people, preferring their intention to that of their legislative agents, it is not easy to show why the intention of the people, directly declared, is not to be preferred to that of their judicial agents. If, on referendum, the voters of the state may disallow the act of one set of agents, it is difficult to see why they may not similarly disallow the act of another set of agents.

It is said, in reply, that a judicial interpretation of law is a different thing from an act of legislation. From the popularsovereignty point of view, however, the difference is immaterial: the interpretation of constitutional law by a court and the act passed by a representative legislature are, each of them, attempts of agents of the people to express the intention of the people. And in fact the distinction between legislation and judicial decision is largely a formal one. Every authoritative interpretation of written law establishes a rule of law; and, if the law to be interpreted be general in its terms, interpretation may make a great many rules of law. This fact is recognized in our legal theory: the Supreme Court of the United States has more than once invalidated a judicial decision rendered by the highest court of a state, on the ground that a state may not make " a law" impairing the obligation of contract.

Recognition, however, that a proposal is logically consistent with accepted political theory has never been regarded by any English-speaking people as a sufficient reason for adopting it. When a change in political practise is proposed, such a people always inquires, first, whether any change is necessary; second, how the proposed change is likely to work.

The first of the questions is answered by the opponents of the "recall" proposal, by pointing out that in most of our states it is easy to amend the constitution. In some of our states such an amendment may be proposed either by the legislature or by a small minority of voters, and may be adopted, with little delay, by a majority of those who vote on the proposal. If in other states the process of amendment is unduly difficult or protracted, the obvious and sufficient remedy is to change the process. On the other hand, the advocates of the "recall" proposal insist that the ordinary process of amendment has unnecessary and undesirable results which the "recall"

avoids. This contention was first clearly and intelligibly presented by Dean Lewis of the University of Pennsylvania Law School. He maintains that the "recall" is not only a more efficient instrument for adapting written constitutions to changing economic and social conditions, but that it preserves the judicial control over unconstitutional legislation which the ordinary form of amendment tends to impair. This last assertion seems to me to be true.

The question how the "recall" will work in other respectsthe question, in particular, of its immediate practical results as distinguished from its ultimate theoretical results-has been little discussed. In endeavoring to indicate how it will probably work, we must, of course, take the proposal as it is now formulated. It is obviously impossible to consider how all other conceivable proposals of a similar general character might operate.

The constitutional restriction upon legislation which has attracted most attention during the past few years, and which has elicited most discussion during the past few months, is found in provisions requiring legislative measures to conform to "due process of law." And, inasmuch as it has long been recognized by our courts that due-process restrictions may cease to be applicable and operative when legislative action is taken under the "police power," this power also has been much discussed. The discussion has not turned upon the proper meaning of due process or upon the extent of the police power. The question is how the interpretation which the courts give to these expressions is to be controlled; how the people are to exercise a power to determine, in last instance, what their constitutional law shall be.

As the "recall" proposal was originally presented by Mr. Roosevelt, the referendum on decisions was not, apparently, to be limited to due-process cases; it was to be available whenever a state law was declared to be invalid because in conflict with any provision of the state constitution. In view, however, of the inconvenient and even absurd results that might possibly be produced by an unlimited "recall," it is now proposed that referenda on decisions shall be limited to due-process cases. The

proposal is so formulated, for example, by Mr. William L. Ransom in his recent book on Majority Rule and the Judiciary, to which Mr. Roosevelt contributes an introduction.

In order that we may see how this limited "recall" would probably work, let us take a situation which has already been much discussed, which has been frequently chosen by the advocates of the "recall" to illustrate the need of a new remedy, and which Mr. Ransom claims would be remedied by a "recall" limited to due-process cases. Let us take the situation which exists in New York as regards compensation of workmen, or of their widows and children, in cases of injury or death caused by industrial accidents. A law passed by the state legislature was pronounced unconstitutional by the Court of Appeals, because it proposed to take the money of the employers without due process and was not within the sphere of the police power. It is claimed that a "recall" limited to due-process cases would have sufficed to validate the law and make it immediately effective. But, in pronouncing this law unconstitutional, the New York Court of Appeals declared that, for the purpose of reaching a decision, it was not necessary for the court to determine whether the act was unconstitutional solely as denying due process. It declined, in particular, to decide whether it was unconstitutional as denying the employers' right to trial by jury. It seems clear, therefore, that even if the decision of the court had been "recalled," the law would have been validated only so far as due process was concerned, and that the question which the Court of Appeals declined to decide would remain open. If, when this issue was raised, the Court of Appeals should again declare the law unconstitutional, it would apparently be necessary, under the limited-recall program, to introduce a new amendment to the constitution, widening the scope of the recall, and to institute a new referendum. If the new amendment were limited in its application to the matter of jury trial, it is not impossible that question might arise whether the law was not in conflict with some other provision of the constitution, or with the general spirit of the constitution.

So numerous are the grounds upon which any law that attempts to realize what is to-day described as "social justice"

may conceivably be pronounced unconstitutional, that it will be no easy task to frame a recall amendment that will cover all these grounds and yet remain limited in its scope. It seems doubtful, to put it mildly, whether any recall proposal thus far formulated promises to secure a more speedy adaptation of our state constitutions to changing conditions than the existing process of substantive amendment. It seems highly improbable that it will bring to a more prompt and satisfactory conclusion any differences of opinion or of sentiment between the state courts and the state electorates. For a single political battle, terminated by a single substantive amendment, the limited "recall" appears to substitute, primarily at least, a prolonged political war, in which the electorate would realize its intention only after several campaigns. Moreover, after

each campaign, the limited recall would be widened in its operation; and it would thus gradually approach that unlimited recall which the supporters of the plan do not at present advocate.

A more fundamental objection, which applies to any conceivable form of referendum on laws pronounced unconstitutional, is that it offers us a crude and unsatisfactory means of obtaining the end desired. The purpose of the proposed referendum is to obtain, particularly in matters of natural right, popular expressions of the sense of social justice. Such expressions are to create precedents which the state courts are to follow. It is, however, extremely improbable that the electorate will consciously attempt to express its sense of social justice. The great majority of the voters will express their varying judgments as to the probable effects, good or bad, of the particular measure submitted to them. If it be replied that the majority judgment will contain, by implication, an expression of its prevailing sense of justice, it may be remarked that implications are matters of opinion, and that widely different implications may be discovered in every such popular decision. After every general election there appear widely different theories as to what was really "the verdict of the people." It may be added that successive referenda on different measures may well contain implications that cannot easily

be reconciled. The difficulty which the courts now experience in determining the true reasons for their own decisions on constitutional questions will appear slight indeed in comparison with the difficulty which they will encounter if they be called upon to determine, first, what intuitions of social justice seem to be implied in a series of popular decisions, and, second, what rules of constitutional interpretation can be formulated that will express these intuitions. To the average lawyer such a process of developing law seems fantastic: hence the generally hostile reaction of the legal profession to the "recall" proposal. To the student of legal history, on the other hand, the process is not fantastic but familiar. It was by this very process the expert interpretation of popular decisions-that law was taking form in the Mediterranean city-states twenty-five centuries ago and again among the Teutons fifteen centuries ago. If the reaction of the legal historian to the proposal is also hostile, it is not because the process seems novel but because it is seen to be archaic. Like the whole direct-government movement, of which it is a product, it is a reversion to the primitive processes of early civilization.

The Development of the Federal Constitution

Closely connected with the problem of adapting our state constitutions to changing social and economic conditions is the problem of the development of our federal constitutional law. The due-process requirement and other restrictions found in our state constitutions are contained in the federal constitution also; and a state court may declare an act of a state legislature invalid because it is in conflict with the supreme organic law of the nation as well as with the constitution of the state. The New York Workmen's Compensation Act, for example, was declared to be in conflict with the federal as well as the state constitution; and it would apparently remain invalid even if the due-process clause of the state constitution were amended or overridden. This being the case, it seems singular that at present there should be so much discussion of means of developing our state constitutions and so little discussion of the possibility of developing the federal constitution.