I

t would not be strictly accurate to say that we take up here for the first time the question of the role and functioning of specialized institutions within a legal system. Hart & Sacks quite properly treat private “orderers” as composing the most important of law making institutions; they treat issues raised by the relations of courts with private law makers as analytically identical with those raised by the relations of courts with administrative agencies and legislatures. It is thus possible to regard the preceding section of this paper as an essay on an institutional “system,” one composed of an official and an unofficial element. The problem, which Hart & Sacks discuss with regard to this system – what relationship of the elements will maximize human satisfactions? – is the same as that which confronts us when we attempt to break down the official element into its part and treat it as a system in its own right.

Yet there is an important non-analytical difference between the detailed inquiry into the organization of officialdom and the attempt to define the proper border between public and private spheres. In short, it is that when Hart & Sacks discuss “government regulation,” they can place themselves within a three hundred year old tradition of political and economic thought.

The issues as between “individualism” and “collectivism,” capitalism and socialism, have been rendered familiar if not clear by public debate. The logical coherence of the utilitarian approach to the economic aspect of the problem has been fully demonstrated by economists, and the close parallel between legal constraints and “prices” was first perceived and exploited by Bentham himself. The value and implications of “freedom of contract” was perhaps the central theme of nineteenth century jurisprudence.

Seen in this light, the indeterminacy (and therefore in Hart and Sacks’ terms the unhelpfulness) of a utilitarian analysis of the relationship of private to public ordering is not surprising, nor, it must be said, is the attempt of any very great intrinsic interest. By contrast, the attempt at a theory of official legal institutions is both more ambitious and more promising, for here we are in the open field of the sociology of organizations, where the tradition is a short one and large systems are rare. It is also in this area that Hart & Sacks have been most influential. Although their theory is no more determinate, nor even more suggestive, in this area than in that of private ordering, it has become one of the principal bulwarks of judges and legal scholars who favor a “passive” role for the judiciary in its relations with other branches of government. But before raising this “ideological” problem, I will try to lay out in some detail the logical structure of the Hart & Sacks theory.

A.

Let us imagine that our planner, Hart & Sacks has somehow brought into being a complete body of substantive and remedial law designed to maximize the valid human satisfactions of their society. Further assume that this body of law is widely accepted by the population – that is, that in general people are willing to abide by its directives; where it is directive, and to pay the “administered” prices of actions which the planner has decided would be incorrectly valued in a system of purely private ordering. Assume that in constructing the body of substantive and remedial law the planner has decided in favor of a regime friendly to private ordering; among the various presumptions about the optimality of free choices which he might have adopted, he has tended to reject those that would lead to large scale direct official action to create welfare (e.g. nationalized industry) and to disfavor those which require detailed regulation of transactions. In short, with little problem of "law and order" and a benign view of the capacities and intentions of the citizenry, the planner has been able to avoid taking a large role. Nonetheless, a system of official institutions will be required in order to administer the legal order.

As compared with the apparatus which would be necessary in a more “collectivist” state, the planner’s institutions can be organized to respond to legal problems rather than to initiate action. At the base of the official machine there will be a mechanism for the referral of private disputes. The official structure can also be relatively decentralized and unarticulated as a system, since it will rarely be necessary to mobilize all official resources for concerted, massive attack on any aspect of the society. Yet the importance of official institutions is not for that reason to be denigrated: unless the three functions of “creative order” are performed, the planner is quite clear that nothing like a maximum of satisfactions will be achieved.

The two most important functions are elaboration (adjudication) of the existing legal arrangements to cover particular circumstances and the free adaptation (legislation) of the legal system to changed circumstances. The third function is subsidiary to the first two; it is the development of factual information and technical knowledge in order to reduce the degree of objective uncertainty under which the other functions must be performed.

Both principal functions are directly related to the goal of maximization – that is, whoever performs them will take this ultimate rather than some intermediate goal as the criterion of the success or failure of his official activity. The functions are conceived as differing in that they are made necessary by two distinct kinds of private disputes requiring resolution in order for the legal order to operate optimally. Elaboration meets the need to resolve disputes between parties who accept the validity of the existing legal system but do not agree as to its proper application to their relationship.

Abstracting from the problem of determining the facts (as I will do throughout), the function of the elaborator is to resolve an uncertainty about what legal rule the goal of maximization requires in a particular situation. The result of elaboration is that “law is made” in some sense, but the law making is (in some sense) “interstitual”. What this means is that the premises to be used in reaching a decision are fixed in advance: they are the elements of the existing body of substantive and remedial law. The correct decision is that which is in “harmony” with existing law in the sense that it will lead to welfare maximization through elaboration of the prescriptions for attaining that end which are implicit in existing law.

Free adaptation meets the need to resolve disputes as to whether existing legal arrangements actually serve to maximize human satisfactions. Again abstracting from the problem of determining the facts, the free adapter is like the elaborator in that his function is to resolve an uncertainty about which rule will maximize, and he too “makes law.” But in adding to or changing existing law, he is not restricted to the premises embodied in that existing law: his decision must be the result of an open-ended consideration of all aspects of the situation which are arguably relevant to maximization. In short, he behaves much as did the planner himself in establishing a legal regime.

The “elaborator” and the “free adaptor” have been described as though they were two different people. There is nothing in the argument thus far that leads to the conclusion that this must be so. It might be argued that while institutional specialization by function is of course not required by the mere existence of two functions, it is, in the United States, compelled by the elemental commitment to “democracy.” The sovereignty of the people may be taken to mean ultimate control of the content of substantive and remedial law by elected officials, and the rule of law to mean the existence of another, independent body of officials charged with administering that law. Hart & Sacks do not take this tack (in fact the argument from “democracy” as a simple value premise or article of faith is barely mentioned in the whole materials). Instead, they develop a three part rationale for specialization based solely on arguments about institutional “effectiveness” in achieving the goal of maximization. The result is a position that goes beyond what “democracy” alone will support: the argument sketched above does not preclude a system in which courts behave exactly as legislators, as long as the legislature had the last word about law making, and in which legislatures behave exactly as courts, so long as courts have the last word about adjudication. In the system sketched below, the rules properly require each institution to keep its own specialized function and actively to reject attempts to involve it in doing the work of other bodies.

The first argument for functional specialization (and the least compelling) is that while any dispute may combine an element of uncertainty about the application of existing law with an element of challenge to that law, cases tend in practice to cluster. A good many disputes are basically about elaboration, with little element of challenge; a good many are basically the opposite. Consequently functional specialization is convenient, if nothing more.

The second argument for specialization is a very important one, and in one form or another it is implicit in most American post-realist legal thought: it is asserted that the skills that are required in order to resolve disputes over elaboration in such a way as to maximize welfare are fundamentally different from those necessary for optimal handling of disputes over the free adaptation of the law. Institutions which are organized (“organization” includes methods of recruiting personnel, obtaining information and enforcing decisions as well as the details of administration) to perform one function “competently” will be “incompetent” to perform the other function. The argument thus supports the strong form of specialization: not only should there be distinct institutions, but they should stick to the tasks which, as Hart & Sacks put it, “they are good for.”

Disputes which are primarily over elaboration, according to the argument, are likely to be conceived by the participants as involving individual interests and past conduct. The area of dispute is likely to be limited, by which is meant that however momentous the principles and interests involved, there will exist a broad area of agreement about values and about law. By definition, elaboration is concerned with the meaning and implications of accepted existing legal arrangements. Yet while the area of dispute may be limited, the parties are likely to find the uncertainty particularly difficult to tolerate. Just because the great body of existing law is accepted as relevant, there is a feeling that there must be a “correct” solution to the question posed, and that a party is in some sense “entitled” to the benefit of application of that correct solution.

According to the theory, these expectations and emotions of the parties are quite justified. If we take the word “correct” to signify not some mechanical sort of precision but “amenability to reason,” then it is true that there are “correct” decisions. The interstitial making of law involves “judgment” and “craft” (i.e. it is unlike natural science) yet there is a determinate method by which the elaborator can best go about attempting to maximize welfare, and the method provides sufficient criteria for evaluating, and controlling, the results reached so that it is reasonable to demand “right” decisions and to criticize “wrong” ones.

The method itself – called “reasoned elaboration” by Hart & Sacks – follows from the nature of the dispute. First, the dispute is by definition one over the meaning of legal directives aimed at the maximization of welfare. If the law made interstitially by the elaborator is to further this large goal, it must be “consistent” with existing law, in the sense that it treats similar cases similarly. To take an example from the text, if a motorcycle is treated as a “motor vehicle” for purposes of a speed statute, but a motorcycle with a side-car is not, the conclusion is inescapable that a mistake in elaboration has been made. Second,

Underlying every rule and standard . . . is at least a policy and in most cases a principle. This principle or policy is always available to guide judgement in resolving uncertainties about the arrangement’s meaning. The uncertainties cannot be intelligently resolved – without reference to it . . .

Not only does every particular legal arrangement have its own particular purpose, but that purpose is always a subordinate one in aid of the more general and thus more nearly ultimate purposes of the law. Doubts about the purposes of particular statutes or decisional doctrines, it would seem to follow, must be resolved, if possible, so as to harmonize them with more general principles and policies. The organizing and rationalizing power of the idea is inestimable.

If the law is conceived in this way, neither apparent inconsistency nor uncertainty nor confusion at the level of specific legal rules should shake the faith of the parties and the elaborator in the existence of a correct solution. In ascending the hierarchy from the specific (e.g. an inheritance statute) to the more general (e.g. the principle that no man should profit from his own wrong) toward the ultimate end of law (maximization of valid human satisfactions), inconsistency, uncertainty and confusion will sooner or later disappear.

The technique of reasoned elaboration bears an interesting resemblance to systems analysis in a system in which goals are given and ranked (by existing legal doctrine). The method requires both technical proficiency of a highly specialized kind (logical distinction of aspects of the problem; ability to relate instances at one level to goals at another, and so forth) and creativity (sense of the coherence of the large system; ability to isolate semi-intuitively the essentially important aspects from large masses of apparently undifferentiated facts, and so forth). It follows readily that the institutional organization of the application of the method should ideally avoid both the "chain of command" model of bureaucracy and the “interest group” model of collegial decision. The number of people making a given decision should be small, conditions of work should favor intense devotion to the development of expertise, written justifications are appropriate, and internal control is best achieved by “review,” that is the reapplication of the identical method by the supervisor.

The nature of the disputes dealt with also has its organizational consequences: since uncertainty over the application of accepted existing arrangements will inevitably be frequent and compose a large proportion of all the disputes the legal system must deal with, they must be dealt with quickly and as simply as possible; there must be “finality” since these disputes involve individual interests and past acts, the procedures for obtaining information should be oriented toward probing the particular historical incident rather than the broad social trend.

In short, the elaborator should be a judge and the specialized institution a court.

Before passing on to the function of free adaptation, it seems well to underline an aspect of this theory of courts which Hart & Sacks tend to blur. It was remarked earlier that parties to disputes about elaboration tend to limit the area of conflict but to feel very strongly about authoritative resolution of that conflict. Hart & Sacks themselves point out that parties are also likely to feel strongly that the requirement of consistency – that like case be treated alike – is a moral one: it is a demand they make of the system without regard to any notion about maximization of social welfare. Nonetheless, neither the technique of reasoned elaboration nor the peculiar organization of judicial institutions is justified in terms of popular demands or expectations. The first is referred to the function of law and the second to the function of official institutions in a system designed by the planner for the maximization of valid human satisfactions through private action.

The justification of specialized institutions of free adaptation ·is similar to that of courts: the demand for “democracy” plays no greater role than that for “justice,” yet the social character of disputes about free adaptation is important. First, such disputes are likely to involve interests conceived as those of groups rather than of individuals, and they focus on future rather than past conduct. The area of conflict between disputants may be broad and deep: basic philosophical or ideological notions come into play and the elemental organization of the society may be questioned. Yet in contrast to disputes over elaboration, the tolerance of uncertainty or ambiguity in the resolution of questions is likely to be relatively great. The parties feel “entitled” not to a correct solution which vindicates their position but to a chance to influence the course of action, to the consideration of interests for which they claim immediacy rather than universality.

As with elaboration, these expectations are in accord with the intrinsic possibilities of the situation; it is a key tenet of the theory of institutions that questions of free adaptation are in general “not amenable to reason.” By this it is meant that the methods available for solution of such problems in the interest of maximizing social welfare do not provide sufficiently precise criteria of evaluation so that it is meaningful, within broad limits, to criticize the results as either “right” or “wrong”. An institution which laid claim to ability to do “justice” in such cases would be inherently unstable because unable to fulfill its promise. Still within broad limits, legitimacy in free adaptation must be based on procedures rather than on the conformity of the results to an ideal.

The fact that we lack a single determinate method for reaching correct decisions does not mean that we are without means to evaluate procedures. We refer first to the ultimate goal of welfare maximization. In an open-ended process designed to this end it is clearly of the utmost importance that the subjective consequences of legal action be communicated as forcefully as possible to the decision makers. The nature and the intensity of the feelings of all individuals and groups affected by proposed action should therefore be fed in. Since the result of decision is to be in change in what has heretofore been the accepted existing legal order, it is important not only that these opinions be known, but that they receive sufficient recognition in fact so that there is a probability that the change will be in turn accepted and complied with. Flexibility of this kind can be bought at some significant cost in “consistency,” since by definition we are adding to the premises of legal argument rather than attempting to act “in harmony” with them.

In contrast with the practitioner of reasoned elaboration, the free adaptor is less the creative specialist th.an the master of compromise and the “art of the possible.” He must have intuition about the subjectivities of social groups combined with the manipulative ability to fashion law which has a chance of acceptance by heterogeneous interests. It would seem to follow that the institutional organization of the application of these skills should be collegial. Since there will be many relevant interests, there should be many decision makers; conditions of work should favor the development of fine sensitivity to the trend of social feeling; the method of decision should be such as to favor compromise and coalition rather than the consistent imposition of a single point of view; a requirement of formal written justification for action would be unreasonable; internal control can be achieved only by the wide dispersal of power within the institution – hierarchy is impracticable.

Since it is assumed that there is a full, accepted legal system in operation, the function of free adaptation is practically by definition one of “backstopping”, sporadic intervention designed to adapt the system to change and correct past mistakes. This is fortunate: it permits the full development of the procedures necessary to achieve the representation of interests, in spite of the inefficiency of such procedures as rapid dispute solvers. Large collegial bodies with dispersed power and unable to proliferate bureaucratic hierarchies of officials controlled by rules could not possibly perform the front-line work of dispute resolution which falls to the courts. They must operate under rules which strictly limit the number of disputes which can be resolved over a given time period.

In short, the free adaptor should be a politician and the specialized institution a legislature.

Courts and legislatures established according to the theory described above should, according to the theory, refrain from trying to perform functions not allocated to them for the simple reason that they are “competent” only in their own areas. Yet even an observer (or back-seat planner) who recognized that this argument has force and will often be determining might hesitate to accept it as conclusive in every case. Will there not be occasions when the results of an appeal to the legislature are highly predictable? Or when legislative procedures are quite evidently failing to function as they ought? The answer of the theory is the third argument for a strong form of institutional specialization: the necessity of “responsibility.”

Thus far we have treated our specialized institutions as though they were to be autonomous – that is, not subject to any form of control by other institutions. Yet a basic problem for the planner is to ensure that his officials, however competent they may be, are committed in practice to achievement of the ultimate goal of welfare maximization. In order to satisfy himself that they will indeed be faithful to his purposes., the planner will certainly subject them to various kinds of pressure; he will attempt to induce in them a sense of responsibility, either in the form of commitment to internal standards, such as “rationality” or “craft,” or in the form of direct checks, such as popular election, veto, or review. The theory asserts that for each specialized institution, a particular set of checks is desirable in the sense of maximizing control with a minimum sacrifice of efficiency in dispute resolution. Once these specialized control systems are installed, the planner can have some confidence that when it is performing its allocated function the institution will do as well as is humanly possible.

It will be obvious to the planner that such specialized controls should in fact be set up. From his point of view a situation in which all institutions perform their functions as well as possible is ideal by definition. But it follows that he must institute a firm rule that specialization by function shall be complete. Since elaborators are controlled by a mechanism geared only to elaboration, it will be disastrous for them to engage in free adaptation, regardless of how competent they may be to do so. And vice versa. The argument from responsibility thus reenforces that from competence and allegedly renders the case for full specialization conclusive.

A catalogue of control mechanisms is one of the more interesting contributions of the Legal Process materials. Hart & Sacks categorize the means to assure the responsibility of officials as “affirmative” and “negative;” the two groups might be more precisely labelled as cultural and professional norms and institutional checks.

In the broadest sense, a system of control through norms depends for its effectiveness on the existence of a high degree of consensus about the values to be furthered through the controlled activity, on a process by which the official “internalizes” the normative prescription and so controls himself (through conscience, a sense of “craft,” or whatever), and on the existence of the social practice of criticism. Hart & Sacks speak often, but exceedingly vaguely, about instilling a general norm of fair dealing in all who participate in the legal process, and the book itself contains many exhortations to this purpose. The more concrete institution of professionalism plays, however, a much more important role in the theory than do more diffuse cultural patterns.

Systems of professional norms are characterized by an ideal of “service” to “clients,” by the independent (non-employee or “tenured”) status of the practitioner, by the existence of an organized body of colleagues consistently engaged in criticizing one another’s work according to the criteria of the method. The planner is likely to consider it self-evident that a norm system of this type is an appropriate mechanism for the control of institutions designed to specialize in the reasoned elaboration of legal arrangements.

First, the nature of the disputes in question suits them to professional resolution: the area of conflict is limited by acceptance of the existing legal order; that order provides the defined body of premises for the application of a single, determinate technique – that of reasoned elaboration – which provides adequate criteria for criticism by a body of colleagues (and academics). Popular expectations that judges should be independent experts rendering “correct” results are in fact expectations of professionalism. Second, institutional organization designed to assure competence in elaboration will be highly compatible with control through professional norms: decision making by individuals or small groups and the requirement of written justification make criticism easy and favor internalization of the norm system: hierarchic structure and the practice of review guarantee full scrutiny of decisions according to the criteria of the method; the absence of means for the gathering of general knowledge about objective and subjective aspects of the society as a society may be an advantage, since it keeps the body of premises available limited, reduces the likelihood of extreme conflict of point of view among practitioners, and thereby preserves the determinate character of the method, its ability to specify “right” answers.

In the terms of the theory, the compatibility of professionalism with institutions of elaboration is so strong that one is tempted to say that professionalism will significantly increase competence as well as guaranteeing responsibility. In any case, it is a tenet of the theory that the effectiveness of control through norms is so great as to reduce to a minimum the necessity of controlling courts through institutional checks. And this, like so many other aspects of the theory, works out for the best, since checks – powers of other institutions to review, redetermine or veto decisions, or to replace decision makers – are not highly compatible with the function of elaboration.

Since all institutions are specialized, the checking institution will inevitably make its decisions by some method other than that of reasoned elaboration; the result is likely to be the introduction of inconsistencies and anomalies into a system of elaboration which takes consistency and reference to the large purposes of law as the very definition of correctness. Moreover, a system of professional norms, if it is to be effective, must be supreme in the consciousness of the body of practitioners: they must regard the duty to serve and to act correctly in conformity with their method as the highest obligation. If they are subject to check by other institutions operating according to other methods, their responsibility will be endangered. A desire to please, or the necessity of pleasing the checking institution may subvert the independence of the checked.

Yet professionalism has its drawbacks and the direct check its virtues. First, the system of professional norms is a closed one: the practitioners themselves control the body of knowledge and are committed to their method. Since they are unlikely to be more than a small group within the society, and will tend to be isolated by their very devotion to calling, there is a danger that they will be insensitive to change and inexpert in sounding the intensities of the preferences their decisions affect. As objective knowledge becomes more complex, more like professional knowledge in its own right, practitioners of a given discipline are likely to become intellectually isolated. Second, the norm system is overwhelmingly effective only in so far as it is internalized; the process of collegial criticism taken by itself bas obvious weaknesses as a control mechanism if the premises of criticism are not accepted. The professional system is therefore vulnerable to bad faith manipulation by those who acquire its skills without acquiring its ideals.

Institutional checks are likely to be strong exactly where professionalism is weak. The checking institution can be relied upon to feed into the decision process of the checked its concerns and its ways of seeing things where the coherence of the professional method and the static quality of “bodies of knowledge” tend to mask change and conflict, the checking institution can be expected to bring them to the surface and force them on the attention of the checked. And of course, powers of redetermination, veto or election of officials may be effective against the bad faith manipulator, the usurper, where criticism cannot be. Checks provide a kind of safety where norms are weak.

These considerations are likely to lead the planner to the conclusion that even his specialized institution for the professional resolution disputes through reasoned elaboration should be subject to some direct institutional checks. The problem is easily solved: the legislature has in any case the function of free adaptation of the law. It can check the judges by changing the interstitial law they make (a power of redetermination). The organization of the legislature – the inherent limits on the amount of business it can transact and the difficulty it must have in dealing with individual cases – make it highly unlikely that this check can ever be so extensive as to seriously interfere with the professionalism of the courts. (The power of executive appointment of judges is a further institutional check.)

The problems presented by the necessity of ensuring legislative responsibility are the inverse of those presented by the courts. It should be apparent that professionalism is unlikely to be effective in controlling an institution dealing with the broadest sorts of problems, in areas where cleavages over values and norms are profound and there is no determinate technique which offers an assurance that answers are “correct.” Further, a legislature organized to be competent at free adaptation cannot hope to be professional: decisions are the product of compromise in a numerous body; the allocation of individual responsibility is enormously difficult; information gathering techniques which allow sensitive response to a vast number of interests and to all aspects of a situation relevant to welfare maximization will be ill-adapted to the creation of finite bodies of knowledge methodically organized.

Conversely, direct institutional checks in the form of popular election of legislators seem appropriate. Sensitivity to the distribution and intensity of the interests of the citizenry is of the essence of the legislator’s art. And the safety provided by checks is particularly important where the decision maker has a mandate to change the law but cannot be effectively restrained by norms.

The planner’s institutional system can now be described in summary form. In order that the functions of official institutions may be performed as competently as possible, they must be specialized: organized so as to perform one function only. In order that officials may be effectively induced to commit themselves wholeheartedly to the proper performance of their duties, each specialized institution must have a control mechanism adapted to eliciting its peculiar excellence. The results are courts specialized in elaboration and controlled by professional norms, and a legislature specialized in free adaptation and controlled by the direct check of popular election. In theory, so long as each institution sticks to its appointed tasks, each function will be performed as well as is humanly possible in the circumstances; the planner will have constructed the most accurate of possible watches.

But how well is “as well as humanly possible in the circumstances?” It has already been pointed out that The Legal Process is an ambiguous mixture of utopian social theory and description of social reality. Hart & Sacks not surprisingly have something to say about what is “humanly possible” in the United States:

The Constitution of the United States and the various state constitutions commit American society, as a formal matter, to the goal of the general welfare, judged on the basis that every human being counts one – which seems only another way of expressing the objective of maximizing the total satisfactions of valid human wants, and its corollary of a presently fair division. But these constitutions do more than this. They distribute power in such a way as to insure a steady pressure for the continued acceptance and the active pursuit of these objectives. Institutional systems are relatively inflexible. Steps of this kind, once taken, are hard to retreat from. Thus, in American society all the forces bath of vested interest and institutional inertia which are on the side of maintenance of existing institutions are on the side of steadily more effective use of the institutions toward the achievement of their settled objectives.

A planner capable of crediting this quotation as an accurate description of the social situation he confronted (and one may ask whether anyone but a fool could so credit it, or could have credited it in 1957, 1937, 1917 or 1897) would have little difficulty in resolving the final question raised in this section. He would institute forthwith a rule of jurisdiction forbidding courts to perform legislative (or administrative or private ordering) functions, and secure, perhaps by constitutional provision, the elaborative (adjudicative) function from legislative interference. The strong form of specialization would be clearly optimal, and nothing short of it acceptable.

B.

In its own terms, the theory set forth above is an attempted resolution through the technique of reasoned elaboration of a set of particularly disturbing legal uncertainties: those which confront officials, and especially judges, when asked to resolve disputes which arguably fall on the borderline of their appropriate institutional sphere. In the case of the judge, such cases involve arguments about “jurisdiction.” A party asks the court simply to refuse to decide an uncertainty on the ground that it is better left to the legislature, an administrative agency, or private law making institutions. Where another institution has already taken legal action arguably determinative of the uncertainty, the problem of reasoned elaboration may be put in terms of “interpretation” (of a statute, administrative rule or private contract), but this is not necessarily the case. Since courts are “front-line” dispute resolvers (“general jurisdiction”) they will be presented with uncertainties not colorably resolvable by reference to the will of other law makers, but nonetheless presenting the question of what is “in harmony” with the accepted body of existing legal doctrine (e.g. is there a common law right of privacy?).

The theory itself is proposed as “law;” that is, as a body of rules designed to resolve particular kinds of disputes in such a way as to maximize valid human satisfactions. As such, it seems to me that it can be criticized on two levels. One may ask whether the theory is “right” in the sense of offering a set of consistent propositions adapted to achieving the specified goal. And one may ask whether it is adequate specifically as “law,” that is whether it offers a method of decision sufficiently determinate to meet the theory’s own demand that courts behave professionally. The two questions are related. If judges are not convinced that the theory is “right,” then to ask them to apply it would be inconsistent with the professional ethic. If, on the other hand, the theory is not sufficiently determinate to provide a useful basis for control through internalized norms and collegial criticism, then its employment by courts controlled in this way would be inconsistent in the theory’s own terms.

Both suggested questions are posed on the assumption that it is meaningful to speak of a legal system as a purposive enterprise designed to maximize welfare the therefore useful to formulate rules for official conduct to that end.

If it turns out that the theory is inadequate in these terms (e.g. is logically inconsistent or assumes so much contrary to fact as to be useless), then our hypothetical planner has failed: he must begin again to construct a better theory. Yet it should be kept in mind that there are other possible approaches. A social theorist less committed to the working out of “reasoned” solutions to social problems might consider a finding that the theory is inconsistent to be of little interest: a far more important question to him might be “Would the use of an internally inconsistent theory of this type cause the legal system to have a different impact on welfare that it would have if a different theory were in use?” A theorist claiming no interest whatever in concepts like “welfare” might ask: “Is the use of an internally inconsistent theory of this particular type likely to be associated with other social traits? Norms? Technologies?”

After attempting to show that in fact the theory is neither “right” in terms of the maximization objective nor adequate as law appropriate for administration by courts, I will suggest an answer to one extra-systemic question of this sort. How, given its inconsistency, can the theory be brought to bear so as to appear to resolve uncertainties about the jurisdictions of institutions?

The message of the theory to the judge can be summed up in three propositions:

(a) “Each individual deciding officer [is] to reach what he thinks is the right answer.” (italics in original). The right answer is defined as that most conducive to the maximization of valid human satisfactions.

(b) In this frame of mind, the judge is to distinguish those questions which are appropriate to resolution through the technique of reasoned elaboration from those which are not.

(c) The judge is to decide those questions appropriate to reasoned elaboration, and to refer the others to other institutions.

Assume for the moment that the judge is in fact capable of performing the operation described in step (b), that is, that he can make the required distinction in a way which is satisfactory to his sense of obligation to decide correctly, and that there are available criteria of criticism sufficiently precise so that the judge can be effectively controlled by his colleagues (and academics). The question is then: Are proposition (a) and proposition (c) consistent with one another?

The following case can be used to test the matter. Suppose a judge confronted with what is to him clearly a “legislative question,” but a judge who is certain, on reflection, that if he refuses to decide, the legislature will reach a result opposite to that which he would reach were he to set himself the task of maximizing human satisfactions in the particular dispute. Suppose that he then goes on to consider the following: (a) the evil to the present and to future parties of the legislative rule; (b) the evil likely to result through other judges than himself misapplying a decision by him to usurp the legislative function in the case at hand; (c) the evil of a diminution of the prestige and effective of the judiciary resulting from such a usurpation; (d) the likelihood of legislative overruling of the usurpation; (e) the likely ill effects and good effects on legislative process in general resulting from usurpation in the particular case, including the indirect effects on the legislature of actions of other judges and private parties influenced by the decision. Finally, let us suppose that with a full and solemn sense of his obligation to render the correct decision, defined as that which maximizes human satisfactions, the judge concludes that the institutional ill effects will be minimal, while the strictly law-making effects will be large and positive: the balance lies quite clearly in his own mind with usurpation, and the rule of functional specialization is in conflict with the norm of welfare maximization.

One possible way of disposing of this case should be dealt with at the outset: a critic might be convinced that as a matter of fact each body of officials inevitably performed its function in every case as well as it was humanly possible to perform that function. For such a critic, the conclusion would be inescapable that any judge who found himself in the dilemma described above was mistaken in his calculations: it would be logically impossible for his resolution of the dispute to be preferable as law to that of the legislature, no matter how special the particular circumstances might appear to him to be. While it might be difficult to persuade the judge that his perplexity was the product of mere error, the body of critics should have no hesitation in condemning the practice of usurpation whenever and wherever it occurred.

This reasoning seems accurate enough, but it is trivial. Neither Hart & Sacks nor any other commentator I know of would argue that the result of strong specialization will be better in every case than would be the result of specialization seasoned with usurpation. The argument is rather that over the long run fully specialized institutions will perform better than specialized institutions regarding themselves as free to make ad hoc exceptions to the rule.

The problem posed by this sort of interpretation of the rule of specialization is perhaps best understood in the context of the philosophical debate over “extreme” versus “restricted” utilitarianism. The extreme utilitarian position is that if a decision is to be justified as “right” in terms of the goal of maximizing welfare, the decision maker must proceed exactly as the judge in our hypothetical: he must examine in detail the utility consequences of alternative dispositions of the specific case at hand. An appeal to a rule, without more, is never an adequate justification. As put by Professor Wasserstrom, the answer of the restricted utilitarian is that:

there are utilitarian arguments which support the thesis that one ought not attempt to justify particular moral decisions by appealing directly to the utilitarian factors of those decisions. In other words, there may be a utilitarian justification for a moral decision procedure which does not itself consist in direct appeals to the principle of utility as a justification for particular decisions. When the merits of extreme and restricted utilitarianism are contrasted, so the argument runs, it is discovered that if a restricted utilitarianism is consistently employed, it will produce a greater quantum of particular decisions, themselves justifiable on utilitari.an grounds, than will any other possible justificatory procedure. The point of the argument is this: If one is consistently a restricted utilitarian, then a greater number of particular results which are themselves justifiable on utilitarian grounds will be achieved than would be achieved were any other procedure of justification to be consistently employed.

It may appear that restricted utilitarianism offers an exit from the impasse in which we placed our hypothetical judge: he is to sacrifice the maximizing result in the particular case to the rule which maximizes over the long run. The problem, as Wasserstrom points out, is that such a procedure implies an inquiry into the utilitarian virtues of the general rule itself. More specifically, if that rule could be changed so as to exclude the class of cases of which our particularly difficult case is a member (for example by the addition to the rule of a more or less narrow exception), then in order to reach the correct, maximizing result, the judge must make the change and decide the particular case contrary to the old rule. Given the great flexibility of rules, it is practically impossible to imagine a case in which the judge would be required to sacrifice the particular result because unable to reformulate. It now appears that far from requiring the sacrifice of the correct maximizing result to the maximizing rule, the restricted utilitarian position does just the opposite: there is not much left of a rule of specialization which the judge is free to refashion in order to satisfy his notion of the best result on all the circumstances of the particular case.

If the analysis thus far is accepted, it should be clear that there is a serious inconsistency in the Hart & Sacks theory of institutional specialization. The requirement that the official behave as a maximizing utilitarian is simply not compatible with the setting up of any “absolute” rule of specialization. The existence of an inconsistency of this kind affords the judge or critic attempting to apply the theory a choice; he can emphasize the obligation of the official to behave as an extreme utilitarian – with all that implies about the legitimacy of “activism” – or he can bear down hard on the virtues of the rule of specialization as an instance where restricted utilitarianism is appropriate – with all that implies about the judicial role. The uses of this “fertile ambiguity” will be discussed in the last part of this section.

At the risk of making the argument over-elaborate, it seems well to consider a final defense for the theory. It might be argued that the inconsistency discussed above would disappear as far as it affects courts if the planner simply defined any change in the rule of complete specialization as “free adaptation” and required such changes to be made by the legislature. The judge confronted with a situation in which a change (say by addition of a narrow exception) in the specialization rule appeared desirable on extreme utilitarian grounds would be obliged to leave the question of whether such a change should be made to the legislature. It should be apparent that this argument involves a petitio principii. Before the judge can accept the rule of legislative decision about the rule of specialization, he must justify that rule, and this involves him again in all the problems discussed above.

The argument thus far can be stated in another way, and thereby summarized. Suppose a system in which a particular value will be maximized if each element in the system behave in a specified way. If all elements in the system but one are behaving as specified, it is obviously desirable in aid of maximization that the final element be required to behave as specified as well. If, on the other hand, two elements deviate from the norm, but only one can be controlled by the maximizer, it does not necessarily make sense to force that one to conform. Given that the system cannot be made to reach a maximum (there is one element which cannot be controlled), the maximizer must make a detailed ad hoc inquiry as to what behavior by the next to last element will be optimal in the circumstances. This is, once again, what economists call the “problem of the second best.” A judge required to define his own jurisdiction in a situation in which other institutions cannot be relied upon (or compelled) to act in conformity with what he considers ideal institutional arrangements faces the second best problem in a particularly acute form. Hart & Sacks are no help to him. Like the neo-classical economists who elaborated the concept of “perfect competition,” Hart & Sacks base their system on a utopia. Their prescription of full specialization is in the end no more convincing than the economists’ argument for unmitigated laissez-faire.

Thus far our analysis has been based on the assumption that the judge faced with a difficult question of jurisdiction (using the term in its widest sense) is able to distinguish in the terms of the theory between “legislative” and “judicial” questions, so that he need only deal with the problem of whether or not to “usurp.” We will now reverse perspective and ask what guidance the theory offers to the judge who is quite convinced that usurpation is undesirable (he accepts a restricted utilitarian approach to the rule of institutional specialization). The question is whether the theoretical distinction between legislative and judicial questions can be applied in practice with sufficient precision so that the judge is subject to some kind of meaningful control, either by his own desire to reach the “right” result or by consensus among professional critics as to what is required in particular cases.

The first difficulty is semantic: a great deal of mischief has been done by calling the classes to be defined legislative and judicial “questions.” The distinction between usurpation and proper exercise of the judicial role lies, according to the theory, not in the nature of the question asked of the court, but in the answer given. When that answer cannot in good faith be reached by application of the method of reasoned elaboration, then there has been usurpation. The judge must have used some other method, in the application of which he is neither peculiarly competent nor made responsible by effective control mechanisms, and this is the evil at which the theory is directed. The question posed in the first paragraph can therefore be made more precise: does the technique of reasoned elaboration provide sufficiently precise answers to uncertainties about legal arrangements so that it will be clear to the judge and to his critics that he has or has not applied the method in good faith in reaching his result?

If the answer to this question is, No, then the distinction between legislative and judicial functions collapses in practice, and the theory will be useless as a means of maintaining institutional specialization, even assuming that all officials believe specialization to be desirable and usurpation to be a vice.

When the problem is formulated in these terms, it becomes apparent that a number of common interpretations of the distinction between legislative and judicial questions are inconsistent with the theory. I will attempt to dispose summarily with three such interpretations before meeting the theory on its own ground. First, it cannot be that a dispute is “judicial” because it is pressed by individual parties, concerns past acts, is argued in terms of existing legal order and evokes a demand for “justice.” Nor do contrary characteristics in the dispute indicate that it is “legislative.” Judges, according to the theory, “make law,” if only interstitially, and their decisions may therefore always affect future interests of groups. A doctrine of jurisdiction which rested on the ability of these groups to mobilize themselves and their willingness to “politicize” the issue would be doctrine at all: the same case might be “judicial” one day and “legislative” the next. No judge could be expected to do the research necessary to make up his mind about the sociological and psychological questions implied in every jurisdictional question – even supposing that social science offers adequate frameworks for such research in the first place.

Second, it is quite obviously not the individual judge’s frame of mind in hearing the dispute – his effort to find the right result or even his belief that he is applying the method – which defines a judicial question. The theory requires him to adopt this attitude in all cases he hears. His problem is to decide when a proposed line of argument would involve him in a departure from the method (reasoned elaboration) which defines the “right” result. If he makes a mistake, and accepts jurisdiction in a case not amenable to resolution through the method, or reaches a result the method does not justify, then he should be criticized, no matter what the frame of mind in which he erred.

Third, it cannot be that because a dispute is argued in whole or in part in terms of “social policy” it is therefore “legislative.” Nor can it be that because in order to reach a particular result the judge must make arguments from social policy, that result represents usurpation. The technique of reasoned elaboration – the only technique appropriate for judicial use – is itself nothing more than the technique of arguing from the “principles and policies” underlying the particular elements of the legal order. A judge who refused to work his way upward, from the rule to its reason, from that reason on toward the ultimate purpose of law, would be incapable of properly fulfilling his specifically judicial function. For Hart & Sacks, every legal question is a question of social policy.

We are left, then, with compatibility with the method of reasoned elaboration as the sole criterion of a “judicial” resolution of an uncertainty. When in a given case there is no resolution possible which is consistent with the method, then the subject matter is “non-justiciable;” when there is a single result consistent with the method, to reach a different result is, to use one of Hart & Sacks’ favorite phrases, “pure fiat” – a bald assertion of an usurped legislative power. Our problem is to evaluate the method in terms of certainty.

It seems best to exclude at the outset two meanings of certainty clearly not intended by Hart & Sacks. They do not mean to suggest that reasoned elaboration provides “answers” to uncertainties in the same manner that logico-deductive method provides answers in mathematics or symbolic logic. The complexity of the social reality with which the planner must deal, and the consequent imprecision in the process of formation of concepts and premises for argument, are fully apparent to them: they are enthusiastic participants in the “revolt against formalism” and “mechanical jurisprudence.” On the other hand, they are equally opposed to the notion that certainty can be reduced to a “high probability that courts will decide a particular question in a particular way.” Their method is to be applied self-consciously by judges and critics striving to find the right answer in the particular case. Its utility depends not on statistical regularity in results, but on the conviction of the planner and the participants that one can reach right answers sufficiently often to make it legitimate to use the method to police the borders of institutional domains. Hart & Sacks are above all leaders in a “revolt against Realism.”

They appeal, then, to our sense of the social possibilities of the use of Reason in human affairs: they suggest the delimitation of the sphere of judicial action in terms of our notions of how far we can get in the resolution of disputes through the working out of widely shared purposes based on the ultimate shared purposes of maximizing valid human satisfactions through law. Without challenging, for the moment, the larger view of American society implicit in this approach, one can criticize Hart & Sacks’ view of reasoned elaboration as over-sanguine in two respects.

First, let us suppose most favorably to the theory, that most disputes involve the reconciliation of conflicting purposes far more specific than “maximization of human satisfactions.” Take the instance of an inheritance statute which directs the disposition of property in accordance with the desires of the deceased as expressed in a written will, but says nothing about the case of a murderer inheriting by will from his victim. Hart & Sacks argue that an asserted uncertainty in the meaning of the statute as to murderers can be resolved by appeal to the legal principle that “no man shall profit by his own wrong.” This principle outweighs in this case the countervailing utilitarian principle that “every alienation imports advantage” because we cannot presume that the testator was aware of the murderer’s intentions at the time he wrote the will. Resort to the purposes of the law as embodied in its unifying principles thus assertedly resolves the uncertainty at a level of abstraction far short of the ultimate purpose of maximization.

There is something fishy here. The utilitarian purpose lying behind the principle that “no man shall profit by his own wrong” is that “the law should not increase the frequency of socially undesirable actions by rewarding them.” Yet we have no more than the most imprecise and intuitive evidence that failure to amend the inheritance statute by judicial decision would have an effect on any kind on the number of murders of testators by heirs. A sketchy knowledge of modern criminology suggests the opposite: the motive of gain seems virtually never present in theories of intra-family crime. In this light, we may wish to reassess the force of the principle that “every alienation imports exchange” as applied in this case. After all, the only hard evidence we have as to the testator’s intent is the will. Do we endanger the general policy of testamentary freedom by importing an exception on so little evidence?

The answer to the problem posed by this case is probably that testamentary freedom is sacrificed not to higher shared purposes but to a higher shared morality. Deterrence aside, it shocks the conscience to reward the murderer. The trouble is that this has little or nothing to do with the method of reasoned elaboration. First, to assimilate “moral action” to “action designed to maximize satisfactions” will appear to all but proponents of utilitarian ethics a gross distortion of one or bath concepts. Second, even if we accept the assimilation, the question of how much “morality” can legitimately be imported into “law” is an extremely difficult one, probably resolvable only at the level of ideologies. If we invite the judge to reach a position on the issue as a step in resolving the uncertainty over the inheritance statute, we will hardly be able to claim a great deal of certainty for the method of reasoned elaboration. If, the other hand, we instruct the judge that the question of the proper role of morality in particular statutory arrangements raises questions “not amenable to reason,” he will be unable to resolve purposively a case which Hart & Sacks consider quintessentially “judicial.”

The point of this first objection to the method can be summarized as follows: a large proportion of the principles and policies to which the method attributes purpose cannot, because of objective and subjective uncertainty about the consequences of decision in particular cases, be applied with any confidence if they are to be justified on utilitarian grounds. If these principles and policies are to be applied without strong evidence as to effects in particular cases, then the judge’s reconciliation of conflicts among them can be justified only in terms of more abstract uncertainties. As to these, we must recognize that the method is either largely indeterminate or not appropriate at all.

The existence of uncertainty of this kind affords the critic or judge applying the theory a useful choice: in the case of virtually every legal uncertainty he will be able to argue that the application of a particular principle or policy is unjustified because “data” is inadequate; alternatively he can argue from the fact that the principle or policy is “deeply rooted” in the law that it is entitled to substantial respect as representing the wisdom of a great tradition of thinkers about the maximization of welfare.

The second, and more basic objection to the method is implicit in the description of the first. It should be apparent that the resolution of uncertainties by the interstitial harmonization of the particular area with larger principles and policies of the law can be successful (or meaningful) only to the extent that the larger principles and policies are consistent among themselves. For example, in the case of the inheritance statute it is possible to interpret the principle “a man should have the power to dispose of his property as he sees fit” as derived from the principle “every alienation imports advantage.” This principle is consistent with the principle “the law should not increase the frequency of socially undesirable actions by rewarding them,” because both are derived from a common (utilitarian and individualistic) view of man’s psychological and social nature. When they conflict, we can “weigh” them against each other in the terms of their common utilitarian framework.

The case is far different when we deal with principles derived from different systems (ideologies, cognitive structures, philosophies, or whatever word is in fashion). “A man should pay damages when he has been at fault,” cannot be “weighed” against “Damages should be allocated among actors so as to maximize deterrence” unless we are willing to ascend simultaneously two distinct hierarchies of purpose to the ultimate conflict over the social function of law. The same is true of “pacta sunt servanda” matched against Holmes’ “the payment of damages is a tariff legitimating a breach of contract.” As for the criminal law, it is notorious that it is grounded in conceptions of man’s nature so radically incompatible that it is impossible for their expositors to hold meaningful conversation. Any comment as to “equal protection of the laws,” or the taking of life, liberty or property without “due process of law” would be superfluous. Two more examples will perhaps be enough to make the point. In antitrust law, there coexist two equally long and equally distinguished lines of cases, one based on the notion that the goal of maximization implies that the economy must be made to correspond as closely as possible to the model of perfect competition, the other on the notion that in an imperfectly competitive, regulated economy the law can achieve no more than atomization of economic power. In business law, the corporation can be conceived either as an association of private individuals or as an arm of the state. (An elaborate and far-reaching case of this kind can be developed from Pound’s brilliant 1917 article on the development of the antithetical concepts of contract and status during the nineteenth century.)

It may be that it is hopeless to argue over the extent of conflicts of this kind among the premises available for the method of reasoned elaboration. I should say that I think such conflicts pervasive, and that in areas where the law appears highly consistent the constant trend is to the development of new models – such as Professor Griffiths’ “family model” of criminal procedure – while there is no over-all trend to “rationalize” the areas where incompatible models coexist. It is clear, however, that some such areas of conflict do exist, and that they tend at present to be centrally important so far as the place of the judiciary in the public consciousness is concerned. The implications for the method of reasoned elaboration are thus interesting even if one denies that alternative “ideologies” are always available to support alternative resolutions of particular legal uncertainties.

First, it should be clear that the judge making a good faith effort to reach the correct decision through reasoned elaboration can perfectly consistently mount the two hierarchies of values implied in two different results until he reaches the choice between ideologies, and that he can then make the choice according to which system will, in his earnest opinion, serve to maximize valid human satisfactions. This choice once made, one or the other of the resolutions of the specific problem will in all probability appear quite distinctly “correct.” Yet if it is open to the judge to make such choices, it is difficult indeed to see how his specialized method of decision differs from that of the legislator: if the category of “choices amenable to resolution through Reason” has any meaning at all, it must exclude the choice of ideologies (cognitive structures, philosophies, world-views, systems). If, on the other hand, we deny the judge the right to choose at this level, require him to halt at some point in the hierarchies of purposes, the method of reasoned elaboration becomes meaningless: everything will depend on the choice of a set of premises within which to “harmonize” the law, but the judge is forbidden to justify that choice.

As with the other aspects of the theory, the indeterminacy of the method when applied to a legal system riddled with inconsistency has its advantages for the critic employing it as a tool in argument. He can assert with equal plausibility about almost any legal dispute, first, that a given result can be reached only by making an essentially “legislative” choice of ideology, and, second, that the result he favors is compelled, since it is the only one the judge can reach if he acts consistently, mounting the hierarchy of purposes within the critic’s system.

It remains to draw the consequences for the theory of institutional specialization. We have not tried to state with any precision the degree of uncertainty of the method as applied to even slightly difficult cases. The reader can form his own estimate on the basis of the degree of inconsistency and reliance on incompatible models he sees in the body of premises for legal reasoning. For the judge or critic who feels the degree of indeterminacy to be great (as I do), the consequences for the theory of specialization are also great: if he cannot, consistently with a strong conviction that he is reaching the correct result, distinguish those questions which are non-justiciable and those results which represent “pure fiat,” the injunction against usurpation becomes meaningless, even supposing that he is convinced by the (specious) argument that usurpation is in all cases undesirable.

As for the appropriateness of a distinction between legislative and judicial questions as a basis for a professional norm of specialization, one would expect the indeterminacy of the method to produce conflict among the body of critics, as indeed it has. In so much as there is consensus, it must be traced to something else than the “correctness” of the theory, e.g. a pre-existing ideological consensus among legal scholars or a structure of professional rewards (from law school through the granting of tenure) which functions, however inadvertently, to deaden independent thinking.

It can be convincingly argued that the theoretical schemas of Hart & Sacks have in fact added an invaluable aura of intellectual respectability to what is essentially a set of ideological biases: the position of the influential group of legal thinkers denominated “the modest” by Professor Shapiro seems to rest, however unstably, on the conceptual framework laid out in The Legal Process. Yet it is important to recognize that Hart & Sacks themselves do not fall within a school. The materials are full of self-righteousness but in no sense predictable, full of elaborately structured arguments but also slippery. The ambiguities and inconsistencies of their theory of institutions are “fertile,” permitting the development of two internally consistent but mutually incompatible sub theories of the judicial role. Much of the force and apparent coherence of the positions taken in the materials derives from the use of one or the other but never both of these sub theories in each particular case.

The Passive Model

When Hart & Sacks are against a proposed judge made change in the law, they tend to make arguments based explicitly or implicitly on the following view of the legal system.

(a) While many areas of legal action are consistently structured in terms of common views about how men should go about maximizing valid human satisfactions, the result urged in this case cannot be reached without an excursion into areas where there are no commonly accepted, consistent legal premises for reasoned elaboration. The existence of areas of consistency is simply asserted the inconsistency in the area in question is demonstrated more or less as was done above.

(b) The available factual information is inadequate in the particular case to permit the intelligent weighing of conflicting policies. This inadequacy is related on the one hand to the inherently “legislative” nature of the question, and on the other to the institutional incompetence of courts as information gatherers.

(c) There are alternative institutions designed specifically to be both competent and responsible in making decisions of the kind proposed, and it is asserted that if judges considered themselves free in every case to question the actual performance of those institutions in the particular disputed area, the result would be chaos and anarchy.

(d) There is an alternative (which may be a holding of non-justiciability) to the proposed “legislative” resolution of the dispute an alternative consistent with the method of reasoned elaboration. The parties have a legitimate expectation that the alternative result will be reached this demand for “justice” is in conflict with the judge’s impulse to treat the case legislatively.

(e) The conclusion is clear that the proposed change would be “pure fiat.”

The Activist Model

When Hart & Sacks are in favor of a proposed judge-made change in the law, they tend to make arguments based explicitly or implicitly on the following view of the legal system.

(a) While there are some areas of legal action so riddled with inconsistency and so devoid of consensus that they are inappropriate for action by courts, the result urged in this case can be reached along a path no more difficult than that normal in the law. True, there are inconsistencies, alternative models and passionate disputes, but many of these are the produce of error or misunderstanding and can be illuminated by the application of Reason. For the rest, no claim has been made that the technique of reasoned elaboration can be applied like a mathematical formula to reach results indisputable by all but fools. The judge’s elemental responsibility is to judge.

(b) The factual data on the impact of a given policy in this particular case may be weak, yet the principles and policies urged upon the court represent the accumulated wisdom of centuries during which the common law has been working itself pure. This wisdom is more enduring and more useful than any number of research projects based on a still nascent scientific theory of society.

(c) Alternative institutions may be of some help should the judge refuse to resolve this uncertainty, yet it must be kept in mind that courts are the front-line dispute resolvers in any society. Legislatures are not and could not be designed to do justice in particular cases, and here it is evident that they are likely either not to act at all or to perform their function with peculiar incompetence and/or irresponsibility.

(d) Both parties to a lawsuit expect to win or there would be no lawsuit. To speak of a “justice between the parties” which is not related to the larger “social justice” which consists in maximizing valid human satisfactions is meaningless. Where a party relies on a doctrine which sacrifices the common good to his particular good, his reliance is “unjustified.” This is especially true since a more spacious view of the role of law would place greatest emphasis not on the division of a static pie among irreconcilably hostile parties but on devising flexible rules which favor a dynamically expanding pie.

(e) The conclusion is clear that to decide against the proposed change would be “denial of justice”.

These two styles of argument are, at this point in history, only too familiar on the American legal scene. The point is not that one is “better” than the other, or even that Hart & Sacks misapply them (although I often disagree with their particular results). It is that they are equally consistent with the method of reasoned elaboration and the theory of institutional specialization. It all depends on whether you admire the “fine spun texture” of the law or rather enjoy its “fertile disorder,” on whether the functioning of legislatures, administrative agencies and private orderers (especially large corporations) gives you a feeling of security or fills you with fear and outrage. Most people seem to feel consistently one way or another about these things: they become activists or pacifists, and work of Hart & Sacks should buoy them equally. But what of Hart & Sacks themselves? If it is not their “system” which turns them first one way and then another, what is it? Two answers are possible, one hopeful, the other not.

It is possible that they are simply “pragmatists,” ambivalent about the workings of the society and the coherence of the law, responding idiosyncratically according to the emotions evoked by the particular case, but altogether unwilling to accept the cosmic angst an admission of freedom would entail. And then it is possible that the logical inconsistency, the naivete, the incoherence of their theory of the legal process is irrelevant: what matters is the cases, for if we analyze them with sufficient skill we will discover the cognitive keying system which turns these consummate judges first to the activist, ten to the pacifist model. In short, behind all the indeterminate talk of Reason we will find the “real” theory which is the legal mentality.

Duncan Kennedy

Duncan Kennedy est Carter Professor of General Jurisprudence à l’école de droit de Harvard (Emeritus). Ayant directement participé à fonder le courant des Critical Legal Studies, le professeur Kennedy a œuvré à proposer une analyse critique du droit privé et de la pensée juridique américaine. Auteur prolifique, il a notamment publié Sexy Dressing Etc. Essays on the Power and Politics of Cultural Identity (1995), A Critique of Adjudication. Fin de siècle (1998) et The Rise and Fall of Classical Legal Thought (2006). Duncan Kennedy est fellow de l’Institut Michel Villey. Il a été en 2018 le titulaire de la Chaire Michel Villey (Paris Jurisprudence lectures).