Paradox of Self-Amendment by Peter Suber Section 6: The Inference and Acceptance Models of Legal Change A. Weaknesses of the inference model cured by the acceptance model Ross's statement of the paradox of self-amendment assumes that the process of legal change is well represented by logical inference. It is important to see that this inference model is just one model of legal change, perhaps the first that common sense would hit upon but also perhaps the least true to the phenomenon. I will call Ross's model the inference model of legal change and validity. The inference model envisions legal change as a series of deductive inferences. An existing procedural rule for legal change is expressed as a conditional statement ("If A, then B," or "If a procedure is satisfied, then a rule is amended"). That conditional statement is made the first premise of the required inference. The fact of procedural satisfaction ("A is true," or "The procedure has been followed") is made the second premise. The validity of the new rule ("B is true," or "A rule is amended") follows as the conclusion. If A (procedure), then B (amendment). A (procedure). ----------------------------------------- Therefore, B (amendment). The inference model has the virtue of capturing our sense that law is logical, or at least that it ought to be, that procedures produce outcomes through a structure of rules, and that a violation at any stage would either fail to produce the same outcome or would do so improperly, with a wink, despite the rules. The limitation of the inference model is that it makes legal change exclusively a matter of rule-governed inference, and ignores the legal realities of historical accident, retroactive justification, informal custom, pressure, politics, ideology, inconsistency, and winks. One of the most famous statements of American jurisprudence is on point.[Note 1] The life of the law has not been logic; it has been experience. The felt necessities of the time, the prevalent moral and political theories, institutions or public policy, avowed or unconscious, even the prejudices which judges share with their fellow-men, have a good deal more to do than the syllogism in determining the rules by which men should be governed. The law embodies the story of a nation's development through many centuries, and it cannot be dealt with as if it contained only the axioms and corollaries of a book of mathematics. Ross does not defend the adequacy of the inference model; he presupposes it. Indeed, one might say he regards the right of logic to rule and overrule law as an a priori truth. Of course, if one held such a view, no empirical inquiry into legal practices would change one's mind. Under the inference model the legal validity of a newly changed rule of law derives its authority from the logical validity of the inference (process) that brought it into being. The inference is demonstrably, necessarily, and eternally invalid if the conclusion contradicts a premise (when the premises are consistent with one another). And it seems this will always be the case when an AC changes itself into a form inconsistent with its original form. If so, then Ross is correct to find necessary contradiction in (substantial) self-amendment under the inference model. Unfortunately, he questioned neither the inference model itself, nor the proposition that logical impossibility implies legal impossibility. If the paradox is real for the inference model, then it must be recognized by any inquirer or discipline that uses formal logic as a standard. While many philosophers have tried to dissolve the paradox even for the inference model, I would like to put those objections aside for a moment (until Sections 10 and 11) in order to introduce the question of the adequacy of the inference model itself, as a model of legal change. If the legal validity of a newly amended rule of law depends on the logical validity of the deductive inference that models its adoption, then let us concede to Ross (for the time being) that self-amendment is paradoxical and unlawful. But if the validity of new law arises from another source, then the inconsistency of the new law with the old rule of change would no longer be an obstacle. Similarly, if the source of authority could tolerate contradiction, then self-amendment could occur without obstacle. Hart provides a model of law in which this is the case; I will call it the acceptance model. The acceptance model is more difficult to state than the inference model because it posits a more complex explanation than a logical form (see Section 7). Essentially the acceptance model locates the authority of all law, including newly changed law, ultimately in the acceptance of the people governed by it, and in the practice and usage of the officials of the system. The acceptance model comes very close to a traditional consent theory of law, but it differs from it in several significant ways. The acceptance model requires a much less explicit form of assent than the consent theory and has fewer contractarian overtones. This is because it requires no mutual advantage for the parties, and because acceptance exists even when people are ignorant or indifferent to the laws that govern them. And at least in Hart's version, the acceptance theory is intended to explain authority descriptively rather than prescriptively, or to explain legality, not legitimacy. An acceptance theory may appeal to acceptance directly for every rule of the system, or only for a few ultimate rules that in turn validate the rest. Hart is often taken to have said the latter. This view has Hart adopt an inference model for all legal rules except the ultimate rule of recognition.[Note 2] Hart's actual position is not so simple. 29