Paradox of Self-Amendment by Peter Suber infinite series. Some power, be it contract or revolution or some other, must make law ex proprio vigore or from its own strength. Only a theory of permissible self-amendment can explain this fully, and therefore only such a theory can explain legality per se. In short, the study eventually reaches foundational moral, legal, political, even theological questions, as well as important issues in logic and epistemology. The detailed study of a paradox —any paradox, I am convinced— will blossom in this way. I say this both to reassure and to warn the reader. At first the paradox looks trivial, perhaps especially to lawyers who are accustomed to dealing with such problems, as noted, without much ado. The detailed exploration of the paradox is not triviality squared. On the other hand, the broader issues of the nature of law and the roles of logic and value in it are not reached soon in the exposition or without a period in purgatory. Law and logic are technical subjects. Necessarily, they must be introduced with appropriate rigor and complexity when the inquiry demands it. Fortunately, this never requires arcana. I have aimed at an audience of educated people who know little or nothing of either law or logic. No special knowledge of these subjects is presupposed, and all that is needed is given in the text. This is an inter-disciplinary inquiry whose results should be brought to workers in law, logic, philosophy, political science, sociology, and history. Readers who happen to know something, or very much, about law or logic must forgive the degree of over-explanation that affects them. No one, I hope, will forgive any degree of under-explanation. One consequence of the need to introduce the logic and law needed for non-experts is a lengthy introduction. Sections 1 and 2 might be skipped by any reader who judges from their titles that they are unnecessary. Each offers reflections considerably beyond a merely factual preparation, but neither is necessary to the development of the paradox, which begins in Section 3. Discussion of the paradox is greatly facilitated if I am allowed to speak of legal rules. However, the paradox itself implicates the question whether and in what sense laws are rule-like. If laws are rule-like in the manner of logical rules, then the requirement of consistency will be absolute and the equation of logical and legal validity will be supported. On the other hand if laws are rule-like in the manner of rules of etiquette or custom, then their partial indeterminacy will prevent any rigid application of a consistency test and, in addition, the acceptance of inconsistent rules may be a genuine part of the phenomenon to be explained rather than explained away. I give myself the license to use the word "rule". But I also alert the reader not to prejudge the question of that rule-like character by my use of words nor to allow me to do so. I have tried not to assume or exaggerate the rule-like character of law in the process of subjecting the paradox of self-amendment to logical analysis. One of my conclusions is that laws lack many of the properties found in the rules of logic, mathematics, and games (see especially Section 21). This conclusion is not subverted by use of the word "rule" to refer to laws, but I would not want the reader to think that this important question was unaddressed or left to be determined by the default of common word usage. Outline Section 1 is a brief introduction to paradoxes, why they cause consternation in logic and mathematics, and why they might or might not do so in law. Section 2 introduces the remarkable clauses in constitutions that authorize amendment, and tries to show why they are important to law and legal change, both in theory and in practice. The importance of amending clauses to sovereignty, justice, democracy, and peaceful evolution will only occasionally be touched in later sections; after Section 2 it will be presupposed. Together Sections 1 and 2 provide the background in logic and law for the rest of the inquiry. Section 3 poses the question whether amending clauses may authorize their own change. If amending clauses do not apply to themselves, then their lawful change is apparently ruled out (Section 4); if they do apply to themselves, then their lawful self-amendment is arguably paradoxical (Section 5). This dilemma of immutability and paradox is as central to the book as the narrower exploration of the logic of self-amendment. Section 6 pauses to outline two rival theories of law essential to the controversy surrounding self-amendment. One, which I call the inference model, is a formalist theory that makes the legal validity of an amendment turn on the logical validity of a deduction that models the amendment. The other, which I call the acceptance model, finds an alegal and alogical source for the authority of law, and hence of amendments, in the sociological fact of a complex sort of public acceptance. Section 7 explicates H.L.A. Hart's version of the acceptance theory. This gives us an articulate alternative to Alf Ross's inference model, and shows us some of the issues that will face anyone who embraces an acceptance theory. Section 8 frames the dilemma of immutability and paradox, and shows where United States courts have constructed their stand. The concept of immutability leads directly to that of omnipotence, for if any legal power is omnipotent (in some sense) then no legal rule is immutable (in a cognate sense), and vice versa. The section examines the attempts by scholars and plaintiffs to identify immutable aspects of the U.S. constitution, hence limitations on the (otherwise) omnipotent power of the amending clause. Section 9 carries the inquiry in Section 8 a step further. While Section 8 looked at entrenchment clauses whose alleged immutability challenged the omnipotence of the amending clause, Section 9 explores the entrenchment of the amending clause itself and the possibilities of self- disentrenchment. Or, while Section 8 examined the alleged irrevocable limits on the amending power, Section 9 examines the revocability of all alleged limits. v