APA Newsletters

Spring 2001
Volume 00, Number 2

Newsletter on Philosophy and Law

Articles

Previous Article | Index | Next Article

The Methodilogy of Theory Construction in Jurisprudence

Jules L. Coleman

For a community to have law is for it to have a practice of a certain kind, identified by familiar institutional forms and constituted by certain structures and principles of governance. The aim of jurisprudence is to illuminate important, perhaps essential features of legal practice-to uncover what is distinctive of law. Jurisprudence proceeds in large part by analyzing the concepts that figure prominently in our legal practice, including, of course, the concept of law itself. The goal of conceptual analysis is not to illuminate the way we speak about law, but to understand the institutions of law. Underlying the project is a view-pervasive in philosophy generally-that the distinctively philosophical way of learning about the world is through an analysis of the concepts by which we comprehend it rationally. Legal Positivism, Natural Law Theory and Dworkinian Interpretivism are all examples of substantive jurisprudential theories that attempt to clarify and to instruct us about this important dimension of our social world.

Recently, the work of Ronald Dworkin, Stephen Perry and Brian Leiter, among others, has stimulated interest in the methodology of theory construction in jurisprudence. Leiter has challenged the fruitfulness of conceptual analysis as a device for illuminating central or important features of legal practice, while Dworkin and Perry have pressed the argument that conceptual analysis of law falls within the province of normative, not descriptive philosophy. Against Leiter, it needs to be said that there is a significant empirical dimension to conceptual analysis generally, and to analytic jurisprudence in particular. Leiter's portrayal of analytic jurisprudence as "intuition pumping" and armchair a priori reflection overstates and misdescribes the role of intuitions in determining the content of our concepts. The picture he draws is mistaken in two ways. First, there is a significant a posteriori dimension to analytic jurisprudence. Second, the choice among conflicting theories of concepts answers to a range of epistemic, theoretical, and especially discursive norms, in the sense that the theory of a concept we adopt represents a decision at one node in our overall theory of the world. We don't simply "pump intuitions" in a theoretical vacuum. Rather, what we are after is an overall theory of the world, and we defend a theory of any particular concept as part of any such theory. We should therefore expect of a theory of the concept of law that it illuminate the relationships between law and other practices in its social neighborhood-for example, the notions of authority, rules, power, coercion, and so on. In engaging in conceptual analysis, we need not be understood as offering arguments to the effect that this or that feature of law is necessary in the sense of constituting a conceptual or analytic truth about the concept; instead, our arguments might well rest on the claim that certain ways of thinking about law illuminate its relationships to concepts in its discursive neighborhood, and to practices in its social neighborhood. This fundamentally Quinean picture of theory construction is very different from the naturalistic side of the Quinean project that Leiter emphasizes. It is this side of Quine-along with the holism and commitment to revisability-that helps to distinguish my work and my methodology of theory construction in law from that of others. Moreover, Leiter's own naturalist project requires so much by the way of familiar conceptual analysis that it could not qualify as an alternative to analytic jurisprudence. His claim that conceptual analysis is fruitless threatens the coherence of his own project.2

The more pervasive, familiar and potentially more powerful objection to the traditional methodology of theory construction in jurisprudence comes from those, like Dworkin and Perry, who claim that conceptual analysis of law is fundamentally normative. These critics do not deny the possibility of fruitful conceptual analysis. They deny only that conceptual analysis is, in any interesting sense, descriptive or value free. Of course there is an unobjectionable sense in which all conceptual analysis-which by its nature seeks to regulate the use of language and concepts and to structure thought accordingly-is a normative endeavor. However, in claiming that the philosophical analysis of the concept of law is normative, Dworkin and Perry mean to make a much stronger and less philosophically innocent claim than this. For them (and for others as well), any philosophical analysis of the concept of law necessarily relies on substantive and contestable premises and principles of political morality. All theory construction in jurisprudence is normative in a sense that extends far beyond the general regulatory aims of conceptual analysis; jurisprudence is, on their view, normative in a way that is incompatible with the methodological claims that Hart and I, among others, have advanced. To engage in the construction of a philosophical theory of law (or of the concept of it) is to engage in first-order, substantive political philosophy-not, to use Hart's phrase, to engage in descriptive sociology.

Though both Dworkin and Perry advance the thesis that jurisprudence is normative in this strong sense (a sense that I shall denote henceforth by the expression "normative jurisprudence"), each defends his claim on very different grounds. Dworkin's most famous argument for normative jurisprudence is the so-called Semantic Sting. It attempts to derive certain claims about the proper methodology of theory construction in law from more general views about the content and application conditions of concepts; in this sense, his methodological claims have a semantic foundation. Perry's arguments have an altogether different character. Instead of focusing on the concept of law itself or on the nature of concepts more generally, Perry explores the way in which Hart develops and defends various claims in the construction of his overall theory of the concept of law. As Perry reads him, Hart's theory of the concept stems both from a substantive and contestable view of law's proper function, and from a morally infused conception of the internal point of view. Where Dworkin defends normative jurisprudence, as it were, from the top down, by deriving it from a general theory of what is involved in grasping the content of a concept (or of a certain "contestable" sort of concept )-Perry defends normative jurisprudence from the bottom up, by showing how theorists (even those who, like Hart, explicitly resist the label of normative jurisprudent)-cannot construct their theories without invoking substantive norms of political morality. If Dworkin and Perry are right, one cannot construct a theory of law, or a theory of the concept of it, without invoking substantive premises of political morality; and in that sense, substantive principles of political morality are necessary to theory construction in law in ways that belie the methodological claims of positivists like Hart and myself.

The thesis that jurisprudence is normative comprises two related claims-one of which is explicit in the prior discussion, the other of which is implicit in it. The already explicit claim is that theory construction in law necessarily invokes substantive and contestable premises or principles of political morality. The implicit claim is that this fact about theory construction in law is in principle consistent with the full range of received substantive theories of law. The first claim expresses the universality of the claim made by normative jurisprudence; the second expresses its supposed neutrality. Let us consider in turn each of these aspects of the position.

The claim that all jurisprudence is normative is by now familiar, though it is not often expressed with adequate precision. All conceptual analysis is normative in the sense that it constructs theories of concepts, and such theories in turn structure thought and regulate use. We can say that our concept of X, in this sense, always depends on what our concept of X ought to be. Thus, our concept of law depends on what our concept of law ought to be. Those who defend normative jurisprudence are making a different claim than this, but how can we formulate that claim precisely. One obvious suggestion is that to say that an analysis of our concept of law must be normative and thus that it invokes substantive principles of political morality is to say that our concept of what law is depends on what law should be. On this view, if ordinary conceptual analysis is normative in the sense that our concept of law depends on what our concept of law ought to be, then the claim that conceptual analysis of law is normative is the claim that our concept of law depends on what law (and not just our concept of it) ought to be. If what law is depends on what it ought to be, then any analysis of what law is will be normative in the sense of invoking substantive principles of political morality. Is this how we should understand the claim of normative jurisprudence?

In general it is not true that our concept of X depends on what X ought to be. The claim that what law is depends on what law ought to be is distinctive, but it is also particularly strong and controversial: strong and controversial because it appears to build the legitimacy conditions of law into our analysis of what law is. While some may defend such an account of law, it may be stronger than necessary to underwrite the normativity of jurisprudence. Let's try a somewhat weaker thesis. Instead of claiming that our concept of law depends on what law ought to be, one could argue that our concept of law depends on a substantive view about why law is a good, valuable or otherwise desirable form of governance. Thus, to understand what law is, we need not have a view about what law ought to be, but we need to understand why governance by law is valuable.3

The second claim of normative jurisprudence is the neutrality claim-the claim that the normativity of theory construction is neutral with respect to the variety of substantive theories of law or of the concept of it. While denying that one can pursue a (morally) value-free project of conceptual analysis, normative jurisprudence also insists that the inevitable normativity of theory construction in law begs no questions regarding which substantive theory of the concept is best. Thus, it claims that the (quite distinct) positivist views that Hart, Raz and I advance, as well as Dworkin's interpretivism and Finnis' natural law view, all partake of materials drawn from political morality; yet this fact about theory construction is not supposed to undermine positivism, or to rule out that positivism in one or another form might be the best theory of our concept of law. In other words, the unavoidable normativity of the method of jurisprudence is not supposed to entail that the correct theory of the concept must assert any necessary connection between law and morality-in any case, not any connection that might be incompatible with legal positivism. The compatibility of the universality and neutrality claims gains some credence from the fact that some positivists, like Gerald Postema, appear to accept the claims of normative jurisprudence. They accept, in other words, the inevitability of political morality in theory construction, but take it to be compatible with defending a substantive positivist position .

The claim that one can have it both ways-maintaining the moral character of the method of theory construction in law while endorsing the possibility of a substantive theory of the concept that rejects the necessity of the relationship between law and morality-is easier made than defended. It is a claim that can be defended, but its defense depends on one's grounds for believing that theory construction in law must involve substantive normative premises. Not every view about why jurisprudence must be normative is compatible with all substantive theories of law. This is easy to see. Were one to defend the normativity of jurisprudence by starting with the premise that law is necessarily the kind of thing whose claim to legitimate authority is justified or prima facie justified, then it would be easy to establish why an analysis of the concept of law would have to invoke substantive moral premises. But that initial premise, though strong enough to make out the universality claim, precludes the neutrality claim. It entails that all substantive theories that rule out the inference from legality to (prima facie) legitimacy are mistaken. It substantiates the normativity of jurisprudence at the cost of ruling out the very possibility of a positivist jurisprudence. If the claims to universality and neutrality are to be reconciled, then some possible lines of defense for normative jurisprudence are ruled out from the start.

Dworkin's Semantic Sting is instructive in this regard. Dworkin maintains that Legal Positivists like Hart are committed to a "criterial semantics." According to criterial semantics, those who share a concept or understand it must share criteria for applying it. They can disagree about whether this or that is an instance of the concept, and they can disagree about what the criteria for the concept ought to be; however, they cannot disagree about the criteria for applying it-for that is what, according to criterial semantics, they must share in order to have the same concept. But, Dworkin argues, those who share a concept can and often do disagree about the proper criteria for applying it. He calls such disagreement "theoretical," and notes that it is a pervasive feature of legal practice. Theoretical disagreement thus cannot be accommodated within criterial semantics-at least for a broad category of contestable concepts, including, for example, art and law. If we are to allow for the possibility-let alone the inevitability and pervasiveness-of disagreement over the criteria for applying concepts like art and law, then our analysis of such concepts cannot consist in identifying shared application conditions. If the semantics of a concept is not criterial, Dworkin contends, then it must be interpretative. Interpretation is normative. The norms governing interpretation depend on the nature of the object of interpretation. The norms governing interpretation of art are (in the first instance) aesthetic; those governing the interpretation of what law is are principles of political morality.4

If law is an interpretive concept and if the norms appropriate to its interpretation are those of moral and political philosophy, then any theory of the concept of law must involve an exercise of substantive moral and political philosophy. Thus, the second claim: any substantive theory of the concept of law-positivist, natural law, realist, whatever-is constructed from premises of first order political philosophy. They rely on a contestable conception of the law's proper function. Thus, all theories of law are constructed from premises of political philosophy. But that fact by itself does not tell us which substantive theory of law is the correct or most illuminating one. It tells us only that any theory of the concept of law draws its lifeblood from substantive views about the proper point or purpose of governance by law. Such a view about law's proper function would be part of a more general account of why governance by law is valuable or desirable.

In The Practice of Principle, I argue against Perry that Hart's claim that the function of law is the guidance of conduct is part of a familiar functionalist explanatory project, and not part of a moral theory of law's proper function. Hart thought that by ascribing to law the function of guiding conduct through reasons, we could explain why legal systems arise and persist over time, and especially why they take the shape they do in their mature form.5 Hart certainly never claimed-nor does any part of his theory require him to claim-that guiding conduct by reasons is law's morally proper function. Moreover, Hart does not see the internal point of view-the point of view of one who acts for the reason that the rule requires her so to act-as morally superior to the point of view of one who acts to avoid the imposition of a sanction. Nothing in his argument depends, one way or the other, on the relative merits of acting for one or the other sort of reason. Rather, the internal point of view is a conceptual necessity within Hart's theory. Legality itself depends, at its foundation, on a conventional rule-the rule of recognition-and the existence of that rule requires, as a conceptual matter, that certain individuals adopt the internal point of view toward the rule-they must act for the reason created by the rule. The internal point of view enters Hart's account as a conceptually necessary condition for the existence of legality as such, and not as a morally attractive form of motivation.6

Dworkin's Semantic Sting argument has been the focus of a great deal of attention, almost all of it critical. Most of the objections to it are familiar, and frankly, well founded. I spend very little time on those. Instead, I note that Dworkin confuses two different projects-the project of retrieving the content of the law of a particular community, and the project of retrieving the content of the concept of law. He regularly draws conclusions about the latter project from arguments that apply only to the former. In addition, he mistakes the kind of coordinative activity that is involved in sharing concepts and in communicating through them. An account of what it is to share or have the same concept can be perfectly compatible with the kinds of theoretical disagreement that we associate with so-called contested concepts. I develop the outlines of one such account in the book.7

Even though none of the arguments offered in defense of normative jurisprudence by its strongest advocates-Dworkin and Perry-are persuasive, we should not let the matter end there. That is because what drives the conviction that there must be something to normative jurisprudence is the very sensible and plausible thought that to understand what law is, we need to understand why governance by law is preferable to other schemes and systems of governance. This is not to presuppose that law is morally legitimate-since it often is not, and certainly need not be-rather, it is simply to acknowledge an important aspect of the way we regard law. If this is at the core of the normative jurisprudential project, then we can see why Perry is drawn to arguments designed to show that Hart's analysis of the concept of law-his theory of law-ultimately must rely on contestable normative premises about the value of governance by reason, rather than by sanctions.

Having found the arguments advanced by others in favor of normative jurisprudence unpersuasive, I want to offer a different and I think more general and more powerful one on behalf of the project of normative jurisprudence. In doing so, I may appear to be arguing against interest, but this would be shortsighted. In fact, though the argument I am about to offer does not in the end amount to a defense of normative jurisprudence, it allows us to see why we are drawn to the thought that we need to understand why governance by law is desirable or valuable if we are to understand what law is. Yet in the end we will see that this implies nothing about whether our concept of law contains substantive political commitments. Thus, the following argument illustrates the insight those who defend a normative jurisprudence are drawn to, and explains why one ought to be drawn to it. At the same time, it explains why this insight is inadequate to make the case for normative jurisprudence.8

When we speak of law, we mean a form of governance that constitutes a distinctive normative relationship between the governing individuals or institutions, and those who are governed by them. The way that law structures this relationship is different from, and in some sense morally preferable to, the relationship between, for example, the ruling powers and the subject population in the case of a military occupation. More generally, governance by law is preferable to governance by force and fear. Any plausible account of law must not only make plain the differences among those forms of governance, it must do so in a way that explains-or enables us to explain-why we believe legal governance is morally attractive. We can capture this condition by saying that law is a predicate of commendation.9

In characterizing law this way, we do not imply that legal authority is always morally legitimate or justified, or even that any actual instance of it is. We certainly need not claim that the law in a particular community is justified merely in virtue of its status as law. Law is, in this sense, a predicate of weak commendation, and we may contrast it with a predicate of strong commendation such as "justice." A theory of justice would be implausible on its face if its extension included morally undesirable social, political, or economic arrangements. The property of moral legitimacy is an essential, or a central, feature of our concept of justice, and an argument to the effect that an analysis of justice picks out some morally illegitimate social arrangement is an argument for the inadequacy of that analysis. By contrast, laws-and perhaps even legal systems-can be morally illegitimate, and more often than we would care to believe, probably are. Nonetheless, we seem inclined to acknowledge that there is something commendable about law as such.

An argument for normative jurisprudence begins with this weak commendation feature of the predicate "law." If law is a predicate of weak commendation, then one could argue that the best explanation of how it is that law plays this role in our normative discourse is that law has a moral property adequate to warrant the discursive role of "law". That is, there must be some moral property that law has that provides the explanation of the fact that "law" plays this commendation role in our normative discourse. Every instance of law must possess this moral property, M-something weaker, perhaps, than either moral legitimacy or prima-facie legitimacy. If law has such a property, an analysis of the concept of law should specify what this property M is-and in doing so, the analysis must appeal to moral argument in two ways: in specifying the content of M, the analysis must reveal how or why that property is morally attractive; and the overall analysis must be such that it picks out only things that have the property M. Thus, in selecting the other elements of the analysis of law, theorists are constrained to select a set of elements that are sufficient for, or at the very least consistent with, M. Succeeding in that is an adequacy condition for an analysis of law, and necessarily entails engaging in substantive moral argument.

This sounds like a case for normative jurisprudence, so let us consider the point more carefully. We have already granted the first premise, that law is a predicate of weak commendation. This is simply a fact about the role that "law" plays in our normative discourse. In acknowledging this, we leave open the question of how or why "law" plays this role. One answer is that provided in the foregoing argument, namely, that there is a moral property, M, that explains how law could serve as a predicate of commendation in our normative discourse. Law plays a commendation role in our normative discourse because the concept of law is the concept of something with a certain morally attractive property.

Clearly, the property of being morally legitimate would suffice to explain law's commendation function, but few normative jurisprudents would wish to endorse the inference from legality to moral legitimacy. There are other values for M that might also explain the commendation feature. More importantly, it has yet to be established that the existence of an essential moral property is the only way we can make sense of the commending role "law" plays in our normative discourse.

It could be argued, for example, that the commendation feature of law is simply an induction over experience. The historical record provides us with examples of a variety of different kinds of governance, legal and otherwise, and the legal ones seem to be preferable on balance. Or perhaps it is something even less creditable: a shortsighted induction. Maybe recent legal systems have been better than their alternatives. For one reason or another, we have formed positive associations about the concept, and that is what explains the role that law plays in our normative discourse. However, this is just an accidental feature of law, and has no bearing on the content of the concept or on the proper method of jurisprudence.

The point is debatable, but I do not want to defend it here. The immediate rejoinder will be that if we have inductively based beliefs about the moral attractiveness of law, these beliefs are not be to explained by mere historical contingency-by brute facts about how various forms of governance happen to have worked out. If law is associated historically with a more humane or just form of governance than its alternatives, that fact (it will be argued) is owing to something inherent in the nature of law. Indeed, it is not clear that we should even accept the claim that our beliefs about the attractiveness of law are inductively based in the first place. It is not obvious that the historical record-whether on a long or short view-presents an unambiguously attractive picture of legal governance, or that the commendation feature really depends on the record's doing so. It seems likelier that the explanation of the commendation feature lies not in what laws and legal systems have actually been, but, rather, in what they can potentially be. That is to say, inherent in the nature of law is the potential for a kind of governance that we feel is more morally attractive than alternatives. Our concept of law is the concept of something that has the inherent potential to achieve, realize, or take the form of a certain ideal of governance.

The suggestion now is that the morally attractive property of law is its inherent potential to realize or manifest an ideal of governance. As a potential, it need not be realized in every instance of law, and that explains once and for all why the argument for normative jurisprudence need not endorse the inference from legality to legitimacy. Yet at the same time, if this potential is an essential or central feature of our concept of law, an analysis of law must appeal to moral argument.

One way of distinguishing different forms of governance is in terms of the structure of the relationship between, as we might put it, "ruler" and "ruled." The idea of law imposes constraints not only on the ruled, but also on the ruler. To be sure, a legal system need not be effective in constraining the exercise of the ruler's power, and may even stipulate that the law imposes no such constraints; but in so far as a ruler exercises purely arbitrary power, he or she does not govern by law. Law thus implies a kind of reciprocity between ruler and ruled. Legal rules are, as such, general in their scope and application, knowable in advance, and susceptible of compliance. These features indicate that under law, the governed are, in some perhaps very modest and limited sense, treated as autonomous agents capable of deliberating and acting on the basis of reason. This normative relationship between ruler and ruled under law is morally preferable to alternatives, and this inherent feature of law explains why it is a predicate of commendation.

We can understand a range of important legal theories as alternative attempts to explicate the inherent potential of law to realize a morally attractive ideal of governance. Dworkin's assertion that law is a practice that aims to justify the state's exercise of coercive force could be seen as a way of explicating a morally attractive potential of law: certainly a state that recognizes an obligation to justify its coercive actions is capable of being better, morally, than a state that fails to recognize any such obligations. Raz's view of law as something that necessarily claims to mediate between persons and the reasons that apply to them contains similar resources, for it conceives of law as capable of serving each citizen's interest in meeting the demands of right reason, conceives of law as figuring in the deliberations of autonomous agents, and so on. Whatever their differences, Dworkin and Raz allow us to understand the moral attractiveness of a kind of governance that law has the inherent potential to realize. At the same time neither view rests on the claim that law must always realize this potential, and thus neither view succumbs to the pitfall of endorsing an inference from legality to moral legitimacy.

Even Hart's analysis enables us to explain the inherent potential of law to realize an attractive form of governance. For in positing, as the function of law, the guidance of conduct by rules that are reasons, Hart posits a function that can be understood, perhaps in a variety of different ways, as morally attractive. It has the capacity to treat individuals as autonomous, to mediate between persons and the reasons that apply to them, to justify the use of coercive force, and to serve a variety of welfare-enhancing ends. The moral attractiveness of law's putative guidance function is perhaps at a higher level of generality than any of these particular accounts of what is morally attractive about legal governance, but all of these accounts are suitable to reveal that law is necessarily the sort of thing with the inherent potential of realizing a morally attractive form of governance.

The argument from commendation may now be summarized as follows. Law is a predicate of weak commendation. This is because it is a part of our concept of law that it is morally attractive as such, from which it follows that every instance of law has some morally attractive property M. That property is the inherent potential of law to realize an ideal of governance. The relevant ideal can be specified in different, perhaps competing, ways, and at different levels of generality; but any analysis of the concept of law must invoke substantive moral premises in order to explain the nature of M, and to orient the analysis toward only those practices that have M. Thus all jurisprudence must be normative.

The flaw in this argument lies in the way we are to understand the idea of the inherent potential of law. Let us grant that law does have the inherent potential to realize a variety of moral ideals that other forms of governance cannot realize, and that this distinguishes law from other forms of governance. Is this inherent potential really a part of our concept of law? We should not be led astray by the metaphysical resonance of an expression like "inherent potential." There are ways of understanding that expression that do have metaphysical implications, but the initial plausibility of the foregoing argument depends on a more straightforward and metaphysically innocent sense of law's "inherent potential."

Law just is the kind of thing that can realize some attractive ideals. That fact about law is not necessarily part of our concept of it. After all, a hammer is the kind of thing that can be a murder weapon, a paperweight, or a commodity. Religion is the kind of thing that can stir murderous passions. Medicine is the kind of thing that can form the basis of a lucrative economic sector-doing so is, in that sense, an "inherent potential" of medicine. However, the fact that a thing, by its nature, has certain capacities or can be used for various ends or as a part of various projects does not entail that all or any of those capacities, ends, or projects are a part of our concept of that thing. The only point we must grant about the "inherent potential" of law to realize an attractive moral ideal of governance is the fact that law is the kind of thing with the capacity to do so. But that alone is sufficient to explain the commendation role that "law" plays in our normative discourse. Nothing follows from this about the content of our concept of law. Thus, the commendation argument errs when it assumes that a particularly interesting capacity of law is in fact a part of our concept of it. An argument is needed to show that that is the case, and none appears to be forthcoming.

Of course, if we were to give an analysis of the concept of a hammer that did not shed light on its capacity to be used as a paperweight, that would be a prima facie inadequacy of the analysis. If we were to analyze the concept of religion in a way that failed to account for the capacity of religion to stir murderous passions, we would have grounds to fault that analysis too. It does not follow that an analysis of either concept must rely on, invoke, or appeal to these capacities in identifying the central features of the concepts. By the same token, an analysis of law should help us to understand what we find morally attractive about it, and an analysis that failed to do so would be lacking. But this condition does not imply that we must appeal to moral argument in order to provide an adequate analysis of law. It is sufficient if, at the end of the day, the analysis we offer helps us to understand the morally attractive capacities of law.

In my own work, especially in The Practice of Principle, I have offered a theory of the concept of law that involves a variety of elements: primary and secondary rules, the rule of recognition, the internal point of view, and the conventionality thesis. These elements can be conjoined in a way that can explain why governance by law is preferable to alternative forms of governance. If one is moved by the moral ideals of autonomy and dignity, then one can see how the elements of my analysis constitute a thing (law) that has the capacity for accommodating those ideals in ways that other forms of governance cannot. If one is moved by the ways in which effective organization can enhance human welfare, then it is plain to see that law, understood in terms of the analysis I offer, can be conducive to those ends. But autonomy, dignity and welfare do not enter at any point into the analysis that I offer, nor do any other moral properties. These ideals are external to the concept of law; law happens to be the kind of thing that can serve them well. The capacity to do so is, in a metaphysically innocent sense, an inherent potential of law. This implies nothing about how the analysis of law must proceed, and the analysis I have offered makes no appeal to any of the values that make law attractive. The same is true of Hart's analysis as well.

In conclusion, while any theory of the concept of law would be wanting if it lacked the resources to explain why governance by law can be morally attractive, it is not true that the moral attractiveness of governance by law is part of our concept of it. The moral attractiveness of law is no basis for the claim that an analysis of the concept must proceed by way of substantive moral or political argument. The project of normative jurisprudence is driven by an important insight-that "law" is predicate of weak commendation-but it is not one that threatens a project of descriptive jurisprudence.

Endnotes

1 . Yale Law School

2. I explore several of these claims in more detail in The Practice of Principle: In Defense of a Pragmatist Approach to Legal Theory (Oxford University Press, 2001), especially Chapter 12. I develop other aspects of my methodological approach-including an important, often missed and invariably misunderstood distinction between two approaches to conceptual analysis-in "Methodology of Jurisprudence" in Coleman, Jules L. & Shapiro, Scott (eds.) & Himma, Kenneth (assoc. ed.), Oxford University Handbook of Jurisprudence and Legal Philosophy (Oxford University Press, forthcoming 2002).

3. Thus, a theorist's view about the value of governance by law is what leads her to emphasize certain features of legal practice at the expense of others. It explains why someone like Hart emphasizes law's capacity to create reasons for actions in explicating law's guidance function, rather than emphasizing, for example, law's power to sanction.

4. I am here merely characterizing Dworkin's argument, not endorsing it. In fact, I have argued strenuously in The Practice of Principle that it is mistaken. All conceptual analysis answers to epistemic, theoretical and discursive norms. This is true regardless of the concept. There is nothing special about law in this sense.

5. Coleman, Jules L., The Practice of Principle: In Defense of a Pragmatist Approach to Legal Theory (Oxford University Press, 2001).

6. This point was first made by Scott Shapiro, "On Hart's Way Out," Legal Theory 4 (1998).

7. See Coleman, The Practice of Principle, Chapter 12.

8 . Op. cit., chapters 7 and 11.

9. This argument in its details is drawn from Coleman, op cit., Chapter 12.

Previous Article | Index | Next Article

APA Newsletters

Spring 2001 Volume 00, Number 2

Spring 2001
Volume 00, Number 2