Paradox of Self-Amendment by Peter Suber area of overlap and inconsistency of the general principles. If the case were resolved by ranking one principle over the other, then inconsistency in application is avoided, even if the ranking is reversed or ignored in the next case. But the ranking leaves inconsistent rules valid and effective; it merely saves defendants from their joint operation. In this sense we show our distaste for contradiction by devices that prevent it from hitting defendants as it were from both sides at once, but we show our toleration of "theoretical" contradiction by devices that leave contradictory rules intact. The common law has innumerable inconsistent rules. That is never denied, and indeed the abundance of inconsistent precedents, authorities on each side of each case, and tortured attempts to distinguish cases contrary to the desired outcome are endemic and notorious features of common law adjudication. We accept the inconsistencies latent in the mass of common law rules at least in part because of the devices that can prevent inconsistent rules from reaching simultaneous application. But this shows that we accept the permanent possibility of the simultaneous application of inconsistent rules. Even when actual reconciliation is attempted, the favored common law technique is piecemeal, or case by case restatement and harmonization. This shows that we tolerate slow reconciliation. Moreover, the piecemeal method results in cases that may stand next to others to form new inconsistencies. We tolerate uncoordinated reconciliation that, if viewed from a distant, exclusively logical standpoint, may take one step backward for every step forward. The devices that prevent the joint operation of inconsistent rules do not change the fact that we accept continuing contradictions in these ways. For such devices do not protect non-parties from the uncertainty and unpredictability of the law. No common law device is used to help the citizen trying to arrange her primary conduct by looking to the law for guidance. Indeed the difficulties of planning are compounded if inconsistent rules are reconciled only at the stage of adjudication. In this sense, merely "practical" avoidance of contradiction for defendants that leaves inconsistent laws intact does not avoid all the "practical" costs of contradiction. The problems of slow, uncoordinated, piecemeal reconciliation of inconsistent rules are often solved at one stroke by codification. The resistance to codification in some fields, such as tort law, shows our preference for "flexibility" and "growth" even at the cost of greater inconsistency and uncertainty. Where codification has occurred, it is safe to assume that whatever inconsistencies remain are rarely if ever accepted and cannot easily be used to show our toleration of contradiction. In these contexts it is important to distinguish contradictions we accept from those we overlook. As noted, I do not claim that all existing contradictions are accepted. Some are pounced upon as soon as they are discovered and were suffered to exist in the first place only because we were unaware of them and because our piecemeal and uncoordinated methods of enacting and amending law cannot be prevented from issuing an occasional (or frequent) unexpected inconsistency. Note that even the view that law is a quasi- logical system in which the criteria of logical validity determine legal validity is compatible with much inadvertent inconsistency, provided that sufficient procedural devices exist to reconcile or invalidate them ab initio whenever discovered. But even in those areas of law where the ideal of consistency is most rigorously upheld, such as criminal law, consistency is not valued for its logical properties, but for policy reasons. Under the right combination of circumstances policy decisions may overrule the ideal of consistency even in an area of carefully codified law. One policy of course is fairness. It is simply unfair to apply rules inconsistent by the direct or compliance test to the same person at the same time. In Lon Fuller's example, U.S. v. Cardiff, 344 U.S. 174 (1952), two sections of a single statute seemed contradictory. One appeared to require that factory owners permit federal inspectors to enter their premises; another appeared to give factory owners a right to refuse consent. As Fuller argues, a logically coherent interpretation was available to the Supreme Court, but the Court perceived the issue ultimately as one of fairness and notice, and voided the conviction of a factory owner who refused consent to an inspection.[Note 20] Statutory and contractual duties may also conflict in a way that makes simultaneous compliance impossible. A.M. Honore gives the example of a drunk truck driver who must return her truck to the depot by 5 p.m. Either she drives and violates the traffic laws, or stays put and violates her contract.[Note 21] The same logic is involved in these cases, but the policy of preventing the application of conflicting rules is considerably weakened by our sense that the defendant got into her dilemma only through her own fault. In different cases we may let the defendant suffer even without a judgment of her fault. For example, if a manufacturer can only meet her sales contracts by polluting in excess of legal standards, then normally we will give priority to the statute expressing the social value of a clean environment. Occasionally, however, a judge will improvise a "compliance schedule" that permits excessive pollution for a generous period. But these examples show that the mere fact that a person is trapped by inconsistent legal rules and must violate one to comply with others is not by itself a sufficient reason to provide judicial relief. Sometimes the defendant will be left to suffer the consequences of her fault, sometimes to suffer the consequences of overriding social values. But whether the defendant is protected from unfairness or left to choose which rule she will violate, we are allowing policy judgments to supersede and govern the application of the ideal of consistency.[Note 22] Another policy served by consistency is predictability. Consistent laws make it easier for citizens to understand their rights and obligations and therefore to make plans in reliance upon the boundaries of their liability. Dennis Lloyd has said that one of the "vital aims" of consistency is to enhance the "security and predictability" of our social and economic life.[Note 23] The policy of predictability, however, like that of fairness, does not universally require consistency. The conflict of the policy of predictability with itself is well illustrated in the ironic attempt of courts to render the law more consistent by overruling anomalous rules when many citizens had come to rely upon those very rules. The disruptive effects of overruling settled but undesirable or inconsistent law is mitigated by making the change of law prospective only. But of course even prospective overruling can upset the reliance of long-range planners. Predictability would be best served by ceasing all change of law, the very absurdity of which shows that predictability is not a supreme policy interest. Freezing the law would also freeze inconsistencies, which suggests that our desire 158