Victoria University of Wellington Law Review
In his lecture Professor Simpson places John Salmond's jurisprudential books, his First Principles of Jurisprudence and his Jurisprudence text book within both the context of the English text book tradition and the more specific context of the English tradition of the general jurisprudence book. Having pointed out that Salmond’s major text book was indeed written for students, Professor Simpson examines Salmond’s explanation of judicial reasoning and whether it can differentiated from other modes of thought. While ultimately Professor Simpson suggests that John Salmond's account of legal thought is not complete, he concludes that it can be favourably compared with other attempts to resolve the question. In particular, Professor Simpson contrasts John Salmond and his contemporaries’ efforts favourably with the efforts of the prominent Oxford legal philosopher HLA Hart. For his own part, Professor Simpson ultimately considers that perhaps there is not so much special about legal thinking when compared to other processes of rational decision making in everyday practical affairs.
I first learnt of Salmond as an undergraduate in Oxford from 1951 to 1954. At this time Salmond's two books, Salmond on Torts and Salmond on Jurisprudence, were still much favoured, and I bought and used both. Salmond on Torts was at this time available in an edition by Stallybrass, always known as Sonners, formerly the Principal of Brasenose College. He achieved a certain distinction by falling or, it was said, being pushed to his death from the 4.45 train from Paddington to Oxford whilst he was Vice-Chancellor. That was in 1948. Salmond on Torts continued to be kept in print for many years later than this, the last editors being Robert Heuston and R A Buckley. Heuston had such respect for its quality that he spoke of assuming the editorship in mystical terms: the mantle fell on him. I understand that there is currently no plan to revive it, but the type of legal literature to which it belongs is, at least outside the American law schools, alive and well. Elsewhere I have attempted to provide a history of the legal treatise. In England treatises predate the revival of organized legal education in or outside the universities. But by the time Salmond on Torts appeared in 1907 – its full title was The Law of Torts. A Treatise on the English Law of Liability for Civil Injuries – the university law schools in England were slowly becoming more prominent, and Salmond's book, as well as being much in use by practitioners, became a standard university text. It enjoyed a life of over close to a century. The curious practice of producing up dated editions of legal works, in which the original text has been almost or even completely buried, is of course related to a somewhat ill thought out desire to retain for the work the personal authority associated with its original author. In some cases the modern editions come to resemble nothing so much as an archaeological site, with layers of rubbish piled on other layers. This never happened to Salmond on Torts, which retained until the end passages by its original author.
Salmond's other treatise, Jurisprudence or The Theory of the Law, which I shall just call Salmond's Jurisprudence, was first published in 1902. It continued to appear in up dated editions until 1972. Its latest editor was Glanville Williams. When I was a student the current edition was that of 1947. It is on this work I shall concentrate my remarks. It belongs to a form of legal literature which is today virtually extinct: the comprehensive treatise on jurisprudence. When it first appeared there were in print, in addition to John Austin's grim Province of Jurisprudence Determined and little read Lectures on Jurisprudence three other works which have some similarity with it. The earliest was Sir William Markby's. Elements of Law Considered with reference to the Principles of General Jurisprudence (1871), then in its 5th. Edition. The latest edition was the 6th of 1905. Next to appear was Sir Erskine Holland's highly successful The Elements of Jurisprudence (1880), and was by then in its ninth edition. It continued to appear until 1924, but was never kept in print after Holland's death. The third was Sir Frederick Pollock's A First Book of Jurisprudence for Students of the Common Law, which first appeared in 1896. Markby, Holland and Pollock were members of the group which founded the Law Quarterly Review; all in various ways were dedicated to the idea that the law could be studied scientifically, as based on a body of principles. They differed in that Markby had extensive judicial experience and a highly critical mind; Holland was basically an expounder of views derived from Austin and Bentham, and the extraordinarily learned Pollock was a somewhat romantic lover of the common law system with a sharp and original mind, but with little direct practical experience of legal practice. All have entries in the new Oxford Dictionary of National Biography; bizarrely Salmond does not. So far as the United States was concerned the earliest such work was JC Gray's The Nature and Source of the Law, which first appeared in 1909. Gray, it will be recalled, invented the rule against perpetuities in its modern form. His book was based on a course of lectures delivered in the Harvard Law School from 1896 onwards. In addition to these rivals Salmond's own The First Principles of Jurisprudence had been published in London in 1893 by Stevens and Haynes; it had been printed however in New Zealand in Temuka, where Salmond was in practice from 1889-1897, by JM Twomey. Salmond's Jurisprudence was to a considerable degree an enlarged and expanded version of this earlier book.
What were these books supposed to be about? Well the first point to be made about them is that they were written not for legal practitioners but primarily for students. Let Salmond explain:
I have endeavoured to make this book useful for more than one class of readers. It is written primarily for the use of those students of the law who are desirous of laying a scientific foundation for their legal education; yet I hope it will not be found destitute of interest by those lawyers whose academic studies lie behind them, but who have not wholly ceased to concern themselves with the theoretical and scientific aspects of the law.
He also hoped lay person interested in ethical and political science might be interested, even if they had:
... no desire to adventure themselves among the repellent mysteries of the law.
They were therefore products of the academy. They were primarily introductions to legal study and reflected the idea that the common law appeared at first sight at least to be extremely complicated and extremely disorderly. The law student needed a theoretical grounding in basic legal ideas and conceptions if he or I suppose she, though at this time there were very few shes, was not be unable to master the science of law. Pollock thought much the same way. He explains in the preface to his book:
It is addressed to readers who have laid the foundation of a liberal education and are beginning the special study of law. Such a reader finds, in the new literature he has to master, a number of leading conceptions and distinctions which are assumed to be familiar, and are so to lawyers, but which, for that very reason, are not often expressly stated, still less often discussed, and hardly ever explained. He has not only to discover for himself, often with much bewilderment, the actual content of legal terms, but to realise the legal point of view and the legal habit of mind.
So one general idea underlying this form of legal literature was that an understanding of the basic ideas and conceptions of the law would enable the student to reduce the disorderly mass of legal materials to order, and thus the study of the law would be made simpler. And this idea was related to a somewhat grander purpose; such literature would ultimately make it possible to reconstruct the law as an orderly, principled and coherent system. There was therefore a loose association between this form of legal literature and the idea that one day the common law would be codified. In England the codification movement died; in the USA it turned into the movement which produced the restatements
A second general underlying idea arose because the common law was now being taught in universities. Thus Markby's Elements of Law was written in much the same spirit as Pollock, as a contribution to what he called "the preliminary training" of lawyers, which he thought should alone be the business of university law schools. But Markby also saw his book as a providing an essential and necessary intellectual contribution to legal study, which justified its being treated as a fit object for study in universities. Law could and ought to be studied as an organised body of knowledge, and not simply picked up as you went along in the manner in which children learn games:
... the only preparation and grounding which a University is either able, or, I suppose, would be desirous to give, is in law considered as a science; or at least, if that is not yet possible, in law considered as a collection of principles capable of being systematically arranged and resting, not on bare authority, but on sound logical deduction...In other words, law must be studied in a University, not merely as it has resulted from the exigencies of society, but in its general relations to the several parts of the same system, and to other systems.
Holland similarly believed that law should be studied as a science, that is to say as a principled body of knowledge, and he called his book "a treatise on legal ideas".
He explained that:
The present treatise is an attempt to set forth and explain those comparatively few and simple ideas which underlie the infinite variety of legal rules
He went on to relate the conception of law as a science explicitly to the idea of codification:
The search for these ideas is not merely a matter of scientific curiosity. The ever-renewed complexity of human relations calls for an increasing complexity of legal detail, till a merely empirical knowledge of law becomes impossible ... an uncodified system of law can be mastered only by the student whose scientific equipment enables him to cut a path for himself through the tangled growth of enactment and precedent, and so to codify for his own purposes.
So Holland thought his book would enable students to think about legal study in the right way, that is as a science.
Salmond's first essay in this genre was his The First Principles of Jurisprudence (hereafter Principles). It was written by June 1893, so he cannot have read Pollock's A First Book of Jurisprudence, though he had read his Essays in Jurisprudence and Ethics.  Nor of course Gray.
He had certainly read Holland, whom he criticised in a note and Markby. His book was devoted to what he called formal jurisprudence:
Formal jurisprudence is the subject matter of the present work, its object being an exposition, not of a legal system or of any part thereof, but of the general idea of law and of all those other ideas that form part of it or are necessarily connected with it.
Unlike Holland he thought that these ideas were extremely complex, and sufficiently so to justify treating formal jurisprudence as a subject in its own right. He called it "a branch of knowledge … fairly entitled to a place of its own in the circle of the sciences". He argued that:
... law is in truth an extremely complex idea, comprising within itself a considerable number of subordinate ideas, that are themselves sufficiently complex, and capable of further resolution into simpler elements. To know, therefore, what law is, we must also know the contents of all the idea involved or comprised in it, for example, political society, government, legislation, the administration of justice, obligations, rights, wrongs, remedies, and many others.
In his Jurisprudence: or the Theory of Law the language changes, and he uses the expression "theoretical jurisprudence" to label the subject matter of his book, which is "the science of the first principles of the civil law":
...it is expedient to set apart, as the subject matter of a special department of study, those more fundamental conceptions and principles which serve as the basis of the concrete details of the law.
These first principles could be studied in three ways – systematically, historically and critically. The function of theoretical jurisprudence was to serve as the introduction to scientific legal study.
When I was a student in Oxford – and this was before the publication of HLA Hart's The Concept of Law in 1961 – you had to study "jurisprudence", for it was one of the compulsory courses in the three year programme. I recall thinking at the time that the topics on which one had to be prepared seemed to me to constitute something of a rag bag. We had to write essays on whether international law really was law, and address other very general questions as to the nature of law, and we also had to write essays on theories of possession, the nature of ownership, on rights and duties, etcetera. In fact the scope of the subject more or less corresponded to the topics which were discussed in England by Markby, Holland, Pollock and Salmond and, in the USA, by Gray. There was a very considerable degree of similarity in the topics dealt with in these books. Lurking behind the course was John Austin's Province of Jurisprudence Determined, for thinking about the nature of law took his command theory as the starting point, a tradition maintained in Hart's The Concept of Law. By the 1950s only eccentrics like myself read Markby or Holland or Pollock, and to be frank I am not really sure I actually did then read Holland. But I and other undergraduates did read the current edition of Salmond on Jurisprudence, and relied on other books which covered all or some of the same ground, such as Gray's The Nature and Sources of the Law. This could lead into some strange waters. One of course involved the writings of Wesley Newcomb Hohfeld, whose more elaborate and indeed bizarre ideas are mainly to be found lurking in footnotes which nobody today ever reads. In addition I recall, for example, wading through Albert Kocourek's Jural Relations (1927) on, for example, possession and the notion of investitive facts. One topic which was not much mentioned in the older literature was natural law. It receives a brief mention in Markby, and there is a considerably fuller discussion in Holland.
In Pollock it is variously mentioned but it is not a major topic. By my time it had somehow crept back into the syllabus and become more prominent, so that in my time we read about the conception of natural law with a changing content, and other bizarre things. I even read G del Vecchio, and it was commonplace for conscientious students to read a book by one AP D'Entrèves, who, I recall, had a very attractive daughter, Lucy.
In the Oxford of the 1950s, as back in the late nineteenth century, the status of the study of law remained uncertain. Some thought it as out of place in a university as plumbing, and some colleges still refused to admit students to read law. In others the law school was used as a dumping ground for sportsmen and founders' kin. The study of jurisprudence was still viewed by defenders of university legal education as essential if academic law was to enjoy any intellectual standing. The remarkable importance attached at this time to the appointment of HLA Hart, viewed as a philosopher, not a lawyer, to the chair of jurisprudence in 1952 reflected this perception. I well recall being assured by John Morris, then well known as an able lawyer who specialised in conflicts and perpetuities, but who certainly was no legal theorist, that it was only the study of jurisprudence that made the Oxford Law School worth preserving.
The study of jurisprudence, with its scope defined by the late nineteenth and early twentieth century works I have mentioned, seems today to be in terminal decline, and the type of literature to which Salmond on Jurisprudence belonged is more or less extinct. What seems to me to have brought about this change was the publication of The Concept of Law in 1961, which triggered a change in academic fashion, marked amongst other things by the rise in the popularity of the title philosophy of law. Back in the days of Markby, Holland, Pollock, and Salmond jurisprudence was the work of persons who saw themselves primarily as reflective lawyers, not as philosophers of any kind, though I suppose a case might be made for treating Holland as an exception, since he taught philosophy at Exeter College for some years after 1859. Hart changed that. He also was influential in attacking the whole tradition of publishing works which purported to be comprehensive treatises on jurisprudence; in particular he launched a powerful and somewhat intemperate attack on Dias and Hughes on Jurisprudence. One such work for which he had, as I know from remarks he made to me, a very low opinion, and which was much used by law students at this time, was GW Paton's A Text-book of Jurisprudence. Some of Hart's objections to works of this kind, in particular that they consisted of discussions of other people's ideas, could not of course have been applied to the original text of Salmond's Jurisprudence, though it could be applied to editions produced after his death.
Now to the substance of Salmond's Jurisprudence. Salmond was born in 1862, and his first jurisprudential book, Principles, was published when he was in his early thirties. Earlier he had published Essays in Jurisprudence and Legal History, in 1891; only one of these essays really addressed legal theory. His Jurisprudence was published when he was forty and a professor of law at Adelaide. Salmond also published a number of articles on jurisprudential subjects.
Let me now try to identify some basic characteristics of Salmond's approach to legal theory. The first point to make is that he was very much a common lawyer, and one with a serious interest in the history of the system. He was writing for students of the common law who were embarking primarily on the study of what we sometimes call lawyers' law. I suspect that one of the reasons why his book was so successful for so long was that it presented a legal theory which appeared to make good sense to someone working within the common law tradition. Here he may be contrasted with two other notable twentieth century analytical theorists: Hans Kelsen and Herbert Hart. The former came of course from the civil law tradition and in all probability neither knew nor cared about the common law tradition. His starting point was a conception of a legal system given structure by a constitution, which he elaborated into the idea of a pyramidical system of norms deriving their validity from a grundnorm. Hart one might have expected to have been different, but the common law is hardly mentioned in The Concept of Law and does not even feature in the index. Salmond appears to have tried to develop a theory of the common law, and he was one of the few theoretical writers – CK Allen, author of Law in the Making, being the most important other example – to have done this. As we all know in the world of the common law the law emerged out of the practices followed in dispute resolution in the Royal Courts. The very first coherent book about the common law, the treatise known as Glanvill, describes what these practices are:
It is, however, utterly impossible for the laws and legal rules of the realm to be wholly reduced to writing in our time, both because of the ignorance of scribes and because of the confused multiplicity of those same laws and rules. But there are some general rules frequently observed in court which it does not seem to me presumptuous to commit to writing.
So the starting point for Salmond's legal theory, and the same is true of C K Allen, is the practice of adjudication. I suppose the same might be said of the writings of Ronald Dworkin, but these are very different in that Dworkin, whatever may be the value of his musings on the matter, has never exhibited the least interest in relating them to the empirical reality of adjudication even in the Supreme Court of the USA much less anywhere else.
So in his Principles, Salmond asked the question, "What is a principle or rule of law" and answers:
A Principle or Rule of Law is a principle or rule recognized and acted upon by the State in the administration of justice. The aggregate of such principles or rules constitute the Law.
In his Jurisprudence this was in effect repeated:
The law is the body of principles recognised and applied by the state in the administration of justice.
Or, more shortly: the law consists of the rules recognised and acted on in courts of justice.
There are obvious objections to this as a comprehensive definition of law, but as a working tool for a law student it clearly identifies what is the object of study. It expresses if not the concept of law yet a concept of law – that of the working lawyer.
His definition of law, for that is what it is, is combined with what seems at first sight to be a curious theory. In Principles he first deals in Chapter I with Justice and then in Chapter II with "Of the Administration of Justice". It is in Chapter III, "Of Law", that we meet his definition. Salmond explained this treatment:
It may be thought that in treating first of the administration of justice and afterwards of law we have reversed the proper order of procedure. It may be considered that the idea of law is logically precedent to that of the administration of justice. But this is not the case: the former must be defined by reference to the latter, not the latter by reference to the former.
In Jurisprudence he puts forward the same argument; paragraph 7 in Chapter II explains that "Law is logically subsequent to the Administration of Justice". For Salmond the administration of justice was a function of the State, and involved:
... the maintenance of right within a political community by means of the physical force of the state. ... The primary purpose of this function of the state is that which its name implies – to maintain right, to uphold justice, to protect rights, to redress wrongs.
He goes on to say that:
Law is secondary, accidental, unessential. It consists of the fixed principles in accordance with which such function is exercised.
What a litigant obtains in the tribunals of a modern and civilized state is doubtless justice according to law, but it is essential and primarily justice and not law...Justice is the end, the law is merely the instrument and the means: and the instrument must be defined by reference to its end.
It would be perfectly possible to have a system for the administration of justice:
... in which right is done to all manner of people in such a fashion as commends itself to the unfettered discretion of the judge, in which equity and good conscience and natural justice are excluded by no rigid and artificial rules. ...
But this would not involve administering justice according to law.
Salmond's idea here has some similarities with the concept of "Khadi justice" as used by Max Weber, but there is no reason to suppose that Weber was influenced by Salmond.
The obvious objection to this is the argument that institutions which administer justice – courts – are themselves the creatures of law, for legal rules are constitutive of such institutions. I am not myself convinced that this quasi logical argument is correct; I think that the existence of institutions with accepted powers can precede the evolution of rules of competence which are evolved to rationalize institutions which already exist, in much the same way that grammatical rules are evolved to explain or rationalize a practice of human communication, not to make it possible. Grammatical and other linguistic rules are not constitutive of the practice of human communication. But that would get me into deeper water than is appropriate here.
Over the years, Salmond argued, the courts of justice have come to adopt the practice of following settled rules and principles:
The development of a legal system consists in the progressive substitution of rigid pre-established principles for individual judgment, and to a very large extent these principles grow up spontaneously within the tribunals. That great aggregate of rules which constitute a developed legal system is not a condition precedent of the administration of justice but a product of it.
For Salmond indeed the "creed of the courts of justice constitutes the law". It is the corpus of rules and principles accepted by the courts which constitute the law, and the function of these rules and principles is that of excluding judicial discretion in the administration of justice.
What he is really talking about here is the evolution of the rule of law. He does not however use this expression, which at first sight seems odd, since the expression must have been made very familiar to all common lawyers through its use by Dicey in his Law of the Constitution (1885).
Apart from those of a very left wing persuasion, who think that law is basically a mechanism for oppressing the lower orders, lawyers and writers about law normally proceed on the assumption that the application of the rule of law to judicial decision making is a Good Thing. Salmond was not so uncritical, and he indeed devotes a considerable portion of his Jurisprudence to discussing the merits and demerits of this reduction in judicial discretion through the evolution of law, so that the justice administered by courts becomes justice according to law.
The advantages were the production of uniformity and certainty, the exclusion of improper motives in adjudication, and protection against errors in individual judgement. Against these he sets the disadvantages – rigidity, conservatism, formalism and undue and needless complexity. He concluded that:
Under the influence of the spirit of authority the growth of law goes on unchecked by any effective control, and in course of time the domain of legal principle comes to include much that would better left to the arbitrium of the courts of justice. At a certain stage of legal development...the benefits of the law begin to be outweighed by those elements of evil which are inherent in it.
So for Salmond, who might be viewed as an early writer in critical legal studies, though he was of course wholly uninfluenced by Marx or by any theory of economic determinism, the evolution of justice according to law was not an unmixed benefit. Indeed he concludes his discussion that some of the vices consequent upon the evolution of justice according to law might be avoided if the rules of law which the judges were bound to follow were flexible rules. They might in effect say to the judges: "Do this except in those cases in which you consider there are special reasons for doing otherwise". In this discussion Salmond was a pioneer; he appears to have been the first writer to use the concept of formalism, which is now commonplace.
Salmond waxed lyrical in his description of the absolute exclusion of judicial discretion by law:
For good and sufficient reasons the courts which administer justice are constrained to walk in predetermined paths. They are not at liberty to do that which seems right and just in their own eyes. They are bound hand and foot in the bonds of an authoritative creed, which they must accept and act on without demur. This creed of the courts of justice constitutes the law, and so far as it extends, it excludes all right of private judgement.
Given this state of affairs Salmond had to provide some account of how this process of binding the judges took place.
It was in this context that Salmond came up with an idea which was later to be much elaborated by Hart in The Concept of Law. This was that some legal rules were about other legal rules. In Principles Salmond called these rules secondary rules, as opposed to primary rules.
He argued that such secondary rules related either to the existence of a primary rule, its application, or its interpretation. So far as secondary rules of the first kind are concerned he thought that some such rules were ultimate in the sense that they do not depend upon any further rule or principle. He gave as examples the proposition that all principles contained in an Act of Parliament are the law, or the rule that local custom has the force of law, or that precedents are a source of law. Salmond combined this with the notion that legal rules normally had legal sources, such sources being the reason why a particular rule counted as a legal rule; we today, influenced by Kelsen and Hart, would say a reason for validity. Salmond explained:
In other words, though the majority of legal rules may be traced to a source, there must be some self-existent rule or rules without legally recognised source, cause or origin, being, so far as legal theory goes, self caused and self-existent; otherwise any investigation of the sources of law would lead to infinity.
In Jurisprudence the passage distinguishing primary and secondary rules is not reproduced. Instead the rules described as secondary in Principles now feature in a chapter on "The Divisions of the Law" in which Salmond offers a general scheme for the arrangement of the law, presumably one which could be used in drafting a code. These ultimate rules are classified as belonging to "The Introductory Portion of the Law". They are said to comprise:
... all those rules which by virtue of their preliminary character or of the generality of their application cannot be appropriately relegated to any special department.
The text makes it clear that these are the secondary rules of Principles, though the term is not used here. But a rewritten passage on fundamental legal principles is included, with the additional argument that legal systems may have one or more than one such principle.
It has been said that Salmond's conception of an ultimate rules was "insufficiently elaborated" and this is no doubt true. But it was to some degree elaborated in an Appendix which discussed" The Theory of Sovereignty" and argued that there was no logical reason why a constitution, that is:
... the existence and observance of a definite scheme or organised structure and operation, …
Which he called a
datum and presupposition ...
might not make no provision for a sovereign legislature. But even if this criticism of Salmond is just one needs to recognize his originality, in which his achievement resembled that of a mountaineer who, like Edward Whymper in the case of the Matterhorn, makes the first ascent. For centuries the feat is viewed as both ridiculous and impossible, like the mating of tortoises. Once the mountain is climbed it all seems obvious that it was always climbable, and mountaineers set about elaborating the possible routes, times of year at which the climb is possible, and so forth. We need to respect the brilliance of the climber who took the first steps even if, as in the case of Whymper, four of his companions were to die on the descent. Salmond left no such trail of destruction behind him.
Secondly he had to try to provide a theoretical account of how the judges forge fetters for their own feet. In Jurisprudence he explains that:
The growth of case law involves the gradual elimination of that judicial liberty to which it owes its origin. In any system in which precedents are authoritative the courts are engaged in forging fetters for their own feet
Here he ran into serious difficulties, as indeed has everyone, from the time of Hale and Blackstone onwards, who has tried to draw a line between the law applying functions of courts and their law making function, a distinction described somewhere by Otto Kahn Freund as one of the fundamental mysteries of the common law system. Slightly simplified Salmond's theory was as follows.. All questions coming before courts for decision can be divided into questions of law and questions of fact. A question of fact is defined as a question which is not a question of law. In the case of the former the proper decision is determined by law, and there is no place for the exercise of judicial discretion. In the latter case the proper decision is not determined by law and so the judge has discretion unfettered by law. As issues of fact are decided by judges precedents are created, and for the future, under the operation of the legal rules governing the authority of precedents, issues of fact turn into issues of law:
The development of a legal system consists in the gradual encroachment of law on judicial discretion, the progressive transformation of questions of fact into questions of law.
He claimed that:
In the strict theory of the law, precedent is wholly constitutive, being quite destitute of abrogative power. Where the law is already settled, the judges have no authority save to obey and administer it. Their power of making new law by judicial decision is limited to those vacant spaces where there is as yet no other law which they can apply. Precedents make law, but cannot alter it.
That Salmond appreciated that there was something not quite satisfactory in this account of the judicial process is brought out by the fact that it was variously, if somewhat discretely, qualified. So he explains that in addition to "filling up with new law the gaps which exist in the old" courts may also engage in "supplementing the imperfectly developed body of legal doctrine". Not all decisions on issue of fact, he explains, turn such issues into issues of law – this is only the case, he argues, if the issue is one that admits of being answered "on principle". Again under the rules governing the authority of precedents some precedents are absolute, and must be "followed without question, however unreasonable or erroneous it may be considered to be". But:
Where ... a precedent possesses mere conditional authority, the courts possess a certain limited power of disregarding it.
Salmond then engaged in an elaborate analysis of the circumstances in which a precedent may be disregarded. Underlying this is the belief that the rules governing the authority of precedents must themselves be precise and clear if they are to perform the function of law, which is to exclude the element of discretion. Even then the court has no power to abrogate law. Salmond was indeed in love with the idea of subjecting judicial discretion by rule.
I think that whatever criticisms may be made of his analysis he deserved the highest credit for the dogged determination with which he tackled this deep mystery of the common law system, one which continues to generate an inconclusive literature, the most bizarre modern contribution coming from Ronald Dworkin, who has argued that the reality of the matter is that law constitutes a comprehensive system so that there is always just one right answer to any legal issue coming before a court.
I think most lawyers, including Salmond, would say, when not wearing a jurisprudential hat, that at any given time there are numerous legal questions to which there is a uniquely correct answer, and others which are arguable in the sense that reasons of a kind acceptable in the practice of the courts can be advanced on either side. Recently I have been much involved in working on a human rights case which has been submitted to the Strasbourg Court, and in the lengthy discussions which took place both face to face, and on email, whilst the application was drafted, we were continually concerned to distinguish issues which, as one said colloquially, were open and shut, and those which were arguable. In relation to the latter we tried to distinguish, often with much disagreement, between arguments which might conceivably fly in Strasbourg, and others which were just dead ducks. This process is made the more difficult because the conventions of legal argument and justification are themselves open to contention, and there is an open ended character to those conventions.
This was essentially the point made by CK Allen in criticising Salmond's attempt to draw a clear distinction between the legal sources and the historical sources of common law.
Salmond had noted the fact that ideas derived from the French jurist RJ Pothier had been adopted, but he had argued that the writings of Pothier were not the legal source of the adopted rule. Salmond explained that a legal source was a source which the law itself recognised as binding on the courts; Salmond called such sources "authoritative", and they were "the only gates through which new principles can find entrance into the law".
Pothier's writings were merely the historical source – where the rule came from:
A source of law, then, is any fact which in accordance with the law determines the judicial recognition and acceptance of any new rule as having the force of law.
It was the case which adopted Pothier's ideas, the precedent, which was the legal source:
is the legal source of the rule, and the others are merely its hisoritcal sources. ...Our law knows well the effect of precedent, but it knows nothing of Pothier, or of Tribonian, or of the Urban Praetor.
Allen argued that:
All this sounds very clear and exact, but when we come to apply it in the courts, its precision becomes blurred.
Citing cases in which Pothier had been treated as authoritative, that is as a legally acceptable argument that a certain proposition was the law, he emphasised the fluidity of the legal practices of the courts in relation to the use of authorities; earlier Markby, an experienced judge, had made much the same point.
But I think we must nevertheless give Salmond great and respectful credit as a pioneer of the idea that legal systems can be viewed as systems of rules within rules, which has proved to be an important tool in the understanding of certain aspects of the concept of law and of a legal system. Yet at the same time we need to recognise that it has certain limitations, and one basic limitation to which we need to attend is that the conception of judicial decision according to rule may be an ideal which, as is the case with ideals pursued in other departments of human activity, never seems to be wholly attainable.
Salmond however thought that the administration of justice was fundamentally different from other human activities in which decisions have to be taken, such in medicine, in flying an airliner, in choosing a restaurant, or a person to marry, or whatever. This is well brought out in Principles where he wrote:
Now the characteristic that most strikingly distinguishes the administration of justice from all other departments of human action, is the extent to which the application of preestablished principles excludes the exercise of free discretion and individual judgement. ... [emphasis added]
And in another passage he said that:
For good and sufficient reasons the courts which administer justice are constrained to walk in predetermined paths. They are not at liberty to do what seems right and just in their own eyes. They are bound hand and foot in the bonds of an authoritative creed, which they must accept and act on without demur. This creed of the courts of justice constitutes the law, and so far as it extends, it excludes all right of private judgement.
I am not convinced that Salmond was right in singling out legal decision taking in this way. The process of legal decision resembles the process of rational decision in everyday practical affairs, where although some solutions are ruled out as silly, we just have to balance considerations, and attend to values which pull us in different ways, and where, at the end of the day, there may be no conclusive reasons for deciding one way rather than the other. It is for example the case that in medical matters well trained doctors follow established rules which often provide clear guidance as to what tests to order, what questions to ask the patient, what drugs to prescribe, and so forth and so on. But at the end of the day medicine is not an exact science, any more than is choosing a restaurant, and I cannot see why law should somehow be different.
But when I suggest that Salmond never managed to provide a satisfactory analysis of the way in which common law emerges from the process of adjudication we must remember that nobody else has ever succeeded either. As I have suggested elsewhere the best that we can hope of from the work on which he was engaged was the development of a keener awareness of the problems and some movement towards a solution, and that he certainly provided.
I these remarks I have concentrated upon only one aspect of Salmond's achievement. In the papers that follow a fascinating account is given of other aspects of his work and career, and of the environment in which he made his contributions not only to the rule of law in New Zealand but more widely to the culture and evolution of the common law tradition. In the history of that tradition he will continue to rank as a major figure, and it was a great honour and privilege to be invited to participate in the Salmond Symposium in August 2006.
[∗] Charles F and Edith J Clyne Professor of Law, University of Michigan Law School. Professor Simpson's participation was made possible by a grant by New Zealand Law Foundation.
 Sir John Salmond The Law of Torts. A Treatise on the English Law of Liability for Civil Injuries (Stevens & Haynes, London, 1907). There were editions of Salmond by the author in 1907, 1910 (2 ed), 1912 (3 ed) 1916 (4 ed) 1920 (5 ed) and 1924 (6ed). WTS Stallybrass edited: 7 ed (1928), the 8 ed (1934) the 9 ed (1936), and the 10 ed (1945). RFV Heuston edited: 11 ed (1953), the 12 ed (1957) 13 ed (1961), 14 ed (1965), the 15 ed (1969), the 16 ed (1973), 17 ed (1977), and with RS Chambers the 18 ed (1981), with RA Buckley the 19 ed (1987) and 20 ed (1992) and 21 ed (1996).
 Stallybrass edited the 7th to the 10th editions.
 Sir John Salmond Jurisprudence: or the Theory of the Law (Stevens & Haynes, London, 1902). It originally cost eighteen shillings.
 There were editions by Salmond himself in 1907, 1910, 1913, 1916, 1920 and 1924. The 8th edition of 1930 was by CAW Manning of New College Oxford (Fellow from 1923-1930), who taught HLA Hart law. The 9th edition of 1937 was by J L Parker, who was a Fellow of Worcester College from 1927-37.Those of 1947 and 1957 were by Glanville Williams. Manning made minor changes and only those of which he thought Salmond would have approved. Parker made more extensive changes, as did Glanville Williams.
 The 10th edition.
 John Austin The Providence of Jurisprudence Determined (John Murray, London, 1832); John Austin Lectures on Jurisprudence, Being the Sequel to "The Providence of Jurisprudence Determined" (John Murray, London, 1863). This latter work was largely the work of his widow, but assisted by Markby.
 Sir William Markby (1829-1914) was educated at Merton College Oxford, where he obtained first class honours, and was a member of the Inner Temple. His wife Lucy was the niece of John Austin's widow, Sarah, and Markby helped to edit Austin's Lectures, which appeared in 1863. He became a judge of the High Court of Bengal in India (1866-1878) and was Vice Chancellor of the University of Calcutta 1877-8. He then became Reader in Indian Law in Oxford 1878-1900 and a Fellow of All Souls. Later he was a Fellow of Balliol and its Bursar. He was knighted in 1878 for his services in India.
 William Markby Elements of Law Considered with reference to the Principles of General Jurisprudence (Clarendon Press, Oxford, 1871). There were subsequent editions in 1871, 1874, 1885, 1889, 1896 and 1905.
 Thomas Erskine Holland (1835-1926) was educated at Balliol and obtained a first in Lit Hum. He taught philosophy at Exeter College for some years from 1859. After practice at the English bar became an academic in Oxford from 1874, holding the Chichele chair of International Law at All Souls from 1874-1911. He was knighted in 1917. He was of an inflexible disposition and in revisions of his book took little note of criticisms; as an international lawyer he was hindered by the belief that international law could be called law "only by courtesy" (9 ed, 125). He did not make any notable contributions to international law.
 Sir Erskine Holland The Elements of Jurisprudence (Clarendon Press, Oxford, 1880). Revised editions in 1882 (2 ed), 1886 (3 ed), 1887 (4 ed), 1896 (8 ed); the other editions were reprints. There was an edition published in New York in 1896 and a pirated American edition in about 1896. The last edition appeared in 1924.
 There were editions in 1882 (2 ed), 1886 (3 ed), 1888 (4ed), 1890 (5 ed), 1893 (6 ed), 1895 (7 ed), 1896 (8 ed with a New York printing), American edition (1896 from 7 ed) 1900 (9 ed), 1906 (10 ed, New York, Oxford University Press), 1910 (11 ed), 1916 (12 ed), 1924 (13 ed).
 Sir Frederick Pollock A First Book of Jurisprudence for Students of the Common Law (Macmillan and Co, London, 1896). There was a reprint in the same year and revised editions in 1904 and 1911.
 A new and revised edition appeared in 1921, edited by Roland Gray, presumably Gray's son, who had Gray's own notes (Gray died in 1915). This was re-issued in 1948.
 The lectures were on "comparative jurisprudence".
 John Salmond The First Principles of Jurisprudence (Stevens and Haynes, London, 1893). This is a small format book of 264 pages; Jurisprudence is a larger format book of 673 pages.
 They had previously in 1891 published JW Salmond Essays in Jurisprudence and Legal History (Stevens and Haynes, London, 1891).
 Preface to the first edition of Salmond Jurisprudence, above n 3, v.
 Pollock, above n 12, Preface xii and xiv; Salmond Jurisprudence, 5 ed, above n 4. References are to the 5 ed, which I happen to possess.
 See Introduction, Markby, above n 8, xiv.
 Ibid, xiii.
 Holland, 1 ed, above n 10
, Preface, vii.
 Holland, 9 ed, above n 10
 Sir Frederick Pollock Essays in Jurisprudence and Ethics (Macmillan and Co, London, 1882).
 Salmond Jurisprudence 5 ed, above n 4
 See ibid, 8 note 45, 97, 187 and 201.
 See ibid, 90, 132, 187.
 Salmond Principles, above n
 Ibid, 5.
 Salmond Jurisprudence, 5 ed, above n 4
 Ibid, 8.
 HLA Hart The Concept of Law (Clarendon Press, Oxford, 1961).
 Albert Kocourek Jural Relations (Bobs-Merrill Co, Indianapolis, 1928).
 Markby, above n 8, 8, 116. In Salmond, above n 3, there is a brief discussion of Bentham's denial of the existence of natural rights (220-223) and a very brief mention of the extensive literature on the subject at 2, discussing "natural jurisprudence".
 Holland, 9 ed, above n 10
 Ibid, 13, 19, 31 and 98.
 Georges del Vecchio Justice, droit, état: études de philosophie juridique (Paris, Librairie du Recueil Sirey, 1938). I used an English translation.
 Alessandro Passerin d'Entrèves Natural Law: an Introduction to Legal Philosophy (Hutchinson's University Library, London, 1951).
 For example Corpus Christi.
 See generally Nicola Lacey A Life of HL A Hart: The Nightmare and the Noble Dream (Oxford University Press, Oxford, 2004) and my review, AWB Simpson "Herbert Hart Elucidated" (2006) 104 Mich LR 1437.
 Morris was a Fellow of Magdalen College and had the status of Reader. He was, it was said, offered the Vinerian Chair of Law but declined through unwillingness to move to All Souls College, an institution he had publicly scorned for many years.
 Presumably on a part time basis and to finance his career at the bar.
 See his "Dias and Hughes on Jurisprudence" (1957-1958) 4 Soc Pub Teachers of Law 143. This work had a very curious history. Originally GBJ Hughes produced a textbook of jurisprudence, but this was withdrawn by the publishers after complaints by RWM Dias that it was based on his lectures in Cambridge. A group of senior members of the Cambridge Law Faculty then negotiated an arrangement under which Dias and Hughes would collaborate in producing a text, and this they did, producing GBJ Hughes Jurisprudence (Butterworth, London, 1955). Thereafter the story circulated that Dias in his turn had taken his ideas from Glanville Williams.
 GW Paton A Text-book of Jurisprudence (Clarendon Press, London, 1946). There were subsequent editions in 1951, 1964 and 1972.
 That on civil liability. See Alex Frame Salmond: Southern Jurist (Victoria University Press, Wellington, 2005) 41.
 Principally "The Law of Nature" (1895) 11 LQR 121; "The Names and Nature of the Law" (1899) 15 LQR 367; and "The History of Judicial Precedents" (1900) 16 LQR 376.
 Kelsen had been involved in the drafting of the Austrian constitution of 1920 and served as a judge on the constitutional court.
 C K Allen Law in the Making (Clarendon Press, Oxford, 1927) was first published in 1927 and went into seven editions, the last being published in 1964. The book originated in the Tagore Lectures delivered in Calcutta in 1926. WL Twining in AWB Simpson (ed) Biographical Dictionary of the Common Law (Butterworths, London, 1984) perceptively describes Allen as "not so much a legal philosopher as a lawyer's jurist", noting his similarity in this respect to his successor in the Oxford chair of jurisprudence, Arthur Goodhart.
 GDG Hall (ed) The Treatise on the Laws and Customs of the Realm commonly called Glanvill (Nelson, London, 1965) 3.
 Salmond Principles, above n 15
 Salmond Jurisprudence, 5 ed, above n 4
 Ibid, Chapter I.
 Ibid, Chapter II.
 Salmond Principles, above n 15.
 Salmond Jurisprudence, 5 ed, above n 4, 15-18.
 Jurisprudence, 5 ed, above n 4, 14, 16. I have put together two passages separated in the original.
 Ibid, 17.
 Note that the somewhat similar view adopted by Markby, 5 ed, above n 8
 See M Rheinstein (ed) Max Weber on Law in Economy and Society (Harvard University Press, Cambridge (Mass), 1954) 213, 317, 351. This edition largely reproduces writings of Weber first published in 1922.
 This notion runs through Hart's Concept of Law, and is applied to linguistic rules, giving the impression, which cannot really be intended, that you have to have rules for the use of the term cat before there can be cats. I think what is meant is that you cannot communicate using the term cat unless there exist rules saying what counts as a cat.
 Jurisprudence, 5 ed, above n 4
 Ibid, 16.
 Salmond Principles, above 15
, 78; ibid, 17 and generally.
 Salmond does not include Dicey's book in the literature of jurisprudence, surveyed in his Appendix IV.
 A view held in part by no less than Adam Smith: "Civil government, so far it is instituted for the security of property, is in reality instituted for the defence of the rich against the poor, or of those who have some property against those who have none at all," Adam Smith The Wealth of Nations, Books IV- V (Penguin Classics, London, 1986) 302.
 Salmond Principles, above n 15
, 90-92, more fully Salmond Jurisprudence, 5 ed, above n 4, 23-34.
 Salmond Jurisprudence, ibid, 33.
 Ibid 34.
 Ibid 17.
 Salmond Principles, above n 15
 Ibid, 220.
 Ibid, 222-3.
 Ibid, 222.
 Salmond Jurisprudence, 5 ed, above 4
, 483. In the 2nd edition the chapter on the divisions of the law is relegated to Appendix IV.
 With the addition of certain unspecified "miscellaneous rules which are of so general an application as not to be appropriately dealt with in any special department of the legal system". Ibid, 109.
 Jurisprudence, 5 ed, above n 4
, 111. This appears unchanged in the 2 ed.
 Hart Concept of Law, above n 23, 292, referring to Salmond, 11ed and Appendix I, which appears in the First as Appendix II.
 Jurisprudence, 5 ed, above n 4
 Ibid, 174.
 I exclude his discussion of mixed questions of law and fact, on which see Jurisprudence, 5 ed, above n 4, at 21.
 Ibid, 19 and 173-4.
 Ibid, 19: "Everything is fact for us, which is not predetermined by legal principles".
 Salmond Principles, above n 15
 Salmond Jurisprudence, 5 ed, above n 4
 Ibid, 170-171.
 Ibid, 75.
 Ibid, 164.
 Ibid, 168-170.
 Ibid, 70-172.
 Allen, 7ed, above n 49
, 270 and following.
 Citation by Allen, ibid, to Salmond Jurisprudence, 11 ed, above n 4, 133, noting that Salmond's text had been somewhat altered in the 11th edition from the original by the editor Glanville Williams. In Jurisprudence, 5 ed, above n 4
, 100 and following the text reads:
The immediate source of it [ie a rule] may be the decision of an English court of Justice. But that court may have drawn the matter of its decision from the writings of some lawyer, let us say the celebrated Frenchman, Pothier; an Pothier may have taken it in from the compilations of the Emperor Justinian ... the precedent is the legal source of the rule, and the others are merely its historical sources. ...Our law knows well the nature and effect of precedents, but it knows nothing of Pothier, or of Tribonian, or of the Urban Praetor.
 Salmond Jurisprudence, 5 ed, above n 4
 Salmond Jurisprudence, 11 ed, above n 4, 133.
 Allen, 7 ed, above n 49, 270.
 Markby, 5 ed, above n 8
, paras 22, 23, 25, and 26-30. Markby's arguments are put forward a criticisms of Austin's account of the nature of judicial decision.
 Salmond Principles, above n 15
 Salmond Jurisprudence, 5 ed, above n 4, 16.
 This view of what legal philosophy is all about is set out in Simpson "Herbert Hart Elucidated", above n 41.