VII.

—CRITICAL NOTICES

The Concept of Law. By H. L. A. HART Clarendon Press:

Oxford University Press. 1961. 21s.

MODERN British philosophy has long lacked any systematic treat-

ment of problems about law that would stand comparison with the
fundamental theories of other times or places—with the philosophies
of Savigny, Austin, Kelsen and others. Professor Hart has now
remedied this lack with a book which has all the virtues of elegance

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

and clarity that characterise British philosophical writing at its best.
If, like any work of first-rate philosophical importance, the book
forces one to re-think one's fundamental conceptions of ita subject-
matter, its value in this respect is not depreciated when one comes
to see that not all of its arguments are conclusive and not all of its
assertions are irrefutable.
Much of the book is taken up with a powerful critique of other
juristic theories, and especially of various positivist and natural law
doctrines. But the book's peculiar interest undoubtedly lies in
Hart's own theory, which gradually emerges from his criticisms of
Austin. Hart's thesis is that the difference between rules which
impose obligations, or duties, and rules which confer powers, is of
crucial importance in jurisprudence (237). Eules of the former kind
he calls primary, and of the latter kind, secondary. Under rules of
the primary type human beings are required to do or abstain from
certain actions, whether they wish to or not. Rules of the other
type are in a sense parasitic upon or secondary to the first; for they
provide that human beings may by doing or saying certain things,
introduce new rules of the primary type, extinguish or modify old
ones, or in various ways determine their incidence or control their
operations. Primary rules impose duties : secondary ones confer
powers, public or private. Primary rules concern actions involving
physical movement or changes : secondary rules provide for opera-
tions which lead not merely to physical movement or change, but to
the creation or variation of duties or obligations (78, 79). Thus
criminal law, apart from its provisions for sanctions, consists of
primary rules (27, 28,31,32). But the legal rulesdefining the ways in
which valid contracts or wills or marriages are made do not require
persons to act in certain ways whether they wish to or not. Instead
they provide individuals with facilities for realising their wishes, by
conferring legal powers upon them to create, by certain specified
procedures and subject to certain conditions, structures ofrightsand
duties within the coercive framework of the law (27). Similarly it is
secondary rules that lie behind the exercise of legislative and judicial
powers (31).
Hart claims that law can best be understood as a union of these
two diverse types of rule (237). He admite, of course, that many
395
396 CRITICAL NOTICES :

other distinctions could and, for some purposes, should be drawn.

But he insists that in the combination of these two types of rule
there -lies what Austin wrongly claimed to have found in the notion
of coercive orders, namely, ' the key to the science of jurisprudence '
(79). Again, his view is not that wherever the word' law ' is properly
used this combination of primary and secondary rules is to be found,
for he thinks it clear that the diverse range of cases of which the
word ' law' is used are not likened by any such simple uniformity,
but by less direct relations—often of analogy of either form or
content to a central case (79). But the union of primary and
secondary rtdes is at the centre of a legal system, even though it is

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

not the whole and as we move away from the centre we have to
accommodate elemente of a different character (96) ; and the
reason why Hart accords this union of elements a central place is
because of their explanatory power in elucidating the concepts that
constitute the framework of legal thought (79). Most of the features
of law which have proved most perplexing and have both provoked
and eluded the search for definition can best be rendered clear if these
two types of rule and the interplay between them are understood (79).
Unfortunately, however, the large claims that Hart makes on
behalf of hia distinction between primary and secondary rules do not
seem altogether justifiable. The distinction turns out to be scarcely
capable of playing the role for which Hart casts it.

II
This begins to become clear when we examine the extent to which
it is possible to class most legal rules as primary or secondary in
Hart's sense. Secondary rules, we are told, confer powers, whether
public or private. But has the concept of power quite so wide an
extension as Hart assumes ? No doubt usage is no more uniform
here than elsewhere, but there does seem something a bit odd in
talking about ' the power . . . conferred on individuals to mould
their legal relations with others by contracts, wills, marriages, etc.'
(28). Normally, legally recognised powers to act in certain specified
ways are conferred only on certain specified persons or categories of
persons. One named person may be given power of attorney over
the estate of another or the power to execute his will; trustees may
be given powers to buy and sell stock ; Chief Constables have the
power (discretion) to determine whether or not a motorist is to be
publicly prosecuted for a parking offence; the Colonial Develop-
ment Corporation has been given power to investigate and formulate
development projects ; and a County Court judge has power (juris-
diction) to try any case for the recovery of land with an annual
value not greater than £100. But when the law talks of things
that anyone may do, if they are properly qualified, the word ' capa-
city ' is more commonly used than the word ' power '. When I
married, no statute, no rule of common law, no private person
empowered me to do so. I did not need any such power : I merely
 L. j. COHEN: The Concept of Law: 397
had the requisite capacity. If you are insane, you have not the
capacity to make a will, and if you are under the age of 16 you
have not, in English law, the capacity to marry. Accordingly
people are elected or appointed to the exercise of powers, but
capacities are merely acquired whether naturally or by a change in
one's social circumstances or by a change in the law. Powers are
sometimes delegated, capacities never. Powers are surrendered or
revoked, but capacities are lost or restricted by disqualifications.
The people who wield powers may be impersonated, but with regard
to a capacity all one can do is to pretend to have it when one has not.
One may or may not have the capacity to exercise a power, but there

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

is no such thing as the capacity to exercise a capacity. In a text-
book on the law of marriage, the heading' capacity ' directs attention
to the rules determining who is capable of marrying ; in a textbook
on the law of local government, the heading ' powers ' would direct
attention to the rules determining what a local authority may do.
I am not suggesting that Hart at any time confuses the concept
of a power with that of a capacity, though he does not in fact articu-
late their differences. Indeed, when he remarks that ' behind the
power to make wills or contracts are rules relating to capacity or
minimum personal qualification (such as being adult or sane) which
those exercising the power must possess ' (28), it is quite clear that
he distinguishes a power from a capacity. Where he creates per-
plexity is about the extension of the term ' power', not about its
intension. Powers to make wills or contracts, or to marry, are,
except in special cases, powers unknown to the law, which treats
rather of capacities here. Presumably the reason for this is that
wills, contracts and marriages are transactions that most people
will want, and are suited, to perform anyway in a modern community,
so that the law treats most people as being capable of performing
them in any case and contents itself, in effect, with laying down
certain disqualifications. But where a power is conferred, as on
Chief Constables, or where provision is made for the conferring of
a power, as for the appointment of executors, the law assumes rather
that most people do not want, or are not suited, to exercise the power
unless certain special conditions are fulfilled. If a grant of power
(from some person or authority) is needed to make some testamen-
tary disposition, it is because the testator would not otherwise be
entitled to dispose of the property in question. Hence to assert that
there are powers of making wills or contracts is to spread an appear-
ance of unity over the whole range of what Hart classifies as secondary
rules, while little unity really enste, for the term ' power ' is extended
by Hart to cover almost everything regulated by law that is not a
duty. In a work designed to further the understanding of law (vii),
in a work that objects to the theory that all rules impose duties
because such a theory achieves uniformity at the price of distortion
(38), in a work that condemns the use of the term ' morality' as a
conceptual waste-paper basket for all rules of social behaviour that are
398 CRITICAL NOTICES

not legal (222)—this treatment of the term ' power' requires a

justification which Hart never adequately provides.
More specifically, in English legal jargon what is called a ' private
power ' is an authority which one person gives to another and is the
only form of power apart from public powers, which are authorities
from the Crown, Parliament, etc. for a public purpose (cf. A. W.
Motion, Pocket Law Lexicon, 1951, and G. R. Hughes, The Student's
Law Dictionary, 1936, s.v. Power). To call an individual's general
competence to make wills, contracts, etc., ' a private power', as
Hart does on several occasions (40, 77, 79, 96, 239), is therefore even
less helpful for students of English Law than it may be for the

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

students of systems in which no such expression as ' private power '
functions as a technical term. There is even a discrepancy between
Hart's use of the term ' public power ' and the ordinary legal use of
it in England. If public powers are only authorities from the Crown,
Parliament, etc., than one can hardly speak in general of what the
Crown or Parliament may themselves do as public powers. Ad-
mittedly usago may not be quite uniform here. Anson, The Law and
Custom of the Constitution (5th edn. 1922, p. 340) does speak of the
Crown's having at one time ' assumed to itself legislative powers ',
and somo dictionaries of English law, e.g. W. J. Byrne's (1923) and
Earl Jowitt's (1959) speak of private powers as those conferred on
private persons, whether by other persons or by statute. But a
power of ordinary citizens to marry, contract or make wills seems
quite unknown to the Common Law.
Certainly Hart errs in good company here. Bentham spoke of
people in general as being ' empowered' by the law to make wills
(Limits of Jurisprudence Defined, ed. C. W. Everett, 1945, p. 54),
and he said that buying aud selling was a mutual exercise of what he
called ' the investitive power' by two persons in favour of each
other (op. cit. p. 75). But Bentham at least had the excuse that he
was fundamentally at variance with the existing legal system and was
planning a new one, whereas Hart claims to be elucidating law as it is.
Again, Salmond instanced the right to make a will or the right to
marry one's deceased wife's sister as a private power (Jurisprudence,
11th edn. 1957, pp. 273 f.); Hohfeld spoke of a citizen's power
to create contractual obligations (Fundamental Legal Conceptions,
1923, p. 51); Roscoe Pound has spoken of the power to acquire
ownership of other people's property through ordinary legal trans-
actions (Jurisprudence, 1959, IV p. 99) ; and Kelsen, General Theory
of Law and State; trans. A. Wedberg 1946, p. 138) spoke of the
law as empowering people to make legal transactions. But all this
shows is that the term ' power' as Hart applies it is a term of
jurisprudence, not that it is a term of law ; and it is remarkable how
successfully English legal terminology has resisted this juristic
usage and retained its own, much narrower application of the term.
Nor is it easy to see how Hart could defend his use of the term' power'
in this juristic fashion (let alone his failure to draw attention to the
 L. j . COHEN : The Concept of Law 399
differences between this use and ordinary legal usage), since it is a
principle purpose of his theory to elucidate ' the specifically legal
concepts with which the lawyer is professionally concerned ' (95).

Ill

Wlat defences could Hart put up here ? I can think of five, but
there may be more, and better, ones.
First, the boundaries of strictly professional legal usage are not
absolutely clear, and Hart might help his case by pointing t.hia
out. No doubt the jargon of parliamentary draftsmen falls well

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

within those boundaries and much of Kelson's or Bentham's ter-
minology falls outside. But, in between, the strictness of legal
usage fades gradually as one passes from the language of textbook
writers and judges through that of barristers and solicitors into that
of jurists and philosophers. It would be surprising if there were not
at least some professional lawyers who from time to time use the term
' power ' as many jurists have done. Nevertheless Hart's use of that
term gets more and more difficult to illustrate as one passes from the
less strict fields of usage to the more strict ones. It is very difficult
indeed to find instances of it in recent English statutes or text-books.
Secondly, Hart might claim that even if modern legal systems do
not explicitly confer powers to marry or make contracte, yet, when
looked at from the outside, this is what they appear to do. ' Even
if the term " power " would be quite inappropriate here as a term of
law ', the defence would run,' yet as a term of sociology it is exactly
right. Weddings and business agreements would have little effect
in a modern community unless they had the force of law behind
them : it is the law that gives them what power they in actual fact
have.' But this defence is not open to Hart, if his distinction be-
tween primary and secondary rules is to fill the role for which he
casts it. For he tells us t h a t ' rules conferring private powers must,
if they are to be understood, be looked at from the point of view of
those who exercise them. They appear then as an additional
element introduced by the law into social life over and above that
of coercive control. This is so because possession of these legal
powers makes of the private citizen, who, if there were no such
rules, would be a mere duty-bearer, a private legislator. He is made
competent to determine the course of the law within the sphere of
his contracte, trusts, wills, and other structures of rights and duties
which he is enabled to build ' (40, 94). Thus for Hart's theory it is
essential that the term ' power ' be a term of law here, not of soci-
ology—a term suited to the internal point of view, not the external.
Indeed, Hart explicitly rejects the view that an analysis of these
powers should be conducted' in the terms of ordinary or " scientific "
fact-stating or predictive discourse'. ' To do justice to their dis-
tinctive, internal aspect', he tells us, ' we need to see the different
ways in which the law-making-operations of the legislator, the
400 curncAi NOTICES :

adjudication of a court, the exercise of private or official powers,

and other " acte-in-the-law " are related to secondary rules ' (96).
Moreover, Hart is careful to point out that his account of primary
rules as duty-imposing is in conformity with Anglo-American legal
usage, since in this usage, he says, the terms' duty ' and ' obligation '
are now roughly synonymous (238). One might therefore suppose
his account of secondary rules as power-conferring to claim an equal
accordance with legal usage.
Thirdly, Hart might defend his use of the term ' power ' with the
argument that, though it does involve some distortion here, yet the
difltortion is illuminating and harmless, because it points to the fact

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

that possession of these legal powers makes of the private citizen a
private legislator (40). But to extend the concept of legislation in
this way, pari passu with the concept of power, so that testators,
contractors, trustors, and even brides and bridegrooms become
private legislators, is radically to blur the concept of legislation.
A legislature may make new laws or repeal old ones. Thereby it
changes the content of a legal system. A private citizen may, by
his wills, contracts, trusts, marriages and other transactions bring
himself and others within the range of certain laws but he does not
change, or even amplify, their content. Hart uses a number of
blanket-phrases which have the effect of destroying this distinction.
Private legislators, like public ones, he says, ' determine the course
of the law ' (40), ' create structures of rights and duties' (27),
' introduce new rules of the primary type, extinguish or modify old
ones, or in various ways determine their incidence or control their
operations' (79). But to lose the everyday distinction between
changing the law, on the one hand, and bringing individual persons
or acts within its range, on the other, is a high price to pay for Hart's
use of the term ' power'. It is difficult to see what illumination
is gained therefrom.
Admittedly there was a time in the mid-nineteenth century when
the influence of Bentham and Ricardo may have favoured the
conception of a contract as a piece of private law (cf. Q. C. Cheshire
and C. H. S. Fifoot, The Law of Contract, 5th edn., 1960, p. 21).
The doctrine of laissez-faire required the utmost freedom of indivi-
duals to contract with one another as they wished, and if contracts,
rather than Parliaments, were thus to regulate how men lived, it
was not altogether absurd to think of contract as an alternative
legislative procedure to Parliament (though if the analogy were
pressed hard the doctrine of consideration in the English law of
contract might seem to suggest that all contractors were corrupt
legislators !) Bentham himself wrote that when the law gives a
power of occupation over a piece of land ' it in a manner adopts
the expression of my will, and turns it into a law' (op. tit. p. 73).
But the ideal as well as the reality of laissez-faire has long since
been destroyed by two important developments in modern society.
Generations of statutory restrictions on freedom of contract in
 L. J. COHEN: The Concept of Law 401
matters of employment, trade, and industry have made it clear
that freedom of contract is not a paramount consideration in the
eyes of the state, and that contract is therefore to be regarded as
an object of legislation rather than a creator of it. In a welfare-
state age there is not much illumination to be gained from thinking
of contracts as pieces of private legislation. In addition, the success
of standardised forms of contract—with their attitude of take it or
leave it—in many fields of social or economic activity has made it
apparent that even within the room for manoeuvre left open by the
state freedom is often completely one-sided. If there is any point
at all in talking about private legislation here, the original act of

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

drawing-up and promulgating a standardised form of contract should
be regarded as being like a legislative activity, not the individual
contracts that are subsequently made in conformity to such a
standard. Indeed in Bonsor v. Musicians' Union, where the rela-
tion between the parties was regulated by a document contractual
in form but in substance a series of peremptory rules, Lord Justice
Denning actually remarked of these rules that' they are not so much
a contract as we used to understand a contract, but they are much
more a legislative code laid down by some members of the union to be
imposed on all members of the union. They are more like bye-laws
than a contract' (cf. Cheshire and Fifoot, op. cit. p. 24). To classify
all contracts as pieces of private legislation, as Hart does, will thus
make it specially difficult to reveal the true nature of a very large
number of contracts in modern law, though even here we must be
careful to say, as Lord Denning did, that the rules of a standardised
form of contract are like bye-laws, not that they are such. Hart
claims in support of his view that Kelsen has shown ' how many of
the features which puzzle us in the institutions of contract or property
are clarified by thinking of the operations of making a contract or
transferring property as the exercise of limited legislative powers by
individuals ' (94, 245). But it seems to me that my criticisms of
Hart here are equally valid against Kelsen. Moreover, when one
looks at the passage of Kelsen to which Hart refers (op. cit. p. 138),
it is difficult to find any substantial use made of the analogy between
contract and legislation except to support the thesis that, in general,
individual legal duties under criminal law are stipulated directly
while under civil law individual legal duties are stipulated only
indirectly, through the medium of legal transactions. But this
general principle does not in fact help us very much in distinguishing
between civil and criminal law. Kelson thought that the typical
exception to his principle was the duty to repair damage caused
illegally, and Hart himself alludes to duties of care under the law of
tort (27). But in fact the principle has been well-nigh destroyed by
modern welfare-state legislation. It is just in the course of legal
transactions—in the sale of goods, the hiring and firing of employees,
the formation of companies, the purchase of overseas investments,
and so on—that so many important duties under criminal law now

26
402 CRITICAL NOTICES :

arise. If Kelsen's principle of distinction between civil and criminal

law is the kind of' clarification ' we shall get by adopting the analogy
between contract and legislation, we shall be better off if we do not
adopt this analogy at all.
Fourthly, perhaps, in order to save his dichotomy of primary and
secondary rules, Hart might claim that the use of the particular word
' power ' in his account of secondary rules was not important. Any
other term would do, he might say, that would cover the same range
of rules. On this defence, however, the supposed analogy between
contracts, wills, marriages etc. on the one side, and legislation, on the
other, would need to be dropped, and it would not be at all clear what

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

there was in common between rules determining capacities for mar-
riage or contract and rules conferring legislative or judicial powers.
The attempt to find another term or expression embracing both, such
as ' right', ' liberty ' or ' privilege ', would encounter similar diffi-
culties ; and in the end it is doubtful whether anything would be found
that all Hart's secondary rules have in common except the more or
less negative property that they are not in themselves duty-imposing
rules though related to them in various ways. For example, at one
point (9) Hart distinguishes between mandatory rules and rules which
indicate what people should do to give effect to the wishes they have.
But, though the latter description applies to many of the roles pre-
scribing the procedures, formalities and conditions for the making of
marriages, wills or contracts, it hardly applies to rules conferring
legislative or judicial powers. Elsewhere Hart points out that if a
transaction does not conform to a relevant secondary rule the trans-
action is invalidated but the non-conformity to rule does not itself
constitute a legal offence (28-31). There is a nullity, but not a
breach of duty. Apart from the ' complication ' that in the interest
of public order a judicial decision that has been given in excess of
jurisdiction stands, at least in England, until quashed by a superior
court, the position is just the same in this respect, Hart argues, for
what he calls ' public powers ' as for what he calls ' private powers '
(30). But the difference Hart mentions seems a little more than just
a complication. The fact is that, even though a decision in excess
of jurisdiction is liable to be quashed on appeal, it will nevertheless
remain valid for ever if no appeal against it is made. But wills and
marriages are not like this at all. If I do not comply with s. 9 of the
Wills Act 1837,1 have not made a will at all: no action by others is
required in order to nullify it. Legislation presents us with yet
another situation. If a measure proposed before a legislature fails
to secure the requisite majority of votes, which Hart regards as a
breach of a secondary rule (31), no action has been nullified or in-
validated. The measure was properly proposed, seconded, opposed
and debated. The vote was properly taken. All that has happened
is that the proposal has been unsuccessful. Admittedly if a local
authority disobeys certain rules about making bye-laws the result
may in some cases be a nullity. But the difference between this
 L. j . COHEN : The Concept of Law 403
nullity and the nullity of an unwitnessed will is as great as the
difference between the corresponding valid transactions, of which one
is an act of legislation and the other, pace Hart, is not. Of course,
Hart would be right to insist here, as he does in his book (15, 23*),
that for a general term, like ' law ' , ' crime ' , ' just', ' good ', etc., to
be correctly used it is not essential that the range of instances to
which it is applied should all share common qualities. But he must
at least suggest some central model, or some range of useful analogies,
which would justify grouping together such prima facie heterogen-
eous rules as those he includes under his rubric o f secondary rules ',
and it is already clear that the alleged analogy between contract and

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

legislation is somewhat remote.
Fifthly, therefore, Hart might be content with a rather more
negative account of what secondary rules have in common, viz.
their not being primary ones though related to them in various ways,
and he would then rely on a disjunctive account of these relations in
order to give body to his distinction. Secondary rules, he might
say, either determine how primary rules, or other secondary ones may
be created, changed, or abolished, or how they may be administered,
or how people may come within their range. They are all' parasitic '
on primary rules (79), though parasitic in different ways. But the
trouble then is that it is difficult to see why one should not treat what
Hart calls secondary rules merely as conditions on primary ones.
If they can thus be regarded as various different kinds of parasites of
primaries, they might just as well be treated as parts of primaries.
After all, Hart himself sometimes speaks of secondaries as ' defining
the ways in which valid contracts or wills or marriages are made ' (27),
or asserts that their concern is ' to define the conditions and limits
under which the court's decisions shall be valid'. The ease with
which it is possible to describe the function of secondary rules as a
denning one suggests that there may be better excuse than Hart
allows for thinking of the law as a system of primary rides only—
as a system of rules that impose duties and determine sanctions for
their breach.
Hart does in fact examine three prima facie plausible ways of
reducing what he calls secondary rules to what he calls primary ones.
He grants that there are pointe of resemblance between these two
sorts of legal rules. In both cases actions may be criticised or
assessed by reference to the rules as legally the ' right' or ' wrong '
thing to do. ' Both the power-conferring rules concerning the mak-
ing of a will,' he says,' and the rale of criminal law prohibiting assault
under penalty constitute standards by which particular actions may
be thus critically appraised ' (32). But he denies the possibility of
reducing secondaries to primaries.
According to one theory the nullity which ensues when some
essential conditions for the exercise of a power is not fulfilled is like
the punishment attached to criminal law. It is a threatened evil or
sanction exacted by law for breach of the rule. We are to think of
404 CRITICAL NOTICES :

the rule providing that a will without two witnesses will be inoper-
ative, as moving testators to compliance with s. 9 of the Wills Act,
just as we are moved to obedience to the criminal law by the thought
of imprisonment. Against this theory Hart rightly objects not only
that in many cases nullity may not be an evil to the person who has
failed to satisfy some condition required for legal' validity (33).
More seriously, he points out, in the case of a rule of criminal law we
can identify and distinguish two things : a certain type of conduct
which the rule prohibits, and a sanction intended to discourage it.
But we can hardly consider in this light such desirable social activities
as men making each other promises which do not satisfy legal require-

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

ments as to form. Again, in the case of criminal law we can, in a
sense, subtract the sanction and still leave an intelligible standard of
behaviour which it was designed to maintain. But we cannot
logically make such a distinction between the rule requiring com-
pliance with certain conditions, e.g. attestation for a valid will, and
the so-called sanctions of' nullity '. In this case, if failure to comply
with this essential condition did not entail nullity, the rule itself
could not be intelligibly said to exist even as a non-legal rule (34).
Hart also considers Kelsen's theory that what is ordinarily
thought of as the content of law, designed to guide the conduct of
ordinary citizens, is merely the antecedent or ' if-clause' in a rule
which is directed not to them but to officials, and orders the latter
to apply certain sanctions if certain conditions are satisfied. All
genuine laws, on this view, are conditional orders to officials to apply
sanctions. The provisions of the Wills Act requiring two witnesses
would appear as a common part of many different directions to
courts to apply sanctions to an executor who, in breach of the
provisions of the will, refuses to pay the legacies : ' if and only if
there is a will duly witnessed containing these provisions and if . . .
then sanctions must be applied to him. The provisions of the U.S.
constitution as to the law-making power of Congress would merely
specify the general conditions under which courts are to apply
sanctions. But, as Hart rightly objects, we shall conceal the
characteristic way in which the rules of criminal law function if we
concentrate on, or make primary, the rules requiring the courts to
impose sanctions in the event of disobedience. These latter rules
are only ancillary, in that they make provision for the breakdown or
failure of the primary purpose of the system. After all, the character-
istic technique of the criminal law is to designate by rules certain
types of behaviour as standards for the guidance either of the members
of society as a whole or of special classes within i t : they are
expected without the aid or intervention of officials to understand
the rules and to see that the rules apply to themselves and to con-
form to them.
But Hart also examines a third way of treating all legal rules as
rules imposing duties, which he attributes to Bentham (239).
According to this theory the criminal law and all other laws imposing
 L. J . COHEN: The Concept of Law 405
duties would be regarded in the normal way, and not as Kelsen
regards them. But ' rules which confer legal powers on private
individuals ' are, for this as for Kelsen's theory, ' mere fragments of
the real complete laws—the orders backed by threats. These last
are to be discovered by asking : what person does the law order to do
things, subject to a penalty if they do not comply V When this is
known, the provisions of such rules as those of the Wills Act, 1837, in
relation to witnesses, and other rules conferring powers on individuals
and defining the conditions for valid exercise of these powers, may be
recast as specifying some of the conditions under which ultimately
such a legal duty arises. Rules relating to the formation of contract

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

will similarly appear as mere fragments of rules ordering persons, if
certain things are the case or have been said or done (if the party is
of full age, has covenanted under seal or been promised consideration),
to do the things which by the contract are to be done. Similarly the
rules conferring legislative powers are represented by the ante-
cedents or if-clauses of rules ordering ordinary citizens, under threat
of sanctions, to do things and not merely, as in Kelsen's theory, as
the if-clauses of directions to officials to apply sanctions (37).
Against this Benthamite theory Hart's objections do not appear
very strong. In regard to rules conferring ' private powers' he
claims that such rules are thought of, spoken of and used in social
life differently from rules which impose duties, and they are valued
for different reasons. What other testa for difference in character,
he asks, could there be ? (41). But the trouble here is that, if I am
right, there are no rules conferring private powers of the kind that
Hart assumes : there are only rules regulating capacity. But there
are also, in regard to contract, say, rules of offer and acceptance,
rules of consideration, rules of content, rules of discharge, and so on.
All these rules are thought of, spoken of and used in social life differ-
ently from rules of capacity and they are valued for different reasons.
So if we are to distinguish rules of capacity from rules imposing
duties for the reasons that Hart gives in respect of his supposed
power-conferring rules, we should also distinguish all these other
kinds of rules both from one another and from rules of capacity and
from rules imposing duties. There would be no neat dichotomy of
primary and secondary rules, but as many different kinds of rules as
the complexity of our legal system allows. Hart seems to treat all
these other rules as conditions on the exercise of a supposed power to
contract. For him they seem to be antecedents of a conditional rule
which has the form : ' A contract may be made if and only if. . . . '
But he gives no reason why we should regard this pattern of condi-
tionalisation as somehow less objectionable than Bentham's. Simi-
larly with regard to wills, marriages, trusts, etc.—if rules of form,
say, may be treated as conditions'on the exercise of a power, then
rules of power (or, more properly, rules of capacity) may be regarded
as conditions on the existence of circumstances in which duties arise.
One can either go all the way in one's logical reconstruction of the
406 CRITICAL NOTICES :

law, and treat all law-sentences as either duty-imposing rules or

conditions (' parasites') on such rules : or alternatively one can
draw as many distinctions and construct as many different schemes of
classification as the richness of the law allows and the subtlety of our
jurisprudence renders profitable. Indeed even in criminal law few
duties or obligations arise that English lawyers would explicitly
recognise as such: murder is a crime but the law imposes no duty
not to murder.
Hart also argues against Bentham's reduction of the rules that
confer and define legislative and judicial powers in the public sphere
to statements of the conditions under which duties arise. He rightly

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

remarks that those who exercise these powers to make authoritative
enactments and orders use these rules in a form of purposive activity
utterly different from performance of duty or submission to coercive
control. But he is not quite right to infer from this that to represent
such rules as mere aspects or fragments of the rules of duty is, even
more than in the private sphere, to obscure the distinctive character-
istics of law and of the activities possible within its framework (41).
For those who wish, like Bentham, to represent such rules as frag-
ments of the rules of duty are certainly not suggesting that the
exercise of legislative power, say, should be regarded as a species
of duty-performance or submission to coercive control. Certainly
Bentham himself does not say, or suggest, in the work to which
Hart refers (Limits of Jurisprudence Defined), that legislators are
' obeying rules when, in enacting laws, they conform to the rules
conferring their legislative powers' (109). Just the opposite:
Bentham and those like him are saying that rules conferring legisla-
tive powers are ' mere aspects or fragments of the rules of duty ', in
Hart's own terms, not complete rules of duty. On the Benthamite
view rules conferring legislative powers turn into the antecedents of
conditional rules with a form like, say, ' if the Queen in Parliament
has ordained that . . . , then it is obligatory that . . .', not into
categorical rules of the form ' It is obligatory on the Queen in Parlia-
ment to . . .'. There need be no distortion or obfuscation here.
Rather the true nature of duty-imposing rules is brought into relief :
they are only valid if the Queen in Parliament . . . or if . . . etc.
After all, the purpose of Hart's legal philosophy is presumably to turn
a searchlight on the nature of the law rather than on the nature of
legislative activity.

IV
Hart tries to sustain his concept of secondary rules by stressing
how important was the introduction into society of rules enabling
legislators to change and add to the rules of duty, and judges to
determine when the rules of duty have been broken. This step, he
says, may fairly be considered as the step from the prelegal into the
legal world (41). Hart remarks that it is possible to imagine a
society without a legislature, co'irts or officials of any kind. Indeed,
 L. j . COHEN : The Concept of Law 407
he says, there are many studies of primitive communities which not
only claim that this possibility is realised but depict in detail the
life of a society where the only means of social control is that general
attitude of the group towards its standard modes of behaviour in
terms of which Hart has characterised rules of obligation (89). Or
at least there are studies of primitive communities which describe
them as approximating to this state (244). Hart refers to such a
social structure as one of primary rules of obligation, and he holds
that if a society is to live by such primary rules alone there are
certain conditions which, granted a few of the most obvious truisms
about human nature and the world we live in, must clearly be

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

satisfied. The rules must contain some restrictions on the free
use of violence, theft and deception; those who reject the rules
except where fear of social pressure induces them to conform must
be in a minority ; and the community must be small, closely knit
by ties of kinship, common sentiment and belief, and placed in
a stable environment (89). Such a system of social control by
primary rules alone has three main defects. If doubts arise as
to what the rules are or as to the precise scope of some given
rule, there will be no procedure for settling this doubt, either by
reference to an authoritative text or to an official whose declarations
on the point are authoritative. This defect in the simple social
structure of primary rules Hart calls its uncertainty. A second defect
he calls the static character of such rules (90): there is no means in
such a society of deliberately adapting the rules to changing circum-
stances, either by eliminating old rules or introducing new ones.
The third defect he discerns in this simple form of social Me is the
inefficiency of the diffuse social pressure by which the rules are
maintained. Punishments for violations of the rules are left to
the unorganised efforts of private individuals ; and, even more
seriously, disputes as to whether an admitted rule has or has not been
violated will always occur and are liable to continue interminably,
if there is no agency specially empowered to ascertain finally, and
authoritatively, the fact of violation (91).
The remedy for each of these three main defects in the simplest
form of social structure consists, according to Hart, in supplementing
the primary rules of obligation with secondary rules. The three
remedies together are enough to convert the regime of primary rules
into what is indisputedly a legal system, and thus it can be shown
how ' law may most illuminatingly be characterised as a union of
primary rules of obligation with such secondary rules ' (91). The
simplest form of remedy for the uncertainty of the regime of primary
rules is the introduction of what Hart calls a ' rule of recognition '.
This will specify some feature or features possession of which by a
suggested rule is taken as a conclusive indication that it is a rule of the
group and is to bo supported by the social pressure which the group
exerts. It may say no more than that an authoritative list or text of
the rules is to be found in some written document or carved on some
408 CRITICAL NOTICES :

public monument. Or in a more developed legal system it may

identify as authoritative any rule enacted by a specified body, any
rule customarily observed in certain fields, and any rule acknow-
ledged in judicial decisions ; and provision may be made for possible
conflict between these criteria by their arrangement in an order of
superiority, as by the common subordination of custom or precedent
to statute (92). The remedy for the static quality of the regime of
primary rules consists in the introduction of what Hart calls ' rules of
change '. The simplest form of such a rule is that which empowers
an individual or body of persons to introduce new primary rules for
the conduct of the life of the group, or of some class within it, or to

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

eliminate old rules ; and more complex rules may limit, these powers
in various ways or define the procedure to be followed in legislation
(93). The third supplement to the simple regime of primary rules,
intended to remedy the inefficiency of its diffused social pressure,
consists of secondary rules empowering individuals to make authori-
tative determinations of the question whether, on a particular occa-
sion, a primary rule has been broken. Besides identifying the indivi-
duals who are to adjudicate, such rules will also define the procedure
to be followed, and they will normally be supplemented by yet further
rules conferring upon judges, where they have ascertained the fact
of violation, the exclusive power to direct the application of penalties
by other officials (94, 95).
But it would in several respects be not quite right to characterise
the shift from a pre-legal to a legal system in terms of the addition
of secondary rules to original primaries. Secondariness is neither
a necessary nor a sufficient qualification for being a rule that needs
to be added here.
It ia not a necessary qualification, since even some of the rules
that Hart mentions are not secondary rules according to his defini-
tion of ' secondary '. The rules he calls rules of recognition are not
rules that confer powers, whether public or private : they set up
criteria. They determine the sources of law : they do not give
power to someone to make it, like the rules that Hart calls ' rules of
change '. Sometimes, e.g. in the Roman-Dutch law of South Africa,
a document or textbook which had little or no authority in a given
community at the time it was written, or could have had no authority
at all because the community scarcely existed, is later recognised by
that community as an authoritative source of law. In such a
case it borders on absurdity to suppose that the relevant rule
of recognition constitutes a retrospective, and probably post-
humous, grant of legislative powers to the author of the document
or textbook. Perhaps Hart does not wish us to suppose this.
But how then can we treat his rules of recognition as being secondary
rules ? Again, the law of evidence is a very important branch of
any modern legal system, and volumes are written on it as large as
any on the law of contract, family or real property. Hart does not
mention it specifically, and presumably it is to be taken as included
 L. j . COHEN : The Concept of Law 409
in what he calls the rules of adjudication, which,' besides identifying
the individuals who are to adjudicate, . . . will also define the
procedure to be followed ' by them (94). But this implies that the
law of evidence should be treated as a mere part or adjunct of the
rules conferring adjudicatory powers, and a student of English
law at least would find out very little about the rules of evidence
if he merely searched the statutes and precedents conferring
jurisdiction. Look through the table of statutes and the table
of cases in a standard work like Phipson on the Law of Evidence
(9th edn. 1952), and you will find that only a minute proportion
of the references there are in any way concerned with matters of
jurisdiction. Indeed it is not only a historical fact that the law of

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

evidence has often developed independently of rules of jurisdiction :
it is also arguable that something important would be lost if the law
of evidence could not so develop. Hart might perhaps object here
that rules of evidence do not lie at what he calls ' the centre of a legal
system ' (96). But rules of evidence are not of much less importance
than rules of jurisdiction in remedying what Hart calls the inefficiency
of a pre-legal system. They are essential if the facts about violations
of primary rules are to be determined in a way that will inspire
general respect. No doubt the rules of evidence, too, could become
antecedents in a systematic reconstruction of the law on Benthamite
lines as a set of conditional rules of duty. But Hart's method of
making rules of evidence the antecedents of conditional rules of
jurisdiction is no better suited than Bentham's to the project of
showing how ' the specifically legal concepts with which the lawyer
is professionally concerned . . . [are] best elucidated ' (95).
These two examples—rules of recognition and rules of evidence
—make it clear that secondariness is not a necessary qualification for
being a rule that normally needs to be added in order to transform a
pre-legal system into a legal one. Nor is it a sufficient one. Certain
rules of a kind that Hart elsewhere classifies as secondary (27) appear
in almost every pre-legal system. Hart himself refers to studies like
B. Malinowski's Crime and Custom in Savage Society (1926) and A. S.
Diamond's Primitive Law (1935) which depict in detail the life of
a society where there are no courts, officials or legislature (89, 244).
But are there very many such studies which depict a society where
there are no rules of marriage ? Indeed it is very often the case
that the more primitive and pre-legal a community appears, the more
elaborate are its rules determining capacity for marriage. No
small part of cultural anthropology is concerned with the structure
of these rules and their relationship to the life of a community
as a whole. Malinowski (op. cit. p. 35) remarks of the culture he
studied that ' marriage establishes not merely a bond between
husband and wife, but it also imposes a standing relation of mutuality
between the man and the wife's family, especially the wife's brother '.
Even Diamond remarks (op. cit. p. 192) that there are already rules
of marriage in the savage societies which do not possess courts.
410 CRITICAL NOTICES :

The moral institution of a promise is thus far from being the only
form of ' power-conferring' rule that is to be found in a pre-legal
society, as Hart at one point implies (94). If there are obligations at
all in such a society, then some of them may be created by marriages,
transfers of property, and so on. It is no more an amenity of law
that it enforces obligations so created, than that it enforces other
obligations : if unorganised social pressure can enforce an obligation
not to murder or steal, it can also enforce the obligations that arise
out of marriage. Nor is it open to a defender of Hart's theory to
argue that even if all actual primitive societies have rules of marriage,
yet one can at least imagine a primitive society that has none.
Hart himself remarks that ' in an extreme case the rules may be

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

static in a more drastic sense. . . . Not only would there be no way
of deliberately changing the general rules, but the obligations which
arise under the rules in particular cases could not be varied or
modified by the deliberate choice of any individual' (90). For, if
people are to be released from certain obligations or if they are to
transfer to others the benefits which would accrue from performance
of their own obligations,' there must be rules of a sort different from
the primary rules ' : there must be rules creating changes in the
initial positions of individuals under the primary rules of obligation
(91). However Hart also recognises that such a state of affairs is
' never perhaps fully realised in any actual community', and it
is doubtful if we can gain much understanding of the actual transition
from pre-legal to legal by considering a state of affairs that is not only
never in fact folly realised but is also fundamentally alien to much of
the primitive human life that is revealed in anthropological studies.
No doubt rudimentary or embryonic organs of legislation or adjudica-
tion may be found in very many primitive societies that we should
still wish to call pre-legal, but if the institution of marriage exists at
all in such societies it is relatively unlikely to be as rudimentary or
embryonic.
It is true that I have myself already argued that the law of
marriage, contract, etc., does not consist of rules conferring powers on
individuals to marry, make contracts, etc., and is therefore mis-
classified by Hart as a set of secondary rules in his sense. But a
defender of Hart's theory would seem to be caught in a dilemma
here. If he gives up treating the law of marriage, contract, etc., as
a set of secondary rules, then the extension of the term ' secondary
rule ' is more or less confined to rules of legislation, jurisdiction and
administration. Since this looks suspiciously like the three branches
of law traditionally treated as constitutional law, the term ' consti-
tutional ' would then perhaps be an apter designation of these rules.
But then Hart's claim that his distinction between primary and
secondary rules is the key to the science of jurisprudence and has
great explanatory power must be reduced here to the somewhat less
exciting proposition that the addition of a constitution, setting up
legislature, judiciary and executive, is the most important element
 L. j. COHEN: The Concept of Law 411
in the transition from a legal to a pre-legal system. On the other
hand, if a defender of Hart's theory insists on treating the rules of
marriage, contract, etc., as power-conferring rules, then he must not
claim that the primary-secondary distinction caste any great light on
this transition. There are secondary rules in any ordinary pre-legal
system as well. What is important about the secondary rules re-
quired for transition to a legal system is not the fact that they are
secondary (indeed, some of the rules so required, as we have seen,
are not secondary ones at all) but the fact that they set up a con-
stitution. To suppose that the secondariness of the rules is im-
portant here, rather than their constitutional character, is like

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

supposing that what is important in creating the traffic problem of
the mid-twentieth century is the invention of the wheel—if I
may adapt a simile of Hart's (41)—rather than the mass-production
of power-driven vehicles. The only way of escape through the
horns of this dilemma seems to be to assume, as Hart sometimes
does (89, 94), not only that the law of wills, contracts, marriages,
property, etc. consists of power-conferring rules, but also that wills,
contracts, marriages, property, etc. are institutions which exist only
in a fully developed legal system, just like legislatures and law-courts.
But, as I have already suggested, such an assumption is not in
accordance with the findings of modern anthropology.

In short, it is not at all clear how anyone could successfully

defend Hart's claim to have found the key to the science of juris-
prudence in his distinction between primary and secondary rules.
But it is very much to be hoped that, if Hart himself still believes
this, he will amplify his theory at some convenient opportunity and
show how it is immune to criticisms of the kind that at present seem
relevant. So far as can at present be seen the main merit of the
theory lies in the fact that it forces anyone who wants to criticise
it to rethink his own attitudes over the whole field. That is no
inconsiderable merit for a philosophical theory to have, but it is
not the merit that Hart himself claims for his theory. Indeed there
may be room for doubt about how far Hart's philosophical ideal of
elucidating legal concepts as they are is compatible with the dis-
covery of any single principle as the key to the science of juris-
prudence.
Because I have been putting forward these criticisms of Hart's
central thesis, I have not mentioned in detail the very large number
of more defensible points that Hart makes in other connections.
Perhaps the final chapter, on international law, is also a little dis-
appointing, since while Hart refutes various arguments for denying
that international law is a legal system, or for denying value to
the question whether it is or is not a legal system, he does not
discuss in detail any analogies of content between municipal and
412 CRITICAL NOTICES :

international law, despite his claim that in such analogy no other social
rules are so close to municipal law as those of international law.
But in his analysis of the difference between a rule and a command ;
in his replacement of the concept of legal sovereignty by the concept
of an ultimate rule of recognition ; in his insistence on the importance
—vis-a-vis juristic positivism—of the difference between the internal
and the external aspect of rules ; in his discussion of the relation
between law and morality ; in his exploration of the narrow path
between formalism and scepticism aa rival attitudes to the relation
of judicial decisions to rules; and in his very fair examination of
how solid a basis there is for theories of natural law—his book

Downloaded from http://mind.oxfordjournals.org/ at Monash University on June 4, 2015

has made a considerable number of profoundly important contribu-
tions to the philosophy of law.
L. JONATHAN COHEN

Induction and Hypothesis : A Study of the Logic of Confirmation.

By S. F. BARKER. Cornell University Press (London : Oxford
University Press). Pp. 203 + xvi. 22s.
THE argument of this study of non-formal reasoning is conducted
largely in abstract philosophical and logical terms. It examines
doctrines and theories about arguments rather than actual instances
of reasoning. In most cases the doctrines and theories are sketched
rather than elaborated in detail and counter-arguments and objections
are treated in the same way. This method of argument allows
Professor Barkor to compress a great deal of material into a short
book, but at the cost of some loss of conviction at key points in the
discussion. In particular the method sometimes leads to a failure to
analyse questions in sufficient detail so that important distinctions
disappear. In this Barker is not alone, since many writers on
induction have created difficulties by seeing only one hard question
where a more detailed analysis can extract smaller questions which
are not hard to answer. The lack of detailed examination of actual
instances of scientific reasoning also contributes to the lack of con-
viction in the argument, since whatever is said about the logic of
generalizing must be tested against instances of the most successful
generalizing we know, the laws of nature.
Four phases can be distinguished in this discussion :
Phase I: Attempts by linguistic analysts to dissolve the tradi-
tional ' problem of induction ' are described and criticized ; with the
conclusion that there is a problem about the legitimacy of generalizing,
and this residual problem is very closely similar to the traditional
' problem of induction ' ; that is the problem is to find a formal
scheme for valid generalizing.
Phase II: The classical solutions to the traditional problem are
examined and dismissed in the standard manner. The introduction
of over-riding premises to construct the inductive syllogism ' Over-