is
not merely a matter of the eficiency or health of the legal system, but
is logically a necessary condition of our ability to speak of the existence
of a single system.
The unity and continuity of the legal system, he goes on, “depends
on the acceptance, at this crucial point, of common standards of legal
validity” (116).
WHAT MAKES LAW32
But if “acceptance” here means what it usually does – treating the
rule as a reason for action – this is puzzling, since “common standards of
legal validity” cannot provide, all on their own, a reason for action.
In Hart’s model of law as the union of primary and secondary rules,
the former are“duty-imposing”and the latter“power-conferring.”4
There
is a sense in which both kinds of rule can guide action. Hart’s secondary
rules of adjudication and change specify how legal disputes are to
be authoritatively resolved and how law is changed in the system. These
rules tell people what oficial role they need to attain and what they need
to do if they are, in accordance with law, to resolve a particular dispute
according to law or to change the law. Similarly, the rule of recognition
tells someone wishing to declare the content of law how to do that. But
these rules guide action hypothetically – “if you want to do that, this is
the way to do it” – just like a recipe or the rules of a game. None of
them provides categorical reasons – “do this!” – for action. Hart’s rule
of adjudication should not be confused with a theory of adjudication as
discussed in the previous chapter, that directs those with the oficial role
of deciding legal disputes how that should be done; it is rather a rule that
tells us who has the authority to decide disputes and according to what
formal processes (97). The rule of change similarly tells us which kinds
of acts of which people or assemblies will be effective in making law, not
that anyone should make law.
Our interest is in the rule of recognition. People can of course accept
that they should comply with or apply the law, and if they do they will
need to know the content of the rule of recognition. But in accepting
a norm – treating it as a reason for action – that I should comply with
or apply law as identiied by the rule of recognition, I am not accepting
the “common standards of legal validity.” I cannot treat a rule that simply
speciies the ultimate criteria of legal validity as itself a reason for
4
Some power-conferring rules, such as rules of contract law, are not secondary rules
according to another way of marking the distinction that Hart sometimes employs –
according to which secondary rules are rules about other rules.The infelicities of Hart’s
distinction have been much discussed, but I won’t explore them here.
LEGAL POSITIVISM 33
action. I treat the norm “feed your children nutritious food” as a reason
for action, but I cannot treat principles of good nutrition themselves as
reasons for action.5
The whole thing makes much more sense if we regard the rule of
recognition as no practical rule at all, but rather a simple statement of
ultimate criteria that one may hold correct or not. From this perspective,
we can read Hart’s use of the word “acceptance” when discussing the
rule of recognition as meaning something other than it usually does for
him. We can say that the oficials must accept that these are the actual
ultimate criteria of validity for this legal system in the sense that they
must believe that they are, or be disposed to treat them as such. It would
be a problem if no one actually thought that the criteria of validity they
appealed to in any particular instance were the real ultimate criteria of
validity, applicable alike to all cases concerning all people. It is plausible
to say that for a legal system to continue and have unity the oficials of
the system, those who have de facto power and administer it in the name
of law, must converge in what they regard the ultimate criteria of validity
within the system to be. If each oficial believed that the criteria applied
“for their part only” we would, as Hart writes, “be in the presence of a
lusus naturae worth thinking about only because it sharpens our awareness
of what is often too obvious to be noticed” (116). So the matter of
external fact that determines whether a rule of recognition “exists” is, we
5
Raz (1979, 92–3) interprets Hart as holding that the rule of recognition is not just a
statement of common standards of legal validity, but a duty-imposing rule, imposing
duties on law-applying oficials. See also Coleman 2001a, 84–6. I argued in the previous
chapter that judges can in principle fully discharge their oficial obligations without
developing a view about the content of the law already in force; so in principle,
positivist judges do not need to know what the rule of recognition is. But supposing
that judges have a legal duty to igure out what the law is, the interpretation makes
sense. But now do legal oficials, as a matter of “logic,” need to accept, in the “treat as
reason-giving” sense, that they must either comply with or apply law, either as subjects
of oficials? It doesn’t seem that the existence of a legal system depends on this, though
certainly its health and eficiency would. What matters is that legal oficials do comply
with or apply the law, not why. If case-by-case self-interest were suficient to ensure
compliance by oficials, that would be enough. See further Chapters 7 and 8.
WHAT MAKES LAW34
should say, convergence in belief or attitude among legal oficials about
the ultimate criteria of validity.
Is this interpretation compatible with Hart’s statement that the rule of
recognition exists as a practice of the “law-identifying and law-enforcing
agencies effective in a given territory”? (Hart 1984,336;see also Raz 1979,
90–7; 1999, 132–48). Let us use Raz’s useful label “law-applying institutions”
to cover both kinds of institution and understand Hart’s “legal
oficials” as the people who staff them. The question is: What exactly is
the “practice” by these oficials that is supposed to reveal the content of
the rule of recognition? As we have seen in the previous chapter, looking
at what courts and other legal decision makers actually do and say will
reveal (if we assume that all these decision makers are conscientious)
implicit theories of adjudication, but not necessarily beliefs about criteria
of legal validity.
The same clearly goes for law-enforcing agencies. The conscientious
police oficer will, we hope, believe that law sets limits to what she can
do in her oficial capacity; but no one would think that her beliefs about
the law could be read off from what she does. Kelsen’s (1967, 349) image
of the law providing a frame only, leaving particular determinations
to the law applier’s discretion, is uncontroversial when applied to law
enforcers.
Hart’s view that the primary source of information about the grounds
of law was the practice of legal oficials is appealing because it suggests
a close connection between de facto political power and the content of
effective law. It also promises to allow us to say that law is grounded in
very concrete social facts – facts about what people do and say. But it
turns out that these two positivist desires cannot be satisied together.
The only practice that would directly and reliably reveal beliefs about
ultimate criteria of legal validity is that of legal experts, particularly legal
academics, in preparing textbooks and other presentations of the state of
current law. This would reduce positivism to the unhappy view that the
ultimate criteria of legal validity are found in a rule that exists as a practice
of law professors.
LEGAL POSITIVISM 35
The better solution is simply to abandon the idea that the rule of recognition
exists as a social practice in any straightforward sense. The rule
of recognition, if it exists, is a set of common beliefs and or attitudes,
perhaps implicit, about the ultimate grounds of law and/or dispositions
about what to regard or treat as the ultimate criteria when iguring out
what the law is.
When we now consider whose beliefs count, Hart’s association of the
foundations of law with legal oficialdom is largely but not wholly vindicated.
Most people who do not work with the law have only vague and
uncertain beliefs about ultimate criteria of legal validity.Among oficials,
some have more comprehensive interaction with law than others, and it
is natural to think that the more comprehensively a person is involved
with the law, the more that person’s beliefs count in ixing the grounds of
law. So Hart’s frequent emphasis on the courts is justiied; but the beliefs
of legal oficials in, for example, law departments of the executive branch
should be regarded as especially signiicant also.
However, though oficials are clearly the central case, it would be
more accurate to say that we are looking, more generally, for convergence
in belief about the ultimate criteria among those who work with
the law in their professional capacities.This expands the net beyond oficials
of the state to include practicing lawyers and academic legal experts.
The intuition here is that even if all state oficials started, for example, to
declare that some formerly embraced constitutional principle was not a
source of law, we would not immediately be inclined to say that the rule
of recognition had changed if all lawyers and legal experts disagreed. At
the limit, too, it is wholly possible for an entire population to have beliefs
about the ultimate criteria of validity, and if its beliefs were suficiently
comprehensive and convergent, that would be relevant to the issue of
what the law is should some oficials start to head off in a different direction
(Adler 2006;Tamanaha 2001, 166–7).
Now the appeal to convergence in belief about criteria of validity
among those with a life in the law may seem hopelessly optimistic.
Dworkin has always denied that any such convergence exists,even among
WHAT MAKES LAW36
judges. One response from positivists has been to point out that there is
an important difference between agreement on the proper standards for
determining legal validity and agreement on the correct application of
those standards in a particular case (Coleman 2001a, 116). This is certainly
true; the positivist need not be embarrassed just by the fact that
two people might apply the same legal rule to the same facts and arrive at
different conclusions. This is as true for standards of legal validity as for
primary legal rules that apply directly to people. Two people may agree
entirely on a criterion for the legal enforceability of agreements (there
must be a mutuality of interest in the exchange, say), but disagree about
whether a particular agreement is enforceable; and there could be consensus
that statutes that have not been enforced for a very long time are
no longer valid but much disagreement about how long, in this context,
is long enough.
This point can only take us so far, however. Even if all acknowledge
the same text as setting out ultimate criteria of validity, disagreement
about the proper method of interpretation of that text can undermine
consensus about the grounds of law. One group might hold, for example,
that a certain eighteenth-century text should be read with an eye to what
would best advance some purpose (perhaps not stated in the text) in the
contemporary world,while another might hold that it must be understood
to mean what a literal-minded eighteenth-century reader would have
thought it meant.When we say that there is agreement about the grounds
of law, we must mean more than that all recognize certain canonical statements
of criteria of validity to be correct.“Manifestly unreasonable legislation
is invalid” for example, is clearly of no use as a criterion of validity
unless we have some ideas about the relevant notion of reasonableness.
If there is to be a rule of recognition, there must be genuine consensus
about the ultimate criteria of legal validity, not nominal consensus
that consists in a general willingness to use the same form of words.6
Of
6
I believe this is the fundamental insight of “The Model of Rules I” (Dworkin 1978,
14–45).
LEGAL POSITIVISM 37
course, a rule of recognition can have gaps – there may be no consensus
on some questions of validity even if there is consensus on many. It is an
open and complicated question how much agreement of the right kind
there is in any particular place. But I will postpone further discussion of
this issue until Chapter 6, where the problem of disagreement will be
taken to the next level, the level of disagreement about the nature of
law itself.
For now, there are other positivist puzzles to consider. On the positivist
account, the foundations of law lie in people’s beliefs and attitudes
about the foundations of law. Many people have nonpositivist views
about law; they may believe that the ultimate criteria are determined, in
the end, by some argument having to do with the legitimacy of the state.
For example, they may believe that a constitution sets out the ultimate
criteria of legal validity just because the scheme of government it sets up
constitutes a legitimate coercive political order. Our positivist does not
have any dificulty with the fact that some beliefs about ultimate criteria
are based on what (to the positivist) are mistaken views about law. All
the positivist looks at is what the beliefs about the grounds of law are, not
the justiication someone might offer for those beliefs. What is interesting
is that it is much harder to make sense of the beliefs of the positivists
among us.
If we are positivists, we think that the grounds of law are what we
think the grounds of law are. It can seem that there is nothing else for
us to be thinking about, when we are thinking about the grounds of law,
other than what we think.
We do have ideas about what legal norms are. At the least, we think
legal norms are directives that present themselves as legitimate demands
on our conduct.Those who create those norms, those whose institutional
role it is to pronounce their application to particular cases, and those
whose institutional role it is to enforce them, all present their institutional
acts as justiied.As discussed further in later chapters, this seems to
be the truth in Raz’s important idea that law claims authority. Something
like these ideas must be in our minds as we distinguish legal norms from
WHAT MAKES LAW38
norms of morality, conventional morality, etiquette, or the rules of a
game.7
But for the positivist, these ideas themselves do not give us any
basis on which to identify the grounds of law. When it comes to that, all
we have to go on is our beliefs about the grounds of law, and those beliefs
are themselves not justiied by any further set of considerations.The puzzle
is that when it comes to the question of what the ultimate criteria are
for determining the content of law, there is nothing substantive for each
of us to be thinking about.
Some have suggested that the rule of recognition is best seen as the
conventional solution to a coordination problem.A coordination problem
is one where the substance of the solution is not what matters (whether
we drive on the left or on the right), but the fact that we all agree on
what it is. If this were a plausible way to think about law, it might make
sense of the puzzle. If law is just a conventional solution to a coordination
problem, it is not surprising that there is no substantive basis to anyone’s
view about the grounds of law, since it wouldn’t matter what the grounds
of law were, just so long as there were some.
But of course this isn’t how we think about law (Dickson 2007; Green
1999). We think it matters a lot whether the king’s command, or the
court’s opinion, or the acts of parliament are sources of law. This is what
makes natural Dworkin’s claim that beliefs about the grounds of law
really are, in the end, grounded in political theory. On Dworkin’s view,
there is something for each of us to think about when we relect on the
foundations of law; we are to think about what might make the resulting
government by law legitimate and potentially just. The positivist agrees
that that’s what we should think about when we are thinking about what
the grounds of law should be, but denies its relevance when we are thinking
about what they are.
The solution to the puzzle of what each positivist is supposed to think
about, when asked what he believes are the ultimate criteria of validity, is
7
In Chapter 8, I discuss further what we might have in mind when we are thinking about
a norm as legal, rather than something else.
LEGAL POSITIVISM 39
found in the history of law.8
In an ongoing legal system, judges, law professors,
and so on are not reduced to asking each other what they think at
any particular time. One thing they all believe, for a variety of reasons, is
that the ultimate criteria of legal validity do not typically change radically
from moment to moment. And so there is something to anchor all our
beliefs about the grounds of law now; it is what we know about what has
been generally believed up until now. There is, in this respect, an important
historicist aspect to positivism, whether it be based on the command
theory (habits of obedience build up only over time) or the model of the
system of norms.
But of course legal orders can undergo revolutionary change. When
change happens, the relevant people can no longer look to history to
ground their beliefs about criteria of legal validity, and so a different
account is needed.The positivist account of radical constitutional change
must be that relevant people somehow reach the position, at roughly the
same time, that it would be better if these were the ultimate criteria of
validity (those set out in the new constitution) rather than those (the constitutional
rules that had been traditionally followed). People’s reasons
for thinking that these criteria would be better than those may or may
not be grounded in moral beliefs; they may also be grounded in a convergence
in people’s sense of their own interests. But again, what matters is
not why they believe this, but that they become disposed to identify law
with reference to the new criteria.
The success of this kind of revolutionary legal reform depends on
everyone connected with law going along with it. If from one day to
the next, these people start treating different factors as determinative
of legal validity, then the revolution has succeeded. When it comes to
fundamental constitutional change, as Hart memorably put it, “all that
succeeds is success” (1994, 153). It should be noted, however, that the
expanded account of whose beliefs count for determining the existence
of the rule of recognition offered earlier forces some changes to Hart’s
8
I am heavily inluenced here and in what follows by Marmor 2009, 155–75.
WHAT MAKES LAW40
account of fundamental constitutional change. On that account, if all the
legal oficials embrace a new rule of recognition, then the rule of recognition
will have changed. But on the broader account I have argued is
necessary, the legal oficials would have to bring everyone else with a
life in the law along with them. And this is intuitively right – a fact that
supports the broader view. For if the oficials embrace one set of criteria
while all the nonoficial lawyers and legal experts reject it, we would be
more inclined to talk in terms of an illegal coup d’état, as opposed to a
successful legal change.
At the moment of revolution, when relevant people become disposed
to identify law by reference to the new criteria, it would be misleading
to say that there was a convergence of belief about ultimate
legal criteria. Deciding to treat certain factors as the grounds of law
is not the same as believing that they are the grounds of law. But this
is not embarrassing for the positivist story. The situation is closely
analogous to inventing rules for a new game, or drastically changing
the rules for a game with an existing name. Andrei Marmor’s (2009)
notion of a constitutive convention is helpful here. We can say that,
at the moment of revolution, the members of the relevant group are
disposed to appeal to certain criteria to determine legal validity, contingent
on their belief that others are similarly disposed. The reasons
people have for being disposed to appeal to a particular set of criteria
are not just that they believe others will – they have substantive reasons,
good or bad, for hoping that others will. But, as positivists, they
will not, at the revolutionary moment, appeal to those criteria unless
they expect others to, because they believe that consensus on the relevant
criteria is required before the new legal order can take effect. If
all goes well for the revolutionaries, there will be a convergence and a
new legal order will have been constituted. Once people believe that
this has happened, then they can straightforwardly be said to have
beliefs about the grounds of law. And from that moment onward, there
is no need to think about the rule of recognition as a convention in
any strong sense that implies that part of the reason people have for
LEGAL POSITIVISM 41
thinking that such and such are the ultimate criteria is that everyone else
does as well.
The situation here is exactly analogous to the emergence of a new
norm of customary international law. A rule of customary international
law exists when there is a general practice of states attended by an
“opinio juris” – the belief that the practice is a case of following the law.
There is an apparent paradox, or circle, in this criterion; it appears that
belief that the practice is legally required has to precede its being legally
required – which, unless law is to be founded on irrational or false beliefs,
seems impossible. It would be irrational to come to believe, for no reason
at all, that a practice until now not legally required is legally required.
The solution is the same as that just sketched for change in the rule of
recognition.
First, there is no problem with rules of long standing. Everyone relevant
may believe that such and such is a customary rule of law just
because that’s what has generally been believed for a long time. There
is nothing irrational or puzzling about that (Tasioulas 2007, 322).As with
the rule of recognition, the problem comes with changes in customary
international law. It would be rational to come to believe that a practice
is legally required on the basis of a reasonable but false belief that there
is a general belief that it is legally required, but it would be troubling if
this were the only way a new customary norm could get off the ground.
But there is a non-troubling possibility, which we might call the orthodox
route to a practice becoming legally required. States may converge in the
belief that it would be good if a certain practice had the status of law, and
so be disposed to treat it as law so long as enough others do. Once the
initial leap is made,and enough others are treating the practice as law,the
belief of any one state that the practice is law need not be contingent on
everyone else’s attitudes or dispositions anymore and the validity of the
rule will just depend on the fact that most states believe it is law.9
9
I return to customary international law in Chapter 8. The literature on the so-called
chronological paradox is voluminous; I ind Lefkowitz 2008 especially helpful.
WHAT MAKES LAW42
Inclusive and Exclusive
The fact that ixes the grounds of law on the positivist account I have outlined
is a convergence of beliefs or dispositions among those with a life in
the law. But now there is an apparent problem linking this view with the
main feature of the picture of law as fact as I have outlined it – the idea
that it is never necessary to appeal to moral considerations to determine
what the law is. For of course many people do believe that moral considerations
are relevant to the determination of legal validity.
On the one hand there is the irm conviction, best expressed in the
works of Raz, that law is found in a morally neutral interpretation of
valid legal materials. On the other hand, there is the basic positivist
position that the validity of legal materials is determined by contingencies
of the beliefs or practices of legal oficials and relevant others.
The problem is that, at the top of the ladder of positivist validity,
beliefs about the ultimate criteria of legal validity might not be consistent
with the idea that morality is never relevant to determining the
content of law.
Positivism splits at this point. If you focus on the motivation for
being a positivist in the irst place – the desire to be able to insist that
the content of the law is one thing, its moral merit another – the exclusivist
position is irresistible. In Raz’s (1994) terminology, the exclusivist
view is that the grounds of law are all socially identiied sources that can
be interpreted without recourse to moral judgment. If we take this view
about the nature of law, moral standards simply cannot be part of law.
So moral language in legal materials will be read not as part of law but
as directing a decision maker to make a judgment, outside the bounds
of law, on the moral merits. If legal oficials and the rest believe that, to
the contrary, those moral-sounding requirements in the materials really
are cases where morality is incorporated into law, that belief must simply
be ignored.A belief that some unwritten overarching moral standard
of validity is among the ultimate criteria must likewise be ignored. On
the exclusive positivist account, we determine the grounds of law by
LEGAL POSITIVISM 43
appeal to a puriication of a consensus among relevant people about the
grounds of law.
There is nothing objectionable about this way of proceeding. The
main positivist idea that law is grounded in fact is not itself justiied by
asserting that everyone believes this. Rather, philosophical argument
is required to get us to that point. The exclusive position is that law is
grounded in social sources, read straight, and that the beliefs of those
with lives in the law about what the social sources are settles the matter.
The inclusive position is that law is grounded in social sources along
with moral considerations that those with a life in the law agree are
relevant. In other words, the content of the rule of recognition itself
will determine whether grossly unjust rules or standards can be valid,
whether a “moral reading” of sources is always appropriate, or whether
“moral language” in sources incorporates just those aspects of morality
into law.10
On each view, certain beliefs of legal participants become relevant
once a certain account of the grounds of law is accepted (that law is
grounded in facts about beliefs and attitudes of those with a life in the
law). On each view, certain beliefs about the grounds of law held by legal
participants must simply be ignored – for example, that there can be no
law that contravenes God’s will. But the views differ on whether beliefs
that certain moral considerations are among the rounds of law must be
ignored. How to choose between these views?
Both versions of positivism are perfectly coherent.The exclusive view
seems overwhelmingly more appealing to someone with an initial picture
of law as fact; but then the inclusive view sits more comfortably when we
are discussing some aspects of legal practice. I will not here address the
10
I leave aside a view that holds that where “moral language” is found in a source, there
morality is incorporated into law by that very fact, irrespective of the content of the
rule of recognition. The problem with this view, as Raz (2004) has discussed, is that
some argument is required, from within the positivist outlook, for assuming that the
moral considerations a judge is, the externalist agrees, thereby directed to consider
when making a decision, are themselves part of law.
WHAT MAKES LAW44
internecine arguments between these two positions, since, as I argue in
Chapter 6, we lack compelling argument at the more basic level: we lack
compelling argument for preferring either version of positivism to any
version of nonpositivism.11
I will, however, make one remark, without further comment, about
the internal debate: the inclusive position seems reactive, driven by the
need to reply to various objections to positivism. The exclusive position,
by contrast, is motivated by a strong sense of law’s validity being entirely
distinct from its goodness. If no one had this initial conviction about the
nature of law, I doubt that the very idea of positivism would ever have
emerged. The nonpositivist position is driven by the opposite conviction
that law occupies the same general space as the good and the required;
that’s just the kind of thing law is. It seems to me that no primitive pretheoretical
motivation for the inclusive positivist position exists. It stakes
out a middle ground that makes positivism seem less jarring when considering,
in particular, constitutional adjudication grounded on bills of
rights (Waluchow 1994). But this middle ground seems unstable, since
once we have become comfortable with the idea that the content of law
is sometimes ixed by the truth about what is good and right, it isn’t so
clear why we have reason to resist the nonpositivist view that considerations
of the good and the right are always relevant to the content of law.
In other words, in the positivist conviction that law is a matter of fact, not
morality, it is the second, negative part of that thought that seems to me
dominant.
11
Inclusive legal positive is defended in Coleman 2001a and Waluchow 1994; Himma
2002 is a helpful shorter discussion.
45
4 Nonpositivism
Good Things about Law
Most people who take any kind of interest in the debate about the nature
of law seem to have very strong views, right from the start. Either positivism
or nonpositivism is obviously right and the other obviously wrong.
What seems obviously wrong about positivism to many is that it misses
the fact that law in its nature is something good, or can be seen as striving
toward being something good – or at the very least, is something that
can’t be very bad. For most people of this inclination, the evident truth
that law is something good, or at least potentially so, is tied up with the
further evident truth that the law is genuinely binding on us, or is usually
so, or is in some sense meant to be so (Greenberg 2011) – it is, as Dworkin
all along insisted, a domain of real rights and obligations. To see law as
ultimately grounded in social fact is to miss these essential moral qualities
of law. From this point of view, it may turn out that the Nazis and the
Taliban have no law, but who cares about that? If something interesting is
going on in this whole domain, something worth relecting on, especially
something worth relecting on philosophically, it must be because there
is something valuable or at least potentially valuable about law, or at any
rate something immediately morally relevant about law, and part of the
philosophical task is to igure out what that is (see, for example, Perry
2001, Soper 1984).
To be clear about the contrast here, we have to remember that positivists
of course agree that there is something potentially valuable about law.
WHAT MAKES LAW46
Good law is good. They may even agree that, wherever there is law, there
you are likely,or even certain,to ind something that is in one way good – an
effective legal system will greatly increase the range of social possibilities,
and we may say that this, in itself, is in one way good. They also are likely
to say that, depending on the law’s content, there are often moral obligations
to obey (some of) it.As we will see in Chapter 7, it is consistent with
positivism to believe that especially certain kinds of law (those that apply
to states) are very likely to generate reasons for action.The disagreement
is that the nonpositivist insists that the inherent moral signiicance of law
must be kept in mind when thinking about what kind of thing law is, and,
in turn,must structure any theory of how to determine legal content in any
particular place.The positivist,by contrast,believes that we can account for
law’s nature while bracketing any moral signiicance it may have.I argue in
Chapters 6 and 7 that the fact that law can have moral signiicance should
be the reason that anyone engages in the exercise of providing a theory of
the grounds of law in the irst place. But for the positivist, that theory itself
is not shaped by any view about law’s value.
That we are describing something of inherent moral signiicance is
a common assumption of nonpositivism, but this broad idea could be
explicated in many different ways. Since the late 1960s, nonpositivism has
rightly come to be associated with Dworkin’s legal theory. This theory is
both deep and complex, and it contains many strong commitments that it
is worthwhile to identify individually, so that we can better see the range
of nonpositivist options.
Morality in the Grounds of Law: The Moral Reading
and the Moral Filter
What I take to be deinitional of all nonpositivist views is that moral considerations
are always among the grounds of law. Morality can igure in
determining law’s content in two main ways.In the irst,it is always playing
a role as a guide to interpretation, resolving indeterminacy and perhaps
also correcting particular mistakes in the law to show the law as a whole
NONPOSITIVISM 47
in its morally best light. This view has its roots in common-law legal theory
and was developed into a comprehensive theory of legal interpretation
by Dworkin.The term “interpretivism” is sometimes (Dworkin 2011,
401; Stavropoulos, 2003) used as a label for Dworkin’s view, but since any
theory of law acknowledges that legal materials require interpretation, I
prefer a more revealing label that Dworkin (1996) also used – “the moral
reading.” As one common formulation of this approach has it, the law is
what follows from the principles that best justify past legal practice.
A different kind of nonpositivism takes a blunter approach.As is well
known to English-speaking legal theorists thanks to discussion in Hart’s
(1958) article “Positivism and the Separation of Law and Morals,” Gustav
Radbruch came to the conclusion that a positivist view of law was partly
to blame for the degree of obedience to Nazi law the German population
displayed during the Third Reich. In response, Radbruch developed the
idea that it was contrary to the nature of law that it could, in its content,
be seriously unjust.Thus much of what was presented as law by the Nazi
regime was not law at all. As noted in Chapter 2, at least one German
court embraced the so-called Radbruch formula in another reconstruction
era – the reuniication of Germany after the fall of the Berlin Wall in
the late twentieth century.
The view that nothing can be law that is very unjust is essentially
(some form of) positivism, with a moral ilter laid on top. The thought
here is that, even though law isn’t always perfect, we rebel at the thought
of what might be called perversions of law. An institutionalized system
of rules and standards that avowedly is aimed at genocide cannot, on this
view, fall within the domain of law. The most prominent defender of the
moral ilter view at the turn of the twenty-irst century is Robert Alexy;
his argument for the moral ilter is discussed briely in Chapter 6.
Legality, Legitimacy, and Justice
Both the moral reading and the moral ilter need to be further speciied,as
perhaps not just any moral consideration is relevant to the interpretation
WHAT MAKES LAW48
of law. One might hold, for example, that what is good about law is that it
regulates social life in a way that respects the autonomy of its members
and treats them all as equals before the law.
That some version of the ideal of the rule of law shapes our understanding
of the nature of law was the view of Lon Fuller (1969). Fuller’s
main line of argument is that unless the entire legal system satisies certain
formal criteria, none of the rules or standards within it can count as
law.This view is discussed further in Chapter 8, where we will evaluate, in
the context of a discussion of international law, various views about the
necessary features of legal systems. But Fuller also thought that considerations
of the rule of law, or legality, should shape legal interpretation
(Rundle 2012, 168–74). Without trying to do justice to the complexity of
Fuller’s actual position, his writings do suggest this view about how to
igure out what the law is: when interpreting legal materials, one must do
so with an eye to making them consistent with the ideal of the rule of law.
This is in fact how Dworkin describes his moral reading, but his account
of the rule of law is expansive enough to encompass all moral considerations
necessary to reach a conclusion about the proper resolution of a
particular dispute (see, for example, Dworkin 2006, 170).
For there to be a distinctive view here, rule-of-law values must be
thought to make up a subset of potentially morally relevant factors – a
distinctly legal morality as Fuller conceived it. This would exclude, for
example, the liberal theory of social justice that is among the resources
of Dworkin’s moral reading. The richest defense of such a view can be
found in the work of David Dyzenhaus (2000, 2007). On such a restricted
moral reading, it would seem that preexisting law would not always be
suficiently determinate to resolve a particular dispute. Reading legal
materials in light of classical rule-of-law values such as publicity, prospectivity,
and equality before the law would clearly reduce indeterminacy as
compared to a positivist nonmoral reading, but it is hard to see that this
limited set of moral considerations would be suficient to allow us to say
that existing law provides an answer to all legal disputes. Dworkin’s more
expansive understanding of the values encompassed by the rule of law
NONPOSITIVISM 49
seems necessary for him to make good on his well-known claim that the
law always already provides a “right answer.”1
As discussed in the next chapter, the content of the ideal of the rule of
law is just as contested as the nature of law. It is therefore not helpful to
sort nonpositivist views according to whether they hold that the grounds
of law include the values of the rule of law or more than that. The point
is simply that within both the moral ilter and moral reading approaches,
a range of options exists about what kinds of moral considerations are
appropriately appealed to in legal interpretation. It seems on the face
of things that just as exclusive positivism best expresses the initial foundational
belief that law is about what is and not about what ought to be
or what is good, a maximalist nonpositivist view such as Dworkin’s best
expresses the initial foundational belief that law is a domain of value. On
the other hand, as Dyzenhaus (2000, 2007) explains, there are advantages
to the restricted moral reading, in that the moral considerations held to
be among the grounds of law emerge from the very idea of legal governance,
rather than a contestable moral and political theory. I will not here
attempt an internal evaluation of these options for nonpositivists.
It does seem, however, that the moral reading is straightforwardly
more plausible than the moral ilter, since once we believe that the content
of law is partly determined by moral considerations, it seems unmotivated
to recognize a discontinuity, such that morality is involved only
when the law would otherwise be very bad. After all, the moral reading,
as a reading of legal materials, preserves the institutional positivity of
law; it does not go all the way to the position that legal norms just are
morally ideal norms.
1
Jeremy Waldron’s (2008) argument that the grounds of law include rule-of-law values
characterizes those values as ones of procedure and form. There are, however, signs
that Waldron is ready to embrace an account of legality as expansive as Dworkin’s.
He writes that when judges disagree about how to resolve hard cases – when there
is no disagreement about the terms of any relevant statute or precedent – judges are
disagreeing about what the law (already) is, by way of disagreeing about the content
of the ideal of the rule of law (48–54). This suggests that any moral factor relevant to
iguring out how to resolve the dispute is a rule-of-law value.
WHAT MAKES LAW50
Justification and Obligation
Among those who approach the theory of law from the nonpositivist
foundational stance, some, such as Dworkin, assert the strong claim that
legal rights and duties are always real, that is moral, rights and duties.
The topic of the normativity of law is discussed in Chapter 7. For now,
the important point is that if we hold as a ixed point that legal rights
and duties are moral rights and duties, the moral reasoning in legal
interpretation may take on a distinctive structure. Whether it will do so
will depend on our moral theory of the duty to obey the law, however.
Thus if we believe that legal rights and duties are real rights and duties
just because we believe that all subjects have, in a relevant sense, consented
or promised to obey, legal interpretation will not be affected. If,
however, we hold, with Dworkin, that legal rights and duties are moral
rights and duties because the law can generate associative obligations
through its creation of a community of principle, then we will interpret
law, if possible, in such a way that it does create a community of principle.
I say more about Dworkin’s theory of associative obligations in
Chapter 7.
A weaker view is that law must be such as would justify its coercive
enforcement, whether or not it provides subjects with genuine obligation.
Thus in a recent article Philip Soper (2007) defends a moral ilter view,
according to which oficial directives that are so unjust that nothing could
justify their coercive enforcement cannot count as law. This defense of
“classical natural law” turns on the claim that law in its nature is not just
that which the state presents as legitimate directives that can legitimately
be enforced, but is that which can, in fact, be legitimately enforced. A
state can justiiably enforce unjust law, but not grossly unjust directives. If
the state does enforce grossly unjust directives, we are dealing with “acts
of violence, rather than laws” – as Aquinas put it.
At the other end of the spectrum, it would be possible for a nonpositivist
to believe, along with positivists, that the question of whether legal
rights and duties are real moral rights and duties is always open – that
NONPOSITIVISM 51
the search for the content of law, though it is always informed by moral
considerations, is not guided by the aim of inding real rights and duties.
Adjudication and Right Answers
As noted in Chapter 2, Dworkin’s legal theory includes the further
strong commitment that all normative considerations appropriately
taken into account in adjudication are ipso facto among the grounds
of law.This commitment is also not essential to nonpositivism as I have
deined it. This is already plain from the discussion of the possible view
that (only) formal rule-of-law values are among the grounds of law. On
such a view, though moral considerations are always among the grounds
of law, not all moral considerations appropriately taken into account in
adjudication are within law’s boundary. To take another example, suppose
that the moral considerations always within the grounds of law
relate only to matters of fundamental individual rights; considerations
of social or economic justice, or of general welfare, by contrast, are
excluded.Taking considerations of individual right into account in legal
interpretation may still leave gaps, or indeterminacy in the law.A judge
who resolves that indeterminacy by appeal to some theory of social
justice or social welfare may be acting appropriately given her institutional
role, on this possible view, but she would be in part making law,
not just applying it.
It is tempting to connect the adjudicatory view with the “right answer
thesis” – a yet further strong feature of Dworkin’s theory.This is because
it is natural to believe that there is always a right or best outcome to
a legal dispute, even if we believe that the law itself is not suficiently
determinate to guarantee that. The adjudicatory view thus seems to go
naturally together with the right answer thesis. But of course everything
depends on the content of the theory of adjudication, on whether
it has the resources to generate a unique best resolution to all disputes.
Adjudication is plausibly thought always to yield a best answer if we
assume that proper adjudication will help itself to unrestricted moral
WHAT MAKES LAW52
argument, even if that is not part of the determination of the content of
law. For moral theory most obviously has the resources necessary to yield
a single best result. That is not essential, however, since other theories
of adjudication could also have adequate resources. If there were such
a thing as “community morality,” Cardozo’s method of sociology would
do the trick as well. As would some theory that holds, for example, that
in cases of indeterminacy the defendant always wins. But the important
point is that we should not assume that all theories of adjudication have
the resources to guarantee a single best result.
The Moral Reading: Nonpositivist Puzzles
Dworkin’s commitments to the right answer thesis, to the adjudicatory
view of law, to the existence of a prima facie duty to obey the law, and to
an unrestricted but appropriately structured (given those commitments)
account of the moral considerations to be taken into account in legal
interpretation – all these commitments hang together to form a coherent
and well-motivated view. But it is important to emphasize that none of
these commitments is essential to nonpositivism as I have deined it. In
chapters that follow, I use “nonpositivism” in the minimal sense of a commitment
to either the moral reading or the moral ilter account of legal
interpretation – and since it seems the better view, I will normally have in
mind the moral reading.
Though it has roots in common-law legal reasoning (Perry 1987), the
moral reading receives its best elaboration in Dworkin’s discussion of
legal interpretation in Law’s Empire. We are to suppose a three-stage
process. In the preinterpretive stage, we start with central cases, cases
that everyone agrees are instances of valid law. This irst stage involves
no moral reasoning; preinterpretive law is a matter of brute fact. In the
interpretive stage, moral considerations are brought to bear to ind the
general principles that best justify the past legal practice that produced
those legal materials; legal rights and duties are those that low from the
best set of general principles that can be said to it or explain past legal
NONPOSITIVISM 53
practice. In the postinterpretive stage, we may, on relection, conclude
that some of the legal materials identiied in the preinterpretive stage
are not legal materials at all; from the point of view of our interpretation,
they are so far in tension with the best interpretation that they must be
banished from law as mistakes. (Recall Blackstone: “For if it be found
that the former decision is manifestly absurd or unjust, it is declared, not
that such a sentence was bad law; but that it was not law; that is, that
it is not the established custom of the realm, as has been erroneously
determined.”)
One important question is how far a good interpretation can go in
declaring preinterpretive material to be mistakes.This is usually posed in
terms of Dworkin’s terminology of “it” and “justiication.”The interpretive
stage aims to justify past legal practice, but for it to do that, it must
after all it past legal practice; we would not be interpreting our law at
all if we simply discarded all prior materials as mistaken. But how much
it is required? And to what extent can judgments of a required degree
of it be separated from the judgments of justiication? These are important
and fascinating questions, but a full discussion would take us too far
aield here.2
Let us simply assume that some plausible account of the twin
tasks of it and justiication can be given that retains the dimension of it
as a genuinely independent criterion.
A somewhat different puzzle does need to be discussed here. It concerns
the preinterpretive stage, the initial identiication of legal materials
that interpretation must (to some degree) it. For this stage, the
moral reading appears to require what we could call a provisional rule
of recognition (Raz 1986). Provisional only because what is initially
treated as a data point may of course later be expelled as a mistake. But
the entire process depends on our ability to ind considerable initial
agreement about clear cases of valid law, which reveals at least some
implicit convergence on criteria of validity. What’s striking here is that
2
See the extended debate between Dworkin (1985, 146–77; 1986, 65–8, 225–75) and
Stanley Fish (1989, 87–119, 356–71).
WHAT MAKES LAW54
nonpositivist legal interpretation takes as its object a matter of brute
social fact; what we are interpreting is just what people agree, for whatever
reason, is part of law. One way to criticize the view, understood in
this way, would be to say that the moral reading takes legal materials
that could have any content at all and pretties them up (Stavropoulos
2014 forthcoming). Why should we make the best of something that, so
far as we know, initially has nothing going for it at all? The conviction
that law is a location of value and a source of practical reasons seems
to require more than that we make something morally indifferent look
good. It would seem to require that we have some reason to think that
there is value, or at least potential value, in the materials that interpretation
must it.
Nowthesituationisnotasstarkasthisobjectionsuggests.Procedurally
valid statutory enactments are, in the moral reading, not just part of the
brute facts to be it; neither need the legal priority such sources have
over judge-made law in common-law systems be treated as brute fact.
For the practice of allowing representative institutions the inal say over
legal rights and duties (within limits) can itself be justiied by democratic
theory.
But suppose no such story were available. Suppose that legislation is
made by lawmakers who inherit their roles; they have neither any claim
to be representative of the people nor any special moral and political
expertise. We may even suppose further that, substantively, the content
of this legislation has systematically favored the interests of the hereditary
ruling class.The idea that, when interpreting the law of this jurisdiction,
interpreters should try to show it in its best light may seem entirely
unmotivated. This relates to the problem of “evil law” for nonpositivists.
If nothing good can be said about some legal tradition in the irst place,
the moral reading, the attempt to show that tradition in its best light,
seems inappropriate. The natural response for the nonpositivist here is
simply to embrace the conclusion that only certain bodies of preinterpretive
legal material can generate legal rights and duties.Any collection of
legal materials that lacks any seed of value, either in the manner of its
NONPOSITIVISM 55
making or its content, will simply be rejected as an appropriate site for
legal interpretation. We could say that the moral reading must apply a
moral ilter to the preinterpretive legal materials, considered as a whole. I
believe this is the most natural way to understand the moral reading (see
Dworkin 1986, 101–13), and generally have this picture in mind in the
chapters that follow.
In his latest writings, Dworkin (2006 34–5; 2011, 400–15) embraced
quite a different approach.3
Nicos Stavropoulos (2014 forthcoming) has
introduced the terminology of “hybrid” and “nonhybrid” interpretivism
that is helpful for making the contrast. The moral reading as I have so
far laid it out corresponds to hybrid interpretivism. It treats legislative
enactments or decisions of courts provisionally as sources of law just
because everyone does (preinterpretive stage) – so long as they pass
through the moral ilter. At the interpretive and postinterpretive stages,
such legal materials may be conirmed as part of the law, or not. On the
reformulated view, it’s a moral inquiry right from the start. The nonhybrid
interpretivist approach treats legal materials as relevant to an investigation
of what the law is not because everyone does, but because there
is a good reason (if there is), grounded in political morality, to treat the
past political decisions that produced those materials as sources of real
rights and obligations. From this perspective, what we are looking for is
not the content of some system of norms whose relation to morality is
so far just that it has passed through a minimal moral ilter; rather, when
we look for the content of law, we are engaging in a moral inquiry at all
times, seeking to determine what effect past political decisions of our
political coercive order have on our moral rights and duties.
The irst thing to notice about nonhybrid interpretivism is that it
presupposes something that, as I have said, is not necessarily part of a
nonpositivist view of the grounds of law, viz. that the outcomes of legal
3
Stavropoulos (2014 forthcoming) and Greenberg 2011 hold that there has been no
change; rather Dworkin has long been misunderstood. I disagree with that interpretation
as, apparently, did Dworkin himself – see Dworkin 2011, 402.
WHAT MAKES LAW56
interpretation are statements of real, moral, rights and duties and that
this end point structures legal interpretation. Dworkin’s reformulated
view puts this position at the center of this theory of law. Legal interpretation
just is a matter of iguring out what our moral rights and duties are,
insofar as they low from past political decisions.
So nonhybrid interpretivism is not a possible general account of the
moral reading, since it depends on commitments that are not essential
to the moral reading. But it is worth asking which view is preferable for
those who accept that legal rights and duties are always real moral rights
and duties.In one sense,nonhybrid interpretivism seems a clear improvement
over Dworkin’s original account of legal interpretation, since if
we start with brute facts about what people think the law is, but insist
that we end up with real rights and duties, it may seem obvious that the
moral argument will always trump the brute facts.Though in the original
account the dimension of it is given moral signiicance through the value
of integrity, or consistency in principle through time and across people,
the morally salient kind of it on that account is what has happened to
people, rather than what people (even judges) have thought the law was
(Dworkin 1986, 248, 284–5). This suggests that the reformulation was
inevitable, as the original account left preinterpretive legal material with
very little force in an interpretation driven by the inal cause of determining
real moral rights and duties.
On the other hand, if we drop any anchoring of the content of legal
rights and duties in brute facts about people’s views about the content
of law, a new problem is immediately presented. Plausible moral arguments
could be made that all kinds of past political (or social, for that
matter) decisions and practices generate rights and obligations, but not
all of them would ordinarily be identiied as legal. Dworkin’s solution to
this problem is to say that legal rights and obligations are those that are
appropriately adjudicated by courts (2011, 405–13). Given his commitment
to the adjudicatory view of law, this is not a surprise, but it does
come at enormous cost. Where actions of the executive or legislative
branches of government are not appropriately reviewable by courts, they
NONPOSITIVISM 57
are by deinition not subject to law, either domestic or international.4
On
the face of it, this seems an extremely high price to pay, for reasons I will
go into in Chapter 7. It also seems to be a dogmatic and unmotivated
piece of taxonomy, and for that reason inconsistent with Dworkin’s idea
(discussed in Chapter 6) that law is an interpretive concept.
More important, taxonomy is usually uninteresting, as Dworkin himself
reminds us (2006, 4–5). Dworkin is not troubled by the fact that
where some presidential decision is not properly reviewable by a court
we must say that no legal obligations constrain the decision, since we can
of course acknowledge the political obligations that apply (2011, 412–13).
But if this is just a matter of labeling, the very natural next step is to say
that it doesn’t really matter which of the obligations people have in virtue
of past political decisions get the label “legal” and which do not. In other
words, I believe that this latest version of Dworkin’s theory leads inevitably
to the eliminativist position I discuss in Chapter 6. There isn’t any
such thing as law as a distinct normative system.There are just actions by
various kinds of institutions that have moral consequences for the members
of that political community.5
In any event, when referring to Dworkin’s version of the moral reading
in the remainder of this book, I have in mind his original account of
legal interpretation from Law’s Empire with all of its three stages.
Mistakes and Scorer’s Discretion
Though there is of course much more that would need to be said to give
a full account of the moral reading, I will end with just one further apparent
puzzle. On the moral reading, judges can make legal mistakes by way
4
Dworkin does not deny international law the status of law on account of the lack
of international courts of general and compulsory jurisdiction. As he makes plain in
(2013), his position is that legal rights and duties are those appropriately adjudicated
by courts actual or hypothetical.
5
These remarks apply equally to the positions laid out in Greenberg (2011) and
Stavropoulos (2014 forthcoming).
WHAT MAKES LAW58
of making moral mistakes. Thus if the death penalty really is, morally
speaking, cruel, the moral reading requires us to conclude that the entire
legal apparatus of the death penalty in the United States is based on a
legal mistake; that it is, in fact, unlawful. This can seem to be an absurd
result,as positivists sometimes insist it is (Marmor 2011,91–2).But in fact
the advantage of positivism over the moral reading in this connection is
at best a matter of degree.
Any theory of law must account for the possibility of judicial error,
and the possibility that judicial error may generate systematic application
of that legal error by other judges and law enforcement oficials.The
alternative is a view of law according to which what the highest court
says is law, is law. Hart criticized such a view by comparing it to a game
of scorer’s discretion – the rules of the game reduce to the decisions the
scorer makes as there is no independent standard against which to criticize
the decisions as mistaken (1994, 142–7).
A decision of the highest court is authoritative for the enforcement
branches and lower courts until overruled or changed by statute. But
though the application of the decision by other oficials is in accordance
with law, it may still have been a legal mistake. If what the highest court
declares to be law is not only appropriately followed by lower courts and
enforcement oficials,but deinitive of the content of the law,then there is
no law, but simply a grant of absolute discretion to highest courts.
Both Hart and Dworkin recognize the importance of the possibility
of legal mistakes in any account of law as a genuine normative system
that effectively constrains legal decision makers (Dworkin 1978,
121; Hart 1994, 141–7). Both also, within the terms of their theories, can
explain how what was once a mistake may become a source of law in the
common law through the operation of the doctrine of precedent. This is
familiar enough in the United States, where nominees for the Supreme
Court frequently tell the Senate Judiciary Committee that though the
decision in Roe v. Wade was a mistake, they would not overturn it given
its entrenchment and consequent signiicant precedential weight. Hart
(1994, 144) also notes the possibility that if we ind ourselves concluding
NONPOSITIVISM 59
that most judicial decisions, most of the time, are mistaken, then the legal
order – the complex set of rules that make up the content of the law in
force on his account – would cease to be effective.
If there is to be a freestanding legal order that provides answers to
legal questions, as opposed to an accepted practice of allowing judges
to create legal rules without constraint, then any legal theory, positivist
or nonpositivist, must accept into its account the legitimate enforcement
of legally mistaken decisions. If there is an objection to nonpositivism
here, it must be grounded on the assumption that if moral considerations
are among the grounds of law, there is more scope for the institutionalization
of judicial error than if they are not. But this assumption seems
unwarranted, or at least in need of defense. The examples that make
nonpositivism look silly – all those executions in Texas are contrary to
law – naturally relate to legal issues with high moral stakes and prevailing
moral disagreement. But there can be disagreement about the best way
to understand a legal rule that does not turn on moral disagreement –
as positivists in other contexts are eager to insist, in order to rebut the
charge that on their view it must be true that everyone agrees what the
content of the law is.
Thus take Lord Mansield’s argument, in a 1765 contract case, that
the ancient notion about the want of consideration was for the
sake of evidence only: for when it is reduced into writing, as in covenants,
specialities, bonds, etc., there was no objection to the want of
consideration.6
This is an excellent argument against the idea that the doctrine of consideration
called for a genuine quid pro quo.Sadly,Mansield was reversed by
the House of Lords in 1778,7
and the requirement of an exchange remains
part of the law in most of the common-law world.8
The point here is that
6
Pillans and Rose v.Van Mierop & Hopkins (1765) 3 Burr. 1663, 1670.
7
Rann v. Hughes (1778) 4 Bro PC 27, 7 TR 350.
8
Though outside the United States it has been largely undermined through judicial
acceptance of the adequacy of a merely nominal exchange. Chappell & Co Ltd v Nestle
Co Ltd [1960] AC 87.
WHAT MAKES LAW60
this issue has enormous signiicance for legal practice; whether Mansield
and his brethren who joined him were right or wrong, lower courts were
bound to follow his decision until overruled. Suppose he was mistaken
about the proper understanding of the doctrine of consideration but he
was not overruled.The common-law world would have rightly enforced a
legal error – until the passage of time converted the error into good law,
through the doctrine of precedent. Nothing in this example requires that
Mansield employed a moral reading of the existing legal materials.I venture
that it is hard to determine in the abstract whether the moral reading
is more or less likely to lead to the conclusion that“a great deal ...of what
people take to be the law ... in a given legal system is legally mistaken”
(Marmor 2011, 91–2) than a positivist approach to legal interpretation.
A lot would depend on the content of the legal materials that are being
read, and also the kinds of moral considerations that a particular nonpositivist
view holds should be taken into account. I see no reason to ind
here a general objection to the moral reading.
Progress?
The disagreements between and among positivists and nonpositivists are
stark. One main question in this book is whether there is any plausible
way to resolve them. But it is important to start with a different question:
Why have philosophers over the past two hundred years or so thought
that this disagreement mattered? Why have they not been happy to allow
what Aristotle appeared to take for granted, that there simply are different
ways of referring to the category of law, and that this is a verbal
dispute of only rhetorical signiicance?
Explaining why that is not so is our irst main task. It will be useful
to start by discussing what might be at stake in disagreements about the
content of some other politically and morally important concepts.