E s s ay 2

Posi tivism and the S e p aration

of Law and Morals

In this article I shall discuss and attempt to defend a view

which Mr Justice Holmes, among others , held, and for which
he and they have been much criticized. But I wish first to say
why I think that Holmes , whatever the vicissitudes of his
American reputation may b e , will always remain for English­
men a heroic figure in jurisprudence . This will be so because
he magically combined two qualities : one of them is imagin­
ative power, which English legal thinking has often lacked ;
the other i s clarity , which English legal thinking usually
possesses . The English lawyer who turns to read Holmes is
made to see that what he had taken to be settle d and stable
is really always on the move . To make this discovery with
Holmes is to be with a guide whose words may leave you un­
convinced, sometimes even repelled , but never mystified. Like
our own Austin, with whom Holmes shared many ideals and
thoughts , Holmes was sometimes clearly wrong; but again
like Austin , when this was so he was always wrong clearly.
This surely is a sovereign virtue in j urisprudence . Clarity I
know is said not to be enough ; this may b e true, but there are
still questions in jurisprudence where the issues are confused
because they are discussed in a style which Holmes would
have spumed for its obscurity . Perhaps this is inevitable : juris­
prudence trembles so uncertainly on the margin of many
subjects that there will always be a nee d for some one , in
Bentham's phrase, 'to pluck the mask of Mystery ' fro m its
face . 1 This is true , to a pre-eminent degree , of the subject of
this article. Contemporary voices tell us we must recognize
something obscured by the legal 'positivists ' whose day is now
over : that there is a 'point of intersection between law and
morals'/ or that what is and what ough t to be are somehow
1 Bentham, A Fragment on Government, in I Works 22 1 , 235 (B owring edn.
1 83 8-43) (preface, 4 1 st para.). All references hereafter to B en th am' s Works are
to this edition.
• D'Entreves, Natural Law 1 1 6 ( 2nd edn. 1 9 5 2 ) .
50 GENERAL THEO RY

indissolub ly fused o r ins ep arab le , ) though the po s i ti vists

denied it. What do these phrases mean ? Or r ath e r which of
the many thi ngs that they could mean , do they mean ? Which
of them do 'p ositivists ' deny and why is it wrong to do s o ?

I shall present the subj ect as p art of the his tory of an idea. At
the c l o s e of the eighteenth c ent ur y and the beginning o f the
nin eteenth the most e arn e st thinkers in England about legal
and social pro b l e m s and the architects o f gre at re forms were
the gr e at Utilitarians. Two of th e m , Bentham and Austin ,
constantly insisted o n the need to distinguish , firmly and with
the maximum of clarity , law as it i s from law as it o u ght to
b e . This the me haunts t h ei r work, and they condemned the
natural-law thinkers p re c is e l y because they had blurred this
ap par ent l y s im p le b ut vital distinction . By contrast , at th e
present time in this country and to a lesser extent in England,
thi s sep aration between law and mo r al s is held to b e super­
ficial and wrong. S o m e c ri ti c s h ave tho u gh t that it blinds men
to the true nature of l aw and its roots in so ciiiI life .4 Others
have thought i t not o nl y intelle ctually mi s lead in g but c o r ­
ru p ti ng in practice , at its worst apt to weaken resistance to
s tate tyranny or ab s ol uti sm , s and at i ts best apt to b r ing l aw
into disrespect. The now p ej orative name 'Legal Positivism',
3 Fuller, The Law in Quest of Itself 1 2 ( 1 940 ) ; Brech, 'The Myth o f I s and
Ought', 54 Harv. L. Rev. 8 1 1 ( 1 94 1 ) , Fuller, 'Human Purpose and Natural Law',
5 3 J. Philos. 6 9 7 ( 1 953).
4 See Friedmann, Legal Theory 1 54, 294-95 (3rd edn. 1 953). Friedmann also
sayS of Austin that 'by his sharp distinction between the science of legislation and
the science of law', he 'inaugurated an era of legal positivism and self-sufficiency
which enabled the rising national State to assert its authority undisturbed by jur­
istic doubts'. Ibid. at 4 1 6. Yet, 'the existence of a highly organised State which
claimed sovereignty and unconditional obedience of the citizen ' is said to be 'the
political condition which makes analytical positivism possible'. Ibid. at 1 63. There
is therefore some difficulty in determining which, in this account, is to be hen and
which egg (analytical positivism or political condition) . Apart from this there
seems to be little evidence that any national State rising in or after 1 832 (when
the Province of jurispnt d en c e Determined was first published) was enabled to
assert its authority by Austin's work or 'the era of legal positivism' which he 'in­
augurated'.
S Se e Radbruch, 'Die Emeuerung des Rechts', 2 Die Wa ndLUng 8 (Germany
} 94 7) ; Radb ru ch , 'Gesetzliches Unrecht und 'Obergesetzliches Recht', I S uddeu tsche
Juristen-Zeitung 1 0 5 (Germany 1 946) (reprinted in Radbruch, Rechtsphilosophie
34 7 ( 4th edn. 1 9 50» . Radbruch 's views are discussed at 7 2-8 below.
 POSITIVISM, LAW, AND MO RALS 51

like m o s t termswhich are used as missiles in intellectual

b attles, has come to stand for a baffling multitude o f di ffer­
ent sins . One of them is the sin , real or alleged, o f insisting, as
Aus t i n and Bentham did , on the separation o f law as it is and
law as it ought to be.
How then has this reversal o f the wheel c o me ab out? What
are the theoretical errors in this distinction? Have the practi­
c al consequences of s t re s sing the distinction, as Bentham and
Austin did, been bad ? Should we now reject it or keep i t ? In
considering these questions we should recall the social phil­
o s op hy which went along with the Utilitarians ' insistence on
this distinction . They stood firmly but on their own utilitarian
ground for all the principles o f lib e rali sm in law and govern­
ment. No one has ever combined , with such even 'minded
sanity as the Utilitarians, the pas sion for re form with respect
fo r law together with due recognition of the need to c ontrol
the abuse of power even when power is in the hands of re­
formers. One by one in Bentham 's works you can id e n ti fy the
elements of the Rech tsstaat and all the princ ipl es for the
defence o f which the t e rmi nology of natural law has in our
day been received. Here are liberty of speech, and of p ress ,
the ri ght o f association ,6 the need that laws should be pub ­
lished and made widely known before they are enfo rce d , 7
the need to control administrative agencies,S the insistence
that there should. be n o criminal li ability without fault ,9 and
the importance of the principle of legality, nulla poena sine
lege. 1 0 Some, I know, find the political and m<;>ral insight o f
the Utilitarians a very simple one, but we should not mistake
this simplicity for superficiality , nor forget how favourably
their simplicities compare with the profundities of other
thinkers . Take only one example : Bentham on slavery . He
says the question at issue is not whether those who are he ld
6 Bentham, A Fragment on Government, in I Works 2 2 1 , 230 (preface, 1 6th
para. ) ; Bentham, Principles of Penal Law , in I Works 3 65 , 5 74-5, 5 16-8 (pt. III,
e. XXI, 8th para. , 1 2th para.).
7 Bentham, Of Promulgation of the Laws, in I Works 155 ; Bentham, Principles
of the Civil Code , in I Works 297 , 323 (pt. I, c. XVII, 2nd para.) ; Bentham, A
Fragment on Government, in I Works 22 1 , 233 n . [m] (preface, 35th para.) .
• Bentham, Principles of Penal Laws, i n I Works 3 65, 5 1 6 (pt. III, c. XX I,
1 0th para. , 1 1 th para. ).
• Bentham, Principles of Morals and Legislation, in I Works I, 84 (c. XIII).
1 0 Bentham, Anarchical Fallacies, in 2 Works 489, 5 1 1 - 1 2 (art. VIII) ; Bentham,
Principles of Morals and Legislation, in I Works I, 1 44 (c. XIX, 1 1 th para.).
52 GENERAL THEORY

as slaves c an reason , but simply wh � ther they suffer. l l Does

this not compare well with the discussion of the question in
terms of whether or not there are some men whom Nature
has fitted only to be the living instruments o f others ? We owe
it to Bentham more than anyone else that we have stopped
dis cussing this and similar questions of social policy in that
form.
So Bentham and Austin were not dry analysts fiddling with
verbal distinctions while cities burned, but were the vanguard
of a movement which laboured with passionate intensity and
much success to bring about a better society and better laws .
Why then did they insist on the separation o f law as it is and
law as it ought to b e ? What did they mean ? Let us first see
what they said. Austin formulated the doctrine :
The existence of law is one thing; its merit or demerit is another.
Whether it be or be not is one enquiry ; whether it be or be not con­
formable to an assumed standard, is a different enquiry. A law, which
actually exists, is a law , though we happen to dislike it, or though it
vary from the text , by which we regulate our approbation and disappro­
bation. This truth , when formally announced as an abstract proposition,
is so simple and glaring that it seems idle to insist upon it. But simple
and glaring as it is, when enunciated in abstract expressions the enu­
meration of the instances in which it has been forgotten would fill a
volume.
Sir William Blackstone, for example , says in his 'Commentaries', that
the laws of God are superior in obligation to all other laws ; that no
human laws should be suffered to contradict them; that human laws are
of no validity if contrary to them ; and that all valid laws derive their
force from that Divine original.
Now, he may mean that all human laws ought to conform to the
Divine laws. If this be his meaning, I assent to it without hesitation . . . .
Perhaps, again, he means that human lawgivers are themselves obliged
by the Divine laws to fashion the laws which they impose by that ulti­
mate standard, b ecause if they do not, God will punish them. To this
also I entirely assent . . . .
But the meaning of this passage of Blackstone, if it has a meaning,
seems rather to be this : that no human law which conflicts with the
Divine law is obligatory or binding; in o ther words, that no human
law which conflicts with the Divine law is a la w . . . . 12

Austin 's protest against blurring the distinction between

what law is and what it ought to be is quite general : it is a
11
Ibid. at 1 4 2 n. § ( c. XIX, 4th para. n. § ) .
12
Aus tin, The Province of Jurisprudence Determined 1 84-5 ( Library of Ideas
edn. 1 954).
 POSITIVISM, LAW, AND MORALS 53

mistake , whatever our standard o f what ought t o b e , what­

ever 'the text by which we regulate our approbation or dis­
approbation'. His examples , however, are always a con fusion
between law as it is and law as morality would require it to
be. For him, it must he remembered, the fundamental pri n ­
ciples of morality were God's commands , to which utility was
an 'index ' : besides this there was the actual accepted morality
of a social group or 'positive ' morality.
Bentham insisted on this distinction without characterizing
morality by reference to God b ut only, of course , by re fer­
ence to the principles o f utility . Both thinkers' prime reason
for this insistence was to enable men to see steadily the pre­
cise issues posed by the existence of morally bad laws , and to
understand the specific character o f the authority of a legal
order. Bentham 's general recipe for life under the government
of laws was simple : it was 'to obey punctually ; to censure
freely 'P But Bentham was especially aware , as an an xious
spectator o f the French revolution, that this was not enough :
the time might come in any society when the law's commands
were so evil that the question of resistance had to be faced ,
and it was then essential that the issues at stake at this point
should neither be oversimplified nor obscured. 1 4 Yet this was
precisely what the confusion between law and m orals had
done , and Bentham found that the confusion had spread
symmetrically in two different directions . On the one hand
Bentham had in mind the anarchist who argues thus : 'This
ought not to be the law, therefore it is not and I am free not
merely to censure but to disregard it . ' On the other hand he
thought o f the reactionary who argues : 'This is the law, there­
fore it is what it ought to b e ' , and thus stifles criticism at its
birth. Both errors , Bentham th ou ght, were to be found in
Blackstone : there was his incautious statement that human
laws were invalid if contrary to the law o f God, 1 5 and 'that
13 Bentham, A Fragment on Governm ent, in I Works 2 2 1 , 230 (preface, 1 6th
para. ).
1 4 See Bentham, 'Principles of Legislation', in The Theory of Legislatio n I, 6 5
n. * (Ogden edn. 1 9 3 1 ) (c. XII, 2d para. n. * ) . 'Here we touch upon the m ost dif­
ficult of questions. If the law is not what it ought to be ; if it openly combats the
principle of utility ; ought we to obey it? Ought we ' .. · · i olate it? Ought we to re­
main neuter between the law which commands an evil, and morality which forbids
it?' See also Bentham, A Fragment on Go vernment, in I Works 22 1 , 2 8 7 - 8 (c. IV,
20th-25th paras. ).
1 5 I Blackstone, Commentaries * 4 1 . Bentham criticized 'this dangerous
54 GENE RAL THEORY

spirit of obsequious quietism that s eems constitutional in o ur

Author' which 'will scarce ever let him recognise a difference '
between what is and what ought to be . ' 6 This indeed was for
Bentham the occupational disease of lawyers : ' [I ] n the eyes
of lawyers - not to speak of their dupes - that is to say, as
yet, the generality of non-lawyers - the is and the ought to
be . . were one and indivisible. ' 1 7 There are therefore two
.

dangers between which insistence on this distinction will help

us to steer : the danger that law and its authority may be dis­

solved in man 's conceptions of what law ought to be and the

danger that the existing law may supplant morality as a final
test of conduct and so escape criticism.
In view of later criticisms it is also imp ortant to distinguish
several things that the Utilitarians did not mean by insisting
on their separation of law and morals . They certainly accepted
many of the things that might b e called 'the intersection of
law and morals'. First , they never denied that, as a matter of
historical fact, the development o f legal systems had been
powerfully influenced by moral opinion, and, conversely , that
moral standards had b een profoundly influenced by law, so
that the content of many legal rules mirrored moral rules or
principles . It is not in fact always easy to trace this historical
causal connection, but Bentham was certainly ready to admit
its existence ; so too Austin spoke of the 'frequent coinci­
dence " s of positive law and morality and attributed the con­
fusion of what law is with what law ought to be t o this very
fact.
Secondly , neither Bentham nor his followers denied that
by explicit legal provisions moral principles might at differ­
ent points be brought into a legal system and form p art of
its rules , or that courts might be legally b ound to decide .in
maxim', saying 'the natural tendency of such a doctrine is to impel a man, by the
force of conscience, to rise up in arms against any law whatever that he happens
not to like '. Bentham, A Fragmen t on Go vernment, in I Works 22 1 , 287 (c. IV,
1 9th para. ). See also Bentham, A Co mmen t on the Co mmentaries 49 ( 1 928)
(c. III). For an expression of a fear lest anarchy result from such a doctrine,
combined with a recognition that resistance may be justified on grounds of utility,
see Austin, op. cit. n. 1 2 supra , at 1 86. .
16 Bentham, A Fragment on Governmen t , in I Works 221 , 294 (c. V, 1 0th
para. ).
1 7 Bentham, A Co mmentary on Humphreys' R eal Property Code, in 5 Works
3 89.
II Austin, op. cit. n. 1 2 supra, at 1 62.
 POSITIVISM, LAW, AND MORALS 55

accordance with what they thought just or b est. Bentham in­

deed recognized, as Austin did not, that even the supreme
legislative power might be subjected to legal restraints by a
constitution 1 9 and would not have denied that moral prin­
ciples, like those of the Fifth Amendment, might form the
content of such legal constitutional restraints . Austin differed
in thinking that restraints on the supreme legislative power
could not have the force o f law, but would remain merely
political or moral checks ;20 but of course he would have rec­
ognized that a statute , for example , might confer a delegated
legislative power and restrict the area of its exercise by re fer·
ence to moral principles .
What both Bentham and Austin were anxious to assert
were the following two simple things : first, in the absence of
an expressed constitutional or legal provision, it could not
follow from the mere fact that a rule violated standards of
morality that it was not a rule of law ; and, conversely , it could
not follow from the mere fact that a rule was morally desir­
able that it was a rule of law.
The history of this simple doctrine in the nineteenth cen­
tury is too long and too intricate to trace here . Let me sum­
marize it by saying that after it was propounded to the world
by Austin it dominated English jurisprudence and constitutes
p art of the framework of most of those curiously English and
perhaps unsatisfactory productions - the omnib us surveys o f
the whole field of jurisprudence. A succession of these were
published after a full text of Austin's lectures finally appeared
in 1 8 63 . In each of them the utilitarian separation of law and
morals is treated as something that enables lawyers to att ai n a
new clarity . Austin was said by one o f pis English successors ,
Amos, 'to have delivered the law from the dead b ody of
morality that still clung t o it ' ;2 1 and even Maine, who was
I ' Bentham, A Fragment on Go vernment , in I Works 22 1 , 289-90 (c. IV,
33rd-34th paras.).
2G
See Austin, op. cit. rio 1 2 supra, at 2 3 l .
2 1 Amos, The Science of Law 4 (5th edn. 1 88 1 ) . See also Markby , Elements
of Law 4-5 ( 5 th edn. 1 896): 'Austin, by establishing the distinction between posi·
tive law and morals, not only laid the foundation for a science of law, but cleared
the conception of law • . .of a number of pernicious consequences to which it
• . •

had been supposed to lead. Positive laws, as Austin has shown, must be legally
binding, and yet a law may be unjust . • .He has admitted that law itself may be
•

immoral, in which case it may be our moral duty to disobey it Cf. Holland,
• • • . '

Jurisprudence 1 -20 ( 1 880).

56 GENERAL THEORY

critical of Austin at many points , did not question this part

of his doctrine. In the United States men like N. St. John
Green ,2 2 Gray , and Holmes considered that insistence on this
distinction had enabled the understanding o f law as a means
of social control to get o ff to a fruitful new start ; they wel­
comed it both as self-evident and as illuminating - as a
revealing tautology . This distinction is, o f course, one o f the
main themes of Holmes 's most famous essay 'The Path of
the Law' ,23 b ut the place it had in the estimation of these
American writers is best seen in what Gray wrote at the tum
of the century in The Nature and Sources o/ the Law. He said :
The great gain in its fundamental conceptions which Jurisprudence
made during the last century was the recognition of the truth that the
Law of a State . . . is not an ideal , but something which actually exists .
. . . [I] t is not that which ought to be , but that which is. To fix this
definitely in the J urisprudence of the Common Law, is the feat that
Austin accomplished. 24

II

So much for the doctrine in the heyday o f its success. Let us

tum now to some of the criticisms . Undoubtedly , when
Bentham and Austin insisted on the distinction between law
as it is and as it ought to be, they had in mind particular laws ,
the meanings o f which were clear and so not in dispute, and
they were concerned to argue that such laws , even if morally
outrageous , were still laws. It is, however, necessary , in con­
sidering the criticisms which later developed, to consider m o re
than those criticisms which were directed to this particular
point if we are to get at the root o f the dissatisfaction felt ;
we must also take account of the objection that, even if what
the Utilitarians said on this particular point were true, their
insistence on it, in a terminology suggesting a general cleavage
between what is and ought to be law, obscured the fact that
at other points there is an essential point o f contact b etween
the two . S o in what follows I c onsider not only criticisms of
the particular point which the Utilitarians had in mind, but
also the claim that an essential connection b etween law and
2 2 See Green, Book Review, 6 Am. L. Rev. 5 7 , 6 1 ( 1 8 7 1 ) (reprinted in Green,
Essays and No tes on the Law of Tort and Crime 3 1 , 35 ( 1 93 3 » .
2 3 1 0 Harv. L . Rev. 45 7 ( 1 89 7 ) .
2 4 Gray , The Nature and Sources of t h e Law 9 4 ( 1 st e dn. 1 909) ( § 2 1 3 ) .
 POSITIVI S M , LAW , AND MO RALS 57

morals emerges i f we examine how laws , the m e anings of

which are in dispute , are interprete d and applied in con crete
cases ; and that this connection emerges again if we widen our
p o in t o f view and as k , n o t whether every p articular rule of
law must satisfy a moral minimum in order to b e a law, but
whe ther a system of rules which alt o gether failed t o d o this
could be a legal syste m .
There is , howeve r , o n e maj o r initial complexity by which
criticism has been much confuse d . We must remember th at
the Utilitarians combined with their insistence on the s e p ar­
ation of law and morals two other e qually famous b u t di stinct
do ctrines . One was the important truth that a p urely analyti­
cal study o f legal con cepts , a study o f the meaning of the
dis tin ctive vocab ulary of the l aw, was as vital to o ur und er­
stan ding o f the nature of law as histori c al or s o cio l o gical
studies , though of co urs e it c o uld not supplant the m . The
other doctrine was the famous imperative theory of law -
that l aw is essentially a command.
These three d o ctrines constitute the utilitarian tradition in
jurisp rudence ; yet they are distinct do ctrines . It is possible to
endors e the separati on between law and m orals and to value
an alytical inquiries into the meaning of legal concep t s and y et
think it wrong to conceive o f law as essentially a command.
One source o f gre at confusion in the criticism o f the s e p ara­
tion of law and morals was the belie f that the falsity of any
one of these three do ctrines in the u tilitarian traditio n sh owed
the o ther two t o be fals e ; what was worse was the failure to
s e e that there were three quite sep arate do ctrines i n th i s tra­
dition . The indi s criminate use of the lab el 'positivi s m ' to des­
ign ate ambiguously each one o f the s e three sep arate d o ctrines
( toge ther with s o me o thers which the Utilitarians never pro­
fes s e d ) has perhaps confused the issue more than any other
single factor.25 Some o f the early Ameri can cri tics of the
2 5 It may help to identify five ( there may be more) meanings of 'positivism '
bandied about in contemporary jurisprudence :
( 1 ) the contention that laws are c·ommands of human beings ; see 5 8-6 2 infra ;
( 2 ) the contention that there i� no necessary connection between law and
m orals or law as it is and ough t to b e ; see 50 6 supra ;
( 3 ) the contention that the analysis (or study of the meaning) of legal concepts
is (a) w�rth pursuing and (b) to be distinguished from historical inquiries into the
causes oi: origins of laws, from sociological inquiries into the relation of law and
other social phenomena, and from the criticism or appraisal of law whether in
terms of morals, social aims, 'functions', or otherwise ; see 64-6 infra ;
58 GENERAL THEORY

Austinian doctrine were , however, admirably clear on just this

matter. Gray , for example , added at the end of the tribute to
Austin , which I have already quoted, the words : 'He may have
been wrong in treating the Law of the State as b eing the com­
mand of the sovereign ' /6 and he touched shrewdly on many
points where the command theory is defective. But other
critics have been less clearheaded, and have thought that the
inadequacies o f the command theory which gradually came
to light were sufficient to demonstrate the falsity of the sep­
aration of law and morals.
This was a mistake, but a natural one. To s ee how natural
it was we must look a little more closely at the command
idea. The famous theory that law is a command was a p art of
a wider and more ambitious claim. Austin said that the notion

of a command was 'the k ey to the sciences of jurisprudence

and morals ', 2 7 and contemporary attemp ts to elucidate moral
judgments in terms of 'imperative ' or 'prescriptive' utterances
echo this ambitious claim. But the command theory, viewed
as an e ffort to identify even the quintessence of law, let alone
the quintessence of morals , seems breathtaking in its simplicity
and quite inadequate. There is much, even in the simplest
legal system, that is distorted if p resented as a command. Yet
the Utilitarians thought that the essence of a legal system
could be conveyed if the notion of a command were sup­
plemented by that of a habit of obedience. The simple scheme
was this : What is a command? It is simply an expression by
one person of the desire that another person should do pr
abstain from some action, accompanied by a threat of punish­
ment which is likely to follow disobedience. Commands are
laws if two conditions are satisfied : first , they must b e gen­
eral ; second they must be commanded by what (as both
(4) the contention that a legal system is a 'closed logical system ' in which
correct legal decisions can be deduced by logical means from predetermined legal
rules without reference to social aims, p olicies, moral standards ; see 64-6 infra,
and
(5) the contention that moral judgments cannot be established or defended, as
statements of facts can, by rational argument, evidence, or proof ( 'noncognitivism'
in ethics ) ; see 82-3 infra. .
Bentham and Aus tin held the views described in ( I ) , ( 2) , and ( 3 ) , but not those
in (4) and ( 5 ) . Opinion ( 4) is often ascribed to analytical jurists ; see 64-6 infra ,
but I know of no 'analyst' who held this view.
26 Gray, The Nature and Sources of the Law 9 4 5 ( 2nd edn. 1 9 2 1 ) .
-

2 7 Austin, op. cit. n . 1 2 supra , a t 1 3.

 POSITIVISM, LAW, AND MO RALS 59

Bentham and Austin claimed) exists i n every p o litical s o ciety

whatever its constitu tional form , namely, a p ers on or a group
of p ersons who are in receipt of hab itual obedience from most
of the society but p ay no such obedience to o thers . These
p ersons are its sovereign. Thus law is the command of the un­
commanded commanders of society - the cre ation of the
legally untrammelled will of the sovereign who is b y definition
outside the law.
It is easy to see that this account of a legal system is thread­
b are . One can als o see why it might seem that its inadequacy
is due to the omission of some essential connection with mor­
ality. The situation which th e simple trilogy of comm and ,
sanc tion , and sovereign avails to describ e, if y o u tak e these
notions at all p recisely , is like that of a gunman saying to his
vic tim, 'Give me y our money or y our life . ' The only difference
is that in the · c as e o f a legal system the gunman says it to a
large number of people who are accustomed to the racket
and habitually surrender to it. Law surely is n o t the gunman
situation writ large , and legal order is surely n o t to be thus
simply identified with compulsion.
This sche m e , despite the p oints o f obvious analogy b e tween
a s t atute and a command, omits some of the m o s t character­
is tic elements of law. Le t me cite a few. It is wrong to think
o f a legislature ( an� a fortiori an electorat e ) with a changing
membership as a group of p ersons habitually ob eyed : this
sim ple idea is suited only to a monarch su fficiently long-live d
for a 'habit' to grow up. Even if we waive this p oint , n o thing
which legislators do m akes law unless they comply with fun­
damental accepted rule s specifying the essential law-m aking
procedure s . This is tru e even in a system h aving a simple uni·
tary constitution like the B ritish. Th ese fundamental accepted
rules specifying what the legislature must do to legislate are
not comm an ds habitu ally obeyed, nor can th ey be expressed
as habits o f obedience to persons. They lie at the ro ot o f a
legal system , and what is most missing in the utilitarian
scheme is an analysis of what it is for a social group and its
officials to accept such rules. Thi s notion, not that of a com­
m an d as Austin claimed, is the 'key to the science o f juris·
prudenc e ' , or at least one o f th e keys.
Again , Austin , in the case o f a dem ocracy , looked past the
legislators to the electorate as 'the sovereign ' ( o r in Englan d
60 GENE RAL THE O RY

as part of it}. He thought that in the United States the mass

of the electors to the state and federal legislatures were the
sovereign whose commands, given by their 'agents ' in the
legislatures, were law. But on this footing the whole notion
of the sovereign outside the law being 'habitually obeyed'
by the 'bulk' of the p opulation must go : for in this case the
'bulk' obeys the bulk , that is, it obeys itself. Plainly the gen­
eral acceptance of the authority of a law-making procedure,
irrespective of the changing individuals who op erate it from
time to time, can be only distorted by an analysis in terms of
mass habitual obedience to certain pers ons who are by defi­
nition outside the law, just as the cognate but much simpler
phenomenon of the general social acceptance
· of a rule, say of
taking off the hat when entering a church, would be distorted
if represented as habitual obedience by the mass to specific
persons.
Other critics dimly sensed a further and more important
defect in the command theory, yet blurred the edge of an im­
portant criticism by assuming that the defect was due to the
failure to insist upon some important connection between
law and morals. This more radical defect is as follows. The
picture that the command theory draws of life under law is
essentially a simple relationship of the c ommander to the
commanded, of sup erior to inferior, of top to bottom ; the
relationship is vertical between the commanders or authors
of the law conceived of as essentially outside the law and
those who are commanded and subject to the law. In this
picture no place, or only an accidental or subordinate place,
is afforded for a distinction between types of legal rules which
are in fact radically different. Some laws require men to act
in certain ways or to abstain from acting whether they wish
to or not. The criminal law consists largely of rules of this
sort : like commands they are simply 'obeyed' or 'disobeyed'.
But other legal rules are presented to society in quite differ­
ent ways and have quite different functions. They provide
facilities more or less elaborate for individuals to create struc­
tures of rights and duties for the conduct of life within the
coercive framework of the law. Such are the rules enabling
individuals to make contracts, wills , and trusts, and generally
to mould their legal relations with others. Such rules, unlike
the criminal law, are not factors designed to obstruct wishes
 POSITIVIS M , LAW , AND MO RALS 61

and choices o f an antisocial sort. On the contrary , these rules

provide facilities for the realization of wishes and choic es .
They do not say (like commands ) 'do this whether you wish
it or n o t ', but rathe r 'if you wish to do this, here is the way
to do i t '. Under these rules we exercise p owers , make claim s ,
an d assert rights. These phras es m ark o ff ch aracteris tic fea­
tures of laws that confer rights and powers ; they are laws
which are , so to sp eak, put at the disp osition of individuals in
a way in which the criminal law is not. Much ingen uity has
gone into the task of 'reducing' laws of this second s ort t o
some c omplex variant o f laws o f t h e first sort. The e ffo rt t o
sh ow that laws conferring rights are 'really ' only conditional
stipulations of sanctions to be e x acted from the person ulti­
mately under a legal duty characterizes much o f Kelsen ' s
work . 2 8 Y e t to urge this i s really just t o exhibit d o gmatic
determination to suppress one asp ect of the legal system in
order to m aintain the theory th at the s tipulation o f a sanc­
tion, like Austin 's c ommand, represents the quintessence o f
law. One might as well urge that the rules o f baseb all were
'really ' only complex c onditional directions to the s corer and
that this showed their real or 'ess ential ' nature.
One of the first j u rists in England to b reak with the Aus tin­
ian tradition, Salm ond, c omplained that the analysis in terms
of commands left the notion o f a right unprovide d with a
place. 29 But he con fused the point. He argued firs t, and cor­
rectly , that if laws are merely commands it is inexplicable that
we should h ave come to speak of legal rights and p ow ers as
conferred or arising under them , but then wrongly concluded
that the rules of a l e gal system must necessarily be c onnected
with moral rules or p rinciples o f ju stice and that only on this
fo oting could the phenomenon of legal rights be e xp laine d.
Otherwise , S almond thought, we would have to s ay that a
mere 'verbal coincidence ' connects the concepts of l egal and
moral right. Similarly , c ontinental critics of the Utilitarians ,
always alive to the c omplexity o f the n o tion of a subjec tive

.1 S ee, e.g. , Kelsen, General Theory of Law and State 58-6 1 , 1 43 - 4 ( 1 945 ) .
According t o Kelsen, all laws, not only those conferring rights and powers, are
reducible to such 'primary norms' conditionally stipulating sanctions •

• 9 Salmond, Th e First Principles of jurisprudence 9 7 - 8 ( 1 89 3 ) . He protested

against 'the creed of what is termed the English school of juris p rudence', because
it 'attempted to de prive the idea of law of that ethical significance which is one of
its mos t essential elements'. Ibid. at 9, 1 0.
62 GENERAL THEORY

right , insis ted that the command theory gave it no place.

Hagerstrom insi s t e d th at i f laws were m e rely commands the
notion of an individual 's right was re ally inexplicable, for
commands are , as he s aid , s o m e thing which we either obey or
we do not obey ; they do not confer rights. 3 0 But he, t o o , con­
clu d e d that moral , or, as he put it, comm on-sense, n o tions of
justi c e must therefore b e necessarily involved in the analysis
of any legal structure elaborate enough t o confer rights. 3 1
Yet, surely these arguments are confused. Rules that c onfer
rights, though dis tinct from c ommands, need not b e moral
rules or coincide with th em . Rights, after all, exist under the
rules o f ceremonie s , gam e s , and in many other spheres regu­
lat e d by rules which are irrelevant to th e question of justice
or what the law ought to b e . Nor need rules which c onfer
rights b e just or m orally goo d rules. The rights of a master
over his slaves s h o w u s th at. 'Their merit o r demeri t ' , as Austin
termed it, depends on how righ ts are distribute d in society
and over whom o r what they are exercised. These critics in­
deed revealed the in adequacy of the simple notions of com­
mand and habit for th e analysis of law ; at many p oints i t is
app arent that the s ocial acceptance o f a rule or standard of
auth ority (even if it is motivated only b y fear or superstition
or rests on in ertia) must be brought into the analysis an d c an­
not itself b e r e d u c e d to t h e t w o simple term s . Y et nothing in
this showed the u tilitarian insistence on the distinction b e­
tween the exis t ence o f law an d its 'm erit s ' to be wrong.

III

I now tum to a distinctively American criticism of the separa­

tion of the law that is fro m the the law that ough t to be. It
em erged from the c ritical study of th e judicial process with
which American jurisprudence has b een on the wh ole so b en­
e ficially occupie d . The m o s t sceptical of these critics - the
l o o sely nam e d ' Realists ' of the 1 9 3 0s - perhap s too naIvely

so Hiigerstrom, Inquiries Into the Nature of Lal}J and Morals 2 1 7 ( Olivecrona

edn. 1 953) : ' [T]he whole theory of the subjective rights of private individuals • • .

is incompatible with the imperative theory '. See also ibid. at 2 2 1 : 'The description
of them [ claims to legal protection] as rights is wholly derived from the idea that
the law which is concerned with them is a true expression of rights and duties in
the sense in which the popular notion of justice understands these terms.'
5 1 Ibid. at 2 1 8.
 POSITIVISM, LAW, AND MO RALS 63

accepted the concep tual framework o f the natural sciences as

adequate for the characterization o f law and for the analysis
of rule-gUided action of which a living system of law at least
partly consists. But they opened men's eyes to what ac tually
goes on when courts decide cases , an d the contrast th ey drew
between the actual fa c t s of judicial decision and the tra­
ditional terminology for describing it as if it were a wholly
logical operation was usually illuminating ; for in spite of some
exaggeration the Realists made us acutely c onscious o f one
cardinal feature o f human l anguage and human thought, em­
phasis on which is vital n o t only for the understanding of law
but in areas of philosophy far b eyond the c onfines of juris­
prudence. The insight o f this sch ool may be presente d in the
following example . A legal rule forbids you to take a vehicle
int o the public p ark . Plainly this forbids an auto m obile, but
what ab out bicycles, roller skates , toy automobiles? What
ab out aeroplanes ? Are thes e , as we say , to b e called 'vehicles '
for the purpose o f the rule or not? If we are to communicate
with each o ther at al, and i f, as in the most elementl;l.ry form
of law, we are to express our intentions that a certain type o f
behaviour be regulate d b y rules, th e n t h e general words we

use - like 'vehicle ' in the c as e I consider - must have some

standard instance in which n o doubts are felt ab out its appli­
catio n . There must be a core of s ettle d meaning, but there
will be, as well , a penumbra of deb atable cases in which words
are neither obviously applicable nor obviously rule d out.
Th ese cas es will each h ave some features in comm o n with the
standard cas e ; they will lack others or be accompanie d by
features n o t pres ent in the standard case. Hum an invention
and natural proce sses continually throw up such variants on
the familiar, and if we are to say that thes e ranges of facts do
or d o not fall under existing rules, then the classifier must
make a decision which is not dictated to him , for the facts
and phenomena to which we fit our words and apply our rules
are as it w ere dumb. The toy automobile cannot sp eak up and
say, 'I am a vehicle for the purpose of this legal rule' , nor can
the roller skates choru s , 'We are not a vehicle. ' F ac t situations
do not await us neatly lab elled, creased, an d folde d ; nor is
their legal classification written on them to be simply read
off by the j u dge . Ins tead , in applying legal rules, s o m e one
must take the responsibility o f deciding that words do or do
64 GENERAL THEORY

not c over some case in hand , with all the practical conse­
quences involved in this decision.
We m ay c all the problems which arise outside the h ard core
of stan dard instances or settled meaning 'problems of the
penumbra' ; th ey are always with us whe ther in relation to
such triv,ial things as the regulation of the use of the public
park or in relation to the multidimensi onal generalities of a
constituti o n. If a p enumbra o f uncertainty mus t surround all
l eg al rules, then their application to specific cases in the pen­
umbral are a cannot be a matter of logical deduction, and so
deductive re asoning, which for generations has b e en cherished
as th e very p erfection of human reasoning, cann ot s erve as a
m o del for what judge s , or indeed anyone, should do in bring­
ing p articular cases under general rules. In this are a men can­
n o t live by deduction alone . And it follows that if legal
arguments and legal decisions of p enumbral questions are to
be rational , th eir rationality must lie in som ething o ther than
a logical relation to premises . So if it is rational or 'soun d ' to
argue and to decide th at for the purposes o f this rule an aero­
plane is not a vehicle , this argum en t must be sound or rational
without b eing logically conclusive. What is it then that m akes
such decisions correct or at least better than altern ative
decisions ? Again, it seems true to s ay that the criterion which
makes a decision sound in such cases is some concept of what
the law ought to b e ; it is easy to slide from that into saying
that it must be a moral judgment ab out what law ough t to b e .
S o here we touch u p o n a p oint o f necessary 'intersection be­
tween law and m orals ' which demonstrates the falsity or, at
any rat e , the misle ading character of the Utilitarians ' em­
phatic insistenc e on the separation of law as it is and ought to
b e . S urely, B en th am and Aus tin could only h ave written as
they did b ecause they misunders t o o d or neglected this aspect
o f the j u dici al process , because they ignored the problems of
the penumbra.
The misconception of the j u dicial process which ignores
th e problems o f the p enumbra and which views the process as
c o nsisting pre-eminently in deductive reasoning is often stig­
matized as the e rror of 'form alism ' or 'literalism '. My question
now i s , how and t o what e xtent does the demons tration o f
this error sh ow th e utilitarian distinction to b e wrong o r
m isleading? Here th ere are m any issu es which have been
 POSITIVIS M , LAW , AND MORALS 65

confused, but I can only disentangle some. The charge of

formalism h as been levelled b o th at the 'p o sitivist' le gal the o r­
istand at the c ourts , but o f course it must be a v e ry different
charge in each case. Levelled at the legal theorist , the c harg e
means that he h as made a theoretical mistake ab out t h e
character of legal decision ; he has thought o f the reasoning
involved as c onsisting in deduction from premises in which
the ju dges ' practical choices or decisions play no p ar t It woul d .

be easy to show that Austin was guiltless o f this erro r ; only

an entire misconception of what analyti cal j urisp rudence is
and why he though t it imp ortan t has led to the view that he,
or any other analys t , believed that the law was a closed logi­
cal system in which ju dges de duced their decisions fro m
premise s . 3 2 O n t h e contrary , h e was very much alive to the
character of langu age , to its vagu eness or open character ;33 he
th ou gh t that in the p enumbral situation judges must neces­
sarily legislate, 34 an d , in accents that sometimes r e c all th o s e

3 2 This misWlderstanding o f analytical jurisprudence i s to be found in, among

others, Stone, The Province and Function of Law 1 4 1 ( 1 950 ) : 'In short, rejecting
the implied assumption that all propositions of all parts of the law must be logi­
cally consistent with each o ther and proceed on a single set of definitions .he . •
·
[ Cardozo, J.J denied that the law is actually what the analytical jurist, for his
limited purposes , assumes it to be. ' See also ibid. at 49, 52, 1 3 8 , 140 ; Friedmann,
Legal Theory 209 (3rd edn. 1 953). This misunderstanding seems to depend on the
unexamined and false belief that analytical studies of the meaning of legal terms
would be impossible or absurd if, to reach sound decisions in particular cases, more
than a capacity for formal logical reasoning from unambiguous and clear predeter­
mined premises is required.
33 See the discussion of vagueness and uncertainty in law, in Austin, op. cit.
n. 12 supra , at 202-5, 2 0 7 , in which Austin recognized that, in consequence of
this vagueness, often only 'fallible tests' can be provided for determining whether
particular cases fall under general expressions_
34 See Aus tin, op. cit_ n. 12 supra , at 1 9 1 : 'I cannot understand h ow any per
son who has considered the subject can suppose that society could possibly have
gone on if judges had not legislated . ' As a corrective to the belief that the
. • .

analytical jurist must take a 'slot machine' or 'mechanical ' view of the judicial
process it is worth noting the following observations made by Austin:
( 1 ) Whenever law has to be applied, the "'competition of opposite analogies '"
may arise, for the case 'may resemble in some of its points ' cases to which the
rule has been applied in the past and in other points 'cases from which the appli­
cation of the law has been withheld '. 2 Austin, Lectures on Jurisprudence 633
(5th edn. 1 885).
(2) Judges have commonly decided cases and so derived new rules by 'build­
ing' on a variety of grounds including sometimes (in Austin's opinion too rarely)
their views of what law ought to be. Most commonly they have derived law from
pre-existing law by 'consequence founded on analogy', i.e. , they have made a new
rule 'in consequence of the existence of a similar rule applying to subjects which
are analogous • ..' 2 ibid. at 638-9.
.
66 GEN E RAL THEORY

of the late J u d ge Jerome Frank, h e b erated the common-law

·
judges for l egi slat in g feebly an d timidly and for blindly rely­
ing on real or fancied analogies with past cases instead of
a d apt in g their decisions t o the growing needs of society as
revealed by the moral stan d ard of utility. 3 s The villains of
thi s piec e , responsible for the conception o f the ju dge as an
automaton, are not the utilitarian thinkers . The resp onsibility ,
if i t is to be laid at the door o f any theorist, is with thinkers
like Blackstone and, at an earlier stage, M ontesquieu. The ro ot
of this evil is preoccupation with the separation of powers
and Blackstone 's 'childish fic tion ' ( as Austin termed it) that
ju dge s only ' find', never 'mak e ' , law.
But we are concerned with 'formalism' as a vice not o f
jurists but of j u dges. Wh at precisely i s i t for a judge to commit
thi s error, to b e a ' formalist', 'automatic', a 'slot machine' ?
Curiously enough the literature, which is full of the denunci­
ation o f these vices, never makes this clear in concrete term s ;
ins tead we h ave only descriptions which c annot mean what
they app e ar to say ; it is said that in the formalist error courts
make an excessive use of logic , take a thing to 'a dryly logical
e xtrem e ',36 or make an excessive use of analytical metho ds.
But just h ow in being a form alist does a judge m ake an ex­
cessive use o f logic ? It is clear that the essence of his error is
to give some general term an interpretation which is blind to
so cial values and conse q uences ( o r which is in some other way

( 3 ) ' [ I] f every rule in a system of law were perfectly definite or precise', these
difficulties incident to the application of law would not arise. 'But the ideal com­
pleteness and correctness I now have imagine d is not attainable in fact though
• • . •

the system h ad been built and ordered with matchless solicitude and skill.' 2 ibid.
at 9 9 7-8. Of course he thought that much could and should be done by codifi­
cation to eliminate uncertainty. See 2 ibid. at 662-8 1 .
3 5 2 ibid. a t 6 4 1 : 'Nothing, indeed, can b e more natural, than that legislators,
direct or judicial (especially if they be narrow-minded, timid and unskilful) , should
lean as much as they can on the examples set by their predecessors. ' See also 2
ibid. at 64 7 : 'But it is much to be regretted that J udges of capacity, experience
and weight, have not seized every opportunity of introducing a new rule ( a rule
beneficial for the future) • This is the reproach I should be inclined to make
. . •

against Lord Eldon • . • [TJhe Judges of the Common Law CQurts would not do
.

what they ought to have done, namely to model their rules of law and of pro­
cedure to the growing exigencies of society, instead of stupidly and sulkily adher
ing to the ol d and barbarous usages.'
36 Hynes v. New York Cen t. R. R. , 23 1 N.Y. 229, 235 ; 1 3 1 N.E. 898, 9 0 0
( 1 92 1 ) ; see Pound, Interpretations of Legal Histo ry 1 2 3 ( 2 n d edn. 1 9 3 0 ) ; Stone,
op. cit. n. 32 supra , at 140- 1 .
 POSITIVIS M , LAW, AND MORALS 67

stupid o r perhap s merely disliked b y c ritics) . B u t logic does

not prescribe interp re tation o f term s ; it dic tates neither the
stupid nor intelligent interpre tation o f any expression . Logic
only tells you hypothetically that if you give a c ertain t e rm a
certain interpretation then a certain conclusion follows . Logic
is silent on how t o classify p articulars - and this is the h e art
of a ju dicial decision. S o this reference to logic and to logical
extremes is a misnomer for something else, whic h mus t b e
this. A judge has to apply a rule t o a c oncrete c a s e - p e rhaps
the rule that one m ay n o t take a stolen 'vehicle ' across State
lin e s , and in this case an aeroplane has been taken. 3 7 He e ither
does not see or pretends n o t to see that the general terms of
this rule are suscep tible o f different interpretations and that
h e has a choice left open uncontrolled by lingu istic conven­
ti ons. He ignores , or i s blind to, the fact that he i s in the area
o f the p enumbra and i s not dealing with a stan d ard cas e . In­
stead of choosing in th e ligh t of social aims th e ju dge fixes
the meaning in a differen t way. He either takes the meaning
that the word m o s t obviously suggests in its ordinary non-l egal
context to ordinary men , or one which the word has been
given in some o ther legal contex t , or, still worse , he thinks of
a standard c a s e an d then arbitrarily identi fies certain fea tures
in i t - for example, in the case of a vehicle , ( 1 ) n ormally used
on lan d , ( 2 ) c apable o f carrying a human pers o n , ( 3 ) capable
of being self-propelled - and treats th ese three as always
necessary and always sufficient conditions for the use in all
contexts of the word 'vehicle ' , irrespective of the social con­
sequences of giving i t this interpretation. This choic e , not
'logic ', would force the judge to inclu de a toy m o tor c ar ( i f
electrically propelled) an d to exclude bicycles and the aero­
plane. In all this there is p o s sibly great stupidity , but no more
'logic ', and no les s , than in cases in which the interpre tation
given to a general term and the consequent applicati on o f
s o m e general rul e to a p articular cas e i s consciously contr olle d
by s o m e identifi e d social aim .
Decisi ons m ade in a fashion as b lind as this would scarc ely
des erve th e name o f decisions ; we might as well toss a p enny
in applying a rule of l aw. But it is at least doubtful whether
any judicial decisions ( even in England) have b een quite as

37 See McBoyle v. United States, 283 U.S. 25 ( 1 9 3 1 ) .

68 GENERAL THE O R Y

au tomatic as this . Rather, either t h e interpretations s tigm a­

tized as automatic h ave resulted from the conviction that it is
fairer in a criminal statute to take a m e aning which would
jump to the m ind of the ordinary m an at the cost even of
defeating other values, and this itself is a social policy (though
p ossibly a bad one ) ; or much more fre quently , what is stigma­
tiz e d as 'mechanical ' and 'automati c ' is a determined choice
made indeed in the ligh t of a social aim but of a conservative
social aim . Certainly many of the Supreme Court decisions at
the tum of the century which have b e en so stigmatized38 rep ­
res ent clear choices in the p enumbral area to give e ffect to a
p olicy o f a c onservative type. This is p eculiarly true of Mr
Jus tice Peckham 's opinions defining th e spheres of p olice
power an d due process . 3 9
Bu t h o w d o e s t h e wrongness o f deciding cases i n an auto­
matic and m echanical way an d the rightness of deciding cases
by reference to social purposes show that the utilitarian in­
sistence on the distinction between what th e law is and what
it ought to be is wrong? I take it that no one wh o wished to
u s e thes e vices of formalism as pro o f that the distinction b e ­
tween what is and what ought to b e is mistaken would deny
that the decisions stigm atized as automatic are law ; nor would
he deny that the system in which such au tomatic decisions
are m ade is a legal system. Surely he would say that they are
law, but they are b ad law; they ought not to be law. But this
would be to use th e dis tincti on , not to refute it ; and of course
b o th B enth am and Austin used it to attack judges for failing
to decide p enumbral cases in accordance with the growing
needs of society.
Clearly, if the dem onstration o f the errors o f form alism is
to sh ow the u tilitarian distinction to b e wrong, the p oint mus t
b e drastically restated. The point m u s t be n o t merely that a
judicial decision to be rational mus t b e m ade in the light of
some conception o f what ought to be, but that th e aim s , the
soci al p olicies and purposes t o which j u dges should app e al if

3. See, e.g. , Pound, 'Mechanical ]urisprudence ', 8 Colum. L. Rev. 605 , 6 1 5 - 1 6

( 1 90 8 ) .
30 See, e.g., Lochner v . New York , 1 9 8 U.S. 4 5 ( 1 905). Jus tice Peckham's
opinion that there were no reasonable grounds for interfering with the right of free
contract by determining the hours of labour in the occupation of a baker may in
deed be a wrongheaded piece of conservatism but there is nothing automatic or
mechanical about it.
 POSITIVISM, LAW , AND MORALS 69

their decisions are to b e rational , are themselves to b e con­

sidered as part of the law in some suitably wide sens e of 'l aw'
which is held to b e m o re illuminating than that used by the
Utilitari ans . This restatement of th e point would have the
following consequenc e : instead of saying that the re currence
of p enumbral ques tions shows us that legal rule s are e ssen­
tially incomplete, and that , when they fail to d e termine de­
cisions, judges mus t l e gislate and so e xercis e a creative c h oice
b etween alternatives , we shall say that the so cial policies
which guide the j u d ge s ' choice are in a sense there for them
to discover; the judges are only 'drawing ou t ' o f the rule what,
i f i t is prop erly unders t o o d , is 'latent ' within it . To call this
judicial legislation is t o obscure some essential c ontinuity be­
tween the clear cases of the rule's application an d the pen­
umbral decisions. I s h all question later wh ether this way of
talking is salutary , b u t I wish at th is time t o p oint out s o m e­
thing obvi ous, but lik ely , if not stated, to tangle the issues.
It does not follow that , because the op p osite o f a decision
reached blindly in the formalist or literalist m anner is a
dec ision intelligently reached b y reference to some concep tion
o f what ought to be, we have a junction of law and m o rals .
We mus t , I think, beware of thinking in a t o o simple-minded
fashion ab out the word 'ough t ' . This is not because there is
no distinction to be m ade b e tween law as it is an d ought to
b e . Far from it. It is b ecause the distinction sh ould b e between
what is and what fro m many di fferent p oints o f view ought
to be. The word ' ough t ' merely reflects the presenc e o f some
standard of criticism ; one o f these standards is a moral stan­
dard, but not all standards are m o ral . We say to our n eigh­
b our, 'You ought n o t to lie ' , and that may certainly be a
moral ju dgment, b u t we should rememb er that th e baffl ed
poisoner may say , 'I ought to have given h er a s econd d o s e ' _
T h e point here is t h a t intelligent d ecisions whic h we oppose
to mechanical or formal decisions are n o t necessarily iden tical
with decisions defensible on m oral grounds . We may say o f
many a decision : 'Ye s , that is right ; that is as it ought t o b e ' ,
and w e may m ean o nly th at s ome accepted purp ose or p olicy
has b een thereby advanced ; we may not m e an t o endors e the
m oral p ropriety of the policy or the decision. S o the c o n tras t
be tween the m echanical decision and th e intelligent one c an
b e reproduced inside a system dedicate d to t h e pursuit o f the
70 GENERAL THEORY

most evil aims. I t does not e xist as a contrast to b e found only

in le gal systems which , like our own, widely recogniz e prin­
ciples o f justice an d moral claims of individuals .
An e x ample m ay m ake this p oint plainer. With us the task
of sentencing in criminal c ases is the one th at seems most
obviou sly to demand from th e ju dge the exercise o f m oral
ju dgm ent. Here the factors to be weighe d seem clearly to b e
moral factors : s ociety must not b e exp o s e d to wanton attack ;
t o o much misery must n ot be inflicte d on either the vic tim or
his dependants ; e fforts mus t be made to e nable him to lead a
better life and regain a position in the s ociety whose laws h e
has violated . To a j u dge striking the b alance among thes e
claim s , with all t h e discretion and p erplexities involved, his
task s eems as p lain an example of the e xercise of m o ral judg­
ment as c ould b e ; and it seems to be the polar opposite of
s o m e m echanic al appli cation o f a tari ff o f p enalties fixing a
sentence careless of th e moral claims which in our system have
to be weigh e d . So here intelligen t and rational decision is
guided however uncertainly by m oral aims . But we have only
to vary the example to see that this n e e d not necess arily b e
s o and surely , i f i t need not necessarily b e s o , the u tilitarian
p oint remains unshaken . Under the Nazi regime men were sen­
tence d b y courts for criticism of th e regim e. Here the choi c e
o f s entence m i ght b e guided ex clusively by c onsideration o f
what was needed to main tain t h e state 's tyranny e ffectively.
What sentence would both terrori z e the public at large and
keep the friends and family of th e p risoner in suspense so that
b o th h op e an d fear would c o op e rate as fac tors making for
s u b servie nce? Th e pris oner o f such a system would be re­
garded s imply as an obj ect to b e used in p ursuit o f these aims .
Y e t , in contrast with a mechanical decisi o n , decision o n these
grounds would b e intelligen t and purp o sive , and from one
p oint of view the decision would b e as it ough t to b e . O f
course, I am not unaware that a wh ole philosophical tradition
has sough t t o demonstrate the fact th at we cannot c orrectly
call decisions or b eh avi our truly ratio n al unless they are in
conformity with m o ral aim s and p rinciples . But the example
I h ave used seems to me to serve at leas t as a warning that we
cannot use the errors of formalism as s o m ething which per se
demons trates the falsity of the u tilitarian insistence on the dis­
tinction between law as it is and law as morally it ought to b e .
 POSITIVISM , LAW , AND MO RALS 71

We can now return to the m ain point. If it is true th at the

intelligent decision o f penumbral ques tions is one made not
mechanically but in the light o f aims , purp o s e s , and policies,
though n ot necessarily in the light o f anything we woul d call
moral principles, is it wise to express this imp ortant fact by
saying that the firm u tilitarian distinction between what the
law is and what it ought to be should be dropped? Pe rhaps
the claim that it is wis e c annot b e theoretically refuted, for it
is , in effect, an invita tion to revise o ur conception o f what a
legal rule is. We are invite d to include in the 'rule ' the various
aims and policies in the light o f which its penumbral c ases are
decided on the ground that these aims have, b ecause of their
imp o rtance, as much right to be called law as the c ore of legal
rules whose meaning is s e ttled. But th ough an invitation c an­
not be refuted, it m ay be refused, and I would pro ffer two
reasons for refusing this invitation. First , everything we h ave
learned ab out the judicial process can be express e d in other
less m ysteri ous ways. We can s ay laws are incurably incom­
plete and we must d ecide the p enumbral cas e s rationally by
referenc e to social aims. I think Holmes, who h ad such a vivid
appreciation of the fact that 'general propositions do not
decide c oncrete c ases ', would h ave put i t that way . S e c ondly,
to insist on the utilitari an distinction is to emphasize that the
hard core o f s ettled meaning is law in some centrally import­
an t s ense and that even if there are b orderlines , there must
first be lines . If this were not so the notion o f rules con trolling
courts ' decisions would be s enseles s , as some of the ' Realists '
- in their m o s t e x treme moods , and, I think , on bad grounds
- claime d .40
By contras t , to s o ften th e distinction, to ass e rt mysteriously
that there is some fused identity between law as it is and as it

40 One recantation o f this extreme position is worth mention in the present

context. In the first edition of The Bramble Bush , Professor Llewellyn c ommitted
himself wholeheartedly to the view that 'what these officials do about disputes is,
to my mind , the law itself', and that 'rules are important so far as they h elp
• . •

you predict what j udges will do

. • • • That is all their importance, except as
• • •

pretty playthings. ' Llewellyn, The Bramble Bush 3 , 5 ( Ist edn . 1 930). In the sec
ond edition he said that these were 'unhappy words when not more fully devel­
oped, and they are plainl y at best a very partial statement of the whole tru th • . • .

[ O l ne office o f law is to control officials in some part, and to guide them even
. • where no thoroughgoing control is possible, Dr is desired
• [T ] he words fail
• • . •

to take proper account of the office o f the institution of law as an instrument

• • .

of conscious shaping .' Llewellyn, The Bramble Bush 9 ( 2nd edn. 1 9 5 1 ) .

. • •
72 GENERAL THEO RY

ought to b e , is to suggest that all legal qu estions are funda­

mentally like those of the p enumbra. It is to assert that there
is no central element of actual law to be seen in the core o f
central meaning which rules h ave, that there i s no thing i n the
n ature o f a legal rule inconsistent with all questions b eing
open to reconsideration in the light of social policy. Of course,
it is good to b e occupie d with the penumbra. Its problems are
rightly the daily diet o f the law sch o ols. But to be occupied
with the penumbra is one thing, to b e preoccupied with it
another. And preoccupation with the penumbra is, if I m ay
say s o , as rich a s ource o f confusio n in the American legal
tradition as formalism in the English. Of course we might
ab andon the n otion th at rules h ave au thority ; we might cease
to attach forc e or even meaning to an argument that a case
falls clearly within a rule and the scope of a precedent. We
might call all such reas oning 'auto m atic ' or 'me chanical ' ,
which is already t h e routine invective o f t h e courts . But until
we decide that this is what we want, we should n o t encourage
it by obliterating the utilitarian distinction.

IV

Th e third criticism of the sep aratio n of law and morals is of

a very different character ; it certainly is less an intellectual
argument against the utilitarian distinction than a passionate
appeal sup p o rted not b y detailed reas oning but by reminders
of a terrible e xperience . F o r it consists of the tes tim ony o f
tho s e wh o h ave descended into Hell , and, like Ulysses or
Dante , brought b ack a message for human b eings . Only in this
cas e the Hell was not beneath or b eyond earth , but on it ; it
was a Hell created on earth b y men for o ther men.
This app e al comes fro m those Germ an thinkers wh o lived
through the Nazi regime and reflected upon its
tations in the legal sy stem . One o f these thinkers, Gustav
Radbruch, h ad himself shared the 'positivis t ' doctrine until
the Nazi tyranny , but h e was converted by this exp erience
and so his app e al to other men to discard the doctrine of the
sep aration o f law and m orals has the sp ecial p oignancy o f a
rec antatio n . What is imp ortant ab out this criticism is that it
really d o e s c onfront the p articular p oint which Bentham and
Austin had in mind in u rging the separation of law as it is and
 POSITIVISM, LAW, AND MO RALS 73

as i t ough t t o b e . Th ese German thinkers put th eir insistence

on the need to j oin together what the Utilitarians separated
just where this separation was o f most imp ortance in the eyes
o f the Utilitarians ; for they were concerned with the problem
posed by the existence of m orally evil laws.
Before his conversion Radbruch held that resis t ance to law
was a m atter for the p ersonal conscience, to be thought out
by the individu al as a moral problem , and the validity of a
law c ould n o t be disprove d by showing that its requirem ents
were morally evil or even by sh owing that the e ffect of com­
pliance with the law would be more evil than the e ffect of
disobedienc e . Aus tin , it may b e recalled, was emphatic in con­
demning those who said that if human laws conflicted with
the fundamental principles of morality th en they ceas e to be
laws , as talking 'stark nonsens e ' .

T h e m ost p ernicious laws, and therefore those which are m o s t opposed

to the will of God , have been and are continually enforc e d as l aws by
ju dicial tribunals. Suppose an ac t innocuous, or p ositively benefic ial , be
prohibited by the sovereign under the penalty of death ; if I commit this
act, I shall be tried and condemned, and if I object to the sentence, that
it is contrary to the law of G od . . . the c ourt of justic e wil dem o nstrate
the inconclusiveness of my reasoning by hanging me up, in pursu ance of
the law of which I have impugned the validity. An excep tion, demurrer,
or plea, founded on the law of God was never heard in a Court o f Jus ­
41
tice, from the creation of the world down to the present moment.

These are strong, indeed bru tal words, but we mus t remember
that they went along - in the c ase of Austin and , of c o urs e,
Bentham - with the conviction that if laws reached a certain
degree of iniquity then there would be a plain moral ob ligation
to resist them and to withhold obedience. We sh all s e e , wh en
we consider the alternatives , that this simple presentation of
th e human dilemma which may arise has much to be said
for i t .
Radb ruch , however, h ad concluded from th e eas e with
which the Nazi regime had exploited subservience to mere
law - expressed, as he th ought, in the 'p ositivis t ' slogan
'law as law' ( Gesetz als Gesetz ) - and from the failure o f the
Germ an legal profession to protest against th e en o rmities
which they were required to p erp etrate in the name of law,

4 1 Aus tin, The Province of Jurisprudence Determined 1 85 ( Library of Ideas

edn. 1 954).
74 GENERAL THEORY

that 'positivism ' (meaning here the insistence on the separ­

ation of law as it is from law as it ought to b e) had powerfully
contributed to the horrors. His considered reflections led him
to the doctrine that the fundamental principles of humani­
tarian morality were part of the very concept of R echt or
Legality and that no positive enactment or s tatute , however
clearly it was expressed and however clearly it conformed
with the formal criteria of validity of a given legal system,
could be valid if it contravened b asic principles of morality.
This doctrine can be appreciated fully only if the nuances im­
ported by the German word R echt are grasped. But it is clear
that the doctrine meant that every lawyer and judge should
denounce statutes that transgressed the fundamental prin­
ciples not as merely immoral or wrong but has having no legal
character, and enactments which on this ground lack the
quality of law should not be taken into account in working
out the legal position of any given individual in particular cir­
cumstances. The striking recantation of his previous doctrine
is unfortunately omitted from the translation of his works ,
but it should be read by all who wish to think afresh on the
question of the interconnection of law and morals.42
It is impossible to read without sympathy Radbruch's
passionate deman d that the German legal conscience should
be open to the demands of morality and his complaint that
this has been too little the case in the German tradition. On
the other hand there is an extraordinary naIvety in the view
that insensitiveness to the demands of morality and s.u bser­
vience to state p ower in a people like the Germans should
have arisen from the b elief that law might be law though it
failed to conform with the minimum requirements of moral­
ity. Rather this terrible history prompts inquiry into why em­
phasis on the slogan 'law is law', and the distinction between
law and morals , acquired a sinister character in Germany, but
elsewhere, as with the Utilitarians themselves , went along
with the most enlightened lib eral attitudes. But something
more disturbing than naIvety is latent in Radbruch's whole
42 See Radbruch, 'Gesetzliches Unrecht und U bergesetzliches R e cht', I Sud­
deutsch ]uristen-Zeitung 1 05 ( Germany 1 946 ) ( reprinted in Radbruch, Rechts
philosophie 347 ( 4th edn_ 1 950». I have used the translation of part of this essay
and of Radbruch, 'Die Emeuerung des Rechts', 2 Die Wandlung 8 ( Germany
1 947 ) , prepared by Professor Lon Fuller of the Harvard Law School as a mimeo­
graphed supplement to the readings in jurisprudence used in his course at Harvard.
 POSITIVIS M , LAW , AND MORALS 75

presentatio n o f t h e issues to which t h e existence o f morally

iniquitous laws give ris e . It is n o t , I think, uncharitable to s ay
th at we c an see in his argument that he has only half diges t e d
th e spiritu al message o f liberalism which he is seeking to c on­
vey to the legal profession. F o r everything that he s ays is
really dependent upon an enormous overvaluation of the im­
p o rtance , o f the b are fact that a rule may be s aid to be a valid
rule of law , as if this , once declared, was conclusive of the
final moral question : 'Ought this rule of law to be obeyed? '
Surely the truly liberal answer to any sinister use of the s l ogan
'law is law ' or of the distinction b e tween law and m orals i s ,
'Very well, b u t that d o e s n o t c onclu de t h e question. L a w i s
not morality ; d o not le t i t supplant morality. '
However, we are not left to a m ere academic discussi on in
order to evaluate the plea which Radbruch made for the re­
vision of the distinctio n between law and m o rals . A fter the
war Radbruch's c onception o f law as containing in itself the
essential m oral principle of humanitarianis m was applie d in
practice by Germ an courts in certain cases in which local war
criminals, spies, and inform ers under the Nazi regim e were
punished. The special imp ortanc e of these cases is that the
persons accused of these crime s claim ed that what they had
done was not illegal under the laws of the regime in force at
th e time these actions were perform ed. This plea was m e t with
the reply that the laws upon which they relied were invalid
as c ontravening the fundamental principles o f morality. Let
'
me cite briefly one o f these cas e s . 43
In 1 944 a wom an , wishing to be rid of her husband,

4 3 Judgment of 27 July 1 949, Oberlandesgericht, Bamberg, 5 Siiddeutsche

Juristen-Zeitung 207 ( Germany 1 950 ) , 64 Ham L. Rev. 1 005 ( 195 1 ) ; see Fried
mann, Legal Theory 45 7 ( 3rd edn. 1 9 5 3 ) . The text has been left as originally
written, but it has been shown by Dr H. O. Pappe of the Australian National Uni
versity in his article 'On the Validity of Judicial Decisions in the Nazi Era' in 23
Mod. L Rev. ( 1 960), 260, that the report of the case in 64 Harv. L. Rev. which
was followed by the author is misleading. As Dr Pappe shows , in the actual case
the German court after accepting the theoretical possibility that statutes might
be invalid if in conflict with natural law held that the N azi statutes in question
could not be held to violate it; the accused was held guilty of unlawfully depriving
her husband of liberty , since she had no duty to inform against him but did so for
purely personal reasons and must have realized that to do so was in the circum
stances 'contrary to the sound conscience and sense of justice of all decent human
beings ', Accordingly, the case as discussed in the text must n ow be regarded as a
hypothetical one_ Dr Pappe's careful analysis of a decision in a similar case in the
German Supreme Court sh ould be studied. (Op, cit., 268 ff. ) .
76 GENERAL THEORY

denounced him to the authorities for insulting remarks he had

made ab out Hitler while home on leave fro m the German
army. The wife was under no legal duty to rep ort his acts,
though what h e h ad said was app arently iIi violation of s tat­
utes making it illegal to make statements d etrim ental to the
government of the Third Reich or to impair by any means
the m ilitary defence of th e German p eople . The husband was
arrested and sentenced to death , app arently p ursu ant to th es e
statutes, though he was not executed but was sent to the
front . In 1 94 9 the wife was prosecuted in a West German
court for an o ffen c e which we would describ e as illegally de­
priving a person o f his free dom (rech tswidrige Freiheitsbe­
raubung) .This was punishable as a crime u nder the German
Criminal C o d e of 1 8 7 1 which had remained in force continu­
ously since its enactment. Th e wife pleaded that her husb and 's
imprisonment was pursuant to th e Nazi statutes and h ence
that she h ad committed no crime. Th e c o urt o f app e al to
which the c ase ultimately came held that the wife was guilty
of procuring th e deprivation of her husb an d ' s lib erty by de­
nouncing him to the German c ourts , even though he had b een
sentenced by a c ourt for having violate d a statute, since, to
qu ote the words of the court , the statute 'was contrary t o the
s ound c onscience an d sense of jus tic e of all decent human
beings '. This reas o ning was followed in m any c ases which have
been h aile d as a triumph o f the doctrines of natural law and
as signalling the overthrow of p ositivism. The unqualified s atis ­
faction with this result seems to m e to b e hysteria. Many o f
u s migh t applaud the obj ective - that o f p unishing a woman
for an outrageously immoral act - but this was secured only
by declaring a statute established since 1 9 3 4 not to h ave the
force of law, and at least the wisdom of this course must b e
doub ted. There w ere , o f course, t w o o ther choices. O n e was
to let the woman go unpunished ; one c an sympa thize wi th
and endorse the view that this might h ave been a b ad thing to
do . Th e o ther was to face the fact that if the woman were to
be punishe d it must be pursuant to the introduction o f a
frankly retrospec tive law and with a full c onsciousness of
what was s acrificed in securing her punishment in this way .
Odious as retrosp ec tive criminal legislation and punishment
may b e , to h ave pursued it openly in this case would at least
have h ad the meri t s of candour. It would h ave made plain that
 POSITIVISM , LAW , AND MORALS 77

in punishing the woman a choice had to b e made between two

evil s , that of leaving h er unpunished and that of sacrificing a
very precious principle of morality endors ed by most legal
systems. Surely if we have learned anything from the history
of morals it is that th e thing to do with a m oral quandary is
n o t to hide it. Like nettles, the occasions when life forces
u s to ch oose the lesser o f two evils mus t be grasp e d with
the c onsciousness that they are what they are. The vice
of this use of th e principle that , at certain limiting p o ints,
wh at is utterly imm oral cannot be law or lawful i s that it will
serve to cloak the true nature o f the problems with which we
are faced and will encourage the romantic o p timis m that all
th e values we cherish ultimately will fit into a single system ,
that no one of them h as to be s acrificed or compromised to
accommo date another.

All Discord Harmony n o t understood

All Partial Evil Univers al G o o d

This is surely untrue, and there is a n insincerity i n any for­

mulation of our problem which allows us to de scribe the
treatment of the dilemma as if it were th e disposition o f the
ordinary case.
It may seem perhaps to make too much of forms, even p er­
haps o f words, to emp hasize one way of dis p o sing o f this dif­
ficult case as compared with ano ther which might have l e d , so
far as the woman was c o ncerned, t o exactly the same re sult.
Why should we dramatize the di fference between them ? We
migh t punish the wom an under a n ew retrospective law and
declare o vertly that we were doing something in consis tent
with our principles as the lesser of two evils ; or we m ight allow
the case to pass as one in which we d o not p o int out precisely
where we sacrifice such a principle . But candour is not just
one am ong many minor virtu es of the administration of law,
ju st as it is not merely a minor virtue of m o rality . For if we
adop t Radbruch 's view, and with him and the German courts
make our protest against evil law in the form of an ass ertion
that certain rules canno t be i aw b ecause of their m oral in­
iquity, we confuse one o f the m o s t p o werful, b ecau s e it is the
simplest , form s of m oral criticism . If with the Utili tarians we
sp eak plainly , we say that laws may be law but too evil t o be
obeyed. This is a m o ral condemnation which everyon e can
78 GENERAL THEORY

understand and it makes an immediate and o bvious claim to

moral attention. If, on the other hand, we formulate our o b ­
j e c tion a s an assertion that these evil things are n o t law, here
is an asserti on which many people do not b elieve, and if they
are disp osed t o consider it at all , it would seem to rais e a
whol e host o f philos ophical issue s b efore it c an be accep ted.
S o p erhap s the most imp ortant single lesson to be learne d
fro m this form o f the denial o f t h e u tilitarian distinction is
the one that the Utilitarians were most concerned to teach ;
when we h ave the ample resources of plain speech we must
not present th e moral cri ticism of institutions as propositions
o f a disputable philosophy.

I have en d e avoured to show th at , in spite of all that has b e en

learn ed and experienced since the Utilitarian s wrote , and in
spite of th e defects of other parts of their doc trine, their pro­
test against th e confusion o f what is and wh at ought to be
law has a m o ral as well as an intellectual value. Yet it m ay
well be s ai d that , though this distinction is valid and import­
ant if applied to any p articular law o f a system , it is at least
misleading i f we attempt to apply it to 'law', that is , to the
notion of a l egal system , and th at if we insist , as I have, on the
narrower truth (or truism ) , we obscure a wider ( or deep er)
truth. A fter all, it m ay be urged , we have learned that there
are m any things which are untrue of laws taken sep arately,
but which are true and imp ortant in a legal system considered
as a whole. F or example, th e c onnection between law and
s anctions and between the existence of law and its 'e fficacy'
mus t b e understood in this mo r e general way . I t is surely not
arguable ( without some desperate extension o f the word
'sanction ' o r arti ficial narro wing of the word 'law ') th at every
law in a municipal legal sys tem must h ave a sanction, yet i t is
at least plau sible to argue that a legal system must, to be a
legal system , provide sanctions for certain of its rules. S o to o ,
a rule o f law may b e said t o exist th ough enfo rc ed o r obeyed
in only a minority of cas es, but this could not be s aid of a
legal system as a whole. P erhap s the differences with respect
to laws taken separately and a legal system as a wh ole are
als o true of th e c onne c ti o n b e twe en moral ( or some other)
 POS ITIVISM, LAW, AND MORALS 79

conceptions of what law ought to b e and law in this wid er

sense.
Thi s line o f argument, found (at least in embry o form ) in
Austin, where he d raw s attention to the fact that every devel­
oped legal system contains certain fundamental n o tions
which are 'necess ary' and 'bottomed in the c ommon n ature
of m an ' ,44 is worth pursuing - up to a p oint - and I s hall s ay
briefly why and how far this is s o .
We must avoid , i f w e c an , the arid wastes o f inapp ropriate
definition , for, in relatio n to a concep t as m any-sided and
vague as that o f a legal system , disputes ab out the 'essential '
character, or nec essity to the whole , o f any single el ement
soon b egin to look like disputes ab out wh ether chess could
be 'chess ' if play e d without pawns . Th ere is a wish, which
may b e understan dable, to cut straight through th e question
whether a legal system, t o be a legal system , must measure up
to s ome m oral or o ther standard with simple statements o f
fact : for e x ample, that n o system which utterly failed in this
respect has ever existed or could endure ; that the n ormally
fulfilled assump tion that a legal system aims at some fo rm o f
justice colours t h e whole way i n which w e interpret sp ecifi c
rules in particular cases , and if this normally fulfille d assump­
tion were not fulfilled n o one would h ave any reason to obey
excep t fear ( and probably not that ) , and still les s , o f course,
any moral obligatio n to obey. The connectio n b e tween law
and moral standards and principles of justice is therefore as
little arbitrary and as 'necessary ' as the connection b etwee n
law and sanctions, and t h e pursuit o f t h e question wh ether
this necessity is logical (part o f the 'meaning' o f law) or
merely factual or c ausal c an safely be left as an innoc ent pas­
time for philosophers.
Yet in two respects I should wish to go further (even
th ough this involves the use o f a philosophical fantasy) and
show what c ould intelligibly be meant by the claim that cer­
tain provisions in a legal system are 'necessary '. Th e world in
which we live, an d we who live in it , may one day change in
many different ways ; and if this change were radic al enough
not o nly would certain statem ents of fact n o w tru e be fals e
and vic e versa, but wh ole ways of thinking and talking which
•• Austin, 'Uses of the S tudy of Jurisprudence', in Th e Province of Jurispru
dence Determined 3 6 5 , 3 7 3 , 3 6 7 9 ( Library of I deas edn. 1954).
80 G E N E R A L THEORY

c onstitu te our present concep tual apparatus , through which

we see the world and each o ther, would lap s e . We have only
to consider how the whole of our s ocial, moral , and legal life ,
as we understand it now, depends on th e contingent fac t that
though our bodies do change in shape, size , and other physical
prop erties they do not do this so drastically nor with such
quicksilver rapidity and irregularity that we canno t identify
e ach o th er as the same persistent individual over considerable
spans of time. Th ough this is but a contingent fact which may
one day be different , on it at present rest huge structures o f
our th ought and principles o f action and s ocial life. Similarly,
consider the follo wing possibility (not b ec ause it is m o re than
a po ssibility but because it reveals why we think certain things
necessary in a legal system and what we mean by this ) : sup­
pose that men were to become invulnerable t o attack by each
o ther, were clad perh aps like giant land crab s with an impen­
etr ab le carapac e , and could extract the foo d they needed
from the air b y some internal chemical process . In such cir­
cu mstances (the de tails of which can be left to science fiction)
rules forbidding the free use o f violence and rules constituting
th e m inimum form of prop erty - with its rights and duties
suffic ient to enable fo od to gro w and be retained u n til eaten
would not h ave the necessary non-arbitrary status which
they h ave for us , c onstitu ted as we are in a world like ours .
At present, and until such radical changes supervene, such
rules are s o fundamental that if a legal system did not have
them there w ould be no p oint in h aving any other rules at all.
Such rules overlap with basic moral principles vetoing murder,
violence, and theft ; and s o we can add to the factual state­
m ent that all legal systems in fact coincide with m orality at
such vital p o in t s , the statement that this is, in this sense,
necessarily so. And why not call it a 'natural ' necessit y ?
O f course even this much depends on t h e fact that in asking
what content a legal system must have we take this question
to b e w orth asking only if we who consider it cherish the
humble aim of survival in close proximity to our fellows.
Natural-law th eory, however, in all its pro tean guises , attemp ts
to push the argument much furth er and to assert th at human
beings are equally devote d to and unit ed in their c o nception
o f aim s (th e pursuit o f knowledge, j u stic e to their fellow
m en ) other than that of survival and these dictate a furth er
 POSITIVIS M , LAW, AND MORALS 81

necessary content to a le gal system (over and ab ove my

humble minimum) without which it would be p o in tl e s s Of .

course we must b e careful not to exaggerate the di fferences

among human beings , but it seems to m e that above this mini­
mum t h e purp oses men have for li v in g in s o c i e t y are too con­
flicting and varying to make possible much extension o f t h e
argument that some fuller overlap of legal rules and m o ral
standards is 'necessary ' in this sense.
Another aspect of the matter deserves attention. If we
attach to a legal system the minimum meaning that it must
consist of general rules - general both in the sense that they
refer to courses of action, not single actions and to mul ti­
p lic i ti es of m e n , not s in gle individuals - this meaning con­
notes the p rin cip le of t reating like cases alike, th ough the
criteria of when cases are alike will be, so far, only the general
elem ent s specified in the rules. It is, however, true t h at o ne
essential element o f the concept of justice is the principle of
treating like cases alike. This is j u s ti ce in the administration
of the law, not justice of the law. So there is , in the very
notion of law cons i st in g o f g e n eral rules, someth in g which
prevents us from treating it as if m orally it is utterly neutral,
without any necessary contact with moral p rin c ip les . Natural
p.r ocedural justice consists therefore of those principles of
objectivity and impartiality in the administration of the law
which implement just this asp ect of l aw and which ar e de­
signed to ensure that rules are appl i ed o nly to what are genu­
inely cases of the rule or at least to mi ni m i z e the ri sks of
inequalities in this sense.
These two reasons ( or excuses) for talking of a ce rt ai n
overlap b e t w ee n l e gal and moral s tandards as n ece s s ary and
natu ral, of course, should not sat is fy anyone who is really
disturbed by the u t ili tari an or 'positivis t' insis tence t h at law
and morality are distinct. This is so because a legal system that
satisfied these minimum requ ire m e n ts m igh t apply , with the
most pe dan t ic impartiality as b etween the persons affected,
laws which were hideously oppressive, and might d e n y to a
vast rightless slave p opul at i o n the minimum benefits of pro­
tection from violence and theft . The stink of such s o c ie ty is,
after all, s till in our nostrils, and to argue that they have (or
had) no legal system would only involve th e rep e titio n of the
argument. Only if the rules failed to p rov i de these essen tial
82 GENERA L THEORY

benefits and protection for anyone - even for a slave-owning

group would the minimum be unsatisfied and the system
sink to the status of a s e t of m eaningless tab o o s . O f course no
one denied those benefits would h ave any reason to obey ex­
cept fear and would have every m oral reas o n to revolt .

VI

I should be less than candid if I did not , in c onclusion, con­

sider something which , I susp ect , most troubles tho s e wh o
react strongly against 'legal p ositivis m ' . Emphasis on the dis­
tinction between law as it is and law as it ought to be m ay be
taken to depen d upon and to entail what are called ' subj ect·
ivist ' an d 'relativis t ' or 'noncognitive ' the ories concerning the
very nature of moral judgments, moral distinctions , or 'values '.
Of course the Utilitarians themselves ( as distinct from later
positivists like Kelsen) did not countenance any such theories ,
however unsatisfactory their moral philosophy may app e ar to
us now. Austin thought ultimate moral p rinciples were the
commands o f G o d , known to us by revelation or through the
'index ' o f u tility , and B entham thought th ey were verifiable
prop ositions about utility . None the less I think ( though I
cannot prov e ) that insistence upon the distinction b etween
law as it is and ought to be has been, under the general head
of 'p ositivism ', c onfused with a m oral the ory according to
which statements of what is the case ( 'statements o f fact ')
belong to a category or type radically different from state­
ments of what ought to be ( 'value s tatements ') . It may there­
fore be well to dispel this source of c onfusion.
Th ere are many contemporary variants of this typ e of
m oral theory : according to some, ju dgments o f what ought
to b e , or ought to be done, either are or include as essential
elements expressions o f 'feeling', 'emotion ' , or 'attitudes ' , or
'subj ective preferences ' ; in others such judgments both express
feelings or emotions or attitudes and enj o in others to share
them. In other variants such judgments indicate that a p ar­
ticular case falls under a general principle or p olicy o f action
which the speaker has 'chos en ' or to which he is 'committed'
and which is itself not a reco gnition of what is the case but
analo gous to a general 'imperative' or command addressed
to all including the speaker himself. Common to all these
 PO S ITIVIS M , LAW , AND MORALS 83

variants is the insistence that judgments of what ought to be

done, because they c ontain such 'noncognitive ' elements, can­
not be argued for or established by rational methods as state­
ments of fact can be, and cannot be shown to follow from
any statement of fact but only from other judgments o f what
ought to be done in conjunction with some statement of fact.
We cannot, on such a theory , demonstrate, e.g. , that an action
was wrong, ought not to have been done, merely by showing
that it consisted of the deliberate infliction o f p ain solely for
the gratification o f the agent. We only show it to be wrong if
we add to those verifiable 'cognitive' statements of fact a gen­
eral principle not itself verifiable or 'cognitive ' that the inflic­
tion of pain ip such circumstances is wrong, ought not to be
done. Togethe r with this general distinction between state­
ments o f what is and what ought to be go sharp parallel dis­
tinctions between statements about means and s tatements of
moral ends. We can rationally discover and deb ate what are
appropriate means to given ends, but ends are n ot rationally
discoventble or debatable ; they are 'fiats o f the will', ex­
pressions of 'emotions ', 'preferences ', or 'attitu des '.
Against all such views (which are of course far subtler than
this crude survey can convey) others urge that all these sharp
distinctions between is and ought, fact and value, means and
ends, cognitive and noncognitive, are wrong. In acknowledg­
ing ultimate ends or moral values we are recognizing some­
thing as much imposed upon us by the character of the world
in which we live, as little a matter of choice, attitude, feeling,
emotion as the truth of factual judgments about what is the
case. The characteristic moral argument is not one in which
the parties are reduced to expressing or kindling feelings or
emotions o r issuing exhortations or commands to each other,
but one by which parties come to acknowledge after closer
examination and reflection that an initially disputed case falls
within the ambit of a vaguely apprehended principle (itself
no more 'subj ective' , no more a 'fiat o f our will' than any
o ther principle of classification) , and this has as much title to
be called 'cognitive ' or 'rational ' as any other initially disputed
classification of particulars.
Let us now supp ose that we accept this rejection o f 'non­
cognitive ' theories of morality and this denial of the drastic
distinction in type between statements of what is and what
84 GENERAL THEORY

ought to b e , and that moral ju dgments are as rationally de­

fensible as any other kin d o f ju dgments. What would follow
from this as t o the n ature o f the connection b etween law as it
is an d l aw as it ough t to be? Surely , from this alone, nothing.
Laws , however m orally iniquitous , would s till (so far as this
point is concerned) be laws . The only difference which the
acc eptance of this view of the nature of moral judgments
would make would be that the moral iniquity o f such laws
would be s o m ething that could be dem onstrat e d ; it would
surely follo w merely from a statement of what the rule re­
quired to be done that the rule was m orally wrong and so
ought not to be law or conversely that it was m orally desir­
abl e and ought to be law. But the demonstration of this would
not show the rule not to be (or to be) law. Proof that the
principles by which we evaluate or condemn laws are ration­
ally discoverable , and not mere 'fiats o f the will ' , leaves un­
touched the fact that there are laws which may have any
degree of iniquity or stupidity and still b e laws . And c on­
versely there are rules that have every m oral qualificati on to
be laws and yet are not laws.
Surely s om ething further o r more sp ecific mus t be said if
disproof o f 'noncognitivism ' or kindred th eories in ethics is
to be relevant to the distinction between law as it is and law
as it ought to b e , and to lead to the ab andonment at some
point or some softening o f this distinction . N o one has d one
m ore than Professor Lon Fuller of the Harvard Law School in
his various writings to make clear such a line of argument , and
I will end by criticizing what I take to be its central point. It
is a point which again em erges when we consider n ot those
legal rules o r parts o f legal rules the m e anings o f which are
clear and excite no debate, but the interpretation of rules in
c oncrete cases where doubts are initially felt and argum ent
develops about their m eaning. In no legal system is the scope
o f legal rules restricted to th e range o f concrete instances
which were present or are b elieved to h ave b e en present in
the minds o f legislators ; this indeed is one o f the imp o rtant
differenc e s b etween a legal rule and a command. Yet , when
rules are recognized as applying to instances b eyond any that
legislators did or could have c onsidered, their extension to
such new cases o ften presents itself not as a deliberate choice
o r fiat on the p art o f those who so interpret th e rule . It
 PO S IT I V IS M , LAW , AND MORALS 85

app e ars neither as a decision t o give the rul e a new o r extended

meaning nor as a gue s s as to what legislators , dead perhaps in
the eigh t eenth century , would h ave said had they b e en alive
in the twentieth century. Rather, the inclusion of the new case
under the rule takes its place as a natural elaboration of the
rule , as s omething implementing a 'purpos e ' which it seems
natural to attribute (in s ome sens e ) to the rule itself rather
than to any p articular person dead or alive. The utilit arian
description o f such interpretative e xtension of old rules to
new cases as judicial legislation fails to do justice to this
phenomenon ; it . gives no hint of the differences b etween a
delib erate fiat or decision to treat the new case in the same
way as past c;ases an d a recognition (in which there is little
that is delibe rate or even voluntary ) th at inclusion of the new
case under the rule will implement or articulate a c ontinuing
and identical purpose, hith erto less specifically apprehen ded.
P erhap s m any lawyers and ju dges will see in this language
som ething ,that precisely fits th eir exp erience ; others may
think it a 'ro m antic glo s s on facts b etter stated in the utilitarian
language of ju dicial 'legislation ' or in the mo dem American
terminology o f 'creative choice'.
To make the p oint clear Professor Fuller uses a non-legal
example from the philosopher Wittgenstein which is , I think,
illuminating.

S o m eone says t o m e : 'Show the children a gam e . ' I teach them gaming
with dice and t h e other says 'I did not m ean that sort of game. ' Must
th e exclusion of the gam e with dice have come before h is m ind when he
45
gave m e the order?

Something important does seem to me to be t ouched on in

this example. Perhap s there are the following (dis tinguishable)
p oints. Firs t , we norm ally do interpret n ot only what people
are trying to do but what they say in the light of assumed
common human obj ec tives , so that unless the contrary were
expressly indicate d we would n o t interp re t an instruction to
show a young child a game as a mandate to introduce him to
gambling even though in other contexts the word ' game'
would be naturally so interpreted. Secondly, very often, the
speaker whose words are thus interpreted might say : 'Yes,
th at 's what I mean [ or 'that 's what I meant all along' ] though

41 Fuller, 'Human Purpose and Natural Law', 53 J. Philos. 6 9 7 . 700 ( 1 95 6 ) .

86 G E N ERAL THEORY

I never th ough t of it until you put this particular case to m e . '

Third , when we thus recognize, p erhaps after argument or
consultation with others, a particular case not specifically en­
visaged b eforehand as falling within the ambit of some vaguely
express ed instruction, we may find this exp erience falsified
b y descrip tion of it as a mere decision on our part so to treat
th e p articular case, and that we can only describ e this faith­
fully as coming to realize and to articulate what we 'really '
want or our 'true purp o s e ' - phrases which Pro fessor Fuller
uses later in the same article.46
I am sure that many philosophical discussions of the charac ­
ter o f m oral argument would benefit from attention to cases
of the sort instanced by Professor Fuller. Such attention
would h elp t o provide a corrective to the view that th ere is a
sharp separation between 'ends ' and 'means ' and that in de­
bating 'ends ' we can only work on each other non-rationally,
and that rational argument is reserved for discussion of
'm eans ' . But I think the relevance o f his point to the issue
whether it is c orrect or wise to insist on the distinc tion be­
tween law as it is and law as it ough t to b e is very small indeed.
Its net effect is that in interpreting legal rules there are some
cases which we find after reflection to be so natural an elab­
oration or articulation of the rule that to think of and refer
to this as 'legislation', 'm aking law ' , or a ' fiat ' on our p art
would be misleading. S o , the argument mus t be, it would be
misleading to distinguish in such cases b etween what the rule
is and what it ought to be - at least in some sense of ought.
We think it ought to include th e new cas e and come to see
after reflection that it really does. But even if this way of
presenting a recognizable exp erience as an example of a fusion
between is and ought to be is admitte d, two caveats must b e
b orn e in mind. Th e first i s that ' ou ght' in this case need h ave
nothing to do with morals for the reasons explained already
in section III : there m ay be just the same sense that a new c as e
will implem ent and articulate t h e purpose o f a rule i n inter­
preting th e rules of a game or some hid eously immoral c o de
of oppression whos e immorality is appreciated by those c alled
in to interpret it. They too c an see what the 'spirit ' o f the
game th ey are playing requires in previously unenvisage d cases.

46 Ibid. at 7 0 1 , 7 02.
 P O SITIVI S M , LAW , AND MO RALS 87

More imp ortant is this : after all is said and done we must
rem ember how rare in the law is the phenomenon held to
justify this way o f talking, how exceptional is this feeling
that one way o f deciding a case is imp osed up on u s as the
only natural o r rational elab oration of som e rule. Surely it
cannot b e doubte d that , for most cases o f interp retation, the
language of choice between alternatives, 'ju dicial legislation '
or even 'fiat ' ( th ough n o t arbitrary fiat ) , b etter conveys t h e
realities o f t h e situation.
Within the framework o f relatively well-settled law th ere
j ostle too many alternatives too nearly equal in attraction
b etween which ju dge and lawyer must uncertainly pick their
way to m ak e a,ppropri ate here language which may well de­
scribe those dxperience s which we have in inte rp reting our
own or o thers ' p rinciples o f conduct , intention, or wish es ,
when we are not conscious of exercising a delib erate choic e ,
b u t rather of recognizing something awaiting recognition. T o
u s e i n the description o f t h e interpretation o f laws t h e sug­
gested tetnliriology of a fusion or inability to sep arate what is
law and ought to be will serve (like earlier stories that ju dges
only find, never make , law) only to conceal th e facts, that
here if anywhere we live am ong uncertainties between which
we have to cho ose, an d that the exis ting law imposes only
limits on our choice and not the choice itself.

POSTS CRIPT

S ee for criticism s and com m ents:

1. L . Fuller, 'Positivism and F idelity to Law : a Reply to P rofessor Hart ',
.
n Harv. L. R ev. 6 3 0 ( 1 958 ) .
2. R. A. D u ff, 'Legal O b li gation and the Moral Nature of Law', Juridical
R ev. 6 1 ( 1 9 8 0 ) .
3 . R. M. D workin, Ta k ing Rights S erio usly ( 2nd impression, 1 9 7 8 ) ,
ch. IV, and pp . xii, xiii, 1 05 8 , 1 24-6 , 3 2 6 - 7 , 3 3 2 -9 , 3 4 1 -3 , 3 48-9.
4 . J . Raz , Th e A u th o rity o f L aw (Oxford 1 9 7 9 ) , 3 7 -5 2 , 1 46 - 9 ;
'The P u rity o f the Pure T heo ry ' , R e vu e Internationale de Philosop hie
44 1 ( 1 9 8 1 ) .
5 . J . F innis, Natural Law and Na tural R ights (Oxfo r d 1 9 8 0 ) , 2 6 , 2 9 , 5 0 ,
3 6 3 ff. , 3 6 7 .
6. N. MacCormick, H.L.A. Hart (London 1 9 8 1 ) , 9 2 1 0 2 .
7 . D. Ly ons, ' On F ormal Justic e ' , 5 8 Co rn ell L. R ev. 8 7 3 ff. ( 1 9 7 3 ) .
8. D . Lyons, 'Moral Aspects of Legal Theory ' , 7 Midwest S tudies in
Philosophy 2 2 3 ( 1 9 82) .