In 1845 a master of English commercial law wrote that there was “no part of the history of English law more obscure than that connected with the maxim that the law merchant is part of the law of the land.” Since then there have been detailed studies of the medieval law merchant and of the later development of English mercantile law, but the precise status of the law merchant in England and the nature of the process by which it supposedly became fused with the common law remain as obscure as they were in 1845. The obscurity begins with the very concept of the “law merchant,” which has been differently understood by different writers and continues to be used in widely divergent senses. Some have regarded it as a distinct and independent system of legal doctrine, akin in status to Civil or Canon law, and perhaps derived from Roman law. Others have supposed it to be a particular aspect of natural law, or the universal ius gentium, and as such akin to international law.

The increase in the powers of public authorities is rightly accompanied by a concern that they should be neither abused nor exceeded. One form of control lies in the judicial remedies, of which the least employed is damages. This is not surprising since, regardless of the loss sustained, the unlawfulness of an administrative act does not automatically confer any right to compensation. Where an administrative act is unlawful because it involves a tort such as negligence, nuisance or trespass, then it is well-established that the public authority responsible will be liable to pay damages to those whom it affects. It is, however, unlikely that one of those torts will be committed where the unlawfulness stems from a breach of the rules of natural justice or a failure to consider all the relevant factors. Nor will they necessarily be involved where an authority misuses or exceeds its powers; for example, by the wrongful refusal or cancellation of a licence. Nevertheless, the extent of governmental regulation means that loss for individuals and businesses is not inconceivable where these types of illegality take place. If there is no right to compensation in such cases, then success in obtaining certiorari or mandamus might well be regarded as a pyrrhic victory.

Section 25 (1) of the Sale of Goods Act 1893 provides:

Where a person having sold goods continues or is in possession of the goods, or of the documents of title to the goods, the delivery or transfer by that person, or by a mercantile agent acting for him, of the goods or documents of title under any sale, pledge, or other disposition thereof, to any person receiving the same in good faith and without notice of the previous sale, shall have the same effect as if the person making the delivery or transfer were expressly authorised by the owner of the goods to make the same.

The word “practice” does not occur frequently in the pages of H. L. A. Hart's book The Concept of Law. Yet the notion of a practice is, if not the key to the science of jurisprudence, at least the key to the central foundations of Hart's theory. The validity of a law, according to Hart, can be established by reference to certain social (or official) practices: in this way questions of legal validity are separated from questions of moral justification. The theory can profitably be read as a study of the connection between the characteristic forms of legal discourse and the existence of social practices.

It is surprising that the already voluminous literature on the Hart v. Dworkin dispute has neglected the issues I intend to raise. One of Dworkin's arguments attacks Hart's claim that the existence of a rule is, in some cases, the existence of a social practice. The argument, if successful, would refute everything Hart has to say about the rule of recognition. Can Hart's theory meet this criticism? I shall sketch the hint of a shadow of an analysis of practices which offers Hart an escape, but an escape which has its price. By adopting something resembling my analysis, Hart can preserve major elements of his theory. I will attempt to isolate those features which are valuable and which can be salvaged. The price that Hart must pay, however, is the abandonment of his extreme positivism. If my suggestions about practices are correct, then the ultimate questions concerning legal validity are moral or political questions.

There can be no doubt that the existence of paradoxes has stimulated vigorous and highly productive activity in philosophy and in logic. Take two famous examples: The Liar Paradox, which arises from a sentence such as

L This statement is false.

and the Russell Paradox which arises from the sentence

R The class of all classes which are not members of themselves

is a member of itself.

The first of these has been a source of anguish for over 2,000 years. The second has engaged the serious attention of logicians for over three quarters of a century. Investigation of paradoxes of this sort has spawned whole new fields of study, such as technical semantics and axiomatic set theory.

It has been claimed that legal reasoning is infected with paradoxes and that these paradoxes are similar in structure to those, like the two we have cited, which are of interest to the logician. If this claim were true one of two consequences would follow. Either the jurisprudent would face what would in all likelihood be a protracted struggle with these legal paradoxes resulting, perhaps, in significant additions to legal theory, or else, if these paradoxes were sufficiently similar to those of the logician, he might try to utilise the logician's results to solve his own legal puzzles.

The first alternative, though attractive to a theoretician, may appear rather dismal to those engaged in the business of law. Whereas reflection on the logical paradoxes can lead to only more refined abstractions—the philosopher's meat and drink—legal theory is rather intimately connected with practical affairs.