Law: Metaphysics, Meaning, and Objectivity

Law Metaphysics, Meaning, and Objectivity

Rodopi Philosophical Studies

8

Edited by

Francisco Miró Quesada (University of Lima)

Ernest Sosa (Brown University)

Amsterdam - New York, NY 2007

Law Metaphysics, Meaning, and Objectivity Social, Political, & Legal Philosophy, Volume 2

Edited by

Enrique Villanueva

INSTITUTO DE INVESTIGACIONES JURÍDICAS UNIVERSIDAD NACIONAL AUTÓNOMA DE MÉXICO

SOCIAL, POLITICAL, AND LEGAL PHILOSOPHY

Edited by Enrique Villanueva Instituto de Investigaciones Jurídicas, UNAM Editorial Advisory Board: Ruth Chang Rutgers University Jules Coleman Yale University Mark Greenberg University of California, Los Angeles Christopher Kutz University of California, Berkeley Thomas Nagel New York University Stephen Perry University of Pennsylvania Ulises Schmill Instituto Tecnológico Autónomo de México Scott Shapiro University of Michigan at Ann Arbor Jeremy Waldron Columbia University

The paper on which this book is printed meets the requirements of “ISO 9706:1994, Information and documentation - Paper for documents Requirements for permanence”. ISBN: 978-90-420-2375-8 ©Editions Rodopi B.V., Amsterdam - New York, NY 2007 Printed in the Netherlands

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii

PART I METAPHYSICS, MEANING, AND OBJECTIVITY 1. Interpretivist Theories of Law . . . . . . . . . . . . . . . . Nicos STAVROPOULOS

3

2. How Facts Make Law . . . . . . . . . . . . . . . . . . . . . Mark GREENBERG

21

3. On the Normative Significance of Brute Facts . . . . . . . . Ram NETA

75

4. On Practices and the Law . . . . . . . . . . . . . . . . . . . Mark GREENBERG

95

5. Supervenience, Value, and Legal Content . . . . . . . . . . 125 Enrique VILLANUEVA 6. Reasons Without Values? . . . . . . . . . . . . . . . . . . . 133 Mark GREENBERG 7. Theory, Practice and Ubiquitous Interpretation: The Basics . . 145 Martin STONE

vi

CONTENTS

8. Law as a Reflective Practice . . . . . . . . . . . . . . . . . 179 Scott HERSHOVITZ 9. On Reflective Practices and ‘Substituting for God’ . . . . . 189 Martin STONE 10. Metasemantics and Objectivity . . . . . . . . . . . . . . . . 215 Ori SIMCHEN 11. Can Objectivity be Grounded in Semantics? . . . . . . . . . 235 Michael S. MOORE

PART II RIGHTS THEORY 12. A Hybrid Theory of Claim-Rights . . . . . . . . . . . . . . 263 Gopal SREENIVASAN 13. Is the Will Theory of Rights Superseded by the Hybrid Theory? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 Horacio SPECTOR 14. In Defence of the Hybrid Theory . . . . . . . . . . . . . . . 299 Gopal SREENIVASAN

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

Preface Some of the main papers in this volume were presented to the I International Conference of the Instituto de Investigaciones Jurídicas, celebrated in Mexico City, July 7-11, 2003. It inaugurated a series of International Conferences to be organized by this Instituto. We express our gratitude to Dr Diego Valadés Ríos former Head of the Instituto de Investigaciones Jurídicas, UNAM, for his overall support that made possible both, first, the celebration of that Conference and, later, the publication of this volume. We also express our gratitude to Dr Héctor Fix-Fierro, Head of our Instituto, for his support in the final stages of the publication process. This is the first volume that appears in a co-edition between the Instituto de Investigaciones Jurídicas and Rodopi Publishing Company. Our gratitude to Prof Ernesto Sosa, Director of the Philosophy Series for his support. Thanks to Rodrigo Ortiz who helped with the organization of that conference. Our gratitude to the Publishing Department, Lic Raúl Márquez, Head, Jaime García Díaz and José Antonio Bautista Sánchez who made the typescript, of our Instituto de Investigaciones Juridicas. Petersburg, Russia, November 2006

PART I METAPHYSICS, MEANING, AND OBJECTIVITY

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

1

Interpretivist Theories of Law* Nicos Stavropoulos** Interpretivism about the nature of law is the view that legal rights and duties are determined by the scheme of principle that provides the best justification of certain political practices of a community: a scheme identifiable through an interpretation of the practices that is sensitive both to the facts of the practices and to the values or principles that the practices serve. Interpretivism has been developed by Ronald Dworkin in a number of publications over the last 30 years or so (see the works of Dworkin cited in the Bibliography). Interpretivism as developed by Dworkin includes the claim that interpretation is sensitive to values in the way just explained, and that it is fundamental to the nature of law. Many theorists accept that, given the law, interpretation that is sensitive to values is necessarily employed in its application (e.g. Brink 2001). In this entry, we shall be concerned exclusively with interpretivism as a theory about the nature of law, and so we shall not consider such views (except as possible misunderstandings of interpretivism). We shall focus on the explanation of the position defended by Dworkin (though not necessarily on his way of defending it), and briefly consider some alternatives in respect of the normative character of legal interpretation. The arrival of interpretivism in the scene, previously dominated by positivist and natural law theories about the nature of law, has stimu* © Nicos Stavropoulos 2003.

Reprinted from The Stanford Encyclopedia of Philosophy (Winter 2003 Edition), Edward N. Zalta (ed.), http://plato.stanford.edu/archives/win2003/entries/law-in terpretivist. ** University Lecturer in Legal Theory, University of Oxford .

4

NICOS STAVROPOULOS

lated a great deal of debate (the following are some examples from the vast secondary literature: Mitchell 1983; Cohen 1984; Hart 1994 (Postscript); Raz 1986; Finnis 1992; Coleman 2001). 1. Propositions and Grounds 2. Propositions and the Nature of Law 3. Interpretive Practices 4. Interpretation 5. Facts and Impact 6. Justifying the Facts 7. Why Political Morality? Bibliography 1. Propositions and Grounds Interpretivism is a thesis about what determines legal rights and duties, i.e. what makes it the case that the law requires what it does. As such, it is a thesis about the nature of law. These questions can be formulated in terms of the grounds of propositions of law. Propositions of law are expressed by sentences that describe legal requirements. For example, such a proposition may be that Gratuitous promises not relied upon are not enforceable in law. Such propositions can be at all levels of generality. They may involve relatively abstract legal requirements, such as the above example, or concrete requirements such as the proposition that Smith is entitled to $100 from Jones. They can be expressed in terms of legal rights or duties: e.g. ‘in law, no duty to do as promised applies to those who promise gratuitously, as long as their promise is not relied upon’, or ‘Jones has a legal duty to pay Smith $100’. What makes a proposition of law true, when it is true? We want to investigate the grounds of law, so it would not help to cite facts that trivially make legal propositions true — we already know that, for any x, the fact that x is capable of making the proposition that x true — and it would not help to cite any legal facts either — doing so would merely postpone, not answer our question. What are the likely candidates?

INTERPRETIVIST THEORIES OF LAW

5

It seems uncontroversial that biographical and other contingencies in the lives of ordinary citizens, such as the fact that Jones was drunk when he met Smith to discuss a deal, are among the facts that make true concrete propositions of law that involve such citizens’ rights and duties. More interestingly, it is uncontroversial that the truth of all legal propositions is, at least in part, dependent upon certain political facts. By ‘political facts’ I mean political decisions and practices, including legislative enactments, custom, or settled judicial practice, understood in non-legal terms, i.e. without reference to what the law requires; and further understood in non-evaluative terms, i.e. without reference to facts about what is right or good. It is uncontroversial that the question whether gratuitous promises bind in law is partly dependent upon political decisions and practices. It is further widely accepted that such decisions and practices crucially include what officials said and did in the past, conceived simply in terms of the texts and utterances they issued and their settled practice of behaving in a certain way, all of which at least in part make propositions of law true. For example, it is widely accepted that the text of a statute that reads ‘gratuitous promises shall not be enforced’, or the fact that judges, as a matter of settled practice, are disposed not to enforce gratuitous promises, would go a long way towards explaining why such promises are not enforceable. Political facts properly conceived may include more than these simple facts. Official utterances and texts are meaningful entities, not mere sounds or ink stains (which could not as such make any legal proposition true) so any proper conception of political facts must rely on some doctrine about what makes texts and utterances mean what they do. But a conception of political facts may be much richer still: for example, it may include counterfactual dispositions of officials, or psychological states of officials such as their intentions, desires, hopes, and expectations not specified in terms of dispositions. Did members of Congress intend or expect that the text of the statute should be construed in a certain way? Would they have formulated the text differently if they had contemplated a contingency they did not in fact contemplate? An appropriate conception of political facts may further include other, more complex facts such as the fact that certain courts possess the power to quash other courts’ decisions. It is more

6

NICOS STAVROPOULOS

controversial whether such complex facts are ultimately reducible to simple facts of linguistic and other behavior of officials or citizens, including their collective dispositions, or whether they are indeed relevant at all. It is however uncontroversial that political facts, conceived in non-evaluative terms, play a critical role in the grounding of legal propositions. For convenience, I shall call political facts so conceived political practice. Interpretivism accepts and provides an explanation for these widely accepted claims, but also defends a view usually attributed to the natural law tradition: that, in spite of their obvious dependence upon political practice, legal requirements further depend, in some complex way, upon certain non-contingent evaluative facts. This part of interpretivism is highly controversial, so we need to clarify and defend it. Interpretivism is the view that, if true, a proposition of law is true in virtue of an interpretive fact: in a nutshell, in virtue of the fact that the proposition follows from the best justification of a community’s political practice. Interpretivism says that it is not political practice alone — such as the fact that Congress voted for a statute that says that gratuitous promises are not enforceable, that its power to enforce its decisions is recognized, or that it is treated as supreme law-maker by lawyers, or the fact that as a matter of settled practice courts do not enforce such promises — that makes legal propositions true, nor is it moral or other evaluative facts alone — e.g. that it is good that gratuitous promises should not be enforceable — that make them true either. Rather, legal propositions are made true by the best justification of political practice, i.e. the scheme of principle that underlies and governs political practice, including e.g. the fact that it is right that Congress should wield the power that it does or that courts’ settled practice should have a large impact on our duties. Interpretivism therefore claims that legal requirements are sensitive to both the facts of the practice and the values served by it, but not fully determined by either. 2. Propositions and the Nature of Law The question of the grounds of legal propositions is not a question about the legal requirements in any jurisdiction. It is not the question of

INTERPRETIVIST THEORIES OF LAW

7

what is the content of the law, here or elsewhere. Rather, it is the question of what makes the content of the law what it is. If the law, for example, includes a requirement to pay capital gains tax at a certain rate, what makes it the case that it includes that requirement? Consider by way of analogy the proposition that a certain course of action, A, is immoral. If this proposition is true, what makes it true? We learn something important about the nature of morality if we learn whether, e.g., this proposition is true in virtue of the fact that A is disapproved by the Gods, or that A is considered immoral by most people. Similarly, in the legal domain we may hope to learn about the nature of law by learning whether legal propositions are made true by contingent political facts — and if so, which ones — or by political values, economic forces, or divine will, or have any other colorable ground. The core at least of any doctrine about the nature of law including all the familiar ones may be formulated as a thesis about legal propositions and their grounds. Put in those terms, some of these doctrines may imply that legal propositions divide into distinct classes, on the basis of the kind of grounds each has. For example, a theory may hold that certain propositions are either made true by a special class of facts of political practice, law-creating facts, alone, or are entailed by propositions that are made true by such facts alone. These propositions constitute a basic class, the law. By contrast, propositions that follow from the basic class with the addition of further contingent or evaluative premises constitute the law’s applications (compare Raz 1980, 70-92; 216-24). Moreover, each theory provides or entails an account not merely of the grounds and the propositions they make true, but also of the nature of the relation between grounds and propositions. For example, a theory may hold that the links between true legal propositions and political practice are conceptual, in the special sense of the shared implicit understanding of the concept of law or other key concepts such as the concept of authority. That is, the theory may say that it is such a conceptual truth that certain aspects of the practices determine the propositions that they do (see Greenberg 2004 for systematic discussion of possible determination relations). Interpretivism may be consistent with a classification of legal propositions by kind of grounds but, since it would not accept that any le-

8

NICOS STAVROPOULOS

gal proposition is made true by political practice (in the narrow technical sense identified above) alone, it need not place any emphasis on such a division. Interpretivism would further say that the relation that holds between the grounds and the propositions is interpretive, in a sense to be explained further. The question of the grounds of legal propositions should not be confused with another question often discussed by legal philosophers: not what makes legal propositions true but rather what makes norms (or legal rules or laws) valid. Some theorists hold that the conditions of validity of norms are exclusively source-based, consisting only in facts about the norms’ pedigree. Other theorists hold that norms’ validity may further depend on certain moral conditions, if such conditions are laid down by more fundamental source-based norms, or by settled judicial practice, or by some other feature of political history. For example, such a view may be that in case the Constitution entrenches some values, the conditions of validity of ordinary legislation include those values. Or it may be that, in case officials happen to use moral tests, e.g. they treat as law whatever Congress says yet only insofar as it is consistent with liberty and equality, these values set partial tests of validity. In both cases the values play the crucial role only because legislators themselves said so or because officials treat them as such — in virtue of political practice. Interpretivism is often taken to represent another possibility along that spectrum: it is recast as the thesis that legal norms’ validity is partly dependent upon the best justification of political practice, in the specific sense that valid norms are those which, in addition to having the right source — their having been promulgated by the right people or institutions — are further consistent with the principles that justify the relevant political practice (even if the principles are not explicitly laid down by any authority). On that reading, interpretation consists in the derivation of valid legal norms through the application of an evaluative filter to a set of given norms. As the metaphor of a filter makes plain, the reading in question presupposes a set of norms (or rules or laws), whose identity does not depend on interpretation, to which the filter is applied at an analytically later stage. The interpretivist, however, need not accept that any legal norms or rules are individuated non-interpretively.

INTERPRETIVIST THEORIES OF LAW

9

On another version, interpretivism does not take any norms as (non-interpretively) given. It rather claims that valid legal norms are identifiable through an interpretation of the legal materials such as authoritative texts — the texts of the constitution, statutes and cases, which canonically formulate the legal decisions — in the light of certain political values. But even on that version, the materials’ legality is given. A broadly positivistic explanation of law in terms of authoritative sources is once again assumed, with interpretation understood as some kind of evaluative processing of the sources. The confusion may be encouraged by the assumption that the interpretivist must rely on or presuppose a division of legal propositions into a special class of propositions grounded on law-creating facts alone, which constitute the law and in respect of which interpretation plays no role, and those that follow from the special class with the addition of further premises, which constitute the application of law. On this assumption, the interpretivist is merely claiming that law must be interpreted to be applied. As we saw however, interpretivism is not restricted in this way. 3. Interpretive Practices The interpretivist claims that legal practice is, in its nature, interpretive. What makes a practice interpretive? Dworkin explains this in terms of a special attitude, which has two components, holding among practitioners. He uses an imaginary example of a practice of courtesy, which includes many rules cited by practitioners and taken by them to set out what courtesy requires. He describes the components of the attitude (Dworkin 1986, 47) as follows: The first is the assumption that the practice of courtesy does not merely exist but has value, that it serves some interest or purpose or enforces some principle — in short, that it has some point — that can be stated independently of just describing the rules that make up the practice. The second is the further assumption that the requirements of courtesy — the behavior it calls for or the judgments it warrants — are not necessarily or exclusively what they have always been taken to be but are instead sensitive to its point, so that the strict rules must be understood or applied or modified or qualified or limited by that point.

10

NICOS STAVROPOULOS

Dworkin’s two components capture two independent conditions, both of which must be satisfied. It is not enough that the practice be thought to serve some value (which would satisfy the first condition); further, the value must be taken to be constitutive of the practice, which is what the second condition amounts to. Together, the conditions have important consequences in respect of the practice’s character. It should be emphasized that the value or principle in question is assumed to be genuine and not trivially satisfied by the practice. The assumption is that the practice is in fact valuable, not that it can be rationalized. Baby-torturers may think that their practice serves some value, and even that its requirements are sensitive to it, but they are wrong: the value in question cannot be the value of torturing babies, a value their practice would trivially satisfy if it existed. Nor can it be the value of inflicting suffering, which would not be trivial but likewise does not exist. In this case, we can confidently assume that no real value or principle is served by what they do. This case is however only a dramatic extreme: the fact that practitioners assume that their practice serves a value does not suffice to make it so, so it could turn out for any practice whatever — including legal practice — that the assumption is false. Practitioners hold themselves responsible to a standard that is set by the value, if any, which in fact justifies the practice, not their individual or collective understanding of the value. This aspect of the attitude introduces a critical dimension to interpretive practices. It implies that the standard that governs such a practice is taken to be in a crucial sense external to it. Participants accept that they are fallible, even collectively, in respect of the standard and thus in respect of the behavior or judgment that the standard entails. Any particular aspect of the practice, however well-settled, may turn out to be a mistake, and the practice could have determinate requirements even where it is unsettled. Consider by way of contrast a conventional practice, as commonly understood: in that case, the standard that governs a practice is defined — exclusively set — by settled practice (compare Hart 1982, 153-61). It follows that, in that case, no part of settled practice could turn out to be a mistake, a failure by practitioners to discharge the duty imposed by the practice. For the interpretivist, however, there is no such guarantee. The question can always be raised in respect of any

INTERPRETIVIST THEORIES OF LAW

11

aspect of the practice: does it really serve the value the practice is meant to serve? But while the standard is in that sense partly external, it is not in another. Participants are open-minded in respect of what makes their practice as it stands now require what it does. When they are inclined to correct their actual practice to conform to a new, better interpretation, they take it that that the new interpretation better articulates the standard to which they were antecedently committed, not that they are persuaded to substitute a new, more attractive standard for the old one (compare Burge 1986). Room for error in the practice entails that disagreement of a special kind is possible. Participants can sensibly disagree about what constitutes courtesy —what the standard of courtesy is, not merely whether some behavior satisfies a given standard— and take it that the right answer to that constitutive question would be right in virtue of capturing the rationale of the practice, whatever it is, not their understanding of the rationale (whether explicit or implicit, individual or collective). This is not to say that it follows from the mere fact that participants appear to disagree in that fundamental way that they do. Rather, the point is that, if participants do hold themselves responsible to the value, whatever it is, that justifies their practice, it is possible for such disagreement to obtain. Whether any actual disagreement has the character in question depends on whether it is in fact best explained in those terms. 4. Interpretation The interpreter’s task in respect of an interpretive practice is to formulate a hypothesis, i.e. to postulate a value that could in fact justify the practice’s requiring what it is held to require. The interpreter’s hypothesis amounts to the attribution of a rationale or point to the relevant practice, an articulation of the values that are served by behaving in the way participants do behave when they are trying to conform to the requirements of the practice in question. It is worth repeating that the practice’s interpretive character merely entails that the interpreter must look for such a rationale, not that there is some guarantee that one exists.

12

NICOS STAVROPOULOS

For example, the rationale of giving up one’s seat to a senior in the name of courtesy (i.e. as a duty entailed by a standard that requires many other things too, such as excusing oneself from certain parties) may be to show respect (assuming that this constitutes a virtue or has some other value); if so, the display of respect would constitute the point of the practice, so that, if faced with the question what the practice requires that one do in a different context, the right answer would be sensitive to that point. For another example, which greatly simplifies a far more complicated practice, the rationale of awarding compensation for injuries caused by defective products may be that manufacturers should bear the cost of the risk associated with the use of the relevant products. Such a rationale would justify awarding damages when such injuries occur, whether or not the manufacturer is at fault — i.e. would justify no-fault liability. Finally, consider the practice of treating statutes as capable of overriding many long-standing arrangements of rights and duties, as having the power instantly to invent new or to extinguish old such arrangements. It would be a major undertaking fully or even adequately to specify the values served by treating the legislature as possessing such sweeping powers; plausibly, they are abstract political values that involve procedural justice. These values, once articulated, would partly determine the impact of a statute that purports to override some fundamental right, with which the legislature had never interfered before. These examples are meant merely to convey the flavor of the interpreter’s task. Given the complexity of legal practice, the legal interpreter’s complete task would be to construct an equally complex hypothesis that can explain and justify, so far as possible, all the ‘data’, the facts of a large number of specific practices, many of which may seem to pull in different directions. The interpretivist then may introduce a superhuman character, the ideal interpreter, who should be able to find a complex hypothesis that achieves equilibrium among all the relevant practices (Dworkin famously calls him ‘Hercules’). Although sometimes presented as a flaw in the theory, the superhuman character merely helps illustrate an aspect of legal practice which does not seem controversial: that any candidate interpretation of the law, however narrowly focused, is open to the challenge that it does not cohere with another part of the law, where both the interpreter and the

INTERPRETIVIST THEORIES OF LAW

13

challenger agree what that part requires. The possibility of this form of challenge implies the formal aspect of interpretation discussed here: that any local interpretation must aspire to be a genuine component of the global, all-encompassing and perfectly calibrated interpretation of the totality of legal practice. The fact that no lawyer or judge can put together such an interpretation is no more disturbing than the fact that no scientist can put together a total explanation of the natural world. Like the total scientific theory, the global interpretation provides the measure of correctness and thus the standard of criticism of ordinary science and ordinary interpretation. 5. Facts and Impact The interpretivist’s central claim is that legal practice is indeed interpretive, in the sense explained above. He claims that it is possible for lawyers to disagree in the special way (and is likely further to claim that they often do disagree in that way): e.g. about what constitutes negligence or legislative intent, each assuming that his or her explanation of the law on the disputed point is the best in virtue of being the only explanation that properly justifies the rest of the practice (the part not in dispute). Most lawyers would agree of course on some very weak proposition about the grounds of law. They would agree that certain aspects of practice — e.g. statutory texts — have large impact on the law, but they only agree because they each rely on a far more elaborate understanding of the precise impact that any aspect of the practice does have. Their thin agreement is only an overlap of very different complete accounts. That is not to say that each lawyer has a well-articulated Herculean account in mind when he engages in a local dispute, only that he assumes that his partial account is good in virtue of being a part of the Herculean account, and hence assumes a responsibility to articulate as best he can the implications or presuppositions of the part he defends, in any direction. Even if lawyers may occasionally have to dirty their hands in the detail, perhaps philosophers could leave the matter at the area of overlap, and be content with the proposition that some role or other must be assigned to political facts. But which facts, more precisely, are we

14

NICOS STAVROPOULOS

talking about? And what makes those facts so central to legal propositions? We can dramatize this question: what makes facts about what members of Congress said and did, though not facts about what each member’s best friend said and did, pertinent to law? And which facts about members of Congress are crucial? Is it exclusively facts about members voting for bills, or perhaps facts about Congressional debates as well? And how about facts about committee activities? And do we only care for the explicit utterances and formal deeds of members of Congress, or are we also interested in any psychological facts about these people (Dworkin 1985b; 1986 Ch. 9)? We can ask such questions in respect of any sub-class of political facts: we can reformulate our questions to focus on the Framers and their friends, the Framers’ speeches and their correspondence, and the political pamphlets some of them wrote; and so on for judges and other political agents. Interpretivism claims that values justify the precise kind of authority that different agents and institutions have and hence the precise way in which their decisions and other acts determine legal duties (including any specific cases in which it may turn out that that the values may dictate that the facts about authoritative decision-making alone determine our duties). Although the interpretivist accepts that propositions of law are founded on political decisions, i.e. that the content of the law is sensitive to what a political community historically decides; and he also accepts that, more abstractly, law is dependent upon certain political practices of a community, including legislation and adjudication; he further says that what makes these decisions and practices the grounds of propositions of law is the normative fact that it is good that they should be such grounds. The interpretivist says that it is in virtue of the fact that our rights and duties ought to flow from past political decisions that they flow from them; or that, more abstractly, legal duties ought to be determined by social practices that they do. And he thinks that they ought to be determined by such decisions and practices because doing so serves certain political virtues such as fairness and procedural justice. The interpretivist says that a fuller articulation of these virtues and of their connections with other political values would yield a more detailed account of the precise way in which the practices determine legal duties.

INTERPRETIVIST THEORIES OF LAW

15

The full account would include the answer to the more specific kind of question that was illustrated above, and so would tell us whether our duties depend on the explicit decisions, only, of Congress and other bodies, or further on the psychological states of the decisionmakers, the decision-makers’ own understanding of the decisions, or any other element of political history and structure. For the interpretivist, the legality of the aspects of political practice that do in fact serve as the grounds of law is settled by reference to what is right or good. It would be a mistake however to think that the only question is which facts count as legal. We do not merely need to work out questions such as why someone ought to make law and who gets to make it (and possibly also what is the broad or rough way in which the words and deeds or other behavior of the relevant persons determine the law’s requirements). The principles of political morality that figure in the best interpretation of political practices do not only dictate that certain facts about the promulgation of statutes and the resolution of cases, or some other aspect of a community’s political practices, are the distinctively legal facts on which rights and duties depend, without helping precisely to determine what impact the facts have on our rights and duties. The question to which the values provide an answer can be expressed more simply (see Greenberg 2004). It is this: in respect of all the facts that could conceivably play a role in determining legal rights and duties, what is the role that each has? This is not simply a question of relevance or even rough weight. Rather, the question invites an answer in terms of the precise impact that a fact has on our rights and duties, i.e. an answer which includes a specification of the impact of the relevant fact, in case different relevant facts pull in different directions. The interpretivist then says that values determine the precise impact that all candidate facts have on legal rights and duties. Legality and impact are not distinct. 6. Justifying the Facts We have just seen that the interpretivist claims that it is in virtue of certain values that the law has the determinants that it does — i.e. that

16

NICOS STAVROPOULOS

its content depends on a community’s political practices. He claims that values do not merely make the political facts relevant but also select the way in which the facts determine legal rights and duties. Simplifying the thesis in that way helps us better to understand Dworkin’s famous claim that interpreting the practices consists in assigning to them the content that best fits and justifies them (Dworkin 1986). While Dworkin’s distinction between fit and justification, as tests of interpretive success, is explicitly a mere heuristic device (Dworkin 1982, 1986 at 229-63), it has received intense critical scrutiny (Raz 1986; Finnis 1992; Hart 1994, Postscript). The distinction seems to suggest that the interpreter has two separate tasks, first to formulate alternative hypotheses that are consistent with the facts of the practice, and second show which one among them provides the most attractive justification of the facts of the practice. The dimension of fit may then be understood as an independent, non-normative constraint, in the sense either that this is how the interpretivist understands it, or that this how he cannot but understand it. In fact the distinction in question is no more than a rough distinction between ineligible and eligible interpretations, which is made within particular interpretations, not prior to any interpretation. It is merely a device for organizing complex interpretations and so for facilitating the representation of alternative interpretations involved in notional or actual disputes. An interpretation will be typically organized in terms of the examples from the practice that it undertakes to justify, and the justification that it offers. But no interpretation need accept that any particular example’s standing as such is immune from interpretive assessment. The interpretation must itself (normatively) explain what makes a putative example a genuine example, if it is indeed such. That implies that no example is immune from discount as a mistake, as a piece of fool’s gold lurking among the genuine samples. The interpretivist does not accept, nor does the fit-justification distinction properly understood entail, a distinction between the data and their reading that is prior to and independent of interpretation. As we have already noticed, common ground among interpretations is mere overlap. No aspect of the practice can be taken as a non-interpretive given. Rather, they are both parts of the interpretive, thoroughly normative account.

INTERPRETIVIST THEORIES OF LAW

17

There is yet another misunderstanding that we must guard against. We concluded that interpretivism claims that neither question — what are the data and what justifies them — can be answered in a non-normative manner. But even if it is conceded that interpretation has no non-normative part, it may still seem that there are two distinct normative questions to be asked, namely what is the justification for taking certain data to be generally decisive in determining our duties, and what is the justification for taking the data to determine them in a particular way, i.e. in a way such that certain duties and not others are identified as legally imposed. And if we think that there are two distinct questions — even if it is granted that they are both normative — it may seem that there could be tensions between them, so that a further question remains which concerns how to balance the two components of the normative account when they conflict. We may dissolve this further misunderstanding by recalling our formulation of the interpretivist point in terms of a single question. Interpretivism says that values select, from among all the facts that could conceivably be determinants of legal duty, the ones that are such determinants, by providing an account that justifies the impact that each precise determinant has on our rights and duties. 7. Why Political Morality? The fact that the interpretivist seeks to identify the foundation of duties that, on their face, seem political — duties owed to political decisions and practice — suggests why that foundation, if it exists, is likely to come from political morality, the domain of duties owed to each other or to institutions or to the community as a whole in virtue of some feature of political life that is capable of providing such a foundation. But the interpretivist commitment is weaker than that. It is rather that, whatever, if anything, is capable of justifying doing as the law requires because it requires it, is an eligible candidate. The interpretivist does not deny that the quest for such a foundation may fail. If so, no value that can play the role of foundation may be attributable to the practice. What the interpretivist must deny is that the attribution of a value or point that is too weak to justify genuine

18

NICOS STAVROPOULOS

duties should always be preferred, even if the attribution of a point that is not weak in that way is available and compelling. For the claim then becomes non-interpretivist, since it assumes that a test other than (and prior to) interpretive success is fundamental. Bibliography Brink, David (2001): ‘Legal Interpretation and Morality’, in B. Leiter (ed.) Objectivity in Law and Morals, Cambridge: Cambridge University Press. Burge, Tyler (1986): ‘Intellectual Norms and Foundations of Mind’, Journal of Philosophy 83, 697-720. Cohen, Marshall (ed.) (1984): Ronald Dworkin and Contemporary Jurisprudence, London: Duckworth. Coleman, Jules (ed.) (2001): Hart’s Postscript, Oxford: Oxford University Press. Dworkin, Ronald (1977): Taking Rights Seriously, London: Duckworth. (1977a): Ch. 2 (The Model of Rules I) (1977b): Ch. 3 (The Model of Rules II) (1977c): Ch. 4 (Hard Cases) (1977d): Appendix (Reply to Critics) —— (1982): ‘Natural Law Revisited’, U. Florida. Law Review, 34: 165. —— (1983): ‘My Reply to Stanley Fish (and Walter Benn Michaels): Please Don’t Talk about Objectivity any More’, in W. J. T. Mitchell (ed.), The Politics of Interpretation, Chicago: University of Chicago Press. —— (1984): ‘A Reply’, in M. Cohen (ed.), Ronald Dworkin and Contemporary Jurisprudence, London: Duckworth. —— (1985): A Matter of Principle, Cambridge MA: Harvard University Press. (1985a): Ch 1 (Political Judges and the Rule of Law) (1985b): Ch 2 (The Forum of Principle) (1985c): Ch 3 (Principle, Policy, Procedure) (1985d): Ch 5 (Is there Really No Right Answer in Hard Cases?) (1985e): Ch 6 (How Law is Like Literature) (1985f): Ch 7 (On Interpretation and Objectivity) —— (1986): Law’s Empire, Cambridge MA: Harvard University Press. —— (1996): ‘Objectivity and Truth: You’d Better Believe It’, Philosophy and Public Affairs, 25: 88. Finnis, John (1992): ‘Natural Law and Legal Reasoning’, in George, R. (ed.), Natural Law Theory, Oxford: Clarendon Press. George, Robert (ed.) 1992: Natural Law Theory, Oxford: Clarendon Press. Greenberg, Mark (2004): ‘How Facts Make Law’, forthcoming, Legal Theory. Hart, H.L.A. (1982): Essays on Bentham, Oxford: Clarendon Press. —— (1994): The Concept of Law, 2nd edition, Oxford: Clarendon Press.

INTERPRETIVIST THEORIES OF LAW

19

Mitchell, W.J.T. (ed.) (1983): The Politics of Interpretation, Chicago: University of Chicago Press. Raz, Joseph (1980): The Concept of a Legal System, 2nd edition, Oxford: Clarendon Press. —— (1984): ‘Legal Principles and the Limits of Law’, in Cohen, M. (ed.), Ronald Dworkin and Contemporary Jurisprudence, London: Duckworth. —— (1986): ‘Dworkin: a New Link in the Chain’, CAL. L. R. 74: 1103. —— (1994): Ethics in the Public Domain, Oxford: Clarendon Press. —— (2001): ‘Two Views about the Nature of Law’, in Coleman, J. ed., Hart’s Postscript, Oxford: Oxford University Press.

2

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

How Facts Make Law* Mark Greenberg** I. Introduction Nearly all philosophers of law agree that nonnormative, nonevaluative, contingent facts—descriptive facts, for short—are among the determinants of the content of the law. In particular, ordinary empirical facts about the behavior and mental states of people such as legislators, judges, other government officials, and voters play a part in determining that content. It is highly controversial, however, whether the relevant descriptive facts, which we can call law-determining practices, or law practices (or simply practices) for short,1 are the only determinants of legal content, or whether legal content also depends on normative or evaluative facts—value facts,2 for short. In fact, a central—perhaps the central—debate in the philosophy of law is a debate over whether value facts are among the deter* For helpful comments on ancient and recent predecessors of this paper, I am very grateful to Larry Alexander, Andrea Ashworth, Ruth Chang, Jules Coleman, Martin Davies, Ronald Dworkin, Gil Harman, Scott Hershovitz, Kinch Hoekstra, Harry Litman, Tim Macht, Tom Nagel, Ram Neta, Jim Pryor, Stephen Perry, Joseph Raz, Gideon Rosen, Scott Shapiro, Seana Shiffrin, Ori Simchen, Martin Stone, Enrique Villanueva, and two anonymous referees for Legal Theory. Special thanks to Susan Hurley and Nicos Stavropoulos for many valuable discussions. I would also like to thank audiences at the University of Pennsylvania, New York University, UCLA, Yale University, the 2002 Annual Analytic Legal Philosophy conference, and the 2003 International Congress in Mexico City, where versions of this material were presented. Finally, I owe a great debt to the work of Ronald Dworkin. ** University of California, Los Angeles. 1 For the moment, I will be vague about the nature of law practices. For more precision, see Section II.B below. 2 For some explanation of what I mean by “value facts,” see note 22 below.

22

MARK GREENBERG

minants of the content of the law (though the debate is not usually characterized in this way). A central claim of legal positivism is that the content of the law depends only on social facts, understood as a proper subset of descriptive facts. As Joseph Raz says, “H.L.A. Hart is heir and torch-bearer of a great tradition in the philosophy of law which . . . regards the existence and content of the law as a matter of social fact whose connection with moral or any other values is contingent and precarious.”3 In contemporary philosophy of law, there are two distinct ways of developing this tradition: hard and soft positivism. Hard positivism denies that value facts may play any role in determining legal content.4 Soft positivism allows that the relevant social facts may make value facts relevant in a secondary way. For example, the fact that a legislature uses a moral term—“equality,” say—in a statute may have the effect of incorporating moral facts—about equality, in this case—into the law.5 On this soft positivist view, however, it is still the social facts that make the value facts relevant, and the social facts need not incorporate value facts into the law. Hence according to both hard and soft positivism, it is possible for social facts alone to determine what the law is, and even when they make value facts relevant, social facts do the fundamental work in making the law what it is—work that is explanatorily prior to the role of value facts. To put things metaphorically, hard positivism and soft positivism hold that there could still be law if God destroyed all value facts. Ronald Dworkin is the foremost contemporary advocate of an antipositivist position. According to Dworkin, a legal proposition is true in a given legal system if it is entailed by the set of principles that best justify the practices of the legal system.6 Since the notion of justification on which Dworkin relies is a normative notion, a consequence of Dworkin’s view is that the content of the law depends on value facts. 3 JOSEPH RAZ, ETHICS IN THE PUBLIC DOMAIN 210 (1994). Raz also puts the point epistemically: the content of the law “can be identified by reference to social facts alone, without resort to any evaluative argument.” Id. at 211. 4 See, e.g., RAZ, supra note 3, at ch. 10; JOSEPH RAZ, THE AUTHORITY OF LAW ch. 3 (1979). 5 See, e.g., Jules Coleman, Negative and Positive Positivism, 11 J. LEGAL STUD. 139 (1982); H.L.A. Hart, Postscript, in THE CONCEPT OF LAW (2nd ed., 1997). 6 See RONALD DWORKIN, LAW’S EMPIRE (1986).

HOW FACTS MAKE LAW

23

Understanding and resolving the debate between positivists and antipositivists requires understanding the nature of the relevant determination relation—the relation between determinants of legal content and legal content. The debate, as noted, concerns whether law practices are the sole determinants of legal content. It is difficult to see how one can systematically address the question of whether A facts are the sole determinants of B facts without understanding what kind of determination is at stake. But the positivist/antipositivist debate has so far been conducted with almost no attention to this crucial issue. A preliminary point is that the determination relation with which we are concerned is primarily a metaphysical, or constitutive, one, and only secondarily an epistemic one: the law-determining practices make the content of the law what it is. To put it another way, facts about the content of the law (“legal-content facts”) obtain in virtue of the law-determining practices. It is only because of this underlying metaphysical relation that we ascertain what the law is by consulting those practices. A second preliminary point, which should be uncontroversial, is that no legal-content facts are metaphysically basic or ultimate facts about the universe, facts for which there is nothing to say about what makes them the case. Legal-content facts, like facts about the meaning of words or facts about international exchange rates (e.g., that, at a particular time, a U.K. pound is worth 1.45 U.S. dollars), hold in virtue of more basic facts. The important implication for present purposes is that the full story of how the determinants of legal content make the law what it is cannot take any legal content as given. It will not be adequate, for example, to hold that law practices plus some very basic legal-content facts (for example, legal propositions concerning the relevance of law practices to the content of the law) together make the law what it is, for such an account fails to explain what it is in virtue of which the very basic legal-content facts obtain. Descriptive facts about what people said and did (and thought) in the past are among the more basic facts that determine the content of the law. I claim that the content of the law depends not just on descriptive facts but on value facts as well. Given the plausible assump-

24

MARK GREENBERG

tion that fundamental7 value facts are necessary rather than contingent, there is, however, a difficulty about expressing my claim in terms of counterfactual theses or theses about metaphysical determination. Even if the value facts are relevant to the content of the law, it is still true that the content of the law could not be different from what it is without the descriptive facts being different (since it is impossible for the value facts, being necessary, to be different from what they are). Necessary truths cannot be a nonredundant element of a supervenience base. Hence both positivists and antipositivists can agree that descriptive facts alone metaphysically determine the content of the law.8 In order to express the sense in which the content of the law is claimed to depend on value facts, we therefore need to employ a notion different from and richer than metaphysical determination. We can say that a full metaphysical (or constitutive) account of the content of the law (of why certain legal propositions are true) must appeal to value facts. I earlier put the point metaphorically by saying that if God destroyed the value facts, the law would have no content. The epistemic corollary is that working out what the law is will require reasoning about value. 7 The point of the qualification “fundamental” is to distinguish basic or pure value facts—that, say, harm is a relevant moral consideration—from applied or mixed value facts—that returning the gun to John tomorrow would be wrong. The fundamental value facts are plausibly metaphysically necessary, while the applied value facts obviously depend on contingent descriptive facts as well as on fundamental value facts. This qualification does not affect the point in the text, since the contingent facts are encompassed in the supervenience base of descriptive facts. That is, if the fundamental value facts supervene on the descriptive facts, the applied value facts will do so as well. 8 The term “metaphysical determination” is typically used to specify a form of modal determination. In this sense, that the A facts metaphysically determine the B facts does not imply that the B facts obtain in virtue of the obtaining of the A facts, or even that the B facts are more ontologically basic than the A facts. Positivists and antipositivists can agree not only that descriptive facts alone modally determine the content of the law but also that the obtaining of the relevant descriptive facts is part of the explanation of the obtaining of legal-content facts. In this paper, we will be concerned only with cases in which the putative determinants are more basic than and part of the explanation of the determined facts. For convenience, I will therefore say that the A facts metaphysically determine the B facts only when the B facts obtain at least in part in virtue of the obtaining of the A facts.

HOW FACTS MAKE LAW

25

As we will see, a full account of what it is in virtue of which legal-content facts obtain has to do more than describe the more basic facts that are the metaphysical determinants of legal content. The relevant determination relation is not bare metaphysical determination. (As we have just seen, if that were the relevant relation, there would be no debate between the positivists and the antipositivists. Positivists would win the debate trivially, since the descriptive facts alone fix the content of the law.) I argue for a particular understanding of the metaphysical relation (between the determinants and the legal content that they determine), which I call rational determination. Rational determination, in contrast to bare metaphysical determination, is necessarily reason-based (in a sense that I elaborate in Section II.B). A quick way to grasp the basic idea is to consider the case of aesthetic facts. Descriptive facts metaphysically determine aesthetic facts. A painting is elegant in virtue of facts about the distribution of color over the surface (and the like). But arguably there need not be reasons that explain why the relevant descriptive facts make the painting elegant. We may be able to discover which descriptive facts make paintings elegant (and even the underlying psychological mechanisms), but even if we do, those facts need not provide substantive aesthetic reasons why the painting is elegant (as opposed to causal explanations of our reactions). On this view, it may just be a brute fact that a certain configuration of paint on a surface constitutes or realizes a painting with certain aesthetic properties. (As noted below, facts about humor provide an even clearer example.) In contrast, if it is not in principle intelligible why the determinants of legal content—the relevant descriptive facts—make the law have certain content, then it does not have that content. Rational determination is an interesting and unusual metaphysical relation because it involves the notion of a reason, which may well be best understood as an epistemic notion. If so, we have an epistemic notion playing a role in a metaphysical relation. (Donald Davidson’s view of the relation between the determinants of mental content and mental content is plausibly another example of this general phenomenon.)9 For this reason, I believe that the rational-determination relation is of independent philosophical interest. 9

See notes 17 and 18.

26

MARK GREENBERG

My main goal in this paper, however, is to show that, given the nature of the relevant kind of determination, law practices—understood as descriptive facts about what people have said and done—cannot themselves determine the content of the law. Value facts are needed to determine the legal relevance of different aspects of law practices. I therefore defend an antipositivist position, one that is roughly in the neighborhood of Dworkin’s, on the basis of very general philosophical considerations unlike those on which Dworkin himself relies. 10 We have two domains of facts: a higher-level legal domain and a lower-level descriptive domain. It is, I claim, a general truth that a domain of descriptive facts can rationally determine facts in a dependent, higher-level domain only in combination with truths about which aspects of the descriptive, lower-level facts are relevant to the higher-level domain and what their relevance is. Without the standards provided by such truths, it is indeterminate which candidate facts in the higher-level domain are most supported by the lower-level facts. There is a further question about the source or nature of the needed truths (about the relevance of the descriptive facts to the higher-level domain). In the legal case, these truths are, I will suggest, truths about value. The basic argument is general enough to apply to any realm in which a body of descriptive facts is supposed to make it the case by rational determination that facts in a certain domain obtain. For example, if the relation between social practices, understood purely descriptively, and social rules is rational determination, the argument implies that social practices cannot themselves determine the content of social rules. (At that point, we reach the further question of the source of the truths needed in the case of social rules; the answer may differ from that in the legal case.) Hence the argument is of interest well beyond the philosophy of law. In this paper, I will largely confine the discussion to the legal case. 10 Dworkin’s theory of law depends on a view about the nature of “creative interpretation.” In particular, he argues that to interpret a work of art or a social practice is to try to display it as the best that it can be of its kind. See DWORKIN, supra note 6, at 49–65. Dworkin’s central argument for the position that legal interpretation is an instance of this general kind of interpretation is that this position is the best explanation of “theoretical disagreement” in law. Id. at 45–96; see also Dworkin, Law as Interpretation, in THE POLITICS OF INTERPRETATION (W.J.T. Mitchell ed., 1983).

HOW FACTS MAKE LAW

27

In Section II, I clarify the premises of the argument and explain that they should not be controversial. In Section III, I examine why there is a problem of how legal content is determined. The content of the law is not simply the meanings of the words (and the contents of the mental states) that are uttered in the course of law practices. Something must determine which elements of law practices are relevant and how they combine to determine the content of the law. Next, in Section IV, I argue that law practices themselves cannot determine how they contribute to the content of the law. In Section V, I consider and respond to three related objections. Finally, in Section VI, I examine what the argument has established about the relation between law and value. 11 II. The Premises In this section, I set out the two premises of the argument and make a number of clarifications. The second premise will require a great deal more discussion than the first. I take both premises to be relatively uncontroversial in many contemporary legal systems, including those of, for example, the United States and the United Kingdom. A. Premise 1: Determinate Legal Content The first premise of the argument is the following: In the legal system under consideration, there is a substantial body of determinate legal content. My use of the term “determinate” (like my use of “determine”) is metaphysical, not epistemic. That is, for the law to be determinate on a given issue is not for us to be able to ascertain what the law requires on that issue (or still less for there to be a consensus), but for there to be a fact of the matter as to what the law requires with respect to the issue. Thus, when I say that there is a substantial body of determinate legal content, I mean roughly that there are many true legal proposi11 There are interesting connections between this paper and G.A. Cohen’s recent Facts and Principles, 31 PHIL. & PUB. AFF. 211 (2003). Cohen’s paper came to my attention too late for me to explore the connections here, however.

MARK GREENBERG

28

tions (in the particular legal system). What do I mean by “legal propositions”? 12 A legal proposition is a legal standard or requirement. An example might be the proposition that any person who, by means of deceit, intentionally deprives another person of property worth more than a thousand dollars shall be imprisoned for not more than six months. For a legal proposition to be true in a particular legal system is for it to be a true statement of the law of that legal system.13 D is consistent with the law’s being indeterminate to some extent, and it is deliberately vague about how much determinacy there is. I think it is obvious that D is true in the legal systems of many contemporary nations. B. Premise 2: The Role of Law-Determining Practices The second premise is: The law-determining practices in part determine the content of the law. The basic idea behind L is that the law depends on the law practices. L thus rules out, for example, the extreme natural-law position that the law is simply whatever morality requires. I take it, however, that very few contemporary legal theorists would defend this position or any other position that makes law practices irrelevant to the content of the law. By the term “law practices” (or, more fully, “law-determining practices”) I mean to include at least constitutions, statutes, executive orders, judicial and administrative decisions, and regulations. Although it is unidiomatic, I will refer to a particular constitution, statute, judicial decision, and so on as a law practice. So a practice, in my usage, need not be a habitual or ongoing pattern of action. I need to clarify what I mean by saying that a practice can be, for example, a statute. Lawyers often talk as if a statute (or other law practice) is simply a text. It is of course permissible to use the word “statute” (or “constitution,” “judicial decision,” etc.) to refer to the corresponding text, and I will occasionally write in this way. But if law practices are to be de12 13

The term is Dworkin’s. See DWORKIN, supra note 6, at 4. I will usually omit the qualification about a particular legal system.

HOW FACTS MAKE LAW

29

terminants of the content of the law, the relevant practice must be, for example, the fact that a majority of the members of the legislature voted in a certain way with respect to a text (or alternatively the event of their having done so), not merely the text itself. So as I will generally use the term, “statute” (“constitution,” etc.) is shorthand for a collection of facts (or events),14 not a text. In general, then, law practices consist of ordinary empirical facts about what people thought, said, and did in various circumstances.15 For example, law practices potentially include the facts that, in a particular historical context, a legislative committee issued a certain report, various speeches were made in a legislative debate, a bill that would have repealed a statute failed to pass, a concurring judge issued a certain opinion, and an executive official announced a particular view of a statute.16 Once I have clarified the claim that law practices partially determine the content of the law, I will be able to say something more precise about what counts as a law practice. When L says that law practices determine (in part) the content of the law, what sense of “determine” is involved? As noted above, a preliminary point is that L’s claim is constitutive or metaphysical, not epistemic. That is, it is not a claim that we use law practices to ascertain what the content of the law is, but that such practices make it the case that the content of the law is what it is. I maintain that the relevant kind of determination is not bare metaphysical determination but what we can call rational determination. 14 I will hereafter ignore the possibility of taking law practices to be composed of events rather than facts. 15 Hypothetical decisions arguably play a significant role in determining the content of the law, but for purposes of this paper they will largely be ignored. Susan Hurley characterizes hypothetical decisions as hypothetical cases that have a settled resolution. See S.L. Hurley, Coherence, Hypothetical Cases, and Precedent, 10 OXFORD J. LEGAL STUD. 221 (1990). Another possibility is to include any hypothetical case that has a determinate right answer, even if there is disagreement on its resolution. There would be disagreement about which hypothetical cases had determinate right answers and therefore about which were determinants of legal content. 16 Nothing turns on how we individuate practices, at least in the first instance. For example, a legislative committee’s issuance of a report could be considered part of the circumstances in which a majority of the legislature voted for a statute or could be considered a separate practice. Once the roles of different elements of law practices are determined, there may be a basis for individuation.

30

MARK GREENBERG

The A facts rationally determine the B facts just in case the A facts metaphysically determine the B facts and the obtaining of the A facts makes intelligible or rationally explains the B facts’ obtaining. Thus, L is the conjunction of two doctrines, a metaphysical-determination doctrine and a rational-relation doctrine. Let me elaborate. I will make the (uncontroversial, I hope) assumption that there are facts that (1) are ontologically more basic than facts about legal content and (2) metaphysically determine that the content of the law is what it is. The metaphysical-determination doctrine is that these more basic facts that determine the content of the law non-redundantly include law practices. Metaphysical determination can be brute. If the A facts are more basic facts that metaphysically determine the B facts, there is a sense in which the A facts explain the B facts, for the A facts are more basic facts, the obtaining of which entails that the B facts obtain. But there need be no explanation of why the obtaining of particular A facts has the consequence that it does for the B facts. To dramatize the point, even a perfectly rational being may not be able to see why it is that particular A facts make particular B facts obtain. The metaphysical-determination doctrine is not enough to capture our ordinary understanding (which L attempts to articulate) of the nature of the determination relation between the law practices and the content of the law. We also need the rational-relation doctrine, which holds that the relation between the determinants of legal content and legal content is reason-based. In the relevant sense, a reason is a consideration that makes the relevant explanandum intelligible.17 Here is one way to put the point. There are indefinitely many possible mappings, from complete sets of law practices to legal content (to complete sets of legal propositions). As far as the metaphysical-determination doctrine goes, it could simply be arbitrary which mapping is 17 I will not attempt to spell out the relevant notion of a reason more fully here. One possibility is that the best way to do so is in terms of idealized human reasoning ability. For example, the idea might be that practices yield a legal proposition if and only if an ideal reasoner would see that they do. The notion of a reason would hence be an epistemic notion. In that case, L would imply that the metaphysics of law involves an epistemic notion; that is, what the law is would depend in part on what an ideal human reasoner would find intelligible.

HOW FACTS MAKE LAW

31

the legally correct one. In other words, the connection between a difference in the practices and a consequent difference in the content of the law could be brute. For example, it is consistent with the truth of the metaphysical-determination doctrine that, say, the deletion of one seemingly unimportant word in one subclause of one minor administrative regulation would result in the elimination of all legal content in the United States—in there being no true legal propositions in the U.S. legal system (though there is no explanation of why it would do so). By contrast, according to the rational-relation doctrine, the correct mapping must be such that there are reasons why law practices have the consequences they do for the content of the law. To put it metaphorically, the relation between the law practices and the content of the law must be transparent.18 (For the relation to be opaque would be for it to be the case that any change in law practices could have, so far as we could tell, any effect on the content of the law. The effects on the content of the law could be unfathomable and unpredictable, even if fully determinate.) It bears emphasis that what must be rationally intelligible is not the content of the law but the relation between determinants of legal content and legal content. L holds not that the content of the law must be rational or reasonable but that it must be intelligible that the determinants of legal content make the content of the law what it is. For example, there must be a reason that deleting a particular word from a statutory text would have the impact on the law that it would in fact have.19 Another important clarification is that it is no part of rational determination that the reasons in question must be value facts. The reasons 18 A useful comparison can be made to certain well-known positions in the philosophy of mind. Donald Davidson’s radical interpretation approach to mental and linguistic content presupposes that behavior determines the contents of mental states and the meaning of linguistic expressions in a way that must be intelligible or transparent. Davidson, Radical Interpretation, in INQUIRIES INTO TRUTH AND INTERPRETATION (1984); and Davidson, Belief and the Basis of Meaning, in INQUIRIES INTO TRUTH AND INTERPRETATION (1984). Similarly, Saul Kripke’s “Kripkenstein” discussion presupposes that we must be able to “read off” the contents of mental states from the determinants of content. SAUL KRIPKE, WITTGENSTEIN ON RULES AND PRIVATE LANGUAGE 24, 29 (1982). See note 25. 19 At this point in the text, I have deleted a paragraph that appeared in the original publication of this paper. I have also made a few changes in the next few paragraphs.

32

MARK GREENBERG

that the determining facts must provide are considerations that explain in rational terms why particular facts of the target domain, as opposed to others, obtain. In general, nonnormative (and nonevaluative) facts can constitute reasons of the relevant kind. Here is an example from a different domain. Consider the facts that (in a particular economy) the total amount of demand deposits is 2 million units and the total amount of currency in the hands of the public is 1 million units. With the fact that the money supply M1 consists in demand deposits plus currency in the hands of the public, these facts make rationally intelligible that the total amount of M1 (in the economy in question) is 3 million units. As far as the rational-relation doctrine is concerned, it is an open question whether there are nonnormative, nonevaluative facts that could constitute reasons for legal facts – and indeed whether there are value facts that could do so. (My ultimate view, of course, is not that value facts could themselves provide such reasons, but that both descriptive and value facts are needed.) The point is important because otherwise positivists could not accept the rational-relation doctrine. The strategy of the argument is to use the rational-relation doctrine, which, I claim below, most legal theorists implicitly accept, to argue for the controversial conclusion that value facts must be determinants of legal content. In principle, conceptual truths (that are not value facts) about law could, with law practices, make rationally intelligible the content of the law. For example, it might be claimed that it follows from the concept of law that a validly enacted statute makes true those propositions that are the ordinary meanings of the sentences of the statute. On this view, the fact that a statutory text says that any person who drives more than sixty-five miles an hour commits an offense, together with certain conceptual truths about law, makes it intelligible that the law requires that one not drive more than sixty-five miles an hour. The general point, again, is that it is a matter for argument, not something presupposed by L, what kinds of facts (if any) must supplement law practices in order for the determining facts to provide reasons that explain why particular legal propositions are true. In particular, L does not presuppose that value facts are needed. 20 20 At a later stage of analysis, we might find that there are other conditions that the determinants of legal content must satisfy. See infra the last paragraph of Section VI.A. L does not presuppose any such conditions, however.

HOW FACTS MAKE LAW

33

Why have I made the qualification that law practices partially determine the content of the law? Law practices must determine the content of the law. But, my argument continues, there are many possible ways in which practices could determine the content of the law. (Put another way, there are many functions that map complete sets of law practices to legal content.) Something other than law practices—X, for short—must help to determine how practices contribute to the content of the law (that is, to determine which mapping is the legally correct one). So a full account of the metaphysics of legal content involves X as well as law practices. This conclusion can be expressed in two equivalent ways. We could say that practices are the only determinants of legal content but that an account of legal content must do more than specify the determinants. This formulation is particularly natural if X consists of necessary truths.21 (A related advantage is that this way of talking highlights that practices are what typically vary, producing changes in the content of the law.) The second formulation would say that X and law practices are together the determinants of the content of the law. Because it is convenient to express this paper’s thesis by saying that X plays a role in determining legal content (and because I want to leave open the possibility that X may vary), this formulation seems preferable, and I will adopt it as my official formulation. Accordingly, I will say that law practices are only some of the determinants of the content of the law. (For brevity, however, I will sometimes omit the qualification “partially” and write simply that law practices determine the content of the law.) C. Law Practices as Descriptive Facts Let me now return to the question of what counts as a law practice. I have said that law practices consist of ordinary empirical facts about what people have thought, said, and done, including paradigmatically facts about what members of constitutional assemblies, legislatures, courts, and administrative agencies have said and done. I want to be clear about the exclusion of two kinds of facts. First, law practices do 21

See text accompanying notes 7 and 8.

34

MARK GREENBERG

not include legal-content facts. Second, law practices do not include facts about value, for example, facts about what morality requires or permits.22 The law practices thus consist of non-legal-content descriptive facts. (For convenience, I will generally write simply “descriptive facts” rather than “non-legal-content descriptive facts”; this shorthand does not reflect a presupposition that legal-content facts are value facts.) Let me explain the reasons for the two exclusions. As I said, I am assuming that the content of the law is not a metaphysically basic aspect of the world but is constituted by more basic facts. The reason for the first exclusion—of legal-content facts—is that law practices are supposed to be the determinants of legal content, not part of the legal content that is to be determined. Suppose an objector maintained that the law practices that determine legal content are themselves laden with legal content. It is certainly natural to use the term “law practices” in this way. After all, the fact that the legislature passed a bill is legal-content laden: it presupposes legal-content facts about what counts as a legislature and a bill. Since legal-content facts are not basic, however, there must be non-legal-content facts that constitute the legal-content-laden practices. At this point, we will have to appeal to descriptive facts about what people thought, said, and did—the facts that I am calling “law practices.” For example, the fact that a legislature did such and such must hold in virtue of complex descriptive facts about people’s behavior and perhaps also value facts. (If, in order to account for legal-content-laden practices, we have to appeal not merely to descriptive facts but also to value facts, so much the worse for the positivist thesis that the content of the law depends only on descriptive facts.) The convenience of 22 By “facts,” I mean simply true propositions. Hence facts about value, or value facts, are true normative or evaluative propositions, such as true propositions about what is right or wrong, good or bad, beautiful or ugly. The fact that people value something or believe something is valuable is not a value fact but a descriptive fact about people’s attitudes. For example, the fact, if it is one, that accepting bribes is wrong is a value fact; the fact that people value honesty is a descriptive fact. This paper does not attempt to address a skeptic who maintains that there are no true propositions about value. One could use an argument of the same form as mine to argue that there must be value facts—for without them there would not be determinate legal requirements. But a skeptic about value facts would no doubt take such an argument to be a case of the legal tail wagging the value dog.

HOW FACTS MAKE LAW

35

talking as if law practices consisted in legal-content-laden facts about the behavior of legislatures, courts, and so on should not obscure the fact that there must be more basic facts in virtue of which the legal-content facts obtain. To build legal-content facts into law practices would beg the question at the heart of this paper—the question of the necessary conditions for law practices to determine the content of the law. (For ease of exposition, I will continue to use legal-content-laden characterizations of the law practices, but the law practices should, strictly speaking, be understood to be the underlying descriptive facts in virtue of which the relevant legal-content facts obtain.) It is uncontroversial that certain kinds of facts are among the supervenience base for legal content: roughly speaking, facts about what constitutional assemblies, legislatures, courts, and administrative agencies did in the past. Of course, as just noted, such characterizations are legal-content-laden and are therefore shorthand for non-legal-content characterizations of the law practices. (I do not mean, of course, that it is uncontroversial exactly which facts of these kinds are relevant; I’ll return to this point shortly.) There are at least two kinds of controversy, however, about the determinants of legal content. First, it is controversial whether value facts are among the determinants of content. The reason for the second exclusion—the exclusion of value facts—is that this paper tries to argue from the uncontroversial claim that law practices are determinants of the content of the law to the conclusion that value facts must play a role in determining the content of the law. If law practices were taken to be value-laden, it would no longer be uncontroversial that they are determinants of legal content. (On the other hand, even those theorists who think that value facts are needed to determine the content of the law can accept that descriptive facts also play a role.) Moreover, unless we separate the descriptive facts from the value facts, we cannot evaluate whether the descriptive facts can themselves determine the content of the law. In sum, by understanding law practices to exclude value facts, I ensure that L is uncontroversial and I prepare the way for my argument that descriptive facts alone cannot determine the content of the law. The second kind of controversy about the determinants of legal content is controversy over precisely which descriptive facts are deter-

36

MARK GREENBERG

minants. I have mentioned some paradigmatic determinants of legal content. But there are other kinds of descriptive facts, for example, facts about customs, about people’s moral beliefs, about political history, and about law practices in other countries that are arguably among the determinants of legal content. Also, somewhat differently, it is controversial which facts about judicial, legislative, or executive behavior are relevant. There can be debate, for example, about the relevance of legislative history, intentions of legislators and of drafters of statutes, legislative findings, judicial obiter dicta, and executive interpretations of statutes. I propose to deal with this second kind of controversy by leaving our understanding of law practices open and nonrestrictive. There are several reasons for this approach. First, my argument is that practices, understood as composed of descriptive facts, cannot themselves determine the content of the law. If I begin with a restrictive understanding of practices, my argument will be open to the reply that I failed to include some of the relevant facts. For this reason, I want to be liberal about which descriptive facts are part of law practices. Second, my argument will not depend on exactly which descriptive facts make up law practices. Rather, I will make a general argument that descriptive facts—in particular, facts about what people have done and said and thought—cannot by themselves determine the content of the law. Therefore it will not matter precisely which such facts are included in law practices. Third, my view is ultimately that the question of which facts are part of law practices—like the question of how different aspects of law practices contribute to the content of the law—is dependent on value facts. (Indeed, I will often treat the two questions together as different aspects of the general question of the way in which law practices determine the content of the law.) As we will see, that we cannot in an uncontroversial way specify which are law practices and which are not is one consideration in support of my argument for the necessary role of value. All we need to begin with is some rough idea of law practices, which can be overinclusive. In sum, let law practices include, in addition to constitutions, statutes, and judicial and administrative decisions, any other non-legal-content descriptive facts that turn out to play a role in determining

HOW FACTS MAKE LAW

37

the content of the law.23 Which facts these are and what role they play are controversial, so we can begin with a rough and inclusive understanding of law practices. One aspect of figuring out how law practices contribute to the content of the law will be figuring out which facts make a contribution and which do not. But there is no reason to expect a clean line between law practices and other facts.24 The exclusion of value facts should not be taken to suggest that law practices are to be understood in solely physical or behavioral terms. To the contrary, as I explain in the next section, I take for granted the mental and linguistic contents involved in law practices. In other words, law practices include the facts about what the actors believe, intend, and so on and about what their words mean. D. Why L Should be Uncontroversial The metaphysical-determination doctrine should be relatively uncontroversial, certainly for those who accept that there are determinate legal requirements. Positivists, Dworkinians, and contemporary natural-law theorists as well as practicing lawyers and judges accept that constitutions, statutes, and judicial and administrative decisions are (nonredundant) determinants of the content of the law. That law practices may also include other descriptive facts to the extent that those facts are determinants of the content of the law obviously cannot make the metaphysical-determination doctrine controversial. More generally, we began with the premise that there are determinate legal requirements. What makes them legal requirements is that they are determined, at least in part, by law practices. Contrast the requirements of morality (or, to take a different kind of example, of a 23 This proviso does not make L the tautological claim that the determinants of legal content determine legal content. L says that constitutions, statutes, judicial decisions, and so on are (nonredundantly) among the determinants of content. 24 One natural understanding of “law practices” is more restrictive than the way I use the term. According to this understanding, law practices are limited to (facts about) what legal institutions and officials do in their official capacities. If we used the term “law practices” in this natural way, we would need, in addition to the category of law practices, a category of other descriptive facts that play a role in determining the content of the law.

38

MARK GREENBERG

particular club). If law practices did not determine legal content, there could still be moral requirements and officials’ whims, but there would be no legal requirements. In order to think differently, one would have to hold a strange view of the metaphysics of law according to which the content of the law is what it is independently of all the facts of what people said and did that make up law practices, and law practices are at best evidence of that content. So I think it should be uncontroversial that law practices are among the determinants of the content of the law. As to the rational-relation doctrine, it is fundamental to our ordinary understanding of the law and taken for granted by most legal theory, though seldom articulated. The basic idea is that the content of the law is in principle accessible to a rational creature who is aware of the relevant law practices. It is not possible that the truth of a legal proposition could simply be opaque, in the sense that there would be no possibility of seeing its truth to be an intelligible consequence of the law practices. In other words, that the law practices support these legal propositions over all others is always a matter of reasons—where reasons are considerations in principle intelligible to rational creatures. (A corollary is that to the extent that the law practices do not provide reasons supporting certain legal propositions over others, the law is indeterminate.) I will not attempt to defend the rational-relation doctrine fully here but will mention a few considerations. Suppose the A facts metaphysically determine the B facts, but the relation between the A facts and the B facts is opaque. In that case, how could we know about the B facts? One possibility is that we have access to the B facts independently of our knowledge of the A facts. An example might be the relation between the microphysical facts about someone’s brain and the facts about that person’s conscious experience. Suppose that the microphysical facts metaphysically determine the facts about the person’s conscious experience but that the relation is opaque. The opaqueness of the relation does not affect the person’s ability to know the facts about his conscious experience, because we do not in general learn about our conscious experience by working it out from the microphysical facts. (Moreover, since we have independent knowledge of conscious experience, we might be able to discover correla-

HOW FACTS MAKE LAW

39

tions between microphysical facts and conscious experience even if those correlations were not intelligible even in principle.) To take a different kind of example, the microphysical facts may metaphysically determine the facts about the weather, and the relation may be opaque, but again, we do not learn about the weather by working it out from the microphysical facts. A second possibility is that we do work out the B facts from the A facts but that we have a nonrational, perhaps hardwired, capacity to do so. For example, it is plausible that the facts about what was said and done (on a particular occasion, say) determine whether what was said and done was funny (and to what degree and in what way). And we do work out whether an incident was funny from the facts about what was done and said. It is plausible, however, that the relation between what was said and done and its funniness is not necessarily transparent to all rational creatures; our ability to know what is funny may depend on species-specific tendencies; that is, there may not be reasons that make the humor facts intelligible; it may just be a brute fact that humans find certain things funny. 25 Law seems different from both of these kinds of cases. First, our only access to the content of the law is through law practices. It is not as if we can find out what the law is directly or through some other route. And the whole enterprise of law-making is premised on the assumption that the behavior of legislators, judges, and other lawmakers will have understandable and predictable consequences for the content of the law. Second, we are able to work out what the law is and predict the effect on the law of changes in law practices through reasons, not through some nonrational human tendency to have correct law reactions to law practices. 25 Compare the issue of how facts about our use of words determine their meaning. Natural languages are a biological creation. Although many philosophers have thought differently (see note 18), we cannot take for granted that the correct mapping from the use of words to their meaning will be based on reasons. How, it may be objected, would we then be able to work out from their use of words what others mean? The answer may simply be that we have a species-specific, hardwired mechanism that rules out many incorrect mappings that are not ruled out by reasons. In that case, an intelligent creature without that mechanism would not be able to work out what words mean.

MARK GREENBERG

40

When lawyers, judges, and law professors work out what the law is, they give reasons for their conclusions. Indeed, if we find that we cannot articulate reasons that justify a provisional judgment about what the law is in light of law practices, we reject the judgment. By contrast, it is notoriously difficult to explain why something is or is not funny, and we do not generally hold our judgments about humor responsible to our ability to articulate reasons for them. A related point is that we believe that we could teach any intelligent creature that is sensitive to reasons how to work out what the law is. It might be objected that although the epistemology of law is reason-based, the metaphysics might not be. It is difficult to see how such an objection could be developed. For present purposes, I will simply point out that when legal practitioners give reasons for their conclusions about what the law is, they believe that they are not merely citing evidence that is contingently connected to the content of the law; rather, they believe that they are giving the reasons that make the law what it is. The point is not that lawyers believe themselves to be infallible. Rather, they believe that when they get things right, the reasons they discover are not merely reasons for believing that the content of the law is a particular way, but the reasons that make the content of the law what it is. Although they would never put it this way, lawyers take for granted that the epistemology of law tracks its metaphysics. And the epistemology of law is plainly reason-based. Legal theorists generally take for granted some version of the claim that the relation between law practices and the content of the law is reason-based. An example is H.L.A. Hart’s argument that the vagueness and open texture of legal language have the consequence that the law is indeterminate.26 If bare metaphysical determination were all that was at issue—if it were not the case that the relation between practice and content were necessarily intelligible—the vagueness of language would in no way support the claim that law was indeterminate. Similarly, when legal realists or Critical Legal Studies theorists argue that the existence of conflicting pronouncements or doctrines in law practices results in underdetermination of the law, their arguments

26

H.L.A. HART, THE CONCEPT OF LAW ch. 7 (2nd ed. 1997).

HOW FACTS MAKE LAW

41

would be beside the point if what was at stake were not rational determination.27 In general, the large body of legal theory that has explored the question of whether law practices are capable of rendering the law determinate (and if so, how determinate) presupposes that law practices determine the content of the law in a reason-based way. If the relation between law practices and the content of the law could be opaque, any set of law practices would be capable, as far as we would be able to judge, of determining any set of legal propositions. (As long as there are as many possible sets of law practices as there are possible sets of legal propositions, there is no barrier to the content of the law’s being fixed by the practices, and we would have no warrant to rely on our assessment of other putative prerequisites for practices to determine the content of the law.) In sum, the doctrine that law practices rationally determine the content of the law captures a basic conviction about the law that is shared by lawmakers, lawyers, and legal theorists and is supported by the epistemology of law. Why does it matter to my argument that the relation between law practices and the content of the law is reason-based? This paper explores the necessary conditions for law practices’ making the content of the law what it is. The central argument is that descriptive facts cannot determine their own rational significance—what reasons they provide. The argument therefore depends on the claim that the descriptive facts determine the content of the law in a reason-based way. It turns out that value facts are needed to make it intelligible that law practices support certain legal propositions over others. 28 E. The Scope of the Argument Premises D and L tell us something about the scope of my argument. The argument is sound only for legal systems in which D and L are See, e.g., Andrew Altman, Legal Realism, Critical Legal Studies, and Dworkin, 15 PHIL. & PUB. AFF. 205 (1986); Mark Kelman, Interpretive Construction in the Substantive Criminal Law, 33 STAN. L. REV. 591 (1981). 28 Suppose that the relation between law practices and the content of the law were necessarily intelligible only in a way that depends on some human-specific tendency. As long as practices must provide considerations that are intelligible (even if only to humans), a version of my argument should still go through. 27

42

MARK GREENBERG

true. So my conclusions are limited to legal systems in which there are legal requirements that are determined in part by law practices. If there is a legal system in which there are no determinate legal requirements, my argument would not apply to it. Similarly, if there is a legal system in which law practices, understood as (facts about) various people’s sayings and doings, do not play a role in determining the content of the law, my argument would not apply to it. For example, perhaps there could be a legal system in which the content of the law is determined exclusively by the content of morality or exclusively by divine will. In this paper, I do not address questions of the necessary conditions for something’s counting as a legal system. It might be argued that a substantial body of legal requirements that are determined by practices of various officials or institutions is a necessary condition for the existence of a legal system, but I do not intend to pursue such an argument. III. Is There a Distinctively Legal Problem of Content? We begin with our two premises: that the law has determinate content, and that law practices in part determine that content. Our question is: What conditions must be satisfied in order for law practices to determine legal propositions? As I said above, since we are interested in problems of the determination of content only to the extent that they are peculiarly legal, we can take for granted the content of sentences and propositional attitudes.29 So the question is: How can a collection of facts about what various people did and said (including the facts about what they intended, believed, preferred, and hoped, and about what their words meant) determine which legal propositions are true? At this point, however, it must be asked whether there is a peculiarly legal problem of content. Once we take for granted the relevant 29 There is no practical problem with taking these matters for granted and proceeding without a solution to basic problems concerning how linguistic and mental content are possible. These problems do not concern difficulties we encounter in practice in attributing linguistic and mental content; the difficulty is in saying what it is in virtue of which a linguistic expression or mental state has its content.

HOW FACTS MAKE LAW

43

mental and linguistic content, it may seem that no problem of legal content remains. Legal content is simply the content of the appropriate mental states and texts. In this section, I consider this possibility and argue that it is not at all plausible. The ordinary mental and linguistic content of utterances and mental states of participants in law practices—nonlegal content, for short—does not automatically endow the law with legal content. Something must determine which aspects of law practices are relevant and how they together contribute to the content of the law. In the next section, I consider the possibility that, given the content of the relevant utterances and attitudes, law practices themselves determine how they contribute to the content of the law and thus can unilaterally determine the content of the law. But before we turn to whether law practices can solve the problem of legal content, we need to see what the problem is—why the nonlegal content of law practices does not provide the content of the law. That is the topic of this section. In legal discourse, both ordinary and academic, constitutional or statutory provisions and judicial decisions are often conflated with rules or legal propositions. For example, lawyers will sometimes talk interchangeably of a statutory provision and a statutory rule, or of a judicial decision and the rule of that case. In nonphilosophical contexts there is generally no harm in this kind of talk. Since our question, however, is how law practices determine the content of the law, it is crucial not to confuse law practices with legal propositions. For example, if one assumed that a statute was the rule or proposition expressed by the words of the statute, one might think that there was no problem of how law practices could determine legal content; or one might think that the only problem was how to combine or amalgamate a large number of rules or propositions. Although it would beg the question to take legal propositions for granted, we do have the propositions that are the content of the utterances and mental states of participants in law practices. What is wrong with the idea that those propositions constitute legal content, so that law practices, once they are understood to include facts about mental and linguistic content, automatically have legal content? I will begin with the least serious problems—those concerning the attribution of nonlegal content. Although we are normally able to at-

44

MARK GREENBERG

tribute attitudes to people based on what they say and do and to attribute standard meanings to a large number of sentences of a language we speak, there are difficulties in attributing nonlegal content to aspects of a putative law practice. Here are a few examples. First, when I say that we can take for granted mental and linguistic content, I mean that we need not ignore the mental and linguistic content that is available. We should not, however, assume that all of the contents of the mental states of all of the people involved in law practices are available. That would obviously be false. In general, what is available in the standard reports of law practices is not sufficient to attribute much in the way of attitudes to the people who actually performed the actions and made the utterances; the fact that a particular legislator voted for a bill or a certain judge signed an opinion is not in general sufficient to attribute beliefs, intentions, hopes, and so on to her. Moreover, the law restricts what evidence of the intentions and beliefs of legislators and judges is acceptable to determine the content of the law. Even when the intentions of a legislator or judge are relevant to the content of the law, it is not the case that, say, her private letters or diary may be a source of that intention. Something must determine which evidence of legally relevant attitudes is legally acceptable. Second, though many sentences of natural languages have standard meanings, it is notorious that this is not true of some of the sentences uttered by those engaged in making law practices. The point here is not that in legal contexts linguistic expressions often have specialized meanings that are not straightforwardly connected to their ordinary meanings. Rather, some of the contorted sentences in the law books have no standard meaning in a natural language. Third, even when sentences taken alone have standard meanings, collections of those sentences may fail to do so. In other words, the property of having a standard meaning (on a notion of standard meaning appropriate for present purposes) is not closed under conjunction (for example, because context may introduce ambiguity into an otherwise unambiguous sentence). Setting aside these problems with ascertaining nonlegal content, we can turn to the more important question of the bearing of nonlegal content on legal content. One problem is that the nonlegal content of some elements of law practices has, or arguably has, little or nothing

HOW FACTS MAKE LAW

45

to do with the legal content determined by those practices. Consider sentences in statutory preambles, sentences in presidential speeches at bill-signing ceremonies, and sentences in judicial opinions that are not necessary to the resolution of the issue before the court. Another example is the actual but unexpressed hopes of the members of the legislature as to how the courts would interpret a statute. Countless sentences are written and spoken at different stages of law-practice-making by people with myriad attitudes.30 Something must determine which sentences’ and attitudes’ contents are relevant. Another problem is that the contribution of a particular law practice to the content of the law may not be the meaning of any text or the content of any person’s mental state. The actual attitudes of appellate judges may be irrelevant; instead the relevant question may be what a hypothetical reasonable person would have intended by the words uttered by the judges or what would be the best, or the narrowest, explanation of the result reached. Another possibility is that aspects of law practices that contribute to nonlegal content in one way contribute to legal content in an entirely different way; facts about what was said and done may have peculiarly legal significance. An obvious example is that common words such as “malice” and “fault” are often used in legal discourse in a technical sense. To take a more subtle instance, when a panel of several judges is badly split, it can be a complex and tricky matter to ascertain the relevance to legal content of the meanings of the words of the different judicial opinions. Similarly, facts about the circumstances in which sayings and doings occurred that have little to do with the nonlegal content of the people’s attitudes and words may significantly affect the content of the law. For example, in a judicial decision, the fact that an issue is not in controversy arguably prevents the court’s statements on that issue from making any contribution to the content of the law. 30 In the case of a judicial decision, for example, the possibly relevant sentences include sentences uttered by the parties to the controversy, by lawyers, and by judges to lawyers and other judges. They include sentences written by judges in orders and judicial opinions. Judicial opinions alone include a large number and variety of sentences: they state facts, give reasons, summarize, make general claims about the content of the law, state holdings; moreover, there are concurring and dissenting as well as majority opinions.

46

MARK GREENBERG

Even when the content of sentences and mental states is relevant to the content of the law, there can be no mechanical derivation of the content of the law. For example, how are conflicting contents to be combined? In general, there remains the problem of how the nonlegal contents associated with different law practices interact with each other (and with other relevant aspects of law practices) to determine the content of the law. We have surveyed a number of reasons why nonlegal content—the meanings of sentences and contents of mental states—does not simply constitute legal content. But this way of thinking about the problem will have an artificial quality for those familiar with legal reasoning. The idea that the nonlegal content of law practices constitutes their legal content presupposes roughly the following picture. Associated with each law practice is a text (and perhaps some mental states). Once we have the meanings of the texts and the contents of the mental states, each law practice will be associated with a proposition or set of propositions. Ascertaining the law on a particular issue is just a matter of looking up the propositions that are applicable to the issue. Even if this picture were accurate, we have discussed a number of reasons why nonlegal content would not automatically yield legal content. But the problem is worse than these reasons would suggest. As I will now suggest, the whole picture is wrongheaded. Law practices do not determine the content of the law by contributing propositions which then get amalgamated. Here is the real problem of legal content. There are many different law practices with many different aspects or elements. There is an initial question of which facts are parts of law practices and which are not. Are preambles of bills, legislative findings, legislative committee reports, dissenting opinions, unpublished judicial decisions, customs, the Federalist Papers, and so on to be included in law practices? In my view, this question is really just part of a second question: Which aspects of, for example, judicial or legislative practices are relevant to the content of the law? Just to suggest the dimensions of the problem, here are some candidates for the relevant elements or aspects of practices. With respect to a judicial decision: the facts of the case, the judgment rendered, the words used by the court in the majority opinion, the reasons given for the outcome, the judges’ beliefs, the

HOW FACTS MAKE LAW

47

judges’ identities, the level and jurisdiction of the court; with respect to a legislative action: the words of the statute, the legislature’s actual intention (if there is such a thing), the purposes that the words of the statute could reasonably be intended to implement, statements by the person who drafted the statute, speeches made during the legislative debate preceding passage, the circumstances in which the legislature acted, subsequent decisions not to repeal the statute. Third, once we know which elements of practices are relevant, the problem of determining the content of the law is not simply a problem of adding or amalgamating the various relevant aspects of practices. One obvious point is that some elements of practices are far more important than others, and elements of practices matter in different ways. But more fundamentally, as anyone familiar with legal reasoning knows, the content of the law is not determined by any kind of summing procedure, however complicated. For example, judicial decisions, constitutional provisions, and legislative history can affect what contribution a statute makes. It is not that those practices contribute propositions that are conjoined to a proposition contributed by the statute. The statute’s correct interpretation may be determined by a potential conflict with a constitutional provision or by the outcome of cases in which courts have interpreted the same or related statutes. To take a different kind of example, constitutional provisions, statutes, and judicial decisions can have an impact on the contribution of judicial and administrative decisions to the content of the law by affecting our understanding of the proper role of courts and administrative agencies. Or, differently, statutes can have an impact on what judicial decisions mean by making clear what the legislature cares about, thus affecting which differences between cases matter and consequently whether past precedents control the present issue. A final example is that the principle that a series of cases stands for is not the conjunction of the propositions announced in each case. It is safe to conclude that the law does not automatically acquire content when actions, utterances, and sentences involved in law practices are attributed content. It is a mistake even to think that the issue is how to convert nonlegal content into legal content. We need to reject the simplistic picture in which each law practice contributes to the content of the law a discrete proposition (or set of propositions),

MARK GREENBERG

48

which is the result of converting the nonlegal content of sentences and mental states into legal content. The bearing of nonlegal content on the content of the law is not mechanical. Once we root out any idea of a mechanical conversion of nonlegal content to legal content, it is clear that something must determine which aspects of law practices are relevant to the content of the law and what role those relevant aspects play in contributing to the content of the law. IV. Can Law Practices Themselves Determine How They Contribute to the Content of the Law? In this section, I consider the possibility that law practices can themselves determine how they contribute to the content of the law. I will argue that without standards independent of practices, practices cannot themselves adjudicate between ways in which practices could contribute to the content of the law. For convenience, let me introduce a term for a candidate way in which practices could contribute to the content of the law. I will call such a way a model (short for a model of the role of law-determining practices in contributing to the content of the law).31 The rational-relation doctrine tells us that there are systematic, intelligible connections between practices and the content of the law. It thus guarantees that there are rules that, given any pattern of law practices, yield a total set of legal propositions. A model is such a rule or set of rules. A model is the counterpart at the metaphysical level of a method of interpretation at the epistemic level. (A model’s being correct in a given legal system is what makes the corresponding theory of interpretation true.) Although the term is not ideal, I use “model” rather than “method of interpretation” to signal that my concern is constitutive or metaphysical, not epistemic; that is, the issue is how practices make it the case that the law’s content is what it is, not how we can ascertain the law’s content from law practices. Because it is more idiomatic, however, I will sometimes write in epistemic terms when discussing models. 31

My thanks to Nicos Stavropoulos for suggesting this term.

HOW FACTS MAKE LAW

49

(By way of analogy, it may be helpful to compare, on the one hand, the relation between practices and the content of the law with, on the other, the relation between words and the meaning of a sentence or group of sentences. The meaning of a sentence depends in a systematic, intelligible way on the arrangement of constituent words; analogously, the content of the law—in a given legal system at a given time—depends on the pattern of law practices. A specification of the meanings of individual words and of the compositional rules of the language is a specification of the rules by which the words determine the meaning of the sentence. Analogously, a specification of a model is a specification of the rules by which law practices determine the content of the law. In this sense, a model is the analogue of the meanings of individual words and the compositional rules for the language.) I will use the term “model” sometimes for a partial model—a rule for the relevance of some aspect of law practices, for example of legislative findings or of dissenting judicial opinions, to the content of the law—and sometimes for a complete model—all of the rules by which law practices determine the content of the law. The context should make clear whether partial or complete models are in question. The legally correct (or, for short, correct) model in a particular legal system at a particular time is the way in which practices in that legal system at that time actually contribute to the content of the law (not merely the way in which they are thought to do so). Which model is correct varies from legal system to legal system and from time to time within a legal system, since, as we will see, which model is correct depends in part on law practices. Models come at different levels of generality. More specific ones include the metaphysical counterparts of theories of constitutional, statutory, and common-law interpretation. Models can also be understood to include very general putative ways in which law practices determine what the law requires. Thus Hart’s rule-of-recognition-based theory of law and Dworkin’s “law as integrity” theory are accounts of very general models. Very general models give rise to more localized models of the contributions made by specific elements of practices. Candidate models are candidate ways in which practices contribute to the content of the law. Since the issue of how practices contribute to the content of the law has several components, models have several

50

MARK GREENBERG

closely related roles: they determine what counts as a law practice; which aspects of law practices are relevant to the content of the law; and how different relevant aspects combine to determine the content of the law, including how conflicts between relevant aspects are resolved. The question of what determines how practices contribute to the content of the law can therefore be reformulated as the question of what determines which models are correct. What settles, for example, the question whether the original-intent theory of constitutional interpretation is true? We can now turn to the main topic of this section: whether law practices can themselves determine which model is correct. Certainly the content of the law, as determined by law practices, concerns, in addition to more familiar subjects of legal regulation, what models are correct. That is, the content of the law includes rules for the bearing of law practices on the content of the law. For example, it is part of the law of the United States that the Constitution is the supreme law, that bills that have a bare majority of both houses of Congress do not contribute to the content of the law unless the president signs them, and that precedents of higher courts are binding on lower courts in the same jurisdiction. The content of the law cannot itself determine which model is correct, however, for the content of the law depends on which model is correct. If, for example, statutes contributed to the law only the plain meaning of their words, the content of the law would be different from what it would be if the legislators’ intentions made a difference. Obviously, which legal propositions are true depends on which model is correct. But as we have just seen, which model is correct depends in part on the legal propositions. The content of the law and the correct model are thus interdependent. This interdependence threatens to bring indeterminacy. Consider the law practices of a particular legal system at a particular time and ask what the content of the law is. Suppose that if candidate model A were legally correct, a certain set of legal propositions would be true, according to which model A would be correct. And if candidate model B were correct, a different set of legal propositions would be true, ac-

HOW FACTS MAKE LAW

51

cording to which model B would be correct. And so on. Without some other standard, each mutually supporting pair of model and set of legal propositions is no more favored than any other pair. 32 Can law practices determine which model is correct? The prima facie problem is that we cannot appeal to practices to determine which model is correct because which model a set of practices supports itself depends on which model is correct. But let us consider the matter in more depth. If practices are to determine which model is correct, there are two possibilities. First, a privileged foundational practice (or set of foundational practices) could determine the role of other practices. This possibility encounters the problem of how practices themselves can determine which practices are foundational. For example, the fact that a judicial opinion states that only the rationale necessary to the decision of a case is contributed to the content of the law cannot determine that that is a correct account of the contribution of judicial decisions to the content of the law. Something must determine that the judicial opinion in question is relevant and trumps other conflicting practices. A putatively foundational practice cannot non-question-beggingly provide the reason that it is foundational. Moreover, it is unwarranted to assume that the significance of a putatively foundational practice is simply its nonlegal content. Its significance depends on which model is correct—the very issue the practice is supposed to resolve. In sum, a 32

This note registers a rather technical qualification and can be skipped without losing the main thread of the argument. A candidate model, given the law practices, may yield a set of legal propositions that lends support to a different, inconsistent model. To the extent that this is the case, we can say that the model is not in equilibrium (relative to the law practices). Models that are in equilibrium (or are closer to it) are plausibly favored, others things being equal, over those that are not (or are further from it). There is no reason to expect, however, that there will be typically be only one model that is closer to equilibrium than any other model. In fact, indefinitely many models are guaranteed to be in perfect equilibrium (yet yield different sets of legal propositions). For example, any model that includes a rule that practices (and thus the true legal propositions) have no bearing on which model is correct is necessarily in perfect equilibrium. Without some independent standard for what models are eligible, there is no way to rule out such models. Hence the varying degree to which different candidate models are in equilibrium does not ensure a unique correct model and determinate legal content. See also the discussion of a coherentist solution in the text below.

MARK GREENBERG

52

foundationalist solution is hopeless because it requires some independent factor that determines which practices are foundational (and what their contribution is). Second, if no practices can be assumed to have a privileged status, the remaining possibility is that all law practices together can somehow determine their own role. Such a coherentist solution might at first seem to have more going for it than the foundationalist one. The idea would be, roughly speaking, that the (total) law practices support the model that, when applied to the practices, yields the result that the practices support that very model. If no model is perfectly supported in this way, the one that comes closest is the correct one. The problem with this suggestion, crudely put, is that without substantive standards that determine the relevance of different aspects of law practices, the (total) law practices will support too many models. For any legal proposition, there will always be a model supported by the practices that yields that proposition. Or to put it another way, the formal requirement that a model be supported by or cohere with law practices is empty without substantive standards that determine what counts as a relevant difference. Suppose a body of judicial decisions seems to support the proposition that a court is to give deference to an administrative agency’s interpretation of a statute. It is consistent with those decisions for an agency’s interpretation of a statute not to deserve deference when there is a reason for the different treatment. Such a reason could be, for example, that the agency in the earlier cases, but not in the present case, had special responsibility for administration of the relevant statutory scheme. But since the facts of every case are different, if a model can count any difference as relevant, there will always be a model that is consistent with all past practices yet denies deference to agency interpretations of statutes. As I have argued more fully elsewhere, such considerations show that practices cannot determine legal content without standards independent of the practices that determine which differences are relevant and irrelevant.33 Hence law practices alone cannot yield determinate legal requirements. The point is a specific application of a familiar, 33

See Mark Greenberg & Harry Litman, The Meaning of Original Meaning, 86 614–617 (1998).

GEO L.J. 569,

HOW FACTS MAKE LAW

53

more general point that Susan Hurley has developed.34 Formal requirements such as consistency are meaningful only in the light of substantive standards that limit which factors can provide reasons. It would be missing the point to suggest that law practices themselves can determine the appropriate standards. Without such standards, a requirement of adherence to practices is empty. In epistemic terms, we cannot derive the standards from the practices because the standards are a prerequisite for interpreting the practices. It may be helpful to notice that the problem has a structure similar to that of two famous philosophical puzzles: Nelson Goodman’s problem about green and grue, and Saul Kripke’s problem about plus and quus.35 In order for there to be legal requirements, it must be possible for someone to make a mistake in attributing a legal requirement (if just any attribution of a legal requirement is correct, the law requires that P and that not P and so does not require anything). One makes a mistake when one attributes a legal requirement that is not the one the law practices yield when interpreted in accordance with the correct model. For any candidate legal requirement, however, there is always a nonstandard or “bent” model that yields that requirement. It is therefore open to an interpreter charged with a mistake to claim that in attributing the legal requirement in question, she has not made a mistake in applying one model but is applying a different model. The proponent of the coherence solution will respond that law practices themselves support certain models. For example, in appealing to practices to decide cases, courts have developed well-established ways of understanding the relevance of those practices to legal content. The problem is that there will always be bent models according to which the judicial decisions (and other practices) support the bent models rather than the purportedly well-established ones. This kind of point See S.L. HURLEY, NATURAL REASONS 26, 84–88 (1989). Hurley credits Ramsey’s and Davidson’s uses of arguments with similar import. See, e.g., Donald Davidson, The Structure and Content of Truth, 87 J. PHIL. 279, 317–320 (1990). 35 See NELSON GOODMAN, FACT, FICTION, AND FORECAST 72–81 (3d ed. 1973); SAUL KRIPKE, WITTGENSTEIN ON RULES AND PRIVATE LANGUAGE 7–32 (1982). These puzzles involve concepts that seem bizarre and gerrymandered. One challenge is to determine what it is that rules such concepts out (at least in particular contexts), for if they are not ruled out, unacceptable results follow. 34

MARK GREENBERG

54

shows that there must be factors, not themselves derived from the practices, that favor some models over others. Here is an example.36 Suppose that on February 1, 2005, a judge in a state court in the United States must decide whether a woman has a federal constitutional right not to be prevented from obtaining an abortion. Imagine that the judge holds that the woman does not have such a right. It seems that the judge has misread Roe v. Wade,37 the seminal decision of the United States Supreme Court. The judge claims, however, that according to the correct model of how judicial decisions contribute to legal content, when constitutional rights of individuals are at stake and strong considerations of justice support the claims of both sides, such decisions should be understood as establishing a form of “checkerboard” solution. According to such a solution, whether a person has the right in question depends on whether the person is born on an odd- or even-numbered day.38 Since Jane Roe was born on an odd-numbered day (let us assume), Roe v. Wade’s contribution to content is that only women born on odd-numbered days have a constitutional right to an abortion. Before discussing the example, it must be emphasized that the point is not that the judge’s position should be taken seriously; on the contrary, the example depends on the fact that the judge’s position is plainly a nonstarter. Since it is evident that the position cannot be taken seriously, there must be factors that rule out models like the one in the example. The example makes the point that these factors must be independent of practices. Since the unacceptable positions that we want to exclude purport to determine what practices mean, the factors that exclude these positions cannot be based on practices. Moreover, there is no way to rule out such positions on a purely logical level, since, as will become evident, it is easy to construct self-supporting, logically consistent systems of such positions. The claim is, then, that our unwillingness to take the judge’s position seriously suggests 36

The example borrows from Dworkin’s discussion of a “checkerboard” solution to the abortion controversy. See DWORKIN, supra note 6, at 178–186. Dworkin cannot be held responsible, however, for my example. 37 Roe v. Wade, 410 U.S. 113 (1973). 38 See DWORKIN, supra note 6, at 178–179.

HOW FACTS MAKE LAW

55

that we must be depending on tacit assumptions independent of law practices in determining which models are acceptable. Let us look at the example to see why practices themselves cannot exclude the judge’s model. The first objection to the judge’s position may be that the Supreme Court in Roe v. Wade said nothing about the abortion right’s depending on birth dates. The judge replies that according to the correct model, the reasons that judges give in their opinions make no or little contribution to the content of the law. A second objection may move to a different level: the practices of the legal system do not support the judge’s model. Judicial decisions, for example, do not interpret the contributions made by other decisions in such a checkerboard fashion, nor do they ignore the reasons judges give. The judge, however, claims that according to his model, judicial decisions have all along been using a bent model, according to which the reasons judges give are significant until February 1, 2005, but not afterwards. Similarly, the model specifies no checkerboard contributions to content until that date, then requires them afterwards. All of the judicial decisions so far are logically consistent with the hypothesis that they are using the bent model. Obviously a third-level objection—that the practices do not support models that give dates this sort of significance—can be met with the same sort of response. In another version of the example, the judge might claim that according to the correct model, in all cases involving the right to abortion, a Supreme Court decision’s relevance to content ends, without further action by the Court, as soon as a majority of the current Supreme Court believes that the decision was wrongly decided. Since the judge believes that that is now the situation with regard to Roe v. Wade, he claims that Roe v. Wade no longer has any bearing on the content of the law. If it is objected that the judge’s position is not an accurate account of how judicial decisions interpret past judicial decisions, the judge will claim that judicial decisions have been following his model all along. Since (let us suppose) it has never been the case before that a majority of the Supreme Court has disagreed with a past Supreme Court decision on the right to abortion, the evidence of past decisions supports the judge’s model, which treats only abortion rights cases idiosyncratically, as strongly as a more conventional one.

56

MARK GREENBERG

The point should be obvious by now: these sorts of unacceptable models are unacceptable because there are standards independent of practices that determine that some sorts of factors are irrelevant to the contributions made by practices to legal content. The practices themselves cannot be the source of the standards for which models are permissible. In this section, I have argued that practices themselves cannot determine how practices contribute to the content of the law. Although I will not discuss the point here, it is worth noting that my argument is not limited to the law. For example, the argument shows that without standards independent of the practices, no set of practices can rationally determine rules. What rules a set of practices rationally determines will depend on what aspects of the practices are relevant and how those aspects are relevant. And the practices cannot themselves resolve those issues. Similarly, my argument does not depend on the complexities of contemporary legal systems. My point therefore holds even for extremely simple cases. Even if there were only one lawmaker who uttered only simple sentences, and even if it were taken for granted that the lawmaker’s practices were legally relevant, the precise relevance of those practices would still depend on factors independent of the practices. For example, there would still be an issue of whether the relevant aspect of the practices was the meaning of the words uttered, as opposed to, say, the lawmaker’s intentions or the narrowest rationale necessary to justify the outcome of the lawmaker’s decisions. V. Objections I want now to consider three closely related objections. First, it may be objected that in practice there is often no difficulty in knowing which aspects of a practice are relevant or which facts provide reasons. Bent models are not serious candidates. Second, it may be objected that practitioners’ beliefs (or other attitudes) about value questions, not value facts, solve the problem of determining how practices contribute to the content of the law. Third, it may be said that in limiting law practices to descriptive facts, I have relied on too thin a con-

HOW FACTS MAKE LAW

57

ception of law practices. Properly understood, law practices can themselves determine the content of the law. I replied to a version of the first objection in discussing the example of the abortion-rights decision, but I will make the point in more general terms here. As I have emphasized, the question of the necessary conditions for law practices to determine the content of the law is a metaphysical, not an epistemic, question. The problems that I have raised concerning how law practices determine the content of the law are not practical problems that legal interpreters encounter in trying to discover what the law requires. Hence it is no objection to my argument that legal interpreters do not encounter such problems. I have argued that there is a gap between law practices and the content of the law that can be bridged only by substantive factors independent of practices. If legal practitioners have no difficulty in crossing this gap—for example, in eliminating bent models from consideration—that must be because they take the necessary factors for granted. With respect to the example of the abortion-rights decision, I argued that practices themselves cannot rule out the judge’s bent models. Therefore our unwillingness to take the judge’s position seriously is evidence that we are relying on tacit assumptions about what models are acceptable. The lack of difficulty in practice suggests not that substantive constraints are not needed but that they are assumed. This point leads naturally to the second objection, which holds that it is the assumptions or beliefs of participants in the practice that solve the problem of how practices determine the content of the law. For example, it might be that a consensus or shared understanding among judges or legal officials determines the relevance of practices to the content of the law. Beliefs about value, not value facts, do the necessary work. As an epistemic matter, of course, we rely on our beliefs about value to ascertain what the law is. But that is exactly what we would expect if the content of the law depended on value facts. After all, in working out the truth in any domain, we must depend on our beliefs. That we do so in a given domain in no way suggests that the truth in that domain depends on our beliefs. Notice, moreover, that if the content of the law depended on beliefs about value, then in order to work out what the law was, we would have to rely on our beliefs about our

58

MARK GREENBERG

beliefs about value. For example, we might ask not whether democratic values favor intentionalist theories of statutory interpretation, but whether there is a consensus among judges that democratic values do so. The most important point is that facts about what participants believe (understand, intend, and so on) could not do the necessary work because such facts are just more descriptive facts in the same position as the rest of the law practices. As with the facts about the behavior of lawmakers, we can ask whether facts about participants’ beliefs are relevant to the content of the law, and if so, in what way. Since the content of the law is rationally determined, the answers to these questions must be provided by reasons. As I have argued, the law practices, including facts about participants’ beliefs, cannot determine their own relevance. More generally, the same kind of argument explains why the questions of value on which the content of the law depends must be resolved by substantive standards rather than by value-neutral procedures. In general, there are procedural ways to resolve value questions—flipping a coin and voting are examples. Such procedures are in the same position as other law practices, however. There have to be reasons that determine that a given procedure is the relevant one and what the significance of the procedure is to the content of the law. The third objection claims that the additional substantive factors are part of law practices themselves. I have already addressed the suggestion that the law practices, conceived as facts about behavior and mental states, determine their own relevance. The present objection is that my conception is too narrow. It somehow fails to do justice to law practices to take them to consist of ordinary empirical facts about what people have done, said, and thought. If the objection is to be more than hand-waving, the objector needs to say what practices consist of beyond such facts and how the enriching factor solves the problem. For example, it would of course be no objection to my argument to claim that the descriptive facts need to be enriched with value facts. Another unpromising possibility, addressed in Section II.C above, is for the objector to maintain that law practices are legal-content-laden. According to this version of the objection, facts about what counts as a legislature, who has authority to make law, what counts as

HOW FACTS MAKE LAW

59

validly enacted, what impact a statute has on the content of the law—in general, legal-content facts concerning the relevance of law practices to legal content—are somehow part of the law practices. As argued, however, unless legal content is to be metaphysically basic, there must be an account of what determines legal content that does not presuppose it. It simply begs the question to take law practices to include legal-content facts. The objector challenges my conception of the law practices on the ground that it is too restrictive. Here is one line of thought in support of my conception. We normally assume that law practices can be looked up in the law books. But all that can be found in the law books, other than legal-content facts, are facts about what various people—legislators, judges, administrative officials, and so on—did and said and thought. If there is something else to law practices, how do we know about it? To put the point another way, if I tell you all the facts about what the relevant people said and did, believed and intended, you can work out what the law is without knowing any more about the law practices. So if there is an aspect of law practices other than these facts, it does not seem to play a role in determining the content of the law. (It is true that you may have to be skilled at legal reasoning to work out the content of the law, and that skill may include an understanding of the significance of the practices to legal content. But I have already addressed the suggestion that it is participants’ understandings, rather than the substantive factors that are the subject of those understandings, that do the necessary work.) VI. The Need for Substantive Factors Independent of Law Practices I have argued that law practices cannot themselves determine the content of the law because they cannot unilaterally determine their own contribution to the content of the law. There must be factors, independent of practices, that favor some models over others. In this section, I sketch where this argument leaves us. In particular, I explain the sense in which the argument requires facts about value, and the nature of the claimed connection between law and value.

60

MARK GREENBERG

A. Value Facts? In order for practices to yield determinate legal requirements, it has to be the case that there are truths about which models are better than others independently of how much the models are supported by law practices. Since practices must rationally determine the content of the law, truths about which models are better than others cannot simply be brute; there have to be reasons that favor some models over others. We have seen that law practices cannot determine their own contribution to the content of the law. By contrast, value facts are well suited to determining the relevance of law practices, for value facts include facts about the relevance of descriptive facts. For example, that democracy supports an intentionalist model of statutes is, if true, a value fact. What about the relevance of the value facts themselves? At least in the case of the all-things-considered truth about the relevant values, its relevance is intelligible without further reasons. If the all-things-considered truth about the relevant considerations supports a certain model of the law practices, there can be no serious question of whether that truth is itself relevant, or in what way. The significance for the law of the fact that a certain model is all-things-considered better than others is simply the fact that that model is better than others. It might be suggested that an appeal to conceptual truth offers a way to avoid the conclusion that the content of the law depends on value facts. The idea would be that the concept of law (or some other legal concept), rather than substantive value facts, determines that some models are better than others. As noted above, conceptual truth is the kind of consideration that could provide reasons of the necessary sort. The question is whether conceptual truth does so in the case of law. My response begins with two points about what notion of conceptual truth this kind of suggestion can rely on. According to what we can call a superficialist notion, conceptual truths are truths about the use of concept-words, truths that are tacitly known by all competent users of those words or are settled by community consensus about the use of the words. Given such a notion of conceptual truth, we should reject the idea that there are conceptual truths that can do the neces-

HOW FACTS MAKE LAW

61

sary work. Ronald Dworkin famously argued that disputes about the grounds of law are substantive debates, not trivial quarrels over the use of words.39 Positivists have generally responded by denying that they hold the kind of view Dworkin was attacking. Thus, both sides agree that questions about which models are better than others are not merely verbal questions that can be settled by appeal to consensus criteria for the use of words. And both sides are correct on this point. When, for example, Justices of the Supreme Court debate whether legislative history is relevant to the content of the law, the dispute cannot be settled by appeal to agreed-on criteria for the use of words. A lawyer or judge who challenges well-established models is not ipso facto mistaken. For example, a lawyer could advance a novel theory according to which New Jersey statutes make no contribution to the content of the law (on the ground, say, that there is a constitutional flaw in New Jersey’s legislative process). The claim would not be straightforwardly wrong merely because it goes against the consensus model, though it is likely mistaken on substantive grounds. Second, we have seen that the practices of participants in the legal system cannot be the source of the standards that support some models over others. It follows that if conceptual truth is to be the source of the standards, conceptual truth must not be determined by the practices of participants in the legal system; it must depend on factors independent of our law practices. The consequence of these two points is that if conceptual truth is to provide the needed standards, it would have to be conceptual truth of a kind that is not determined by consensus about the use of words and is not determined by our law practices. I am sympathetic to such a notion of conceptual truth. Given such a notion, however, it is not clear that an appeal to conceptual truth is a way of avoiding the need for substantive value facts. Instead, the conceptual truths in question may include or depend on value facts, for example, facts about fairness or democracy. At this point, the burden surely rests on a proponent of the conceptual-truth suggestion to offer a position that avoids the two problems that I have just described without collapsing into a dependence on substantive value facts. 39 DWORKIN,

supra note 6, at 31–46.

62

MARK GREENBERG

A different kind of appeal to conceptual truth is possible. It could be argued not that there are conceptual truths about which models are better than others, but that conceptual truth determines that such issues are determined by a specific internal legal value. This appeal to conceptual truth does not attempt to avoid the need for value facts; it attempts to explain those value facts as internal to the law. I will turn now to the nature of legal value facts. It is worth noting, however, that an appeal to conceptual truth as the source of internal value facts will encounter the same challenge as the appeal to conceptual truth to avoid the need for value facts. Such an appeal requires an account of conceptual truth according to which truths about the concept of law are independent of our law practices yet also independent of genuine value facts. I have argued that the content of the law depends on substantive value facts. What is the nature of those value facts? The most straightforward possibility is that, other things being equal,40 models are better to the extent that they are favored by the all-things-considered truth about the applicable considerations—the Truth, for short. In other words, the legally correct standard or value is simply the truth about value. On this view, there is no special legal standard or value. For example, the bearing of legislative history on the content of the law depends on considerations of democracy, fairness, welfare, stability—on every consideration that is in fact relevant to the issue. A second possibility is that, in the special context of the law, the allthings-considered truth about the relevant considerations is that the standard for models is not the general, all-things-considered truth about the relevant considerations but some different standard. For example, it might be that, taking into account all relevant considerations, the Truth is that the legally correct resolution of value questions is the one that maximizes community wealth. According to this second possibility, special legal value facts are genuine value facts; they are the consequence of the application of genuine value facts—Truth—to the specific context of law.41 On this view, the fact that, say, wealth 40 “Other things being equal” because practices also play a role in determining which models are better than others. See Section VI.B. 41 The position Dworkin calls “conventionalism” could be advanced as a version of possibility two, though that is not exactly the way in which he presents it. See DWORKIN, supra note 6, at 114–150.

HOW FACTS MAKE LAW

63

maximization is the virtue of models is a genuine value fact. A version of this possibility would allow the special legal value facts to vary from legal system to legal system. On the first and second possibilities, the content of the law depends on genuine value facts in a way that is inconsistent with both hard and soft positivism. A positivist might try to argue that even if my argument so far is sound, there is a third possibility. According to this possibility, there are substantive standards that within the law do the work of value facts in resolving value questions but are not genuine value facts. We might describe this possibility by saying that legal value facts are internal to the law. The hypothetical positivist’s suggestion that legal value facts are internal to the law would have to mean more than that they have no application outside of law. There could be legal value facts that were genuine value facts applicable only in the legal context. In that case, the second possibility would be actual, and the content of the law would depend at base on genuine value facts. The third possibility is supposed to avoid the conclusion that the content of the law depends on genuine value facts. Perhaps the idea would be that legal value facts matter only to those who are trying to participate in the legal system (and only to that extent). (As with the second possibility, a version of the third possibility would allow that the internal legal value can vary from legal system to legal system.) I do not mean to suggest that the idea of internal legal values is unproblematic or even fully coherent. I therefore do not need to explain exactly what it would mean for there to be internal values. Nor do I need to explain what, other than the Truth, could make it the case that there is a special legal value. I mention the idea only because it seems to have some currency in philosophy-of-law circles. My point is simply that I do not claim in this paper to have ruled out the view that the content of the law depends on internal value facts rather than genuine ones. I will briefly comment on the problems facing this view. We have already ruled out the possibility that law practices determine their own relevance to legal content. Therefore something other than law practices would have to determine the internal value standard—to make it the case that this standard was the relevant one for the law (or for the

64

MARK GREENBERG

particular legal system). It is difficult to see what that could be other than the relevant considerations—the Truth. If we appeal to the Truth, however, we have returned to the first or second possibility. Any account of internal value facts thus faces a challenge of steering between the law practices on the one hand and the Truth on the other. I have already described the way in which an attempt to ground internal legal facts in conceptual truth faces this challenge. But the challenge confronts any account of internal value facts. For example, suppose a theorist appeals to the function of law or legal systems to ground internal value facts. On the one hand, as we saw with conceptual truth, if the law’s function is going to provide the value facts necessary for practices to determine the content of the law, that function must be determined by something independent of the law practices. On the other hand, if the law’s function is determined by the all-things-considered truth about the relevant factors, an appeal to function is not a way of avoiding an appeal to genuine value facts. Until we have an account of internal value facts that meets the challenge, it is difficult to evaluate the potential of an appeal to internal value facts. An internal-value view faces a more substantive challenge as well. Internal value facts would have to have appropriate consequences for the nature of law. In a normal or properly functioning legal system, the content of the law provides reasons for action of certain kinds for certain agents. Whether the content of the law can provide such reasons may depend on the nature and source of the legal value facts. For example, it is plausible that for a legal system to be functioning properly, the content of the law must provide genuine reasons for action for judges. An internal-value theorist must explain how legal content determined exclusively by law practices and internal value facts can provide genuine, as opposed to merely internal, reasons for action. More generally, we can investigate the nature of legal value facts by asking what role such facts must play in a theory of law. B. The Role of Value Facts Let us now turn to the role of value facts in determining the content of the law. Since I do not want to beg the question against the possibility

HOW FACTS MAKE LAW

65

of a special legal value (whether internal or not), I use “X” for that property in virtue of which models are better than others. X might be, for example, (the promotion of) wealth maximization, the maintenance of the status quo, security, fairness, or morality. (If there is no special legal value, X is the Truth, in the technical sense explained above.) Note that the fact that a particular model is favored by X may be a descriptive fact (e.g., if X is wealth maximization). In that case, the relevant value fact is that X is what the goodness of models consists in. I will make two clarifications about the role of X and then consider the implications for the relation between law and value. The first point is that X only helps to determine which models are correct. X’s favoring model A over model B is neither necessary nor sufficient for A to win out over B. As we saw in Section IV, practices play a role in determining which model is better. Hence the model that is best all things considered may not be the same as the model that is ranked highest by X alone. (For simplicity I sometimes omit this qualification.) In Section IV we discussed the interdependence between models and legal content. We saw that if we hold law practices constant, different candidate models yield different sets of legal propositions. Without X, each mutually supporting pair of model and set of legal propositions is as favored as any other such pair, and indeterminacy threatens. X’s independence makes it possible for the interdependence of model and legal content not to lead to global indeterminacy. In particular, what bearing practices have on the legally correct model depends on which model is most X-justified in advance of any particular practices. For X constrains the candidate models of practices and thus makes it possible for practices to determine anything. Practices themselves have something to say about the second-order question of how practices contribute to the content of the law. But X helps to determine what practices have to say on that question. Roughly speaking, the legally correct model is the one that is most X-justified after taking into account practices in the way that it is most X-justified to take them into account.42 In other words, the legally correct model is the one that is most X-justified, all things considered. 42 In many legal systems, the practices, when taken into account in the way that is most X-justified in advance of the practices, will support a model that is not the most X-justified in advance of the practices. And when taken into account in accordance

66

MARK GREENBERG

The second point can be brought out with an objection. Suppose it is objected that X need determine only what considerations are relevant to the content of the law but need not go further and determine how conflicts between relevant considerations are to be resolved. According to this suggestion, X would eliminate some candidate models as unacceptable but would have nothing to say between models that give weight only to relevant aspects of law practices. The objector grants my argument that without an independent standard of relevance, practices could not determine which models were correct. The objector points out, however, that once we have an independent standard of relevance, practices themselves might be able to determine which models are correct. Here is a brief sketch of a reply to the objector. In order for there to be determinate legal requirements, X must do more than determine what considerations are relevant; X must favor some resolutions of conflicts between relevant considerations over others. Otherwise, given the diversity of relevant considerations and the complexity of factual variation, law practices will not yield much in the way of determinate legal requirements. Inconsistent propositions of law (and inconsistent models) will typically have some support from relevant aspects of law practices. Therefore, in order for there to be determinate legal requirements, X must not only help to determine what considerations are relevant but must also help to determine the relative importance of elements of law practices and how such elements interact. In fact there is a deeper problem with the objection. It assumes that there are discrete issues of what considerations are relevant to the content of the law and how the relevant considerations combine to determine the content of the law. It may be convenient to separate the two kinds of issues for expository purposes, but we should not be misled with that model, the practices may support yet a different model. The question therefore arises of how important it is for a model to be supported by the practices (taken into account in accordance with that model). (In the terminology of note 32 above, the more that a model is supported by the practices, the more the model is in equilibrium.) Since X is the virtue of models, X is what determines how important it is for a model to be supported by the practices. This is why it is fair to say, as I do in the text, that the legally correct model is the one that is most X-justified after taking into account the practices in the way that it is most X-justified to take them into account.

HOW FACTS MAKE LAW

67

into thinking that they are resolved separately. It is not the case that there is an initial, all-or-nothing determination of whether a type of consideration is relevant and then an independent, further determination of the relative importance of the relevant considerations. Rather, the reason that a consideration is relevant determines how and under what circumstances it is relevant and how much force it has relative to other considerations. For example, legislative history’s relevance to the content of the law derives, let us suppose, from its connection to the intentions of the democratically elected representatives of the people. Thus, in order to determine how important legislative history is relative to other factors, we need to ask exactly how it is related to the relevant intentions and what the importance of those intentions is. The point is that the contribution to content of some aspect of a law practice and how it interacts with other relevant aspects depend on why the aspect is relevant. If this suggestion—that relevance and relative importance are not independent questions—is right, then in helping to determine the relevance of various considerations, X will necessarily be (helping to) resolve conflicts between relevant considerations. I have argued that there is a certain kind of connection between law and value. I would like to conclude by saying something about the implications of this connection. Just for the purpose of exploring these implications, I will assume that X is morality. The point of this assumption is to make clear that even if morality were the relevant value, the consequences for the relation between law and morality would not be straightforward. As I will show, it would not follow that the content of the law would necessarily be morally good or even that the moral goodness of a candidate legal proposition would count in favor of the proposition’s being true. First, although (by assumption) morality provides legally relevant reasons independent of the content of the law, the legally correct model is not simply whatever model is morally best (or most justified). “Morally best” here means most supported or justified by moral considerations in advance of consideration of the practices of the legal system. The legally correct model need not be the morally best one in this sense because, as we have seen, practices also have an impact on which model is legally correct.

MARK GREENBERG

68

Second, morally good models do not guarantee morally good legal propositions. Even if the legally correct model was a highly morally justified one, the content of the law might be very morally bad. A democratically elected and unquestionably legitimate legislature could publicly and clearly promulgate extremely unjust statutes, such as a statute ostensibly excluding a racial minority from social welfare benefits. The judicial decisions may rely on highly morally justified models, ones that, among other things, give great weight to such morally relevant features of legislative actions as the clearly expressed intentions of the elected legislators. The most justified model, all things considered, will be a morally good one yet will yield morally bad legal content. In fact, in such a legal system less justified models could yield morally better legal content than more justified models. (In such cases, a judge might sometimes be morally obligated to circumvent the law by relying on the less justified model.) 43 Although morally justified models do not guarantee morally good legal propositions, it might be suggested that part of what makes a model morally justified is that it tends to yield morally good legal requirements.44 For example, assume that, other things being equal, a legal requirement is morally better the more it treats people fairly. Some models will in general have a greater tendency to yield legal requirements that treat people fairly. According to the suggestion under consideration, that a model has such a tendency would be one factor supporting that model. Suppose that the suggestion were correct. According to one line of thought, it follows that the content of the law would simply be whatever 43

The relation between a judge’s moral obligations and morally justified models raises interesting issues, but space does not permit discussion. 44 At the extreme, for example, a model could hold that in some circumstances the goodness of a candidate legal proposition tips the balance in favor of that legal proposition and against competing candidates. (A different way to describe such a position would be to say that value not only can help to determine which model is best, thus indirectly favoring some candidate legal propositions over others, but also can favor candidate legal propositions directly. I will not use this terminology.) As I say in the text, such a model may be less supported both by morality and by practices than models that give less weight to content-oriented considerations. I suggest below (see the last four paragraphs of Section VI.B), that the role that such a model assigns to value facts is outside the role that this paper’s arguments support.

HOW FACTS MAKE LAW

69

it would be morally good for it to be (or more generally, whatever it would be most X-justified for it to be). In that case the practices would be irrelevant. This line of thought might therefore be taken to provide a reductio of my argument for the role of value in determining the way in which practices contribute to the content of the law. The line of thought is not sound, however. First, even if the tendency of a model to yield morally good legal propositions counts in favor of that model, a variety of other moral considerations favor models that make the content of the law sensitive to relevant aspects of law practices. A model may be morally better, for example, to the extent that it respects the will of the democratically elected representatives of the people, protects expectations, enables planning, provides notice of the law, treats relevantly similar practices similarly, minimizes the opportunity for officials to base their decisions on controversial beliefs, and so on. Roughly, we have a distinction between content-oriented considerations and practice-oriented considerations. The relative weight accorded by morality to these two kinds of considerations is a question for moral theory that I will not take up here. On any plausible account, however, morality will give substantial weight to practice-oriented considerations. So the morally best model (considered in advance of law practices) will make the law sensitive to relevant aspects of law practices. Second, as we have seen, the legally correct model also depends on the law practices. Apart from the weight that morality gives to practice-oriented considerations, the practices themselves may support models that make the law sensitive to practices. (Contemporary positivists, my primary target in this paper, are likely to be sympathetic to the view that practices support models that make the law sensitive to practices.) For example, although I will not defend the claim here, in the U.S. and U.K. legal systems, practices themselves strongly support models that make the law sensitive to law practices. Practices are thus a second reason that the role of value need not have the consequence that the all-things-considered best model will be one that tends to yield morally good legal propositions. (Also, even a model that has a tendency to produce morally good legal propositions may not do so, given the law practices of a particular legal system.)

70

MARK GREENBERG

Third and finally, if we reflect on the argument for value’s role in determining the content of the law, we see that it supports only a limited role for value, one that does not involve supplanting law practices or making them irrelevant. Our starting point was that law practices must determine the content of the law and that they must do so by providing reasons that favor some legal propositions over others. The crucial step in the argument was that law practices cannot provide such reasons without value facts that determine the relevance of different aspects of law practices to the content of the law. The argument thus supports the involvement of value facts in determining the content of the law only for a limited role: determining the relevance of law practices to the content of the law. We can apply this point to the specific question of to what extent a legal proposition’s goodness can help to make it true: the goodness (in terms of morality or of value X) of a candidate legal proposition is relevant to the proposition’s truth only to the extent that its goodness contributes to making it intelligible that an aspect of a particular law practice has one bearing rather than another on the content of the law. I will call this the relevance limitation. I want to emphasize that the point is only that the argument of this paper supports no more than such a limited role for value facts; the argument does not show that the role of value facts must be so limited. Whether there is some other or more expansive role for value in determining the content of the law is left open. This paper’s argument for the conclusion that value facts play a role in determining legal content is that value facts are needed in order to determine the relevance of law practices to the legal content. The argument therefore supports only that role for value facts. There might, of course, be a different argument that shows, say, that morality or some other value supplants the law practices (though of course almost no contemporary legal theorist, least of all one of my positivist targets, thinks that there is such an argument). Let us consider more specifically the implications of the relevance limitation. The limitation does not imply that the goodness of a legal proposition can never be relevant to its truth.45 The goodness of a le45 It is easy to see that the goodness of a legal proposition could have evidentiary relevance to the content of the law. Suppose that the intention of legislators matters to the content of the law. If there is reason to believe that the legislators would have

HOW FACTS MAKE LAW

71

gal proposition will be relevant to the extent that it has a bearing on the intelligibility of law practices’ supporting that legal proposition over others. A Dworkinian theory of law provides a helpful example.46 Consider a model according to which law practices contribute to the content of the law precisely that set of legal propositions that best justifies those law practices. Whether this model respects the relevance limitation will depend on the notion of justification involved in the Dworkinian model. Consider a simplistic understanding of justification that has the following implication: the set of propositions that best justify the law practices is that set that results from taking the morally best set of propositions and carving out specific exceptions for the law practices of the legal system—exceptions tailored in such a way as to have no forward-looking consequences. On this understanding of justification, the model would not respect the relevance limitation, because value facts would not determine the significance of the practices; instead, the practices would simply be denied any significance by a kind of gerrymandering. On a more sophisticated notion of justification, to the extent that a legal proposition is bent or gerrymandered, it will be less good at justifying law practices. (In the extreme case just considered, where a particular law practice is simply treated as an exception without further application, that practice is not justified at all by the propositions to which it is an exception.) I think it is plausible, though I will not argue the point here, that, given a proper understanding of justification, the Dworkinian model I have described respects the relevance limitation. (Below I will consider a different model, often attributed to Dworkin, that arguably does not respect that limitation.) The relevance limitation implies that the goodness of a legal proposition is never sufficient to make it true. That value facts are needed to intended what is morally better (at least other things being equal), the moral goodness of candidate legal propositions will have a bearing on their truth because it will have a bearing on what the legislators intended. The discussion in the text concerns the question of whether the goodness of candidate propositions can have constitutive rather than evidentiary relevance. 46 I say “a Dworkinian theory” rather than “Dworkin’s theory” to avoid questions of Dworkin exegesis. I believe that the position I describe is the best understanding of Dworkin’s position. See also note 47.

72

MARK GREENBERG

determine the contribution of law practices to the content of the law does not provide a basis for making law practices irrelevant. To put it another way, that a candidate proposition is a good one does not make it intelligible that the law practices, regardless of what they happen to be, support that proposition. It might be tempting to regard a model on which the goodness of a legal proposition can, at least in some circumstances, be sufficient to make it true as the degenerate or limiting case of a model that determines the relevance of law practices to the content of the law. The model determines that in the relevant circumstances, practices have no relevance. But though this description may be formally tidy, the argument that value facts are needed to enable law practices to determine the content of the law provides no support for a model on which value facts can make practices irrelevant. In other words, though we can describe a putative “model” according to which practices provide a reason favoring any particular set of legal propositions (the morally best ones, for example), it does not follow that practices could provide such a reason. What reasons practices provide is a substantive, not a formal, question. We can apply this point to an intermediate case. Consider a model that includes rules for the contribution of law practices to the content of the law but also includes a rule of the following sort: If more than one legal proposition is supported by the (total) law practices (given the other rules of the model) to some threshold level, the legal proposition that is morally best (of those that reach the threshold) is true.47 47 Dworkin sometimes seems to suggest such a rule. See, e.g., Dworkin, supra note 6, at 284–285, 387–388; DWORKIN, TAKING RIGHTS SERIOUSLY 340, 342 (1977). And his commentators typically interpret him in this way. See, e.g., LARRY ALEXANDER & EMILY SHERWIN, THE RULE OF RULES ch. 8 (2001); John Finnis, On Reason and Authority in Law’s Empire, 6 LAW & PHIL 357, 372–374 (1987); RAZ, supra note 3, at 222–223. I think that this is not the best understanding of Dworkin’s view (and Dworkin has confirmed as much in conversation). On the best understanding, fit is merely one aspect of justification, there is no threshold level of fit, and how much fit matters relative to other aspects of justification is a substantive question of political morality. (The idea of a threshold of fit that interpretations must meet to be eligible and beyond which substantive moral considerations become relevant should be taken as merely a heuristic or expository device.) See DWORKIN, A MATTER OF PRINCIPLE 150–151 (1985); Dworkin, “Natural” Law Revisited, 34 U. FLA. L. REV. 165, 170–173 (1982); DWORKIN, supra note 6, at 231, 246–247. A

HOW FACTS MAKE LAW

73

I suggest that R is not supported by this paper’s argument for the role of value. In general, that legal proposition A has morally better content than legal proposition B does not ipso facto make it intelligible that law practices support A over B. Adding the hypothesis that law practices provide strong support for both A and B—support above some threshold level—does not change this conclusion. A moral reason for favoring proposition A over proposition B is not itself a reason provided by law practices, since it is independent of law practices. If this argument is right, my argument for the role of value facts does not support a role like that captured by R—one in which there is room for value facts to favor one legal proposition over another independently of law practices. (Again, however, the point is only that this paper’s argument does not support such a role for value facts, not that such a role is necessarily illegitimate.) In sum, even if value X were morality, it would not follow that the most morally justified model would be legally correct, and even a morally justified model would not guarantee morally good legal requirements. It is no part of the role of value argued for in this paper that the goodness of a proposition ipso facto counts in favor of the proposition’s truth. The role of value is in determining the relevance of law practices to the content of the law. VII. Conclusion I have argued that law practices, understood in a way that excludes value facts, cannot themselves determine the content of the law. Different models of the contribution of practices to the content of the law would make it the case that different legal propositions were true, and a body of law practices cannot unilaterally determine which model is different point is that Dworkin sometimes seems to suggest that there is an aspect of the question of the extent to which interpretations fit law practices that is purely formal or at least not normative. See, e.g., DWORKIN, TAKING RIGHTS SERIOUSLY, at 107 (suggesting that how much an interpretation fits is not an issue of political philosophy); see also DWORKIN, TAKING RIGHTS SERIOUSLY, at 67–68 (perhaps suggesting that there are aspects of institutional support that do not depend on issues of normative political philosophy).

74

MARK GREENBERG

correct. In order for there to be determinate legal requirements, the content of the law must also depend on facts about value. What is the role of such value facts? I have suggested that they support some models over others—that is, they help to determine which features of law practices matter and how they matter. It is not that the goodness of a candidate legal proposition counts in favor of its truth. Rather, the role of value is in helping to determine how practices contribute to the content of the law. This paper does not attempt conclusively to rule out the view that the needed legal value facts are internal to law. I have argued, however, that the proponent of such a view must overcome significant obstacles to explain how internal legal value facts could be independent of both law practices and genuine value facts. The paper also suggests a way forward: We can ask what the nature and source of legal value facts must be in order for law to have its central features, for example, for a legal system to be able to provide certain kinds of reasons for action.

3

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

On the Normative Significance of Brute Facts Ram Neta Introduction Sometimes, there are reasons for us to think or act in certain ways. We pay our taxes, we show up on time for our classes, we refuse to assent to claims that we recognize to be inconsistent, and we refrain from wanton violence, and we do each of these things because – we take it – there are reasons for us to do them. I’ll express the general point by saying that there are norms that apply to us, and to our thought and action. For a norm to apply to a person is for there to be a reason for that person to think or act in a particular way, the way indicated by the norm. Perhaps this reason is one that the person herself doesn’t in any way recognize or acknowledge. But the reason itself – whether or not it is recognized or acknowledged by the person for whom it purports to be a reason – is a norm. For instance, I might not recognize or acknowledge various norms concerning theft. But those norms still exist, and they still apply to me; there is a reason for me not to steal, whether or not I recognize or acknowledge that reason. My failure to recognize or acknowledge such reasons doesn’t make them any less real or any less applicable to me. The fact that a norm applies to someone – the fact that there is a reason for someone to think or act in a certain way – is what I’ll call a “normative fact”. All other facts I will call “non-normative”. This distinction between normative and non-normative facts has often been thought to have great metaphysical importance. In order to explain why it has been thought to have this importance, I should first draw a different distinction between two mutually exclusive and jointly ex-

RAM NETA

76

haustive kinds of facts.1 There are the “evaluative” facts, which are facts about what is good, what is bad, what is better than what, what is worse that what, and so on. All other facts are “non-evaluative”. So there are normative facts and non-normative facts (that is one exclusive and exhaustive way of dividing up the facts), and there are evaluative facts and non-evaluative facts (that is another exclusive and exhaustive way of dividing up the facts). For present purposes, we need not take a stand on how the first pair of categories is related to the second pair of categories. Now, philosophers have often been inclined to think of the world as consisting fundamentally of nothing more than the non-normative, non-evaluative facts.2 I’ll use Anscombe’s phrase “brute facts” to denote all and only those facts that are both non-normative and nonevaluative.3 Using this terminology, I will say that philosophers have often struggled to understand how the brute facts can somehow add up to normative facts of various kinds. How – they have wondered – can the brute facts make it the case that we have reason to think or act in particular ways? Typically, philosophical attempts to address this question lead to one of three results: reductionism, eliminativism, or emergentism. Reductionists attempt to show how a conglomeration of brute facts can somehow add up to a normative fact of some kind. Thus, we might try to reduce moral facts to facts about what behavior would maximize utility or fitness, epistemic facts to facts about the reliability of our belief-forming processes, semantic facts to facts about the covariation of neural events and external events, and so on. Eliminativists claim that such reduction is impossible, and so conclude 1

I hope and believe that all this talk of different “kinds of fact” is neutral with respect to various metaphysical disputes concerning facts. In any case, I seek a neutral formulation, and invite the reader to suggest one if she takes the formulation in the text to be non-neutral. 2 I will not attempt to document this historical claim, nor will I attempt to explain it. I refer the interested reader to the classic work on this topic, Burtt 1924. 3 Anscombe 1958. As Anscombe uses the term, a fact A is “brute” only relative to another fact B. She leaves it open whether there are facts that are brute relative to any other facts. She also leaves it open whether the relative bruteness of a fact has to do with its being normative or non-normative, evaluative or non-evaluative. So I’m not sure that my use of the term “brute fact” bears any significant resemblance to her use. Nonetheless, the term strikes me as both convenient and appropriately evocative.

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

77

that there really are no facts of the normative kind in question (i.e., no moral facts, no epistemic facts, no semantic facts). And finally, emergentists claim that reduction is impossible, and so conclude that the world contains facts over and above the brute facts. In his paper “How Facts Make Law”,4 Mark Greenberg offers an elegant and apparently compelling argument for emergentism about legal normative facts, or what I will call “legal emergentism”. In other words, he argues that facts about what we have legal reason to do or not to do cannot be reduced entirely to the brute facts. This is not to suggest that Greenberg thinks that the legal normative facts are metaphysically basic or primitive: he explicitly denies this. Rather, what Greenberg claims is that, while the legal normative facts might be reduced to some other facts, they cannot be reduced entirely to the brute facts. According to Greenberg, unless there were some evaluative facts, there could be no legal normative facts. In this sense, then, Greenberg’s metaphysics includes more than simply the brute facts. It also includes the evaluative facts, and it must, for Greenberg, include the evaluative facts if it is to include the legal normative facts. This is what I am calling his “legal emergentism”. Greenberg’s argument for legal emergentism seems to have very radical implications, for it suggests a more general argument for emergentism concerning all normative facts, or what I will call “normative emergentism”. If there is a sound, general argument for normative emergentism, that would be news of the very greatest importance to philosophy, for we would then know that the sparse metaphysical picture that includes nothing more than the brute facts leaves out something. If Greenberg’s argument really does give us a way to show something of this sort, then we should find this out. In this paper, I intend to find out. Specifically, I will do two things. First, I will argue that the generalization of Greenberg’s argument is not sound, and so does not establish normative emergentism. But the flaw in the generalized version of Greenberg’s argument reveals something important about his local argument for legal emergentism. And this brings me to the second goal of this paper, which is to show that the compellingness of Greenberg’s local argument for legal emergentism depends upon contingent and possibly unknown facts of legal history. If Greenberg’s 4

Greenberg. This volume.

RAM NETA

78

argument establishes its conclusion, then the fact that it establishes its conclusion cannot be known a priori. Before I begin, I must issue a brief warning about terminology. As I use the term “norm”, a norm is a reason to think or act in some way. But many philosophers – especially philosophers of law – use the term “norm” differently, so that, from the proposition that there is a norm that we V, it doesn’t immediately follow that there is a reason for us to V.5 Such philosophers might use the term “norm” to denote a certain class of actual beliefs, or statements, or sentences, or intentions, or practices. They might then sensibly ask whether there is any reason to do what certain legal norms enjoin us to do. Now, I use the term “norm” in a different sense than do these philosophers; I cannot sensibly ask whether there is any reason to do what a legal norm enjoins us to do: if there is a norm that enjoins us to do it, then it immediately follows – on my use of “norm” – that there is reason to do it. This may strike some readers as perverse stipulation on my part, but by way of excuse I note that there is no way to avoid some stipulation in the use of the term “norm”: it is a term of art, and different authors use it to mean different things. Furthermore, in using the term as I do, I follow what I take to be Greenberg’s usage. Since he is my chief target in this paper, it seems to me sensible to follow his usage. Greenberg’s Argument for Legal Emergentism In this section, I’ll state Greenberg’s argument for legal emergentism. First, I’ll briefly summarize Greenberg’s explanations of the terminology that he uses in his argument: “Legal decisions” are decisions that legislators, judges, and other people make, as well as other legally relevant historical events that can be fully characterized in brute terms. (I leave aside the difficult but irrelevant issue of what’s involved in being able to understand a historical event “fully”: however we end up settling that issue, Greenberg can then define “legal decisions” in the appropriate terms.) “Legal content” is the normative content of the law, i.e. what is legally forbidden or required, or more generally, what legal norms there are. 5

This is true of Hart, Finnis, and Raz, among others.

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

79

I should note that Greenberg distinguishes legal content from semantic content: the former is what is forbidden or required by a law, whereas the latter is what is meant by a particular sentence or thought. Using this distinction between legal content and semantic content, we can say that it is open question – and one that we don’t need to discuss here – whether the legal content of a particular law is identical to the semantic content of a sentence or thought that expresses, or lays down, that law. “Legal propositions” are propositions articulating legal content. If the entire legal content of a particular legal system were articulated in legal propositions, then the semantic content of those legal propositions would fix the entire legal content of that particular system. It is possible, though, for some of the legal content of a particular legal system to be unarticulated. From its being the law that P, it doesn’t immediately follow that anyone has articulated any legal proposition that means that P.6 Finally, to say that one thing A “provides reason for” another thing B is to say that A makes it the case that B, and makes it the case in a way that makes it at least somewhat reasonable for it to be the case that B. Equivalently, we can say that A “rationally determines” B. To illustrate: there is nothing reasonable or unreasonable about the fact that water boils at 212 degrees Fahrenheit, and so whatever makes it the case that water boils at 212 degrees Fahrenheit does not rationally determine that fact. Nothing rationally determines the fact that water boils at 212 degrees Fahrenheit. In contrast, it is at least somewhat reasonable for the law to require that people who are not convicted of crimes not receive punishment. It’s not just a fact that the law requires this, but it is a reasonable fact. Thus, whatever makes it the case that the law requires this, provides a reason for the law to require it, and so rationally determines that the law requires it. These examples should provide one with a general sense of how Greenberg is using the terms “provide a reason” and “rationally determine”. Admittedly, I have not given a rigorous account of these notions, but then Greenberg doesn’t offer a rigorous account either, and I’m following his practice for now in order to state his argument. It will turn out that 6

point.

I am grateful to two referees for Legal Theory for pressing me to clarify this

80

RAM NETA

his argument is subject to criticism no matter how precisely these notions are explicated. Using the terminology above then, here is Greenberg’s argument: (1) Premise D: In the legal system under consideration, there is a large body of determinate legal propositions. (2) Premise L: The legal decisions in part determine the content of the law. (3) The legal decisions can determine the content of the law only by providing reason for the content of the law being what it is. (In other words, the legal decisions can determine the content of the law only by rationally determining it.) (4) The legal decisions provide reason for the content of the law being what it is. (From 2, 3) (5) In a legal system with a large body of determinate legal propositions, the legal decisions by themselves cannot fully determine what reason they provide for the content of the law being what it is. (That is, the legal decisions cannot fully determine exactly how they rationally determine the content of the law.) (6) Something else besides the legal decisions must help to determine what reason the legal decisions provide for the content of the law being what it is. (From 1, 4, 5) (7) Evaluative facts are the only things that can play the role of helping to determining what reason the legal decisions provide for the content of the law being what it is. (8) Evaluative facts enter into determining the content of the law in the legal system under consideration. (From 6, 7) If this argument is sound, then all attempts to reduce legal content to the history of legal decisions, or to the semantic contents of written and spoken texts, or to what judges had for breakfast, must fail. For instance, all versions of legal positivism and legal realism fail. None of the brute facts can, by themselves, fully rationally determine the content of the law. That’s because they can enter into determining the content of the law only by providing reason for the law being what it is. But these facts cannot fully rationally determine their own rational significance. That is, they cannot fully rationally determine what reason they have the power to provide. And that would be true no matter how

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

81

broadly we extend this range of facts, so long as they exclude the evaluative facts and the facts about legal content. That totality of facts still could not fully determine its own rational significance for the content of the law. And so we would, according to Greenberg, need to add something to it in order fully to determine the content of the law. Here’s one intuitive way to think about Greenberg’s thesis: The law requires us to act in all sorts of determinate ways, and it forbids us to act in all sorts of determinate ways. But there must be some reason for the law to require some things and forbid other things – the law’s requirements, unlike the boiling point of water, are either reasonable or unreasonable, and something makes them either reasonable or unreasonable. The brute facts by themselves cannot provide reason for the law to issue these determinate requirements. The brute facts cannot, on their own, makes the legal requirements either reasonable or unreasonable. So there must be something over and above the brute facts that provides reason for the law to issue these determinate requirements. And this extra factor is the value facts. Greenberg thus argues that no version of legal reductionism can be right. If we grant premise 2, and so reject legal eliminativism, Greenberg makes it look as if we must accept legal emergentism. The Generalization of Greenberg’s Argument Part of what makes Greenberg’s argument so important is that it suggests a more general argument concerning all normative systems of any interest whatsoever – moral, epistemic, semantic, and so on. To see this, let’s consider whether the premises of Greenberg’s argument for legal emergentism apply more generally. Premise 1 says that in the legal system under consideration, there is a large body of determinate legal propositions. But something analogous will be true of any interesting normative system. For instance, any interesting moral code will include a large body of determinate moral requirements. Any interesting methodology will include a large body of rules for theory-choice. Any interesting linguistic system will include a large body of semantic rules. And so on. So it seems that we can generalize premise 1 as follows:

82

RAM NETA

Premise 1’: In the normative system under consideration, there is a large body of determinate normative propositions (hereafter, “norms”). Premise 2 of Greenberg’s argument says that certain brute facts – here, the legal decisions – in part determine the content of the law. But again, it seems that something analogous will be true of any interesting normative system. For instance, the normative content of a moral code depends to some extent upon various brute facts about the creatures to which the code applies (e.g. that they are mortal, that they are capable of suffering pain, that they can communicate, they are susceptible to certain kinds of temptation, and so on). The normative content of a particular scientific methodology depends to some extent upon various brute facts about the theoretical practice and practitioners to which that methodology applies (e.g. that they have certain sense organs and not others, that they are capable of making certain sorts of calculations easily but others only with great difficulty, that their sense organs can be trained to respond reliably to certain ranges of energies and not others, and so on). And the normative content of a particular linguistic system depends to some extent upon various brute facts about the creatures that employ that linguistic system (e.g. that they have a certain universal grammar hard wired, that they have learned to speak a SVO language instead of a SOV language, that they have learned to pronounce certain phonemic combination and not others, and so on). So it seems that we can generalize premise 2 as follows: Premise 2’: The brute facts in part determine the norms. Now it may be objected against premise 2’ that there are some normative facts that are metaphysically basic: they do not depend on any contingent features of the creatures to whom they apply. For instance, one might think that the categorical imperative, or modus ponens, is such a rule. I will not contest such claims here (even though I do think that they are false).7 But even if they are true, we can still accept that premise 2’ holds for all normative systems of which premise 1’ is true, 7 Briefly, the categorical imperative is not a reason for anyone to do anything: rather it is a constraint upon something’s being a good practical reason. Again, modus ponens is not a reason for anyone to think anything, but, in tandem with the laws of logic, places a constraint upon what it is for something to be a good theoretical reason.

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

83

i.e., all normative systems that generate many determinate norms. The categorical imperative, all by itself, doesn’t generate many determinate norms. It can only generate many determinate norms when it’s conjoined with lots of contingent facts about the features of the actual agents to whom it applies. The same holds of any other allegedly necessary and basic normative fact. So, even if there are necessary and basic normative facts, this does not threaten the substance of premise 2’. Premise 3 of Greenberg’s argument says that the legal decisions can determine the content of the law only by providing reason for the content of the law to be what it is. Recall that the phrase “providing reason” is here being used to signify a metaphysical relation: X provides reason for Y just in case X makes Y obtain and also makes it reasonable for Y to obtain. In this sense, nothing provides reason for water to boil at 212 degrees Fahrenheit: it just does boil at that temperature. But something does provide reason for the law to require that people not convicted of a crime not receive punishment. Instead of using the phrase “providing reason” to designate this metaphysical relation, we might equally well use the phrase “make it reasonable”. Now, we might worry that these examples do not give us a firm grasp on the metaphysical notion of “providing reason” or “making it reasonable”. Greenberg tells us a bit more about this notion, at least in its application to the law. Here’s what he says: “The basic idea is that the content of the law is in principle accessible to a rational creature who is aware of the relevant law practices. It is not possble that the truth of a legal proposition could simply be opaque, in the sense that there would be no possibility of seeing its truth to be an intelligible consequence of the law practices. In other words, that the law practices support these legal propositions over all others is always a matter of reasons – where reasons are considerations in principle intelligible to rational creatures.” I’m eventually going to raise a question about how to interpret this passage. But for now, I’ll allow the passage to stand without comment, and I’ll ask whether something analogous to this passage could be said about the determination of norms in other kinds of normative system. For instance, do the brute facts that enter into determining the norms of a moral code make it reasonable for the norms to be what they are? Do the brute facts that enter into determining the norms of a

84

RAM NETA

particular methodology make it reasonable for the norms to be what they are? Do the brute facts that enter into determining the semantic norms of a particular linguistic system make it reasonable for the norms to be what they are? I think it’s not entirely clear how to answer these questions. But I shall now argue that there is at least some plausibility in answering each of them affirmatively. Here’s my argument: If the brute facts do not make it reasonable for the norms to be what they are, then either nothing makes it reasonable for the norms to be what they are, or else the reason for the norms to be what they are includes something independent of the brute facts. Let’s consider what follows from each of these two hypotheses. If the first hypothesis is right, then nothing makes it reasonable for the norms to be what they are. In that case, it’s arbitrary that the norms are what they are, i.e., there is nothing reasonable about the norms being what they are rather than some other way; they just are that way. But if there is nothing reasonable about the norms being what they are, then those norms are like the rules of a game that there is no reason to play, or the rules of a practice that there is no reason to participate in. That is not merely to say that some particular person has no reason to follow those norms. It is rather to say that there is no reason to follow those norms. But recall that a norm, as I’m using the term here, is just a reason: if there is a norm that says to think or act in a certain way, then that is just for there to be a reason to think or act in that way. If there is no reason to follow a norm, then it is not really a norm at all. But by hypothesis, we are talking about norms. Therefore, the first hypothesis cannot be right, and so something must make it reasonable – at least more or less reasonable – for the norms to be what they are. There cannot be norms which are such that nothing makes it reasonable for them to be what they are.8 This line of reasoning will give rise to two objections: first, it may be objected that it generates an infinite regress of norms. But this isn’t so. We can avoid the infinite regress either by appeal to a big circle of 8 I remind the reader that the term “norm” is used in different ways in the literature, and on some of these uses, the reductio argument that I’ve just given will seem to be a non sequitur. It’s important, then, in going through this argument, to recall that, as I use the term “norm”, it is just a reason to think or act in a certain way.

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

85

norms, or by appeal to some foundation that makes it reasonable for the norms to be as they are, but is not itself a norm (i.e., is not itself a reason for thinking or acting in any particular way). We needn’t choose between these strategies here. Second, it might be objected that, many of the norms that actually do apply to us are arbitrary, but are, for all that, still norms. For instance, it may be said, we have reason to follow the particular linguistic norms of our language community, even though those norms are arbitrary. Again, the law gives us reason to stay within the speed limit, or to drive on a particular side of the road, even though that is arbitrary as well. The problem with this objection is that these are not cases of there being reason to follow arbitrary norms. Rather, they are cases in which the specific norms that there is reason to follow are determined by more general non-arbitrary norms, together with various non-normative contingencies. For instance, there is a general norm to the effect that we should speak in such a way as to make ourselves understood. But this general norm makes it reasonable for us to comply with the more specific norms of our linguistic community, whatever those happen to be. Again, there is a general norm to the effect that one should promote social coordination. This general norm makes it reasonable for us to comply with the laws of our land, whatever those happen to be (at least within limits). In each of these cases, one has a reason to do some specific thing by virtue of one’s reason to do some more general thing, along with the contingencies of one’s particular situation. These are not cases of having a reason to do something for no reason at all. And so these cases do not invalidate the principle used in the preceding argument: if nothing makes it reasonable for a norm to be as it is, then there is no reason to follow the norm, and so it is not really a norm at all. I conclude that the first hypothesis cannot be right. If the second hypothesis is right, then the reason for the norms to be what they are is independent of the brute facts that determine those norms. In that case, it’s arbitrary that the norms are binding on all and only those creatures of which the determining brute facts obtain. For instance, it’s arbitrary that the norms are binding on mortal creatures who are capable of feeling pain, rather than on angels. And in that case, again

86

RAM NETA

there’s nothing that makes the norms in question binding on creatures like us. And that is just to say that, while the norms might really be norms for some creatures, they are not really norms for us. There is nothing about us that engages the binding power of the norms. They are like rules of a game that creatures like us have no reason to play. On the strength of these considerations, I provisionally conclude that neither the first nor the second hypothesis can be right. (I will raise a question about this conclusion below, when I reconsider the question of what it is for one thing to “provide reason” for another thing.) It seems then, that for any normative system of which premises 1’ and 2’ are true (i.e., any normative system that has determinate norms that are binding on us), the brute facts that determine the norms of that system must provide reason for us to comply with those norms, and so provide reason for those norms to be what they are. If the preceding line of thought is correct (and, as I said, I will re-examine it below), then we can generalize premise 3 as follows: Premise 3’: The brute facts can determine the norms only by providing reason (i.e., making it reasonable) for the norms being what they are. Premise 5 of Greenberg’s argument says that, in a legal system with a large body of determinate legal propositions, the legal decisions by themselves cannot fully determine what reason they provide for the content of the law being what it is. There are many possible mappings from legal decisions to legal content, and the legal decisions, according to Greenberg, cannot themselves determine which possible mapping is the correct one. That’s why other facts, besides the legal decisions, are needed to determine the correct mapping. Now, so far as I can see, whether or not premise 5 is true depends upon nothing that is peculiar to legal normativity. Whatever it is that makes it the case that legal decisions cannot determine the correct mapping from themselves onto the facts of legal content, that same thing makes it the case that brute facts cannot determine the correct mapping from themselves onto the normative facts. If there is supposed to be something special about the determination of legal norms in this regard, it’s not at all clear what it could be. This suggests that, if premise 5 is true, then so is.

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

87

Premise 5’: In a normative system with a large body of determinate norms, the brute facts by themselves cannot fully determine what reason they provide for the norms being what they are. Finally, premise 7 says that evaluative facts are the only things that can play the role of helping to determine what reason the legal decisions provide for the content of the law being what it is. Greenberg explains this point in the following passage: “In order for practices to yield determinate legal requirements, it has to be the case that there are truths about which models are better than others, independently of how much the models are supported by law practices. Since practices must rationally determine the content of the law, truths about which models are better than others cannot simply be brute; there have to be reasons that favor some models over others.” The reason that favors some models over others includes facts about which models are better. And these are evaluative facts. But for the same reason that evaluative facts are needed fully to determine the correct mapping from legal decisions onto legal norms, so too, it seems, evaluative facts will be needed fully to determine the correct mapping from brute facts onto other norms generally. If premise 7 is true, then so is Premise 7’: Evaluative facts are the only things that can plausibly play the role of helping to determine what reason the brute facts provide for the norms being what they are. With all this in mind, we can now consider the following generalization of Greenberg’s argument: (1’) Premise D: In the normative system under consideration, there is a large body of determinate norms. (2’) Premise L: The brute facts in part determine the norms. (3’) The brute facts can determine the norms only by providing reason for the norms being what they are. (4’) The brute facts provide reason for the norms being what they are. (From 2’, 3’) (5’) In a normative system with a large body of determinate norms, the brute facts by themselves cannot fully determine what reason they provide for the norms being what they are.

88

RAM NETA

(6’) Something else besides the brute facts must help to determine what reason the brute facts provide for the norms being what they are. (From 1’, 4’, 5’) (7’) Evaluative facts are the only things that can play the role of helping to determine what reason the brute facts provide for the norms being what they are. (8’) Evaluative facts enter into determining the norms in the normative system under consideration. (From 6’, 7’) By generalizing Greenberg’s argument, we’ve constructed an argument for normative emergentism. The argument assumes (premise 2’) that normative eliminativism is false, but from this assumption it argues against normative reductionism. What’s Wrong with this General Argument for Normative Emergentism? I will now argue that this general argument for normative emergentism is not compelling, for either premise 5’ is false or else we have no reason to accept premise 3’. I’ll begin by leveling an objection against premise 5’, and then I’ll argue that the only way to save premise 5’ from this objection is to interpret the notion of “providing reason” in such a way that we have no reason to accept premise 3’. Premise 5’ says that in a normative system with a large body of determinate norms, the brute facts by themselves cannot fully determine what reason they provide for the norms to be what they are. But let’s consider whether or not this general claim is borne out by cases. Of course, any discussion of such cases – any discussion of why we have reason to do one thing or another – is bound to be controversial. I cannot hope to find a case that will uncontroversially tell for or against the generalization of Greenberg’s argument. But we can test that argument against what seem to us to be plausible views concerning familiar examples of normativity. To this end, I will now briefly consider the normativity of promising. Suppose that Alice mowed the lawn because John promised to pay her 400 dollars if she mowed the lawn. Now, what makes it wrong for John to break his promise? It seems to me that any plausible answer to

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

89

this question will involve the following claim: what makes it wrong for John to break his promise in this case is that it would be a case of breaking a promise, and it’s generally wrong to break promises. That’s an essential feature of the practice of promising: making a promise places one under an obligation to keep it (except in very special circumstances). The practice of promising essentially involves norms that are binding on all those who participate in that practice, and these norms include the norm that it’s wrong to break a promise. When I say that the practice of promising “essentially” involves these norms, I mean that it’s not just true by convention that the practice of promising involves the norm that it’s wrong to break one’s promise. For instance, it’s not just that we happen to use the word “promising” to designate practices that have this feature. Rather, the practice of promising essentially involves the norm in question because there’s no way for the practice of promising to exist over time without that norm. If people were generally permitted to break their promises, then people would know (or soon come to know) this fact, and so would not expect other people to keep their promises, and so would tend not to act on any such expectation. Everyone would soon come to know all of this, and so no one would continue to expect her own promises to have any impact. Each person would thereby lose any motivation to make promises. In short, the institution of promising would quickly cease to exist if people were generally permitted to break their promises. For there to be an institution of promising, it must be impermissible – and known to be impermissible – for people to break their promises. Thus, it’s an essential feature of the practice of promising that it involves the norm that it’s wrong to break your promises. 9 But this just pushes our original question back a step. Now, instead of asking why John must keep his promise, we can ask why John has reason to participate in this essentially norm-governed practice of 9 Again, although this point about promising seems very plausible to me, I recognize that it is not uncontroversial. (For an instance of an opposing view, see Ardal 1968.) I don’t know of any examples that will serve my dialectical purposes at this point and that are entirely uncontroversial. But I hope and believe that, in order to motivate the present dilemma to the generalization of Greenberg’s argument, I needn’t rely on there being any uncontroversial examples of norms fully determined by the brute facts. It should suffice if I can display the epistemic possibility that there are such examples.

90

RAM NETA

promising. And there could be any number of answers to this question. For instance, John might have an interest in having his lawn mowed, and he recognizes that the only way that he can get his lawn mowed is by promising some able-bodied person that he’ll pay them if they mow it. Or John might, like some children, simply enjoy participating in a social practice that affords him opportunities for market interactions with others. But whatever the story, so long as John has some reason to participate in the practice of promising, he has reason to comply with the norms of that practice, and so they are norms for him. Now let’s consider whether I have indeed provided “the reason”, in Greenberg’s sense, for why it’s wrong for John to break his promise. We can assess this issue by considering what it is for a normative fact to obtain “for a reason”, on Greenberg’s use of that phrase. Paraphrasing the passage from Greenberg quoted above, we can say this: for norms to obtain for some reason is for the contents of those norms to be in principle accessible to a rational creature who is aware of the relevant brute determinants of the norms. It is not possible that the obtaining of those normative facts could simply be opaque, in the sense that there would be no possibility of seeing them to be intelligible consequences of the relevant brute determinants. In other words, that the relevant brute determinants support these normative propositions over all others is always a matter of considerations in principle intelligible to rational creatures. Now, it seems that, in the story that I’ve told about promising, I’ve shown how the wrongness of John’s breaking his promise follows from the general prohibition against breaking promises. But that general prohibition is an intelligible consequence of various brute facts about the practice of promising. And John’s reasons for engaging in that practice are themselves intelligible consequences of various brute facts about John (e.g., what he likes and doesn’t like). Thus, it seems, I’ve articulated the reason why it’s wrong for John to break his promise. I’ve provided an explanation of what makes it wrong for John to break his promise, and my explanation adverts solely to brute facts about the social function of the institution of promising, and about John. If my explanation is correct, then premise 5’ is false. (And if you disagree with my explanation of what makes it wrong for John to break his promise, then I invite you to consider whether there isn’t another case that would falsify premise 5’.)

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

91

Of course this is not because of something special about promising. Whether or not my story about promising is correct, it seems that many other norms may be equally susceptible of explanation in terms of brute facts. In order to rescue premise 5’ from all such apparent counterexamples, the proponent of the argument above seems to have only one way out, and that is to claim that these explanations are not explanations of the right kind at all – they do not explain what reason there is for the norms being what they are. Now, I’m not sure how strong a case can be made for or against this way of avoiding the objection that I’ve just leveled, and that’s because I’m not sure exactly what’s involved in the metaphysical constitutive relation that Greenberg uses the term “providing a reason” to designate. What exactly is involved in the relation between normative facts and their brute determinants being “a matter of reasons” or “in principle intelligible to a rational creature”? Since I’m not sure how to answer these questions, I will not attempt to argue against this proposed way of avoiding my objection to premise 5’. Instead, I’ll point out that it does the proponent of the argument above no good. For if my aforestated explanation of why it’s wrong to break a promise doesn’t “provide a reason” (in the relevant sense) for why it’s wrong to break a promise, then I don’t see why we should think that anything else “provides a reason” (in the relevant sense). In other words, if I haven’t provided a reason for why it’s wrong to break a promise, then why should we think that there is any reason (in the relevant sense) for why it’s wrong to break a promise? Why shouldn’t we just think that the wrongness of breaking a promise is a metaphysically basic fact of the world? Or it obtains not by virtue of something else providing a reason for it, but rather by virtue of something making it obtain, in some non-rational way? Now recall that we have already considered a plausible argument against these apparent possibilities: if there is no reason for the norms to be as they are, then we have no reason to comply with those norms, and so they are not really norms. Now, I do find something plausible about this argument, when the notion of “reason” that it employs is broad enough to include the kind of reason that I gave above for why it’s wrong to break a promise. But if the notion of “reason” is interpreted more narrowly than that, then the argument seems to me to lose

92

RAM NETA

its plausibility. Why should there be a “reason”, in this special, narrow sense, for why it’s wrong to break a promise? If there is no “reason”, in this special narrow sense, for why it’s wrong to break a promise, then how does it follow that we have no “reason”, in a broad, ordinary sense, for not breaking our promises? Pending an answer to this question, there seems to be nothing to favor premise 3’. In sum, if we grant that premise 5’ is false, then we have no reason to believe premise 3’. Either way, the general argument above for normative emergentism is not compelling. Maybe normative emergentism is true, but its truth is not proven by the general argument constructed in the previous section. Now what, if anything, does this show about Greenberg’s local argument for legal emergentism? The Consequences for Greenberg’s Local Argument for Legal Emergentism In the preceding section, I argued that the generalization of Greenberg’s argument is not compelling: either premise 5’ is false (for the brute facts do fully determine that John must pay Alice 400 dollars), or else we have no reason to believe premise 3’ (for we don’t know enough about the rational determination relation). I will now argue that an analogous, but somewhat weaker, conclusion is true of Greenberg’s local argument for legal emergentism: either premise 5 is subject to historical falsification (for all we know a priori, the legal decisions may be such as to fully determine their own rational significance), or else we have no reason to believe premise 3 (for we don’t know enough about the rational determination relation). Recall that premise 5 says that in a legal system with a large body of determinate legal propositions, the legal decisions by themselves cannot fully determine what reason they provide for the content of the law being what it is. Now, in defending this premise, Greenberg considers a foundationalist challenge to it and a coherentist challenge to it. I would like to focus on the foundationalist challenge, and Greenberg’s response to it. Here is the relevant passage: “If practices are to determine which model is correct, there are two possibilities. First, a privileged foundational practice (or set of foun-

ON THE NORMATIVE SIGNIFICANCE OF BRUTE FACTS

93

dational practices) could determine the role of other practices. This possibility encounters the problem of how practices themselves can determine which practices are foundational. ...A putatively foundational practice cannot non-question-beggingly provide the reason that it is foundational.” Now, I’d like to ask why a putatively foundational decision cannot provide the reason that it is foundational. Suppose that, when the framers drafted the American Constitution, they had included a clause that stated explicitly and precisely how the content of the law was to depend upon the legal decisions. Couldn’t their decision to include this clause be a foundational decision, and provide the reason why it is foundational? Perhaps Greenberg would object that their decision to include this clause could not have non-question-beggingly provided the reason why it is foundational: if their decision is foundational, that’s just because the decision says that it is foundational, and so its foundational character is founded in a question-begging way. But then I ask: why must the reason why a decision is foundational be non-question-begging, in this sense? I can imagine, on Greenberg’s behalf, the following response to this question: suppose that there are two putatively foundational but inconsistent decisions. In that case, neither decision can provide a reason why it, rather than the other, is really the foundational decision. And so neither decision can be foundational, except by dint of the help of some additional factor. In that case, each decision has its rational significance for the content of the law only by dint of the help of this additional factor. I’ll grant that this is true in the case in which we have two putatively foundational but inconsistent decisions. But I don’t see why it should also be true for the case in which we have only one putatively foundational decision, the dictates of which are consistent with all of the legal decisions. In short, I don’t see why there cannot be a genuinely foundational decision that provides the reason why it is itself foundational. Of course, different philosophical accounts of law will provide us with different understandings of what sort of event can qualify as a genuinely foundational decision, and how such a foundational decision can provide the reason why it is itself foundational.

RAM NETA

94

But we do not need to canvass these differences here in order to appreciate the force of the general challenge to Greenberg’s argument. Now, it is open to Greenberg to reply to this general challenge as follows: in the case that I’ve described, what we have is a legal decision that metaphysically determines the correct model, but not a legal decision that provides a reason why one model is correct. Since I don’t have a fully firm grip on the notion of “providing a reason”, I will not object to this response. But then, if that is the response, I wonder what reason we have to think that decisions determine the content of the law by “providing reason” (in the relevant sense) for that content being what it is. So my challenge to Greenberg stands as follows: If, as a matter of contingent historical fact, the framers had included a clause explicitly stating precisely how legal decisions were to determine legal content, and this clause was consistent with all other legal decisions, then either premise 5 of Greenberg’s argument is false, or else we don’t have any reason to accept premise 3. Either way, the compellingness of Greenberg’s argument depends on a matter of contingent, and possibly unknown, fact of legal history. 10 Works Cited Anscombe, G.E.M. 1958. “On Brute Facts”. Analysis 18: 69-72. Ardal, Pall. 1968. “’And That’s a Promise’.” The Philosophical Quarterly 18: 225-237. Burtt, E.A. 1924. The Metaphysical Foundations of Modern Physical Science. Doubleday: New York. Greenberg, Mark. Forthcoming. “How Facts Make Law.” Legal Theory.

10

I am grateful to Mark Greenberg, Doug Lavin, Eric Marcus, Ori Simchen, and two anonymous referees for Legal Theory for helpful discussion of an earlier draft of this paper, and for much useful discussion of the relevant issues. I am also grateful to Enrique Villanueva for organizing the very useful conference for which this piece was written.

4

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

On Practices and the Law1 Mark Greenberg In a recent article, “How Facts Make Law” (Greenberg 2004; hereafter HFML), I launch an attack on a fundamental doctrine of legal positivism. I argue that nonnormative facts cannot themselves constitutively determine legal facts. In response, Ram Neta (this volume) defends the view that nonnormative social facts are in general sufficient to determine2 normative facts, including moral and legal facts. Neta’s paper provides a useful opportunity to address a spelled-out version of this view, which in various forms is widely held in philosophy of law and other areas of philosophy. 3 Some terminology will be helpful. A legal fact4 is a true legal proposition – a fact about the content of the law. That contracts for the sale of land must be in writing is a legal fact in many legal systems. A normative fact is a true normative proposition, such as a true proposition about what is good or bad, right or wrong, fair or 1 I am grateful to Ram Neta for his rich and interesting paper, to which I am not able to do full justice here. Thanks to Pamela Hieronymi, Kinch Hoekstra, Herb Morris, Stephen Munzer, Ram Nea, Keemin Ngiam, Scott Shapiro, Seana Shiffrin, and Nicos Stavropoulos for comments on drafts and helpful discussions. 2 Throughout, unless otherwise specified, the relevant kind of determination is constitutive, not modal. For more on constitutive and modal determination, see Greenberg 2005. 3 For examples of writers who hold that nonnormative social facts are sufficient to determine legal facts, see Coleman (2001, chaps. 7, 11); Hart (1994, chaps. 5-6); Himma (2002); Raz (1983, chaps. 3-4; 1994); Shapiro (1998). As noted below, not all of these writers would agree with Neta that legal facts are normative facts. See below note 7 and accompanying text. For examples of writers who hold that nonnormative facts are sufficient to determine normative facts (outside the legal domain), see Hart (1994, chaps. 4-5, pp. 256-257); Gilbert (2000, chap. 5); Schapiro (2003). 4 In HFML (p. 158), I used the uglier term “legal-content fact.”

96

MARK GREENBERG

unfair.5 A descriptive fact is a nonnormative, contingent fact. Lawdetermining practices, or law practices for short, are descriptive facts – paradigmatically facts about the attitudes, sayings, and doings of members of constitutional assemblies, legislatures, and courts – that are constitutive determinants of the legal facts. I call the relevant facts “law-determining practices” rather than “legal decisions” because the term “decisions” tends to suggest judicial decisions in particular. (In this somewhat unidiomatic usage, law practices are facts.)6 One clarification may be helpful at the start. Unlike some legal positivists, Neta holds that legal facts are normative facts. In particular, he holds that part of what it is for a legal standard to apply to someone is for there to be a reason for that person to act in accordance with the standard.7 Indeed, a main theme of Neta’s paper is that the le5 I use the term “normative facts” here in the way that I used “value facts” in HFML. Thus, I use the term to include what are sometimes called “evaluative facts,” such as facts about what is good or bad. For further explanation of the notion of a normative fact, see below note 7 and HFML, p. 167 n. 22. It is worth noting that some nonnormative facts are not descriptive facts, for descriptive facts, as defined in the text immediately below, do not include necessary truths. Therefore, even if descriptive facts cannot alone determine the content of the law, it does not immediately follow that normative facts must play a role. See HFML, pp. 165-166, 187-188. 6 For elaboration of the notion of a law practice, see HFML, Sections II.B. & II.C.. For ease of exposition, I use legal-content-laden terms, such as “legislature” and “court” in characterizing law practices. Strictly speaking, however, the law practices should be understood to be the underlying descriptive facts in virtue of which the relevant legal facts obtain. Since legal facts are not basic, there must be non-legal facts that constitute the legal-content-laden practices. These more basic facts will include descriptive facts—in particular, the facts that I am calling “law practices.” For example, the fact that a legislature enacted a statute must hold in virtue of complex descriptive facts about people’s attitudes and behavior and perhaps also normative facts. (If, in order to account for legal-content-laden practices, we have to appeal not merely to descriptive facts but also to normative facts, so much the worse for the positivist thesis that the content of the law depends only on descriptive facts.) The convenience of talking as if law practices consisted in legal-content-laden facts about the behavior of legislatures, courts, and so on should not obscure the fact that there must be more basic facts in virtue of which the legal facts obtain. 7 See Neta (2004, pp. 75-76). Neta defines “normative fact” as a “fact that there is a reason for someone to think or act in a certain way” (p. 75). See also text accompanying notes 22-24 below. As I am using the term, the category of normative facts is broader than the category captured by Neta’s definition. (For one thing, as

ON PRACTICES AND THE LAW

97

gal domain and a range of normative domains are on a par. He takes a view about all normative facts (or at least about many kinds of normative facts) that is parallel to the legal positivist’s view about legal facts. The project of HFML and the present paper is to argue that normative facts determine legal facts without assuming that legal facts are themselves normative facts. A consequence of these terms of engagement is that my case against Neta’s view is more straightforward outside the legal domain, for example in the case of moral norms. In the case of non-legal norms, I simply argue that Neta’s account fails to explain the existence of reasons for action. In particular, I show that there are many cases in which his purported sufficient conditions for the existence of norms are satisfied, though there is no reason to act in accordance with the putative norms. In the legal case, I forswear a parallel strategy because, as noted above, my goal is to show that normative facts are constitutive determinants of legal facts without assuming that legal facts are normative facts. In HFML, I therefore offer a different kind of argument – an argument driven by a claim about the nature of constitutive determination in the legal domain. In the present paper, we see that the argument of HFML provides leverage in the legal domain that is neither available nor necessary in the moral domain. 1. Rational Determination I begin by summarizing my argument in HFML and Neta’s response. My argument starts from the relatively uncontroversial premise that descriptive facts are among the constitutive determinants of the content of the law. What is controversial is whether descriptive facts can alone determine the content of the law or whether normative facts must also play a part. I argue that a full account of what makes the law have the content that it does will have to appeal to normative facts. remarked in note 5 above, I include what are sometimes called evaluative facts among the normative facts.) But any fact that is constituted in part by the fact that there is a reason for someone to think or act in a certain way is certainly a normative fact. Any differences between Neta’s and my understanding of the term will not matter for purposes of this paper.

98

MARK GREENBERG

A variety of domains – the aesthetic, the political, the legal – are plausibly not metaphysically or ontologically basic. That is, the facts of those domains are constituted by more ontologically basic facts. In each case, the obtaining of the more basic facts makes it the case that the facts of the target domain obtain. For example, perhaps facts about the arrangement of paint on a surface make it the case that a painting is elegant or clumsy. I take for granted that legal facts are not a metaphysically basic aspect of the world (HFML, pp. 158-159). In the legal case, the more basic determining facts include law practices – facts about the attitudes and behavior of various people – and may or may not include various normative facts as well.8 The nature of the making-it-the-case or constitutive determination relation may vary from domain to domain. To make progress on the central issue of the determinants of legal content, we need to investigate the nature of the determination relation in the legal domain. I claim that this relation is what I call rational determination. Suppose that we are trying to give a constitutive account of the facts of a target domain. Suppose further that we are able to specify constitutive determinants of the target facts that modally entail those target facts. In some domains, this may be all that a constitutive account can do. The constitutive determinants need not provide reasons why the target facts obtain. A reason, in the relevant sense, is a consideration that makes an explanandum intelligible in rational terms (as opposed to, say, emotional or aesthetic ones) – or, as I will sometimes write, that makes an explanandum rationally intelligible.9 By contrast, if the nature of the determination relation in a domain is rational determination, the determining facts must make it rationally intelligible that the facts of the target domain are thus and so. We can call this the rational-relation requirement.10 8 To be precise, for a given domain, there may be more than one (sufficient and non-redundant) set of determining facts at different levels of ontological basicness. For example, in the legal case we could consider the relation between microphysical facts and legal facts, rather than the relation between facts about behavior (and attitudes) and legal facts. In the present paper, I focus on the relation between facts about behavior and legal facts. See also note 12 below. 9 For more on the notion of a reason, see HFML, pp. 164-166. 10 In HFML, I use the term rational-relation doctrine. That certain facts provide reasons for the target facts is only a necessary, not a sufficient, condition for their

ON PRACTICES AND THE LAW

99

Let us consider an example of a domain in which there need be no reasons of the relevant sort. I mentioned above the view that aesthetic facts obtain in virtue of physical facts. A painting is elegant or clumsy in virtue of the arrangement of paint on the canvas.11 But arguably the facts about the arrangement of paint need not provide a reason for the painting’s elegance or clumsiness (though those facts will no doubt be part of a causal, for example psychological, explanation of our reactions). For example, a small difference in the arrangement of paint might make a clumsy scene elegant, without providing a reason for the difference. (Imagine trying to explain the impact on the painting’s beauty to a rational Martian or to a fully rational human being who lacked the relevant sensitivity.) The claim that the relevant determination relation in law is rational determination is a claim that the relation is different in this respect from the relation between physical and aesthetic facts.12 There must be reasons why the relevant facts, for example about decisions of legislatures and courts, have the effects on the law that they do. (This metaphysical truth is reflected in the requirement that judges must articulate reasons for their decisions.) For example, it could not be a brute fact – one that cannot be rendered rationally intelligible – that a particular change in the wording of a statute produces a particular change in the content of the law.13 rationally determining those target facts. See HFML 163-164; Greenberg 2006, section II. 11 On another possible view, facts about the artist’s intentions are also determinants of the aesthetic facts. On that view too, it is plausible that the determining facts need not provide reasons for the obtaining of the aesthetic facts. 12 It might be asked whether, for a given domain, the nature of the relation between determining facts and target facts could be different for different sets of determining facts at different levels. My view, which I will not argue for here, is that rational determination requires only that the legal facts be intelligible in light of some set of non-legal facts. At any rate, as noted above, when I discuss the nature of the determination relation in the legal domain, I mean to consider the relation between, on the one hand, facts about behavior and attitudes and, on the other, legal facts. 13 This point will not make sense unless one is careful to distinguish between, on the one hand, legal texts (and the meaning of those texts) and the content of the law. Neither a statutory text, nor its meaning, is the content of the law, though they may be determinants of the content of the law. See HFML, Section III.

100 MARK GREENBERG

Two clarifications will be important for what follows. First, rational intelligibility is a relation between the determinants of legal facts and the legal facts, not a monadic property of the legal facts.14 The obtaining of particular legal facts, as opposed to others, must be rationally intelligible in light of the determining facts. The rational-relation requirement imposes no requirement on the content of the law independent of the determining facts. For example, it does not demand that the content of the law be reasonable or rational. Second, the rational-relation requirement does not specify that the reasons it requires must be normative facts. The reasons that the determinants of legal content must provide are considerations that explain in rational terms why the law has the content that it does. As far as the rational-relation requirement is concerned, it is an open question whether there are nonnormative facts that could constitute such reasons – and indeed whether there are normative facts that could do so.15 (The point is important because otherwise positivists could not accept the rational-relation requirement. The strategy of the argument is to use the rational-relation requirement, which I claim most legal theorists implicitly accept, to argue for the controversial conclusion that normative facts must be determinants of legal content.) The present paper creates a special risk of confusion with respect to this point because I also use the term “reason” in another context. A crucial issue in the debate between Neta and me concerns reasons for actions, and in fact I distinguish two types of reasons for action.16 I therefore want to emphasize that the reasons that the rational-relation requirement demands are not reasons for action. Again, they are considerations that explain in rational terms why particular legal facts obtain.17 14

See HFML, p. 165. See HFML, pp. 165-166. My ultimate view, of course, is not that normative facts could themselves provide such reasons, but that both nonnormative and normative facts are needed. Law practices are essential. 16 See section 2 below. 17 I do not mean to suggest that the term “reason” is ambiguous. As noted below, Pamela Hieronymi (MS) helpfully offers an account of a reason as a consideration that bears on a question. See below note 31. In these terms, the reasons that the rational-relation requirement demands are considerations that bear on the constitutive 15

ON PRACTICES AND THE LAW 101

In HFML, I sketch some arguments for the proposition that the relation between law practices and legal facts is rational determination, at least in the U.S. and U.K. legal systems and perhaps in all legal systems,18 though I do not attempt anything like a full defense of that proposition. I also give reasons for thinking that the proposition is fundamental to our ordinary understanding of the law and taken for granted by much contemporary legal theory, though seldom articulated (HFML, pp. 170-172). 19 I then argue that law practices cannot themselves rationally determine legal facts, or indeed norms of any kind. In brief, the argument is that law practices cannot unilaterally determine their own relevance. Factors independent of law practices must be part of what makes their contribution intelligible in rational terms. In the case of law, the best candidates for such factors are normative facts. (In other domains, there may be other candidates (see HFML, p. 161, Section VI.A.).) For example, facts about democracy and fairness explain the relevance of the publicly promulgated decisions of elected legislators to the content of the law.20 question of why particular legal facts, rather than others, obtain. The two types of reasons for action that I distinguish in section 2 below are considerations that bear on the question of what to do (“normative reasons”) and considerations that bear on the question why a person acts in a particular way (“motivating reasons”). 18 To be more precise, all those in which the legal facts are determined by law practices. 19 I am not especially concerned with the question of whether rational determination is the determination relation in all possible legal systems. Readers who believe that rational determination is the determination relation only in some legal systems can understand my arguments as applicable only to those legal systems. 20 The rational-relation requirement demands that there be reasons for legal facts, not that there be reasons for the relevance of law practices to legal facts. So why do I say that normative facts explain the relevance of law practices to legal facts? An intuitive summary of why law practices cannot themselves provide reasons for legal facts is that there are many possible ways in which law practices could bear on the legal facts, and they cannot determine their own relevance. For this reason, I sometimes express the rational-relation requirement by saying that there have to be reasons that determine the contribution of law practices to the legal facts. This way of putting things should not mislead us into thinking that the rational-relation requirement is a requirement not only that there be reasons for the legal facts, but that there be reasons that explain why those reasons are reasons. That line of thought could suggest that an explanatory regress lurks. What the rational-relation requirement demands is not

102 MARK GREENBERG

Neta’s leading idea is that descriptive facts21 can alone provide a full account of normativity. He first argues that, in all normative domains, descriptive facts in part determine the content of the norms, and the relevant determination relation is rational determination. Next, he tries to show that a full account of moral norms need appeal only to descriptive facts. In particular, he offers an account of the wrongness of promise-breaking entirely in terms of descriptive facts. He concludes that it does not follow from the fact that rational determination is the relevant determination relation that normative facts must play a role. Neta also gives a purported counterexample in the legal domain – a case in which he claims that descriptive facts can themselves rationally determine the content of the law. I now turn to Neta’s attempts to show that descriptive facts can alone determine norms in the moral and legal domains. First (section 2), Neta’s account of why it is wrong to break promises fails. It begs the question by taking for granted that a person’s desires or other motivational states necessarily justify the actions that they motivate. Second (section 3), I turn to Neta’s alleged legal counterexample. Here I draw on the notion of rational determination to move the argument beyond an appeal to intuitions. I show that the facts to which Neta appeals cannot be reasons of the required sort, for they already take for granted what needs to be explained – the bearing of law practices on the content of the law. Also, Neta fails to consider the possibility that normative facts are playing a silent role in his example. Finally, in section 4, I turn to the larger issue of how far my argument applies to other normative domains.

higher-order reasons, but determining facts that together provide reasons for the legal facts. In Greenberg 2006, I discuss objections that claim that if normative facts are needed to explain the relevance of law practices, an infinite regress is generated. See Greenberg 2006, section VI, especially note 57 and accompanying text. 21 Neta uses “brute fact” instead of “descriptive fact.” For the notion of a descriptive fact, see above p. 1. It is important to note that in HFML and here, I use “brute fact” in a way different from Neta’s. See above text accompanying notes 13 and HFML, pp. 160, 171. To avoid confusion, I replace Neta’s “brute fact” with “descriptive fact” in quotations from his article (using square brackets to indicate the change).

ON PRACTICES AND THE LAW 103

2. Neta’s Account of Promising Neta uses the case of promising to argue that the full metaphysical explanation of normative facts need appeal only to descriptive facts. He offers an example in which “Alice mowed the lawn because John promised to pay her four hundred dollars” and tries to explain “what makes it wrong for John to break his promise.” (p. 88).22 According to Neta, “any plausible answer” involves the claim that “what makes it wrong for John to break his promise ...is that it would be a case of breaking a promise, and it’s generally wrong to break promises” (pp. 88-89). This first step appeals directly to a normative fact – that “it’s generally wrong to break promises.” The success of Neta’s argument depends on whether he can explain, without appealing to any normative (or evaluative) facts, why it is generally wrong to break promises.23 Let’s call the moral requirement that one not break promises – the requirement that Neta is trying to explain – “the promising norm.” Before proceeding, we need to clarify one terminological matter. As I will use the term “norm,” a necessary condition for there to be a norm is that there actually be reasons for agents to act in the way required by the norm. (Occasionally, I will use “genuine norm” rather than merely “norm” as a reminder of my usage.) Other writers sometimes use the term in a purely descriptive way according to which, roughly speaking, for there to be a norm in a particular group requiring that people act in a certain way, it is sufficient that members of the group think and act as if there were a genuine norm requiring that people act in that way. Consider the putative norm of a crime gang that its members must commit an assassination once a year. In my usage, the fact that the gang members accept the putative norm – regularly comply with it, enforce it by punishing those who fail to comply, believe that they are justified in doing so, and so on – does not make it the case that the putative norm is a norm. Neta is very clear that he uses “norm” in my sense.24 I

22

Page numbers standing alone are references to Neta (this volume). When Neta writes that it is “generally wrong to break promises,” I assume that he means that it is wrong to break promises without good enough reasons, or that it is wrong prima facie or pro tanto. For brevity, I will follow his way of talking. 24 See Neta (this volume, pp. 78, 84 & n. 8, also p. 75). 23

104 MARK GREENBERG

will continue to use “putative norm” to express the purely descriptive sense of “norm.”25 Neta’s attempted explanation appeals crucially to the existence of “the practice of promising.” In what follows, I show that Neta’s practice-based explanation of the wrongness of breaking promises doesn’t work. I want first to note a problem with the whole approach, however. Neta seems to assume, as many writers do, that promising could not take place in the absence of a particular kind of social practice or convention.26 In my view, this is a mistake. A Robinson Crusoe who has never made a promise and has no practices or conventions with regard to promising can, when he encounters a castaway for the first time, promise to help him (and it would be wrong for Crusoe to break that promise without good enough reasons). The fact that breaking promises is wrong in the absence of any practice supports the view that the existence of a practice is not what accounts for the wrongness of breaking promises even in cases where there is a practice. On this view, which I do not attempt to defend here, the appeal to practices is misguided from the start. But for present purposes, I consider Neta’s practice-based explanation on its own terms. Neta’s attempted explanation begins with the claim that “[t]he practice of promising essentially involves” the promising norm (p. 89). This claim might be taken to suggest that he is using the notion of a practice in such a way that the existence of a practice is constituted in part by the obtaining of normative facts. If that were the case, it would be question-begging to appeal to a practice as a necessary part of the explanation, for the goal is to explain normative facts in nonnormative terms. (See HFML, Sections II.C, V). Neta soon makes clear, however, that the claim is that “the institution of promising would quickly cease to exist if people were gener-

25 Other pairs of terms, such as “social rule”/“normative rule” and “effective norm”/“valid norm” have been used to mark distinctions in the general neighborhood of the distinction that I explain in the text. I prefer my terminology because it does not suggest that norms and putative norms are two species of a single genus. 26 There is a large literature on promising. Three recent papers, which provide citations to the literature, are Seana Shiffrin (MS); David Owens (2006); Jay Wallace and Nico Kolodny (2003).

ON PRACTICES AND THE LAW 105

ally permitted to break their promises” (p. 89).27 There is no suggestion here that the existence of the practice is constituted by normative facts. Rather, when Neta claims that the norm is essential to the practice, he means that as a causal or historical matter, if people did not treat promises as creating obligations, the practice would cease to exist. We need to be clear what exactly is necessary for the practice to exist. Is it that breaking promises be wrong? Or is it that people believe that it is wrong, criticize others for breaking their promises, think it is appropriate to criticize others for this reason, and so on? Neta uses phrases that are ambiguous with respect to this point. That it is “permitted” to break promises could mean either that people permit each other to break promises with impunity (i.e., that they act as if promising were not wrong) or that breaking promises (even without good reason) is in fact morally permissible. What is required in order for people’s expectations not to be disappointed and thus for the practice to continue is that many people believe it is wrong to break promises, not that it actually be wrong. In fact, even this formulation seems too strong: what is necessary is not that people believe that it is wrong to violate the norm, but that people act as if they believe it is wrong – or, to oversimplify in order to save words, that people generally comply with the promising norm. Whether in practice people will not comply with the norm unless they believe that it is wrong to violate it (as opposed to believing, for example, that others believe it is wrong and will punish violations) is an empirical question. In sum, all that Neta can take to be necessary for the promising practice is that people often act in accordance with the putative promising norm, not that it be wrong to violate it. Neta sees that this kind of dependence of a practice on a putative norm does not suffice to show that violating the putative norm is wrong, i.e., that it is a genuine norm. But he thinks that all that needs to be added to complete the account is that a person has a reason to participate in the practice: 27 We need a qualification similar to that of the previous note. I assume that Neta means that the institution would quickly cease to exist if people were generally permitted to break their promises without good enough reasons, or if people did not even count the existence of a promise as a prima facie or pro tanto reason for acting in accordance with it.

106 MARK GREENBERG

[That the norm is essential] just pushes the original question back a step. Now, instead of asking why John must keep his promise, we can ask why John has reason to participate in this essentially norm-governed practice of promising. And there could be any number of answers to this question. For instance, John might have an interest in having his lawn mowed, and he recognizes that the only way that he can get his lawn mowed is by promising some able-bodied person that he’ll pay them if they mow it. Or John might, like some children, simply enjoy participating in a social practice that affords him opportunities for market interactions with others. But whatever the story, so long as John has some reason to participate in the practice of promising, he has reason to comply with the norms of that practice, and so they are norms for him (pp. 89-90).

At this point, Neta takes himself to have established what he set out to show – that it is wrong to break promises (p. 90). In short, Neta’s argument is that if a person (believes he) could obtain something he wants28 or enjoys by participating in a practice that would not exist if people did not generally comply with certain putative norms, then those putative norms are genuine norms, and it is wrong, even morally wrong, not to comply with them. Neta’s account cannot be right. There are many silly, corrupt, or evil practices that would “cease to exist” if the participants did not generally comply with certain putative norms. The practices of operating fraudulent investment schemes, hazing new recruits, and selling children into sexual slavery are examples. People can obtain things they want by participating in these practices. It hardly follows that it is wrong, even pro tanto, not to comply with the putative norms of the practices. 28 I interpret John’s having “an interest in having his lawn mowed” as John’s wanting his lawn mowed (or wanting other things that require having his lawn mowed), as opposed to some evaluative fact, e.g., that it is good for John to have it mowed. Neta is supposed to be appealing only to descriptive facts. Moreover, if Neta intended to appeal to the goodness of John’s lawn being mowed, as opposed to John’s motivation, there would be no need for him to appeal to John’s recognition “that the only way that he can get his lawn mowed is by promising some able-bodied person that he’ll pay them if they mow it.” (p. 90).

ON PRACTICES AND THE LAW 107

It is useful to develop one example a little more fully. Consider the protection racket – organized crime’s practice of providing “protection” for restaurants, shops, and other businesses in return for regular payments. It is essential to the existence of the practice that many of the lower-level criminals who carry out the work of extorting protection money follow putative norms requiring them to use violence against shop owners who fail to pay protection money, and requiring them not to inform on their bosses. Similarly, it is essential to the continued existence of the practice that many of the small business owners follow putative norms requiring them to pay on time and not to complain to the police. Therefore, the protection-racket practice “essentially involves” the putative norms that it is wrong to refuse to pay protection money and that it is wrong to report the organizers to the police in exactly the sense in which Neta claims that the promising practice essentially involves the promising norm.29 The criminals who extort the protection fees can obtain things they want by participating in the practice – they earn a living and avoid being punished by their bosses. Even the business owners have self-interested reasons for participating in the practice. They avoid having their shops burned down or worse, and they may in fact receive some protection from petty criminals and other benefits from being on good terms with the organized crime gangs. According to Neta’s position, it is enough that they have “some reason to participate,” not that the all-things-considered balance of reasons supports participating (though in some cases it may do so). It hardly follows that the putative norms of the practice are genuine norms, let alone that it would be wrong not to comply, for example, by reporting the organizers to the police or refusing to pay protection money. Where does Neta’s argument go wrong? First, it is not warranted to move from John’s wanting to have his lawn mowed to John’s having 29 It might be suggested that it matters to Neta’s account not only that participants comply with putative norms but that they believe that it would be wrong not to do so. In response, we may suppose that it is essential to the protection racket that many of the lower-level criminals believe that it would be wrong of them to refuse to extort money, to report their bosses to the police, and so on. We could even suppose that it is essential to the continued existence of the practice that many of the small business owners who pay for protection believe that it would be wrong of them not to pay. It still would not follow that it would be wrong for anyone to violate the putative norms.

108 MARK GREENBERG

reason to comply with the essential norms of the practice of promising. That someone could obtain something he wants by participating in a practice does not show that he has even a narrowly self-interested reason for complying with all the putative norms necessary for the practice to exist. One might get what one wants by making promises without keeping them. Perhaps in the case of promising it is plausible that one must generally comply with the promising norm in order to be able to use promises effectively to get what one wants. It is easy, however, to imagine practices such that one can obtain their benefits through partial or apparent participation, i.e., without complying with all of their putative norms. Moreover, at most John’s wanting to have his lawn mowed gives him a reason for keeping his promises when not doing so will hurt his reputation (or hurt his reputation more than the benefit from breaking it). It is a commonplace that there may be cases in which it is evident that breaking a promise will not harm his reputation. For instance, John might make an unwitnessed promise to a dying friend moments before the friend dies. Neta might respond that John’s behavior does not really count as promising unless he intends to keep his promises, believes that it is wrong to break them, and so on. In order to promise, one must be participating in the practice, and the promising norm is “an essential feature” of the practice.30 As emphasized above, however, Neta’s argument that the promising norm is essential establishes only that the practice wouldn’t exist if many or most people didn’t comply with (or believe in) the putative norm. It doesn’t follow that John or any particular person must believe it is wrong to break promises or intend to keep his promises in order to count as promising. But suppose that we waive this point and grant that one does not count as promising unless one accepts the promising norm. In that case, however, Neta has not shown that John has reason to promise, only that he has reason to make apparent or pseudo-promises. For pseudo-promises may well be enough for John to get what he wants. Hence, even given the assumption that one does not count as promising unless one accepts the promising norm, it does not follow, so far as Neta has argued, that John has 30 As noted above, in my view, one need not be participating in a practice in order to make a promise. For purposes of argument, I set this point aside.

ON PRACTICES AND THE LAW 109

any reason to accept the promising norm. Neta here runs into a problem that dogs attempts to show that it is rational to act morally. That one would be worse off if no one complied with a putative norm does not show that one has reason to comply (even if others do comply). Neta is not free to assume away the free-rider problem. A second problem with Neta’s argument is the move from John’s having a reason to comply with a putative norm to its being a genuine norm that is applicable to him (“he has reason to comply with the norms of that practice, and so they are norms for him”). This step in the argument seems to depend on an equivocation on the term “reason.” Michael Smith helpfully distinguishes between “normative reasons,” which are considerations that count in favor of, or justify Fing, and “motivating reasons,” which are factors that (potentially) motivate someone to F. (Smith 1994, pp. 14, 94-98).31 Neta’s examples of John’s reasons are John’s wanting to have his lawn mowed and his enjoyment in participating in the practice. Neta takes it as obvious that these are reasons, and offers no further explanation. Consequently, in pointing out that John has reasons to engage in the practice, Neta must be using the term to mean motivating reason. For the fact that John has a desire that would be satisfied by Fing or enjoys Fing shows immediately that John has a motivating reason to F but, without further argument, does not show that he has a normative reason to F. But the plausibility of moving from John’s having a reason to comply with a putative norm to its being a genuine norm depends on John’s reason being a normative reason. That someone believes that he can get what he wants by Fing cannot be sufficient to show that there is a norm requiring him to F. Since the issue is whether normative facts can be accounted for entirely in nonnormative terms, it obviously begs the question to move without argument from the existence of a moti31 For convenience, I will use Smith’s terminology in this paper. But my point does not depend on accepting his view of the underlying issues. It suffices for my purposes that a desire or other factor that explains an action does not necessarily justify the action or count in favor of it. Pamela Hieronymi (MS) gives an illuminating account of reasons, according to which there is only one kind of reason. On her view, the relevant distinction is between two types of questions on which reasons can bear. For my purposes, we can be neutral between Smith’s, Hieronymi’s, and other accounts. For helpful discussion of different accounts, see Hieronymi (MS).

110 MARK GREENBERG

vating reason – a descriptive fact – to the existence of a normative reason – a normative fact. It might be objected that the term “reason” is not simply ambiguous. Many motivating reasons are, or involve, normative reasons. For example, at least one kind of motivating reason to act in a particular way is in part constituted by a belief that there is a normative reason to act in that way. There are certainly important connections between motivating reasons and normative reasons. But because motivating reasons need not be normative reasons, an explanation of the existence of a normative reason must do more than point to the existence of a motivating reason. The critical question is whether it is necessary to appeal to normative facts as well. Even when a motivating reason constitutes a normative reason, the question remains whether normative facts are part of the explanation of what makes the motivating reason a normative reason. Taking for granted that no explanation is needed to get from the existence of a motivating reason to the existence of a normative reason begs the question of whether the wrongness of promise-breaking can be explained without appeal to normative facts. It might be wondered whether Neta’s position could be defended by appeal to a substantive theory according to which any desire or other state that (potentially) motivates someone’s taking a certain action counts in favor of the person’s taking that action. In the first place, such a theory is highly implausible. Moreover, the theory makes a normative claim, a claim about what kinds of things provide justifications. So an account of normativity that relies on the theory does not restrict itself to descriptive facts. At this stage, an obvious suggestion is that only good practices, or practices that it is right to engage in, generate genuine norms. The problem, again, is that the suggestion explains normativity only by appealing to normative facts – facts about which practices are good or which courses of action are right. There are further problems with Neta’s account. That it is wrong to break promises (even prima facie or pro tanto) is a stronger claim than that there is some normative reason for not doing so. This is not the place to attempt a full account of exactly how it is stronger. One type of case will suffice. I have reasons to have annual medical tests.

ON PRACTICES AND THE LAW 111

It does not follow that it is wrong for me not to do so, not even pro tanto. Certainly, it doesn’t follow that it is morally wrong.32 Furthermore, it would not make a difference if there were a practice to which the putative norm of annual medical tests was essential in Neta’s sense. In fact, even if we granted for the sake of argument that all motivating reasons were normative reasons (and assumed that normative facts were not an essential part of the explanation of this proposition), Neta’s account would seem to lack the resources to explain specifically moral wrongness. John’s wanting to have his lawn mowed, even if it is a normative reason, is the wrong kind of material with which to explain the moral wrongness of his breaking promises. And since Neta’s project is to explain the wrongness of breaking promises without appealing to normative facts, Neta obviously cannot appeal to moral facts. 3. Neta’s Legal Counterexample I now turn to Neta’s purported counterexample to my claim that normative facts are needed to determine the contribution of law practices to the content of the law – or, as I will say, to determine which model (of the contribution of law practices to the content of the law) is correct.33 He supposes that the framers of the U.S. Constitution “had included a clause that stated explicitly and precisely how the content of the law was to depend upon the law practices” (p. 93). In the last section, we were able to undermine Neta’s claim that nonnormative facts are sufficient to determine normative facts simply by considering the evidence of actual and hypothetical cases. Examples like that of the protection racket are difficult to reconcile with the claim that practices (and motivating reasons to participate in them) 32 Even moral reasons to F do not necessarily make it wrong not to F. For there can be moral reasons to take actions that are supererogatory. For example, I have moral reasons to give nine-tenths of my income to charity. It does not follow that it is morally wrong for me not to do so, not even pro tanto. 33 For more on the term model, see HFML, pp. 178-179. The legally correct (or, for short, correct) model is the way in which practices in a given legal system actually contribute to the content of the law (not merely the way in which they are thought to do so).

112 MARK GREENBERG

necessarily generate normative reasons. (It is for similar reasons that Hart’s critics were able so conclusively to dispatch Hart’s practice theory of rules as a general theory of moral and other norms.) 34 As noted, Neta assumes that legal facts constitute normative reasons for action.35 Given that assumption, it would be straightforward to mount an argument against Neta’s legal counterexample parallel to my argument against his account of normative facts generally. That a practice involves a document that explicitly and precisely (and without contradiction from other documents) specifies how the practice is to generate rules does not ensure that the practice generates normative reasons for action. As noted in the Introduction, however, I want to avoid relying on an assumption that legal facts constitute reasons for action (or, more generally, that legal facts are normative facts), so as not to beg any questions against legal positivists who do not accept the assumption. Consequently, my argument that descriptive facts cannot themselves determine legal facts must take a different form than my argument that descriptive facts do not determine moral and other normative facts. HFML introduced a new way of answering the question whether descriptive facts can themselves determine legal facts. This approach appeals to an argument about the properties the determinants of legal facts must have. In particular, I argue in HFML that in law, in contrast to many other domains, the determinants must make the facts of the target domain rationally intelligible. In Neta’s example, it is plausible, though by no means obvious, that the explicit constitutional clause would be foundational – that it would determine the relevance of other law practices to the content of the law, and that its contribution to the content of the law would not be affected by other law practices. The relevant question, however, is whether the existence of that clause (and other descriptive facts, such as the absence of competing clauses) could be the full explanation of why that clause was foundational. As we will see below, an important possibility, which Neta does not consider, is that normative facts are part of the explanation. For example, values of democracy, stability, 34 See Dworkin 1977, pp. 48-58; Raz 1999, pp. 53-58. See also Hart 1994, pp. 255-256. 35 See above note 7 and accompanying text.

ON PRACTICES AND THE LAW 113

and fairness support giving great weight to what the ratifiers of the Constitution and the public believed the ratifiers to be agreeing on. Should we accept Neta’s view that the descriptive facts of his hypothetical case are sufficient to determine that the imagined constitutional clause is foundational? Or do normative facts do part of the work? We can adjudicate between these and other competing accounts of legal facts by asking whether their proposed determinants satisfy the rational-relation requirement. Let us begin by distinguishing two very different approaches that a positivist could take in attempting to show that normative facts are not necessary to satisfy the rational-relation requirement. According to the first approach, the explanation of why descriptive facts are able to satisfy the rational-relation requirement is special to the legal case. Such an explanation appeals to an alleged basic truth about the nature of law (or perhaps a conceptual truth about law). An example of such an alleged truth would be the Hartian thesis that officials’ “acceptance”36 of a given rule of recognition makes that rule the correct model in the community’s legal system. A theorist might argue that understanding this thesis – grasping the explanatory significance of officials’ acceptance of a rule – is part of understanding what law is. In “Hartian Positivism and Normative Facts” (Greenberg 2006; hereafter “Hartian Positivism”), I consider an attempt to defend a Hartian account of law – the most influential contemporary version of legal positivism – in this way. I argue that the Hartian thesis is not part of our reflective understanding of law. 37 The second approach makes no appeal to specifically legal truths. Ordinary descriptive facts, for example about the behavior and attitudes of people, are sufficient to satisfy the rational-relation requirement. Neta’s view falls in this second camp, as he makes no appeal to specifically legal truths and in general assimilates the legal domain to other normative domains. In discussing Neta’s example here, we should therefore set aside any ideas we might have of basic truths about law. The question is whether the ordinary empirical facts about 36

Hart 1994, pp. 55-57, 255. I say “a Hartian account” rather than “Hart’s account” because I try to address the most powerful and plausible version of a position in the neighborhood of Hart’s, rather than to be faithful to the details of Hart’s own view. 37

114 MARK GREENBERG

the sayings and doings of various people are themselves able to satisfy the rational-relation requirement.38 In HFML (pp. 180-181), I considered the possibility of a “foundational practice” – one that determines the bearing on the content of the law of other practices, but whose own bearing does not depend on that of any other practice. I argued that a putative foundational practice could not by itself rationally determine that it was in fact foundational. More fundamentally, it could not rationally determine its own significance to the law. Neta concedes that if there were “two putatively foundational but inconsistent practices,” some additional factor would be needed to make intelligible why one practice rather than the other is foundational. He focuses, instead, on a case in which “we have only one putatively foundational [practice], the dictates of which are consistent with all of the law practices.” In this case, Neta suggests that the sole putatively foundational practice can provide the reason “why it is itself foundational” (p. 93). Neta’s description of the case takes three elements as given. First, the practice in question is putatively foundational. Second, no other practice is putatively foundational. Third, the content of the putatively foundational practice (its “dictates”) is consistent with the content of all other law practices. Stipulating these three elements begs the question of whether the descriptive facts can themselves determine which model is correct. For these three elements are not descriptive facts, but facts that are logically posterior to what is at issue – the correct model in the legal 38 Another version of the second approach – the one that makes no appeal to specifically legal truths — would appeal to truths about the way in which social facts determine rules in a domain or class of domains that is broader than and includes the legal domain. I consider this approach and comment on the obstacles it faces in Greenberg 2006 (section III). As I explain there, legal positivists have not seriously pursued this approach. (It is important to note that H.L.A. Hart’s (1994) account of law does not treat legal rules as an instance of a more general case of social rules. See, e. g. Hart (1994, p. 256). For on Hart’s account, roughly speaking, what makes a standard a legal rule is not that it is accepted, but that it is specified by an accepted rule. See Greenberg 2006, text accompanying notes 30 and 34.) Neta does not attempt to show that legal norms are an instance of the (putative) norms of practices that he discusses, and such an attempt would not be plausible.

ON PRACTICES AND THE LAW 115

system. For example, for a law practice to be the only putatively foundational practice in a legal system, it must be the case that no other law practice has a bearing on what the putatively foundational practice contributes to the content of the law. Whether that condition is satisfied depends, however, on which model is correct. Similarly, whether the putatively foundational practice is consistent with all other law practices depends on the legal significance of all the law practices. And, again, the legal significance of law practices depends on which model is correct. In sum, the elements on which Neta relies are not the descriptive facts to which he is entitled, but facts that take for granted a model. What has gone wrong with Neta’s argument? It is a descriptive fact that only one legal practice in the legal system has associated with it a sentence asserting that the practice has a foundational role. And it is a descriptive fact that no other legal practice has associated with it an inconsistent sentence. Neta seems to equate these descriptive facts with the facts he stipulates – that there is only one putatively foundational practice, and that its content is not in conflict with the content of any other practice. Thus, he implicitly assumes that the significance of a law practice is no more and no less than the ordinary non-legal content of the sentences associated with it. We know that this assumption is false, however. For example, the fact that judicial decisions have interpreted other law practices in a particular way bears on the contribution of law practices to the content of the law, regardless of whether the interpreting judicial decisions explicitly say anything about how law practices are to be interpreted. More generally, as I argue at length in HFML, the contribution of law practices to legal content is not constituted by the conjunction of the contents of sentences associated with those practices (HFML, Section III). The significance of law practices depends on reasons, and there are reasons for law practices to contribute the content of the law in ways that do not track the meaning of sentences associated with those practices. Another problem with Neta’s account is that he seems to assume that it is a matter of contingent historical fact whether a constitutional clause of the sort he describes would rationally determine that the clause was foundational. He says that the example of the constitutional

116 MARK GREENBERG

clause shows that my account is “historically falsifiable” and “dependent on a matter of contingent, and possibly unknown, fact of legal history”. This claim seems to confuse the question of whether the clause would be foundational with whether the descriptive facts (of the existence of the clause and the absence of a conflicting practice) are the full explanation of its being foundational.39 As noted above, Neta neglects the possibility that normative facts play a role in explaining the bearing of his hypothetical constitutional clause. Again, it is plausible that the clause would have foundational significance. But the best explanation may not be simply that the clause says so. There are powerful reasons, for example of democracy and fairness, for giving great weight to the plain meaning of the constitutional text. In order to see that such reasons may be playing a silent role in Neta’s example, consider how much less effective the example would be if the single putatively foundational clause were in a decision of a minor administrative official. Neta’s explanation of the foundational nature of his clause may legitimately appeal only to purely descriptive facts. Given the important normative considerations that support the Constitution’s special status, it is difficult to be sure that our evaluation of the example is influenced only by the descriptive facts. 40 It might be objected that the problem with Neta’s example is that constitutional requirements are themselves part of the content of the law and therefore cannot ultimately determine how practices determine the content of the law. Perhaps a version of Neta’s example would fare better if it involved not a constitutional provision, but the attitudes and dispositions of legal officials that, on a Hartian account, make a rule of recognition the correct model for a legal system.41 As noted above, in “Hartian Positivism,” I consider whether a Hartian account can satisfy the rational-relation requirement. I show that the arguments of HFML yield the same conclusion with respect to officials’ acceptance of a rule of recognition as with respect to Neta’s constitutional clause. 39 Another problem is that it is a subtle legal issue, not a straightforward matter of descriptive fact, whether a clause is foundational. 40 For discussion of a related objection, see Greenberg, this volume, pp. 137-141. 41 I am grateful to an anonymous referee for Legal Theory for the suggestion that I consider this case.

ON PRACTICES AND THE LAW 117

4. Other Normative Domains I have thus far criticized Neta’s attempts to explain moral and legal facts entirely in terms of nonnormative facts. The relevance of the moral domain to my original paper depends on a more general claim that Neta makes. The claim is that my argument generalizes to all normative domains: more precisely, all normative domains have the structural features of the legal domain that are necessary to generate my original argument. In order to defend the argument of HFML against Neta, I don’t need to deny this claim, for I have already argued against Neta’s attempt to explain moral norms in nonnormative terms. But the issue of how far my argument generalizes is of independent interest. In this final section, I first consider and reject Neta’s argument that purports to show that all normative domains have the relevant features of the legal domain. I then offer a competing picture of some other normative domains. A crucial premise of my original argument is that descriptive facts are among the determinants of the legal facts, and the determination relation between the determinants and the legal facts is rational determination (see HFML, section IIB). A version of the argument can be developed only in domains in which a corresponding premise is true. I suggest, for example, that the determination relation in the domains of aesthetics, humor, and mental content is not rational determination (HFML, pp. 160, 170-171 & fns. 18, 25). If that suggestion is correct, a version of the argument cannot be developed in those domains. 42 Neta claims that all normative domains have the structure that I attributed to law. Call this claim the generalized premise: In all normative domains, descriptive facts are among the determinants of the content of the norms, and the relevant determination relation is rational determination.

I am sympathetic to a view that generalizes my conclusion to all normative domains. In particular, I am sympathetic to the view that a full explanation of the existence of any genuine norms cannot be 42

II.E.

For discussion of the scope of the argument, see HFML, p. 161 and Section

118 MARK GREENBERG

given entirely in terms of descriptive facts. I did not argue for this view in the original article, however, and will not do so here. As I will show, Neta’s argument for the generalized premise is not successful. Moreover, the generalized premise is not plausible. Hence, the specific argument that I gave does not generalize in the way that Neta claims. A number of passages in Neta’s comments suggest that he is working with a determination notion different from the notion of rational determination. For example, when he explains the determination notion, he writes: there is nothing reasonable or unreasonable about the fact that water boils at 212 degrees Fahrenheit, and so whatever makes it the case that water boils at 212 degrees Fahrenheit does not rationally determine that fact. ... In contrast, it is at least somewhat reasonable for the law to require that people who are not convicted of crimes not receive punishment. ... Thus, whatever makes it the case that the law requires this, ... rationally determines that the law requires it (p. 79).

Similarly, he writes: “there must be some reason for the law to require some things and forbid other things – the law’s requirements, unlike the boiling point of water, must be at least somewhat reasonable.” (p. 81). Neta here contrasts domains in which the notion of reasonableness applies to the content of the relevant propositions and domains in which it does not. It is neither reasonable nor unreasonable that water boils at a certain temperature. It is reasonable that the law imposes certain requirements and not others. Moreover, the quoted passages taken together suggest that Neta thinks the content of the law is rationally determined if and only if the content of the law is “somewhat reasonable.” Neta’s notion of the reasonableness of a norm’s content is not the same as my notion of rational determination.43 As noted above, to say 43

If Neta has misunderstood what I mean by “rational determination,” my lack of clarity in HFML is surely to blame. Neta tells me (personal communication) that he was assuming that if the A facts rationally determine the B facts, the B facts “must be at least prima facie somewhat reasonable.” I disagree. Suppose that facts about

ON PRACTICES AND THE LAW 119

that the A facts rationally determine the B facts is to say that the obtaining of the B facts must be intelligible in rational terms in light of the obtaining of the A facts. Consider the legal requirement that contracts for the sale of land be in writing. The intelligibility issue is not whether this requirement is a reasonable or wise one, but whether the existence of the requirement is explained by, or intelligible in the light of, for example, statutes and judicial decisions. Rational intelligibility concerns not the content of the higher-level facts, e.g., that contracts for the sale of land must be in writing, but the relation between the determining facts and the higher-level facts (HFML, pp. 164-165), e.g., between statutes and judicial decisions and the fact that contracts for the sale of land must be in writing. The reasonableness of the content of a norm is consistent with its being opaque why the determinants have the consequences that they do for the content of the norm. For example, one could consistently think that it is a good idea for the Constitution to ban the death penalty, yet that the relevant facts, e.g., about the text of the Constitution or the intentions of its ratifiers, do not make the existence of such a ban rationally intelligible. Conversely, even if it is a terrible idea to allow a tax deduction for mortgage interest, the existence of that deduction might be perfectly intelligible in light of the Internal Revenue Code. (Indeed, the relation between the arrangement of microphysical particles and macrophysical facts, such as the fact of water’s boiling point, might be intelligible even if the notion of reasonableness is not applicable to the macrophysical facts.) In evaluating Neta’s argument for the generalized premise, we need to ensure that we evaluate it with respect to the correct notion. microphysics rationally determined facts about the boiling points of liquids. It would not follow that the notion of reasonableness even has application to facts about the boiling points of liquids. Whether rational determination places some constraint on the content of the law is an interesting and complicated question. For discussion, see HFML, VI.B. But such subtleties are unnecessary for present purposes. As we will see shortly, Neta’s argument on the first horn of his dilemma relies not on the assumption just mentioned, but on its converse - that if the relevant determination relation is not rational determination, then the content of the norms is not reasonable. Given my notion of rational determination, there is no basis for this assumption. With respect to the second horn of the dilemma, the problem that I identify with Neta’s argument would remain even if both the assumption and its converse were true.

120 MARK GREENBERG

Neta’s argument takes the form of a reductio. He begins with the following conditional: “If the [descriptive] facts do not make it reasonable for the norms to be what they are, then either nothing makes it reasonable for the norms to be what they are, or else the reason for the norms to be what they are includes something independent of the [descriptive] facts.” He then argues that both possibilities lead to contradictions, so the antecedent of the conditional must be rejected. The argument is phrased as if the question is the reasonableness of the content of the norms. But as I have just explained, that is not the right question. So it is important to reformulate the argument in terms of rational determination. We first need to clarify one point about the antecedent of Neta’s conditional. Substituting rational determination for reasonableness (and switching to my terminology), the antecedent becomes: “the descriptive facts do not rationally determine the normative facts.” Does this mean that the descriptive facts do not even in part rationally determine the normative facts or that they do not alone do so? Remember that the crucial premise of my argument in the legal case is that the descriptive facts in part rationally determine the normative facts. Since Neta is trying to show that the corresponding premise is true in all normative domains, the antecedent of the conditional (i.e., the premise of his reductio) must be the negation of that premise. Therefore, the correct understanding of the antecedent has to be that the descriptive facts do not even in part rationally determine the normative facts. We can now address Neta’s dilemma. As to the first horn, he argues that if “nothing makes it reasonable for the norms to be what they are,” then “there is no reason to follow those norms,” and thus they are not really norms at all (p. 84). The relevant condition is not, however, that nothing makes it reasonable for the norms to be what they are. It is that the determination relation in the domain in question is not rational determination – i.e., that it is possible that the obtaining of the norms is not intelligible in light of the determinants of the norms. Why should we assume that, in that case, there is no reason to follow those norms? From the fact that the determination relation is not rational determination, nothing follows about the content of the norms. Hence one possibility is that the content of a norm can be the source of reasons, quite apart from the relation between the determining facts and the content of the norm. For example, it might be that the reasons

ON PRACTICES AND THE LAW 121

to follow moral norms derive from their content. (Indeed, it is common to contrast legal and moral norms with respect to whether the content of a norm is the source of the relevant reasons.) More generally, it cannot simply be assumed that putative norms have no normative force if the relevance of the determinants of those norms to the norms need not be reason-based. For that is to assume precisely what is at issue – that rational determination is the relevant determination relation. With respect to the second horn, Neta argues that “if the reason for the norms to be what they are is independent of the [descriptive] facts that determine those norms ... it’s arbitrary that the norms are binding on all and only those creatures of which the determining [descriptive] facts obtain.” In that case, he concludes, “there’s nothing that makes the norms in question binding on creatures like us,” which is to say again that they are not norms after all (pp. 85-86). Neta seems to think that the relevant condition is that 1) descriptive facts determine the norms, and 2) the relevant determination relation is rational determination, yet 3) something independent of the descriptive facts provides the reasons that explain the existence of the norms. But it is hard to see why one would take this to be the relevant condition. If rational determination is the relevant determination relation (so we are not on the first horn of the dilemma) and the descriptive facts do not even in part rationally determine the content of the law (by hypothesis), then descriptive facts are not among the determinants of the normative facts at all. For example, the normative facts might be necessary truths that do not depend on contingent practices. (I sketch a picture in which fundamental moral and epistemic norms have such a status.) If descriptive facts are not determinants of the normative facts, there is no concern that “it’s arbitrary that the norms are binding on all and only those creatures of which the determining [descriptive] facts obtain.” So far I have criticized Neta’s argument that rational determination is the determination relation in all normative domains. I now say something about a few illustrative cases. Consider first the moral domain. According to a plausible and widely held view, fundamental moral truths44 are necessary truths that are independent of any descriptive facts. Applied (or “mixed”) moral truths are applications of the fun44

For what I mean by “fundamental moral truths,” see HFML, p. 159 fn. 7.

122 MARK GREENBERG

damental moral truths to particular circumstances. Perhaps it is a fundamental moral truth that it counts in favor of an action, other things being equal, that it would reduce suffering. In combination with contingent facts about the sources of suffering (and perhaps also with other fundamental moral truths), that fundamental truth yields many applied moral truths, for example, that we should provide infants with nurturing environments. I take this picture for granted here (though Neta would not accept it). On this picture, the interesting metaphysical question is the explanation of the fundamental moral truths. Given that explanation, the explanation of the applied moral truths follows straightforwardly. And it is obvious that normative facts – the fundamental moral truths – play a role in that explanation. I will not of course offer an account of fundamental moral truths here. The present point is that the legal case is a non-starter as a model for such an account. It is uncontroversial that the content of the law obtains in virtue of law-making practices. Fundamental moral truths do not obtain in virtue of practices. That reducing suffering counts in favor of an action (say) is a necessary truth, not something that is contingent on what people do or decide. There is no legal analogue of fundamental moral truths – legal truths that could not be otherwise and that yield all the more specific legal facts when applied to particular circumstances.45 The question of the explanation of the fundamental moral truths is not one on which my argument is intended to shed light. An interesting example is provided by the assumption that all desires or motivating states justify the actions that they (potentially) motivate. This assumption, we saw, was necessary to Neta’s account of promising. Notice that it is better understood as a candidate for a necessary truth than a contingent one. It would be very strange to maintain that some contingent social practice makes it the case that all desires justify actions. The defender of such an assumption would presumably take it to be a necessary truth, which, in combination with facts about what desires people happen to have, yields truths about what justifies particular actions. Hence, Neta’s account of the As an anonymous referee for Legal Theory pointed out, there may be some legal facts that could not be otherwise, e.g., that what is not prohibited is permitted. But even if there be such non-contingent legal facts, they do not, when applied to particular circumstances, yield all the more specific legal facts. 45

ON PRACTICES AND THE LAW 123

normativity of promises, once we make explicit the assumption that all desires justify actions, can be seen to instantiate the picture of the moral domain that I have sketched.46 In my view, the epistemic case is similar to the moral one. Fundamental epistemic norms do not depend on contingent social practices. Again, more specific, applied epistemic norms are the result of applying the fundamental norms to particular circumstances. Neta also considers the case of “the normative content of a particular linguistic system” (p. 82). I am not sure what he has in mind, but we can consider an example I mention in passing in HFML (fns. 18, 25). It is plausible that facts about a person’s or community’s use of words are among the determinants of what those words mean. I do not think that the facts about what words mean are normative facts, so this issue may be a point of disagreement between Neta and me. But this case nonetheless provides a helpful comparison with the legal case. As I say (HFML, fn. 25), it may be that the relation between facts about our use of words and facts about meaning is not rational determination. In sharp contrast to the legal case, I see no reason to think that the relation between the use of words and their meaning is necessarily rationally intelligible. It may be that many possible meanings are ruled out arbitrarily (i.e., in ways for which reasons cannot be given) and that we have non-rational mechanisms that exclude these possibilities from consideration. I can sum up these brief remarks in an intuitive way. At bottom, we do not create epistemic and moral norms. Fundamental moral and epistemic normative truths are necessary, and do not depend on the contingencies of our practices. We do create the facts about what words mean, but we do so in a way that need not be rationally accessible. In the legal domain, we create the facts, and we do so in a way that must be rationally accessible. References Coleman, J. 2001: The Practice of Principle. New York: Oxford University Press. Dworkin, R. 1977: Taking Rights Seriously. London: Duckworth. 46

I thank Seana Shiffrin for pointing this out.

124 MARK GREENBERG

——1986: Law’s Empire, London: Fontana Press. Gilbert, M. 2002: Sociality and Responsibility, Lanham: Rowman and Littlefield. Greenberg, M. 2004: “How Facts Make Law”, Legal Theory, 10, 157-198, reprinted in E. Sosa and E. Villanueva, eds., this volume, and in 1) Hershovitz, S., ed., Exploring Law’s Empire (Oxford University Press, 2006); 2) E. Cáceres, I. Flores, J. Saldaña, and E. Villanueva, eds., Problemas Contemporáneos de la Filosofía del Derecho (Universidad Nacional Autónoma de México, 2005); and 3) Problema, 2 (forthcoming) (cited as “HFML”). ——2005: “A New Map of Theories of Mental Content: Constitutive Accounts and Normative Theories,” Philosophical Issues, 15, p. 299. ——this volume: “Reasons Without Values?,” in E. Sosa and E. Villanueva, eds., this volume. ——2006: “Hartian Positivism and Normative Facts: How Facts Make Law II” (cited as “Hartian Positivism”) in Scott Hershovitz, ed., Exploring Law’s Empire (New York: Oxford University Press). Hart, H.L.A. this volume The Concept of Law. Oxford: Oxford University Press (2nd ed.). Hieronymi, P. MS: “Reasons, Actions, and Attitudes” (unpublished manuscript). Himma, K. 2002: “Inclusive Legal Positivism,” in J. Coleman and S. Shapiro, eds., Oxford Handbook of Jurisprudence and Philosophy of Law. New York: Oxford University Press. Neta, R. this volume: “On the Normative Significance of Brute Facts,” Legal Theory, 10, 199-214, reprinted in E. Sosa and E. Villanueva, eds., this volume. Owens, D. 2006: “A Simple Theory of Promising,” Philosophical Review, 115, pp. 51-77. Raz, J. 1983, The Authority of Law. Oxford: Oxford University Press. ——1994: “Authority, Law and Morality,” in Ethics in the Public Domain. New York: Oxford University Press. ——1999: Practical Reasons and Norms. Oxford: Oxford University Press (2nd ed.). Schapiro, T. 2003: “Compliance, Complicity, and the Nature of Nonideal Conditions,” The Journal of Philosophy, 100, pp. 329-355. Shapiro, S. 1998: “On Hart’s Way Out,” Legal Theory, 4, pp. 469-507. Shiffrin, S. MS: “Promising, Intimate Relationships, and Conventionalism” (unpublished manuscript). Smith, M. 1994: The Moral Problem. Oxford: Blackwell. Sosa E. and Villanueva E., eds., this volume. Wallace, J. and Kolodny, N. 2003: “Promises and Practices Revisited,” Philosophy & Public Affairs, 31, pp. 119-154.

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

5

Supervenience, Value, and Legal Content Enrique Villanueva* I congratulate Prof Mark Greenberg for writing this enlightening paper1 and for bringing into the fore some Metaphysical issues of law. Prof. Mark Greenberg is concerned with fundamental issues among them the place value facts have in law’s content. He states his view in the face of both Legal Positivists and Strong Legal Moralists. His view enlightens the issue and constitutes a new exiting proposal. I have quite a number of questions that spring out of this rich paper but I have to restrain myself to two of them. Mark Greenberg’s view on the relationship between legal content and value facts steers between Positivism and Strong Moralism. Legal Positivism holds that law’s content is propositional and is a social fact and relates to value only contingently. There could be law even where there is no value or where there are negative values or counter values. It holds further that any values that may play a role in law had to be internal to law, they have to be expressly stated or implied by law’s content. Legal Moralism holds that legal content has to do with moral value facts or keeps a conceptual, non-contingent, connection with moral value facts; it is a sort of necessary condition for legal content. Strong Legal Moralism holds that moral values are sufficient for legal content. There are two sorts of value-facts, namely, genuine or external to law and internal to law. *

Instituto de Investigaciones Jurídicas, UNAM. HOW FACTS MAKE LAW, included in the present volume. My comments address a previous version of Mark Greenberg’s paper that was titled “The strange and Intelligible Metaphysics of Law”. 1

126 ENRIQUE VILLANUEVA

Mark Greenberg opposes both, Strong Legal Moralism and Legal Positivism when they claim to provide sufficient conditions for legal content: neither legal decisions nor moral values are sufficient for legal content, according to Greenberg. I. On the Supervenience of Law My first question concerns a general ontological question, namely, What is the ontological status of Law? Is Law a separate entity vis a vis the physical world? Or is Law something that supervenes from the physical world? Thus I am concerned with Mark Greenberg’s view on supervenience. He holds that law facts supervene on physical facts. But then when considering the role value-facts play in law’s content he introduces the rational determination relation (RDR) and holds that this relationship is not captured by supervenience and thus should be considered apart from supervenience. Let’s see how he states his view. Mark Greenberg introduces two sorts of supervenience one s1 that goes from physical facts into moral and legal facts and another s2 that goes from legal decisions or practices into law’s content. We can infer by transitivity that law’s content supervene on physical facts. If there is a difference in descriptive facts there will be a difference in legal facts. More locally, if there is a difference in legal decisions there will be a difference in legal content. And if there is a difference in legal content we can infer that there was a difference in both, legal decisions and physical facts. Now Mark Greenberg says that legal content is made out not only of legal decisions but of values too and that these values impose an epistemic constraint for unless reasons come into (and with reasons enter value) there will not be determinate content (will be vague or non-existent). Mark Greenberg tries to capture reasons’ role by the rational determination relation (RDR). But now, he holds that supervenience fails to capture RDR and thus that RDR does not supervene s2. Or is it that besides supervening s2 RDR gets an ‘extra’ status and that this extra status is not captured by supervenience? But couldn’t it be instead that the RDR is the form that takes the supervenience relation in this case? Or does he want to hold that the RDR is another relation ontologically wholly independent of supervenience2?

SUPERVENIENCE, VALUE, AND LEGAL CONTENT 127

I fail to see why RDR cannot also supervene on physical facts. For even if RDR is an epistemic relation, as Mark Greenberg holds, there is no obstacle to think that epistemic facts also supervene from physical facts and thus that RDR does equally supervene. I must underline here that the metaphysical relationship of supervenience does no prohibit differences between the subvenient base and the resulting supervenience; the only difference excluded is an ontological difference between subvenient base and what supervenes. As RDR does not entail an ontological difference, only an epistemic one, we can conclude that RDR also supervenes on the subvenient physical base and on the legal facts that make the legal decisions. There is no absurdity in holding that legal decisions, reasoning, legal facts, all contribute to fix legal content and thus that legal content supervenes from all that subvenient base. If Materialism is not to be rejected out of hand it should be agreed that reasoning on legal value or RDR does supervene. Reasoning and giving reasons cannot be kept apart from what is material but rather should be kept continuous with it. II. Reasons, Value and the Determination of Legal Content The second thesis I want to consider is that legal content will remain undetermined unless value facts are called into to make it determinate. That legal reasoning had to incorporate value facts in order to provide legal reasons and thus to reach some legal content. That unless resource is made to value facts there will be no legal content and no legal proposition (individual norm). Contra Positivism Mark Greenberg holds that social practices or meanings of words cannot determine the content of law; more generally, that legal decisions cannot by themselves determine law’s content. That if we restrict ourselves to legal practices no determinate content will be reached. The view is that legal decisions together with value facts determine which elements of legal decisions are relevant and how they combine to determine the content of law. Value facts by themselves cannot ei-

128 ENRIQUE VILLANUEVA

ther determine law’s content. This is the necessary role value facts play. It is it a heuristic or interpretive role the one granted to value facts. But Mark Greenberg wants to hold further, that by thus entering into legal content value facts are not just heuristic but become an ontological constituent of law. Thus value facts are both epistemological and ontological. And then, thirdly, he entertains that value facts may have to be external or foreign (moral) to legal decisions for otherwise they cannot fulfil their criterial role in the practice of legal interpretation. I find the transition from knowledge into legal content obscure. The question can be set in this way: why is it that value facts foreign to legal decisions have to be brought into in order to make determinate legal content? Why is it necessary to import these foreign value facts in the legal case? Thus, according to Mark Greenberg value facts are (metaphysical) constituents of law. They also have heuristic relevance and because of that somehow they get into the content of legal propositions. How exactly they reach into the content together with legal decisions? Here is an answer: value facts have a relevance limitation, namely, that value facts determine the relevance of different aspects of legal decisions to the content of law. That when reasoning one has to determine which features of legal decisions matter and how they matter and in order to sort this out by legal reasoning resource has to be made to some value-facts, to what is worth in such and such legal decisions, and thus to what provides reasons to settle the issue one way or another. Mark Greenberg states this much and leaves open what sort of value facts are required to decide if the values required have to be internal or external to law; further argument is needed to decide this last issue but he says that he argues for the need to bring into moral values in another paper. Let us look further into this issue: on the one hand, value facts are criteria and thus have to be independent facts not included in the legal decisions themselves; on the other, they are worth, they have a worthy content. Independence and worthiness are independent features of value facts. That in order to fix a standard that allows sorting out some feature of legal decisions the standard has to enjoy some independence from these legal decisions, seems to be a purely logical point.

SUPERVENIENCE, VALUE, AND LEGAL CONTENT 129

The standard meter (in Paris) is a meter long but it also plays the role of being the standard to determine if some length is one meter long or not; besides being a meter long it plays the role of determining lengths. It is one and the same piece of metal that plays two independent roles, there is no need to have two different pieces of metal, one for each role. On the other hand, this purely logical point has nothing to do with value facts. If we want to say that the standard sets the value of one meter long, that sense of value has nothing to do with being worth, being morally worth, for example. We have to see how being independent or playing as criteria and having value work together. Worth relates to what is (intrinsically) good in (a given) value. Mark Greenberg attributes to value facts a double function, namely, they are criteria that help to choose between the many different features of legal decisions and they are (intrinsically) worth. Value facts are necessary to make law determinate but they are necessary because of their worthiness or because they are (independent) criteria? Let us consider a case. A judge may consider that such and such features of the legal decisions are the ones that are relevant, salient, to reach a sentence; but the judge does not consider these features because they are worth but, for example, just because they fit with the content of the general norm which in its turn fits with the social conventions (and these may be unworthy), or because they will produce such and such behavioral effects (without value), etc., and the judge may choose these features even if they are counter-values or they are morally wrong; he will choose them because he considers that they fit best the legal norms. Again, the sentence issued by the judge may or may not be morally worth but the judge didn’t consider any value-aspect but reached his sentence only on the basis of legal technical reasons. But does the fact that the judge chooses these features and not others make the judge’s decision a value fact? It seems as if by judging the official has to choose among features of legal decisions and choosing involves giving or providing reasons for his decision and in order to do this he must make resource to something independent and what Mark Greenberg holds is that this independent something must be a value fact and his argument seems to be that if it is not a value fact there will be no reason at all and no determinate content of law. The judge will not be using reasons unless the one he surmises includes or en-

130 ENRIQUE VILLANUEVA

tails some value fact and an external value at that. These reasons are value-loaded. But it seems that one can grant that there is legal reasoning that appeals to (independent) criteria but that using that criteria does not imply making resource to some value facts. Legal content will be made determinate without any appeal to value facts. Let us turn again to the judge’s case. Our legalistic judge makes a reasoned choice among the features of the legal decisions but his choice does not bring into any moral or legal values. He just moves inside the content of the legal decisions and chooses the ones that he considers better. Now, better to what? Better to the legal text stated in the norm. His is an internal reasoning that moves inside the legal decisions that help the judge to apply what the norm states in general, to a given case. Reasoning does not involve values but providing reasons in favor of X or Y feature of legal decisions seems to involve some value facts, a minimum of appraisal and value. It seems that giving reasons imply, in the case of law, taking into account some value, something worth. But this minimum worth that the judge, for example, has to take into account does not imply that the judge thinks that it is worth but that it is in fact something worth. Mark’s view states that there will not be legal reasoning, no evaluation or sorting of possibilities, without resource to independent value-facts. Does this mean that an act of choosing whose reason didn’t involve some value facts wouldn’t be intelligible? That providing legal reasons implies making resource to value facts? That without value facts there are no reasons. That choosing without appeal to value facts will be an irrational choosing or no choosing at all. That reasoning or choosing has built into it some value facts at least. There is something obscure, (mystical?) in this conception of ‘providing reasons’, something that the RDR relationship is far from making clear; but far from bridging the gap between ‘reasons’ and ‘values’ only obscures it. This crucial point demands fundamental elucidation. There is something importantly obscure under the name ‘value fact’ for this ‘fact’ has not to be conceived as non-brute or non- conventional, but whatever its ontological status that ‘fact’ has got to provide an explanation of the constituents that will be selected when choosing one model among many other models available.

SUPERVENIENCE, VALUE, AND LEGAL CONTENT 131

But this is precisely what the case of the legalistic judge brings out. Our legalistic judge makes a reasoned choice between alternatives and claims that under the circumstances his is the best possible choice for in reaching it he sticks to the letter both of the text of the legal decisions and of the general norm. This judge made a legally grounded choice but did not take into account any particular value. He may have a general conception according to which observing the text of the legal decisions and norms preserves social coherence, or so, but no particular value follows from it. This judge claims that he made a better choice because it was better that a number of alternatives available (among them some morally informed ones according to which resource had to be made to foreign moral values, that is, foreign and even opposite to the text of the legal decisions and norms). It was better because it fitted better the legal texts than any of the alternatives available. I think this is a very common case in legal practice. It is true that there is some value involved in the judge’s choice, namely, that he considers worthy to keep as close as possible to the law’s text. This value is a general precept or rule, something rather instrumental lacking in intrinsic worthiness. This is a minimum value, not a moral one, and could even be a morally evil value as it was the case in Nazi Germany. Summing up this second point I think that there are two situations in the legal cases, namely, one in which there is no value involved, just the letter of the legal texts and another in which there are some values, and this last possibility branches into two cases, namely, one in which the values are contained already into the legal decisions or in the norm and another of values brought from outside of these legal decisions. I think Mark Greenberg accepts only the last possibility because he feels a double pressure, namely, one coming from the need to interpret the legal texts and another second one coming from the need of independent criteria. I submit that in the legal case there is a gap between reasoning, providing reasons and appealing to value facts and some additional argument has to be supplied in order to fill in that gap. That you may reason a legal case without making any resource to any moral or otherwise external-to-legal-decisions intrinsically worth value facts. That there is no compulsion to abide for such kind of value facts when interpreting the law.

132 ENRIQUE VILLANUEVA

I cannot even begin to provide an account of what is involved in bridging the gap between legal reasons, values, and legal content and neither can claim that such an account is reachable. A first step would be to fix the adequacy conditions that such account should fulfill if it is to claim to be worth. But I am grateful that Mark Greenberg paper touches that central question with enlightening clarity and allows us to see into the problem.2 Tepoztlán, 7/5/03, Petesburgo, 9/1/04

2 I want to express my gratitude to Mark Greenberg and to Ram Neta, my co-symposiasts, who read some previous versions of the present text and help me to correct some unhappy and muddled sentences.

6

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

Reasons Without Values?1 Mark Greenberg In “How Facts Make Law” (Greenberg 2004; hereafter HFML), I argue that a constitutive account of legal facts must appeal to normative facts.2 In the present paper, I take up a challenge raised by Enrique Villanueva (this volume). My argument in HFML begins from two premises. First, the constitutive determinants of legal content include non-normative, contingent facts – descriptive facts, for short.3 The relevant descriptive facts are law-determining practices (or law practices), for example facts about decisions of legislatures, courts, and administrative agencies.4 Second, the determinants of the content of the law rationally determine the content of the law.5 A consequence is that the determinants of the content of the law must include reasons6 why law practices make the contribution that they do to the content of the law. Law practices – and, indeed, descriptive facts more generally – cannot themselves provide the necessary reasons.7 1 I am grateful to Enrique Villanueva for his challenging and thought-provoking paper. I would also like to thank Andrea Ashworth, Kinch Hoekstra, and Seana Shiffrin for valuable comments. 2 I develop the argument in Greenberg (2006a) and Greenberg (2006b). 3 I use the term “normative facts” here in the way that I used “value facts” in HFML (n. 22). See also Greenberg (2006a, n. 5 and accompanying text). 4 See HFML (pp. 162-163, Section II.C.). 5 For elaboration, see HFML (pp. 160, 163-166, 170-172) and Greenberg (2006a, section 1). 6 For more on the relevant notion of a reason, see HFML (pp. 164-166); Greenberg (2006b, section II). 7 For a fuller summary of my argument, see Greenberg 2006a; 2006b.

134 MARK GREENBERG

In section 1 of this paper, I address a preliminary question raised by Villanueva, concerning the relationship between rational determination and supervenience. On this matter there is no fundamental disagreement between us, and I am grateful for the opportunity to clarify the nature of rational determination. In section 2, I turn to Villanueva’s main challenge. He suggests that, to put it very briefly, descriptive facts can be reasons of the relevant kind. Therefore, even if the content of the law depends on reasons, it does not follow that law practices cannot themselves determine the content of the law. Villanueva proposes a value-neutral criterion – textualism. In other words, he suggests that the descriptive facts about the meaning of legal texts are themselves reasons that determine the contribution of law practices to the content of the law. This suggestion depends on too shallow a conception of the requirement of reasons. For the law to be rationally determined, it is not enough that there be some value-neutral criterion that specifies that law practices have certain consequences for the content of the law. There have to be reasons that explain why that criterion, as opposed to others, is the legally correct one – the one that, in the relevant legal system, determines the contribution of law practices to the content of the law.8 Villanueva is right that, in general, normative facts are not the only reasons of the relevant kind. In the case of a constitutive account of legal facts, however, normative facts are the best candidates for such reasons. And, in fact, Villanueva’s textualist criterion derives its appeal from normative facts. 1. Rational Determination and Supervenience It is uncontroversial that the content of the law supervenes on descriptive facts – that is, that the descriptive facts modally determine the content of the law (HFML, p. 159). I claim, however, that the descrip8 The notion of legal correctness here is metaphysical not epistemic. In other words, the legally correct criterion is not the one that judges use to ascertain the content of the law, but the one that determines the content of the law. Of course, judges’ beliefs and practices will have an impact on what is legally correct. See also HFML (p. 179).

REASONS WITHOUT VALUES? 135

tive facts are not the whole story about what constitutively determines, or makes it the case,9 that the law has its content. The making-it-thecase relation in the legal domain is rational determination. And descriptive facts alone are inadequate to rationally determine the content of the law. Villanueva questions the need for me to appeal to rational determination. His suggestion seems to be that I could substitute a thesis about the supervenience base of the content of the law – i.e., about the modal determinants of the content of the law – for a thesis about the rational determinants of the content of the law (Villanueva, this volume, pp. 126-127). Rational determination is a kind of ontological relation, more specifically a kind of constitutive or making-it-the case relation. Supervenience, by contrast, is a modal relation. In my view, it is a mistake to try to cash constitutive claims in modal terms, though this is not the place to argue the point (see Greenberg 2005). In fact, it is natural to appeal to constitutive truths to explain modal ones. For example, that the A facts constitute the B facts explains why the B facts supervene on the A facts. I maintain that descriptive facts and normative facts together rationally determine the content of the law. It follows that the content of the law supervenes on descriptive facts and normative facts (HFML, p. 163).10 But normative facts supervene on descriptive facts (see HFML, p. 159). Hence, it follows from my claim that descriptive facts and normative facts rationally determine the content of the law that the content of the law supervenes on descriptive facts. 9 In this paper, I will use the term “makes it the case” for constitutive rather than modal determination. Also, I use “determination” (“determines,” “determinants,” etc.) for constitutive, rather than modal, determination. (When I mean modal determination, I will be explicit.) For discussion of constitutive and modal determination, see Greenberg 2005. 10 In HFML (pp. 163-164), I make a supervenience condition an explicit part of the account of rational determination. I now think that a better explication of rational determination, along the lines indicated in the text, would not use any modal notion. In order for the A facts (wholly) to rationally determine the B facts, the obtaining of the A facts must make it the case that the B facts obtain. This condition entails that the B facts supervene on the A facts. For suppose that the B facts do not supervene on the A facts. In other words, it is possible for there to be a difference in the B facts without a difference in the A facts. In that case, however, the A facts must not be the whole story about what makes it the case that the B facts obtain.

136 MARK GREENBERG

There are at least two reasons why a supervenience thesis is inadequate for my purposes. First, my central question is whether a full constitutive account of legal facts must make reference to normative facts as well as to descriptive facts. This question cannot even be posed in terms of supervenience, since normative facts are modally redundant (HFML, p. 159). Second, I argue that descriptive facts cannot themselves provide the reasons why certain legal facts, as opposed to others, obtain. (HFML, pp. 160-161; 172-173; 184). Since (bare) modal determination does not require reasons – the subvening facts need not provide reasons why they have the implications that they do for the supervening facts – substituting modal determination for rational determination in my argument would produce a non-starter of an argument. Villanueva suggests that rational determination “is the form that . . . the supervenience relation” takes in the case of law (Villanueva, this volume, p. 126). In light of what I have just said, it would be better to say that rational determination is, in the case of law, the form taken by the constitutive relation between determining facts and the facts they determine. 2. Reasons All the Way Down Villanueva begins the second, main part of his paper by raising a question about the nature of my argument that legal facts depend on normative facts. He takes me to begin from the claim that judges, in ascertaining the law, must appeal to normative facts. He then finds it “obscure” how the argument moves from this claim about the epistemology of law to a conclusion about the metaphysics of law—that normative facts are part of what makes the content of the law what it is (Villanueva, p. 128). My argument does not move from a claim about the epistemology of law to a claim about its metaphysics, but is metaphysical through and through. Rational determination is an ontological relation, not an epistemic one. (See HFML pp. 158, 163.) The claim that law practices cannot, without normative facts, rationally determine the content of the law is not a claim about what is necessary to ascertain the content of the law, but a claim about what is necessary to make the content of

REASONS WITHOUT VALUES? 137

the law what it is.11 The role of normative facts is, accordingly, a metaphysical one: normative facts help to make the content of the law what it is. Some side remarks in HFML are no doubt the source of the misunderstanding (pp. 160, 164 & fn. 17). I point out that my explanation of the rational determination relation employs the notion of a reason, which may well be best understood as an epistemic notion. Thus, one interesting feature of rational determination is that it may be an example of an ontological relation that has an epistemic component. As an example, I consider the possibility that the notion of a reason might be spelled out in terms of the notion of an ideal human reasoner. In that case, “what the law is would depend in part on what an ideal human reasoner would find intelligible” (HFML, p. 164 fn. 17). The situation is therefore somewhat delicate. Rational determination is an ontological relation. To say that the A facts rationally determine the content of the law is to say something about what makes the content of the law what it is, not to say anything about how we discover that content. But the way in which the making-it-the-case gets accomplished may involve an epistemic element.12 Villanueva next argues that my conclusion that normative facts must play a role in determining the content of the law does not follow from my claim that the determination relation in the legal domain is rational determination. A judge can appeal to value-neutral criteria in order to decide the relevance of the law practices to the content of the law. Villanueva’s objection is based on an important point: it is not built into the notion of rational determination that the necessary reasons must be normative facts. But he underestimates the difficulty of finding an alternative to normative facts that could do the necessary work with respect to the constitutive determination of legal facts. 11 As I say in HFML: “It is only because of this underlying metaphysical relation [between law practices and the content of the law] that we ascertain what the law is by consulting those practices” (p. 158). 12 Another example of a position that makes the metaphysics of a domain depend on an epistemic element is Donald Davidson’s well-known view of the mental. According to Davidson (1984a; 1984b), what beliefs and desires a person has depends constitutively on which overall interpretation would make the person most intelligible. See HFML (p. 164 n. 18 and p. 171 n. 25).

138 MARK GREENBERG

A preliminary point is that the relevant question is not whether a judge can reach a decision about what the law is without appealing to normative facts, but whether the content of the law depends on normative facts. In explaining a decision about what the law is, a judge can often provide reasons that are adequate for ordinary purposes without appealing to normative facts. But this observation does not resolve the question whether normative facts are among the determinants of the content of the law. For one thing, a judge may make a mistake about what makes the law what it is. More importantly, in practice a judge typically gives only an incomplete account of what makes the law what it is. Even if a judge does not mention normative facts, they may be among the reasons supporting the relevance of the factors that the judge does mention. In fact, as Ronald Dworkin (1977; 1986) has powerfully argued, when judges in our legal system address basic questions about the grounds of law, it is common for them to appeal to normative facts.13 In light of this clarification, Villanueva’s objection has to be not that a judge could cite only value-neutral criteria, but that value-neutral criteria could by themselves determine the contribution of law practices to the content of the law. In HFML, I introduced the notion of a model of the role of law practices in contributing to the content of the law – a model, for short (pp. 178-179). In these terms, Villanueva urges that the legally correct model – the one that determines the contribution of law practices to the content of the law14 – can be (or can be supported by) a value-neutral criterion. Villanueva’s main example of a value-neutral criterion is “keep[ing] as close as possible to the law’s text” (Villanueva, this volume, p. 131).15 A judge can give as a reason for a conclusion about the content of the law that the conclusion “fitted better the legal texts than any of the alternatives available” (p. 131).

13

See Greenberg (2006b, text accompanying n. 49). For the notion of legal correctness, see above note 8. 15 For a related objection, see Ram Neta’s purported legal counterexample to my argument (Neta 2004, 12-13). My reply to Neta is relevant to Villanueva’s objection as well (Greenberg 2005a, see especially sections 2-3). 14

REASONS WITHOUT VALUES? 139

Let us waive internal problems with the fitting-the-text criterion,16 and assume for purposes of argument that we can construct a coherent and workable model that resolves questions about the relevance of law practices to the content of the law in a value-neutral textualist way. Villanueva suggests such a model for law practices generally, but in order to make the discussion concrete, I will focus on statutes and judicial decisions. Presumably a fitting-the-text model would take a statute to contribute to the content of the law the plain meaning of the statutory text. The case of judicial decisions is more complicated. A decision can involve multiple opinions or no opinion, and judicial opinions do many things other than (purport to) announce what the law is. To fix ideas, we can understand a fitting-the-text model for judicial decisions as follows: in a judicial decision where the majority opinion makes a clear statement of the decision’s relevance to the law, the decision’s contribution to the content of the law is that stated by the majority opinion.17 (This model is of course not the correct or operative one in Anglo-American law; consider, for example, a judicial opinion that purports to announce a legal rule that goes beyond what is necessary to resolve the case before the court.) Why isn’t the fitting-the-text model a reason that could determine the contribution of law practices to the content of the law? That the model would, if applied to law practices, yield certain legal facts does not by itself provide a reason why law practices support those legal facts.18 What is further needed is a reason why the significance of law 16 Fitting the text does not seem to be a criterion that does the work required of it. Even if we ignore aspects of law practices other than texts, the law does not have a text. Rather, there are a vast number of texts associated with law practices. In HFML, I show that there is no direct or automatic route from the meaning of texts to the content of the law. The problem of legal content is not the problem of somehow amalgamating the meanings of myriad texts (HFML, Section III). 17 This model could not by itself determine the relevance of all judicial decisions to the content of the law, but I will focus on it for simplicity. 18 A judge might well cite plain meaning as a reason for taking a statute to make a particular contribution to the law, and might feel no compulsion to give reasons why plain meaning matters. But, as noted above, even assuming that a judge’s account of the relevant reasons is correct, the account will typically be incomplete. (If one is at all tempted by the position that no reason is needed for taking a text to contribute its plain meaning to the content of the law, consider the application of the fitting-the-text

140 MARK GREENBERG

practices is determined by the fitting-the-text model, as opposed to competing models. In the case of statutes, there are competing models that would make the contribution of a statute depend on, for example, the intentions of legislators, legislative history, or the principles that would best justify the enactment of the words of the statute. In the case of judicial decisions, competing models would make the contribution of a judicial decision depend on the narrowest rationale that would justify the result in the case, the rationale that would best justify the result, or the reasoning that the judges employed (which may be in tension with what the judges say the impact of the decision is). What sorts of reasons could support the fitting-the-text model over others? The content of the law could itself support the model. The content of the law has a great deal to say about the relevance of law practices to the content of the law (HFML, pp. 179). But since the present issue concerns how it is possible for the law to have content at all, we cannot appeal to the content of the law. It is easy to find reasons, other than ones deriving from the content of the law, for and against the fitting-the-text model. In the case of statutes, for example, it might be argued that democratic values support the fitting-the-text model. According to this argument, very roughly, the will of the people’s elected representatives should govern, and the plain meaning of a statutory text is the best embodiment of that will. Similarly, it might be argued that the model is fair because it gives citizens the best chance of having notice of what the law is. By contrast, in the case of judicial decisions, democratic values arguably cut against the model, for the model makes it possible for judges to lay down legal standards that are not required for the resolution of cases that come before them. The point is that these reasons for and against the model are, or appeal to, normative facts, such as facts about whose will should govern or about what is fair. These normative facts lie behind the plausibility of the fitting-the-text model with respect to statutes. When one appeals to the plain meaning of the words of a statute, one implicitly appeals to the familiar reasons why the meaning of statutory texts model to judicial decisions, which is not only plainly in need of support, but is incorrect in our legal system. Or, for that matter, consider the fitting-the-text model as applied to legislative history.)

REASONS WITHOUT VALUES? 141

should be respected. Villanueva neglects the possibility that normative facts are playing this role.19 In HFML and another recent article, “Hartian Positivism and Normative Facts,”20 I argue more generally that normative facts are the best candidates for reasons favoring some models over others. For example, I discuss the possibility that conceptual truths (that are not normative truths) could determine which models are correct (HFML, pp. 185, 187-188). Similarly, I discuss the idea that it is a basic truth about the nature of law that the correct model is determined by officials’ “acceptance” of a rule of recognition, as specified in Hart’s (1994) account (Greenberg 2006b, sections III-V). Villanueva offers no new candidates for the necessary reasons. I close by briefly addressing an objection that arises naturally at this juncture. By reasoning parallel to my own, an objector might claim, an appeal to a normative fact would get us nowhere because we would need a reason for the relevance of the normative fact. Hence, the objection would continue, my argument launches us on a vicious regress. I discuss this and related objections more thoroughly in Greenberg (2006b).21 Here I merely summarize the gist of that discussion. In HFML, I argue that normative facts and law practices together, unlike law practices alone, provide the requisite reasons for the law facts. The objection we are considering grants this claim, but maintains that the appeal to normative facts generates a further, higher-order explanatory demand. (And satisfying that further demand will generate a still higher-order explanatory demand, and so on.) I sometimes summarize the argument that shows that law practices alone do not constitute sufficient reasons for legal facts by writing that we need reasons that explain the contribution of law practices to the content of the law.22 In putting the point this way, I may have misled readers into thinking that rational determination requires not only reasons for the legal facts, but also reasons why those reasons are rele19 20 21 22

See Greenberg (2006a, section 3). Greenberg (2006b). See especially Greenberg (2006b, section VI). See, e.g., text accompanying note 6.

142 MARK GREENBERG

vant. The requirement is only that the constitutive determinants of the legal facts together provide reasons for the legal facts. A constitutive account appeals to normative facts not to satisfy a second-order explanatory requirement, but simply to meet the first-order explanatory requirement that law practices do not meet themselves. Saying that we need reasons for the contribution of law practices to the content of the law is just an intuitive way of summarizing why something must be added to the law practices if the determining facts are to constitute reasons for the legal facts. If the law practices and the normative facts together provide reasons why the legal facts are what they are, there is no further explanatory requirement. *** Reasons play a central role in the ontology of law. The determinants of the content of the law, which include law-determining practices such as statutes and judicial decisions, influence the content of the law in a systematic way. But their influence on the content of the law cannot be brute: the determining facts must constitute reasons why particular legal facts obtain. Descriptive facts cannot themselves provide the necessary reasons: for any descriptive fact that is a candidate reason, there are many possible models of its significance for the legal facts. Given the descriptive facts alone, it is arbitrary which of the possible models is correct, and therefore what the legal facts are. Descriptive facts therefore cannot alone determine the content of the law. Normative facts are the best candidates for what needs to be added to the law practices in order for the determining facts to make rationally intelligible why particular legal facts, as opposed to others, obtain. References Davidson, D. 1984a: “Radical Interpretation,” in his Inquiries into Truth and Interpretation. Oxford: Oxford University Press. ——1984b: “Belief and the Basis of Meaning,” in his Inquiries into Truth and Interpretation. Oxford: Oxford University Press.

REASONS WITHOUT VALUES? 143

Dworkin, R. 1977: Taking Rights Seriously. London: Duckworth. ——1986: Law’s Empire. London: Fontana Press. Greenberg, M. 2004: “How Facts Make Law,” Legal Theory, 10, 157-198, reprinted in E. Sosa and E. Villanueva, eds., this volume and in S. Hershovitz, ed., Exploring Law’s Empire. Oxford: Oxford University Press, 2006 (cited as HFML). ——2005: “A New Map of Theories of Mental Content: Constitutive Accounts and Normative Theories,” Philosophical Issues, 15, 299-320. ——2006a: “On Practices and the Law,” Legal Theory, 12 (113-136), reprinted in E. Sosa and E. Villanueva, eds., this volume. ——2006b: “Hartian Positivism and Normative Facts: How Facts Make Law II,” in S. Hershovitz, ed., Exploring Law’s Empire. Oxford: Oxford University Press. Neta, R., this volume: “On the Normative Significance of Brute Facts,” Legal Theory, 10, 199-214 (2004), reprinted in E. Sosa and E. Villanueva, eds., this volume. Sosa, E. and Villanueva, E., eds., this volume. Villanueva, E., this volume: “Supervenience, Value, and Legal Content,” in E. Sosa and E. Villanueva, eds., this volume.

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

7

Theory, Practice and Ubiquitous Interpretation: The Basics Martin Stone I. Introductory Throughout his work, whatever the topic, Stanley Fish is preoccupied with a question concerning the basis of our entitlement, in various domains of discourse, to the notions of correctness and objectivity in judgment. Literary criticism and legal analysis supply his main examples. In virtue of what, he often asks, is one reading of a literary text or one application of a legal rule correct, and not another? This question is already present (though outside the main focus) in Surprised by Sin, with its perception of Milton as a writer concerned with severe disagreements: cases where someone’s access to how things are, or to what is a reason for what, appears to depend on their acceptance of a premise which is un-demonstrable (since the compellingness of any demonstration seems to depend on it).1 Next, Self-Consuming Artifacts argues for the role the reader in determining the meaning of a literary text–and this in something more than the trivial sense that insofar as there are works of literature there must be readers of them. (The idea, rather, is that something about the reader or her situation explains why a literary text means what it does.) This view gets articulated in Is There a Text in this Class as a kind of “conventionalism,” according to which communal accreditation determines what counts as correct in judgments about literary texts. And this thesis is extended to other subjects (especially law) in Doing What Comes Naturally, where Fish portrays a number of otherwise diverse theorists (Unger, See also Fish, The Trouble With Principle (Cambridge: Harvard University Press, 1999), 243-248, 263-70. 1

146 MARTIN STONE

Dworkin, Hart, Posner) as seeking–often against their own declared intentions–an Archimedean standpoint for judgment. Doubting the availability of such a standpoint, Fish proposes s a “pragmatic” alternative in which the notions of interpretation and community again help to secure the notion that a text sometimes means one thing rather than another.2 In this essay, I sketch the main argument running through these works, address a few misunderstandings, and indicate, somewhat programmatically, how the argument deserves to be criticized.3 The grounds for my criticism are not exactly foreign to Fish’s work. In fact, my main premise might be described as “the priority of the practical point of view,” something Fish himself seems to favor in his criticism of the theorists mentioned above. Thus, my hope is that Fish will be able to see my criticism as a friendly one: as clarifying and extending a valuable line of his thought, notwithstanding that the theory to be exorcized, in this case, is his own. Otherwise put, my criticism finds a conflict within Fish’s work; and the part of a friend, in such a case, is naturally to be an ally–giving reinforcement or resolution—to the better side. II. The Skeptical Argument To start things off, we might remember that Milton writes in an historical moment of crises having the form of a perceived gap between the desire for a just order of society and historical experience, or be2 More recent work by Fish extends these themes by proposing that we see “Liberalism” (as he writes it) as presenting a political analogy to the suspect forms of theory depicted earlier. See, e.g., The Trouble With Principle, ibid. This proposal is off the main line of my discussion. 3 The following remarks stem from seminar I taught in Spring 1998 in Duke’s Program in Literature; they retain here the style and sound of their pedagogical origin. My focus is only on the main nerve of Fish’s argument. A broader treatment can be found in my contribution to John Gibson and Wolfgang Huemer, Literature after Wittgenstein (forthcoming, Routledge). Related issues are also touched on in my “Wittgenstein on Deconstruction,” in Alice Crary and Rupert Read, eds., The New Wittgenstein (London: Routledge Press, 2000), and my “Focusing the Law: What Legal Interpretation is Not,” in Andrei Marmor, ed., Law and Interpretation: Essays in Legal Philosophy (Oxford: Oxford University Press, 1995).

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 147

tween moral value and social fact–a gap figured by Milton as the relation between God and man.4 The project of Paradise Lost–“to justifie the ways of God to men”–is motivated by a sense of the apparent failure of God, and hence implies the need for human speech and judgment to bring moral intelligibility to historical experience. Thus, what attracts Fish to Milton is that Milton treats poetry’s religious theme as a secular, not a religious poet. That is, unlike his predecessors in this theme, Milton’s focus is not the difficulties man encounters in keeping to the demands of God, but the potentially false surmise–the idolatry, as it were–involved in talk of the specific content of divine imperatives.5 Two ideas may be extracted from this general picture of Milton—ideas, which structure a great deal of Fish’s subsequent work: (1) (the idea of correct judgment or right social order as that judgment or order which is in accord with God’s will; and (2) the idea that what makes a particular judgment one that accords with (keeps faith with) God’s will is itself a question which calls for human judgment. The first of these ideas expresses what a “judgment” is for the human being laboring in history, a creature capable of departing–through responsible exercises of his conceptual capacities—from the divine will. The second idea might be thought of as a form of “antinomianism.” Putting these together, we have the thought: Although correct judgment is judgment in accord with God’s will, there is no means of externally validating (i.e., from a point external to our judgments) whether one has got things right. In Fish’s words: 4 See Allen Grossman, “Milton’s Sonnet, “On the Late Massacre in Piemont’: The Vulnerability of Persons in a Revolutionary Situation,” in The Long Schoolroom: Lessons in the Bitter Logic of the Poetic Principle (Ann Arbor: University of Michigan Press, 1997). On the general connection between Fish’s reading of Paradise Lost and his later interest in “interpretation,” see the Preface to the second edition of his Surprised by Sin (Cambridge: Harvard University Press, 1997). 5 The idea of idolatry–the false representation of divinity—provides the point of intimacy, which Fish later explores, between Milton and the politics of Roberto Unger. See “Unger and Milton” in Fish, Doing What Comes Naturally: Change, Rhetoric, and the Practice of Theory in Literary and Legal Studies (Durham: Duke University Press, 1989), esp. pp. 403, 412.

148 MARTIN STONE

The doctrine of the inner light marks out the area of interpretive labour; the doctrine of the single Truth names the goal of that labour, but withholds explicit directions for attaining it.6

In saying that a thought along these lines structures Fish’s work, what is meant is that Fish is concerned to respond to a certain felt difficulty with this thought. The difficulty is characteristic of modern philosophy. People are apt to feel that, in the absence of “directions” for certifying disputed judgments as correct, the very idea of correctness in judgment—of their being something to “get right”—must come under threat. The difficulty arises on the assumption that our entitlement to regard one of two conflicting judgments as “objectively correct” requires that there be some means of demonstrating its correctness through premises which do not presuppose either of the positions in question. It appears questionable, to say the least, whether this requirement can be satisfied in such contest-laden domains as law or literary criticism, where judgment nonetheless has objective purport. So someone working in these domains might come to feel the pinch of the present difficulty. And they might then naturally move in one of two opposing intellectual directions:(a) They might attempt to vindicate the objectivizing view of discursive participants by supplying a theory of what makes judgments in the relevant domain correct. That is, they might construct a theory of validity for the discourse in question. (Or, if we call different judgments–e.g., about literature or law—different “interpretations,” then we may speak here of a theory of validity in interpretation). (b) Alternatively, they might come to deny that talk of correctness has the substance which discursive participants are inclined to credit it with. In its extreme versions, this view says that we are not really entitled to talk about “getting things right,” only about what people take to be right. This means that engaged participants are prone to an illusion of some kind. For they take their judgments to be not merely their way of “taking” things, they take them–this defines the participant perspective–to be true; whereas according to the theorist, such claims 6

Fish, Surprised by Sin, ibid., p. Xliv.

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 149

need to be accounted for in other terms, e.g., as ideology or the rhetorical camouflage of power, and so on. Since (b) evidently expresses a form of skepticism, it is tempting to think of those embracing (a) as anti-skeptics. This is right in one way, but misleading in another. It accurately records the way (a) and (b) are interlocked voices in a single argument, as well as the intellectual derivativeness of (a): proponents of (a) are often defending certain commonplaces of the practice against the threat of skepticism. However, in a somewhat broader sense, one might regard both positions as “skeptical” ones. That is, “skepticism” might be taken to refer not just to arguments which deny that correctness in judgment is possible, but also to attempts to refute those arguments. Why speak this way? The point is to mark an alternative point of view from which (a) and (b) look intellectually intimate with each another; from which it appears that what these views share in common is larger than the point over which they disagree. What these views share in common is the premise: (P) Our entitlement to see one of two conflicting judgments as objectively correct requires some means, independent of those judgements, for validating one or another of them as correct. The intimacy between positions (a) and (b) is that neither so much as sees a question to be asked about (P). That is, for each of them, (P) is invisible as a premise. Thus, each position takes itself to be the only alternative to the other. Where in this landscape does Fish belong? The answer is: Fish wishes to reject both (a) and (b). So he is essentially an anti-skeptic in the broad sense of the term. “Various characterizations of me as a skeptic–as someone who disbelieves in truth or relatives value...or is unconfident in his judgments, follow from the confusion between a very limited denial of a universal mechanism of validation and the denial, which I do not and never would make, of just about everything”7–of just about everything (I take this to say) about which people judge and, often enough, agree. As this makes clear, Fish’s intended target is not the possibility of true or confident judgment just as such, but only what he regards as one misFish, “A Reply to My Critics,” The Responsive Community, Vol. 12, no. 3 (2002), p. 65. 7

150 MARTIN STONE

taken defense of this possibility–one which seeks a “universal mechanism.” Fish sometimes calls the mistaken defense a “foundationalist” one. His continuous theme is that the options described by (a) and (b)—the Scylla of foundationalism and the Charybdis of skepticism, as it were—are not exhaustive; and that we can locate another possibility if we allow ourselves to question (P), the premise which the skeptic and her traditional opponent share in common. Now it should be noted that “questioning (P)” evidently means for Fish: constructing an alternative explanation of the basis of our entitlement to the notion of correctness in judgment. (This means giving an alternative account, as we shall see, of how there could sometimes be “plain meanings”–obvious cases which no one disputes.)8 But I shall suggest below that this misses a more radical possibility (§§5-6). One doesn’t need to read Milton, of course, in order to feel that (P) is not an innocuous premise. Indeed, one might see (Fish’s) Milton’s central thought–that we endeavor to judge the “single Truth” without anything standing surety for our judgments–as a variation on a point of Aristotle’s: namely, that the practically wise person doesn’t have a recipe (or a set of deductively applicable instructions) for living well, but rather is able to see the significance of the details of practical situations in light of a correct grasp of the relevant ethical concepts. Aristotle’s remarks trace a circle that never leaves the domain of ethical thinking. For if we ask, “what makes a grasp of the relevant ethical concepts a correct one,” Aristotle is apt merely to refer us to the judgments of the practically wise person, just where a more modern (and in the broad sense, skeptical) line of thought would expect to find an attempt at external validation.9 At this point, someone might wish to object, however: “Surely Fish is a skeptic. Doesn’t he everywhere say that every judgment is contestable and that there may be no means, independent of the dispute, for settling the matter; that what will count as evidence in favor or one or another judgment, for example, may itself be a function of the position one holds? What is this if not a skeptical challenge to our notion that some judgments are objectively correct?” Cf. Fish, “Force” in Doing What Comes Naturally, esp. p. 513. See John McDowell, “Some Issues in Aristotle’s Moral Psychology” in Mind, Value, and Reality (Cambridge: Harvard University Press, 1998). 8 9

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 151

This is a misunderstanding of Fish’s aim, though it may not be one—I’m inclined to think—of his achievement. As I will later explain, there are genuine difficulties in Fish’s argument, difficulties which understandably lead his readers to take him for a skeptic (§5). (Thus disclaimers like the one quoted above need to be continuously re-issued.) This recurrent mistake on the part of Fish’s readers needs a more careful account, however. For the present objection merely records the fact that Fish does indeed set his face against attempts to defend “objective correctness” along the lines of (a). The objection thus testifies to the tenacious hold of premise (P). For if one unquestioningly accepts (P), then one is bound to hear Fish’s opposition to (a) as incurring a commitment to the skepticism of (b). This misses the general alternative Fish has in mind: The failure of foundationalism, rather than affording a reason for embracing skepticism, should, given the practical intolerability of the skeptic’s position, provide a reason for questioning (P), the premise which makes it appear as if these were the only options. III. “Theory” Fish sometimes calls (a) “theory hope” and (b) “theory fear.” (What I call (P) therefore exhibits the common genus.) These labels imply that someone questioning (P) is seeking freedom from a way of thinking which makes having a “theory” a prerequisite to our entitlement to take up the participant or objectivizing point of view. That of course is how Fish often presents himself: as being against “theory” in some sense of the word.10 This way of talking won’t do any harm if one bears in mind what “theory” stands for–the requirement expressed by (P). Yet it is not especially perspicuous either, if only because the word “theory” is so beloved by academics today as to be almost devoid of significant contrast: it comes to appear that to think or reason about anything at all is to “theorize.”11 (The causes of this emptiness are, I think, significant: 10 See, e.g., “Consequences” and “Dennis Martinez and the Uses of Theory,” both in Doing What Comes Naturally. 11 Cf. “Dennis Martinez,” in Doing What Comes Naturally, p. 378.

152 MARTIN STONE

it expresses difficulties we have with the thought that certain forms of reasoning are distinctively “practical”: see below.) Here the following points ought to be kept in view. First, “theory/practice” is not to be construed as a contrast between what we should ordinarily call “doing something” on the one hand, and reasoning or reflecting, even very abstractly, on the other. Fish’s anti-theoreticism is not the view—which might describe certain spiritual exercises like Zen or Pyrrhonism—12 that it is useless to reason or ponder. It is true that in some activities–playing basketball, for example–one’s performance is apt to be hindered if one reflects on what one is doing while doing it. This shows up in forms of training: one learns basketball by playing it, not (as one first learns, say, law) by studying it. Nonetheless, this is simply a special feature of certain activities, related to the kinds of performances and skills they require. (For this reason, Fish’s use of the game of baseball as an analogy for legal “practice” is as obscuring as it is clarifying.)13 In the case of other activities—legal argumentation, for example—to perform successfully to is to reason at a high level of abstraction; it is to advance, as lawyers say, a “legal theory.” Being a plumber is perhaps an intermediate case, somewhere between basketball and law. Much of what a competent plumber does he could do “in his sleep.” Yet a competent plumber ought also to be capable, when the occasion demands, of posing alternative hypotheses about the source of a problem and considering different ways of proceeding. (The best course may not be the one that “comes naturally” or prior to deliberation.) The general point here is two-fold: first, it does not make sense to contrast doing something and reasoning or reflecting in general; second, where such a contrast can be drawn (e.g., in describing two aspects of the plumber’s job), the contrast will clearly be seen to be beside Fish’s point. 12 Or perhaps the experience of clinical depression–though it seems more accurate to say that depression is more often the feeling that it is useless to act (Hamlet). 13 See “Dennis Martinez” in Doing What Comes Naturally. Part of what makes sports both important and pleasurable is that the meaning (and other effects) of action are completely tractable within the game. Related to this is the fact that the point or goal of action is in general completely perspicuous–to win. Most everyday activities are not like this, and the law is a far cry away.

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 153

Why should premise (P) be associated with a kind of “theoreticism”? The idea calls for a contrast, as I have hinted, not between doing something (“practice”) and reasoning about it (“theory”), but between two forms of reasoning–namely, theoretical and practical. Consider a judge who endeavors to apply a legal rule, say one requiring good faith in one’s dealings with others. The judge must think about what this concept requires; he must determine what, in particular cases, would be in accord with the rule. Now we might wish to call this movement from the abstract to the concrete “practical” reasoning because it involves capacities of thought which are distinct from drawing logical inferences and from thinking about what will lead to what.14 This will be true if, as here, the rule in question cannot be expected to function as part of a premise from which, given the facts, one could simply deduce the desired conclusion. Of course, there may be room for explanations of “good faith dealing”–explanations which might even be called (in an anodyne sense) a “theory” of it. But it may also be that correctly to grasp the concept of “good faith”–or some concept used to explain it—just is, in part, to be able to see that this and not that is required in circumstances like these. Applications of theoretical rationality cannot in general tell one what it is to get things right in such applicative judgments, or how to recognize particular acts or circumstances as instances of general classifications. Nor can correct judgment be generally explained by means of rules for making those judgments, for then we should need rules for correctly applying those rules, and so on, in a hopeless regress. 15 Now suppose that a dispute breaks out about what the rule requires. Premise P says we are not entitled to think of either view as being genuinely “correct” unless we have a validating argument from the outside. Applied quite generally (i.e., not just to “good faith” but to 14 Reasoning about what will lead to what often comes into reasoning about what to do in a particular situation. But it is not distinctly practical in the present sense. 15 This was noticed by Kant. See Critique of Pure Reason, trans. Norman Kemp Smith (New York: St. Martin’s Press, 1929), A133/B172; and Kant, “On the Common Saying: ‘This May be True in Theory, but it does not Apply in Practice’” in Kant, Political Writings, ed. Han Reiss (Cambridge: Cambridge University Press, 1970). The point also arises in Wittgenstein’s discussion of “accord with a rule” in L. Wittgenstein, Philosophical Investigations, trans. G. E. M. Anscombe (Oxford: Blackwell, 1958), §§138-202.

154 MARTIN STONE

every concept which can be used to explain it), this can only be a demand that “correct” judgment be made deductively accessible: the correct resolution of the issue would be expressible as the conclusion of an argument which would be compelling to anyone who can draw logical inferences and recognize what the facts are.16 (If someone still persisted in not “getting it,” they could be convicted of irrationality by established standards of theoretic rationality; so this would comprise a “universal mechanism.”) To be “against theory,” in this context, is to allow that there may be useful explanations of “good faith”; but it is to reject the thought that no applicative judgment can be regarded as “objectively correct” without an explanation of this sort: one which makes judgment available in a way that, in principle, obviates the need for practical discernment.17 IV. “Interpretive Communities” The example of applying a legal rule brings out the fact that any account of how correct judgment is possible (and that is Fish’s general question) must explain also how it is that certain actions or events can be “in accord” with a rule, or indeed with any bit of intentionality. Meaning has a normative aspect: we could not speak about texts (or 16 I develop this idea in the context of debates in legal theory in my “Legal Formalism: The Task of Judgment,” in Jules Coleman and Scott Shapiro, eds., Oxford Handbook of Jurisprudence and Legal Philosophy (Oxford: Oxford University Press, 2002). 17 For purposes of simplicity, I am not questioning the thought that explanations which make judgment deductively available would obviate the need, in principle, for “practical discernment.” But the better view is that all judgments–even deductive ones–rely on something like the kind of discernment which is out in the open in cases of practical conflict. As John McDowell has argued, this may be taken as one of the lessons of Wittgenstein’s remarks on “accord with a rule.” (There is a similar point in Stanley Cavell, The Claim of Reason, 2d ed. (Cambridge: Harvard University Press, 1999). As McDowell suggests, this lesson should allay the temptation to think that judgment in hard cases (where the need for discernment is conspicuous) must suffer in its credentials of objectivity by comparison with a deductive case, conceived as a paradigm of objective judgment. For the lesson is that even the deductive case does not live up to the notional ideal of a discernment-free path to judgment which structures the invidious comparison. See McDowell “Virtue and Reason” in Mind, Value and Reality (Cambridge: Harvard University Press, 1998).

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 155

correct judgments about them) if we couldn’t make use of such notions as “accord.” Difficulties we get into over such normative notions are thus at the core of Fish’s argument. To illustrate, consider a simple statement describing how things are in the world, for example “There is a vehicle parked on Elm Street.” This sorts the world into states of affairs which are in accord with it and those which are not; it makes a demand, one may say, on how the world must be if the judgment is to be correct. A similar point applies to any item— e.g., a rule, judgment, wish, order, thought, expectation, belief, etc.–which carries meaning: generally speaking, meanings sort things out. One important bit of the world which meanings sort-out is of course human linguistic behavior itself. For example, if someone grasps the meaning of the word “vehicle” then she is required, if she is to act “in accord” with what she has grasped, to reach certain determinate verdicts when the world presents her with circumstances which bring this concept into play. Naturally, there may be borderline cases–Is it still a “vehicle” even though it lacks a motor? Is it “on Elm” when it is abutting the corner–but insofar as these words can be used to communicate anything at all, there must also be plain cases, cases in which no classificatory hesitation arises. This commonplace idea–viz., that meaning has a normative bearing on linguistic performances–is related to other commonplaces concerning truth and objectivity, for example, that the world can be such as to make it correct or incorrect to say certain things about it. The very idea of “something which can correctly be said about the world” presupposes that there is a normative pattern in our use of words, a pattern that a particular use can (or can fail) to keep faith with. If that weren’t the case, then anything could be said about anything—so nothing could be said at all. The upshot is that should we begin to loose our grip on normative notions like “accord,” then our notions of meaning, objectivity and truth will come under threat as well. And this is just what is happening in the skeptical currents in which Fish is swimming. Keeping a grip on the notion of “accord” (or related normative notions: “misuse,” “misunderstanding,” “misapplication,” etc.) turns out to be a difficult thing to do. For there is a tempting line of thought which seems precisely to unhinge us here. And this line of thought provides the right context in

156 MARTIN STONE

which to understand the general work which Fish sees “interpretive communities” as doing. The unhinging line of thought begins with the notion of a “sign” or a “text.” A sign or a text is anything which carries linguistic meaning.18 Thus, the first line of Milton’s Paradise Lost–“Of Mans First Disobedience....”—comprises a sign or text, as does also a road sign pointing out the direction in which one is to go (——>). It will be useful to take the latter as our example because it makes immediately perspicuous that a sign carrying meaning sorts behavior into that which accords with it and that which does not. (This, in fact, is the chief difference between signs carrying meaning and mere doodles or noises: signs are “alive” in that they have such normative reach. But this is also the thought on which we seem, under theoretical pressure, to loose our grip.) Before proceeding, two comments may help to set things up more clearly. First, concerning exemplification. Fish initially introduced the notion of “interpretive community” to address a set of questions arising within literary studies: e.g., the relevance of authorial intentions, the distinctive “literariness” (if there be any) of literary texts, the semantic multiplicity or univocity of such texts, the reasons for interpretive disagreement, the status of appeals to the “text itself” in resolving disagreement, the possible innovativeness of literary interpretation, and so on. Our board (——>) would be a poor example for discussing such issues.19 But it becomes clear, in the evolution of Fish’s work (§1), that insofar as literary texts always require interpretation, they are to be regarded as merely exemplary of how it is with discourse in general: “Communications of every kind are characterized by exactly the same conditions–the necessity of interpretive work...and the construction by acts of interpretation”.20 I find this development dis18 “Non-natural meaning” would be more precise. See Paul Grice, “Meaning” The Philosophical Review 64: 377-388 (1957). The idea, at any rate, is just to focus on the concept of meaning at stake when one says, e.g., “That is not what the text means,” as opposed to e.g., “These tracks mean that a lion was here.” 19 At least without some further details. 20 Fish, “With the Compliments of the Author: Reflections on Austin and Derrida,” in Doing What Comes Naturally, pp. 43-44.

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 157

appointing because, for reasons which will appear, I think the necessity of interpretation could only have been plausible as a restricted thesis about the meaning of literary texts. (In this role–as opposed to the perfectly general role Fish gives it–the necessity of interpretation might also have told us something genuinely informative about “literature,” or about the nature of our interest in it.)21 In any case, given the generality of Fish’s thesis, a basic example (——>) is just what is wanted for discussing it; the more basic the better. Second, concerning skepticism. The line of thought to be considered may be called “skeptical,” though it doesn’t aim towards a skeptical conclusion. It does not aim to deny, for example, that it is sometimes perfectly plain what road signs or other texts mean. The question is merely how such “plainness” is possible. What Fish (and others today) wish to deny is not that there are perfectly plain meanings (that would make them not just skeptical but mad),22 but only a purportedly suspect conception–naive or metaphysical–of such phenomena or of their conditions of possibility. Provisionally, we may say that on the suspect conception, a case of plain meaning is apt to be considered “inherently plain, plain in and of itself” or plain as a simple fact-of-the-matter.23

21 The implications of the idea that the interpretability of literary works is a function of the kind of interest we take in (what we call) “literature” are developed in my contribution to Wittgenstein after Literature, note 3 above. The idea is not completely foreign to Fish–see e.g., “Fish v. Fiss,” in Doing What Comes Naturally, p. 137 (contrasting literature and law)—but it never leads him to question the generality of the interpretivist thesis. To the contrary, such differences as may appear between literature and law are, for him–given that thesis—to be considered as effects of interpretive activity. 22 See e.g., “Force” in Doing What Comes Naturally, p. 513: “The question is not whether there are in fact plain cases–there surely are–but, rather, of what is their plainness a condition and a property”; see also “Working on the Chain Gang” in Doing What Comes Naturally, p. 101. 23 Fish, “Force” in Doing What Comes Naturally, p. 513. I say that Fish’s formula is “provisional” (for us) because its sense is part of what needs to be investigated here. In the end, I think there should be no problem saying that some cases are “inherently plain”; that could strike us as just a bit of practical commonsense. See §§5-6 below.

158 MARTIN STONE

Briefly, then, the “skeptical” line of thought unfolds like this: 24 1. Considered just “in and of itself” (say, as an inscribed piece of wood), the sign (——>) does not determine what is in accord with it and what is not; it does not determine, say, whether one is to go in the direction of the arrow or in the opposite one. (This is true of any text: just in itself, it is dead matter, powerless to determine its own meaning or how we are to understand it; powerless, as Fish sometimes likes to say, to “execute” its own meaning.) 2. To animate the sign into meaning something–i.e., to get the normative notion of “accord,” and hence of “meaning” into play–we need to consider the sign not “in itself” but under some interpretation that has been put on it. That is: we need to interpret it–e.g., as a road sign saying that one is to go in a certain direction. Signs mean what they do only by way of some interpretation. 3. This seems clear enough. But is it? If a sign or text cannot “in itself” determine what accords with it, how does it manage to do so when considered under an interpretation? In fact, it looks like there’s a problem here. 4. Suppose that an “interpretation” involves some further sign or text. For example, one might “interpret” the sign (——>) by using the spoken words “this way,” accompanied by a pointing gesture. Surely, the original sign is now alive with meaning, no? Well, no. For according to step (1), this new text cannot “in itself” determine how it is to be followed. It too seems dead–a bit of sound and fury. It seems that if the interpretation (“this way”) is really to animate our original sign into meaning something, we shall first need an interpretation of this interpretation, and so on. A hopeless regress–not to mention an hysteria of gesticulation (“this way, I mean THIS way”)—looms before us. 5. We had better back-up. Why did we think–in step (2)–that an interpretation could help bring “meaning” into the picture? The answer seems clear. There was a doubt about how the sign was to followed, and we know that “interpretations” do sometimes successfully function to remove or avert such doubts. “Interpretations” in a familiar, sense are a kind of explanation: they come into play when the mean24 The dialectic sketched here retraces a few passages in my “Wittgenstein on Deconstruction,” in The New Wittgenstein, ibid. It is under investigation in Wittgenstein’s discussion of the concepts of meaning and understanding.

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 159

ing of a text isn’t fully clear.25 Thus, it was hoped that we could get “meaning” into our picture by making a quite general use of this familiar function of interpretation. But it appears now that the notion of “interpretation” is unsuited for this general role. Rather than animating our original sign, the requirement of interpretation seems only to redouble the problem of its impotence. 6. But wait, someone will say. When we put someone pointing and saying “this way” into the picture, we didn’t just introduce another inert block of wood, or even (comically) any number of inert items, each one standing behind the next; nor did we just introduce some noises, such as a person might make. Instead, we put a person into the picture, a living being. One wants to say: surely that makes a difference; surely meaning, in all its splendid animation, is somewhere at hand! The thought which is apt to occur now is that a person makes a difference, not as a potential source of sound and other commotion (many things are that), but as the locus of a mind. Thus, the demonstrative utterance “this way” introduces meaning into the situation because it introduces someone who thinks or intends the sign in a certain way. Indeed, it seems that thought is really the essential thing here, for it might have occurred—and meaning might come into the picture—even without the giving of any further signs. The motive here is understandable enough: since further signs (texts or linguistic items) merely re-double the problem of the sign’s impotence, it becomes tempting to think that “interpretation” must refer to some essentially mental act of thinking the sign one way or another. What we need, the thought goes, is not another inert bit of nature, but a mind; not essentially interpretations (qua signs) but an interpreter, alive and present. 7. Alas, this solution can make us happy only for a moment. Suppose it is asked, “What does his thinking or intending the sign this way rather than that way consist in?” After all, if we can doubt what the original sign requires, it should be possible to raise a question 25

This seems to be true even of “performing interpretations” (e.g., Gould’s interpretation of the Goldberg Variations) which don’t at first look like explanations. Even here, however, the sense is that a performance helps to elucidate aspects of a work of art which would not otherwise be fully perspicuous.

160 MARTIN STONE

about what he intends or what accords with his intention. There seem to be only two general possibilities. (A) We might say that his intending the sign this way consists in his meaning that, or meaning that. Clearly, this answer goes nowhere: it merely re-uses the very notion–“meaning”–which “interpretation” was supposed to explain.26 (B) We might try to identify something which went on “in his mind,” considered as a region of goings-on that is left over once we abstract from the world and all the (“in themselves” meaningless) items to be found in it. However, this option looks no less hopeless. For one thing, if someone always follows the sign in the direction of the arrow (or points out the mistake when other people don’t), then we shall say that he grasps its meaning no matter what actually goes on “in his mind.” In fact, the search for a meaning-creating item in his mind only returns our original problem. Lots of things might have occurred to him, some of which (like the pang of hunger or impatience he felt) seem irrelevant. But if anything occurs to him which does seem relevant (perhaps he thought “to the right” or perhaps he saw a picture of the traveler turning right in his mind’s eye), it is bound to disappoint us. For it is just one more discrete item which, like our original sign, can always be projected and applied in different ways. So it too stands in need of an interpretation. This result shouldn’t surprise us. For it amounts to what much 20th century philosophy has told us anyway: viz., that we can’t really make intelligible to ourselves how a thought occurring in someone’s mind–e.g. ”turn right”–can be such as to be any more determinate, or less in need of interpretation, than a text representing that thought. For we can’t really understand what it would be for there to be an item in the mind that had the requisite normative properties of meaning but that was not, from the get-go, 26 The point here is not that we can’t give analytic explanations of what “intending” or “meaning” something consists in; Grice and others do that. The point is that, given the dialectical set-up (one suggesting that meaning is in fact impossible), the kind of explanation we need here must be one that does not make use of any normative notions closely related meaning. Gricean and other analyses of “meaning” do not meet this requirement. This is the answer to a question that George Wilson asked me at a talk I gave at Johns Hopkins several years ago. I regret that I was only able to give a confusing answer at the time.

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 161

subject to the conditions or requirements of representability (or communicability) in signs.27 Actually, there is a third possibility for attempting to answer the question “what does his thinking or intending the sign...consist in?” We might try to identify something that goes on “in his mind,” considered not as a region apart the world, but as something that includes a lot of happenings there, for example what other people say and do. Fish’s idea of an “interpretive community”—or of a subject whose mental life is what it is only “by virtue of his membership in a community of interpretation”28 —is a version of this idea. I shall postpone commenting on it because my endeavor is first to get into a position to describe its appeal more fully. 8. Someone might throw up their hands at this point and say something like this: “Clearly, an interpretation is needed to get meaning into this picture. But the idea of “interpretation” needed is just that of a very unique and remarkable spiritual power to make signs mean this rather than that–somewhat like the power to give life to dead matter. ‘The Mind’ is that unique kind of thing which has such remarkable powers. “To interpret” is mentally to present oneself with a Meaning. And Meanings sort things out in a way that is–we know (today) not how—immune to any further interpretation.” This response is not just a re-use of the term “meaning” a la (6A). When it is seriously entertained, it purports to explain the meaning of signs through the mind’s grasp of entities called “Meanings.” Such a proposal–often dubbed “platonism” by its detractors–is sometimes thought to arise as a kind of self-standing conception at the beginning of philosophical inquiry into meaning; it is what the skeptic, in pointing to the indeterminacy of texts “in themselves” is supposed to be reacting against. But it should be clear at this point that “platonism” can just as well, or better, be given a different pedigree. It can be understood as a late (and desperate) product of a way of thinking which begins with a skeptical thought, a thought which then creates a felt need to explain the general possibility of meaning because it latently sug27

Of course, there is more to say on this point. It must suffice here to note that it is common ground between both Wittgenstein and Derrida, different as they are. 28 “Why No One’s Afraid of Iser,” Doing What Comes Naturally, p. 83.

162 MARTIN STONE

gests, contrary to everyday experience, that meaning is in fact impossible.29 9. A metaphysically occult idea of meaning does indeed seem to be forced upon us here. This is so because we are eventually led to see that in order for “interpretation” to function as a general condition of the possibility of meaning, there has got to be some last or final interpretation–i.e., an interpretation not in need of any further interpretation– which is what we call “the meaning.” In other words, our starting point commits us, after whatever twists and turns, to looking for some item which a meaning-endowing interpretation can consist in, but which, unlike ordinary linguistic signs, will function as a regress-stopper. It comes to seem that only a unique act of mind can do that. Of course, there still remains another bold option: bite the skeptical bullet and accept “the endless movement from sign to sign.”30 But the choice, within the present set-up, seems clear: either (1) indulge a platonistic mythology, allowing that there just are Meanings, remarkable normative entities channeled through a mind; or (2) accept rampant interpretivism, admitting that anything can be made to accord with any interpretation of a text (on some interpretation of that interpretation). Which is to say, for some people today, the choice is clear: We should avoid mythology, and accept the ubiquity of interpretation. So much for the skeptical recital. It presents ubiquitous interpretation as the realistic or demystifying alternative to a suspect metaphysics of meaning. But doesn’t demystification here look unnervingly like decapitation? Hasn’t this line of thought in fact destroyed the very possibility (i.e., of plain meanings) which it was supposed merely to account for? If anything can be made to accord with a text (on some interpretation), it looks like we simply can’t talk about accord or con29 “Latently”: The point here is that this pedigree can be historically accurate even if the implications of the originating skeptical thought are not recognized until later on–even after the “platonistic” moment. John McDowell develops a similar idea in Mind and World (Cambridge: Harvard University Press, 1994), to which I’m indebted here. 30 This phrase is meant to echo Derrida in the context of a similar dialectic. See, e.g., Speech and Phenomena, David B. Allison, trans. (Evanston, Northwestern University Press, 1973), pp. 103, 149; “Sign, Structure and Play” in., Writing and Difference, Alan Bass, trans. (Chicago, University of Chicago Press), p. 292.

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 163

flict, and therefore can’t talk about meaning.31 Are we really to accept that on a clear-headed view of things, nothing really means anything; and that, could we but see to the bottom of things, we should see that our most everyday concourse with meaning (“Please come next weekend”) is unreal? It is just at this point that Fish wishes to make use of the notion of an “interpretive community.” That notion comes into play as an attempt to hang onto the idea that “to understand is to interpret” while avoiding the skeptical consequences that seem to come in the wake of this idea.32 According to Fish, the impression of skeptical consequences–the decapitating “denial of just about everything”33–arises only because we have not really disabused ourselves of a longing for a metaphysical ideal, by comparison with which the available notions of meaning and truth seem disappointing. To purge immodest hope, however, can be at once to allay unfounded fear. We can purge the longing by recognizing that there are no “interpretation-free” facts about meaning. But we can allay our fear by noting that in place of such facts, there is always something coming, for all practical purposes, close enough: a story to be told about our membership in communities of mutual recognition, about how we achieve good standing and credit in such communities, about the sanctions which attach to deviance, and so on. Essentially, Fish’s thought is that is that the source of norms relevant to meaning is the community itself: someone who does not behave (e.g., follow the sign) as the community does is in violation of one of its norms and may justifiably be said to “misunderstand” the relevant text. And given the mutually sustained communal framework, any dues-paying member will find the meaning of “——->” and many other signs to be plain as day: they will go right, naturally. So, in the end, the attack on “theory-hope” is not a destructive one—or it is destructive only of bits of philosophy (“texts which are clear in and of 31 Cf. Ludwig Wittgenstein, Philosophical Investigations, trans. G. E. M. Anscombe (Oxford: Blackwell), §202. 32 The general possibility of such a move was indicated at the end of step (7) above. 33 Fish, “A Reply to My Critics,” The Responsive Community, Vol 12, no. 3 (2002), p. 65; quoted in §2 above.

164 MARTIN STONE

themselves”)34 which we have no need for anyway. Thus, we can come to love interpretation, not fear it. For it suffices to account for all the plainness and stability we could intelligibly ask for to see how meanings are–by a kind of groundless self-enactment of the community–socially constructed and maintained. V. Spectators and Agents, Theory and Practice The foregoing dialectic elaborates two paradigmatic moments which appear in a typical Fish essay. The first moment invokes a notion of possibility: “it is always possible” for a text to mean something else, no matter what–or how richly specified—the context; every text (and context) is surrounded by a space of interpretive possibilities;35 it does not just by itself determine, etc....36 Anemic as the notion of possibility invoked here is–that a doubt is possible doesn’t mean anyone does doubt–this may seem alarming. At the second moment, however, we learn that we needn’t worry about the first moment. For after the interpreter is located within a community, we are supposed to have the materials we need to reconstruct such normative notions as are indispensable to our everyday talk of texts and their meanings. Indeed, it turns out that we not only needn’t worry, but needn’t even be interested, unless we are interested in theoretical questions. For, according to Fish, the reconstruction of “plain meaning” (as an effect of interpretation) would leave everything, practically speaking, as it was: “When you come to the end of the antiformalist road, what you will find waiting for you is formalism.”37 (“Anti-formalism,” as several Fish essays make clear, is another term for “rampant interpretivism.”)38 “Force” in Doing What Comes Naturally, p. 513. See, e.g., “With the Compliments of the Author,” in Doing What Comes Naturally, p. 51. 36 For examples of the appeal to notional “possibility”–or its cousin, the “absence of necessity”— see Doing What Comes Naturally, pp. 296, 512; The Trouble with Principle, p. 271. 37 Trouble with Principle, p. 294-95. 38 See e.g., “Introduction: Going Down the Anti-Formalist Road” in Doing What Comes Naturally, esp. pp. 4-6. 34 35

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 165

Remarks like the preceding one–archetypal in their structure (they tell of a journey and a return, of something lost and regained)–deserve special comment. At such moments, Fish goes so far as to invite us to regard rampant interpretivism as just a transitional step–a self-consuming artifact, if you like–the effect of which is merely to remove some mistaken bits of philosophy, but without consequences for practice.39 On the other hand, Fish doesn’t think that his interpretivism is completely self-consuming; for he clearly thinks that it is to be endorsed, at the journey’s end, as the right view in place of the wrong one. This is evident from the way the foregoing passage continues: “...what you will find waiting for you is formalism; that is, you will find the meanings that are perspicuous for you, given your membership in what I have called an interpretive community” [my emphasis]. To anticipate what I will have to say about Fish’s view, it is worth noting an ambiguity here. Is “your interpretive community membership” within the intensional scope of “what you will find” at the end of the road, or is it merely the general pre-condition—itself unfound or unrecognized—of everything else you will find? The ambiguity isn’t surprising, ultimately, because neither option should sit well with Fish. If communal interpretivism is part of what you will find, in what sense have you come back to “formalism”? If, on the other hand, communal interpretivism is only the (unthought) pre-condition of what you will find, why should anyone accept it as true? How could they? 39 Martha Nussbaum notices this structure in Fish and finds an analogy in the notions of “epoche”–suspension of commitment” and “ataraxia”–“freedom from disturbance” from Pyrrhonian skepticism. See “Skepticism about Practical Reason in Literature and the Law,” 107 Harvard Law Review 714 (1994). This somewhat obscures the philosophical register in which Fish is operating. What Fish seeks to gain freedom from is not, as Nussbaum says, “all normative judgment” (Nussbaum, p. 726), but rather certain philosophical accounts of its possibility. A better analogy for Fish’s gesture of self-consumption would be the use of a similar self-conscious literary strategy in, e.g., the early Wittgenstein: “My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical....(He must, so to speak, throw away the ladder after he has climbed up it.)” Tractatus Logico-Philosophicus, §6.54. On eludicatory nonsense, see James Conant, “The Method of the Tractatus” in From Frege to Wittgenstein: Perspectives on Early Analytic Philosophy, Erich Reck (ed.), (Oxford, Oxford University Press); Cora Diamond, “Throwing Away the Ladder” in Realism and The Realistic Spirit (Cambridge, Mass., MIT Press, 1991).

166 MARTIN STONE

The significance of these questions will become more apparent in a moment. For now it may be remarked simply that, at the last stop (or what he regards as the last), Fish is prepared to find not just practice unchanged, but practice regained through the right bit of theory. This is not place to examine every stop on this road. I limit myself to three observations. These observations will lead to the following conclusion. Fish’s interpretivism would be more satisfying if it were utterly self-consuming: something to be recognized, at the last stop (if not before), as complete nonsense. In this role–as a piece of transitional nonsense–the interest of interpretivism obviously could not be, as Fish thinks, that it shows us the truth about texts and their meaning: nonsense is nonsense. Rather, its interest would be that it shows us something about ourselves, namely (1) that we are sometimes prone to imagine that we are making sense when we aren’t, and (2) that this illusion is connected to our wish to say something philosophical, our wish, that is, for there to be a philosophical perspective on things. The recognition of ourselves as harboring this wish–hence as calling on words like “interpretation” outside the practical settings in which they have their significance–is the last stop. From such a recognition (of our wish to speak philosophically as one that would not be satisfied insofar as what we said made sense), there follows a loss of attraction to philosophical investigation–or not. In any case, my remarks here amount to friendly encouragement to Fish to take another step along the road and not chicken out. My reasons for so encouraging him will shortly become apparent. 1. To begin with, we need to see why the interpretive community story cannot provide a satisfactory account of meaning. Such a story is clearly aspiring to be a kind of down-to-earth pragmatism, as against metaphysically suspect conceptions of meaning. But it is really lacking in the perspicuity it would need to be that. Some of Fish’s commentators have drawn attention to problems in the definition of “community”–what constitutes the relevant community? can there be different but equally “right” answers for different communities?40 if so, can we really not intelligibly aspire to any more full-blooded obCf. Trouble with Principle, p. 295: “Of course, members of other communities will not see what you point to or will see something else, but that’s life.” 40

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 167

jectivity than this?–but the problems I have mind here are much more basic: they arise no matter how the community is spliced. Notice how things appear from the first-person point of view, that of the agent engaged in hands-on transactions with meaning. The question is what the text means, not what other people think it means. For a judge, the question is what the rule requires not what other people (or other judges) think it requires. Of course, it is possible for a judge to ask himself what other judges think, but this is only because someone is asking the hands-on question—someone is having thoughts about what the rule requires. The hands-on question is a critical question, as we may call it, not a sociological one.41 Of course, all of us are, everyday, such hands-on agents. For, as indicated, there would be no sociological questions to ask about texts if there weren’t, in the first place, agent’s whose relation to them is the engaged or critical one. (The sociology of meaning, if there is one, concerns the thoughts of such agents.) In this sense, the agent perspective on meaning is primary: It is possible to think of a world in which people only ask the critical questions, but not a world in which the only questions about texts are sociological ones. Now this is not, just by itself, likely to be perceived as an objection by Fish. He is apt to say that the competent agent has internalized the community’s way of seeing things, and so has no need to consult anything but the “rule itself” as it appears within the relevant communal-interpretive framework. So the perspective of practical agency, Fish will say, is preserved in his story. But things are not that simple. We might ask: How does such a picture manage to be a picture of meaning (with its normativity intact) at all? For that a certain decision is required by a rule consists, according to this picture, in nothing 41 This could be refined to accommodate the fact that in many legal systems, judicially correct judgment involves following precedent, even when prior decisions are “wrong” on the merits. In such cases, the critical question isn’t abandoned in favor of a question about what other judges think. Rather a higher-order norm is applied, namely the norm that precedent is binding; and the critical question is what precedent requires, not what other judges think it requires. Conventional rules–like “drive on the right” are also not an exception to the present point. It is true that the reasons for following such a rule depends on whether other people follow them. But this is different from saying that what the rule means depends on what other people think it means.

168 MARTIN STONE

more than the community agreeing that it is required. How does this differ from the picture of a community merely pretending to agree–or collectively sustaining the myth—that something is required by the rule?42 How, this is to ask, can an agent so much as “agree” that something–anything at all–is genuinely required by the rule, if he is not entitled to the view that the rule requires this no matter what others might think? The most he can say, it would seem, is something about what the community thinks. But no one in the community is any better position. They, likewise, are not entitled to the view that the rule imposes a determinate requirement regardless of what anyone else might think. And since no one in the community is in a position to say anything stronger than something about what the community thinks, the perspective of agency vanishes here. (And of course, when this vanishes, there can also be no “communal way of seeing things,” internalized or not.)43 Thinking about the primacy of the agent-perspective is instructive here. It means that what communal agents are being asked to agree to–the subject of their possible agreement–is, in the first place, what the rule requires, not what the community thinks it requires. The later is a possible question only because there are agents in the community who are not asking it, who have other things on their mind. But the problem, for Fish’s story, is to see what entitles any of community’s agents to represent (to themselves or to others) that the rule (e.g.) “no 42 See John McDowell, “Wittgenstein on Following a Rule,” Synthese, 1984 vol. 58, pp. 325-63 for a helpful exploration of these questions. 43 To put this another way, if talk about a “community framework” makes sense here, it records the fact that insofar as it is possible to speak at all, it must be possible to speak for others (“this is what we call a vehicle”) without having to consult them or do a bit of socio-linguistic research first. But talk of “internalizing” the community’s perspective doesn’t itself put us in a position to see how such “speaking for” is possible once we are obliged to suppose that whether, e.g., something really is a vehicle is just a matter of whether the community agrees it is one. To the contrary, it now looks positively irresponsible to speak for other community members without checking with them first. On the idea of “speaking-for” as inherent in speaking, and on the centrality of this idea to the procedures of “ordinary language philosophy” see Stanley Cavell, The Claim of Reason, ibid., ch. 1, and “Must We Mean What We Say” in Must We Mean What We Say, 2d ed. (Cambridge: Cambridge University Press, 2002).

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 169

vehicles in the park” genuinely prohibits such-and-such events, once it is understood that whether one has gotten things right must ultimately be a question of what the community agrees the rule prohibits. All of this points to the same general conclusion. The ersatz notion of “correctness in judgment” provided by the interpretive community story can’t really sustain the notion of their being meaning (i.e., of there being agents subject to genuine normative requirements) once it is seen, that at the basic level, underneath talk of “what a rule requires,” there is nothing but mere convergent behavior—or (one might again say: §4) mere soundings-off. Fish’s account of meaning, someone might feel, is really no less skeptical than the skeptic’s account it is meant to combat. (As will become clear in a moment, it is also no less “theoretical” than the theoreticism it is meant to combat: §6). The objection advanced here is essentially that the interpretive community story does not make room for the attitudes towards meaning which agents must have if there is to so much as be an interpretive community—a community agreeing in its critical judgments–at all. For the story tells agents that in making such judgments–viz., that the rule requires such-and-such no matter what others might think–they are engaging a philosophical illusion. Given this structure, an understandable way of trying to defuse the objection would be to acoustically separate the judgments of practical agency from the deliverances of “theory.” Thus, it might be said that theory (i.e., the right story about the possibility of plain meanings and correctness in judgment) is one thing, and practice (i.e., the engaged concourse with texts and their meanings) is quite another. According to this rejoinder, an agent can judge (and represent to all the world) that the “no vehicle in the park” rule genuinely prohibits such-and-such events because, in making this judgment, she is acting (and seeing the world) in her capacity as agent, not as a detached theorist of meaning. She is playing the legal game, as Fish is apt to say, not the theory game. The motive here is obvious: Those who judge that such-and-such is required by a rule–required as a plain fact–had better not be those who also see, by means of the right theory, to the deeper level of things, at which it becomes apparent that such notional “requirements” and “facts” are only such by courtesy of interpretation. The theoretical truth about meaning, in other words, had better not get around. For it hardly seems

170 MARTIN STONE

clear how the attitudes agents must have in their practical concourse with meanings could be psychologically stable ones once it does. But is this separation plausible? What is supposed to stop the theoretical truth from getting around? One way of trying to stave off the possibility of reflective conflict would be simply cleaving everyone in two. Thus, it might be said: “Let the truth get around as it will. Still, at any moment, we are either having transactions with meaning qua practical agents or theorizing about it qua knowing-spectators; but these two parts of ourselves can never shake hands, for they may never be present at once.” Of course this looks desperate. Why can’t they be present at once? Simply because agents would be psychologically unstable in their attitudes if they were? That is our point: It argues against a theoretical view which puts agents in the way of such conflict, not against the evident possibility of conflict on the grounds that practices and their agents are, after all, reflectively stable ones. However these problems are to be developed, we might note that it is surprising to see Fish appealing to the split-agent picture as much as he does. For this is precisely a picture of a self-standing realm of “theory,” laying bare the grounds of possibility of practice, yet somehow separated or detached from the judgments of practice. Isn’t this ground-giving just what Fish everywhere says is “impossible,” a hopeless attempt to look at oneself (qua agent) from sideways-on? (Evidently, Fish doesn’t see this, so I will address this question—and not just ask it rhetorically—in a moment.) 2. The inadequacy of the interpretive community story about meaning would explain the matter I mentioned earlier: why Fish’s work is liable to give some readers the impression that it intends a kind of skepticism (§2). Fish’s argument depends on following a skeptical (interpretivist) progression of thought up to a certain point, and then heading off its apparent unacceptable consequences by appeal to the notion of “community” as a source of (“always-already”)44 interpre44 Fish often finds his critics to be mistakenly positing a moment of interpretive freedom which then needs to be constrained, whether by texts, rules, conventions, communities and so on. See e.g., “Working on the Chain Gang: Interpretation in Law and Literature” and “Critical Self-Consciousness” (esp. pp. 458-59) in Doing What Comes Naturally. I insert the parenthesis–“always-already”–to make clear that nothing in my criticism incurs a commitment to this allegedly suspect conception of

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 171

tive stability. The trouble is that although Fish’s intention is a non-skeptical one, the solution remains too much in league with the skeptic, follows him too far down the road. Fish might ask himself whether such an account of “correct judgment” (as that view which the community realizes from among the interpretive possibilities) really squares with the Miltonic notion of correctness as judgment in accord with God’s will (§2).45 It would seem that a notion of correctness as “community agreement” must remain a second-best notion, something falling short of some ideal.46 To my ears, Fish is ambivalent on this point: sometimes he presents the interpretive conditions of judgment as requiring that we understand every judgment as falling short of some ideal; at other times, he presents the interpretive conditions of judgment as requiring us to abandon the notion of such an ideal as illusory (and not just unattainable). 3. How far down the road should one follow the skeptical progression of thought? A satisfactory response to it requires questioning, sooner or later, its very first step, the step at which the notion of a sign or text “considered in itself” is introduced. If one accepts that step unfreedom/constraint. That is, I’m happy to follow Fish in saying that a community’s interpretive framework is always-already internalized by its agents, or that subjects are always already “inscribed” within an interpretive framework. My question is how, on Fish’s story, there could so as much as be an interpretive framework to be inscribed in. 45 Fish cites Richard Rorty, who in response to Alasdair MacIntyre having said “In your view, the worst thing someone can say about the Soviet Union is that it is un-American,” shrugged and replied, “What could be worse?” Fish approvingly glosses Rorty’s response as follows: “I would be hearing in [Rorty’s]...line a thicker statement and a serious question. The statement would be a rehearsal of the interlocking values, investments, and social commitments...we implicitly refer to when we say ‘America.’ The serious question would be, ‘What could be worse than a state and an ideology opposed in every way to everything we cherish and believe in?’” “A Reply to My Critics” in The Responsive Community, op. cit., p. 63. The question seems easily answered: Worse than a state opposed to everything we believe in is a state that conforms to everything we believe in when our beliefs are evil. Clearly, there is something better than being guided by our most cherished beliefs, something most of us would want more: being guided by just and correct beliefs, or at least by not unjust or incorrect ones. Even Hitler (or Satan) was guided by his most cherished beliefs; there’s nothing especially good about that just as such. 46 See, e.g., the quotation from the Preface of Surprised by Sin in §2 above.

172 MARTIN STONE

questioningly, then it will be natural to accept the thesis that to grasp a meaning is to interpret; and if one accepts this thesis, it will be natural to feel obliged to choose between “platonism” and some social-pragmatic story about meaning. But there is another option. We might come to see that we have no use for such notions as signs or texts “in themselves” unless we are trying to give a philosophical account of the meaning of a sign or text. And (taking a hint from the proverbial man suffering from carrying around a heavy rock who found an astonishing solution: drop it) we could simply stop trying to give such an account. What is meant by this can be indicated by thinking about what happens when we give everyday explanations of meaning–i.e., explanations in situations where questions about the “conditions of possibility” of meaning are not in play. Generally, we rely on the responses and uptake of others: we count on them, for example, to follow in the direction of the pointed finger, not the opposite one. Everyday explanations are (thus) directed towards removing or averting such doubts as, under the circumstances, actually arise—not every “possible” doubt, whatever that might mean. In contrast, a philosophical account seeks to explain how the meaning of a sign gets fixed from among all the possibilities. (Remember, what justifies talk of ubiquitous interpretation, for Fish, is not that there are always real doubts—some cases are plain ones—but that doubts are “always possible.”) The burden of a philosophical account, to put this another way, is not simply to rule out such doubts as might, under the circumstances, arise, but to specify the meaning of a sign absolutely.47 Now the notion of a sign or text “in itself” is a natural starting point for such an endeavor. Why? Because this notion is formed by abstracting the sign from our practical concourse with it–i.e., the natural (circumstantial) responses and uptake on which everyday explanations rely. By means of this abstraction, we, in effect, represent doubts which are merely notionally possible (they might arise in some circumstances) as somehow already present to an agent considering the sign. Such a representation is clearly the mirror image of the I owe a debt to Cora Diamond for this formulation. See her Realism and the Realistic Spirit, ibid., pp. 68-69. 47

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 173

“platonistic” notion that grasping a sign’s “Meaning” determines (in the present moment) its application in all possible circumstances, excluding all possible doubt. What the platonist and the interpretivist have in common, then, is the endeavor to give an account of the fixity of meaning, as it were, in light of all the possibilities. (Both express what someone might call “the metaphysics of presence.”) What happens if, in contrast to both, we were to free ourselves of the felt need for such an explanatory endeavor? We should have no use for speaking of signs “in themselves”–save perhaps in the practically useful way that (e.g.) lawyers sometimes do, namely to distinguish between a text (“the rule itself”) and someone’s gloss on it. And if we had no use for such an abstraction, we should also have no use for the thought that there must always be an interpretation that fixes a sign’s meaning. “What gives life to signs,” we will be inclined to say (if we must say something about this), “is that they are part of the weave of our lives. It is we who are the life of signs.” This is to be heard not as a further bit of rock-carrying theory, but simply as expressing that one no longer feels compelled to try to account for the normative aspect of signs by means of whatever materials remain in view after one abstracts from the sort of practical activities and attention which comprise our sign-filled lives. Given this possibility (of rejecting the argument’s first-step), it seems clear that we are not compelled to embrace the interpretive community story in order to head-off the argument’s nihilistic consequences. We might instead come to recognize, in light of those consequences, that the thought “to understand is always to interpret” is–just as it intuitively seemed to be—an absurdity. The ubiquity of interpretation, communalized or not, is not intrinsically plausible. At best, it comes to look plausible as the result of a philosophical set-up which makes it look as if “platonism” and “nihilism”–“theory hope” and “theory fear”—were the only other options. In such cases, the solution is always to figure out what we need to do to scrap the set-up. Here, this would mean asking what has happened–what doubts have arisen—to make an account of the very possibility of meaning seem like something we need. (That there must or could be such an account—a substantially correct one—is of course not something that

174 MARTIN STONE

Fish, for all of his good cautionary advice about “the unavailability of cosmic doubts,”48 ever questions.)49 Why doesn’t Fish recognize that his “interpretivism” is cut from the same philosophical cloth as the “platonism” it would oppose? Recognizing this would require that he see more clearly what is wrong with the platonistic or “foundationalist” idea of “absolutely fixed meaning.” Fish tends to speak as if the trouble with this were just a suspect wish to find a “universal mechanism”–to give assurance to judgment from the outside. But a more fundamental question is whether we can so much as make out what is being said here—“fixed in light of all the possibilities”? If intelligibility (rather than substantial truth or falsity) is the trouble, no satisfaction is to be gained from denying that meaning could be so fixed by asserting that, on the contrary, all meaning is subject to interpretive conditions. To assert the later (i.e., that it is always possible for the text to mean something else) is to join ranks with what one means to be opposing: it is to suppose that one has managed to make sense of the idea of an absolute space of meaning-possibilities, the space of what a text could mean. (It is to entertain “cosmic doubts.”) An example of Fish’s might help to make this clearer: “As yet two plus two equals four has not become...a flash point of disagreement, but it could....Until two plus two equals four crosses someone’s ambi48 See Fish, “Theory Minimalism” ibid., p. 772: “Schlag’s mistake can be seen by considering the nature of the ‘doubts’ he considers ‘requisite’....They are cosmic doubts, not doubts about this or that, but doubts about the entire cognitive structure within which ‘this’ or ‘that’ emerge as objects of inquiry. That form of doubt is not available to situated beings....” Fish ought to have seen this his own interpretivism falls by this axe. My argument here may be expressed, in the terms of this passage, by saying that: (1) rampant interpretivism presupposes the intelligibility of doubt not about this or that text (there are plain cases), but about the possibility of textual meaning as such; and (2) such a form of doubt does not appear intelligible from the point of view of “situated beings”–it requires a notional God’s-eye point of view. 49 “What was required,” Fish writes, “was an explanation that could account for both agreement and disagreement, and that explanation was found in the idea of an interpretive community.” “Change,” in Doing What Comes Naturally, p. 141. Fish is speaking here about introducing the notion of “interpretive community” to address certain issues in literary theory. But at the time this was written, this was also the central notion, for him, in an account of the possibility of the determinate meaning of any text. See, e.g., “With the Compliments of the Author” in Doing What Comes Naturally, p. 43.

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 175

tion, it is a fact agreed on by all the parties, but this doesn’t mean that there are truths above ideology but that there are (at least by current convention) truths below ideology.”50 The passage is virtuostic, as so much else in Fish, in purporting to exhibit how virtually anything, including the supposedly hard facts of mathematics, can be reconstructed as effects of interpretation. (No facts, to put this somewhat less benignly, are capable of getting in the way of Fish’s interpretivism.) But to pursue further the intuition of “possibility” invoked here (“it could...”), we might ask: From what point of view does this possibility— that “two plus two equals four” could (intelligibly) cross someone’s ambition—appear? From what point of view does it appear, for that matter, that a case which is perfectly plain under a rule could (tomorrow) come in for doubt?51 Certainly not our point of view as practical agents, at least if “could” means that we can make sense of these possibilities. (And it if doesn’t mean that, what does it mean?) After all, some cases are so clear that to “doubt”—or to try to doubt—merely announces to others that you are not someone with whom it is going to be possible to speak; and if there is no speaking with you, there is no disagreeing, or feeling crossed by you, either. God’s point of view then? A deeper diagnosis of what is wrong with foundationalism–its reliance on such a notional point of view–should have led Fish to see that his own interpretivism falls by the same axe. Interpretivism is the negative image of foundationalism. And a general moral to be grasped here is that one does not get rid of philosophical foundations by denying that there are any. That is merely a way of preserving the structure of the question–i.e., the demand for an explanation of how plain meaning is possible–which foundational views take themselves to be answering. VI. Practice, or the View from Straight-on To conclude, it is worth recalling a point mentioned earlier (§4), namely that the everyday idea of interpretation is at home in cases of real doubt or uncertainty–cases which occur against a background of 50 51

Trouble with Principle, 271 [emphasis on “could” is mine]. Cf. “Force” in Doing What Comes Naturally, p. 512.

176 MARTIN STONE

“plain cases” in which there is no call for interpretation. I have been arguing that to assert (in contrast to this) that an interpretation is always required because no text is immune to possible doubts, is essentially to entertain, in league with one’s philosophical opponent, the idea of a “philosophical perspective” on meaning–an account of how meaning is fixed from among “all the possibilities.” When we give up this idea, we can return the word “interpretation” to its ordinary use, whereby interpretation is sometimes needed and sometimes not (it is no longer a general requirement). By the same token, we can return the expression “text itself” to its ordinary use, which marks a distinction between a text and an interpretation or gloss someone has put on it. For all of his pragmatic aspiration, Fish misses this possibility—the possibility, you might say, of trusting in how things ordinarily appear. At the last stop, it seems he wants there to be a philosophical perspective on meaning, an account for him to be “right” about; that idea–philosophy’s traditional idea of itself—never ceases to attract him. His attraction to it, and his blindness to the intellectual possibilities it occludes, are ironic, of course, because the rejection of philosophical dogmas–including the dogma that there must always be room for good answers to philosophy’s ‘how possible” question—is one of his big themes. The conflict I’m describing comes directly into view in remarks like this: “Theory’s project–the attempt to get above practice and lay bare the grounds of its possibility–is an impossible one”.52 How, we will want to know, is the general account Fish seeks to be described if not as an attempt to give grounds of the possibility of our concourse with meaning? Isn’t that just what is in the offing when a question like “What makes it the case that this action is in accord with the utterance ‘a diet Coke please’” meets with answers like “some community-informed interpretation,” as opposed to the sort of answers that actually figure in our practical activity, answers which merely direct attention to features of the text in question or to the situation in which it was uttered? In contrast to what I have called the everyday use of interpretation, “interpretation” as an ubiquitous requirement begins to look like another name for–an occupant of the same explanatory place as–divin52

“Change,” in Doing What Comes Naturally, p. 156.

THEORY, PRACTICE AND UBIQUITOUS INTERPRETATION 177

ity: it is it the terminus of all other explanations of meaning. So this looks like “theory’s project” more or less as Fish describes it: not a “universal mechanism,” to be sure, but still an attempt to get above practice and exhibit its grounds of possibility; an attempt, in Fish’s words, to look “sideways at oneself.”53 (Looking from straight-on—to continue the metaphor—it will appear that an interpretation is needed only when there is some actual doubt, not the mere notional possibility of doubt, to be cleared-up or averted.) Fish’s mistaken sense of his own philosophical radicalism comes out again when he quotes a remark of Hilary Putnam’s which is a modern variation on Plato’s myth of the cave: “What if all the philosophers are wrong,” Putnam asks, “and the way it seems to be is the way it is?” Fish approvingly glosses the question like this: “What if the answers philosophers come up with are answers only in the highly artificial circumstances of the philosophy seminar, where ordinary reasons for action are systematically distrusted and introduced only to be dismissed as naive?”54 But now it should be plain that, with respect to the refusal of philosophical tradition gestured toward in this question, Fish’s interpretivism is on the wrong side. For that an interpretation is required in every case is not how things appear from the (naive) point of view of practical activity. (Imagine the server, with no special (circumstantial) prompting, replying, “I interpret that to mean you’d like a certain beverage now.” Is she mad? Or just doing a bit of literary theory?) In fact, it is only in the caves of the seminar room that the term “interpretation” shows up as part of an account of how it so much as possible for certain signs to be meaningful and hence to afford agents with reasons for action. Thank God, wise is He, for that. From the straight-on perspective, the answer to a question (should it arise) like “What makes it a fact that he ordered a diet Coke?” must surely be: not a communal interpretation, but rather (perhaps after reminding the questioner of the richly-woven world of restaurants, menus, orders, servers, meals, preferences, beverages, and so on) “look, you can that’s what happened yourself.”55 A final point to be “Theory Minimalism,” ibid., p. 772; see also Trouble with Principle, pp. 305-6. The Trouble with Principle, p. 294. 55 I’m indebted to David Finkelstein’s voicing of a similarly flat-footed response in “Wittgenstein on Rules and Platonism” in The New Wittgenstein, ibid. 53 54

178 MARTIN STONE

dealt with involves being clear about the status of the material I have put in parentheses here. The parenthetical material reminds of us the practical situation or setting. But the point is not to say that it is really, in the end, practice which determines a text’s meaning, or which mediates between a text and what accords with it. That would be another bit of theory; and we should then have to ask (a la §4), whether we really have a notion of “practice” as behavior which is describable without attributing “meanings” to anyone, and also (a la §5) whether behavior, so described, is really sufficient to give us the notion of “accord,” and hence of meaning with its normativity intact. After everything I have said, it must be clear that I don’t mean the parenthesis in this way, as finally the best theory of all—the “practice theory”! The parenthesis is there rather to remind us that from the practical (engaged, straight-on) perspective, no general gap between an order and what accords with it appears. (It only looks like there is a general gap when we consider the order “in itself.”) Since there is no general gap, there is no general need for the explanatory (gap-filling) work of “interpretation,” “practice” or anything else. Speaking as practical agents, what we shall say is simply this: sometimes there is a gap (and an interpretation is useful in bridging it) and other times there isn’t (and then there is no call for interpretation). Of course, this is not a philosophical remark. It is merely something that practical agents can see and (often enough) agree on, in just the way that they (often enough) agree about such things as his having ordered a diet Coke. No explanation of the possibility of such agreement in judgments, or the possibility of “plain meaning” is on offer here, no attempt to go deeper than the fact that we do (often enough) agree. Is an explanation therefore missing here? A certain traditional philosopher is apt to be certain it is. Of course. But Fish? He ought to have said that it isn’t compulsory to think so. For from the primary standpoint of practical engagement, an explanation is not merely not needed–it isn’t even wanted. Despite his anti-theoreticism, Fish never really gets this intellectual possibility fully into view. Yet, from much of what he says, I think it is what he was after.

8

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

Law as a Reflective Practice Scott Hershovitz In “Theory, Practice and Ubiquitous Interpretation: The Basics”, Professor Martin Stone accomplishes something rare for an analytic legal philosopher. He takes Stanley Fish seriously. This is not an easy thing to do, as Fish’s caustic style and literary flair often obscure the depth of his argument. Not only has Stone taken Fish seriously, he has managed to say something deeply illuminating about Fish’s project. Stone advances two criticism of Fish’s work, one small, one big. The small criticism is only small by comparison, because it is nothing less than a challenge to Fish’s central claim—that interpretive communities are the source of the standards by which interpretations are properly judged. While the small criticism addresses Fish’s solution to the task he sets himself, the bigger criticism calls the task into question. According to Stone, if Fish understood the full implications of his own arguments, he would never ask the question which leads him to the interpretive community view in the first place. This comment explores Stone’s criticisms of Fish’s work. I shall suggest that both criticisms fall short of their marks, that Fish’s project can withstand them. However, this should not obscure the importance of the contribution that Stone’s essay makes. It deeply illuminates the nature of Fish’s project and its weak points. I. Preliminaries Before we can assess Stone’s criticisms of Fish, we need first to understand the views that are under attack. I shall present only a short summary here due to the thorough job that Stone does in his essay. As Stone tells us, Fish rejects both foundationalism and skepticism.

180 SCOTT HERSHOVITZ

Foundationalism here is the idea that it is possible to supply or construct a theory of what makes judgments in a given domain correct that is independent of those judgments. In this essay, we will be most concerned with judgments about meaning. The skepticism Fish is concerned with denies that it is possible to speak of getting things correct and counsels instead focusing on what people take to be correct. All this is quite arid, so an illustration is in order. Suppose Aaron has an interpretation of Hamlet on which the protagonist is paralyzed by indecision (originality is not Aaron’s forte). How do we know whether Aaron’s interpretation is correct? Well we might think that we should compare Aaron’s interpretation to the text of Hamlet to see if he has got things right. But for a variety of reasons we do not have the space to explore here, Fish says we cannot do this. We cannot consult the text to see if an interpretation is correct, because our only access to the text is through interpreting it (at least according to Fish). If the text of Hamlet cannot be consulted to determine the validity of an interpretation of Hamlet, it might seem that there is nothing to be consulted which could provide assurance that an interpretation of the play is a good one. Indeed, this is just what the skeptic concludes. The skeptic concludes that no interpretations of Hamlet are really correct or incorrect; some are just taken to be correct. But Fish rejects this skeptical view of things as well. One of the important contributions of Stone’s paper is to clarify just what it is that foundationalism and skepticism share in common, and hence just what it is that Fish is rejecting. Stone identifies the following premise lurking behind foundationalism and skepticism: (P): Our entitlement to see one of two conflicting judgments as objectively correct requires some means, independent of those judgments, for validating one or another of them as correct. Foundationalists think that judgment-independent means of verification are available; skeptics, while agreeing that such means are necessary to validate judgments as correct, deny that that they are available. Fish thinks that he can provide an alternative account of the correctness of our judgments. Although he rejects foundationalism, he aims to avoid skepticism. His solution is that interpretive communities determine whether an interpretation is correct or incorrect, successful or unsuccessful. So in order to know whether Aaron’s interpretation

LAW AS A REFLECTIVE PRACTICE 181

of Hamlet is correct, in Fish’s view, we do not ask whether it matches the text of Hamlet, we ask instead whether Aaron’s interpretation is accepted by the community as correct. Now notice that this method of determining the correctness of judgments is not judgment-independent. Whether a judgment is correct depends on whether the community accepts the judgment as correct. This, of course, depends on what judgments the community makes. Thus, the interpretive community view allows us to speak of correctness and incorrectness where someone who accepted (P) would not. The view is neither foundationalist (because it does not posit a judgment-independent test of correctness), nor skeptical (because it preserves a notion of correctness). This is the view that Stone’s small criticism is targeted at. II. Interpretive Communities and Normativity Stone’s small criticism consists of three objections to the interpretive community view. Stone’s first objection to the interpretive community view is that from the perspective of an agent engaged with a text, the primary question always has to be what the text means and not what people think the text means. In Stone’s view, it simply does not make sense to wonder about what people think a text means unless someone out there is wondering about what the text means. Now as Stone notes, Fish is not apt to consider this an objection, because he will say that people simply internalize the community’s perspective. They will of course speak as if they are addressing the hands-on question of what the text means, even though the standard for correctness is what the community thinks it means. I think that Stone is right that Fish will not count this as an objection, and I want to sharpen just why he will not. According to Fish, interpreters share strategies for writing not reading texts.1 Now this is an odd metaphysical view, but it is Fish’s. Because of it, he is likely to say that people 1 “Interpretive communities are made up of those who share interpretive strategies not for reading but for writing texts, for constituting their properties.”

STANLEY FISH, IS THERE A TEXT IN THIS CLASS: THE AUTHORITY OF INTERPRETIVE COMMUNITIES 327

(1980).

182 SCOTT HERSHOVITZ

address the hands-on question of what a text means only after the text has been constituted by communal interpretation. Stone’s second objection is much deeper. Stone argues that in Fish’s picture there is no normativity for communities. There is no such thing as the community getting Hamlet right or wrong. Whatever the community accepts as a correct interpretation of Hamlet just is correct on Fish’s view. As Stone points out, this makes it impossible to distinguish a community which agrees on the meaning of a text from one which is, perhaps, merely pretending to agree. Stone is correct to suggest that there is no normativity at the community-level on Fish’s view, but once again Fish is unlikely to see this as an objection. After all, what Fish is concerned with is normativity for individuals. That is, he wants to explain how individuals can get things right or wrong, and he may see no need to tarry over the fact that his answer obviates any sense in which an entire community could be right or wrong. After all, Fish might fairly ask why it is important to maintain an ability to distinguish between a community which accepts that Hamlet is paralyzed by indecision and a community which merely pretends to accept it. Such a distinction seems quite foreign to our literary practice. Stone’s most serious objection is that there is no normativity at the level of individuals in Fish’s picture either. If Stone is right about this, then Fish will be defeated, after all it is just this sort of normativity that Fish aims to account for. Stone’s argument is simple. He says one will not be entitled to say, “this text means X,” once one understands that whether this claim is true does not depend on facts about the text, but rather on whether the claim is regarded by the community as true. That is, Aaron cannot maintain that his interpretation that shows Hamlet as paralyzed by indecision is correct, once he understands that to be correct in this case is merely to be regarded by the community as correct. While Stone’s final objection would be decisive if successful, I think it fails. Consider the following analogous case. Imagine a group of people who believe that any rule requires whatever God regards it as requiring (which should not be too much of a stretch). On this picture, there is no normativity for God; God cannot be wrong about what a rule requires, because it requires whatever God believes that it

LAW AS A REFLECTIVE PRACTICE 183

does. But for individuals within the group, there can be normativity. They can be right or wrong about what a rule requires (the test of whether they are right or wrong is simply whether their belief about what the rule requires matches God’s). Now if we substitute “the community” for “God” in this story, we end up with Fish’s picture of how interpretive communities provide normativity for their members. I see no reason to accept Stone’s assertion that normativity disappears as soon as people learn that the test for correctness is conformance to the community’s view. As long as it makes sense to talk of correctness and incorrectness, there is genuine normativity in the picture. I am inclined to agree with Stone, however, that Fish’s interpretive community view fails, but on different grounds than Stone suggests. I think the real flaw is that the view fails to provide a plausible account of disagreement within communities. I have developed this argument elsewhere,2 but it is unnecessary to explore it here. Doing so would merely delay getting to Stone’s bigger, more important criticism of Fish’s project. III. Fish and Wittgenstein According to Stone, the deepest problem with Fish’s project is that that the interpretive community view is offered as a “philosophical account” of meaning. A philosophical account of meaning, in Stone’s terminology, is one that “seeks to explain how the meaning of a sign gets fixed from among all the possibilities.” Stone does not believe that we can have such an account, and indeed, the most provocative part of his essay is his suggestion that should stop seeking philosophical accounts. In Stone’s view, everyday explanations of meaning are “directed towards removing or averting such doubts as, under the circumstances, actually arise.” Fish’s interpretive community view, on the other hand, purports to be a general explanation of how meaning is fixed. It purports to secure meaning against all possible doubts, not just the ones we actually have. In Stone’s view, Fish would have been better off if he 2

Scott Hershovitz, Judging Interpretations 54-58 (2001) (unpublished D.Phil. thesis, University of Oxford) (on file with author).

184 SCOTT HERSHOVITZ

had simply rested on his rejection of (P), his denial that judgment-independent means of verification are required in order to be entitled to claims of correctness of judgment. Fish’s view fails, according to Stone, because like foundationalism before it, it seeks to do too much. Stone’s critique of Fish draws heavily on the work of the later Wittgenstein. Space does not permit a full exploration of Wittgenstein’s views, but we need a cursory examination in order to place in context the worry that I shall raise about Stone’s invitation to stop seeking philosophical accounts of meaning.3 The relevant portions of Wittgenstein’s writings appear mostly in the Philosophical Investigations;4 they are sometimes known as the rule-following remarks. Wittgenstein is investigating, among other things, how we know what actions accord with rules, and he focuses on rules of mathematics and language. He is concerned to show that one does not always need to interpret language in order to understand it. By “interpretation”, Wittgenstein means something a bit idiosyncratic: replacing one expression of a rule with another.5 Sometimes interpretation of this sort is called for, Wittgenstein admits, but it cannot be the case that we always need to interpret to understand, because interpretations themselves are just further bits of language. If we could not at some point understand language without interpreting it, we could never understand language at all; we would be stuck in a endless regress, needing to interpret our interpretations. Clearly we can understand language, and so it must be possible to understand without interpretation. 3 For a detailed account of Wittgenstein’s rule-following remarks and their relevance to law, see Scott Hershovitz, Wittgenstein on Rules: The Phantom Menace, 22 OXFORD J. LEGAL STUD. 619 (2002). 4 LUDWIG WITTGENSTEIN, PHILOSOPHICAL INVESTIGATIONS (G.E.M. Anscombe trans., Blackwell 3d ed. 1998)(hereinafter “PI”). Following convention, citations will point to section numbers. 5 PI 201. This is not what we mean by “interpretation” generally. When we interpret a painting, we do not replace one expression of the painting with another. Similarly, interpreting a law need not involve replacing on expression of the law with another, though it may. The idiosyncratic use of “interpretation” is another cause for concern in extending Wittgenstein’s arguments about interpretation and language beyond their intended scope.

LAW AS A REFLECTIVE PRACTICE 185

As Stone explains, Fish’s rampant interpretivism (to use Stone’s phrase) is in part a consequence of his view that one can always raise doubts about the meaning of a text.6 The interpretive community view is meant to explain how meaning gets fixed from among all the possibilities. But Wittgenstein rejects the idea that “secure understanding is only possible if we first doubt everything that can be doubted, and then remove all those doubts.”7 Explanations of meaning, he says, “remove or avert a misunderstanding . . . that would occur but for the explanation; not every one that I can imagine.”8 In the normal case, we understand without an explanation of meaning. Importantly, we may not be able to give an explanation of meaning if called upon to do so. According to Wittgenstein, we often follow rules or use language without being able to provide reasons to justify what we do. We might say that, in this way, our use of the rules of language is unreflective. I shall not take issue with the Wittgensteinian picture of language that Stone draws upon. Rather I want to grant that it is true of language in general, but question whether it teaches us anything about law or literature, the two areas Fish is most interested in. I shall argue that it does not, and moreover, that we cannot take up Stone’s invitation to stop seeking philosophical accounts of meaning within literature and law, even if we might be able to do so for language in general. Even if Wittgenstein’s picture of rule-following within language and mathematics is correct, it is not necessarily true of all rules, not even all rules within language or mathematics. Colin McGinn, in outlining Wittgenstein’s position, writes ... where the bringing to bear of reasons is appropriate the possibility of doubt is correspondingly real. For when reasons are appropriately brought to bear we are dealing with beliefs and actions which are reflec6 And in this way, Fish shares a lot in common with the skeptic Saul Kripke sees in Wittgenstein’s writings. Kripke’s skeptic also exploits that fact that, on some interpretation of a rule, any action can be made out to accord with it. See SAUL

KRIPKE, WITTGENSTEIN ON RULES AND PRIVATE LANGUAGE: AN ELEMENTARY EXPOSITION 7 8

PI 87. Id.

(1982).

186 SCOTT HERSHOVITZ

tive, with respect to which reasons may be weighed and evaluated; and where the question of the goodness of a reason is appropriately raised it will be appropriate to entertain doubts about the quality of the reasons one has. But when an activity is as undeliberative as using language is it lies outside the sphere of the reason based and doubt ridden.9

Our question is what do Wittgenstein’s remarks about rules tell us about law and literature? If McGinn is right, and I think he is, they tell us almost nothing about law and literature. Both are among our most reflective activities. Consider the lawyerly activity of distinguishing one case from another. Faced with a rule in a prior case adverse to one’s client, a lawyer seeks to distinguish the prior case from the one at hand. In a case I recently worked on, the following question arose: is an admission made in the course of a summary judgment proceeding binding on a litigant? Multiple authorities stated clearly that such admissions were binding in the jurisdiction. However, the inquiry could not stop there. I had to anticipate how opposing counsel might distinguish the prior cases so as not to be bound. One difference jumped out—the previous decisions were all made during summary judgment proceedings; none of the cases addressed whether or not the admissions were binding in subsequent phases of litigation. But then, none indicated that admissions would not be binding in subsequent phases either. To convince a judge, we would have to be prepared to offer arguments for applying the rule so as to bind litigants in all proceedings, rather than only in summary judgment proceedings. The practice of distinguishing cases demands that lawyers offer reasons for applying a rule one way rather than another. This sets law apart from our normal use of language. To see this, think of the skeptic Saul Kripke sees Wittgenstein’s writings. Kripke’s skeptic is engaged in a practice of distinguishing much like the lawyer’s practice. Faced with the problem 68+57, the skeptic tries to distinguish the instant case of addition from all previous cases in order to convince us that the answer is 5 rather than 125. He tells us that these numbers are larger than we have faced before, and that with such large numbers, 9 COLIN MCGINN, WITTGENSTEIN ON MEANING 22-23

(1984).

LAW AS A REFLECTIVE PRACTICE 187

the correct answer is always 5. He challenges us to give a reason for thinking that the answer is 125, and the force of the skeptic’s argument is our inability to do so. Our short patience with Kripke’s skeptic comes, I think, from his attempt to distinguish one case from another when doing so does not make sense because it has no place in the practice. The Wittgensteinian reply to the skeptic is that we do not need reasons; it is enough that we apply rules in the way we find natural given our training. In law, however, distinguishing one case from another and giving reasons for going on in a particular way are central parts of the practice. Our inability to give reasons for the way we apply rules in language or mathematics need not trouble us, but a lawyer who cannot give reasons for the way she recommends applying a legal rule is inviting trouble. Law is, to use McGinn’s words, reason based and doubt ridden. One last thought experiment will help us to see that the Wittgenstenian picture of language has limited applicability to law. Wittgenstein explains that underlying our ability to use language is the fortunate fact that we all find it natural to go on in the same way despite our inability to justify how we go on. He writes: Disputes do not break out (among mathematicians, say) over the question whether a rule has been obeyed or not. People don’t come to blows over it, for example. This is part of the framework on which the working of our language is based . . . 10 As an experiment, let’s rework this passage. Disputes do not break out (among lawyers, say) over the question whether a rule has been obeyed or not. People don’t come to blows over it, for example. This is part of the framework on which the working of our law is based . . .

The rewritten passage is absurd. Not only do disputes break out among lawyers over whether rules have been obeyed or not, these disputes are central to the working of law. The point of this long digression into the applicability of Wittgenstein’s picture of language to law is to suggest that the strategy Stone adopts to critique Fish’s project has limited appeal. Even if 10

PI 240.

188 SCOTT HERSHOVITZ

Stone is right that we cannot and need not seek philosophical accounts of meaning in general, he may not be right about law. When it comes to language, we may not need an account which explains how meaning gets fixed from among all the possibilities; such a thing may not even be possible. But Fish’s search for a means of verifying judgments may make sense for law, given law’s status as a reflective practice, as a practice in which reason-giving and doubt have a central role. Fish intends his interpretive community view to apply generally, to be an account of meaning not just within law and literature, but everywhere. But given a more modest scope, Fish’s project can survive the objection Stone offers. In the end, Stone’s major contribution is demonstrating that Fish has overextended himself, that he has strayed too far from his roots in law and literature.

9

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

On Reflective Practices and >Substituting for God= Martin Stone 1. Stanley Fish says an interpretation is required if anything is to mean anything: Communications of every kind are characterized by exactly the same conditions B the necessity of interpretive work...and the construction by acts of interpretation. 1 This sounds alarming B of every kind? B but it is nothing worrisome, Fish says, once we realize that interpretation always take place in a community.2 Meanings are as plain as anyone could want B even if not Aplain in and of themselves@3 B when the community moves in uniform step. Indeed, from this same idea, Fish suggests, we get all the resources we need (or have ever had) for talking of people getting things objectively right or wrong. Such talk comes to nothing more mysterious than this, that interpreting agents are everywhere subject to correction by that larger agent of meaning, their Community. Fish B not alone among theorists today B calls this Apragmatism.@ In my essay AOn Theory, Practice and Ubiquitous Interpretation,@4 I said this doctrine is strange and not any down-to-earth kind of pragmatism: It is exactly as strange as the Afoundationalist@ or Aformalist@ accounts of meaning which Fish believes he is combatting. For, grasped at the right level of depth, what is wrong with those accounts 1 Stanley Fish, AWith the Compliments of the Author: Reflections on Austin and Derrida,@ in Doing What Comes Naturally (Duke University Press, 1989), 43-44. 2 Fish, Doing What Comes Naturally, ibid., p. 83. 3 Fish, Ibid., p. 513. 4 In this volume; also in Gary Olson and Lynn Worsham (eds)., Postmodern Sophistry: Stanley Fish and the Critical Enterprise (State University of New York Press, 2004).

190 MARTIN STONE

is also what is wrong with Fish=s doctrine: They are endeavors to imagine how the meaning of a Asign@ gets fixed from among all the possibilities (as it were), after one dislodges, or abstracts away from, the very standpoint which is needed to identify a sign as meaning something at allB viz., the practical standpoint of agents, those employing the sign or having concourse with it. There is good news, however. Since Fish already has a lively feeling for what is wrong with the doctrines he is combatting, he has only a small step to take B a slight deepening of his best ideas B in order to abandon completely all the strange talk about ubiquitous interpretation for which he is known. He is close; and it even emerges, in a response written by Fish, that B in saying that interpretation is often superfluous B I was expressing his view: Better, says Stone, that we Areturn the word >interpretation= to its ordinary use, whereby interpretation is sometimes needed and sometimes not B it is no longer a general requirement.@ Once again I find myself in agreement with an argument mounted against me and inclined to claim it as my own.5

By all means, claim it! I=m hardly surprised by Fish=s recognition of his own view in what I said (even if it reverses the thesis for which he is known!) since expressing his exact view (expressing it exactly) was just my intent. When the critical point is already present in the form of a conflict someone is in (when the conflict is a philosophical one), all that is needed is a perspicuous representation of their own doctrine B one which manages to make apparent the connections among the different conflicting thoughts involved. 5 See Stanley Fish, AOne More Time@ in Olson and Worsham (eds.), Postmodern Sophistry, op cit., p. 280-81. Fish essentially retracts his view, it seems to me, by saying that the Ainterpretive community@ idea was meant to be part of a Asociological@ account, not a constitutive account meaning (p. 277). Two points about this: First, it is exactly the distinction I said needs to be drawn here: viz., one between (A) questions about what people believe or how they take things (even whole communities of people) on the one hand, and (B) questions concerning what is so. Second, one can find plenty of passages in Fish=s previous work which appear to argue against both the need and even the possibility of drawing this very distinction.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 191

Fish makes only one small mistake: My argument wasn=t Amounted against him.@ His courtroom style requires two parties. Mine was an argument against something I knew he didn=t mean (something he will perhaps not ever have meant), hence an argument in favor of him, or his better half. Scott Hershovitz wasn=t convinced, however, and his objections to my argument now threaten to keep Fish=s conflict alive B if not for Fish, at least for others holding (or not holding as the case may be) similar views.6 2. The main issue is this. Hershovitz seeks to confront my argument with differences between Law and Literature on the hand, and other (Aunreflective@) activities of Ausing language@ on the other. My argument, he feels, relies on a AWittgensteinian picture@ which concerns the later, but which Adoesn=t teach us anything@ about Law and Literature, the Areflective@ activities AFish is most interested in.@ The upshot, for Hershovitz, is that my critique needs to be reigned-in: On the main topics, or his main topics, Fish=s view Asurvives@ it. Of course, this must mean, happily, that my critique survives in some down-sized way too. It doesn=t dispose of Fish=s interpretivism completely, but it does show, Hershovitz allows, that Fish Ahas overstepped his bounds.@ My essay mentioned Wittgenstein only in a few footnotes because my aim was to give a perspicuous representation of Fish=s doctrine, not to talk about one. If I mis-described something or drew a faulty inference, it should be possible to see this without bringing Wittgenstein into it. But now I feel I should say that I wasn=t relying on any representation Wittgenstein makes about Alanguage-use,@ because he doesn=t, as I read him, make any.7 Indeed, he harps on this point throughout his work. He has fewer positive characterizations of what he is up to, but one of them is this:

6 Before diving in, a word of thanks: The length of this response to Hershovitz shows just how provocative (and clarifying) his comments were. As will be seen, he connects my criticism of Fish to questions about law which need to be considered on another occasion. You=ve got to be grateful to a commentator who gives you that can of helpB I truly am. 7 See my AWittgenstein on Deconstruction,@ in Alice Crary and Rupert Read (eds.), The New Wittgenstein, Routledge Press, 2000.

192 MARTIN STONE

A perspicuous representation produces just that understanding which consists in >seeing connexions=....The concept...is of fundamental significance for us. It earmarks the form of account we give, the way we look at things.8

Remarks like this are central to the way I understand Wittgenstein: What does he mean in depicting his own originality so emphatically as a particular Aform account@? He wishes to say, in part, that one would be missing his thought, or what is original in it, if one read him as making some representation about Alanguage-use@ which is self-standing in the sense that it can be extracted from his work and Aapplied@ to other topics. This is connected to the fact that the Aform of account@ in question is one which begins with the words of another. Or (better put) with the words of a not-me: with a Atemptation to say something@ (PI §254) which is not fully owned but rather presented as troubled B Ain some sense,@ Ain a queer way,@ Ait were as if@ B as expressing conflict. This has a number of implications which Wittgenstein stresses: that the sort of Arepresentation@ sought is one responsive to intellectual difficulties informing a certain dialogic context (PI §§133, 132); that (therefore) almost no remark of his stands-up by itself outside of these dialogues; that it is therefore pointless to make arguments in philosophy which depend on the authority of anything he says (the right way to use his work is rather Ato have thoughts of one=s own@: PI p. viii); that the kind of instruction his work offers is only that of a Aseries of examples@ which Ademonstrate a method.@ (PI §133) A further implication might be mentioned here: That a AWittgensteinian picture of language doesn=t teach us anything about law@ is exactly rightB but only if one adds that it doesn=t fail to teach us about it either. For it doesn=t teach us about something other than law either. Its teaching us something in that way (about anything) isn=t B if we=ve begun to understand it B even in question. I=ll need to develop this a little (below). Hershovitz=s accommodation of my claim and Fish=s claim has a classically judicious feel to it B survival is assured to each in his proper sphere B but I think it also has the Solomanic madness to it, 8 Ludwig Wittgenstein, Philosophical Investigations, G.E.M. Anscombe, trans. (Blackwell, 1958), §122 [hereafter API@].

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 193

which threatens to leave no one happy after the issue is cut in two. So I=m going to object: That AFish has overstepped his bounds@ is the one thing my criticism cannot have shown, since it belongs to the nature of Fish=s view to be quite unbounded. Once it is broken up, or limited to special Areflective@ cases, it has essentially been defeated, or it is no longer Fish=s view. And Fish is everywhere on record as agreeing on just this point.9 How is it supposed that Fish can be defended by drawing a distinction between different kinds of discourse when Fish=s claim (and the very one I was criticizing) is that, as far as Athe necessity of interpretive work@ (§1) goes, no such distinction can be drawn? Hershovitz is making exactly the right point: Everyday communications are one thing, Literature another, and Law still another. And even within these domains B it might be added B there are going to be sub-domains, and within those, cases, not all of which will be the same. The only difference between this and my own (more long-winded) efforts is that this alone gives no clue as to how Fish could ever have got himself into position to deny such things. To succeed in recalling someone to their Aown@ view, you have to take the longer way of course and retrace how they got so beside themself; it won=t do merely to assert the thing they deny, however obvious it is. But these discursive differences are obvious enough. (So is Awhat following a rule is,@ for that matter.) Or rather: If such things were hidden from Fish, this is not because they weren=t in plain view.10 But surely there must be more to it than this, for this is very puzzling! How has Heshovitz managed to think he was defending Fish? And did anything in my critique really suggest that people don=t reflect about law? There is an issue here, I think, on which such summary judgment isn=t available. I myself won=t have an answer, but it would be good to make the question perspicuous. 3. A main source of Hershovitz=s sphere-dividing approach is something Colin McGinn says in explicating Wittgenstein: 9 Besides the quotation which begins this paper, see AIntroduction: Going Down the Anti-Formalist Road,@ in Fish, ibid., p. 4: AThe implications [of the ubiquitous need for interpretation] are almost boundless, for they extend to the very underpinnings of the universe.@ 10 I am alluding here to Wittgenstein=s remarks about the character of philosophical problems (or what he means by that) at PI §129.

194 MARTIN STONE

...where the bringing to bear of reasons is appropriate the possibility of doubt is correspondingly real. For when reasons are appropriately brought to bear we are dealing with beliefs and actions which are reflective, with respect to which reasons may be weighed and evaluated; and where the question of the goodness of a reason is appropriately raised it will be appropriate to entertain doubts about the quality of the reasons one has. But when an activity is as undeliberative as using language is, it lies outside the sphere of the reason based and doubt ridden.11

Hershovitz sees me as drawing on AWittgensteinian@ ideas which pertain B Hershovitz says B to Ausing language.@ He takes McGinn, however, to be of help in explaining why Wittgenstein=s ideas aren=t going to be instructive about Law or Literature B these being Aamong our most reflective activities.@ This is hard to follow. Wittgenstein=s focus is on Ausing language.@ But it isn=t thereby on Aan activity@ contrastable with activites involving doubts and giving reasons, for the later are uses of language too. By the same token, Ausing language@ isn=t, as such, unreflective or reflective. Certainly it can be called an Aactivity@ since people usually speak or write intentionally; the exceptions B like talking in one=s sleep or Aautomatic writing@ B are presumably understood as Alanguage-use@ on the basis of their resemblance to the intentional forms, the Alanguage games@ as Wittgenstein calls them. But there are many activities which involve Ausing language@ (PI §23); and considered as Aan activity,@ using language must surely contrast not with Law or Literature but with, say, hunting and fishing12 B or maybe drawing, if we define Alanguage@ narrowly. Why does McGinn speak of Ausing language@ as an activity Aoutside the sphere of the reason based and doubt ridden@? (Does an activity have to be Aridden with doubts@ to be based on reason?) The Awhy@ in the sentance starting this paragraph, let it be noted, seeks reasons and is, therefore, an interpretive Awhy.@ It comes to asking AHow so?@ or AWhat is he on to?@ Perhaps by Aoutside the sphere of the reason based and doubt ridden@ McGinn just means that this reason-seeking Awhy@ doesn=t come up every time anyone uses language. For sure: 11 12

Colin McGinn, Wittgenstein on Meaning (Blackwell, 1983), pp. 22, 23. As long as you=re not hunting for the truth or fishing for a compliment.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 195

Not all language-use is as obscure this. But then Aspheres@ makes the wrong suggestion here. There are simply cases in which the interpretive questions don=t come up, not a sphere of Ausing language@ in which they don=t. Draw a circle around Alanguage use@: Outside it, I assume, are the less reflective activities, the ones we share with creatures who don=t use language; and inside will be found the activities based on reasons, so that doubts B requests for reasons — could come anywhere here; why not? 13 Whatever McGinn had in mind, the idea Hershovitz takes away (a doubt-free zone within language-use, reflection roaming outside) is, I think, the very opposite of what Wittgenstein was getting at. To make good on this idea, one would need to do some philosophical zoning; judicious laws would be needed concerning where or when doubts may arrise. But ...if there were, not a single signBpost, but a chain of adjacent ones or of chalk marks on the groundCis there only one way of interpreting them?CSo I can say, the sign-post does after all leave no room for doubt. Or rather: it sometimes leaves room for doubt and sometimes not. And now this is no longer a philosophical proposition but an empirical one. (PI §85)

Putting a main point of Wittgenstein=s Arule-following@ discussion at its briefest, we might say this, that with one exception it isn=t for philosophy to say where an explanation or an interpretation is going to be required. The exception is that an interpretation isn=t required everywhere. That is something we can remind ourselves of through philosophy, if something happens to make it loose its obviousness. 14 13 One reason to think they could flows from what was just said: Using language is generally done intentionally, and an intentional action just is B or so it might be thought B one to which the question Awhy,@ in roughly the sense indicated, has application. This is G. E. M Anscombe=s thesis, first stated in ‘5 of Intention, 2d ed. (Cambridge: Harvard University Press, 2000). 14 The question in the first sentence of PI §85 recognizes that (1) it belongs to the kind of thing which a sign bearing meaning is that it is fit for the interpretive question B a meaning-bearing sign operates in the space of reasons. At the same time, we know, from the famous regress which threatens, that (2) the uptake of meaning doesn=t always require an interpretation, whether this be another meaning-bearing

196 MARTIN STONE

Wittgenstein aside, it is hard to see how Hershovitz=s thesis could be right, since it implies there are no Aunreflective@ uses of language in Law. What about when the judge orders the defendant to pay a $100 fine? Or if that still requires reflection to see what is required, what about when the defendant=s lawyer explains it thus: AYou just write a check, Pete, and you=re done.@ The thesis has an unnerving implication: People must reflect and resolve doubts just to comply with the law. Reflecting on this, it seems fortunate that almost all cases of law are very easy ones and require no reflection at all. Or at least this is so in any sense of Acase@ suited for describing the practical reasoning of judges. (This primary sense will have to be distinguished from an Aeconomic@ sense, parasitic on it, meaning a situation of sufficient uncertainty to make it rational to incur the expense of litigation. Without the primary sense, there is no explaining why this is sometimes rational and sometimes isn=t.) To keep to the idea of spheres, one would need to say either that the cases instanced here belong to the sphere of the doubtful and reflection-requiring, or else that they are not Law. Neither option is palatable. This said, I think Hershovitz is making an important point, which can be put B without any activity losing touch with reason B like this: Our reasons for interpreting are different in different domains of discourse. Moreover, as people interpret for different reasons, what they are doing varies accordingly. When a judge interprets the law, for example, he is deciding the issue of right raised by the parties; and that is obviously not an apt description of what literary critics or performers do in interpreting. So Ainterpretation@ isn=t everywhere the same. It might turn out B unsurprisingly then B that in some kinds of discourse (literary criticism, for example), interpretation is, in some good sense, ubiquitous; whereas in others (the Law, one hopes), interpretations either aren=t needed (as they generally aren=t with, say, cooking recipes) or they come to an end when things are clear. To say that these differences in the employment of interpretation are explained by the presence or absence of Areason@ seems to get sign or some mental equivalent. The accommodation of these two points Bfirst of (2) into a formulation which then gets revised to accommodate (1) B is the work of the rest the passage. The rule following considerations B and their conclusion without any sort of Aphilosophical proposition@ B appear in miniature here.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 197

things the wrong way around: It is the reasons people have for understanding these kinds of texts (and hence for interpreting them) which makes them the kind they are. Roughly, a legal, literary or culinary text is one we have a certain use for; and the interest and attention we give it flows from this, from what it is good for. 15 4. Does Wittgenstein have anything Ato teach us about law or literature@? His work seems, recurrently, to contain many negative lessons for legal and literary theoryB one of which begins to appear in the paragraph above. Hershovitz=s different take seems partly due to his having a narrower notion of what it is for a philosophical text to Ateach us@ somethingB he is apparently looking for a doctrine to apply.16 He presents his point through a juridical re-write of Wittgenstein=s well-known remark B ADisputes do not break out (among mathematicians, say) over the question whether a rule has been obeyed or not@ (PI §240): Disputes do not break out (among lawyers, say) over the questions whether a rule has been obeyed or not. People don=t come to blows over it, for example. This is part of the framework on which the working of our law [for Wittgenstein=s Alanguage@] is based. The absurdity of this illustrates Athe limited applicability of Wittgenstein=s picture of language to law.@ For Hershovitz wishes to affirm its negation: ADisputes [do] break out among lawyers...@, and Athese disputes are central to the working of law.@ The re-write is absurd all right. But it might also be noted that its negation B ADisputes do sometimes break out among lawyers...@ B isn=t exactly any less absurd. Who would this be said to and when? The absurdity B either way B is a function of one already there in Wittgenstein=s original. In case we fail to hear it, Wittgenstein is helping us out B drawing attention to it B by adding Acome to blows.@ Mathematicians do not dispute or come to blows....? Was someone thinking they did? Where is this Areminder@ (PI ‘127) a useful thing to have on record? Evidently Ain philosophy@ B in the context of some special intellectual 15 I have developed this in my “On the Old Saw, ‘Every reading of a text is an interpretation’: some comments” in John Gibson and Wolfgang Heumer (eds.), The Literary Wittgenstein (Routledge Press, 2004). 16 This is also something one can learn about from Wittgenstein, as what sort of teaching is involved when we are not given doctrines is one of his themes, beginning with the first sentences of the Preface to his Tractatus Logico-Philosophicus (London: Routledge & Kegan Paul Ltd, 1961).

198 MARTIN STONE

problems. But treated as a self-standing observation about mathematics or language-use, Wittgenstein=s remark is already as fully bizarre as the remark about law (or its negation) which Hershovitz constructs from it. Missing this, it becomes easier to read PI §240 as if it were the sort of thing which Wittgenstein everywhere says his work is not: a bit of doctrine, something endeavoring to inform us about how things are here and thereB in mathematics, language-use, and so on;17 as if PI §240 might really be Aapplicable@ somewhere else besides the law, (to mathematics? linguistics?), or as if the (negated) juridical version of it might come upB where? In the remarks to the entering class? (ANow, Ladies and Gentlemen, I should mention B since your previous activities of Ausing language@ will not have prepared you for this — that lawyers sometimes dispute....@) What is going on here, since it clearly isn=t this? To begin with, PI §240 is far from any doctrine of spheres. For it remembers not that there is an activity called “using language” which, unlike law, is free of dispute, but rather that such disputes as might occur anywhere among talkers take place against a background of agreement among them (cf. PI §241-2). ADisputes do not break out (among mathematicians, say) over the question whether a rule has been obeyed or not@ B this says: AMathematicians do not dispute about that; that is why they can dispute Band sometimes come to blows B about other things.@18 So the parallel point is not that rule-following disputes never break out among lawyers, but that they do not always or everywhere break out. Does that even have to be said? Well, in philosophy it does carry a tune. It comes to saying that there are legally easy cases B something H.L.A Hart had occasion to remember too, when some theory-motivated lawyers seemed to have forgotten it.19 How it comes 17

Cf. Wittgenstein, Tractatus, op. cit (Preface: not Aein lehrbuch@); cf. PI §§128,

599. That is what happened, for example, in the 19th century battle over Cantor’s transfinite set theory. But we needn=t get stuck on particular spheres here, since Amathematicians@ is clearly only an example. Disputes don=t generally break out (among citizens of our noble republic, say) over whether the traffic or health codes have been obeyedB if they do, then those codes should be re-written immediately! And when the citizens do dispute, it often isn=t over the meaning of the code, but B precisely because they agree on that — over what happened in a particular case. 19 H. L. A Hart, The Concept of Law (Oxford University Press, 1961), ch. 7. 18

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 199

about that we forget such things B the ones which stand plainly in view (PI §129) B is the interesting question here. The forms of thought and imagination responsible for it are the ones which find expression in an interpretivist doctrine like Fish=s; and that is the context in which Wittgenstein=s remarks have to be read, if they are to have any work to do.20 Reflecting a little further, however, on how good it is when the law does manage to make a few vital matters easy, to put them beyond the vicissitudes of immediate reflection and dispute, Hershovitz=s further assertion concerning the centrality of deputes among lawyers seems no longer quite trivial. Surely such disputes are peripheral and rare occurrences in the law. This follows from two premises: (1) the purpose of legal rules isn=t so lawyers can have disputes; (2) legal rules usually achieve their purpose. In this respect at least, legal rules are like those of a game like soccer: The disputes which arise tend to be breakdowns in the law=s primary purpose, not what the law is furnished for. Of course, a professional Soccer Umpire might write a book on the casuistry of the Ano hands@ rule, asking what makes one or another applicative decision a true statement of what the Game requires. And 20 I suspect that Hershovitz was probably responding here to recent accounts of Wittgenstein in law journals, for there the idea is sometimes put forward that Wittgenstein has a Apractice theory@ of rules, according to which what a rule means just is, or is constituted by, the way people collectively follow it. The strange result is that rampant disputes about what a rule requires tend to suggest that there isn=t a rule at all; and this of course implies that there are never rampant disputes about what a rule requires. If this is what Wittgenstein is saying, then it becomes only natural to want to respond as Hershovitz does: That may be true in some regions of language use, but it isn=t true in law. (A taste of the discussion can be found in the papers collected in Dennis Patterson (ed.), Wittgenstein and Law (Ashgate, 2004), which contains a paper of my own.) The mistake here lies in supposing that Wittgenstein is putting Apractice@ forward as part of a constitutive explanation of the possibility meaning; his point is rather that from the point of view of practice, no such explanation is wanted or needed (see ‘8 below). When moving between Wittgenstein and the law, it should also be born in mind that Wittgenstein has no special interest in Arules@ as this term is used in law and legal theory. He is not interested, for example in their existence conditions, their relation to reasons, their role in practical reasoning, their availability for systematization, and so on. ARules@ come into his discussion of merely as an example of a type of normative relationB Aaccord@B which is present in any intentionality: in wishes, orders, expectations, beliefs, statements, etc. (PI §437-38).

200 MARTIN STONE

this focus might understandably give someone the impression that disputes are Acentral to the workings@ of Soccer. This is what tends to happen in legal theory, since it is written mainly by lawyers, and since adjudication (not everyday law-following) is the lawyer=s professional concern. Maybe disputes among lawyers are always breakdowns of the law=s primary purpose. But there is no need to lay it down here that they can never be just what the law has in mind. The authors of Amendment VIII to the U.S. Constitution (Acruel and unusual punishments@) may have felt that this formulation was an especially good one because it would oblige people, as circumstances arose, to come into disputes about what was meant.21 In that case, furnishing occasions for disputes among lawyers will have to be counted as part of the law-giver=s purpose B part of the Aoriginal intention@ B and not just (from our knowledge of the Aindeterminacy@ of legal language) a generally foreseeable effect. The only point here is that the propositions Alaw is reflective,@ or Alegal disputes are the central thing,@ are doubtful if they are supposed to mean that such a constitutional rule illustrates the whole of the law.22 This seems to be the trouble with Ainterpretivst@ doctrines generally (even when limited to just the Adoubtful@ regions of Law and Literature!): They tend to make it hard to see the distinctions that are there. Literary interpretation, to take one further example, isn=t a response to any sort of Adoubt@ about meaningB though there is a contemporary tendency to mis-describe it in terms of this essentially legal-hermeneutical construction. Terms like Adoubt@ or Aindeterminacy@ are at home in the law, where they get defined in terms of the operation of applicative judgment; literary texts aren=t doubtful or indeterminate, for the reason that there is no such thing as Aapplying@ them, following their meaning through in a particular case.23 Wittgenstein=s remark that the philosophical task is just to describe things (cf. PI 21 Ronald Dworkin suggests something like this with his characteristic power in AComment@ in Antonin Scalia, A Matter of Interpretation: Federal Courts and the Law, Amy Gutmann, ed. (Princeton University Press, 1989). 22 Amendment VIII presupposes rules defining punishable criminal offenses, for example, to which no dispute-engendering intention can be attributed. 23 I develop this in my AOn the Old Saw: Every reading of a text is an interpretation@ in The Literary Wittgenstein, op. cit.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 201

§124) sometimes rubs people the wrong way B as if it were too modest. But it is to be taken against the background of his showing how hard it is for us B when things get a little complicated B to do just that. 5. Hershovitz is right to call attention to the special attitudes involved in law; it is only odd it is that he doesn=t take this as one of the morals of my criticism of Fish. Fish gives a numbingly general application to the term Ainterpretation@; breaking this up, we can begin to see again as we look for interpretation under our feet rather than over our heads. Hershovitz says he finds my way locating Fish (roughly near Kripke=s Wittgenstein) Adeeply illuminating.@ But his sense that his complaint is with me rather than Fish stems, I think, from a failure to appreciate how strangely general Fish=s question must be, if this way of locating him is correct. Law and Literature, says Hershovitz, are Athe two areas Fish is most interested in.@ If only that were true! Law and Literature are Fish’s main examples, but it is easy to miss the fact that B in the strand of his work concerning Ainterpretation@ B they are only examples. This is easy to miss, not just because Fish elsewhere does wear the critic=s and the lawyer=s hat, but because his questions B viz., Awhat makes this or that reading of a text correct@ B are often ambiguous. They can sound like questions about norms of explanation which are special to a field, questions properly addressed to legal and literary experts. But Fish=s own answer calls on no expert knowledge, and its operative idea B Ainterpretive community@ B doesn=t come out of the study of legal or literary texts.24 His answer shows what the question is. The literary or legal text are merely occasions B important but not essential ones B for raising what Fish conceives to be a quite general question about correctness in interpretation tout court. More than law or literature, it is philosophy which Fish cannot get done with. When he discusses law or literature, he is almost always endeavoring to present a philosophical idea about meaning B by negating it, of course.25 24 The point is not that one can=t encounter the Ainterpretive community@ idea in law school or literature departments, only that it doesn=t come up when one is asking what a literary or a legal text means. 25 This isn=t to say that Fish denies that there is a more nuanced knowledge of Ainterpretation@ to be had. He himself draws attention to different ways of “reading”

202 MARTIN STONE

To illustrate the difficulty which arrises here, consider the question: “What makes this or that reading of the Negligence Rule / of Hamlet correct?@ This formulation is fishy, because it can be used to make different inquiries. Lawyers and critics do have answers to some of them. But the irrelevance of their answers (to Fish=s question) is assured by the fact those answers always refer to other meaning-bearing itemsB e.g., Athe principle laid down in Dreadlock v. Dreadlock,@ or AHamlet=s impression that the ghost looked more sorrowful than angry,@ etc. These answers, in other words, refer to Aa text,@ and, about any text, there is no mistaking Fish=s signature line: It is Aconstructed by acts of interpretation@ (§‘1) and does not constrain them. Fish=s critics don=t always see clearly the implications of this. They take Fish to be saying that their professional judgments lack objectivity, or at least a certain kind of objectivity. That much is right.26 But then they show their misunderstanding by endeavoring to meet this challenge to the status of their judgments by application of one or more standard candidate for establishing hermeneutical validity: intentions, ideal authors, purposes and principles, cannons of interpretation, professional norms, Apractices,@ and so on. None of this advances significantly beyond the original reference to Athe text,@ however. For the salient fact about Atexts,@ for Fish, is just that they are bearers of meaning, and all of these further entities (whatever they are) are also bearers of meaning. What needs to be understood is this. The fact that Fish=s critics go wrong again and again (in Fish=s view)27 isn=t something that just happens here. It is structurally immanent to this discourseB it is the repetitive structure of an interpretive regress. Following out the regress which are instinctB as a sense of Aprofessional correctness@ B among lawyers and literary critics. See e.g., Fish, Doing What Comes Naturally, op. cit., pp. 54, 137. But Fish never allows these differences to get in the way of his general interpretivism; they merely illustrate and confirm his thesis that the way any text is read is a matter of Ainterpretive practice.@ Hence they merely illustrate his point that his question (about interpretive validity) isn=t answered, but only deferred, by well-informed answers coming out of specialized fields. 26 If the theory of the interpretive conditions of all communications isn=t at least purporting to say this much, it isn=t saying anything. 27 See Fish, AWrong Again@ and AStill Wrong After All These Years,@ in Doing What Comes Naturally, op. cit.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 203

which begins with Fish=s response to the well-informed or knowledgeable answer, one will eventually have to see that there is nothing left in the end for his question to be except: How is the meaning of a sign or text possible just as such? (Or: How is it so much as possible that some noises or marks have meaning?) That this is what the question comes to B indeed, what any general question heard today about the Adeterminacy of meaning@ must come to B is apt not to be apparent at first. That is just the problem. Representing matters so that it becomes apparent is thus a step towards the solution. This explains why it was appropriate for me to discuss Fish=s view through the example of how to follow the sign-postB a most unpromising case from which to launch a teaching about law. Such signs generally give no occasion for reflection or dispute, but merely fulfil B even more reliably than the law B their purpose. They are the super-easy cases, deep background. Yet this is just what makes them ideal in the present context. Being easy and beyond all suspicion of needing expert handling, the danger is minimized, in such cases, of confusing the question Fish is asking with one it merely sounds like he is asking. 6. A representation is needed, then, which makes apparent: (1) what the question is, (2) its relation to other questions Fish has been heard as asking, (3) how the project of constructively answering the question conflicts (whatever the answer is) with the strand of Fish=s work which stresses the primacy of Apractice,@ and (4) why the later strand is one we ought to endorse. Before turning to (3) and (4), I want to comment briefly on the way Hershovitz formulates the overall point here (he calls it my Abig@ criticism) because I think his formulation makes it seem unnecessarily threatening to the law=s Areflectiveness@: AIf Fish understood the full implications of his own arguments,@ Hershovitz writes, he would never ask the question which leads him to the interpretive community view in the first place.@ This has a funny air of paradox to it, since it could be reversed: AIf Fish hadn=t asked those questions, he would never have understood the full implications of his arguments.@ But I think I know what Hershovitz is getting at: Elsewhere, he hears me issuing an Ainvitation to stop seeking philosophical accounts of meaning.@ If that was the point, then Hershovitz was right to ask where it leaves us with the law, which has always invited philosophical inquiries, especially those which continue the rea-

204 MARTIN STONE

son-seeking activity of judges. Since I may unfortunately have suggested these formulations, I=ll have to ask for leave to amend here. Hershovitz shows how misleading what I said can be if it isn=t heard B its issue contained B within the present context. First, about Aphilosophical accounts of meaning.@ No: What is at stake is more specific. It is a sought-for account of the very Apossibility of meaning.@ That phrase refers to a problem which becomes urgent when signs or texts Ain themselves@ appear normatively inert: they can be understood or applied like this or like that. Not every Aphilosophical account of meaning@ moves within this framework. Grice=s explanation of Anon-natural meaning,@ for example, relies on the notion of a communicative Aintention,@ something with a normative aspect itself, which Grice isn=t endeavoring to explain.28 Yet I wouldn=t hesitate to call this a Aphilosophical account,@ and I=d even suggest it has things to tell the law. Second, I=m not against anyone asking philosophical questions. If I were, I too would be sawing off the branch on which I=m now sitting. I=m the opposite of against that, thanks in part to Wittgenstein, who has shown how valuable a philosophical investigation of meaning can be, even if it is perhaps also inevitable for us. It is just that, if Wittgenstein is right, what we stand to learn here is not what makes meaning possible, but something about the nature of our attraction to this question, something about ourselves. So part of the value of philosophy, for him, is that it can sometimes change our sense of philosophy=s value or importance.29 This will explain why I want to put the point about Fish a little differently. Here=s my version, spelled-out to make clear what the Atwo@ criticisms (by Hershovitz=s count) are: —Big Criticism: Fish sometimes speaks in a doubtful way about Aphilosophy,@ which he associates with a drive toward Ageneral theory.@ By this criterion, Fish=s question has the unmistakable pitch of philosophy. 28 See Paul Grice, AMeaning@ in Studies in the Way of Words (Cambridge: Harvard University Press, 1989). Grice is endeavoring explain the concept of meaning, not the very possibility that one could mean or intend this rather than that; so he isn=t about to about to be troubled by the signature Fishian objection mentioned in §5. 29 See Wittgenstein, Tractatus Logico-Philosophicus, op. cit. (Preface); PI §§118, 108.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 205

—Small Criticism: Here is why [§7 below] the conflict should be resolved, in this case, in favor of Fish=s suspicion of general theory.

My essay=s portrayal of an endeavor to win a friend over to his better self, I should add, doesn=t imply a merely therapeutic stance toward another; for nowhere than in philosophical criticism is it more true, I think, that the friend is another myself (as Aristotle says). In other words, I could hardly have any interest in criticizing Fish=s views if they didn=t at least have some imaginative charge for me. Wittgenstein writes: AThere is an inclination to say: Every application of a rule is an interpretation of it@ (PI §201). And that is what Wittgenstein=s Ainterlocutors@ do say in various forms. But Wittgenstein couldn=t know of such things B not in the right way B if his source wasn=t himself. Someone might feel I=m being fussy: This is just AStop asking...@ more artfully put. But I think there is a difference B relevant to Wittgenstein B between quasi-legislative pronouncements concerning what people should do or say and a responsive interest in the uneasy words of Aanother@ B a not-me (§2). Moreover, it should be clear that my Abig criticism@ doesn=t say anything, on its own, about what anyone should do. That comes only with the reasons offered for resolving the conflict one way rather than another; and those reasons stand open to discussion. So there aren=t two criticisms here, unless just pointing out that someone is in conflict is a Acriticism@ of them. (Perhaps so, but it=s not that big.) 7. Now about those reasons which stand open to discussion Following some lines of criticism of Kripke=s Wittgenstein,30 I claimed that the notion of Ainterpretive community@ doesn=t do what it is needed to do: It doesn=t show how we can make use of notions like Aaccord@ after the thesis that Ato understand is to interpret@ has brought this into doubt. If it doesn=t do this, and if we nonetheless have to be able to make use of such normative notions wherever there is meaning (that is the premise), it follows that we don=t have a picture of meaning here at all. So Ainterpretation@ isn=t really in the picture either. That is: If everything can be interpreted, then nothing can be B the placement of interpretive activity within a community notwithstanding. 30 See, especially, John McDowell, AWittgenstein on Following a Rule,@ Synthese 58 (1984): 325-63.

206 MARTIN STONE

But why doesn=t Fish=s theoretical use of Acommunity@ make normative notions like Aaccord@ or Acorrect understanding@ available? Originally, I said that no propositions about meaning are left within this picture for members of the community to endorse as their own view. For whatever judgment they make, the theoretical orientation in question tells them that its claim to be considered true depends on the community=s taking it to be true. That puts everyone in an intolerable situation. But perhaps it would make things clearer to start at the other end: Alternative formulation: Start with the fact that members of the community do have a view, and then ask whether the theorist of meaning could himself adopt it B the very view he privileges as determining the correctness of attributions of meaning. He couldn=t. For he is bound to regard it as involving an essentially illusory notion of the determinacy of meaning. He is obliged to think: AIt is not true B what those agents with a view think B that the rule demands this and not that. That all depends on the interpretation that is put on it; tomorrow a different interpretation may prevail.@ So the theorist is confined, by his philosophical scruples, to the role of a detached observer: He can issue reports about the way the community cooks-up its meaning, but he cannot, without abandoning his theoretical orientation, join in the feast. 31

This shows how Fish can resolve his conflict: by adopting, without philosophical scruples, the practical or agent perspective which he emphatically recommends in theory. When one accesses the truth about practical agents in this way B assesses it practically rather than speculatively B one is an agent oneself and joins in the everyday feast.32 Hershovitz sees the point here as turning “decisively” on the claim that “normativity disappears as soon as people learn that the test for 31 David Wiggins formulates a structurally similar incoherence which arises in the region of value-theory, which I=m following here. See ATruth, Invention and the Meaning of Life,@ in Needs, Values, Truth, pp. 97-98 (and more generally, 98-108). 32 Fish=s mistake is to try grasp or represent his point about the primacy of practice from the observer=s position. For what needs to be grasped is not what an agent is qua object of knowledge (someone situated in a community and specifiable in many other particular ways) but qua subject or agent; and this everyone knows without observation or study just by being one.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 207

correctness is conformance to the community’s view,”33 and he presents an argument against this claim. Imagine, he says, a group of people who believe that any rule requires whatever God regards it as requiring. On this picture, there is no normativity for God; God cannot be wrong about what a rule requires, because it requires whatever God believes that it does. But for individuals within the group, there can be normativity. they can be right or wrong about what a rule requires (the test of whether they are right or wrong is simply whether their belief about what the rule requires matches God’s). Now we substitute “the community” for “God” in this story, we end up with Fish’s picture of how interpretive communities provide normativity for their members.

I must admit that the possibility of defending Fish like this wouldn=t have occurred to me had Hershovitz not suggested it. But isn=t there a short answer? Surely it is significant that in order to imagine this possibility, Hershovitz wants to speak about God! Ironically, this defense of Ainterpretive community@ lays bare what is wrong with Fish=s use of this idea. Whenever the minds of human agents are philosophically mis-represented so that it (eventually) becomes apparent that they aren=t going to be up to the normative task at hand, another Agent tends to be sought for whose Ideas have a little more kick to them. While this chapter of philosophy evidently isn=t over yet, the work has started to decline. For Ainterpretive community@ is an inferior Asubstitute@ for AIdeas in God >s Mind@ as an explanation of the normative aspect of meaning: It is apt to make us think that we really understand something here, whereas talk of AGod@ at least had the virtue of making clear that we don=t. Beyond this, I feel as if a game is being proposed without any very clear rules. How am I to draw inferences about what is possible for human beings, or communities of them, on the basis of what is (allowed to be) possible for God? A few of the better known differences might cause trouble when the time comes to replace Him with us: 33 Although I can recognize it as a version of my point, I wouldn=t myself make this formulation Adecisive@ of anything. I=ll have to leave the explanation of its defects for elsewhere.

208 MARTIN STONE

(1) The community=s beliefs are those of its members, its human agents (unless the community is itself some mysterium); (2) In endorsing certain thoughts as those they Abelieve,@ human agents are trying to get something right (and only in special cases the community=s own beliefs); whereas the distinction between “believing” and “decreeing@ doesn=t have any application to GodB i.e., there isn=t anything He believes.). These points locate the problem with the Ainterpretive community@ picture exactly. Insofar as a community Abelieves@ something, its members make normative assessments. They choose some of their thoughts or responses as ones they endorse: this is to be (is worthy of being) believed. But the theoretical orientation challenges this assessment: AYou needn=t believe it; in truth, it is a matter of interpretation.@ An agent who continues to endorse her belief from within such an orientation would be putting herself, precisely, Ain the place of@ God: She now treats the irrelevant biographical fact of her own believing something (or her own inclination to act in a certain way) as a ground or a reason for the other=s belief or action.34 Clearly the rule of secular replacements can=t be: AIf it makes sense to believe it is possible for God, you can believe it about a >community= as well@B though this does seem to be Hershovitz=s premise. I won=t ask what the rules for identifying what is possible for God are in the first placeB I assume that wouldn=t be in the right spirit here. That is, when Hershovitz has people Abelieving@ that AGod cannot be wrong... etc.,@ I assume it would be irrelevant to object that people can only Abelieve@ what it makes sense to believe. The rule here must be that you get to say whatever you like about God; challenges get to be heard only at the substitution phaseB in the Kafkaesque form, ASure it=s possible B only not for us.@ Such rules of play may have an important place in religious practice, but solving our recent intellectual problems is another matter. For one thing, there is no substitution phase B or at least not one in which 34 After the apparent exhaustion of explanations, at the point he calls Areaching bedrock,@ Wittgenstein says: AThen I=m inclined to say: >This is simply what I do.=@ (PI §217) The Ainterpretive community@ story must hear this, when it has the support B or Aauthority@ as Fish says — of the community behind it, as the secular equivalent of divine decree: We say, so it is so.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 209

we get to pipe in B in the religious context. If there were, some strict rules of inference would be needed to keep from going off the deepend. It mustn=t turn out, for example, that AHe is present in the bread@ is a possibility for us, just because it finds application in beliefs about God.35 Liberal deployments of the method of divine analogy wouldn=t get us far anyway, since the more Apossibilities@ it proved, the less creditable any would be. How tight should the substitution rules be? Just tight enough, it stands to reason, so we could skip the step about God and consider directly the question of what makes sense for us B that is, what makes sense. 8. Let me return then to the attempt to understand diagnostically why God is coming up here. It has to do with a kind of philosophical account apt to fancy itself a down-to-earth pragmatism, but which merely puts one term in place of another (replaces foundations, for example, with the Aabsence of foundations@), while keeping the structure of the question in tact. Now, it stands to reason that any argument which depends on Asubstituting for God@ can be turned around B anything you can put in place of God must be metaphysically on a par with Him. So, taken in the context of the difficulties with the explanatory use of Ainterpretive community@ (‘7), what the argument suggests is not that a more down-to-earth explanation is possible after all, but that any entity which purports to function in such an explanation can seem to do so only insofar as we have managed to disguise from ourselves what we are really thinking of. What are we thinking of when we speak of an Ainterpretive community,@ of Ataking people into our community@ and the like? It is not the kind of community (a collection of agents) you can invite into the great hall for roll call, that=s clear B but is there another kind? Cutting through the irrelevant imaginative associations which can arise here (does Robinson Crusoe lack a Acommunity@?),36 we might simply give 35 By the same token, it is pointless to object that what the Catholic believes is really impossible. The response will be: AYou=re rightB it isn=t possible for one of us to be present in a space like that.@ See G. E. M. Anscombe, AOn Transubstantiation,@ in Ethics, Religion and Politics: Collected Philosophical Papers, Volume III (Blackwell, 1981), pp. 108-09. 36 In the critical literature, one can find two approaches to the use of the word Acommunity@ in Kripke=s claim that we can Athink of Crusoe as following rules@ only by Ataking him into our community@ (Saul Kripke, Wittgenstein on Rules and Private

210 MARTIN STONE

the answer which is determined by the structure of the problem and by the logical contribution which any term would have to be making to its solution. Structurally, what needs to be delivered into any picture of meaning which has inert Atexts in themselves@ (and therefore ubiquitous interpretation) among its elements is some further entity, however named, which possesses two related properties: first, it is fitted to be a regress-stopper, bringing to an end the reference from text to text;37 second, it is (thereby) able to infuse life B the power to mean B into those inert bits of textual matter. God fits this bill, for these are His traditional powers. Replacing Him with us (but keeping the question in tact) doesn=t give you pragmatism. It turns Ainterpretive community@ into something sublime: another name for Divinity, the terminus of all (Atext-referring@) explanations of meaning, their ground of possibility. As indicated, I would prefer the out-right God-mentioning option (among the present choices, including Platonism and its deconstruction) because this would put my audience on notice that I don=t know B or claim to know B what I=m talking about. But I think there is a better way to avoid misunderstanding here, and that is just to permit oneself to say, without philosophical scruples, what we say anyway, as agents, everyday: viz., that it is the rule itself which (sometimes) determines what one is to do. The Arule itself@ here just means Athe rule@ B as opposed to the rule with some further gloss on it. (Nothing occult: Lawyers draw applicative conclusions from Arules themselves@ everyday; they even print them and send them through the mail.) So such a commonplace should not be confused with any APlatonism@ about meaning. What is remembered in saying AThe rule itself determines how it is to be applied@ is merely what a Language (Harvard University Press, 1982), p. 110. One begins by assuming that we know what a Acommunity@ is B a collection of persons B and finds Kripke=s claim absurd; the other infers from the absurdity of the claim, so construed, that Acommunity@ is something mysterious here. Of course, these approaches aren=t exclusive. For an example of the first, see A.J. Ayer, Wittgenstein, (Random House, 1985), pp. 73-4; for the second, see Noam Chomsky, Knowledge of Language, (Praeger Paperback, 1986), p. 233. 37 See David Finkelstein, AOn Rules and Platonism@ in Alice Crary and Rupert Read, eds., The New Wittgenstein, op. cit., and my essay AWittgenstein on Deconstruction@ in the same volume.

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 211

rule is, for agents who have use for one. No explanation of how rule-informed judgments can be correct or true is offered here: There only seems to be question about this when words are considered in isolation from the applications agents have for them B considered (in this imaginary way) Ain themselves.@38 AThe rule itself determines....@ is what Wittgenstein calls a grammatical remark, one which tells Awhat kind of thing@ something is.39 No doubt, one could thank God that such a thing as a rule is possible, seeing how much good it does (in the law for example). But one=s attitude would no longer be that of someone thinking about some particular intellectual problem, but of someone being thankful B for rules or (it comes to the same) for being so much as able to thank.40 Perhaps we could thank ourselves too. A deeper sort of Apragmatism@ comes into view here, one having the sound of an impeccable naivete. Wittgenstein puts it like this: AWhat we do is to bring words back from their metaphysical to their everyday use.@ (PI §116) Terms like Aeveryday use@ are themselves notoriously subject to irrelevant associations. When Wittgenstein speaks of this (or speaks of Aobeying a rule@ as Aa practice@: PI §202), it must above all not be supposed that he is proposing a constitutive account of Ameaning@ in terms of some items B ordinary uses, practices, etc. B which have to be observationally located. His thought is rather that from the practical standpoint (that of someone deciding what a rule requires, or what to say or do more generally), the sort of knowledge which any such account could supply B something essentially speculative B is neither needed or wanted.41 A less misleading rendering of Aback to their everyday use@ B might thus be: Aback to the applications that living beings make of them.@42 Or B preserving Wittgenstein=s stress on the 38 AThe arrow points only in the application that a living being makes of it@ (PI §454; cf §432); for development of this, see Finkelstein, ibid. 39 PI §373. The remark shows Awhat goes to make up what we call >obeying a rule= in everyday life@ (PI §235). 40 See Martin Heidegger, What is Called Thinking (Harper Perennial, 1976). 41 This seems to be among the most difficult points in Wittgenstein to keep in focus; doing so would put both the notion of Aapplying his rule-following considerations to law@ and denying them application there out of the question. 42 PI §454; cf §432.

212 MARTIN STONE

first-person — Aback into our lives@: the Awe@ of Awhat we do@ (or Awhat we call@) being just that person=s use of language which is known of, by each of us, without having to observe or gather it. While these formulations might no doubt give rise to other misunderstandings, they will at least help to dispel that socio-theoretical one which has today all but rendered Aordinary language@ a useless term in philosophy.43 No one should be tempted to ask, for example, whose use? which lives? (Answer: that of which the speaker knows B knows its sayings and doings — without evidence, which-ever one that may be.) It is in this way that the words Arule itself@ can be brought back into an activity of using language: We could give it a job to do, as reasons require; Ainterpretation,@ Acommunity@ and Areasons@ too. Of course, so conceived, our everyday naivete isn=t naive at all, since it expresses what we get to through philosophical investigation, not merely the point at which we began. AThe place I really have to get to is a place I must already be at now,@ Wittgenstein remarked.44 It is the place he will have been, where he will recognize himself, when he isn=t beside himself. The return to that place of practical self-knowledge need not be any repetition, if this is to exclude a turn to something new.45 For it hardly needs saying that we often fail to know ourselves. Likewise, the place of return needn=t be anything Aordinary,@ if this means not just quotidian in structure but humdrum or familiar. Do we know what it would be to show up for the everyday feast without those reluctances and denials which sometimes find their chance in the scrupulous gaze from side-ways on?46 In the present instance, this would involve a different relation to a Aquestion@ asking 43 For an attempt to stem the tide, see Stanley Cavell, AMust We Mean What We Say@ in Must We Mean What We Say (Cambridge University Press, 2002). 44 Ludwig Wittgenstein, Culture and Value (University of Chicago Press, 1984), p. 7. The rest of the remark is very suggestive: AAnything that I might reach by climbing a ladder does not interest me.@ 45 The Aturn@ in Wittgenstein=s Areturn to the everyday@ is one of Stanley Cavell=s themes. See e.g., Cavell, In Quest of the Ordinary: Lines of Skepticism and Romanticism (University of Chicago Press, 1994); The New Yet Unapproachable America: Lectures after Emerson after Wittgenstein (Living Batch Books, 1994). 46 For a recent study of some of the problems which on-lookers of our kind get themselves into, see Richard Moran, Authority and Estrangement: An Essay on Self-Knowledge (Princeton University Press, 2001).

ON REFLECTIVE PRACTICES AND ‘SUBSTITUTING FOR GOD’ 213

for an Aaccount,@ and not just another account which leaves the structure of that question unchanged. 9. The judicious reader might ask B the question at the start (§2) B why Hershovitz would want Fish=s view to Asurvive in a more limited form@? To leave room for reflection and dispute of course; but some special motivation must have been required to read me as taking that away when my aim was only to get it into clearer view than is possible through the mind-numbing doctrine that everything is an interpretation. How is there a dispute here? Perhaps like this. Hershovitz=s point about some discursive regions being more doubt-ridden than others could be extended, I think, to various sub-regions, and from there to types of cases, and so on, down to the event of someone speaking about something, under the circumstances, somewhere or when; so that, in the end, there would be nothing of a philosophical (or a priori) nature to say about where doubts or reflections must come up. Some rough generalizations are all that could be meant by Adoubt-ridden discourse.@ Uptake of the Atext@ tends to be more reliable in communications from the flight tower to the pilot, we may observe, than in legal theory. But that isn=t to say that something couldn=t be perfectly clear in law, or that the tower=s landing instructions couldn=t be doubtful. Hershovitz might have a different view. His willingness to follow the line of discursive differences might stop at whole regions of discourse like ALaw,@ so that another interpretivism is afoot here. There are interpretive conditions of discourse B he might be thinking B only their jurisdiction is more limited than Fish thinks: some kinds of talk, not all kinds. This view is in fact suggested by a word which appears in his titleB practice. Much of the time, Hershovitz calls law a reflective activity. Would I be missing something if I agreed it was but put no special emphasis on law=s being a reflective practice as well? The answer for Ronald Dworkin, we know, is yes. The law is reflective in a rather special way, Dworkin wants to say, and the word Apractice@ helps capture this. It betokens an entity with some peculiar properties, not generally shared by Aactivities,@ or even Apractices@ as these might be mentioned in other contexts— a traveller=s report, for example. ALegal practice,@ in his special sense, seems to be only imperfectly presentB as something being practiced, an activity underway

214 MARTIN STONE

but never fully unfolded. And related to this, Aa practice@ is essentially fit to be understood through (what Dworkin calls) a Aconstructive interpretation@: a mode of understanding which exhibits the unity of its parts or instances on the basis of both evaluative and source-based criteria.47 My guess is that enthusiasm for a conception of law Aas a practice,@ in this sense, affected the shape of Hershovitz=s response. For it does seem to go with this conception that an interpretation B of that unfolding whole B is required in every case. Suppose this is right. Then I wonder: Does Hershovitz think that defending such a conception of law requires defending Fish? Or is the idea merely that it would be nice if some support could be had from him? I myself would not have thought there was any need, from the point of view of Dworkinian legal theory, for Fish to be right. What needs explaining is merely why, notwithstanding the absurdity of Fish=s general interpretivism, the law (and perhaps other qualifying Apractices@) are special.48 But maybe there is more to it. In allowing my critique of Fish to be partly right, Hershovitz is renouncing Fish=s own idea that legal interpretivism might flow merely from the conditions of possible meaningful concourse with a text. But once one gives up full-Fish, an explanation is needed of why limited-Fish should be right, for Fish himself doesn=t even purport to explain this. If there is something right in Fish=s view when it comes to Areflective@ activities like law, this would have to be, strangely enough, for reasons Fish has never contemplated. Why should interpretivism be specially fitted for Law? Perhaps Hershovitz= impossible suggestion B that Fish=s interpretivism is true but only for Areflective@ activities like law B is the result of a conflict: a felt need to support a practice conception of law out of the resources of general interpretivism, along with a recognition of the unavailability of such fishy resources. 49 See Dworkin, Law=s Empire (Belknap Press, 1988), ch. 2. Mark Greenberg=s paper (in this volume) seems to be in agreement on this point, since it attempts to distance its interpretivist view of law from any quite general (or Asemantic,@ as Greenberg calls it) interpretivism. In this sense, it is an endeavor to explain why the law is special. 49 These last issues obviously call for consideration on another occasion. It is only possible to note them here. 47 48

10

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

Metasemantics and Objectivity Ori Simchen* 1. If Jones is found guilty of negligence, and the verdict is never overturned, does it follow that the judgment that Jones was negligent is objectively true? If it does, then it is presumably a fact that Jones was negligent. What kind of fact is it? On the other hand, if it does not follow that the judgment is objectively true, what might still falsify it? Is there a ‘way things are’ legally speaking that goes beyond actual judicial decisions upheld by the courts? Such questions as these are most often raised regarding specific domains of judgment, such as the domain of legality or of morality or of science. In this paper I propose to examine the issue of objectivity more generally, with the hope of shedding some light on domain-specific concerns. Writers on objectivity typically set things up in the following oppositional way: Here we are with our X-type judgments. In order for such judgments to be true or false, there has to be something ‘over there’ by virtue of which they are so – call it the ‘X-facts’.1 Now the question arises as to the nature of these X-facts: Are X-facts really there to be discovered by us, or are they actually here in some sense, constituted by us? Are they judgment-independent or judgment-dependent? (I will not even pretend to do justice to the plethora of alternative ways of framing this type of query.) Once we have settled on the metaphysics of X-facts, on the approach being considered, we can then turn to the question of epistemic access to them. *

Department of Philosophy, University of British Columbia. The very idea of facts as truth-makers for judgments has been the topic of much heated controversy in contemporary philosophy. It has been called into question by Frege, Gödel, and Davidson, among others. For a recent round in the debate, see Stephen Neale, Facing Facts (Oxford: Oxford UP, 2001). 1

216 ORI SIMCHEN

By characterizing the relation between the thinker and the X-facts in such a way, we already open up a divide that any account of the bearing of the world on our judgments will be hard-pressed to close satisfactorily. Yes, X-facts are there to be discovered by us to the extent that we can get the facts wrong; but, No, they are not entirely independent of us to the extent that they attest to our conceptual involvement. And so the familiar back-and-forth refinements can continue without apparent end in sight. The history of philosophy is replete with illustrations of the moral that if the metaphysical and the epistemic are pried apart at the very outset of accounts of the objectivity of our judgments, then putting them back together again will prove to be a formidable task. In this paper I hope to make some headway towards resisting the common temptation to pry them apart at the very outset. The strategy I will employ is best characterized by fastening on an imagistic contrast. As against setting up a yawning gap between the thinker and the facts and then turning to ask the metaphysical question about the nature of those facts, followed by the epistemic question about the thinker’s access to them, I begin by focusing on those aspects of contact between the thinker and the world through which content emerges. My strategy for broaching objectivity will be metasemantic and will follow in the footsteps of the so-called new theory of reference. We begin by considering the thinker in her worldly surroundings. We then ask the metasemantic question: How do the thinker’s terms happen to gain the content that they do? Any plausible answer here will allude to what the worldly surroundings of the thinker actually are and to what the thinker’s overall epistemic situation actually is. But the metasemantic explanatory strategy does not begin by addressing the nature of X-facts independently of the thinker’s epistemic access to them and then proceed, as a separate project as it were, to address (or to set aside, as the case may be) the issue of epistemic access. Rather, the metasemantic strategy takes its point of departure from the basic idea that terms typically have content. This is sometimes referred to as their ‘intentionality’, or ‘aboutness’. And it is the possibility of such endowment that can be shown to require objectivity. Or so I will argue. The argument to follow has a distinctly transcendental, and so Kantian, flavor. But a rather more direct route runs back from it to the

METASEMANTICS AND OBJECTIVITY 217

work of the later Wittgenstein.2 I will try to show that the very possibility that terms in a given domain should have content, the very possibility that they should contribute to the truth-conditions of claims in which they partake, depends on there being a distinction between what is relevantly the case and what only seems to be the case. More specifically, content-determination will be claimed to depend on the existence of an objective measure of correctness in extension-determination. For example, that a term such as ‘negligence’ should have content demands that there be an objective measure of relevant similarity to paradigmatic instances of negligence, a measure of similarity that must be capable of transcending what merely seems to be relevantly similar to paradigmatic instances of negligence. If this is correct, then we still face a choice. We can either affirm the requisite measure of objectivity, or else we can deny the possibility that our terms have content after all. But such is an inevitable feature of transcendental arguments. Kant, for example, offers an argument purporting to show that if experience is at all possible, then the objects of experience must conform in certain elaborate ways to our cognition rather than the other way around. Assuming that this argument is successful, it is still open to us either to affirm Kant’s consequent or else to deny his antecedent and conclude that experience is not possible after all. We may thus think of Kant’s effort in this area as purporting to illustrate the heavy price we incur by denying his consequent, his so-called Copernican revolution. Similarly in this case, if the argument to the effect that the possibility of content depends in certain elaborate ways on the 2 As in passages peppered throughout Ludwig Wittgenstein, Philosophical Investigations, Third Edition (Oxford: Blackwell, 1953), such as the following: Let us imagine a table (something like a dictionary) that exists only in our imagination. A dictionary can be used to justify the translation of a word X by a word Y. But are we also to call it a justification if such a table is to be looked up only in the imagination? –‘Well, yes; then it is a subjective justification.’ – But justification consists in appealing to something independent. – ‘But surely I can appeal from one memory to another. For example, I don’t know if I have remembered the time of departure of a train right and to check it I call to mind how a page of the time-table looked. Isn’t it the same here?’ – No; for this process has got to produce a memory which is actually correct. If the mental image of the time-table could not itself be tested for correctness, how could it confirm the correctness of the first memory? (As if someone were to buy several copies of the morning paper to assure himself that what it said was true.) (§265).

218 ORI SIMCHEN

relevant objectivity is successful, then it is still open to us to deny objectivity. But the price of such denial is the denial of the possibility that our terms have content. And that is a heavy price indeed. 2. Our terms have content. How do they gain it? This basic metasemantic question has a prima facie intelligibility. In what follows I will consider the extent to which it can be respected in its own terms. There is a certain philosophical tradition that considers such a question to be misdirected, ill framed, or beg fundamental issues of philosophical methodology. This is a tradition that takes the facts of meaningfulness to derive from the interpretative situation. It is a tradition that takes the basic question in the area to be something along the following lines: What makes it the case that terms have the content that they do? Donald Davidson clearly falls within this camp, as does, albeit in a different way, David Lewis. Both take their most immediate inspiration from the work of W.V. Quine. On Davidson’s view, that expressions should have their content is a matter of being interpretable in this way according to a suitable (‘interpretative’) Tarskian truth-definition for the linguistic corpus to which they belong, given the attitudes that speakers are likely to have in their actual surroundings. On Lewis’s view, that expressions should have their content is a matter of the existence of an eligible mapping that so assigns contents to them, where eligibility is a matter constrained both by the attitudes of speakers, appropriately interpreted in turn, and by what the plurality of possible worlds is really like. For both thinkers, what makes it the case that terms have their content is at bottom a matter of how they are interpreted. Call this way of thinking about meaningfulness ‘metasemantic interpretivism’. If we situate speakers in their environment and raise the philosophical question of how it is that their terms come to have their content without privileging the interpretative situation, we part company with the above tradition.3 Call the alternative approach ‘metasemantic productivism’. Unlike the philosophical query after the determinants of the semantic 3 The qualification ‘philosophical’ before ‘question’ is important because neither Davidson nor Lewis would deny the existence of interesting empirical questions in the general area of content-determination. On the other hand, the qualification should not be taken as a tacit endorsement of a contentious analytic-synthetic distinction – as if philosophical questions can be sharply distinguished from empirical ones. These matters are far subtler than first impressions reveal and I cannot deal with them in a satisfactory way in the scope of this paper.

METASEMANTICS AND OBJECTIVITY 219

state of meaningfulness (‘How is it that terms have their content?’), the question raised by the rival approach (‘How is it that terms come to have their content?’) targets the determinants of a process – the process of gaining semantic content. To this alternative approach belong first and foremost the efforts of Keith Donnellan, Saul Kripke, and Hilary Putnam. Various current attempts to naturalize the mind a la Jerry Fodor or Fred Dretske also belong here, where what is sought is a naturalist reduction of the intentional to the non-intentional. But reductive naturalism is not the only option in metasemantics. We can acknowledge that no such reduction of the intentional to the non-intentional is available while making genuine metasemantic explanatory progress within the general framework of metasemantic productivism. To see how this might be so, it is helpful to consider an analogy. The Russell of ‘Knowledge by Acquaintance and Knowledge by Description’ asks how it is possible for anyone other than Bismarck to grasp a proposition about Bismarck.4 Most of us have never met the man, nor would we recognize him had we encountered him. Worse still, even if we had met him in the past or were perceiving him at the very moment of thinking or talking about him, while knowing full well that the man in front of us is Bismarck, we would not thereby be acquainted with the man himself but only with how he appears. In short, by Russellian lights we have no direct epistemic access to such items as Bismarck. Yet it is precisely such direct epistemic access to each constituent of a proposition that is required, according to Russell, to be in a position to grasp the proposition. So Russell’s answer to the aboutness question regarding Bismarck is roughly this: There is no possibility for anyone other than Bismarck of grasping a proposition that has Bismarck himself as a constituent. One can only grasp a proposition each element of which is an object of one’s acquaintance, and only one’s own sense data, one’s self, and various universals, qualify as objects of one’s acquaintance. However, we can grasp various propositions that describe Bismarck, propositions that Bismarck himself uniquely satisfies and that are composed of elements with which we are acquainted. It is only a proposition of this second type that anyone other than Bismarck can expresses with the words ‘Bismarck 4 Bertrand Russell, ‘Knowledge by Acquaintance and Knowledge by Description’, Proceedings of the Aristotelian Society 11(1910), 108-128.

220 ORI SIMCHEN

was an astute diplomat’. Aboutness regarding Bismarck for anyone other than Bismarck is, on this view, a species of satisfaction. Let us set aside the question of whether or not Russell’s theory is correct. Consider someone who objected to it on the following grounds: Rather than offer a genuine answer to how aboutness regarding Bismarck is possible, the theory merely pushes back the question of aboutness to aboutness regarding one’s own sense data, one’s self, and various universals – in short, to aboutness regarding items with which one is acquainted. But the theory offers us nothing at all when it comes to how we can entertain propositions about objects of acquaintance; how, for example, one can entertain a proposition that has oneself as a constituent. Given this glaring lack, so the complaint concludes, no explanatory progress has been made by Russell’s theory after all. But the complaint is misguided. Explanatory progress would have been made by Russell’s theory – if only it were otherwise plausible. To claim that Russell’s explanation of aboutness regarding the likes of Bismarck is unsuccessful because it has not succeeded in eliminating any trace of aboutness from the explanans is to set the bar of explanation far too high. If Russell’s theory were plausible, it would succeed in explaining how aboutness regarding all things reduces to aboutness regarding items of our acquaintance. That would have been a significant explanatory achievement. As it happens, the theory has little to recommend it on other grounds. But an account of aboutness need not culminate in a reduction of the intentional to the non-intentional in order to make genuine explanatory progress. 3. In order to begin to see how the possibility of endowment with content requires objectivity, we need to enlist the distinction between semantics and metasemantics and focus on the latter. 5 As it is commonly understood, semantics is concerned with specifying semantic contents and their modes of composition, whereas metasemantics is concerned with the general issue of content-determination. An easy illustration of the dis5 The distinction is discussed (under a slightly different terminology) in Joseph Almog, ‘Semantical Anthropology’, Midwest Studies in Philosophy 9(1984), 479-489; David Kaplan, ‘Afterthoughts’, in Joseph Almog, John Perry, and Howard Wettstein (eds.), Themes from Kaplan (Oxford: Oxford UP, 1989), 565-614, especially at 573-576; Robert Stalnaker, ‘On Considering a Possible World as Actual’, Proceedings of the Aristotelian Society Supplementary Volume 75(2001), 141-156; and Jules Coleman and Ori Simchen, ‘”Law”’, Legal Theory 9(2003), 1-41.

METASEMANTICS AND OBJECTIVITY 221

tinction is afforded by the semantics and metasemantics of proper names. Direct reference theorists claim against descriptivists that the contents of names are simply their bearers. So the content of ‘Bismarck’ in my mouth, say, is the man Bismarck. Such identification belongs to semantics. It specifies the content of the name as the entity named. But now the metasemantic question arises: How does the name ‘Bismarck’ in my mouth come to have the man Bismarck as its content? And here the prevalent metasemantic view that accompanies the semantic theory of direct reference is the causal-historical chain view. It holds that the name ‘Bismarck’ in my mouth comes to have Bismarck himself as its content by virtue of a causal-historical chain running back from my current employment of the name, via my own cognitive history, via the sources from whom I acquired the name, be they a history teacher or an author of a book I have read, and their cognitive histories in turn, and further back via sources of sources of sources all the way down to some initial act of naming Bismarck ‘Bismarck’. This causal-historical chain view is a metasemantic thesis, a thesis that must be distinguished from the semantic thesis that the content of the name ‘Bismarck’ is simply the man Bismarck.6 The new theory of reference, which is the general framework of metasemantic productivism I employ, initiated an externalist revolution in our thinking about the cognitive relations between the mind and the world, specifically the relation between the contents of terms and what they are about, namely, their extensions.7 Traditionally, the relation between contents and extensions was thought to be a species of satisfaction in the formal sense. On this view, the content of a term poses a mere condition that specifies what the term is about by way of satisfaction of the condition.8 Such contents were thought to be immediately 6 The most influential statement of the metasemantic thesis regarding proper names is found in Saul A. Kripke, Naming and Necessity (Cambridge, MA: Harvard UP, 1980), 90-97. 7 I aim to remain as neutral as possible on how exactly to think about what contents are. This semantic issue, while a crucial ingredient in any overall metasemantic story concerning content-determination, lies outside the scope of my immediate concerns. 8 By ‘mere condition’ I mean to rule out de re conditions, conditions that depend for what they are on the objects that satisfy them, such as the condition expressed by ‘identical with O’ where the name ‘O’ is understood to be contributing its bearer directly.

222 ORI SIMCHEN

accessible to the mind of the agent, whereas the portions of the world that terms are about were thought to be cognitively once removed. In this way, cognition was thought of as inevitably mediated by conditions entertained in the mind. It was a crucial feature of this outlook that contents do not depend for what they are on what the terms are about, or even on whether they are about anything at all, just as a mere condition can be the condition that it is whether or not anything satisfies it. On an externalist view of the relation between the mind and the world, aboutness is no longer thought to depend on satisfaction of conditions. The basic idea is that a term is about whatever it is about by virtue of being of it. To get an intuitive handle on what this ofness amounts to, it is useful to consider the aboutness of photographs.9 Consider the case of a photograph taken of one of two identical twins. Suppose further that the twins are so similar (or the photograph so imprecise) that had a photograph been taken of the other twin under suitable conditions, it would have been molecule-for-molecule identical to the actual photograph. Thus, as a mere visual condition, the photograph does not discriminate between the two twins. Yet for all that, it is only about one of them. We simply do not think of the aboutness of photographs as a matter of satisfaction of visual conditions. Rather, we think of this aboutness as having to do with the photograph’s ofness. The photograph is about the twin it happens to be of. It is about whichever of the two twins was the relevant causal-historical antecedent to the photograph’s formation as consequent. One thing suggested by such cases is that we do not in general individuate photographs in abstraction from what they are about. A photograph of the other twin might have been molecule-for-molecule identical to the actual photograph, yet it would still be a different photograph by virtue of being of – and thus about – someone else. Semantic externalism considers the ofness of expressions to be essential to their aboutness. Whatever their content is ultimately held to be, the aboutness of terms is achieved via their ofness. This means, among other things, that we have a reversal of the traditional view of the relation between content and extension. Traditionally, contents were thought to specify extensions as mere conditions that do not de9 Such heuristic appeal to photography is inspired by a similar appeal made in David Kaplan, ‘Quantifying In’, Synthese 19(1968), 178-214.

METASEMANTICS AND OBJECTIVITY 223

pend for what they are on what the terms are about. But on the externalist outlook, content crucially depends for what it is on what the term is about – so much so, in fact, that many versions of semantic externalism simply identify contents with extensions. To summarize the contrast in a word we might say that whereas the traditional view thought of content as determining extension, the new orthodoxy thinks of extension as determining content. With these cursory remarks on the general framework in hand, we can now turn our attention to the metasemantics of general terms such as common nouns and adjectives. How are we to think of content-determination for a typical general term? To the extent that we think that those contents depend for what they are on what the terms are about, i.e. on their extensions, extension-determination is going to play a crucial role in the overall account of content-determination. Take the familiar example of ‘water’. To the extent that the content of ‘water’ depends for what it is on its extension, an account of how ‘water’ comes to apply to all and only samples of water will occupy a central role in any plausible story about the way in which ‘water’ gains its content. For the remainder of this paper, my earlier antireductionist remarks must be borne in mind. Specifically, the explanatory burden of a metasemantic account of general terms should not be thought of as the reduction of the intentional to the non-intentional. 10 Elsewhere Jules Coleman and I have defended an account of extension-determination that takes its main cue from Putnam’s work in ‘The Meaning of “Meaning”’.11 As against the traditional view that knowing the content of a typical general term is a matter of knowing an extension-fixing criterion that all and only samples of the relevant kind satisfy, Putnam (and, independently, Kripke) has argued convincingly that 10 One reason for being pessimistic about the prospects of a naturalistic reduction of aboutness is that for many terms content-determination proceeds by way of linguistic deference to a relevant expertise, as we shall see. Such deference implicates an elaborate authority structure, and there is good reason for being pessimistic about the explanatory prospects of attempting to account for such social phenomena naturalistically, within the vocabulary of cognitive science, say. Such pessimism carries over to a general pessimism about a cognitive-scientific reduction of aboutness. This is an area of heated controversy that obviously demands far more attention than I can devote to it here. 11 See Jules Coleman and Ori Simchen, ‘”Law”’. Putnam’s classic paper is collected in Hilary Putnam, Mind, Language and Reality: Philosophical Papers, Volume 2 (Cambridge: Cambridge UP, 1975), 215-27.

224 ORI SIMCHEN

there is little reason to think that proficient speakers are in possession of any such criteria. For example, adult speakers of English who are proficient with the noun ‘gold’ seldom know a general criterion that applies to all and only instances of gold. But for all that, ‘gold’ applies to all and only instances of gold. How is this determination achieved? It is achieved in two ways: socially and environmentally. Socially: Let us grant that proficiency with ‘gold’ does not entail possession of an extension-fixing criterion. Yet I am a proficient speaker who would be quite easily taken in by samples of iron pyrite (‘fool’s gold’). Does it not follow that ‘gold’ in my mouth picks out anything that I would be inclined to regard as gold, including samples of fool’s gold? – Not at all. ‘Gold’ in my mouth still applies to all and only samples of gold because what ‘gold’ in my mouth applies to is not just a matter of how things are with me considered in isolation from the rest of my linguistic community. Rather, it is an intricate matter of social exchange that Putnam dubbed ‘division of linguistic labor’ and Gareth Evans likened to the relation between producers and consumers. I will follow received practice and refer to this phenomenon as ‘linguistic deference’. The basic idea is that ordinary speakers (‘novices’) successfully refer to gold and not to fool’s gold by employing ‘gold’ via their tacit reliance on a relevant expertise, in this case metallurgy. By placing their trust in an expert doctrine, speakers can employ general terms to refer determinately to things regarding which they are relatively ignorant. Reference cannot be such a difficult cognitive task so as to require each and every member of the linguistic community to become an expert on what is being talked about. Environmentally: Had speakers’ surroundings been relevantly different, say with some distinct yet superficially indiscernible water-like substance occupying the role of water, or with cat-like demons occupying the role of cats, the extensions of ‘water’ and ‘cat’, and so their contents, would have been different from what they actually are. Speakers employ such terms to speak about whatever in the world is around them. We employ ‘cat’ to refer to cats and not to cat-like demons. Our counterparts in a cat-like demon world use ‘cat’ to refer to cat-like demons and not to cats. The difference is in what is around. Moreover, if we bear in mind that our cats are not demon-cats (assuming such things are possible) and if we consider our own intuitions regarding whether or not ‘cat’ as spoken by us applies also to demon-

METASEMANTICS AND OBJECTIVITY 225

cats, the answer is a resounding No, even under the further assumption that we would never be able to tell them apart from cats. ‘Cat’ as spoken by us applies to cats and to nothing else. Similar intuitions can be elicited for other general terms. This suggests that there is an indexical element in extension-determination for a typical general term. To fall under the extension of ‘cat’ is to bear some relation – a relation that demon-cats, for example, do not bear – to paradigmatic cats in the environment that are referred to indexically, say by employing ‘this[the furry meowing thing over here]’. In other words, the actual interaction between speakers and their environment determines what ‘cat’ applies to at every possible world.12 As Frank Jackson puts it regarding the example of ‘water’: ‘The reference in all worlds is settled by what is watery and the subject of the relevant acquaintance in the actual world’ (39).13 Putting the above two points together yields the following schema of an account of extension-determination for general terms. The extension of a typical general term ‘N’ is specified by the following condition: (†)

N(x) « = (x,this)14

‘N’ applies to all and only those items, at any possible world, that bear a relevant similarity relation =to whatever is indexically referred to by ‘this’ as spoken in the actual world. Take ‘gold’. It applies to all 12 (For those interested in semantic scruples:) When considering whether ‘cat’ applies to occupants of other possible worlds that are relevantly similar to actual cats picked out indexically, or whether instead it applies to occupants of other possible worlds that are relevantly similar to items picked out indexically in those other worlds, the first option seems to be supported by, and the second option to conflict with, a basic semantic fact about indexicals, namely, that indexicals take large scope relative to intensional operators. Thus, (i) is consonant with the logic of indexicals as it is commonly construed while (ii) is not (we let the square brackets indicate scope and ‘˜’ stand for the relevant similarity relation): (i) [this]?(“x)(cat(x) « = (x,this)) (ii) ?(”x)(cat(x) « [this] = (x,this)). If this is correct, then the rigidity of ‘cat’ depends on a feature of its metasemantics, a feature that depends, in turn, on a feature of the semantics of indexicals. 13 Frank Jackson, From Metaphysics to Ethics (Oxford: Clarendon Press, 1998). 14 Nothing requires that ‘this’ be the particular indexical expression involved in extension-determination for a typical general term. The choice of a specific indexical is only for heuristic purposes, as should become clear from the discussion to follow of the role of (†) within the overall metasemantic account.

226 ORI SIMCHEN

and only samples of a substance, at any possible world, that bear a certain similarity relation – in this case microstructural similarity – to paradigmatic samples of the substance indexically referred to by ‘this’ as spoken in the actual world. In short, ‘gold’ refers to whatever is microstructurally close enough to paradigmatic samples of gold in speakers’ actual environment. And discerning that the relevant similarity relation (call it ‘=M’, for ‘metal’) indeed obtains is something that is left to metallurgical expertise to decide. In other words, the social aspect of extension-determination for ‘gold’ enters primarily15 in discerning that =M obtains between putative instances of gold and paradigmatic samples, whereas the environmental aspect enters in the employment of the indexical ‘this’ in the actual presence of paradigm instances of gold. 4. Before moving on to consider further wrinkles, I need to say more about the role of (†) itself within the overall account of content-determination on offer. First and foremost, (†) is not meant itself to capture the content of ‘N’. It is offered as an extension-fixer that speakers are tacitly committed to, given their actual referential intentions. (†) figures in a metasemantic, rather than a semantic, account of extension-determination. The idea is that speakers employ a typical general term ‘N’ as if they are committed to (†) as an extension-fixing stipulation. And the force of this ‘as if’ claim is just that speakers employ ‘N’ with the referential intention to pick out whatever is in fact relevantly similar to paradigmatic instances of the kind in their environment. Now, (†), all by itself, is completely schematic. Precisely how we are to think of the implicated referential intentions in concrete instances is a subtle matter, but this much is relatively clear. Referential intention attribution is a species of intention attribution more generally, which is, in turn, a species of attitude attribution. It is a commonplace in attitude attribution that we attribute to agents only those attitudes they can be expected to have given their overall epistemic situation. What is rather surprising and seldom noticed is that this constraint, mild as it may seem, actually renders certain common 15 It is sometimes suggested that linguistic deference enters not only in discerning relevant similarity to paradigmatic instances of the kind, but also in identifying the paradigmatic instances themselves. This seems correct for some deferential terms and incorrect for others. The general issue need not concern us here.

METASEMANTICS AND OBJECTIVITY 227

philosophical attributions of referential intentions highly implausible. One such implausible attribution is the attribution to ordinary speakers of the intention to employ ‘water’ to refer to anything relevantly similar to paradigmatic instances of water from the standpoint of the world as it is in itself, beyond whatever we might come to believe about the matter. Another implausible attribution, from the other end of the philosophical spectrum, is the attribution to ordinary speakers of the radical subjectivist intention to employ ‘water’ to refer to whatever the agent herself would regard as relevantly similar to paradigmatic instances of water in a way that is not susceptible to any external check on the matter. (I will come back to radical subjectivist referential intentions below.) Neither attribution is plausible in light of speakers’ other attitudes. Referential intention attribution is an exercise in making sense of speakers’ attitudes within an overall metasemantic story about how a given term comes to possess the content that it has. More specifically, it consists in squaring referential intentions with speakers’ other intentions, beliefs, desires, hopes, fears, etc. It should thus aspire to remain as true as possible to speakers’ actual attitudes and refrain as much as possible from subjugating them to extrinsic philosophical agendas. A related point concerning the role of (†) is that (†) is neutral as to whether ‘N’ is linguistically deferential or not. Certain terms, such as the natural kind term ‘gold’, are linguistically deferential if any term is. As mentioned above, individual proficiency with ‘gold’ does not require of speakers to be capable of discerning that a given sample of substance is relevantly similar to paradigmatic instances of gold. Rather, speakers are best understood to be implicitly deferential to metallurgy to decide on such matters. But other terms do not exhibit linguistic deference – the non-natural kind term ‘chair’ is one salient example. It is no part of our linguistic practices vis-à-vis ‘chair’ that we are deferential to some chair-expertise to discern relevant similarity to paradigmatic chairs. In ‘”Law”’ Jules Coleman and I argue that despite initial appearances to the contrary, the deferential-non-deferential distinction cuts across the natural-non-natural distinction. It is just not the case that the natural kind terms are the deferential ones whereas the non-natural kind terms are the non-deferential ones. Some natural kind terms, such as ‘puddle’, are not deferential, whereas

228 ORI SIMCHEN

some non-natural kind terms, such as ‘carburetor’, are.16 This makes the question of what determines whether or not a given term is linguistically deferential more demanding than is often presumed. Much of the discussion in ‘”Law”’ is devoted to answering this question and drawing implications from the answer for the recent contention in the philosophy of law that we are as linguistically deferential to jurisprudential expertise with respect to ‘law’ as we are to metallurgy with respect to ‘gold’.17 To summarize this all-too-brief metasemantic sketch: Extension-determination for a typical general term ‘N’, which plays a crucial role in its content-determination, is achieved via referential intentions to pick out whatever is relevantly similar to (i.e. bears ˜ to) paradigmatic instances of N in speakers’ environment. (The role of ‘this’ in (†) is simply to make the environmental aspect of extension-determination vivid.) If ‘N’ is linguistically deferential, speakers leave it up to some relevant expertise to discern whether or not ˜ obtains. If ‘N’ is not linguistically deferential, speakers are relatively self-reliant in this regard. Either way, ‘N’ gains its extension, and consequently its content, via referential intentions that specify that it is to apply to anything bearing =to paradigmatic instances of N in speakers’ environment. 5. We are now finally in a position to explore the bearing of this metasemantic story on the question of objectivity. If the above sketch is on the right tracks, then content-determination depends on extension-determination. And extension-determination depends, I now claim, on an objective measure of similarity ˜ to paradigmatic instances of the relevant kind. In other words, it is built into the possibility that the general term ‘N’ should have whatever content it happens to have that there be an objective measure of similarity to instances of the kind. Extension-determination for a typical general term depends on the existence of some independent standard that can facilitate a genuine difference between cases where instances only 16 I use ‘natural kind term’ in a way that purports to remain neutral with respect to further metaphysical commitments regarding natural kinds. Puddles are not non-natural; hence, they are natural. I realize that there is a lot more that can be said here concerning what in general counts as natural, but it is not required for present purposes. 17 For a defense of this contention see Nicos Stavropoulos, Objectivity in Law (Oxford: Clarendon Press, 1996).

METASEMANTICS AND OBJECTIVITY 229

seem to be relevantly similar to one another and cases where this is in fact the case. To see this, we turn to consider some examples. In the metasemantic literature it is often presumed that in the case of substance terms such as ‘water’ or ‘gold’, such a standard is provided by the microstructure of the substance. Take ‘gold’ again. It is thought to apply to all and only samples that are microstructurally close enough to paradigmatic samples of gold. But this is a matter that can and does easily transcend mere seeming similarity to gold. Ordinary proficient speakers are not privy to the procedures whereby experts distinguish samples of genuine gold from samples that only seem to the unaided mind to be samples of gold. It is precisely here that linguistic deference enters the picture. For it is built into our linguistic practices vis-à-vis these substance terms that in determining whether or not such a term applies in a given case, an ordinary speaker is beholden to a relevant expertise on the matter. This means that distinguishing similarity to paradigm instances in such cases from mere seeming similarity is a matter that novices leave to experts to decide. The case of simple artifact terms such as ‘chair’ or ‘hammer’ is more difficult, but here, too, there are means for distinguishing similarity to paradigm instances of the kind from mere seeming similarity. As mentioned above, a term such as ‘chair’ is not linguistically deferential, so such means are not provided by an expert doctrine. In addition, while as a linguistic community we are obviously successful in classifying things under the label ‘chair’, how we determine relevant similarity to paradigmatic chairs is not entirely transparent to us. However, there are compelling empirical reasons for thinking that relevant similarity in such cases is heavily informed by the intended function of instances of the kind.18 Let us assume that this is correct: A proficient speaker employs ‘chair’ with the referential intention to pick out anything that is relevantly similar to paradigm instances, where relevant similarity has much to do with being intended to serve the same function as paradigmatic chairs. But whether or not a given item is intended to serve the same function as a paradigmatic chair is a 18 The psychological literature on conceptual development abounds with attempts to identify features of artifacts that are generally considered to be essential to them. See, for example, F. C. Keil, Concepts, Kinds, and Cognitive Development (Cambridge, MA: MIT Press, 1989).

230 ORI SIMCHEN

matter that can easily transcend what merely seems to be the case. If it is indeed true, as findings on conceptual development suggest, that relevant similarity to chairs is a matter that is heavily informed by intended function, then in the case of an item that despite appearances to the contrary has no intended function, speakers would stand corrected if they initially classified it as a chair and were then informed that in fact the item has no intended function. Be that as it may, whether we are dealing with linguistically deferential kind terms or with ones that are not, without facilitating a distinction between genuine similarity to instances of the kind and mere seeming similarity, no extension could be secured for the kind term in question, and, consequently, no content. For without any means of effecting a seems-is distinction in extension-determination, a term would apply to anything seeming to be relevantly similar to (what seem to be) paradigmatic instances of the kind. This means that no possibility of erroneous application of the term would be facilitated. But without the possibility of genuine error in application, there is no possibility of genuine correctness in application either. In this way, the term would not contribute to the truth-conditions of claims in which it partakes; that is to say, it would lack content altogether. This line of reasoning bears repetition in a more orderly and schematic fashion: Let ‘N’ be a putative kind term for some putative kind N. Suppose that (i) there is no objective measure of similarity among instances of N. Then, (ii) there is no objective measure for membership in N. Thus, (iii) there is no possibility of genuine error in the application of ‘N’. But then, (iv) there is no possibility of genuine correctness in the application of ‘N’ either. Therefore, (v) ‘N’ does not contribute to the truth-conditions of claims in which it partakes. And therefore, (vi) ‘N’ lacks content. A standard subjectivist response to this argument is to claim that nothing in it effectively rules out the possibility of endowment with subjective content. In other words, so the objection presses on, nothing that has been said so far rules out that a general term may have such content that is shaped by the radical subjectivist referential intention alluded to above, the intention to pick out anything that merely seems to be relevantly similar to what merely seem to be paradigm instances of the kind. All that is needed for endowment with subjective

METASEMANTICS AND OBJECTIVITY 231

content is that the relevant similarity relation itself be subjective. For ‘N’ to be endowed with subjective content, its extension need only be fixed by the subjective inclination to regard things as relevantly similar to whatever one is subjectively inclined to regard as paradigmatic instances of seeming-N. As long as this remains a standing possibility, it is just wrong to claim, as I have, that endowment with content requires an objective measure of similarity in extension-determination. Let us examine what would transpire if we withdrew the requirement that there be an objective measure of similarity in extension-determination, in the way suggested by the objection. Let us assume for the sake of argument that the attribution of the radical subjectivist referential intention is in fact adequate for some substance term that applies to all and only instances of seeming water-like. Let this term be ‘water*’. We are supposing, then, that ‘water*’ applies to whatever is deemed by the speaker to be relevantly similar to (i.e. whatever bears =* to) other instances of water*, where being an instance of water* is constituted by merely seeming water-like. Now, in order for any general term ‘N’ to gain a determinate extension, and so to contribute to the truth-conditions of claims in which it partakes, there has to be some way of effecting a seems-is distinction that would allow us to say that something can only seem to be relevantly similar to instances of N without actually being relevantly similar to those instances. We just saw how a seems-is distinction is provided for in the examples considered above of substance terms and simple artifact terms. In the case of a term such as ‘water*’, what would be required is a facility to distinguish cases of merely seeming to bear =* to instances of water* from cases of genuinely bearing =* to them. But here comes the crucial point: If something seems to bear =* to instances of water*, then ipso facto it bears =* to them! For to bear =* to something is to seem relevantly similar to it. But to seem to bear =* to something is to seem to seem relevantly similar to it. But seeming to seem relevantly similar to something is just to seem relevantly similar to it all over again. Seeming does not genuinely iterate. Consider any case of seeming to f. If something seems that it seems to f, then it also thereby seems to f. If something seems that it seems red, then it also seems red; if something seems that it seems sweet, then it also seems sweet; if something seems that it seems painful, then

232 ORI SIMCHEN

it also seems painful. In other words, there is no place to insert the requisite seems-is wedge when it comes to seeming-to-f. If I believe that something is red then I may be mistaken, for it may only seem to me that it is red while being some other color. But if something only seems red to me – where ‘seems red’ is not just a stylistic variation on ‘is red’ – I cannot be mistaken about that. What this means, in effect, is that extension-determination for ‘water*’, and so content-determination, cannot take place after all. So the metasemantic question regarding ‘water*’ remains unanswered. If this is correct, then despite initial appearances to the contrary, a term such as ‘water*’ cannot gain content. The possibility of endowment with content requires objectivity in extension-determination. 19 6. I began this paper by briefly considering the oppositional way that writers on objectivity typically set things up, using their strategy as a point of contrast to my alternative metasemantic strategy. Rather than focus on judgments in a particular domain and raise the question of the metaphysical status of the facts that constitute their subject matter, followed perhaps by the question of epistemic access to those facts, I chose to focus on the world-thinker interplay that undergirds endowment with semantic content. Given the framework of metasemantic productivism sketched above, it turned out that the very possibility of such endowment requires that there be some independent measure to facilitate the difference between genuine similarity to instances of a given kind and mere seeming similarity. In this way, objectivity is required for extension-determination, and so required for 19 The transition from the claim that if ‘water*’ seems to apply to something then it does in fact apply to it, to the claim that ‘water*’ has no content, might give rise to the following worry. Suppose that ‘water*’ applies by seeming to apply. Could I not still misapply it, say by intending to misapply it? But in that case, it would seem that a genuine contrast between application and misapplication for ‘water*’ can be facilitated, in which case ‘water*’ can gain a determinate extension, and so a determinate content, after all. However, further reflection will reveal this to be gratuitous. Under the conditions specified above, what might it mean to say that we can misapply ‘water*’? Suppose I resolve to misapply it in a given instance. In what (or against what) might my misapplication of it consist? The only available answer is that ‘water*’ seems to misapply in this given instance. In other words, the term in question applies by seeming to apply and misapplies by seeming to misapply. And this can only mean that there is no talk of genuine application or misapplication here. Thanks to Mark Greenberg for drawing my attention to this worry.

METASEMANTICS AND OBJECTIVITY 233

content-determination more generally. But curious minds still want to know: What kind of objectivity do legal facts, let us say, enjoy, as opposed to the objectivity of the facts of natural science, or of moral facts, or of mathematical facts? Is the objectivity of legal facts not ‘softer’ in some sense than the objectivity of natural-scientific facts? Yet for all that such questions may strike us as gripping and unavoidable, it is far from obvious that anything useful can be said about objectivity as a feature of facts considered as truth-makers for our judgments. We can, however, turn our attention to objectivity as presupposed by the possibility of endowment with content of specific terms and perhaps learn thereby something important about their employment. By way of conclusion, let us turn again to ‘negligence’. If the overall argument of this paper is sound, and to the extent that the metasemantic framework sketched above is independently plausible, then in order for ‘negligence’ to gain its content there must be an independent measure to distinguish genuine similarity to paradigmatic instances of negligence from mere seeming similarity. The presumption that ‘negligence’ is endowed with content, that it contributes in relevant ways to the truth-conditions of claims in which it partakes, entails, among other things, that ‘negligence’ can apply wrongly. As long as speakers – and this includes judges – consider the term ‘negligence’ in their mouth to have content, as long as they consider it as contributing in relevant ways to the truth-conditions of claims in which it partakes, they must also consider it to be susceptible to misapplication in specific cases. If the term is not susceptible to misapplication, it cannot have content. And it is, in fact, overwhelmingly likely that speakers, including judges, consider such terms to be susceptible to misapplication. In this, ‘negligence’ is not so different from ‘water’. As long as speakers – and that includes chemical experts – consider the term ‘water’ in their mouth to have content, as long as they consider it as contributing in relevant ways to the truth-conditions of claims in which it partakes, they must also consider it to be susceptible to misapplication in specific cases. If the term is not thus susceptible, then it cannot have content after all. Where does this all leave us? One thing that can be said about objectivity is that a measure of independence from what we happen to

234 ORI SIMCHEN

deem relevantly similar to paradigmatic instances of kinds is presupposed by the very possibility that our kind terms are endowed with content. And this includes what we deem relevantly similar to paradigmatic instances of kinds even at some hypothetical end of inquiry, or under ideal epistemic conditions. The epistemic conditions under which the verdict that Jones was negligent was reached may have been ideal. Yet for all that, in order for the term ‘negligence’ in the court’s mouth to have content, that is, to contribute to the truth-conditions of claims in which it partakes, the court must be regarded as susceptible to error in application. This means that despite all the epistemic ideality in the world that may happen to obtain as regarding the court’s employment of ‘negligence’, it is compulsory to treat it as answerable to some independent standard that can facilitate the distinction between being relevantly similar to paradigmatic instances of negligence and merely seeming to be so. Does this conclusion saddle us with some version of untenable Platonism? – Hardly. For the metasemantic strategy for broaching objectivity does not begin with the customary examination of the metaphysical underpinnings of the relevant facts, followed by a treatment of epistemic access to them. The above argument, if indeed successful, illustrates that objectivity is presupposed by the possibility of endowment with content. To my mind, given the attractiveness of the metasemantic picture offered above, such a conclusion is on much firmer ground than are customary defenses of objectivity, defenses that tend to raise far more perplexities than they actually succeed in quelling.

11

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

Can Objectivity be Grounded in Semantics? Michael S. Moore* I. Introduction There are two topics examined in Professor Simchen’s paper. The first and metaphysical topic is the sense in which our judgments in science, ethics, and law might be objective and how one should argue for objectivity in that sense. The second topic, this one in the philosophy of language, is about the central question for that discipline: what is the meaning of the terms used in a natural language such as English, and how did such terms acquire such meanings? It seems to be a central organizing principle of the paper that answering the questions in the philosophy of language will help in answering the questions in metaphysics. At the close of these comments I shall return to discuss the connection between the two topics. Before doing that it would be well to discuss the topics separately, which I plan to do, starting with the metaphysical question of objectivity. II. Simchen’s Transcendental Deduction of Objectivity Simchen puts aside what he calls the “oppositional approach” to metaphysical questions.1 On this approach – of which I take myself to be * Charles R. Walgreen Jr. University Chair, Professor of Law, Professor of Philosophy, and Co-Director of the Program in Law and Philosophy, University of Illinois. 1 Ori Simchen, “Metasemantics and Objectivity,” this volume, pp. 215-234.

236 MICHAEL S. MOORE

an exemplar2 – one isolates a class of judgments, seeks the facts that could make such judgments true, asks whether such facts exist and if so, whether their existence is independent of human belief and convention. The existential and mind-independence questions classify one as a realist, a skeptic, or an idealist of some kind about such facts. One separately addresses the epistemological question of how one discovers and grasps facts of this nature and how one justifies belief in them. Simchen’s motives for putting aside this traditional approach to metaphysics appear to be three fold. First, he doubts “that anything useful can be said about objectivity as a feature of facts considered as truth-makers for our judgments.”3 Call this the intelligibility worry. Second, the history of metaphysics carried on in this way is not reassuring to Simchen: “the familiar back-and-forth refinements can continue without apparent end in sight.”4 Call this the non-resolvability worry. Third and most important to Simchen, there is what I shall call “the gap worry”: if we separate questions about the nature of certain facts from questions of how we discover, grasp, and justify our beliefs about such facts (as is done on the “oppositional” approach), then we will have opened up a gap (between what there is and what we can discover/understand/justifiably believe) that we cannot bridge.5 Because of such a gap, we will always be open to the skeptic’s taunt that the evidence we possess is insufficient to justify belief in the things that evidence supposedly evidences. Simchen’s approach, as he himself characterizes it, “has a distinctly transcendental, and so Kantian, flavor.”6 Let us first be clear what that flavor is before we probe Simchen’s variety of it. On my under2 See my “Moral Reality,” Wisconsin Law Review, Vol. [1982], pp. 1061-1156; “Moral Reality Revisited,” Michigan Law Review, Vol. 90 (1992), pp. 2424-2533; “Legal Reality: A Naturalist Approach to Legal Ontology,” Law and Philosophy, Vol. 21 (2003), pp. 619-705. These and like-minded essays are collected in Moore, Objectivity in Ethics and Law: Essays in Moral and Legal Ontology (U.K.: Ashgate Press, 2004). 3 Simchen, “Metasemantics,” p. 237. 4 Id., p. 216. 5 Id., pp. 215-234. 6 Id., p. 216.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 237

standing, a transcendental deduction has three steps to it.7 First, one does a deep analysis of some body of alleged fact, seeking its presupposed deep principles. We ask for the conditions of the possibility of those facts being as we believe them to be. At this step we do not ask whether such facts or their presuppositions are true; only if the one indeed presupposes the other. Second, we count the cost of denying such presuppositions, i.e., can we countenance or even imagine a world in which such presuppositions are not true? Third, if we find the cost of abandoning such presuppositions to be too great, we commit to the facts from which they are taken. That is, we affirm such facts to be true. Simchen’s use of this broadly Kantian schema is as follows (using legal judgments as our exemplars). First, Simchen observes that we experience law as a set of statements using terms like “negligence” in a way that endows them with semantic content; such terms, in other words, apply by virtue of their meaning to some things and not to others. In this endowment of our terms with content we presuppose the objectivity of such terms in the sense that we presuppose a distinction between what such terms actually apply to and what they only seem to apply to. Second, we can’t imagine experiencing law in any way that did not endow legal terms with content, that did not make them applicable to the world. Third, we should therefore commit to the objectivity of law in the sense presupposed by our practices, viz, that there is an objective truth about whether such terms apply or not to a given state of affairs, a truth not captured by the subjectively experienced applicability of such terms by certain legal actors. We should commit to judgments of the form, “x is negligent,” as capable of being objectively true in this sense because not to do so would rob “negligence” (and like terms of law) of their content. 7 See generally Lewis White Beck, A Commentary on Kant’s Critique of Practical Reason (Chicago: University of Chicago Press, 1960), p. 170: “The process of transcendental deduction is not that of linear inference from a premise to its logical consequence. It is a process of taking some body of alleged fact (e.g., mathematics or science) which has been challenged and showing (a) what its necessary presuppositions are and (b) what the consequences of denying these presuppositions are.” I would add: “(c) one affirms either the facts in question or their presuppositions because the costs of denying one or the other is seen to be too great.”

238 MICHAEL S. MOORE

Putting it this explicitly reveals my first query about Simchen’s argument. Put simply, the analysis in the first step of the deduction doesn’t seem very deep: it ends pretty much where it begins. That is, the presupposition of our experience of law (that at the end of the deduction we are to affirm) is pretty much right on the surface of the experience itself. The idea that legal terms have content – that they apply to the world – seems very close to the idea that we must distinguish actual from merely seeming applicability of such terms. This seems close to saying that we experience law as objective in its character, and (in that trivial sense) our experience of law “presupposes” its objectivity. In which case, wouldn’t it be more accurate for Simchen to argue: we experience law as objective in character, and we can’t imagine that it isn’t? To the extent Simchen intends a truly Kantian form of argument here, the closeness of what is presupposed to that from which it is presupposed robs the argument of its force. A transcendental deduction is only persuasive when there exists some differential plausibility between some facts that we experience, on the one hand, and their presuppositions, on the other. It is this difference in plausibility that allows the deduction to endow the one with the greater plausibility of the other. When the presupposition and the facts experienced are one and the same, there can be no such enhancement of plausibility. Supposing that Simchen’s argument can be taken to be an instance of a transcendental deduction, my second query is whether such deductions can ever succeed in establishing what they seek to establish. My general take on such modes of argument is that they either prove too little or they prove too much. Take the “too little” horn of the dilemma first. Such deductions threaten to collapse into an analytic psychology, of the kind Kant in his ethics dubbed a “metaphysical deduction.”8 They so collapse when the costs of giving up some concept is bearable, and so the conclusion sought to be established – the presupposition of the concept – is hardly one we must affirm. 8 In Kant’s ethics a “metaphysical deduction” seeks a kind of pure, a priori knowledge of concepts (such as the concept of duty), without showing us that our concepts give us knowledge of the objects to which such concepts purport to refer. As Beck puts it, “The metaphysical deduction is Kant’s effort to discover what the categories are; the transcendental deduction is his effort to show that they are valid.” Commentary, p. 110.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 239

As an illustration, consider Ronald Dworkin’s early foray into this kind of analytic psychology. In his famous hard cases argument,9 Dworkin noted that we conceive of law in such a way that litigants never approach judges as supplicants of a favorable exercise of judicial discretion; they always appear as claimants of legal rights even in the hardest of hard cases. A condition for the possibility of there being legal rights in hard cases is that there is law governing such cases; if there is no obvious law governing the case (which there is not because it is a hard case), then there must be a “seamless web” of unobvious law, consisting of moral principles having some institutional support in other parts of the obvious law. There is nothing in this argument as stated10 that compel us to affirm either our concept of law or its presupposition. For we can easily imagine revising our concept slightly; the rhetoric of lawyers arguing hard cases is seen as just that, rhetoric, not to be taken seriously as requiring beliefs by anyone that there really are legal rights in hard cases. In which case the cost of not affirming the presupposition of these practices – a kind of natural law view of the law governing hard cases – is one we might easily pay. That would leave Dworkin’s “deduction” as only a bit of uncommitted, third person, psychological analysis: we conceive of law and law practice in ways presupposing a seamless web view if it. I take it that Simchen does not think we could imagine that legal terms like “negligence” are not endowed with content in the way that we could imagine that there were no legal rights in hard cases. So let us swing to the other horn of the dilemma. Suppose we find it literally inconceivable that we did not experience sequences of events (for example) in ways that did not divide them up between those orderly ones supporting induction and those variable ones that do not.11 Such inconceivability would mean that we find unintelligible any ways of 9 Ronald Dworkin, Taking Rights Seriously (Cambridge, Mass.” Harvard University Press, 1978), Chs. 2-4. 10 One can perhaps beef up Dworkin’s counting of the costs of giving up the practice in question. For some suggestions, see Moore, “Legal Principles Revisited,” Iowa Law Review, Vol. 82 (1997), pp. 867-891, reprinted in Moore, Educating Oneself in Public: Critical Essays in Jurisprudence (Oxford: Oxford University Press, 2001). 11 Kant’s transcendental deduction against Hume’s causal skepticism.

240 MICHAEL S. MOORE

thinking that fails to draw this distinction. (As Willard Quine put it (in a quite different context): if there were a question to beg, we would be begging it.)12 That would mean we cannot self-consciously affirm either this distinction or the presupposition of it (viz causation), not at least as an act of affirmation conditioned on acceptance of the distinction. Because we cannot imagine not accepting the distinction in question, so we cannot imagine having to proclaim that we must accept the distinction and its presupposition because it is too costly not to. Rather, we simply affirm that causation exists. We affirm, analogously, that the statement, “x is negligent,” is objectively true; we do not affirm that it must be regarded as true because otherwise we must change our ways of thinking in certain ways. When successful, in other words, transcendental deductions seem to prove too much. Such successful deductions are indistinguishable from standard realist arguments framed, not in terms of experiences and conceptualizations it would be costly to give up, but in terms of reasons for believing certain propositions to correspond with how things (objectively) are.13 In addition to these two queries about Simchen’s use of a transcendental deduction-like argument, I also wonder about his motives for wanting to use such a form of argument. About the first too motives mentioned above – the unintelligibility and non-resolvability worries – I shall say little. Partly this is due to the fact that Simchen himself does not say enough to reveal why he is skeptical of the traditional metaphysical enterprise; also, I have elsewhere laid out why I think 12 W.V.O. Quine, Word and Object (Cambridge, Mass.: MIT Press, 1960). Quine was referring to our inability to see others linguistic behavior in a way that did not indicate a distinction (between affirmation and denial) being drawn by those where behavior it was. As another example, consider the way that Simon Blackburn (Ruling Passions (Oxford: Calrendon Press, 1998), pp. 54-59) regards the principle of charity in attributing representational states of belief and desire to another creature: Blackburn transforms the principle from a heuristic into an “a priori principle of interpretation” so that we cannot see a creature that did not conform to the principle as even having beliefs or desires. 13 I am not so naïve as to think that any dedicated Kantian would be convinced by these abbreviated remarks. Simchen himself would weaken the deduction so that it displays impossibility but not inconceivability; this, so that we can conceive of what is impossible.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 241

the enterprise is worthwhile, susceptible of progress, and productive of not only an intelligible answer, but a true one as well. 14 Simchen’s gap motivation merits closer attention. I have two questions here. One is whether the gap (between the realist’s truth conditions, on the one hand, and graspability, discoverability, and justification conditions on the other) should force a change in our metaphysical modes of argument. Simchen, like the skeptic and the constructivist in metaphysics, worries about being ship wrecked on the side of metaphysics if ever he admits of a kind of metaphysics that allows a gap to exist between epistemology and metaphysics. I should have thought that the lesson of books like Gil Harman’s Thought15 was that such fears are groundless. They are the product of notions of graspability, discovery, and justification that are too demanding. The lesson we should take from such a gap is a lesson in epistemology, not in metaphysics: we should relax what we require in order to understand something, discover something, justify belief in something, so as to accommodate the fact that our conclusions (in some sense) always outrun our evidence for them. Given the relative firmness of our belief in induction, physical objects, other minds, and the past, in the face of what we recognize to be less than overwhelming evidence for any of them, a proper understanding of a non-foundationalist epistemology is to relax our epistemological beliefs.16 Second, supposing the worry about the gap to be a genuine worry, I wonder whether Simchen’s response could adequately assuage such a worry. We certainly have the experiences on which Simchen’s transcendental deduction rests, viz, we ascribe content to our concepts. In so doing we do indeed seek those “objective similarity relations” between particulars we call kinds, different from the merely “seeming similarities” that historical persons see (or even epistemically idealized persons would see). Yet in that seeking are we not presupposing just the gap Simchen fears, reappearing now within his own preferred mode of argument? Is there not still a gap between what there is (obSee note 2 supra. Gilbert Harman, Thought (Princeton: Princeton University Press, 1971). 16 An argument laid out at somewhat greater length in Moore, “The Plain Truth About Legal Truth,” Harvard Journal of Law and Public Policy, Vol. 26 (2003), pp. 23-47, reprinted in Moore, Objectivity. 14 15

242 MICHAEL S. MOORE

jective similarities), and what our actual (and even our best) theories can accommodate (actual or hypothetical seeming similarities), that Simchen’s approach does not close? In which event, I fail to see why a fear of gaps would motivate Simchen’s approach (again, even conceding arguendo that the gap fear is a genuine fear). III. Simchen’s Semantics I turn now to Simchen’s semantic project. Professor Simchen has joined myself,17 David Brink,18 Nicos Stavropoulos,19 and others20 in thinking that “K-P” Semantics (after Kripke-Putnam):21 (1) is the correct semantics for most terms in a natural language like English; and (2) in the correct semantics for legal terms like “negligence.” I shall consider each of these two issues in turn. A. The Search for the Correct General Semantics Simchen argues that the target of the new, K-P semantics is criterial semantics, according to which the meaning of any term is given by a set of individually necessary and jointly sufficient conditions for its correct usage. Actually the target of K-P semantics is considerably 17 Moore, “The Semantics of Judging,” Southern California Law Review, Vol. 54 (1981), pp. 151-295, “A Natural Law Theory of Interpretation,” Southern California Law Review, Vol. 58 (1985), pp. 277-389; “Do We Have an Unwritten Constitution?,” Southern California Law Review, Vol. 63 (1989), pp. 107-139. 18 David Brink, “Legal Theory, Legal Interpretation, and Judicial Review,” Philosophy and Public Affairs, Vol. 17 (1988), pp. 105-148; Brink, “Realism, Naturalism, and Moral Semantics,” Social Philosophy and Policy, Vol. 18 (2001), pp. 154-176. 19 Nicos Stavropoulos, Objectivity in Law, Oxford: Clarendon Press, 1996). 20 E.g., Leo Katz, Bad Acts and Guilty Minds (Chicago: University of Chicago Press, 1987), pp. 82-96. 21 Saul Kripke, Naming and Necessity (Oxford: Blackwell, 1970); Hilary Putnam, “The Meaning of ‘Meaning,’” Minnesota Studies in the Philosophy of Science, Vol. 7 (1975), reprinted in Putnam, Mind, Language, and Reality (Cambridge: Cambridge University Press, 1975). By “K-P semantics,” I mean to include both what Simchen would classify as Kripke’s and Putnam’s semantics and their metasemantic story as to why their semantics is correct.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 243

broader than that,22 and since it matters to understanding both K-P semantics and Simchen’s version of it, I shall first portray against what K-P semantics was directed. 1. The Target: Conventionalist Semantics On the traditional view of semantics, the meaning of words is a matter of convention. The conventions of our linguistic community have assigned certain properties as fixing the extension of a word like “gold,” or they have assigned the word “gold” to name certain particular hunks of metal (whatever their properties might be). In either case, there are certain analytically necessary truths, statements that are true by convention: “gold is a yellow, ductile metal,” or “the stuff in storage at Ft. Knox is gold,” might be an example of such truths. Conventionalist semantics comes in quite a few varieties. A useful way of organizing those varieties for present purposes is by the resources available to answer the critique of K-P semantics. Let us accordingly group conventionalist semantics into three levels, the levels organized by the degree of reconstruction contemplated for the facts of raw, linguistic usage. At the first and most shallow level, there is what might be called the behavioral semantics of ordinary language philosophy. On this view of semantics, the conventions that give a word its meaning are those conventions accurately generalizing how most native speakers use the word. What it would be odd and not odd to say is used as the touchstone of the meaning of words. Consider the word, “voluntary.” Gilbert Ryle urged that it would be odd to call an action voluntary if it were not up for some kind of appraisal; from this usage fact Ryle concluded that “voluntary” could not mean, a willed bodily movement – for many of such movements are not up for appraisal.23 I call this a behavioral semantics because it does no reconstructive work on the raw data of linguistic usage. (It does not even divide conventions of usage between the semantic conventions related to truth, 22 Putnam, for example, quite explicitly directs his argument against criteriological as well as criterial semantic theories. See Putnam, “Is Semantics Possible?,” in Mind, Language, and Reality, at p. 139. 23 Gilbert Ryle, The Concept of Mind (London: Hutcheson, 1949), at p. 69.

244 MICHAEL S. MOORE

and the pragmatic conventions related only to appropriate utterance.) The second level of conventionalist semantics does some reconstruction of raw usage facts. It distinguishes semantic conventions from merely pragmatic conventions, regarding the semantic conventions as extension-determiners. (An extension in semantic theory is the class of things of which a predicate is true.) At this level one parses usage into one of two kinds of extension-determiners. One is in terms of definitions, which are lists of properties anything within the extension of some predicate analytically must possess. The other is in terms of paradigmatic exemplars, particulars that analytically must be within the extension of the predicate for which they are paradigms.24 I shall describe each briefly in turn. The criterial theory that Simchen mentions is one kind of definitional theory of semantics. It holds that the meaning of a term like “bachelor” is given by a crisp definition: anything that is unmarried, male, and a person is a bachelor. Such a definition gives three properties, possession of each of which is individually necessary and possession of all of which is jointly sufficient for the correct usage of the word, “bachelor.” Another definitional theory is the criteriological theory, according to which there is a list of properties analytically connected to each meaningful word; only, the properties are not individually necessary, and no subset of the properties is jointly sufficient, for correct application of the word.25 Rather, there is simply an overlapping of properties, some determining the extension on some occasions while other properties determine the extension on other occasions. Still, even on this less crisp definitional theory, the entire list of properties is jointly necessary and jointly sufficient for correct use of the word. The paradigm version of this second level of conventionalist semantics is known as the Paradigm Case Argument, or PCA seman24 I explore these in greater depth in my “Semantics” pp. 281-292, and “Natural Law Theory,” pp. 291-292 n. 25, 295-291. 25 A view often attributed to Wittgenstein in his Philosophical Investigations ‘ 67 (Oxford: Blackwell, 3rd edit., 1958). See Carl Wellman, “Wittgenstein’s Conception of the Criterion,” Philosophical Review, Vol. 71 (1962), p. 433; Bill Lycan, “Non-Inductive Evidence: Recent Work on Wittgenstein’s ‘Criteria’,” American Philosophical Quarterly, Vol. 8 (1971), p. 109; Richard Rorty, “Criteria and Necessity,” Nous, Vol. 7 (1973), p. 313.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 245

tics.26 Here it is not words but things that are linked by convention to the word whose meaning is in question. On this view, the meaning of a word like “gold” is given by the things (pieces of gold, presumably) early speakers noticed and baptized with the label, “gold.” “Gold” necessarily applies to those things; if one didn’t apply the word to those items, he would be said not to know the meaning of the word – because it is those items that give the word its meaning. The extension of “gold” includes more than these paradigmatic exemplars. It also includes those items that are similar to the paradigm cases of gold. Such similarity is not to be cashed out in terms of certain properties that the similar items share. For if this were possible, then one could frame a definition out of such properties.27 Rather, the analogies between paradigmatic and penumbral instances within the extension of “gold” are based on a primitive similarity relation, a relation not limited to a few properties in respect of which two things might be similar. The third level of conventionalist semantics is what I have called “deep conventionalism.28 Here one iterates the reconstructions of usage done at level two, so that a term’s usage yields two layers of semantic conventions. There are conceptions (or interpretations) of the meaning of some words given either in terms of definitional properties or in terms of paradigmatic exemplars; and there are concepts giving the meaning of words, again conceived either in terms of properties or exemplars.29 The definitions or exemplars giving the meaning of concepts are more general, deeper, more agreed-upon, than are the definitions/exemplars making up the conceptions of such concepts. The idea is to accommodate considerable disagreement between the conventions that constitute distinct conceptions while preserving the idea that See the citations in my “Semantics,” p. 286. A point much stressed by legal theorists who adopted PCA semantics. See H.L.A. Hart, “Positivism and the Separation of Law and Morality,” Harvard Law Review, Vol. 71 (1958), pp. 593-629; John Borgo, “Causal Paradigms in Tort Law,” Journal of Legal Studies, Vol. 8 (1979), pp. 419-455, at p. 437. 28 See Moore, “Natural Law Theory,” at pp. 298-300. 29 See generally W.B. Gallie, “Essentially Contested Concepts,” Proceedings of the Aristotelean Society, Vol. 56 (1956), pp. 167-198; Ronald Dworkin, A Matter of Principle (Cambridge, Mass.: Harvard University Press, 1985), pp. 128-131. 26 27

246 MICHAEL S. MOORE

there is still a convention-based meaning for every word in terms of that word’s concept.30 2. The K-P Critique of Conventionalist Semantics There are two shoals on which all forms of conventionalist semantics founder. One has to do with when a word should be said to change its meaning. Both disagreements within a culture at a time, and disagreements between cultures over time, are hard to make sense of on conventionalist accounts of meaning. If you (or the ancient Greeks) mean by “whale,” “a big fish,” and I mean something mammalian, how can we disagree? After all, you and the Greeks have fixed the meaning of “whale” one way, and I have fixed it another, so we will just talk past each other even though both sides use the same word, “whale.” Or you (and Norman Malcolm)31 mean by “dreaming” the only criterion we had for dreaming prior to 1950, viz, a waking remembrance of occurrences during sleep known not to be real. Certain scientists discover REM and EEG patterns usually accompanying dreaming, and hypothesize that we do not remember all that we dream. If you fixed the meaning of “dreaming” by the criterion of waking remembrance, then the idea of an unremembered (and certainly an unrememberable) dream is literally senseless.32 As Kripke and Putnam originally pointed out, this inability to capture our sense that these disagreements are meaningful (because the words in terms of which such disagreements are carried on do not change their meaning when used by the opponents in such disagreements), and that one side of such disagreements is or at least can be right (so that science is capable of progress), is a damning indictment of conventionalist semantics. Equally damning is another implication of conventionalist semantics, this one having to do with the idea, not of changing meaning, but of running out of meaning. Suppose one co30 The use to which Dworkin puts such deep conventionalism in Law’s Empire (Cambridge, Mass.: Harvard University Press, 1986). 31 Norman Malcolm, Dreaming (London: Routledge, 1959). 32 Hilary Putnam, “Dreaming and Depth Grammar,” in R.J. Butler, ed., Analytic Philosophy, First Series (Oxford: Oxford University Press, 1962), reprinted in Mind, Language, and Reality.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 247

mes across a piece of metal that is white and ductile; on the criteriological and PCA versions of meaning, there is no answer as to whether or not this piece is or is not gold, for it shares only some of the properties definitive of gold (on the criteriological view) and it is only somewhat analogous to paradigmatic instances of gold (on the PCA view). The word is vague, meaning we have run out of conventions sufficient to place the item definitely in or definitely out of the extension of “gold.” Yet most of us sense that there is an answer as to whether the thing is or is not gold. “Gold,” that is, seems to have a meaning sufficient to determine whether or not these items are within its extension; since we have run out of conventions, meaning must be constituted by something other than these conventions. These two theoretical considerations militate strongly against any form of conventionalist semantics, at least for any discourse where: (1) meaningful, theoretical disagreement about the extension of some predicate exists in the face of there being differing definitions, paradigms, or other supposedly extension-fixing conventions; and (2) meaningful questions about the extensions of some predicate exist in the face of there being no relevant or non-vague definitions, paradigms, or other supposedly extension-fixing conventions. Such considerations point to an alternative semantics, K-P semantics. Simchen helpfully divides this into semantic and metasemantic theses. The semantic thesis is that the meaning of a word is given by (or at least heavily influenced by) its extension. The metasemantic thesis explains how this could be so: speakers discover certain exemplars they provisionally think might be instances of a kind; they baptize the kind with a word (e.g., “gold”); there is a causal chain of usage with each succeeding speaker intending to refer to the kind first baptized with the label, “gold”; expertise develops about what that nature is and what are its exemplars; paradoxically, it may turn out that the initial items people took to be exemplars of the kind are in fact not such exemplars, but merely lucky heuristics to the discovery of the kind. The main payoff of such a semantics is its ability to handle the two theoretical concerns that sink all forms of conventionalist semantics. People can meaningfully disagree because the terms they employ have the same meaning, i.e., the same thing to which their words refer. Peo-

248 MICHAEL S. MOORE

ple do not run out of meaning as fast as they run out of conventions because meaning is a function of the world and its nature, which may only be partially known and thus only partially reflected in conventions. 3. Three Versions of K-P Semantics in Legal Theory It is an interesting question as to what is required to make the K-P metasemantics story plausible for some realm of discourse. Let me distinguish three possibilities here, roughly corresponding to three generations of K-P semanticists within legal and moral theory. a. The standard model. Those of us who were Hilary Putnam’s students in the early 1970’s when he was writing “The Meaning of ‘Meaning’” took away the following interpretation, what I shall call the “standard model” of K-P Semantics.33 On this model K-P Semantics is appropriate when but only when two sorts of facts are true. First, there is (what Simchen aptly calls) an environmental fact: the world must contain the item to which apparent reference is being made in the use of the word in question. In the case of singular terms, these items will be particulars; in the case of general predicates, these items will be universals, namely, kinds.34 For such kinds to be apprehended there must in addition be particulars whose similarities inter se suggest that there is a kind of which they are instances. Such kinds, on the standard model, must not be mere aggregation of individuals; rather the kind must have a nature sufficiently robust and unitary that it can be referred to without mention of its instances. Secondly, there must be what I have elsewhere called “facts of usage,”35 and what Simchen terms, “social facts.”36 On the standard model there are three such usage, or social, facts: 33 See Moore, “Semantics”; “A Natural Law Theory.” See also Mark Platts, Ways of Meaning (London: Routledge, 1980). 34 Thus Putnam carefully separates the question of whether speakers intend to refer to a kind, from the question whether there is in fact a kind, using the example of jade. See his “Meaning.” 35 Moore, “Hart’s Concluding Scientific Postscript,” Legal Theory, Vol. 4 (1998), pp. 301-327, reprinted in Moore, Educating Oneself, at pp. 100-102. 36 Simchen, “Metasemantics,” pp. 215-234.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 249

(i) Speakers must use the predicate in question with indexical intentions, that is, an intention to name whatever local stuff happens to be around them. In Putnam’s famous example, we on Earth use “water” to refer to the local stuff around here (H2O, as it turns out), but the speakers of English on Twin-Earth use “water” to refer to their clear, colorless, life-giving, etc., local stuff (XYY, as it turns out).37 “Water” is thus indexical in the same was as “I,” (Bar-Hillel’s original example of indexicals),38 depending for its reference on whatever is in the environment of the original speakers. (ii) Speakers must use the predicate in question with referential (versus attributive) intentions.39 The distinction is most easily grasped with singular terms, so to use one of Leo Katz’s examples:40 your wife directs you to meet “the man in the Brooks Brother suit, the Yves St. Laurent tie, and the Gucci shoes.” If her intention is to refer to some one particular person, no matter what he is in fact wearing, then her intention is referential; if her intention is to refer to whoever is wearing these three items, then her intention is attributive. For predicates like “gold” analogously, if we speak intending to name a kind whatever its properties turn out to be, our intentions are referential; if we speak intending to name whatever class of individuals turns out to possess the gold-making properties we take to be definitive of gold, then our intentions are attributive. (iii) Speakers must be willing to defer to any well-evidenced expertise others may possess about the true nature of the kind to which all refer. Putnam calls this the “division of linguistic labor,”41 while Simchen relabels it “linguistic deference.”42 It is this deference that makes it plausible how any individual Putnam, “Meaning.” Bar-Hillel, “Indexicals,” Mind, 39 Keith Donellan’s distinction. See his “Reference and Definite Descriptions,” Journal of Philosophy, Vol. 75 (1966), pp. 281-294. 40 Katz, Bad Acts, p. 85. 41 Putnam, “Meaning.” 42 Simchen, “Metasemantics,” p. 224. Simchen pays greater attention to linguistic deference in Coleman and Simchen, “Law,” Legal Theory, Vol. 9 (2003), pp. 1-41. 37 38

250 MICHAEL S. MOORE

speaker can “mean more than he knows” because he can rely on and incorporate the knowledge of others in his referential intentions. One might well call the standard model of K-P semantics the realist model for that semantics’ applicability. For the environmental fact required is a realism about kinds, and the social facts required are facts true only of speakers who are themselves some kind of realist in their metaphysics. My own early application of K-P semantics to legal and moral theory,43 together with like applications by David Brink44 and Mark Platts,45 illustrate this realist understanding of K-P semantics. b. The pedestal model. A less robustly metaphysical view of K-P semantics has been developed by Nicos Stavropoulos46 (and perhaps Ronald Dworkin, if one accepts Stavropoulos’ likening of Dworkin’s interpretivism to K-P semantics).47 On this model, the environmental fact needed to ground K-P semantics is quite modest: only that there be certain particulars that can be classed together, and a concept so grouping these particulars together and which is accepted by the speakers who use the word.48 The usage facts are also comparatively modest: speakers need to have indexical intentions, making their words’ reference hostage to the accidents of environment; and speakers must be willing to defer to expertise when it is possessed by those with theories about the concepts that group the particulars in question. 43 Moore, “Semantics”; “Moral Reality”; “A Natural Law Theory”; “Unwritten Constitution”; “Moral Reality Revisited.” 44 Brink, “Legal Theory”; Brink, Moral Realism and the Foundations of Ethics (Cambridge: Cambridge University Press, 1989). David is more guarded in the metaphysical commitments of K-P semantics in his “Semantics and Legal Interpretation (Further Thoughts),” Canadian Journal of Law and Jurisprudence, Vol. 2 (1989), pp. 181-191. 45 Platts, Ways of Meaning. 46 Stavropoulos, Objectivity in Law. 47 Id. at pp. 129-136, 160; compare Moore, “Postscript,” p. 102 n. 7 (in Educating Oneself). 48 On Stavropoulos’ version of K-P semantics, “key concept-words are intended to pick out the concepts they stand for, whatever their content may be,” and “the content of the relevant concepts is determined by substantive theory, which is constrained by paradigmatic applications and abstract characterizations of the relevant practice of application.” Objectivity in Law, p. 160.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 251

(This latter feature is what leads Simchen to accuse Dworkin and Stavropoulos of placing theorists on a pedestal, thus my name for this model).49 On the “pedestal” view, there is no need for speakers to have referential intentions with respect to a kind; it is enough that they are willing to defer to experts about a concept of the kind. c. Simchen’s Version of K-P Semantics. Like Stavropoulos, Simchen dispenses with any metaphysics of kinds as a presupposition of K-P semantics. The only environmental fact needed is that there be some particulars susceptible to grouping into a class by speakers. (There apparently need not even be a shared concept of how such grouping should be defined.) Simchen also dispenses with referential intentions with respect to kinds, again like Stavropoulos. But unlike Stavropoulos, Simchen rids K-P semantics of any need for linguistic deference; speakers need not be willing to defer to anyone, even in principle, because they rightly think themselves to be in possession of the correct classificatory scheme grouping certain particulars together. Such classificatory scheme is “correct” only in the sense that correctness is here a matter of convention: if some scheme is in accord with the classificatory abilities possessed by most native speakers, then it is “correct.”50 Coleman and Simchen, “Law,” pp. 10-11. See id. at p. 20: “Schematically, to be a chair is to be taken by the average speaker as having the same intended function, general appearance, and so on, as paradigmatic chairs. Determining whether or not some item bears the sameness relation to a paradigmatic chair is something which an average speaker can be expected to do…the ‘essence’ of chairs depends on ordinary speakers’ everyday classificatory capacities.” See also id., p. 22 (“the equivalence relation itself was determined by speakers’ ordinary stuff-involving classificatory capacities”); p. 28 (“Whether or not sameness obtains between a given item and a paradigmatic instance of law is determined by the average speaker’s ordinary classificatory capacities”); p. 28 fn. 39 (“whether or not the relevant similarity relation obtains…is determined by the average speaker’s ordinary classificatory capacities”); p. 30 (“Something belongs to the extension of ‘law’ just in case it would be deemed by the average speaker as relevantly similar to paradigm cases”); p. 33 (“the extension...is fixed by the average speaker’s classificatory tendencies”); p. 39. It is a bit of an understatement to say that I am at a loss to reconcile this strait-forward conventionalism with the objective (i.e., non-conventional) similarity relations that determine extension in Simchen’s present paper. 49 50

252 MICHAEL S. MOORE

4. Queries About Simchen’s Version of K-P Semantics Simchen and I are in agreement in rejecting Stavropoulos’ version of K-P semantics, although our reasons for doing so are somewhat different. We both think that Stavropoulos has failed to justify any deference to theorists by ordinary users of English predicates. I think that, however, because I link the justifiability of such linguistic deference to there being something, a kind, with a deep nature amenable to theoretical treatment; Stavropoulos’ requirement that there be a concept justifies only a theory of the nature of that concept, a kind of deep conventionalist semantics at best. Whereas Simchen rejects Stavropoulos’ theorizing on the ground that no such theorizing or deference is required (nor is there any in fact for many words) for K-P semantics to be applicable; all that is needed are the ordinary classificatory abilities of native speakers, which abilities require and in fact generate no deference to experts. Turning to Simchen’s version of K-P semantics, my first query is whether his version can garner for itself the two theoretical advantages that K-P semantics possesses vis-à-vis all forms of conventionalist semantics. Consider first the constancy of meaning that makes possible radical disagreement. Without the environmental fact that reference to a genuine kind has succeeded, and without the usage fact that speakers typically intend to refer to such kinds in their usage of the relevant words, I don’t see how meaning remains constant across very divergent beliefs. All Simchen’s version of K-P semantics has to work with is the environmental fact that there are certain particulars picked out by a term, grouped into the extension of that word by the normal classificatory abilities of native speakers, and the usage fact that speaker’s intentions are indexical, i.e., the word is intended to pick out whatever particulars are in the vicinity of native speakers. These two facts are perhaps sufficient to reject criterial semantics, for the actual and intended indexicality of a term makes the ordinary criteria for use inadequate to determine reference; “water” as used on Earth has the same criteria of use as on Twin-Earth, and yet the reference is different in the two different environments. Yet these two facts are not sufficient to show how there can be the converse situation, namely, where the reference is the same but the criteria are different, as in the “whale” and “dreaming” examples earlier. And it is this latter

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 253

kind of example that is needed to show constancy of meaning despite very different criteria for use. Even if we and the Greeks both happened to stumble across genuine instances of gold, our differing beliefs about the stuff could generate sufficiently different classificatory schemes by the two groups of speakers that there was little overlap – in which case the extension of “chrysos” and of “gold” would differ and we and the Greeks would be talking past each other when we disagreed about the nature of gold. Now consider the second theoretical advantage of K-P semantics, that of successful reference despite vague, non-existent, or conflicting conventions guiding usage. Only the intent to refer to a kind whose nature outstrips current convention, together with the existence in fact of such a kind, generate the needed implication about successful reference despite indeterminate conventions. The classificatory abilities shared by native speakers is just a convention, and like other conventions it is no more comprehensive than the behavior from which it is constructed. Where ordinary speakers’ classificatory tendencies are confused, fragmented, or conflicted, there will be no answer to the question of whether some item is within the extension of some predicate. K-P semantics in Simchen’s version fares no better here than any other form of conventionalist semantics. Which introduces my second query: isn’t Simchen’s version of K-P semantics just a reversion to some form of conventionalist semantics? Not the definitional form of such semantics, for the reason mentioned earlier (indexicality prevents sameness of definition from guaranteeing sameness of extension). Yet how does Simchen’s semantics differ from either PCA semantics or the shallow, behavioral semantics of ordinary language philosophy? Consider each in turn. Whether Simchen’s semantics collapse back into the old PCA semantics depends on how Simchen regards the paradigms which speakers baptize with the name of a class. Many years ago I distinguished strong from weak paradigms.51 A strong paradigm is a particular that is (analytically) necessarily within the extension of the predicate for which it is a paradigm. Such paradigms are tied by convention to words, so that anything that is a paradigmatic exemplar of blue, or is relevantly similar to such exemplars, is necessarily blue. Weak para51

Moore, “Semantics,” pp. 287-288.

254 MICHAEL S. MOORE

digms, by contrast, are no more than heuristics: they indicate to speakers that they are instances of a kind. But howevermuch such paradigms were the original evidence for the existence of a kind, howevermuch they are the standard learning tools by which a culture teaches the use of the kind-word, weak paradigms may turn out not to be instances of the kind at all. The original exemplar of flat may have been the ocean, the way “flat” is taught may be by pointing at the ocean, yet the ocean is not in fact within the extension of “flat.” It just looks flat. I am unclear how Simchen’s semantics permits any but strong paradigms. After all, if there need be no kind referred to by some predicate, only some particulars sharing whatever properties as happen to be picked out by the ordinary classificatory abilities of native speakers, then what would make any paradigm only provisionally within the extension of some predicate? To what deeper insight is its status hostage, in the absence of any but the most conventional nature? If Simchen’s semantics does rely on there being strong paradigms, similarity to which determines the extension of each predicate, then Simchen’s semantics are indistinguishable from the conventionalist PCA semantics of fifty years ago, a semantics to which K-P semantics was supposed to be the antidote. Perhaps, however, Simchen wishes to regard paradigms as weak, which is to say, provisional and defeasible. Perhaps the linguistic dispositions of native speakers is held to trump the paradigmatic status of any given particular (although it is hard to see how such classificatory dispositions could dispense with all such paradigms, as it should in principle be able to do if the paradigms are truly weak paradigms). Yet then, what distinguishes Simchen’s semantics from the behavioral approach of ordinary language philosophy? Common to both is ultimate reliance on what people are disposed to say, such common, classificatory sayings determining what they are talking about. My third query has to do with why Simchen is attracted to a version of K-P semantics that is stripped of the metaphysics of kinds, stripped of intentions to refer to such kinds in the typical uses of words, and stripped of any deference by ordinary speakers to the expertise others may possess about the nature of such kinds. One temptation for this stripped-down version of K-P semantics could be ontological: one

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 255

could doubt the realist (i.e., anti-nominalist) metaphysics of kinds, either across the board or at least for many of the predicates making up a language. Simchen’s motives, however, do not seem to stem from ontological parsimony. Rather, his doubts are rooted in the usage facts depended upon by the standard version of K-P semantics. He doubts whether ordinary speakers have the metaphysical views he thinks they would have to have in order to intend to refer to kinds with a nature others may know better than do they. It is worth quoting Simchen here, since he seeks to load the dice a bit. Simchen thinks that any attribution “to ordinary speakers of the metaphysical realist intention to employ ‘water’ to refer to anything relevantly similar to paradigmatic instances of water from the standpoint of the world as it is in itself, beyond whatever we might come to believe about the matter,”52 is highly implausible. Or again: “such a view attributes to speakers, when using a kind term N, the intention to refer to everything having the same underlying nature as some paradigmatic sample of N quite apart from what any expert doctrine about the nature of N does or would reveal.” 53 There are several things to untangle in these rather exaggerated characterizations of the referential intentions needed by the standard version of K-P semantics. To begin with, in this context the distinction between Peircean (or Putnam’s “internal”) realism and metaphysical (or “external”) realism, is a red herring. I doubt (as does Simchen) that ordinary speakers’ referential intentions are sufficiently fine-grained so as to pick out one or the other of these metaphysical views. Fortunately, however, this does not matter to the issues at hand, which are (1) whether such speakers presuppose that there is a kind to which they intend to refer and about the nature of which they intend to defer when confronted with a better theory (the usage fact); and (2) whether there is in fact such a kind (the environmental fact). As I have argued elsewhere,54 the internal realist can match the metaphysical realist stride for stride in these commitments, so a presupposition of either form of realism (or the undifferentiated combination of both) is suffi52 53 54

Simchen, “Metasemantics,” p. 227. Coleman and Simchen, “Law,” p. 36 n. 43. Moore, “Legal Reality,” p. 694.

256 MICHAEL S. MOORE

cient to support the referential intentions and metaphysical presuppositions I argue are needed to use K-P semantics. The second clarification has to do with the place of paradigmatic examples and similarity functions in ordinary speakers’ referential intentions. We should distinguish the referential intentions of the original baptizers of a kind, from those far down the causal chain of reference. Only the baptizers need have before them puzzlingly similar particulars from which they self-consciously hypothesize a kind; later users need not think about (or even believe in the existence of) any paradigmatic exemplars.55 Their intentions can be simpler: to refer to a kind with their general predicates in the same way they refer to a particular with their singular terms. Such later users need have in mind no complex function of similarity relations over certain particulars. Notice that both of these points simplify considerably the content of the referential intentions the standard version of K-P semantics would attribute to contemporary language users. The first rids that content of any total independence “from what any expert doctrine about the nature of N does or would reveal.” The second rids that content of any isolation of paradigmatic samples and universal quantification over particulars sharing the nature of such samples. The content of the relevant referential intention is easier: it is to refer to a kind that exists independently of whether the speaker or her community thinks it exists (a realism about universals); and such intention is accompanied by the belief that the nature of the kind may only be partially revealed (to either the individual speaker or to her linguistic community). That these simpler psychological states are required on the standard version of K-P semantics makes the latter version more plausible because these states are more easily ascribed to ordinary speakers. Having clarified the content of the requisite referential intentions, it remains to clarify the nature of the claim made when it is claimed that a speaker has the requisite referential intention and accompanying belief. As Simchen recognizes, this is of a piece with one’s general views on what is required to ascribe intentional attitudes to another. 55 Other than for the word, “meter,” can one identify plausible paradigms? Surely any original bits of metal, for example, have been lost to us even if they were our initial samples of gold.

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 257

One thing that is not required (on this I assume Simchen agrees) is that there be some Joycean phenomenology explicitly containing the content of the intention and the belief. We do not require such conscious recitations to ascribe intentions and beliefs generally, so there is no warrant for requiring such here. What is required to ascribe intentions and belief is that there be certain dispositions, which is to say that certain counterfactuals are true of the individual who mental states they are. In the case of referential intentions, the most pertinent dispositions are verbal dispositions, specifically: what would the speaker think and say on learning certain surprising facts about some subject of his discourse? To enlist an old intuition pump of mine,56 suppose the speaker has pronounced as dead a person who has lost consciousness and whose heart and lungs have ceased spontaneous functioning because he has been immersed under very cold water for thirty minutes. What would such a speaker think if presented with the conclusion that the drowning victim “is not really dead” and if presented with the medical evidence supporting that conclusion (intact brain function, revivability, etc.)? If the speaker’s intentions had been attributive in using the word “dead” – so that anything that possesses the properties definitive of “death” for the speaker is necessarily dead – then he should refuse the conclusion as senseless. Such a victim may not be smead (a new state defined by brain function), but that victim is necessarily dead? Whereas if the speaker’s intentions were referential, then he would readily accede to the meaningfulness of the conclusion and to the relevance of the evidence for sustaining it; he would thus regard his own conclusions about death as fallible and recognize that experts’ views about death might well be better than his, even though theirs too are fallible. These are the beliefs of a realist about the kind, death, even though such a speaker is wholly ignorant of the realism/anti-realism debate in philosophy. My own empirical intuition is that such referential intentions are quite widespread, both as to people holding them and as to words with respect to which they are held. Reverting to the death example, surely few native speakers of English will cut the organs out of a drowning victim who meets the prevailing definition of “death” but who is not really dead. 56

Moore, “A Natural Law Theory,” pp. 293-294, 297-300, 322-328.

258 MICHAEL S. MOORE

B. The Reach of K-P Semantics: the Search for the Correct Semantics for Terms Used in Law and in Legal Theory Simchen and I apparently agree that law is an artifactual notion (what I call a functional kind),57 that K-P semantics applies to the terms referring to artifacts, and because of these facts, that K-P semantics applies to the terms used both in propositions of law and propositions about law. Yet because of our disagreement about what is required for K-P semantics to be appropriate, this apparent agreement masks some serious disagreements about the semantics of legal terms. My view about legal terms is the view of the standard version of K-P semantics: (a) Environmental facts: kinds such as contracts and law exist in the world irrespective of us thinking that they exist and words like “contract” used in propositions of law, and words like “law” used in legal theory, take their meaning from the nature of these kinds and not from conventional guides to usage; (b) Usage facts: ordinary users of “contract” and “law”: (i) have indexical intentions to refer to the contract-stuff and law-stuff exemplified around here; (ii) have referential intentions to name these kinds whatever their properties turn out to be; and (iii) have the conditional intention to defer to any well developed expertise of lawyers or legal theorists about the nature of these kinds should any such expertise appear. Simchen, by contrast, is agnostic about (a) and denies (b)(ii) and (b)(iii). Although disputes about empirical facts between philosophers firmly planted in their armchairs are doubtlessly to be viewed with some skepticism, my own sense is that ordinary language users are actually deferential to legal professionals about terms like “contract” appearing in propositions of law.58 H.L.A. Hart noticed a facet of this 57 Compare Coleman and Simchen, “Law,” with Moore, “A Natural Law Theory,” p. 301 n. 44, and Moore, “Law as a Functional Kind,” in R. George ed., Natural Law Theories (Oxford: Clarendon Press, 1992), reprinted in Moore, Educating Oneself. 58 Simchen appears not to disagree about deference to experts on terms of law (Coleman and Simchen, “Law,” p. 26), only about deference to experts for terms of legal theory such as “law.”

CAN OBJECTIVITY BE GROUNDED IN SEMANTICS? 259

many years ago when he pronounced such legal terms to be “defeasible.”59 My sense also is that ordinary language users and lawyers are potentially deferential to legal theorists about terms of legal theory like “legal right,” “legal duty,” “second-order reasons,” “sovereignty,” “rule of recognition,” “legal system,” and ultimately, “law” itself. Remembering that the relevant question is the counterfactual one – what would such speakers think if they were fully informed about a better theory of law, etc.? – the question is not actual deference to legal theorists by lawyers and laypersons. Nor is the question whether there is currently any such expertise in jurisprudence, nor even whether there will ever be. It is after all quite possible that the lawyers’ and the laypersons’ current understanding of law (assuming arguendo there is one) just happens to be correct. Then there can be no better theory to which one will or should ever defer. Yet the question is how such language users regard their knowledge of the nature of law. If they regard themselves (collectively) as infallible – because the meaning of “law” is given by conventional criteria and they have grasped those criteria—then they lack the referential intention and deferential attitude I think necessary for K-P semantics. But if they regard their beliefs about law as in principle correctible, then they do not think the meaning of “law” is given by conventional criteria but by the nature of the kind to which reference is made. So clarified, I thus think that even for the terms of legal theory like “law,” the usage facts needed on the standard model of K-P semantics are present. IV. Connecting Simchen’s Semantics to Simchen’s Transcendental Argument for Objectivity I find Professor Simchen’s version of K-P semantics worth discussing for its own sake. Almost none of the details of those views, however, have much bearing on the larger issue of objectivity. More specifically, one could adopt any of the three versions of K-P semantics 59 H.L.A. Hart, “Ascription of Responsibility and Rights,” Proceedings of the Aristotelian Society, Vol. 49 (1949), pp. 171-194. I reallocate Hart’s defeasibility conclusion towards a K-P semantics basis in “A Natural Law Theory,” pp. 337-338, and in “Legal Reality,” pp. 666-669.

260 MICHAEL S. MOORE

above outlined and base Simchen’s objectivity argument on it. For this larger purpose, our differences are thus bootless. Much the same I fear is also true about the difference between all versions of K-P semantics, on the one hand, and almost all versions of conventionalist semantics, on the other.60 For the only mileage Simchen takes out of K-P semantics is the supposition of the latter that the similarities between particulars making up a kind must be objectively existing similarities, mere subjective (or “seeming”) similarities not being acceptable. Yet surely this reality/appearance distinction is available on virtually any of the semantic possibilities we have canvassed. If there are conventional criteria, for example, that give the meaning of “gold,” then presupposed is that it is only when those criteria are actually (objectively) satisfied is the word used correctly and that a mere seeming satisfaction is insufficient. Perhaps Professor Simchen would not wish to deny this, allowing that any semantics even plausibly endowing our terms with content must presuppose objectivity. This includes, but is not limited to, the K-P semantics Simchen favors. If this is Simchen’s belief, then it is only the organization of the paper that seems to make a stronger claim. The organization – moving from objectivity to K-P semantics and back again – might suggest that the justification for the excursion into K-P semantics is that only that form of semantics presupposes objectivity. Whereas in reality perhaps the justification for the excursion is the intrinsic interest of K-P semantics itself.

60 Only the supposed semantics of a private language (against which Wittgenstsein protested) seems ruled out as a possible basis for Simchen’s transcendental argument.

PART II RIGHTS THEORY

12

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

A Hybrid Theory of Claim-Rights Gopal Sreenivasan* The language of rights is pervasive. As commonly used, however, it is also ‘indiscriminate’ and ‘loose,’ as Wesley Hohfeld complained long ago.1 In a celebrated effort to forestall confusion, Hohfeld distinguished various senses of ‘a right,’ which he regimented in terms of four jural equivalences or correlatives. Of these, he himself regarded the equivalence in which rights are correlative with duties as regimenting rights ‘in the strictest sense’: if X has a right against Y that he shall stay off the former’s land, the correlative (and equivalent) is that Y is under a duty toward X to stay off the place. If, as seems desirable, we should seek a synonym for the term ‘right’ in this limited and proper meaning, perhaps the word ‘claim’ would prove best (Hohfeld, p. 38).

Now few dispute that Hohfeld’s ‘claim’ at least marks a central and important sense of ‘a right.’2 Indeed, it is perfectly standard to define claim-rights, as they are more often called, on the model of Hohfeld’s equivalence: X has a claim-right against Y that Y f if and only if Y is under a duty toward X to f.3 *

Department of Philosophy, University of Toronto. W.N. Hohfeld, Fundamental Legal Conceptions, ed. W.W. Cook (New Haven: Yale University Press, 1919), p. 36. 2 Some dispute that rights have ‘a’ strictest sense. See, e.g., L.W. Sumner, The Moral Foundation of Rights (Oxford: Clarendon Press, 1987), ch. 2. 3 See, e.g., J. Feinberg, Social Philosophy (Englewood Cliffs, N.J.: PrenticeHall, Inc., 1973), ch. 4; J. Waldron, “Introduction” to his (ed.) Theories of Rights (Oxford: Oxford University Press, 1984), p. 8 [hereafter, TR]; Sumner, pp. 25-7; J. Thomson, The Realm of Rights (Cambridge, Mass.: Harvard University Press, 1990), 1

264 GOPAL SREENIVASAN

So defined, one person’s possession of a claim-right is equivalent to someone else’s possession of a duty—a duty, moreover, with the same content.4 But neither this fact nor anything else Hohfeld says, for that matter, tells us how we are to identify the relevant pair of persons. If I have a claim-right to some land, then someone else has a duty to stay off the land. But who has this duty exactly? Similarly, if I have a duty to pay my taxes, it may be asked who, if anyone, holds the correlative claim-right. In certain cases, we may think this is easy to say. As far as my claim-right to land goes, for example, we may think the bearer of the correlative duty is everyone. But other cases will not be so easy. How, in general, are we to identify the bearer of the duty that correlates with a given claim-right? Or the holder of the claim-right that correlates with a given duty? If a correlation between the right-holder and the duty-bearer belongs to the nature of claim-rights, as the standard definition suggests, an adequate understanding of claim-rights requires an understanding of the basis of this correlation. Rival accounts of the correlation are offered by two well-established theories of rights, the Will theory and the Interest theory.5 Yet neither theory, it seems to me, is ultimately satisfactory. To begin with, I shall present each theory, together with the main problems it faces. In my view, the best objection that each theory wields against the other is unanswerable. More constructively, I shall then suggest a hybrid of the two theories. I shall argue that it solves the main problems confronting the Will and Interest theories. We should therefore prefer the hybrid theory. pp. 41-3; M. Kramer, N. Simmonds, H. Steiner, A Debate Over Rights (Oxford: Clarendon Press, 1998) [hereafter, DR]. 4 In Hohfeld’s example, the claim-right and the duty share the content ‘that Y stay off X’s land.’ To preserve idiom, we could also say, alternatively, that they share a content that is satisfied by ‘Y’s staying off X’s land.’ 5 In presenting the debate between these theories in terms of their account of the correlation between claim-right holder and duty-bearer, I follow Waldron, TR, pp. 8-9, and Sumner, pp. 24 and 39-45. While there are other ways to frame the debate, I do not think the choice of frame affects the argument at any point. My choice is based on independent grounds, which I discuss in a companion paper, “Duties and their direction.”

A HYBRID THEORY OF CLAIM-RIGHTS 265

I §1. Let me begin with the Will theory.6 The following rough statement will serve our purposes: (WT) Suppose X has a duty to f. Y has a claim-right against X that X f just in case Y has some measure of control over X’s duty.

To explain what is meant by a ‘measure of control’ over a duty, we should turn to H.L.A. Hart, one of the foremost proponents of the Will theory. According to Hart, the full measure of control over X’s duty comprises three powers: (i) the power to waive X’s duty or not; (ii) the power to enforce X’s duty or not, given that X has breached it; (iii) the power to waive X’s duty to compensate, which is consequent upon his original breach. 7

Note that the power to enforce X’s duty in (ii) includes both the power to sue X for compensation and the power to sue for an injunction against X. I think it is fair to say that our clearest paradigms of a claim-right are the claim-rights recognized in property and contract law. The Will theory bases itself closely on these paradigms. Indeed, it holds, in effect, that they present the necessary and sufficient conditions for claim-right holding. It therefore stands to reason that, in property and contracts, the duties that correlate with claim-rights are duties over which the claim-holder typically has the full measure of control encompassed by the powers (i)-(iii). It is the signal advantage of (WT) that Y’s having the full measure of control over X’s duty to f gives a readily comprehensible sense to 6 Throughout the discussion, I shall ignore the distinction between moral rights and legal rights. That is, I shall not pay special attention to it. While this will sometimes involve me in (minor) infelicities, it conveniently enables me to concentrate on the core features of a claim-right, which hold in common between law and morals. 7 H.L.A. Hart, “Legal Rights” in his Essays on Bentham (Oxford: Clarendon Press, 1982), pp. 183-84 [hereafter, “LR”].

266 GOPAL SREENIVASAN

the statement that X’s duty is owed to Y, and so to the statement that it is Y who holds the correlative claim-right against X. Lesser measures of control can be accommodated as approximations to the case of full control. Thus, as Hart says, ‘duties with correlative rights are a species of normative property belonging to the right holder, and this figure becomes intelligible by reference to the special form of control over a correlative duty which a person with such a right is given by the law’ (“LR,” p. 185). But (WT) confronts two serious objections. One concerns inalienable rights.8 Sometimes a claim-right holder is disabled from waiving the duties that correlate with his claim-right. Typically this is done for the right-holder’s own good and protection. Moreover, as Neil MacCormick observes, the protective disability is typically also seen as strengthening the claim-right. A dramatic example is the claimright not to be enslaved. Less dramatic examples include the claim-right not to be operated upon without informed consent;9 and the claim-right not to be employed in unsafe working conditions. (In some ways, the less dramatic examples are actually more important, since they exhibit the fact that inalienable claim-rights need not be correlated with extremely weighty duties. Hence, the strength added by a protective disability is distinct from the weight, in that sense, of the original claim-right.)10 In any case, we likely do not wish to deny either that Y has a claim-right against X that X not enslave Y or that Y 8 See D.N. MacCormick, “Rights in Legislation,” in P. Hacker and J. Raz (eds.) Law, Morality, and Society (Oxford: Clarendon Press, 1977), pp. 195-99 [hereafter, “RL”]. 9 The inalienable claim-right here, to be precise, is the claim-right to receive a standard disclosure prior to consenting (to an operation or other medical treatment). In U.S. law, this claim-right may not be inalienable, since it seems that a physician’s duty to disclose can actually be waived. See J.W. Berg, P.S. Appelbaum, C.W. Lidz, and L.S. Parker, Informed Consent: Legal Theory and Clinical Practice, second ed. (New York: Oxford University Press, 2001), ch. 4. But given the standard analysis of the requirements of informed consent—on which they include understanding the standard disclosure—a patient’s having the power to waive the physician’s duty to disclose is incoherent. On the standard analysis, therefore, the claim-right is inalienable. 10 The second of the less dramatic examples also makes it clear that what the protective disability protects need not be the autonomy, specifically, of the claimright holder.

A HYBRID THEORY OF CLAIM-RIGHTS 267

has a claim-right against X that X not employ Y in unsafe working conditions. It is worth emphasizing that the crucial question here concerns the possibility, rather than the fact, of inalienable claim-rights. (WT) makes inalienable claim-rights incoherent in principle. A second criticism concerns incompetent adults.11 Say Y is incompetent to exercise any of—and therefore lacks, in the relevant sense—the powers (i)-(iii) because, for example, Y is in a coma. Do we wish to say that Y no longer has a claim-right against X that X not assault Y or steal from Y? Or that X no longer owes Y duties not to assault or steal from Y? Presumably not. But (WT) implies that someone with no measure of control over a duty lacks the correlative claim-right. In the paradigm cases of a claim-right, the interests of the right-holder are advanced, on balance, by the fact that he or she is empowered to waive the correlative duty. One way to regard these objections is this: they present cases in which the interests of the person in question are not advanced, on balance, by being so empowered. Indeed, they present cases in which the person’s interests are advanced, on balance, either by not having the power to waive the relevant duty (first objection) or by someone else’s having that power (second objection). Yet it seems intuitive to regard the person disabled from waiving the relevant duties as still holding the correlative claim-right. The fundamental difficulty with (WT), then, is that it prevents us from generalizing the notion of a claim-right from the paradigm cases to cases of inalienability and incompetence, cases to which we clearly should be able to generalize it. It seems to me that this difficulty cannot be overcome. §2. By way of illustration, let us consider Nigel Simmonds’ recent response to the inalienability objection.12 Simmonds discusses partial and complete inalienability separately. In the partial case, the agent lacks (i) the power to waive a duty, but retains (ii) the power to sue for enforcement and (iii) the power to waive compensation. In the complete case, the agent has no control over the duty whatever. Here Simmonds’ discussion concerns criminal law prohibitions against murder and assault. 11 A better known variant of this objection concerns children. But I think this variant is liable, with reason, to greater controversy. 12 N. Simmonds, “Rights at the Cutting Edge,” in DR, pp. 225-232.

268 GOPAL SREENIVASAN

Simmonds argues that the Will theory does, in fact, vest agents in the partial case with a correlative claim-right, since they retain a residual measure of control over the duty, of precisely the kind Hart describes. Furthermore, he denies that this reduction from the full measure of control is inconsistent with strengthening the claim-right, on the ground that it is over-simple to identify the strength of someone’s claim-right with the measure of her control over the correlative duty. To handle the case of complete inalienability, Simmonds invokes the distinction between legal rights and moral rights. On the legal side, he is content to affirm (WT)’s implication that the criminal law confers no claim-right against murder or assault. On the moral side, he is content to insist that the duties not to murder and not to assault really can be waived. Neither of these replies withstands scrutiny. It is true that (WT) vests agents who have even a residual measure of control over a duty with a correlative claim-right. But it cannot seriously be maintained that reducing the right-holder’s measure of control is consistent with strengthening her claim-right, at least not on the Will theory’s conception of a claim-right. Someone with only a residual measure of control over my duty to f lacks the ability to exert her will in certain ways—notably, to make it the case that my failure to f does not count as a breach of my duty. How can this not weaken her ability to exert her will, and so not weaken her claim-right on (WT)? Simmonds’ second reply is beside the point. In either law or morals, one might debate whether or not a certain claim-right is completely inalienable. Let the example be one’s favourite. For present purposes, it is simply irrelevant to insist that one position or the other in the resultant debate is correct. The question is whether the debate itself is coherent; and the very fact that Simmonds can engage in the debate shows that it is. But this contradicts (WT), which excludes the coherence of asserting that any claim-right is completely inalienable. II §1. Let me now introduce the Interest theory. The following rough statement will serve our purposes to begin with:

A HYBRID THEORY OF CLAIM-RIGHTS 269

(IT) Suppose X has a duty to f. Y has a claim-right against X that X ö just in case Y stands in a sanctioned relation to benefiting from X’s f-ing.

This formulation is slightly odd. But it allows (IT) to cover a number of subtle variations in the basic structure of the Interest theory. For now, we can think of ‘standing to benefit from X’s f-ing’ and ‘being intended to benefit from X’s f-ing’ as the sanctioned relations. (IT) also gives a comprehensible sense to the statement that X’s duty is owed to Y, and so to the statement that it is Y who holds the correlative claim-right against X: namely, the duty is for Y’s benefit. (IT) is more general than (WT) in two important respects. First, it extends the notion of a claim-right to a wider range of cases than (WT) does. In particular, (IT) extends the notion to the cases of inalienability and incompetence that motivated the objections to (WT). On the plausible assumption that X’s enslaving Y or employing Y in unsafe working conditions or assaulting a comatose Y sets Y’s interests back, (IT) yields the verdict that X’s duties not to perform any of these actions are still owed to Y, and so that Y still holds the correlative claim-right. Second, (IT) is associated with a more general account of the justification of claim-rights. On the account associated with (WT), the justification for empowering Y to waive the duty correlative to her claim-right, and so for vesting her with the claim-right, lies in the fact that so doing serves Y’s interest in autonomous choice.13 In the paradigm cases, empowering Y to waive this duty also advances her interests on balance. By contrast, on the account associated with (IT), the justification for the structure of Y’s normative standing, as we might put it, lies in the more general fact of what advances Y’s interests on 13 Hart, “LR,” pp. 188-89. On some versions of the Will theory, it would be objectionable to characterize the justification of Y’s power to waive in terms of her interest in autonomous choice. But nothing turns on this formulation, at least not for my purposes. I could as well describe the justification associated with (WT) as ‘appealing in some fashion’ to the value of (individual) autonomy. The formulation in the text makes the greater generality of the justification associated with (IT) explicit on the surface of the two accounts. However, the facts about which account is more general hold independently of this formulation (see the following note).

270 GOPAL SREENIVASAN

balance. It is not tied to the more specific fact of what advances Y’s interest in autonomous choice.14 When these facts coincide, as they do in the paradigm cases, (IT) yields the same results as (WT) (cf. MacCormick, “RL,” pp. 207-8). But when they diverge, as they do in the cases motivating the objections, (IT) classifies duties that advance Y’s interests as owed to Y, even if Y has no measure of control over them, as long as Y’s disability with respect to these duties advances her interests on balance. At least one serious objection can be raised against (IT). This concerns the problem of third party beneficiaries.15 In simple form, the problem is as follows. Suppose you promise your brother to pay your sister $100. Ordinarily, we would say that your brother now has a claim-right against you or that you now owe a duty to your brother to pay your sister $100. Hart questions whether (IT) yields this verdict.16 However that may be, (IT) certainly yields the verdict that your duty to pay your sister is (also) owed to your sister, and so that your sister (also) has a claim-right against you, since she benefits from the $100. Hart also maintains that this verdict is incorrect.17 However that may be, it would certainly be the wrong verdict if your duty to pay your sister were also owed to your sister’s child, on whom—let us say—she will spend the $100, so that her child had a 14 The justification of claim-rights associated with (IT) can let whatever it is about the value of (individual) autonomy that grounds the justification associated with (WT) weigh in favour of empowering Y to waive X’s duty to f. Its greater generality consists in the fact that it also allows other factors—to wit, aspects of Y’s well-being that are independent of her autonomy—to weigh against empowering Y to waive this duty. It is irrelevant to this claim whether advocates of (WT) would assign justificatory weight to these other factors, so long as you and I do. 15 See, notably, H.L.A. Hart, “Are there any natural rights?”[1955] in Waldron, TR, pp. 81-2; and “LR,” pp. 187-8. 16 See previous note. Some interest theorists reply by affirming that your brother has an interest in your fulfilling your promise to him, in which case (IT) will vest him with a claim-right against you. See, e.g., D. Lyons, “Rights, Claimants, and Beneficiaries,” [1969] reprinted in his Rights, Welfare, and Mill’s Moral Theory (New York: Oxford University Press, 1994), pp. 42-4; and M. Kramer, “Rights Without Trimmings,” in DR, pp. 79-80. 17 See note 15; cf. H. Steiner, An Essay on Rights (Oxford: Blackwell, 1994), pp. 61-2. Kramer protests that it begs the question to deny that your sister has a claim-right against you (DR, pp. 66-8). Cf. Lyons, pp. 37-41.

A HYBRID THEORY OF CLAIM-RIGHTS 271

claim-right against you. But (IT) clearly appears to yield that verdict as well. More generally, the objection is that, intuitively, there is a limit to the number of people to whom duties are owed, and so to the number of claim-rights that arise, under a third party promise or contract. Indeed, for many duties, there is an intuitive limit to the number of people to whom the duty is owed. It is therefore a condition of adequacy on (IT) that its generalization of the notion of a claim-right suitably limit the number of people it classifies as correlative claim-holders. For the most part, however, this condition of adequacy has not been met.18 §2. By way of illustration, let us consider a promising defence of the Interest theory recently offered by Matthew Kramer.19 To his credit, Kramer confronts the central problem head on, explicitly acknowledging that we have to distinguish the relevant beneficiary from other people whose well-being may be advanced by the execution of the contract. ... [Must the Interest theory] ascribe a right to anyone who might benefit from the carrying out of the contract? If the answer here were ‘yes,’ then the Interest Theory would merit no further consideration as a serious theory of rights (DR, pp. 80-1).

The theory Kramer defends is somewhat different from (IT). In particular, it offers different sufficient conditions for holding a claim-right. Kramer adapts his preferred sufficient conditions from Bentham’s test for the assignment of rights under a law, as glossed by Hart (“LR,” pp. 177-9). According to Hart, Bentham’s test identifies holders of a claim-right correlative to a given duty by asking what findings are necessary to establish a breach of that duty by the duty-bearer. In particular, it asks whether detriment to the candidate right-holder is necessary to establish a breach. Kramer adapts the test by substituting ‘sufficient’ for ‘necessary’ (DR, pp. 81-2). 18 MacCormick replies that the third party objection can also be re-modelled to tell against the Will theory, and thus proves too much (“RL,” pp. 208-9). As it happens, his re-modelling is not effective. But even if it were, it would still not show that (IT) had itself satisfied the condition of adequacy. 19 M. Kramer, “Rights Without Trimmings,” in DR, pp. 66-8 and 79-84.

272 GOPAL SREENIVASAN

Thus, on Kramer’s test, if detriment to X is sufficient to establish a breach by the duty-bearer, then X holds a correlative claim-right and otherwise not (DR, p. 81). In terms of our example, we are to ask what findings are sufficient to establish that you have breached your duty to pay your sister $100. Since proof that your sister suffered the detriment of ‘not having been paid $100 by you’ suffices to establish that you breached this duty, it follows on Kramer’s test that she holds a claim-right correlative to your duty. By contrast, proof that her child suffered the detriment of ‘not having been given a $100 present’ does not suffice to establish a breach of your duty. Hence, the child does not hold a correlative claim-right. So Kramer’s test certainly rules some beneficiaries out as claim-right holders; and may even seem to draw the line in the right place. Appearances, however, can be deceiving. To begin with, we should ask how, on Kramer’s test, your brother—the promisee—qualifies as a claim-right holder. Presumably, it is because of his interest in seeing his wishes fulfilled (cf. DR, pp. 79-80). The idea is that proof of your brother’s detriment of not seeing his sister get her $100 will suffice to establish a breach of your duty. But it is not clear that this will do the trick. Imagine that your grandmother has an interest in seeing her grandchildren behave themselves, get benefits, and so on. Will proof of the detriment to her—of not seeing her granddaughter get her $100—suffice to establish a breach of your duty? If so, your grandmother will also hold a correlative claim-right. An adequate test of claim-right holding should draw the line between your grandmother and your brother. Yet it is not clear how Kramer’s test can exclude the former without also excluding the latter. I can make out three options, none of them satisfactory. First, your grandmother may be excluded because she is not a party to the promise. But this fails to distinguish her from your sister, who is said to hold a correlative claim-right. Second, your grandmother may be excluded because her interest is parasitic—it smuggles in reference to your sister’s detriment. But this fails to distinguish her from your brother, who would otherwise fail to hold a correlative claim-right.20 20 Changes in the description of your brother’s interest can be mirrored by changes in your grandmother’s interest. In principle, the description of his interest should not explicitly refer to the breach or fulfilment of your promise, since this

A HYBRID THEORY OF CLAIM-RIGHTS 273

Third, your grandmother may be excluded because her interest is not important enough. But the one clear way of interpreting this option is not available to Kramer. One might require the detriment to be so important that proof of it is necessary to establish a breach of the relevant duty. However, this would be to adopt precisely the structure of Hart’s gloss on Bentham’s test, which Kramer explicitly rejects. Furthermore, if we examine the notion of what ‘suffices to establish’ a breach a little more closely, a different sort of trouble soon emerges. Consider the special case where your brother waives your duty to pay your sister.21 In this case, your sister’s detriment is not ‘sufficient to establish’ a breach of your duty. Having once seen this, we should then recognize that her detriment does not suffice even when your brother does not waive your duty, since he might have done.22 In fact, even your brother’s parasitic detriment does not really suffice to establish your breach, since detriment on his part does not, strictly speaking, entail that he did not waive your duty. Kramer’s test therefore fails to vest the one uncontroversial claim-right holder—the promisee—with a claim-right against you.23 III One version of the Interest theory is plausibly regarded as exempt from the third party beneficiary objection. I think Joseph Raz’s version may be seen as having solved the problem, which is somewhat ironic, since as far as I know he does not discuss it.24 Still, let me briefly adapt his definition of rights to this end (MF, p. 166): would make Kramer’s test vacuous (cf. note 23). For that matter, however, your grandmother may also have an interest in promises to her grandchildren being kept. 21 This case is also discussed by Hillel Steiner, who makes somewhat different use of it. H. Steiner, “Working Rights,” in DR, pp. 285-6. 22 Kramer later mentions, but does not resolve, a related issue about when a given proof is sufficient (DR, pp. 90-91). 23 Kramer does sometimes slip in the qualification ‘unexcused detriment’ (e.g., DR, pp. 82-3), which might be exploited to cover cases where the duty is waived. But this makes his test vacuous. Compare the equivalent notion of a ‘detriment in breach,’ which explicitly drains the test of content. 24 J. Raz, The Morality of Freedom (Oxford: Clarendon Press, 1986) [hereafter, MF].

274 GOPAL SREENIVASAN

(RZ) Y has a claim-right against X that X f just in case, other things being equal, an aspect of Y’s well-being (his interest) is a sufficient reason for holding X under a duty to f.

The threshold requirement that Y’s interest must itself suffice, other things equal, to justify X’s duty may be regarded as formidable enough to set a suitable limit on the number holding a claim-right correlative to X’s duty to f. Is your nephew’s or niece’s interest in the $100, for example, itself sufficient to ground a duty on your part, other things equal, to pay your sister? Probably not. So far, so good. However, Raz’s account faces another problem, as he himself concedes.25 I believe Raz’s solution to this other problem is objectionable. Moreover, it is arguable that he is forced into the objectionable aspect of that solution by the structure of his solution to the third party beneficiary problem. Thus, while (RZ) avoids the objection against (IT), the price it ultimately pays for its solution is unacceptably high. The problem Raz admits he faces is that of explaining the apparent mismatch between the weight of many rights and the weight of the right-holder’s corresponding interest. The weight of the claim-right—that is, the weight of its correlative duties—is often, so it seems, much greater than the weight of the right-holder’s interest. Consider, for example, a journalist’s claim-right to withhold the names of her sources (cf. MF, pp. 247-8). We may suppose that the duties that correlate with this claim-right—duties that protect the freedom of the press—have great weight. By contrast, the interest an individual journalist has in protecting her sources is often, if not always, comparatively slight. But it is unclear how this can be, given that on (RZ) the journalist’s interest is meant to be sufficient to justify the correlative duty (not to require journalists to disclose their sources). Raz’s solution may be described, in a nutshell, as a piggy-backing solution: He allows that, sometimes, the importance of an individual’s interest can, for the purposes of assessing its contribution to the justi25 See J. Raz, “Rights and Individual Well-being,” in his Ethics in the Public Domain (Oxford: Clarendon Press, 1994), p. 45. (Cf. MF, p. 187). My discussion in the remainder of this section refers to this article.

A HYBRID THEORY OF CLAIM-RIGHTS 275

fication of someone else’s duty, be augmented by taking the interests of third parties into account. Specifically, it can be thus augmented just in case the interests of the third parties are served precisely by serving the relevant interest of the individual in question. That is, the importance of the individual’s interest can be augmented just in case the third parties’ interests can piggy-back on it; and then its importance is augmented by crediting it with the weight of the piggy on its back. Applied to the case of the journalist, Raz’s solution is to allow the general public’s interest in a free press—including its interest in living in the kind of society made possible by freedom of the press—to weigh in favour of the journalist’s claim-right; and to do so just because the public’s interest is served precisely by securing the journalist’s own interest in protecting her sources. More generally, Raz argues, the great weight of many fundamental civil and political rights is to be explained by the fact that the distinctive common goods of a liberal culture are riding piggy-back on the individual interests that correspond to these rights. My objection to this solution is that it instrumentalizes the individual’s status as right-holder.26 By using the individual to enable others to grace their cause with the banner of right-holding, Raz’s solution fails to take the status of right-holder seriously enough. Assignments of this status—that is, the vesting of an individual with a given claim-right—should reflect nothing apart from the intrinsic standing of the individual who is to possess it.27 Consequently, if claim-rights are vested on the basis of the weight of an individual’s interest, as they are on (RZ), then the individual’s interest has, for these purposes, to be weighed simply on its own. In the case of the journalist, that is to say: if her claim-right to protect her sources is to prevail in the social calculus, and prevail on account of the journalist’s status as a rightholder, that must be because the journalist’s own interest has sufficient Compare Simmonds, DR, pp. 195-200; and Thomson, p. 152. At a minimum, this requirement is a desideratum for a theory of claim-rights, one that derives from the aim of preserving the connection between the language of rights and liberal individualism. In its weakest version, my argument against (RZ) is that it fails this desideratum, whereas (as we shall see) my hybrid alternative satisfies it. I actually believe, more strongly, that the requirement stated in the text is a condition of adequacy on a theory of claim-rights. But I shall not argue for this here. 26 27

276 GOPAL SREENIVASAN

weight to defeat the interests others have in learning the identity of her sources.28 If the journalist’s interest lacks this weight, then either she has to reveal her sources or, more plausibly, freedom of the press will have to be regarded as (at least, largely) a matter of net social utility, rather than as a matter of individual rights. Notice that insisting that the status of right-holder reflect only the intrinsic standing of the relevant individual leaves it entirely open what significance that standing actually has. In particular, it remains open whether an individual’s own interest (always) has sufficient weight to prevail against the onslaught of the social calculus. Perhaps no individual’s interest can be worth that much. I do not know—that is another matter. But if it is not worth that much, we should not pretend that individual claim-rights have as much weight as we ordinarily suppose they do.29 Moreover, a theory of rights should not vindicate that pretence on the underside of a piggy. (RZ) avoids the objection confounding (IT) because of its requirement that Y’s interest suffice, other things equal, to justify X’s duty in order for Y to hold the correlative claim-right. However, given the mismatch between the great weight ordinarily attached to the duties that correlate with certain claim-rights and the limited weight that can be justifiably accorded to any one individual’s interests, an account with (RZ)’s structure faces a dilemma. Either Y’s interest is weighed strictly on its own or it is not. If it is, then X’s duty must itself have 28 The interest of the journalist that must have sufficient weight here is her interest as an individual person (albeit, one who is a journalist). As an anonymous referee has observed, the individual journalist might also be thought to have interests as the occupant of a certain office (i.e., that of journalist), interests that are independent of her interests as a person and that reflect—by definition (of the office), rather than by instrumental alignment—the interests third parties have in a free press. We need not decide whether this alternative analysis provides a better account of the freedom of the press. Even if it does, the account it provides either makes no appeal to the journalist’s status as a right-holder (as distinct from her status as an office-holder) or else it, too, makes assignments of that status reflect something in addition to the journalist’s intrinsic standing as an individual. In the first (more likely) case, the account on offer is not enough like Raz’s account to help him; and, in the second case, too much like it to satisfy our desideratum. 29 Either that or we should reject (RZ)—we should deny, that is, that claim-rights are vested on the basis of the sufficiency of an individual’s interest to justify the correlative duty. But this option is not open to Raz.

A HYBRID THEORY OF CLAIM-RIGHTS 277

limited weight, which contradicts our ordinary assumption in certain important cases. If it is not, then the correlation of X’s duty with Y’s claim-right is due to an instrumental—albeit non-fortuitous—alignment between an individual’s interests and those of certain third parties. Raz appears to embrace the second horn of this dilemma. But neither strikes me as very comfortable. IV It seems to me, then, that we lack a satisfactory solution to the debate between the Will theory and the Interest theory. We therefore lack a satisfactory understanding of the correlation between the right-holder and the duty-bearer that is constitutive of claim-rights. I should like to propose a new understanding. The account I propose is a hybrid of the Will theory and the Interest theory. I shall first present a rather crude hybrid, which is nevertheless adequate to meet the objections faced by (WT) and (IT). Then I shall refine my proposal, and explain how it avoids the objection I made to (RZ). §1. Consider a Simple Hybrid model of claim-rights. (SH) Suppose X has a duty to f. Y has a claim-right against X that X ö just in case: either Y has the power to waive X’s duty to f or Y has no power to waive X’s duty to f, but (that is because) Y’s disability advances Y’s interests on balance.

This model has various advantages over (WT). First, (SH) can handle classical inalienable rights—for example, a claim-right not to be enslaved or a claim-right not to be subject to unsafe working conditions. In these cases, Y’s disability is imposed to secure Y’s own position on balance and is standardly taken to strengthen Y’s claim-right. Second, (SH) can handle various forms of incompetence to waive a duty without having to dissolve the correlative claim-right. Here, again, the disability to waive the relevant duties secures the person’s own position on balance. In both cases, Y qualifies as a claim-right holder under the second disjunct.

278 GOPAL SREENIVASAN

The Simple Hybrid model also has an important advantage over (IT). (SH)’s advantage is that it solves the infamous third party beneficiary problem. Say that B promises A to do something that explicitly favours C and implicitly favours D (whom C will favour, as it happens, if B performs). On (SH), A qualifies as a claim-right holder under the first disjunct, since we may assume that A has the power to waive B’s duty. D does not qualify as a claim-right holder since he fails both disjuncts. Ordinarily, the same holds of C. It is natural to see the merit of the Interest theory as being that it is more general than the Will theory. At the same time, however, it is possible to see its demerit as being precisely that it over-generalizes from the Will theory. The correct theory clearly has to be more general than the Will theory because there are important cases of claim-rights that the Will theory mishandles. One way of approaching the problem of a theory of claim-rights, then, is to look for a way of generalizing from the Will theory’s treatment of the paradigm cases of contractual and property rights that manages not to over do it. The Interest theory fails at this because it awards too many claim-rights. I contend that (SH) succeeds, by generalizing the Will theory but only within clear limits. There is also a generic advantage to (SH), that is, an advantage with respect to both the Will and the Interest theories. (SH) gives us some independent purchase on the question of whether there are individual claim-rights under the criminal law. This question is standardly treated as a matter of bare judgement. Those who think there obviously are claim-rights under the criminal law treat their judgement as a basis for criticizing the Will theory,30 whereas those who think there plainly are not claim-rights under the criminal law, at least not for the most part, treat their judgement as a basis for criticizing the Interest theory.31 On (SH), the answer turns on the reasons for vesting control over criminal law duties in the public prosecution service. If the public prosecutor has this control because that is what best secures the interest of individual members of the public, then criminal law duties do correlate with individual claim-rights. But if the justification for vest30 For example, MacCormick in “RL” and Kramer in “Rights Without Trimmings.” 31 For example, Hart in “LR.”

A HYBRID THEORY OF CLAIM-RIGHTS 279

ing control with a public prosecutor is, for example, to secure consistency in the administration of justice, then criminal law duties do not correlate with individual claim-rights. What if both justifications apply? Then we should consider whether a given justification would itself be sufficient to overturn an opposite verdict from the other. Say, for example, that individuals’ interests in security are best advanced by vesting individuals with the power to waive criminal law duties, but the administration of justice is best advanced by vesting that power in the public prosecutor instead. In this case, which justification would prevail? If the latter would prevail, then even if individuals’ security and the administration of justice are both advanced by vesting control in the prosecution service, the latter justification is what settles the question. According to (SH), therefore, criminal law duties do not correlate with individual claim-rights, since individuals are not disabled from waiving those duties because it advances their interests on balance. Notice that, on this view, the weight of the relevant duties is irrelevant to the question of whether they correlate with claim-rights (at least, it is not a sufficient condition). There is no incongruity, it seems to me, in saying that weighty duties that are not owed to individuals do not correlate with individual claim-rights. §2. Let us now consider a series of pair-wise comparisons between (SH) and various complications on it that serve to identify defects in (SH) upon which we can improve. The variations from (SH) appear in bold face. Variant A. Suppose X is duty-bound to f. Y has a claim-right against X that X f just in case: either Y has the power to waive a duty of X’s to f or Y has no power to waive a duty of X’s to f, but (that is because) Y’s disability advances Y’s interests on balance.

(SH) may suggest or imply either that X can only have one duty with a given content—for example, to f—or that, in order for X to owe Y a duty to f, Y must have the power to bring it about that X has no duty to f at all. By contrast, we ordinarily suppose that someone can owe duties with identical contents—even identical non-indexical

280 GOPAL SREENIVASAN

contents—to different people. For example, I can owe both my fiancée and my mother a duty to make it to the church on time. Variant A makes clear that X’s owing Y a duty to f does not exclude X’s also owing Z a duty to f. Likewise, Y’s waiving X’s duty to f need not bring it about that X has no duty to f. It may well be, for example, that X still owes Z a duty to f. This ‘limit’ on Y’s power to waive a duty of X’s to f does not prevent Y’s power from making it the case that at least one of X’s duties to f is owed to Y; and so does not prevent Y from holding a correlative claim-right. Variant B. Suppose X has a duty to f. Y has a claim-right against X that X f just in case: either Y has some measure of control over X’s duty to f or Y has no control over X’s duty to f, but (that is because) Y’s disability advances Y’s interests on balance.

Recall that Hart distinguishes three levels of control in the ‘full measure’ of control over a duty recognized by the Will theory. (SH) fixes narrowly on the first level of control, whereas variant B ranges more widely. On (SHB), the primary non-contractual beneficiary of a contract—C in my earlier example—may also qualify as a rightholder if he or she has the power to enforce the duty, even if he or she lacks the power to waive the duty itself (cf. Hart, “LR,” p. 187). Variant C. Suppose X has a duty to f. Y has a claim-right against X that X f just in case: either Y has the power to waive X’s duty to f or Z has the power to waive X’s duty to f, but (that is because) so empowering Z advances Y’s interests on balance.

When Y is competent to waive X’s duty, but nevertheless disabled from doing so, (SH) allows Y to qualify as a right-holder under its second disjunct as long as Y’s disability secures her own position on balance. In these cases, Y’s position is secured, on balance, by preventing her from exercising this power to her own detriment. When Y is incompetent, there is no need to prevent her from exercising any

A HYBRID THEORY OF CLAIM-RIGHTS 281

such power. (SH)’s second disjunct is trivially satisfied, and so functions simply to preserve Y’s correlative claim-right. But (SH) thereby appears to assimilate cases of incompetence to those of inalienability, by making no provision for X’s duty ever to be waived. This leaves open the case in which Y is incompetent to waive X’s duty, but where Y’s interests are advanced on balance by vesting some third party, Z, with a power to waive X’s duty. (SHC) makes it explicit that powers of waiver can be exercised in trust and that those on whose behalf they are placed in trust can still qualify as claim-right holders. 32 Variant E. Suppose X has a duty to f. Y has a claim-right against X that X f just in case: Y’s power (or disability) to waive X’s duty to f matches (by design) the outcome (having the power or not) that advances Y’s interests on balance.

This has two advantages over (SH). First, it tidies up the disjunctive formulation. The parenthetical reference to design is meant to preserve (SH)’s sensitivity to the justification underlying Y’s position. Second, the one small change in content effected by the tidying is actually an improvement. On (SH), possession of the power to waive X’s duty to f suffices to qualify Y as a claim-right holder, regardless of whether having that power advances Y’s own interests on balance. By contrast, on (SHE), having the power to waive X’s duty qualifies Y as a claim-right holder only if having it advances Y’s own interests on balance (and that is what justifies Y’s power). Thus, someone who exercises a power of waiver in trust for someone else—for example, either a surrogate decision-maker in clinical care or a parent—will thereby count as a claim-right holder on (SH), but not on (SHE). In preferring (SHE), I assume that the surrogate or parent is vested with a power of waiver in order to advance someone else’s interests on balance— namely, those of the incompetent patient or minor. (SHE) has a further advantage over (WT), which will agree with (SH) here. 32 We might also consider a variant D: the same as (SHC), except that ‘Z’ is interpreted to cover both third parties and the null party, i.e. ‘no one.’ (SHD) would extend the rationale in (SHC) to cover the case in which Y’s interests are advanced on balance when no one at all has the power to waive X’s duty to f.

282 GOPAL SREENIVASAN

Although the change in content introduced by variant E is no doubt an improvement, it also mildly complicates the hybrid theory’s solution to the third party beneficiary problem. I shall explain the point simply in relation to the status of the promisee (i.e., the second party). In discussing (SH), I said we could assume that the promisee has the power to waive the promisor’s duty to fulfill the promise. Since the warrant for this assumption is that it reflects a standard feature of promising as ordinarily understood, I shall continue to regard it as an assumption we are entitled to make. As previously explained, the assumption suffices, on (SH), to qualify the promisee as a claim-right holder. However, it does not suffice on (SHE). On (SHE), as we have just seen, the promisee (like anyone else) qualifies as a claim-right holder only if her power to waive the promisor’s duty advances her own interests on balance (and is justified on that basis). The mild complication, then, is that the adequacy of variant E’s solution to the third party beneficiary problem depends on the justification for empowering the promisee to waive the promisor’s duty; and this, in turn, will depend upon our preferred theory of promising. To avoid having to enter into that subject here, let me re-state the complication as follows: Variant E constrains the theory of promising (or contracts) to empower the promisee to waive the promisor’s duty because the promisee’s own interests are advanced on balance by having that power. I describe it as a mild complication because it seems to me that any plausible theory of promising will plainly satisfy this constraint.33 §3. We can now roll the Simple Hybrid model and its variants (SHA)-(SHE) into an omnibus formulation, or Complex Hybrid model: (CH) Suppose X is duty-bound to f. Y has a claim-right against X that X ö just in case: Y’s measure (and, if Y has a surrogate Z, Z’s measure) of control over a duty of X’s to f matches (by design) the measure of control that advances Y’s interests on balance. 33 Or, more precisely, either the correct theory of promising will satisfy this constraint or it will license us to abandon the ordinary assumption that the promisee has a claim-right against the promisor. Notice that the constraint says nothing about how the promisor’s duty has to be justified. In relation to the central burden of a theory of promising, then, it is no constraint at all.

A HYBRID THEORY OF CLAIM-RIGHTS 283

I maintain that (CH) avoids Raz’s problem of instrumentalizing the status of claim-right holder. To see this, we should first notice that (CH) treats the question of whether a given duty correlates with an individual’s claim-right as strictly independent of the question of what justifies the given duty.34 In particular, the question of correlation—and hence, assignment of the status of right-holder—is independent of what justifies either the existence or the weight of the given duty. Having disconnected these questions, (CH) then confines the tradeoffs affecting whether an individual is empowered to waive a given duty to the sphere of that same person’s interests. The balance of interests that settles this question is nothing but a balance of the individual’s own interests. Thus, the trade-offs that affect an individual’s status as a claim-right holder, as someone to whom a particular duty is owed, do not have to be aligned with the interests of third parties. Nor need they be aligned with any other aspect of the larger trade-offs inherent in the social calculus. The perils of instrumentalization are thereby avoided. Of course, the nature of a particular duty’s justification will inevitably constrain whether a given individual can justifiably be empowered to waive the duty in question. Thus, Y will be justifiably empowered to waive X’s duty to f only if two further conditions are both satisfied: (i) the justification of X’s duty to f is consistent with X’s not f-ing; and (ii) Y is a good judge of whether condition (i) holds in any given case.

(If we want Y to be uniquely empowered to waive X’s duty, then we have to add: (iii) Y is the only person who is a good [enough] judge of whether condition (i) holds in any given case.)

Now there are various cases in which these further conditions may be satisfied. The most familiar case is where the decisive consideration in 34 Indeed, all the versions of my hybrid begin with some such formulation as ‘Suppose X is duty-bound to f.’ This is meant to signal that the justification for the duty to f is taken as given.

284 GOPAL SREENIVASAN

the justification of X’s duty to f simply is what advances Y’s interests on balance. Given the traditional liberal dogma that an individual is the best judge of what advances his or her own interests, Y will easily satisfy condition (ii). Indeed, Y will also satisfy the additional condition (iii), and hence be uniquely empowered to waive X’s duty. A second case in which conditions (i) and (ii) may be satisfied is where third party interests play a significant role in the justification of X’s duty to f (for example, in explaining its great weight), but where these third party interests are served precisely by advancing Y’s interests on balance. We might call this the ‘constrained piggy-back scenario’ because it describes a sub-set of Raz’s piggy-back cases— namely, those which also satisfy constraints (i) and (ii). Strictly speaking, this is simply a subtle instance of the previous case.35 It therefore avoids concerns about instrumentalizing the individual’s status as right-holder, since the decisive balancing of interests for vesting Y with control over X’s duty remains a balancing of Y’s own interests. Notice that, here, the weight of X’s duty to f depends on the interests of third parties, but Y’s status as a claim-right holder according to (CH) does not. Y’s status is independent of any alignment with the interests of third parties. Finally, there may be cases where the justification of X’s duty to f has nothing to do with Y’s interests, but where conditions (i) and (ii) nevertheless hold on account of Y’s excellence in judgement. Y may be a superb judge of social utility, for example. These cases may well raise concerns about instrumentalizing Y in some fashion. However, we need not worry about them, since they are not cases in which (CH) vests Y with a correlative claim-right. It may be useful to consider an example of the constrained piggy-back scenario. For technical reasons,36 the journalist’s case dis35 In this case, the decisive considerations justifying X’s duty are the third party interests. Ex hypothesi, however, these decisive interests are served precisely by serving the balance of Y’s own interests. That makes the balance of Y’s interests decisive for justifying X’s duty, which is why this is actually an instance of the previous case. See also the discussion below of the artist’s right of integrity. 36 The journalist’s right to withhold the names of her sources is actually a cluster-right: a combination, that is, of a liberty-right to withhold (or to disclose) these names and of a claim-right not to be required to disclose them. This is a familiar

A HYBRID THEORY OF CLAIM-RIGHTS 285

cussed earlier is not really suitable to examine in this context. So let us consider a structurally similar case instead—say, the claim-right of artists to the integrity of their work.37 The right of integrity correlates with a duty not to distort, dismember, or misrepresent a work of art. It was violated, for example, by the owner of a refrigerator, bought at a Paris charity auction, on six panels of which Bernard Buffet had painted a composition. Although Buffet considered the six panels to be a single work (and had signed only one of them), the owner offered one of the panels for sale at another auction half a year later. (Buffet sued to prevent the separate sale of the panel and won). There are various interests served by the right of integrity. Clearly, these include the interests of individual artists, such as their interests in communication and in reputation. But the right of integrity also serves significant third party interests. In addition to audience interests, these third party interests importantly include collective social interests—for example, the interest in preserving a culture’s artistic heritage.38 Together, the common good of an artistic heritage and the audience interests ride piggy-back on the individual artist’s interest, since they are served precisely by protecting that individual’s (e.g., Buffet’s) own interest in communication and reputation. 39 arrangement, in which the claim-right serves to protect the liberty-right. However, it makes the relation between a power to waive the correlative duty (not to require disclosure) and the journalist’s primary interest here (i.e., the liberty to withhold or disclose the relevant names) unnecessarily complicated for our purposes. This is less of an issue for Raz, for whom the specific notion of a claim-right, as distinct from a liberty-right, is not at the forefront of attention. 37 For discussion of this right, see (e.g.) J.H. Merryman, “The Refrigerator of Bernard Buffet,” Hastings Law Journal 27 (1976), pp. 1023-49. The details of the case that follow in the text appear on p. 1023. I owe the example and the reference to Charles Beitz. 38 Compare Merryman, p. 1041: ‘The machinery of the state is available to protect “private” rights in part because there is thought to be some general benefit in doing so. Thus the interests of individual artists and viewers are only a part of the story. Art is an aspect of our present culture and our history; it helps tell us who we are and where we came from. To revise, censor, or improve the work of art is to falsify a piece of the culture.’ 39 To tighten the parallel with the journalist’s claim-right, we can assume, plausibly, that these third party interests have a greater weight in the social calculus than the interests of individual artists.

286 GOPAL SREENIVASAN

Does the right of integrity illustrate the constrained piggy-back scenario, more specifically? It depends, in the first instance, on whether the common good of an artistic heritage or the audience interests are consistent with individual artists’ waiving the duty protecting the integrity of their own works when they see fit. Suppose the third party interests were not consistent with this outcome. In that case, we would have to give up the idea that artists have a claim-right to the integrity of their work. For artists should then be disabled from waiving the relevant duty and their disability would not be designed to advance the balance of their own interests. Rather, it would simply be required by the superior weight of third party interests inconsistent with artists’ having the power to waive this duty. Hence, on (CH), individual artists would have no claim-right. But now suppose, not implausibly, that the common good and the audience interests are (at least, on balance) actually consistent with individual artists’ being empowered to waive the duty not to distort, dismember, or misrepresent their work. From the standpoint of the social calculus, there would no longer be an impediment to artists’ having this power. However, while it may then be justifiable to empower artists to waive this duty, that does not suffice to vest them with a claim-right: On (CH), it still matters whether this power of waiver advances their own interests on balance. For simplicity, let us say that artists have the power to waive the relevant duty and that (is because) it advances their interests on balance.40 Then (CH) will assign them a claim-right to the integrity of their work; and this claim-right will illustrate the constrained piggy-back scenario. The complications here highlight an important point. When the status of claim-right holder is not treated instrumentally, piggy-back riding is a fairly precarious enterprise. To stick with the right of integrity, the third party interests must be aligned with more than an artist’s interest in actually protecting the integrity of her work. They must be 40 In some jurisdictions (e.g., France), the right of integrity is actually partially inalienable. The artist lacks the power to waive the duty not to distort her work, but retains the power to enforce the duty (as Buffet did) or not. This arrangement remains in the spirit of (CH), since it reflects the same concern for the balance of the artist’s interests (cf. Merryman, p. 1044). (We need not enquire here what measure of control over this duty best satisfies that concern.)

A HYBRID THEORY OF CLAIM-RIGHTS 287

aligned, more precisely, with the sub-alignment of the artist’s own balance of interests on having the choice to protect that integrity or not. The artist’s freedom of choice must serve both the balance of her own interests and (the balance of) the common good and audience interests. Without the first alignment, the artist will not qualify as the correlative claim-right holder on (CH); and without the second alignment, it will not be justifiable to empower her to waive a duty of such great weight. On the other hand, if this elaborate alignment of interests does hold, then the duties protecting the artist’s interest can have a greater weight than the artist’s interests themselves justify, and this without instrumentalizing her status as the correlative claim-right holder.41

41 This paper originated in a graduate seminar on rights I gave at Georgetown University during the Spring semester of 2001. Successive versions were delivered at a conference on rights at the Hebrew University in Jerusalem; as a philosophy colloquium at York University in Toronto; and at a conference on contemporary problems in the philosophy of law at the Universidad Nacional Autónoma de México in Mexico City. I am grateful to the audiences on all these occasions for helpful discussion. For comments on previous versions, I should also like to thank Brian Kierland, Robert Myers, Hanoch Sheinman, Martin Stone, Wayne Sumner, Judith Thomson, and Leif Wenar. I am especially grateful to Alon Harel and Horacio Spector, my commentators in Jerusalem and Mexico City respectively.

13

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

Is the Will Theory of Rights Superseded by the Hybrid Theory? Horacio Spector* I In his interesting and thoroughly argued paper, Professor Sreenivasan provides an account of rights that seeks to avoid the main difficulties that affect the Will Theory and the Interest Theory. Drawing on Hohfeld’s classical analysis, Professor Sreenivasan identifies rights with claim-rights. He says “the best objection that each theory wields against the other is unanswerable” (p. 264).1 Gopal formulates the Will Theory in this way: (WT) Suppose X has a duty to F. Y has a claim-right against X that X F just in case: Y has some measure of control over X’s duty.

Citing Hart, he takes the full measure of control over X’s duty to comprise the power to waive X’s duty, to enforce it or not, given that X has breached it, and the power to waive X’s duty to compensate. As Gosal recognizes, the paradigms of the Will Theory are the claimrights recognized in property and contract law, where right holders have typically the full measure of control. Gopal argues that the Will theory is too restrictive because it cannot accommodate inalienable rights and the rights of incompetent adults. Thus he says: “The Will Theory makes inalienable claim-rights incoherent in principle” (p. 267).While the holders of inalienable rights *

Universidad Torcuato Di Tella, Argentina. Page numbers refer to Gopal Sreenivasan’s. A Hybrid Theory of Claim-Rigths, this volume. 1

290 HORACIO SPECTOR

lack the power to waive the correlative duties, incompetent rights holders cannot exercise the powers of control on which the Will Theory relies. This objection is not compelling, however. As Gopal himself concedes, “lesser measures of control” can be accommodated as approximations to the case of full control. In fact, though holders of inalienable rights do not have the full measure of control over the correlative duties, they typically retain the power to enforce correlative duties and the power to waive compensation duties. Since the inalienability objection is not compelling, Gopal must complement it with what I will call the strength difficulty. According to the Will Theory, inalienability must weaken rights, because inalienable rights do not include the power to waive correlative duties. Yet the disability associated with inalienable rights — says Gopal — is typically seen as strengthening those rights. Replying Simmonds, Gopal sets out the strength difficulty as follows: …it cannot seriously be maintained that reducing the right-holder’s measure of control is consistent with strenghening her claim-right, at least not on the Will theory’s conception of a claim-right. Someone with only a residual measure of control over my duty to Æ lacks the ability to exert her will in certain ways — notably, to make it the case that my failure to Æ does not count as a breach of my duty. How can this not weaken her ability to exert her will, and so not weaken her claim-right on (WT). (p. 268).

Gopal formulates the Interest Theory in these terms: (IT) Suppose X has a duty to F. Y has a claim-right against X that X F just in case: Y stands in a sanctioned relation to benefiting from X’s F-ing.

The Interest Theory covers easily the cases of inalienability and incompetence, but is subject to the well-known third-party beneficiary objection. As is well known, Raz’s formulation of the Interest Theory is meant to overcome this objection: (RZ) Y has a claim-right against X that X F just in case:

IS THE WILL THEORY OF RIGHTS SUPERSEDED 291

other things being equal, an aspect of Y’s well-being (his interest) is a sufficient reason for holding X under a duty to F.

As Raz himself acknowledges, this formulation must confront a new difficulty, which I will call the weight difficulty. Gopal expounds it in these terms: The weight of the claim-right — that is, the weight of its correlative duties — is often, so it seems, much greater than the weight of the rightholder’s interest. But it is unclear how this can be, given that the right-holder’s interest is meant to be sufficient to justify the correlative duty”. (p. 274)

To deal with the weight difficulty, Raz proposes a “piggy-back solution”: the weight of an individual’s interest may be sometimes augmented by taking the interest of third parties into account. An example is a journalist’s claim-right to keep her sources secret. Against this solution, Gopal says: My objection […] is that it instrumentalizes the individual’s status as right-holder. If fails to take his or her status as a right holder seriously enough. If an individual’s claim-right is to prevail against the onslaught of the social calculus, it must do so of its own accord – or, at least, substantially of its own accord. (p. 275)

Gopal proposes a Simple Hybrid Model that is supposedly resistant to the objections against the Will Theory and the Interest Theory: (SH) Suppose X has a duty to F. Y has a claim-right against X that X F just in case: either Y has the power to waive X’s duty to F or the justification of Y’s disability to waive X’s duty is settled by considering whether vesting Y with a power to waive X’s duty would, on balance, advance Y’s interests.

To solve a number of technical problems – which I will not consider here – Gopal turns the Simple Model into the Complex Model:

292 HORACIO SPECTOR

(CH) Suppose X is duty-bound to F. Y has a claim-right against X that X F just in case: the question of whether Y (or his surrogate Z) is vested with some measure of control over a duty of X’s to F (and if so, of which one is and with what measure ) is settled by the consideration of what would, on balance, advance Y’s interests.

II I take Gopal’s paper as seeking to provide an analysis of both moral and legal rights. On the one hand, he adopts the Hohfeldian notion of a claim-right, applicable centrally to legal rights, and borrows many examples of claim-rights from the law. On the other hand, he adopts the well-established usage in moral philosophy of rights as trumps or constraints on the utilitarian calculus. In the absence of this usage, it would be difficult to make sense of the instrumentalization problem. Therefore, Gopal’s project should be assessed in terms of its ability to account for both moral and legal rights. It is not clear to me that any conceptual analysis of rights can meet the above standard of success at the present stage of legal evolution. In the seventeenth and eighteenth centuries, the situation was probably different. At those times moral and legal rights were essentially associated with the value of personal autonomy. According to the Kantian doctrine of right, for instance, legal rights were public and institutional ways of recognizing the status of persons as ends-in-themselves. Rights were intelligible concepts against the background of a fundamentally non-consequentialist moral outlook. In turn, legal rights recognized individual autonomy by vesting in individuals the powers that Will Theory picks out. It is no surprise that Hart mentions rights in property and contract law as the central examples of claim-rights. Those legal rights constituted the basic legal machinery of the Kantian conception of law. The Will Theory is an incomplete and fragmentary way of articulating the classical conception of rights. In fact, the theory focuses on the powers of right holders, but ignores the underlying autonomy-based justification. Gopal cites Hart’s suggestion that the justification of claim-rights associated with the Will Theory is “the interest in autono-

IS THE WILL THEORY OF RIGHTS SUPERSEDED 293

mous choice” (p. 269). On the Kantian view, however, rights cannot be grounded on any interest – not even an interest in autonomous choice – because that would amount to disregarding the value of autonomy. Rights should be based on the status of individuals as autonomous agents. A formulation of the Will Theory faithful to the moral and legal tradition from which it takes its cue should refer both to the powers present in classical legal rights and to its associated justification on basis of the value of individual autonomy. Toward the end of the nineteenth century, the harmonious picture of legal rights and underlying autonomy-based moral rights broke down. In a well-known process that spanned over decades, the individualistic paradigm of legal rights was gradually replaced by a paradigm centered on interests and social goals. Legal rights were no longer a way of recognizing the status of autonomous agents but instruments for enhancing the interests of individuals – particularly the worst-off – and for implementing social goals. Jhering’s famous theory of rights as legally protected interests exemplifies this new approach to the law. The Interest Theory is, of course, an attempt to account for this new conception of legal rights, which disconnects them from the value of individual autonomy. Though the Interest Theory is sensitive to the greater variety of legal rights in modern legal systems, like the right to a minimum wage or the right to education, it is at odds with those features of moral and legal rights that answer to the classical paradigm. These features are emphasized in the contemporary notion of rights as trumps or constraints. Given the historical linkage between rights and non-consequentialism, it is natural that the Interest Theory cannot provide a satisfactory account of rights. The instrumentalization problem is just an instance of this general truth. Even if we leave aside Raz’s “piggy-back solution”, the Interest Theory cannot accommodate the anti-consequetialist overtones of the concept of rights. As Eric Mack has recently shown, the Interest Theory is in tension with two central features of rights-based theories: the impermissibility to trading off rights against aggregative goals and the principled rejection of paternalism.2 The Interest Theory – particularly in 2 Erick Mack, “In Defense of the Jurisdiction Theory of Rights”, The Journal of Ethics 4 (2000), Nos. 1-2; reprinted in G. Pincione and H. Spector (eds.), Rights, Equality, and Liberty, Dordrecht, Kluwer, 2000.

294 HORACIO SPECTOR

Raz’s version – is incompatible with those features because it understands the value of rights as based on the interests of rights holders. These interests can have either agent-neutral (impersonal) or agent-relative value. If they have agent-neutral value, rights fall prey to the utilitarian calculus. In effect, given that the normative force of rights is grounded on the right holders’ interests, there is always the possibility that those interests be outweighed by other people’s interests. Alternatively, if the rights holders’ interests only have agent-relative value (i.e., they only provide reasons for the right holder), the theory cannot explain how those benefits or interests can justify holding someone else under a duty. In both variants (agent-neutral or agent-relative value), rights could not block (in a principled way) interferences intended to advance more successfully the right holder’s interests. For instance, if the right to reject medical treatment were based on the patients’ interests, it might be possible to justify the imposition of a blood transfusion on a Jehovah’s Witness. I believe that an illuminating theory of rights should take into consideration the changing paradigms in which the notion of rights is embedded. It seems obvious that while the Will Theory tries to capture the classical features of rights, the Interest Theory seeks to accommodate the widely different kinds of rights that modern legal systems recognize and their different underlying justifications. Therefore, it may be impossible to provide an analysis of rights that does full justice to the various and changing ideas with which the concept of rights was associated at different points of its historical evolution. A partial success may be the most we can aspire to. Conceptual analysis is very often criticized by its lack of historical awareness. Though I am not in general sympathetic to this kind of criticisms, usually linked to an anti-intellectual bent, I do believe that the analysis of rights can enrich itself by paying more attention to the facts that I have outlined. III If we are confronted with the need to opt for one theory, I think that the Will Theory carries the day. Consciously or unconsciously, Hart dropped from the explicit formulation of the theory any reference to

IS THE WILL THEORY OF RIGHTS SUPERSEDED 295

its underlying autonomy-based justification. We should celebrate this because, as I suggested, legal rigths are embedded in different normative paradigms. For instance, we cannot assume that the justification of modern welfare rights matches the justification of classical rights in private law. Unlike the Interest Theory, the Will Theory is neutral with respect to justificatory matters and, therefore, is in a better position to account for different kinds of legal rights. So what are the obstacles that stand in the way of embracing the Will Theory? As I said above, Gopal tries to rebut the Will Theory by means of the strength difficulty. He claims that typical inalienable rights, like the right not to be enslaved, are stronger than alienable rights because of the disability such rights incorporate. This conclusion seems to contradict the Will Theory. Since this theory equates rights with powers, a right comprising fewer powers — so the argument goes — must be weaker than a right including all the relevant powers. In my opinion, the argument from the strength difficulty relies on a controversial assumption. It is true that inalienable rights are typically stronger than alienable rights. However, the explanation of this fact need not lie in the disability, as the argument assumes. Classical inalienable rights, like the right to freedom, are “stronger” than alienable rights not because of the disability, but because of the weight of the duties that correlate with them and the importance of the value that such duties secure. Even if the holder of an inalienable right has lesser power to control the correlative duty than the holder of an alienable right, the greater weight of the former right might simply mirror the greater weight of the correlative duty. This is clear in the case of the right not be enslaved. In effect, enslaving a person destroys her status as an autonomous agent, regardless of whether she voluntarily consented to become a slave. On the classical Kantian view of rights, the inalienability of the right to freedom is in accord with the autonomy of the will. Thus, the second formulation of the Categorical Imperative forbids us to treat not only the humanity in others but also the humanity in ourselves as only a means. Consenting to become a slave is a direct violation of this prohibition. Classical inalienable rights are stronger than ordinary rights not because but despite of the disability they include.

296 HORACIO SPECTOR

As soon as we dismiss the strength difficulty, inalienability poses no serious problem to the Will Theory. It can handle inalienability because holders of inalienable rights have residual control of the correlative duties. We can reinforce this conclusion by noticing that Hart’s enumeration of the powers associated with rights leaves aside an essential ingredient illuminated by Feinberg. When we say that Y has a claim-right against X that X F we imply that Y is in a position to make a valid claim that X should F and to make a complaint if his claim is not satisfied. IV Even if my treatment of the strength difficulty were unsuccessful, Gopal should still prove that the Hybrid models can handle this difficulty more successfully than the Will Theory. This is far from obvious. Gopal says that the disability of inalienable rights is typically seen as strengthening those rights. The hybrid models fail to explain this fact (if it is fact). Why should rights become stronger when the second disjunct of the Simple Hybrid model is true (and weaker when the first disjunct applies)? I can see no reason for this asymmetry. Nor does the Complex Hybrid model explains why inalienable rights are stronger than alienable rights. (Notice that Gopal could not explain the difference in terms of the importance of the interests or values that the rights protect. This would bring us back to my analysis of strength as weight, which solves the strength difficulty.) To show that the Hybrid models are superior to the Will Theory, Gopal should also prove that such models are immune to the difficulty that threatens the Interest Theory (and to which the Will Theory is not subject). Thus, he argues that the Hybrid models avoid disregarding the status of individuals as rights holders (i.e., the problem of instrumentalization). This is not clear to me. On the one hand, according to the Complex Model, the question of whether the right holder is vested with a measure of control of the correlative duty must be settled by considering the balance of his interests. Therefore, the attribution of rights under that model is compatible with illiberal paternalistic interferences. For instance, if the State of Sonora passed a law

IS THE WILL THEORY OF RIGHTS SUPERSEDED 297

abrogating the powers of citizens to protect their liberty to smoke, it could well claim that the law respects citizens’ rights because it is justified on the balance of each citizen’s interests. Gopal could reply that the duty to respect citizens’ liberty to smoke is justified on independent grounds. But the paradox remains that the Complex Hybrid model does not allow citizens to invoke the most natural defense they have in this case, namely, that the law violates their right to smoke. On the other hand, individual autonomy is the most natural justification of the legal powers on which the Will Interest focuses. The Complex Hybrid model excludes a Kantian justification of those powers and, therefore, fails to do justice to their traditional anchorage in the value of individual autonomy. Accordingly, the instrumentalization problem could reappear under a different shape. The balance of interests that justifies powers might be traded off against the interests of other people, or of society at large. It is true that the model restricts trades offs to the right holder’s balance of interests, but this restriction seems arbitrary. The Complex Hybrid model suggests that the attribution of legal powers to individuals must be grounded on their interests. Once this is admitted, it is difficult to stop the sliding into the utilitarian calculus. For example, the arrest of certain individuals (e.g., Arabs) could be justified if it is needed to preserve public tranquility. (Notice that constitutional guarantees are best conceived as Hohfeldian powers.) V Gopal also suggests that admitting claim-rights under the criminal law runs afoul of the Will Theory. I disagree. In fact, right holders do have the full measure of control under the criminal law because they can waive criminal prohibitions, just as they can waive private law prohibitions. For instance, if Alice waives her ownership right over her piano (and no one else claims it), Martin’s taking control of it cannot constitute robbery. What rights holders cannot typically do under the criminal law is to cancel the offender’s liability to punishment, because criminal prosecution is a public matter. In contrast, rights holders can waive the obligation to compensate under private law. But this difference does not affect the Will Theory, for it does not pick out the power to cancel criminal liability once the offence has been performed.

298 HORACIO SPECTOR

True, in some cases, like murder and assault, rights holders cannot even waive criminal prohibitions. This is probably the thrust of the claim that there are no claim rights under the criminal law. However, the inalienability of those rights is not necessarily connected to their criminal law protection. Individuals lack the power to waive their rights against murder and assault even under private law. These are unlawful acts even though the victims (or their heirs) have the power to waive the duty to compensate. I concede, however, that the holders of inalienable rights have narrower residual powers under criminal law. This is not a serious difficulty for the Will Theory. The Hybrid models also imply that there are fewer rights in the criminal law, because the standard argument for denying victims the power to cancel the offender’s liability to punishment relies on the public interest, rather than on the victim’s balance of interests.

14

SOCIAL, POLITICAL, & LEGAL PHILOSOPHY, 2 Law: Metaphysics, Meaning, and Objectivity, 2007

In Defence of the Hybrid Theory Gopal Sreenivasan There is a lot with which I agree, and still more that I admire, in Horacio Spector’s sophisticated and insightful defence of the Will theory against my advocacy of a hybrid alternative.1 Implicitly or explicitly, Spector appeals to four conditions of adequacy on a theory of claim-rights in the course of his discussion. I accept two of these, as well as a weaker version of a third. But, as we shall see, this is not enough either to ground an objection to my Hybrid theory or to vindicate the Will theory. Since it will serve as a useful frame of reference for later points in the argument, I shall begin by reviewing the nature and merits of Spector’s conditions of adequacy. Then I shall consider Spector’s treatment of what he calls the ‘strength difficulty,’ a problem for the Will theory he attempts to dismiss. Finally, I shall turn to what Spector calls the ‘instrumentalization problem’ and dispel the doubts he wishes to cast upon the Hybrid theory’s ability to improve on the Interest theory. §1 Spector’s four conditions of adequacy on a theory of claim-rights may be summarized as follows: 1. The theory must account for both moral and legal rights. 2. To account for moral [claim-]rights, the theory must account for them as trumps or constraints on the consequentialist calculus. 1 H. Spector, “Is the Will Theory of Rights Superseded by the Hybrid Theory?” this volume. References to Spector’s commentary are given as bare page numbers in the text.

300 GOPAL SREENIVASAN

3. The theory must be sensitive to the history of rights. 4. The theory must build a principled rejection of paternalism into the attribution of rights. 2 Of these conditions, I accept (1) and (3), but reject (2) and (4). The challenge in (3) arises from the fact that, as Spector notes, developments in different historical periods correspond respectively to the Will theory of rights and to the Interest theory of rights. Fidelity to history, then, requires the correct theory of rights somehow to embrace both the Will theory and the Interest theory. So it is no wonder that he cautions that ‘it may be impossible to provide an analysis of rights that does full justice to ... its historical evolution’ (p. 294). While I used to despair of the possibility myself, I now believe that the two theories can be reconciled. Of course, any reconciliation will have to be somewhat revisionary, since the historical evolution of rights has arguably overshot the mark. Legal and social discourse, that is to say, now contains such a proliferation of rights claims that the language of rights, as many have lamented, is in serious jeopardy of losing its distinctive content. As I understand it, the question is therefore whether a principled intermediate point can be found in this historical evolution, in terms of which the notion of a claim-right may be rehabilitated. My proposal is that the Hybrid theory represents this point. There are various reasons to reject (2). I shall mention a couple of them, but only rely upon one here. To begin with, it is worth pointing out that (1) and (2) are actually inconsistent. Like Spector, I assume that legal claim-rights are Hohfeldian in structure—they are defined, that is, in terms of their correlation with a duty that is owed to the claim-right holder. Presumably, therefore, the moral claim-rights to which (1) refers have the same structure. However, to function as a constraint on the consequentialist calculus, as (2) requires, a claim-right has, as I argue elsewhere, to be defined in terms of something rather different, namely, the all-things-considered moral ought.3 Unfortunately, there is no suitable logical bridge between an Hohfeldian claim-right’s correlative duty 2 This is one of two conditions of adequacy Spector adopts from Eric Mack. I ignore the other only because it is effectively a re-statement of (2). See E. Mack, “In defense of the jurisdiction theory of rights,” Journal of Ethics 4 (2000), pp. 71-98. 3 See my “Duties and their direction” (in preparation), §2.

IN DEFENCE OF THE HYBRID THEORY 301

and the all-things-considered ought.4 So the moral claim-rights to which (2) refers must not be Hohfeldian in structure.5 Hence, there is no univocal sense of ‘moral claim-right’ on which both (1) and (2) are true. A simpler reason to reject (2) will suffice, as it removes Spector’s ground for imputing the condition to me. (2) is not, in fact, necessary ‘to make sense of the instrumentalization problem’ (p. 292). Spector refers here to my criticism of Raz’s version of the Interest theory, and of his piggy-backing solution in particular. At bottom, however, my criticism of Raz turns on the premiss that assignments of a claim-right should ‘reflect nothing apart from the intrinsic standing of the individual who is to possess it.’ While I take it that Spector accepts this premiss, the crucial point, for present purposes, is that the premiss is considerably weaker than the anti-consequentialism of (2). It is one thing to restrict the basis of a claim-right to an individual’s intrinsic standing, and quite another to say how much that standing is worth. As I emphasized in the paper, my criticism of Raz leaves it open whether an individual’s intrinsic standing ‘(always) has sufficient weight to prevail against the onslaught of the social calculus.’ It thereby leaves open whether consequentialism is true. In different ways, Spector’s anti-consequentialist (2) and the premiss of my criticism of Raz both express the historical connection between the language of rights and (liberal) individualism. The version of individualism expressed by anti-consequentialism is much stronger, as I have said, than the version on which I rely to criticize Raz. In the light of (3), I cannot resist adding that, historically, anticonsequentialism is also a rather later version. The connection with individualism is there from the beginning of the long history of (natural) 4 The obvious candidate is the idea that the duty to f that correlates with an Hohfeldian claim-right entails that its bearer ought all-things-considered to f. But this is demonstrably false. See J. Thomson, The Realm of Rights (Cambridge, Mass.: Harvard University Press, 1990), ch. 3. A logically impeccable alternative is that the duty to f that correlates with an Hohfeldian claim-right entails that, other things being equal, its bearer ought all-things-considered to f. But this is not suitable either, since the bridge it provides is too weak to satisfy (2). For elaboration, see the previous note. 5 I agree with Spector that there is a sense of ‘moral claim-right’ on which (2) is true and that it is commonly invoked in moral philosophy. But this is a distinct sense of the term—the deontological sense, rather than the Hohfeldian sense. It is imperative not to confuse the two, not least when trying to satisfy (1).

302 GOPAL SREENIVASAN

rights discourse,6 well before the advent of utilitarianism or even the landmark formulations of Kant. It strikes me as controversial, if not simply anachronistic, to read specifically anti-consequentialist overtones into the natural lawyer’s figure of the sovereign individual, even though an emphasis on individual autonomy later came to acquire them.7 Finally, I also reject (4). In brief, my reason is that it excludes a perfectly eligible rationale for making a claim-right inalienable. Since inalienability is another feature of rights that has been there from the beginning, (4) cannot be a condition of adequacy on a theory of claim-rights. But I shall return to spell this out in the context of a fuller discussion of inalienability below. §2 My criticism of the Will theory in the paper consisted of two standard objections, the inalienability objection and the incompetence objection. The inalienability objection itself has two parts, corresponding to the two kinds of inalienable claim-right, partially inalienable claim-rights and completely inalienable ones. Spector’s efforts to vindicate the Will theory concentrate exclusively on the first part of the inalienability objection. In the case of a partially inalienable claim-right, the right-holder lacks the power to waive the correlative duty, but retains the power to sue for enforcement and to waive compensation. As I freely granted in the paper, the Will theory is entirely capable of explaining how someone who retains a residual measure of control over a duty can be vested with a correlative claim-right. In the partial case, the objection, following MacCormick, is rather that the Will theory cannot explain how the inalienability of the claim-right strengthens the right.8 Spector calls this the ‘strength difficulty.’ 6 For the classic account, see R. Tuck, Natural rights theories: their origin and development (Cambridge: Cambridge University Press, 1979). 7 Of course, I do not mean to suggest that individualism in its original version is pro-consequentialist either—simply, pre-consequentialist. 8 This objection also applies in the case of complete inalienability. But there it supplements the objection that the Will theory cannot explain how someone with no

IN DEFENCE OF THE HYBRID THEORY 303

To rebut this objection, Spector makes three points. First, he distinguishes different senses of ‘strength.’ In addition to the <degree of control over the correlative duty>, there is also the <weight of the correlative duty>. Second, he argues that the greater strength of an inalienable claim-right should be understood in line with the latter sense—in terms, that is, of the greater weight of its correlative duty. The Will theory would have no particular difficulty explaining strength so understood. Third, Spector complains that the Hybrid theory does not handle the strength difficulty any better than the Will theory. Spector’s distinction is well-taken. Moreover, the strategy implicit in his second point is, in principle, a good one. Partisans of the Will theory are free to re-interpret the sense of ‘strength’ at issue, as long as they can both preserve and explain the facts in question. The trouble with Spector’s second point is that it gets the facts wrong: the duties correlative to inalienable claim-rights do not always have greater weight than those correlative to alienable ones. There is no such fact to explain. Spector’s strategy works well enough with the claim-right not to be enslaved. But it does not work with other, more mundane inalienable claim-rights, such as the claim-right not to be employed in unsafe working conditions (or, for that matter, the artist’s claim-right to the integrity of her work, which in certain jurisdictions is partially inalienable). In these examples, which the paper included by design, the correlative duties are not especially weighty; and it is not hard to imagine alienable claim-rights that have weightier correlative duties (e.g., the right to property). It is important to keep the comparison class straight here. Some inalienable claim-rights are not stronger than other alienable claim-rights. But every inalienable claim-right is stronger than it would be without the disability imposed on the right-holder. Of course, Spector is quite right that the Hybrid theory must itself be able to explain this fact. But the task is easily discharged. Inalienable claim-rights are stronger than their alienable counterparts in the sense that they afford a greater degree of protection to the rightholder’s interests on balance. In certain cases, vesting the right-holder with a lesser degree of control (including zero) over the correlative control over a duty can be vested with a correlative claim-right. Whereas in the partial case, it is the only objection on offer.

304 GOPAL SREENIVASAN

duty, i.e. imposing a disability on her, actually results in a greater degree of protection to her interests on balance—notably, cases in which the right-holder is expected to exercise the relevant powers (e.g., to waive the correlative duty) to her own detriment on balance.9 Moreover, these are precisely the cases in which the Hybrid theory makes a claim-right inalienable—in which, that is, it recognizes an individual who is disabled from (fully) controlling a duty as holding a correlative claim-right. This explanation naturally invites the observation that an obvious rationale for making a claim-right inalienable, whether partially or completely, is simply to protect the right-holder. More specifically, it is to protect the balance of the right-holder’s interests from, among other things, certain actions of her own. It is therefore a classically paternalist rationale. Now I do not say that the paternalist rationale is the best, still less the only, rationale for making a claim-right inalienable. But it is a sufficiently obvious and straightforward rationale that its eligibility as a rationale for inalienability is really beyond question.10 Admitting its eligibility, however, rules (4) out as a condition of adequacy on a theory of claim-rights. 11 Before leaving the subject of inalienability, I should perhaps say that, in my own view, the strength difficulty is not the most fundamental part of the inalienability objection. The fundamental objection is the one that only applies in the case of complete inalienability (see note 8). Since Spector does not discuss that case, nor indeed the incompetence objection, his argument would not have vindicated the Will theory, even if it had succeeded in dismissing the strength difficulty. 9 So, in these cases, it is actually correct to say that someone vested with a claimright under (SH)’s second disjunct thereby has her interests protected to a greater degree on balance (i.e., has a stronger claim-right) than she would have if she had the powers described under (SH)’s first disjunct. 10 If the paternalist rationale were conceptually ineligible as a rationale for making a claim-right inalienable, then assertions of inalienable claim-rights should, throughout history, have been accompanied by references to an alternative rationale. Otherwise they risked not making sense. 11 Substantive moral objections to paternalism are beside the point here, which is that the falsity of paternalism is not guaranteed by anything in the concept of a claim-right.

IN DEFENCE OF THE HYBRID THEORY 305

§3 Let me close by returning to ‘the instrumentalization problem,’ i.e. my criticism of Raz’s theory (RZ). The nub of my criticism, recall, is that (RZ) makes the assignment of a claim-right depend, in some cases, upon a contribution from select third party interests. (RZ) therefore allows such assignments to depend upon something other than the intrinsic standing of the (relevant) individual, which is objectionable. I claimed that my Complex Hybrid theory (CH) is immune from that problem. Although (CH) makes the assignment of a claim-right depend upon a balance of interests, the interests in question are all interests of the same individual. The required alignment of interests, and hence the basis of (CH)’s assignment, is therefore an aspect of that individual’s intrinsic standing and not something other than it. To illustrate the crucial difference, contrast the role of third party interests in (CH)’s constrained piggy-back scenario with their role in (RZ)’s piggy-back scenario. Recall the example of the artist’s right to integrity from the paper. Say individual artists have the power to waive the duty not to distort their works of art; and that the common good and audience interests not only support (the artist’s interest in) this duty, but are also consistent with the artists’ power of waiver. In (RZ)’s piggy-back scenario, the role of these third party interests is to qualify individual artists for a claim-right to which their own interests, considered by themselves, would not entitle them. It is to push the individual over the line into right-holder status, a line she would not cross on her own.12 By contrast, in (CH)’s constrained piggy-back scenario, we must first ask whether, on balance, an artist’s own interests favour empowering her to waive the duty in question. If they do, then artists qualify for a claim-right entirely on the basis of intrinsic facts about them as individuals: they qualify, according to (CH), because their actual degree of control over the (correlative) duty and the degree of control favoured by their own interests is the same, i.e. the full measure of control. Here the role of third party interests is 12 Moreover, it pushes the individual over the line in order to grace the third party’s cause with the banner of right-holding. That is where the instrumentalizing aspect enters.

306 GOPAL SREENIVASAN

simply to augment the weight of the duty that correlates with a claim-right to which the individual artist is independently entitled. Against this defence of (CH), Spector raises two principal complaints. First, he charges that (CH)’s restriction of the trade-offs taken into account to trade-offs among the individual’s own interests is ‘arbitrary’ (p. 294). By this, he could mean either that the restriction is unmotivated or that it is unjustified. But neither charge stands up. Far from being unmotivated, (CH)’s restriction is positively required, as we saw earlier, to preserve the connection between the language of rights and (liberal) individualism—to satisfy the weaker version of Spector’s condition of adequacy (2). Nor is the restriction unjustified. To assign a correlative claimright, (CH) requires an individual’s actual degree of control over the duty in question to match the degree of control that advances her own interests on balance. (CH) focuses, that is, on whether she has the degree of control that would be justified overall if her interests were the only relevant interests. Obviously, however, the individual’s interests might not be the only relevant ones. The salient case here is where relevant third party interests are not only weightier than the individual’s own, but favour vesting her with a different degree of control from that favoured by her own interests (e.g., none instead of some).13 In this case, if the individual’s actual degree of control over the duty did match the degree of control that advances her own interests on balance, the outcome would be unjustified overall. Still, there is no objection to (CH) to be found here, since it does not require the unjustified outcome. (CH) says nothing about what an individual’s actual degree of control over a duty should be. A fortiori, (CH) does not say that her actual degree of control should match the degree that advances her own interests on balance. The individual’s actual degree of control over a given duty is what it is. (CH) ‘requires’ her actual degree of control to match the degree that advances her own interests on balance only as a condition of assigning her a correlative claim-right. If this match is not justified overall, then (CH) does not 13 In the paper, this case was illustrated by the initial discussion of the right of integrity, conducted under the assumption that the audience and common good interests are inconsistent with artists’ being empowered to waive the duty not to distort their work.

IN DEFENCE OF THE HYBRID THEORY 307

require the individual to have an unjustified degree of control over the duty. It simply fails to assign her a correlative claim-right (as in the case in note 13). Spector’s second principal complaint is that the ‘balance of interests that justifies powers might be traded off against the interests of other people’ (p. 294). In other words, the possibility remains that the (balance of the) right-holder’s interests may be overwhelmed by social interests. This could mean one of two things. It could mean that the degree of control favoured by the right-holder’s interests is overwhelmed by social interests. In effect, this is the scenario we have just discussed.14 Alternatively, it could mean that the correlative duty protecting the right-holder’s interests is overwhelmed by social interests. On this reading, which is suggested by Spector’s arresting example, his complaint reiterates the charge that (CH) fails his condition of adequacy (2). Now Spector is right that (CH) does not exclude the possibility that the duty correlative to a given claim-right may be over-ridden by competing interests in the social calculus. However, this has nothing in particular to do with my criticism of (RZ). More importantly, it provides no basis for objecting to (CH). Simplifying matters somewhat, consequentialism is either true or false. If it is true, then the possibility of a given (correlative) duty’s being permissibly over-ridden is genuine, in which case (CH)’s failure to exclude it is no defect. If consequentialism is false, then (correlative) duties cannot be permissibly over-ridden, in which case there is no genuine possibility that we still need (CH) to exclude. In neither case does (CH) lead to an objectionable result. It does, of course, leave us to decide the merits of consequentialism on independent grounds, but that is as it should be.

14 Notice, however, that even in this scenario (CH) does not trade the powers favoured by the individual’s interests off against social interests. Rather, (CH)’s matching requirement simply turns out to be impossible to satisfy—impossible, at least, on the assumption that the individual’s actual degree of control over the duty is justified overall. Hence, no claim-right is assigned.