Law and Philosophy [1 ed.]
 0199237158, 9780199237159

Citation preview

LAW AND PHILOSOPHY CURRENT LEGAL ISSUES 2007 VOLUME 10

CURRENT LEGAL PUBLICATIONS Editor-in-Chief Michael DA Freeman Editorial Board Sir John Baker E Barendt IH Dennis D Galligan Dame Hazel Genn Sir Bob Hepple J Holder J Jowell N Lacey Sir Neil MacCormick A Le Sueur ADE Lewis E McKendrick R Mokal RW Rideout P Sands Lord Steyn Lord Woolf of Barnes

Law and Philosophy Current Legal Issues 2007 VOLUME 10 Edited by

M IC H A E L FR E E M A N Professor of English Law University College London

and

ROSS H A R R ISON Former Professor of Philosophy University of Cambridge

1

3

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Th ailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © Oxford University Press 2007 The moral rights of the authors have been asserted Database right Oxford University Press (maker) Crown copyright material is reproduced under Class Licence Number C01P0000148 with the permission of OPSI and the Queen’s Printer for Scotland First published 2007 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Law and philosophy / edited by Michael Freeman and Ross Harrison. p. cm.—(Current legal issues ; v. 10) Includes bibliographical references and index. ISBN 978–0–19–923715–9 (hardback : alk. paper) 1. Law—Philosophy. I. Freeman, Michael D. A. II. Harrison, Ross. K235.L335 2007 340'.1—dc22 2007031589 Typeset by Newgen Imaging Systems (P) Ltd., Chennai, India Printed in Great Britain on acid-free paper by Biddles Ltd., King’s Lynn ISBN 978–0–19–923715–9 1 3 5 7 9 10 8 6 4 2

General Editors’ Preface This collection arises from the inter-disciplinary colloquium on law and philosophy held in University College London in July 2006 the tenth to be held in this series. The papers are revised versions of a selection of the presentations to the conference, as well as two papers by people who were expected but prevented from coming, and a paper given earlier that year at UCL. The collection displays a wide spectrum of the ways in which philosophy can be applied to legal questions, as well as the interconnections between them. The two central themes are the lively and contentious contemporary debate about the nature of law and the always relevant normative debate about what the state should do (that is, what the law ought to be). We would like to thank everyone who gave papers at the conference and helped with its organization, in particular Lisa Penfold and Claire Grant. Lisa Penfold also helped with the subsequent processing of papers and Samantha Watters aided their final assembly. The 2007 colloquium is on ‘Law and Bioethics’ and the 2008 colloquium (to be held on 7 and 8 July 2008) is on ‘Law and Anthropology’. Offers of interest in this should be directed to Michael Freeman at . Ross Harrison and Michael Freeman May 2007

Notes on Contributors Christopher Bennett is a lecturer in the Department of Philosophy at the University of Sheffield. His book on the theory of punishment, The Apology Ritual, will be published by Cambridge University Press in 2008, . Kimberley Brownlee is a lecturer in political theory and philosophy in the Politics Department at the University of Manchester, . Brian Burge-Hendrix is a Research Fellow in law and philosophy at Churchill College, Cambridge. Sylvie Delacroix was a lecturer in law at the University of Kent at Canterbury when she wrote this, but from September 2007 is a lecturer in the Faculty of Laws at University College London. Her Legal Norms and Normativity was published in 2006, . RA Duff is a Professor of Philosophy at the University of Stirling. Claire Grant is a Reader in Social & Political Philosophy at the University of Warwick. At the time of the conference she was a senior lecturer in legal and political philosophy at Birkbeck College London. She is foundation co-editor with RA Duff of the journal Criminal Law & Philosophy. Stephen Guest is Professor of Legal Philosophy in the Faculty of Laws at University College London. He is currently British Academy/Senior Research Fellow working on a project on humanity and law. Ross Harrison organized this conference when he was Quain Professor of Jurisprudence at University College London. Antony Hatzistavrou is a lecturer in the Department of Classics and Philosophy at the University of Cyprus and in 2006–7 came to the conference while he was a visiting fellow at Clare Hall, Cambridge, . Kenneth Einar Himma is in the Department of Philosophy, Seattle Pacific University, . Christoph Kletzer is a lecturer in jurisprudence in the Faculty of Law at Cambridge and Director of Studies in Philosophy at Queens’ College and Christ’s College Cambridge, . George Letsas is a lecturer in the Faculty of Laws at University College London, . SE Marshall is a Professor of Philosophy at the University of Stirling, . Robert Morris wrote his paper on the basis of research conducted at King’s College London in the Philosophy and War Studies Departments, .

x

Notes on Contributors

John Oberdiek is in the Rutgers University School of Law-Camden and also Associate Graduate Faculty in the Rutgers-New Brunswick Department of Philosophy. He is Codirector of the Rutgers Institute for Law and Philosophy, . Dennis Patterson is Distinguished Professor in the Rutgers University School of LawCamden, . Mark R Reiff is a lecturer in legal and political philosophy at the University of Manchester School of Law. He is the author of Punishment, Compensation, and Law: A Theory of Enforceability (Cambridge University Press, 2005), . Tanja Staehler is a lecturer in philosophy at the University of Sussex, . Emmanuel Voyiakis is a lecturer in law at Brunel University, .

1 Reconsidering a Dogma: Conceptual Analysis, the Naturalistic Turn, and Legal Philosophy* Kenneth Einar Himma

Quine’s ‘Two Dogmas of Empiricism’ is thought to have revolutionary implications for philosophical methodology.¹ Quine’s view that there is no non-circular way to explicate the notion of analyticity, together with his view that any claim can be revised in the face of recalcitrant experience, seems to have refuted traditional views about conceptual analysis and metaphysics. Quine’s analysis shows, on this view, that philosophy lacks a distinct methodology and should be continuous with scientific theorizing. There are a number of legal philosophers who agree that philosophy should abandon traditional conceptual analysis (TCA) for scientific methodology. Brian Leiter, for example, complains that legal philosophy has ignored the damning Quinean criticisms of TCA: ‘While every major area of philosophy—metaethics, philosophy of language, epistemology, philosophy of science, philosophy of mind—has undergone a naturalistic turn over the last quarter-century, Anglo-American legal philosophy has remained untouched by these intellectual developments.’² Although naturalistic philosophy thrives in many areas, not much has changed. TCA, conceived of (or implicitly treated) as involving a priori reasoning, continues to flourish. Metaphysics—which means ‘beyond the physical’ and hence not amenable to empirical analysis—continues to thrive, as does TCA in epistemology, meta-ethics, philosophy of mind, and philosophy of law. The concepts of law, free will, mind, and goodness are explored in many papers every year—and * An earlier version of this paper was presented at the 22nd World Congress of the International Association for Philosophy of Law and Social Philosophy in Granada, Spain in 2005. I am grateful to Stanley Paulson, Jan Wolenski, Svein Eng, and George Pavlakos for their comments there. I am also indebted to Brian Leiter, Larry Alexander, Dennis Patterson, Julie Dickson, William Talbott, Laurence BonJour, Stephen Layman, Patrick McDonald, and Phil Goggans. ¹ WVO Quine, ‘Two Dogmas of Empiricism’(1951) Philosophical Review, 60 (1951). ² B Leiter, ‘Naturalism and Naturalized Jurisprudence’ in B Bix (ed), Analyzing Law (OUP, 1999) 80.

Reconsidering a Dogma: Conceptual Analysis, the Naturalistic Turn, and Legal Philosophy. Kenneth Einar Himma. © Oxford University Press 2007. Published 2007 by Oxford University Press.

4

Kenneth Einar Himma

the relevant methodology seems distinct from scientific methodology with its reliance on induction and empirical observation. This need not mean much. If naturalists are correct, TCA is flawed at its very foundations. Such analysis might be ‘thriving’ in that it is commonly done, but not in the sense that it is solving any problems. One way or another, the naturalist’s challenge requires an answer that engages the most influential arguments against TCA—which include Quine’s arguments. In this essay, I respond to the most important considerations adduced against TCA: (1) the denial of the distinction between analytic and synthetic truths; (2) recent research on the fallibility of ordinary intuitions; (3) the idea that all claims are revisable in the face of recalcitrant experience; and (4) the claim that even putatively conceptual claims are contingent in character. I show that each argument is vulnerable to serious objections.

I. What is TCA, Anyway? A. What are Concepts and Conceptual Analysis? Conceptual analysis is the analysis of concepts; and this raises the issue of what is a concept. This much is obvious: concepts are or correspond to mental ingredients needed to think about things of the relevant sort. For example, it is not possible to think about electrons without having or grasping the concept of an electron. This says little about concepts. It does not tell us anything about the nature of a concept; the way in which concepts are acquired; or the extent to which they are shared among members of a community. But it is clear that the possession or apprehension of a concept is a necessary condition for being able to think about the thing of which the concept is a concept. There are different views about what concepts are: (1) concepts are psychological states representing ideas or things; (2) concepts are abilities of a special kind—namely, the ability to discriminate one kind of thing of which a concept is a concept from another kind of thing of which another concept is a concept (having the concepts of tree and bush is being able to distinguish trees from bushes); and (3) concepts are meanings or ‘senses’ of words. The methodology of conceptual analysis must be responsive to what a concept is. For example, if concepts are psychological representational states, the proper analysis of any concept would involve saying something about the content of the representation as well as about the nature of the relevant state (which might be a compound state involving beliefs and dispositions of various kinds). Regardless of what concepts turn out to be, they are intimately associated with language. People use language to express concepts. We may not have a word associated with every concept we have. But we have a large number of words to

Reconsidering a Dogma

5

express, refer to, convey, or pick out concepts: we use ‘love’ to think or talk about love; ‘law’ to think or talk about law, and so on. Our ability to talk to each other about some thing that a concept is a concept of (say, law) says something about the relationship between the content of a concept and our use of the associated concept-term. It is reasonable to think that we have some pre-linguistic capacity to develop concepts and actually develop some long before we learn the associated word; without such a capacity, we could not learn a language. But the contents of our concepts are highly responsive to the linguistic practices involving the associated words because we cannot share an understanding of how to use a word without sharing a grasp of the core content of a concept.³ If you and I have radically different concepts of law, we will not be able to understand what each other is saying when using the term ‘law’. This is not to deny that disagreements arise within a community of speakers about the contents of concepts. Ask people whether the Pope is a bachelor and you will get the entire range of responses: ‘yes’, ‘no’, and ‘I don’t know.’ But it does mean that competent speakers of a language share an understanding about the core of a concept’s content. Disagreements arise only on penumbral issues. We might, for example, disagree on whether Bruce Willis is bald, but surely agree that Ben Kingsley is bald and that Patrick Dempsey is not. Disagreement about Bruce Willis is compatible with sharing a concept and an understanding of the term ‘bald’; disagreement about Ben Kingsley and Patrick Dempsey is not. If we do not agree on the latter cases, we do not share a concept of baldness. This has an important implication. Although we have a pre-linguistic ability to form concepts, the content of our concepts is determined (or defined) in part by the linguistic practices regarding the associated concept-terms. I might start out with a concept of water as being any clear, tasteless, scentless liquid but will eventually wind-up understanding the concept as picking out a particular liquid, one with a particular chemical structure, as I come better to understand the linguistic practices associated with the concept-term ‘water’ and learn more about the world.

B. What is Traditional Conceptual Analysis? The claim that concepts are associated with linguistic practices does not imply that concepts are linguistic entities; in particular, it does not imply that concepts ³ I am agnostic with respect to how these practices arise or should be characterized beyond claiming that (1) they are social practices that arise and are maintained in part because they are shared across a community and that (2) core elements of this practice yield norms that have prescriptive force among members of the relevant community. For example, one might think nothing more specific regarding our practices for using ‘bachelor’ than that they comprise various patterns of usage shared among a community. Alternatively, one might think these practices constitute conventions. Either way, the practices are social in character and express a shared norm that ‘bachelor’ is properly used only of men: it is uncontroversial that someone who uses ‘bachelor’ to talk about a dog has made a mistake.

6

Kenneth Einar Himma

are meanings. There is nothing in the claim that the content of our concepts is shaped by the content of linguistic practices that would entail that concepts are meanings, rather than mental representations or abilities. My concept of water might simply be an ability to distinguish things that are water from things that are not, even if that ability is responsive to how people around me use the term ‘water’. Still, people who do TCA generally presuppose that concepts are meanings. On this understanding, TCA seeks to identify the content of the sense of the relevant word—eg the meaning of ‘law’; and it does this by identifying properties that distinguish things that fall under the relevant concept from things that do not. These properties are usually thought expressible in the form of necessary and sufficient conditions for applying the concept. A classic example would be the notion of bachelor: X is a bachelor if and only if X is an unmarried adult male. The classic view is that descriptive conceptual analysis is a priori. While it is true that one cannot identify the core content of the patterns of usage that link words and concepts without empirical experience, no further experience is needed to justify true conceptual claims. Once it is observed that people use the word ‘bachelor’ only to refer to unmarried men, no further experience is needed to justify the claim that no woman is a bachelor. This claim is justified as a valid logical deduction from the conceptual claims that only men are bachelors and that no woman is a man. Not every purely deductive argument, of course, will be as easy as those described in the last paragraph. This will be obvious to anyone who has read Hart’s Concept of Law or Raz’s The Authority of Law or to anyone who has done any serious mathematics. It took hundreds of years to find a proof for Fermat’s Last Theorem; and the paper proving it was more than 100 difficult pages long— and something that no layperson could hope to understand.⁴ Of course, there is more to TCA than simply making inferences from core patterns of linguistic usage. Frank Jackson describes another methodological element in discussing the concept of free will: What we are seeking to address is whether free action according to our ordinary conception, or something suitably close to our ordinary conception, exists and is compatible with determinism, and whether intentional states according to our ordinary conception, or something suitably close to it, will survive what cognitive science reveals about the operations of the brain. . . . But how should we identify our ordinary conception? The only possible answer, I think, is by appeal to what seems to us most obvious and central about free action, determinism, belief, or whatever, as revealed by our intuitions about possible cases.⁵ ⁴ The view that TCA is a priori depends on the idea that the epistemic justification for logically valid rules of inference does not depend on experience; while I believe this is true (just as I believe this is true of the claim that 2+2=4), I don’t think anything depends on that assumption. Beyond challenging the arguments against TCA, I claim no more here than TCA is methodologically distinct from science. ⁵ F Jackson, From Metaphysics to Ethics: A Defence of Conceptual Analysis (OUP, 1998) 31.

Reconsidering a Dogma

7

TCA attempts to identify conceptual content that goes beyond the most obvious features of our shared practices by considering intuitions about possible cases. While the relevant intuitions are ordinary in the sense of being widely shared, they frequently make explicit features of our practices of which ordinary speakers might not be cognizant until they consider such cases. For example, most speakers will probably not realize that our concept of law does not entail that law must conform to morality⁶ until asked whether the Nazis had a legal system or whether the Jim Crow enactments were laws. Such cases help to flesh out the content of our shared practices in more specificity, identifying features of the concept that will have to be explained theoretically. On the traditional conception, the relevant intuitions are purely descriptive and do not include moral intuitions about right or wrong, good or bad. Identifying the content of a concept, on this view, is not governed by moral principles,⁷ even when the concepts are used to evaluate behaviour. For example, it is irrelevant in evaluating two proposed analyses of the concept of law whether one analysis is morally better than another—though moral standards obviously should play a role in deciding what should be enacted into law. TCA is purely descriptive and purely general in the sense that these intuitions range over all possible cases.⁸ Jackson distinguishes a modest from an immodest conception of conceptual analysis, claiming that TCA should be thought of as modest. According to the immodest conception, conceptual analysis gives us insight into what the world is like; that is, an analysis of the content of our concept of law, for example, would give us insight into the essential nature of law as it really is independent of our linguistic practices and conceptual frameworks. According to the modest conception, conceptual analysis merely ‘tells us what to say in less fundamental terms given an account of the world stated in more fundamental terms’ (Jackson, 44). The talk of moving from more to less fundamental terms reflects the commonsense view of analysis as attempting to resolve and explain what is complex in more simple terms. The idea is that when we analyse a concept, we break it down into simpler constituents so as to display its logical structure and explain its central features—an idea that is quite common historically; though theorists have frequently supplemented and modified the analysis of analysis, the basic features have nonetheless remained largely intact.⁹ ⁶ There is no disagreement among legal theorists, legal practitioners, or even laypersons on this issue. Indeed, natural law theorists in the Thomist tradition, like John Finnis, deny that Aquinas claimed that unjust norms could not be law. ⁷ There are two kinds of norm that might figure in analysing a concept: epistemic norms, like that requiring consistency, and moral norms, including those that govern states. There is no controversy about whether epistemic norms should play a role in theorizing. See J Coleman, The Practice of Principle (OUP, 2001). ⁸ Many theorists, including Ronald Dworkin, believe it is not possible to give a purely descriptive analysis of evaluative concepts like law. See, eg Dworkin, Law’s Empire (Harvard University Press, 1986). ⁹ M Beaney, ‘Analysis’ in E Zalta (ed), Stanford Encyclopedia of Philosophy; ; see also S Blackburn, Oxford Dictionary of Philosophy (OUP, 1996).

8

Kenneth Einar Himma

But the important point here is that modest conceptual analysis presupposes an account of the world ‘stated in more fundamental terms’ as its starting point. This account, however, must be someone’s account; and, on Jackson’s view, it is a shared account grounded in shared intuitions and is thus our account of the relevant thing to which the associated term refers. TCA thus presupposes a certain story about things—one that is our story—and does not purport to describe the world as it is independently of our stories about things (though our story might be the objectively correct story about things; modest TCA does not rule that possibility out). Jackson understates the role of linguistic practices in determining the content of our concepts. While it might be true that all possible concepts exist in logical space along with every other abstract object, the content of our concepts is at least partly fixed by our linguistic practices. Indeed, as critics of TCA like Quine concede, if we used the symbol ‘water’ to refer to something other than the clear liquid to which it is used to refer, it would not be a conceptual truth that water is H2O—though there would undoubtedly be some conceptual truth corresponding to the relationship between some term and H2O. Concepts might be abstract objects independent of our social activities about which there are utterly mindindependent objective truths, but language is a social construct grounded in a social practice; and which concepts our words pick out or express is partly defined by the content of those practices.¹⁰ This suggests that explanations of the traditional methodology for descriptive conceptual analysis are not entirely accurate. Although philosophers frequently justify conceptual claims by an appeal to ‘ordinary intuitions’, the intuitions that matter are shared views reflecting the core practices for using the terms. If, as Jackson maintains, ‘[t]he business of consulting intuitions about possible cases is simply part of the overall business of elucidating concepts by determining how subjects classify possibilities (33),’ then shared views about language will condition the relevant intuitions because how subjects classify possibilities depends on the core practices for using the words. We classify things by using words, which in turn, on the traditional view, express, mean, or convey concepts. Analysing the content of a concept, on this view, requires identifying shared practices concerning the word.¹¹ Some of the relevant views might be non-linguistic in character and express some widely accepted belief about a mind-independent feature of reality to which we want to refer. But, in such cases, these non-linguistic ‘intuitions’ about the ¹⁰ The nature of these linguistic practices is a highly contentious matter. I take no view here on which view is correct. See n 3, above. ¹¹ As Hart put it: ‘Many important distinctions, which are not immediately obvious, between types of social situation or relationships may best be brought to light by an examination of the standard uses of the relevant expressions and of the way in which these depend on a social context, itself often left unstated. In this field of study it is particularly true that we may use, as Professor JL Austin said, “a sharpened awareness of words to sharpen our perception of the phenomenon”’. HLA Hart, The Concept of Law (revised edn, OUP, 1994) v.

Reconsidering a Dogma

9

world will likely play an important role in defining core features of our practices regarding a concept-term. For example, part of what will explain why we use the term ‘water’ to refer to only clear liquids with the chemical structure of H2O is the scientific discovery that water is H2O; this non-linguistic consideration helps to explain why we have converged on using ‘water’ to apply only to H2O. The content of natural-kind concepts frequently changes in response to scientific discoveries. But the relevant views are usually informed by our core linguistic practices for using the associated term. I have, for example, the intuition that all bachelors are adult males because, as an empirical matter, no one uses the term ‘bachelor’ to refer to boys or women, and not because I have some non-linguistic intuition about the nature of bachelors. Our shared practices converge on calling only men ‘bachelors’—and my intuition that only men are bachelors is informed by and expresses those shared linguistic practices. Of course, I use the term ‘bachelor’ to think about bachelors; but what I take to distinguish bachelors from non-bachelors is conditioned by our practices regarding ‘bachelor.’ Indeed, it is hard to imagine how one could come to have a concept of bachelor except by learning two different kinds of social practices regarding ‘bachelor.’ While we have a pre-linguistic capacity for forming concepts, bachelor is not the kind of concept that can be developed without understanding social practices defining the institution of marriage, as well as social practices defining the terms we use to pick out and express these institutions. Either way, these views matter because, and only because, they are related in the right kind of way to the core practices for using the relevant terms. What concept a term picks out is defined, in the first instance, by our shared social practices for using it. For this reason, an analysis of the concept must be grounded in those views that define or express the core content of those shared social practices. While people have idiosyncratic intuitions that go beyond these core practices, appeals to intuition are relevant only if shared across a linguistic community; and intuitions will be shared across a linguistic community only if they fall within the core of our practices for using the term. When people widely disagree about intuitions, the issue is not settled by these practices. In these cases, the practices are vague or indeterminate with respect to the issue. TCA, then, has an undeniably empirical (or ‘naturalistic’) element.¹² What views are shared among a community is a contingent matter that cannot be determined without empirical observation. While it is true that one reliable way to do the empirical work, as Jackson suggests, is to take an opinion poll, it is not usually ¹² Frederick Schauer makes exactly this point: ‘Like his philosophical colleagues at the time, [Hart] examined ordinary usage closely for the distinctions it embodied and rigorously analysed and defined the terms used to mark legal concepts. . . . Conceptual analysis, as Hart surely must have recognized, inevitably rests at least in part on empirical observation, and this empirical foundation of conceptual analysis could also implicity undergird his claim to be doing descriptive sociology.’ Schauer, ‘(Re)taking Hart’ (2006) 119 Harvard Law Review 852, 860–61.

10

Kenneth Einar Himma

necessary. Armchair sociology of the sort at which philosophers excel (usually without realizing that they are doing sociology) is sufficiently reliable as long as one belongs to the linguistic community under consideration. It is tempting to think this deprives TCA of its distinctively philosophical character. Leiter, for example, argues that TCA is nothing more than ‘glorified lexicography’: Conceptual analysis, as Jackson conceives it, becomes hard to distinguish from banal descriptive sociology of the Gallup Poll variety. Indeed, Jackson says explicitly that he advocates, when necessary, ‘doing serious opinion polls on people’s responses to various cases’! But this now seems to blur the line between conceptual analysis and lexicography: for does not lexicography aim to track statistically normal usage of words or concepts, precisely the pattern of usage a well-designed opinion poll would detect (Leiter, 41 SSRN)?

This is a mistake. Conceptual analysis might start from something that resembles lexicography, but it is considerably more than this. Conceptual analysis, as it is done in metaphysics, meta-ethics or philosophy of law, goes deeper than just identifying shared views; that, of course, is the job of a lexicographer who records empirical patterns of word-usage. Conceptual analysis attempts to theorize these views by identifying deeper metaphysical commitments that they imply or presuppose, as well as more general principles that explain them. While this might or might not be a distinctively philosophical enterprise, it goes well beyond the empirical task of identifying shared intuitions or core features of our linguistic practices. There is an easy way to see this point: simply compare what lexicographers have to say about the word ‘law’ with what Hart has to say by way of explication of the concept of law. Here is how the lexicographers for the Oxford American Dictionary have defined law: law | noun 1 (often the law) the system of rules that a particular country or community recognizes as regulating the actions of its members and may enforce by the imposition of penalties : they were taken to court for breaking the law | a license is required by law | [as adj. ] law enforcement. • an individual rule as part of such a system : an initiative to tighten up the laws on pornography. • such systems as a subject of study or as the basis of the legal profession : he was still practicing law | [as adj. ] a law firm. Compare with jurisprudence. • a thing regarded as having the binding force or effect of a formal system of rules : what he said was law.

The differences between the philosophical analysis of law and the lexicographical analysis of law are conspicuous. First, there is no mention here of many pieces central to Hart’s analysis: social practices; the rule of recognition; secondary and primary rules; validity; etc. Second, the lexicographer’s job is accomplished in a few lines, while Hart took

Reconsidering a Dogma

11

more than 200 pages in analysing law in The Concept of Law. If Hart starts from the shared views constituting meaning, it should be clear that he is also doing something radically different from what lexicographers are doing. Part of the reason that there is so much more to a philosophical analysis of a concept than a lexicographical analysis of a word’s meaning is that modest conceptual analysis cannot avoid making metaphysical claims—though those claims are concept-dependent. Although this is rarely discussed, our linguistic practices have metaphysical implications. For example, the sentence ‘2 is an even number’ implies that there exists something that is fairly named by the term ‘2’; insofar as ‘we cannot touch 2’ is true, it follows that 2 is an abstract object—and that is a metaphysical claim. Likewise, if it is a conceptual truth that every bachelor is unmarried, it follows that it is not logically possible for someone to be a married bachelor. Claims about what is or is not causally possible, of course, are not metaphysical; it is an empirically verifiable matter that it is not causally possible for me to jump unaided to the moon. But claims about what is or is not logically possible are metaphysical in character because they go beyond making claims about features of this specific world with its natural laws; no matter what else the world might look like in terms of the laws of nature describing the casual behaviour of physical objects, there could not be a married bachelor given the core features of our linguistic practices for using ‘bachelor.’ Modest conceptual analysis cannot be only about conceptual content; it cannot avoid making metaphysical claims—though these claims need not (but could) be the absolute objective truth about reality an sich as it is independent of our conceptual framework. Indeed, an analysis of a concept is intended to make claims about the nature of something. People who analyse the concept of law, for example, take themselves to be explaining the nature of law. But, again, the nature of law, as we understand it, is partly fi xed by our shared practices, both institutional and linguistic, for using the relevant vocabulary. The fact that we use the term ‘law’ to refer to institutional normative systems, instead of to water, plays a role in determining the nature of the thing to which we refer by using ‘law’.¹³ It is, after all, our concept of law that the theorist attempts to analyse; and the content of our concepts is at ¹³ The nature of law will be determined, in part, by a second set of social practices. Since the relevant linguistic conventions dictate that the term ‘law’ applies to certain social institutions, the nature of law is determined at least in part by the social practices that bring those institutions into existence and constitute their operations. Accordingly, a conceptual analysis of law will have to be minimally responsive to these conventional practices—as well as core features of the associated linguistic practices—because these practices contribute to determining the very nature of law. Natural kind terms like ‘water’ and evaluative terms like ‘good’ are different: there is an objective truth that is at least partly independent of our conceptual practices about what objects are picked out by these terms. The substance to which we refer by use of ‘water’ is not constructed by our social practices in the way that law is constructed by certain social practices; law is, while water is not, a social institution. Natural kind terms and evaluative terms need not be as tightly responsive to social practice, conceptual and otherwise, as social kind terms like ‘law’. I am grateful to Larry Alexander for helping me to see this.

12

Kenneth Einar Himma

least partly fi xed by our shared linguistic practices. Even so, the goal of conceptual analysis is to identify the nature of the thing to which we use the conceptterm ‘law’ to refer. Modest conceptual analysis is not glorified lexicography. Conceptual analysis does not stop with the identification of shared views about the application of a word; lexicography does. Looking just at his ‘modest’ claims about law, the kind of thing that Hart did in The Concept of Law, which was surely grounded in armchair sociology directed at identifying core linguistic practices, goes so much deeper than the polling used by lexicographers that it is obviously false to suggest that they are doing the same kind of thing. Ironically, conceptual analysis as glorified lexicography is more likely to result from naturalizing philosophical method than it is to result from TCA. Once we remove the most abstract and speculative elements of Hart’s theorizing (ie those elements supposed to be the distinctive characterizing features of TCA’s methodology) the result is a theory that consists of little more than the shared views that can be experientially confirmed by empirical observation of the sort made by lexicographers—who are ‘scientists’. At least, one scientific approach to analyzing concepts, the lexicographical approach, yields nothing more than lexicography. More importantly, it is just not clear that the most abstract and speculative elements of traditional philosophical methodology are properly characterized as ‘continuous with scientific methodology.’ The problem is that the most ‘abstract and speculative areas of science’ naturalists believe should absorb philosophy are regarded by many scientists as being too far removed from experiential confirmation to count as ‘science’. For example, although string theory has drawn much interest as potentially unifying Einstein’s theory of relativity and quantum mechanics by eliminating certain tensions between them, many physicists reject the methodology as too abstract and removed from experiential confirmation to be science.¹⁴ Indeed, some highly respected physicists have rejected string theory as a research agenda for physics, characterizing it instead as ‘metaphysics’ and ‘philosophy’.¹⁵ Naturalists simply assume that abstract, speculative theorizing on non-linguistic matters that cannot be confirmed by experiential observation is compatible with an uncontroversial account of scientific methodology; and this is far from obvious. Although TCA and science surely share much by way of methodology, it is disputed whether the abstract speculative focus of metaphysical and conceptual theorizing is also an appropriate focus for scientific methodology. Many ¹⁴ For an accessible discussion of string theory and the associated controversies, see B Greene, The Elegant Universe (Vintage Books, 2003), esp 211–14. ¹⁵ See, eg L Smolin, The Trouble with Physics: The Rise of String Theory, the Fall of a Science, and What Comes Next (Houghton Mifflin, 2006); P Woit, Not Even Wrong: The Failure of String Theory and the Search for Unity in Physical Law (Basic Books, 2006); for a review of these books see, J Holt, ‘Unstrung’ New Yorker, (2 October 2006; available at .

Reconsidering a Dogma

13

reputable scientists reject that assumption—and plausibly so. Perhaps this is nothing more than a semantic dispute, but it seems clear that traditional philosophical theorizing (including TCA) has a different focus than scientific theorizing as currently understood.

II. Quine’s Circle-of-Terms Argument in ‘Two Dogmas’ The possibility of a unique methodology for conceptual analysis rests on the distinction between analytic and synthetic truths. Analytic truths are true in virtue of meanings, while synthetic truths are true in virtue of other facts that must be empirically observed. TCA purports to identify, explain, and theorize the semantic content of concept-terms; while the content of a concept is not exhausted by the ordinary meaning of the associated word, it is fi xed by the meaning. Accordingly, a conceptual analysis purports to justify analytic truths. Thus, if there are no analytic truths, then TCA is problematic because the methodology is contrived to do what cannot be done—namely, specify the content of a concept in terms of analytic truths about that concept. In ‘Two Dogmas’, Quine attempts to show that there are no analytic truths and hence that TCA should be ‘abandoned’ in favour of an empirical methodology continuous with science. The argument proceeds in two steps. The first tries to show that analyticity cannot be grounded in the notion of meaning. The second tries to show that every other possible formulation of the distinction is circular in that we cannot make sense of the terms supposed to clarify the analytic-synthetic distinction without presupposing the very distinction they are supposed to clarify.

A. An Initial Difficulty: What Exactly is Quine Trying to Show? Quine’s position is ambiguous.¹⁶ At the beginning of the essay, he states it in very strong terms, suggesting there is no distinction whatsoever of this kind: ‘[The] belief in some fundamental cleavage between truths which are analytic, or grounded in meanings independently of matters of fact and truths which are synthetic, or grounded in fact . . . [is] ill founded . . . [and should be] abandoned.’ Elsewhere he makes the weaker claim that such a distinction cannot be given an adequate philosophical grounding and can be accepted only as ‘a metaphysical article of faith’. These are two different claims. The first sentence asserts there is no difference between true statements called ‘analytic’ and those called ‘synthetic’. The second asserts we cannot formulate a rigorous explanation of the analytic-synthetic ¹⁶ The following discussion is clearly indebted to Scott Soames’s excellent discussion in Soames, Philosophical Analysis in the Twentieth Century, vol 1 (Princeton University Press, 2003).

14

Kenneth Einar Himma

distinction and hence are not justified in accepting it; if such a distinction is to be accepted, it must be accepted as an ‘article of faith’. This weaker claim says nothing about whether there are any analytic truths; it claims only that our belief that ‘every bachelor is a man’ expresses an analytic truth must be accepted on faith because we cannot explain the distinction between analytic and synthetic truths in a philosophically rigorous way that would show there are analytic truths. Quine’s arguments, if sound, justify at most the second weaker claim. If we are unable to give an adequate theoretical ground for a distinction, it might follow we are not epistemically justified in accepting that ‘every bachelor is unmarried’ expresses an analytic truth and that ‘some bachelors are taller than 6 feet’ does not. It does not follow that there really is no distinction between the propositions expressed by these sentences; that we cannot rigorously explain a distinction between a set we refer to by the term A and a set we refer to by the term B doesn’t imply there is no distinction between the sets marked by A and B; we make mistakes about what we think we can and cannot make sense of all the time. It is hard for us to make rigorous sense of Einstein’s claim that space is curved, but it doesn’t follow that space is not curved—as naturalists would point out. Our epistemic abilities just cannot do that kind of heavy lifting. Indeed, HP Grice and PF Strawson argue that even if we cannot rigorously ground the distinction, there is still good reason to think the intuitive distinction marks a difference between analytic and synthetic truths. There is widespread agreement among competent speakers on which statements are analytic and which are synthetic. For example, any two competent speakers agree, once they get an intuitive grasp of what is being asked, that ‘bachelors are unmarried’ is analytic, while ‘the sun rises in the east’ is synthetic. Moreover, they agree on not only the standard examples, but also on any new examples they are asked to consider. This suggests that the ability to distinguish statements into these categories is ‘genuine’ and that there is a difference between them; the ability to see a difference could not be ‘genuine’ if there were no relevant difference expressed by the terms ‘analytic’ and ‘synthetic’.

B. A Missing Argument Quine’s first step, as noted above, is to show that analyticity cannot be adequately explained in terms of the ordinary notion of meaning. The idea here is thus to repudiate what is perhaps the most natural approach to distinguishing analytic from synthetic statements: analytic truths, on this view, are true wholly in virtue of the meanings of the terms, whereas synthetic truths are not.¹⁷ ¹⁷ It seems to me that one can define the distinction without presupposing there are meanings: A proposition is an analytic statement if and only if a competent speaker can discern its truth-value simply in virtue of being a competent user of the words in the sentence; otherwise, the proposition is synthetic. I will not pursue this issue further in this paper.

Reconsidering a Dogma

15

Everything Quine has to say about meaning is contained in the following passage: For the theory of meaning a conspicuous question is the nature of its objects: what sort of things are meanings? A felt need for meant entities may derive from an earlier failure to appreciate that meaning and reference are distinct. Once the theory of meaning is sharply separated from the theory of reference, it is a short step to recognizing as the primary business of the theory of meaning the synonymy of linguistic forms and the analyticity of statements; meanings themselves, as obscure intermediary entities, may well be abandoned (TD 22).

Quine is certainly correct about this much: there is much obscure and difficult about the ordinary notion of meaning. It is not at all clear what sort of entity a meaning would be—beyond the obvious fact that it would have to be an abstract object. It is likewise unclear whether meanings are vague or whether vagueness owes to our lack of understanding about a word’s meaning. But none of this justifies abandoning an intuitive notion of meaning deeply entrenched in our ordinary pre-philosophical understandings. As Laurence BonJour observes: This is not much of an argument for ‘abandoning’ meanings as a type of entity; and in any case it is not at all obvious that giving up the idea of meanings as entities requires giving up the whole idea that words have meaning, which is in effect what Quine proceeds to do. It is the rather casual dismissal of the idea of meaning that makes it seem plausible that synonymy must be explained by appeal to the idea of definition or that of interchangeability salva veritate or ultimately by appeal to analyticity itself. . . . But this is at least largely misdirection, for surely the most natural and obvious course of explanation is to explain synonymy as sameness of meaning and then explain analyticity in terms of synonymy in the Fregean way.¹⁸

The idea that meaning is ‘obscure’ is not enough, by itself, to justify abandoning it—especially given the counter-intuitive quality of the position that Quine is defending here; after all, the claim that ‘bachelor’ means ‘unmarried adult male’ seems much more plausible from an intuitive point of view than anything that Quine can offer by way of rebuttal. Indeed, the idea that we should reject obscure notions has absurd consequences. For example, the concept of a number is obscure. It is not, for example, at all clear what the content of the concept would be. Numbers have been defined in set theory as equivalence classes: the number ‘2’ picks out the set of all objects that can be placed in one-to-one correspondence with the set {∅, {∅}}. Intuitively, the number 2, then, is the set of all sets with 2 members. This might not be terribly obscure, but it is sufficiently counter-intuitive to leave us without an adequate understanding of the concept. Yet ‘abandoning’ the notion of a number—or the equally unclear notions of a set, morality, etc—would not only have devastating ¹⁸ L BonJour, In Defense of Pure Reason (CUP, 1998) 71.

16

Kenneth Einar Himma

consequences, but is also an overreaction: the proper reaction is to continue to think about these concepts. This renders the Quinean project a non-starter. If he cannot give an argument that justifies giving up the concept of meaning, then he can give no reason to give up the distinction between analytic and synthetic truths because there is a plausible way to explain the distinction: analytic truths are distinguished from synthetic truths in that the former, but not the latter, can be transformed into logical truths by substituting one term for another with the same meaning. While the second step in his argument is intended to show we cannot explain the distinction between the two kinds of truth this way because we cannot explain sameness of meaning (ie synonymy) in a non-circular way, this argument ultimately accomplishes nothing because it falsely presupposes that Quine’s first argument succeeds in showing we should give up the notion of meaning because it is obscure.¹⁹

C. The ‘Circle-of-Terms’ Argument Although the considerations adduced above suffice to reject the Quinean argument in ‘Two Dogmas’, it is worth considering whether Quine succeeds with his second step. Quine’s strategy here is to consider all the possible explanations of analyticity and to show that each implicitly presupposes the very notion of analyticity it purports to explain. Since we cannot explain analyticity in terms of meaning or any other consideration, Quine concludes we have no adequate explanation of analyticity and that the analytic-synthetic distinction should be abandoned. Quine begins by rejecting the idea that we can distinguish analytic from nonanalytic statements in that the former, but not the latter, can be transformed into logical truths simply by substituting synonyms for synonyms. The problem, according to Quine, is that the notion of ‘“synonymy” is no less in need of clarification than analyticity itself.’ Likewise, we could not explain analytic statements as being ‘true by definition’ because the notion of definition presupposes the notion of synonymy, which, again, is itself in need of clarification. Further, Quine denies we can provide a non-circular account of ‘synonymy’. The most plausible account, on Quine’s view, is the idea that A and B are synonymous if and only if A and B can be substituted for one another in any statement without change of truth-value. On this line of analysis, then, ‘bachelor’ and ‘unmarried man’ are synonyms because we can substitute the latter for the second instance of the former in the logical truth ‘Necessarily, all and only bachelors are bachelors’ and get a true statement—namely, ‘Necessarily, all and only bachelors are unmarried men.’ This shows both that ‘bachelor’ and ‘unmarried man’ are synonyms and that the statement ‘all and only bachelors are unmarried men’ is analytic. ¹⁹ I am indebted to Laurence BonJour for this important point.

Reconsidering a Dogma

17

Quine responds that this explanation is circular because it relies on the notion of necessity: The condition of interchangeability salva veritate varies in its force with variations in the richness of the language at hand. The above argument supposes we are working with a language rich enough to contain the adverb ‘necessarily,’ this adverb being so construed as to yield truth when and only when applied to an analytic statement. But can we condone a language which contains such an adverb? Does the adverb really make sense? To suppose that it does is to suppose that we have already made satisfactory sense of ‘analytic.’ Then what are we so hard at work on right now?

On Quine’s view, we can explain ‘synonymy’ in terms of interchangeability only in languages containing the notion of ‘necessity’, but the notion of ‘necessity’ itself requires the notion of ‘analyticity’ for its explanation. Accordingly, such an explication is, like the others, circular. There are two problems here. First, Quine incorrectly presupposes that the only necessary truths are analytic truths: the notion of necessity, he believes, should be construed ‘so narrowly . . . as to be applicable only to analytic statements’ (TD 29). It is true that the logical positivists, the targets of Quine’s attack in ‘Two Dogmas’, believed that necessity and a prioricity must be explained in terms of analyticity, but this view has fallen out of favour with most theorists since Kripke’s Naming and Necessity seemed to show the notions are independent. Second, the claim that we cannot explicate necessity without reference to analyticity does not show there is no distinction between analytic and synthetic statements. If the notion of analyticity were incoherent, that would be a good reason to think that there are no analytic truths and hence that all statements are synthetic (or empirical in character). Unlike a showing that a notion is incoherent, however, the fact that we cannot give a non-circular analysis of a concept does not entail that there is nothing in the world to which that concept applies. To see why, it is helpful to note that the sort of circularity that Quine attributes to the notion of analyticity is a common property of concepts. We cannot, for example, give a non-circular account of modal terms like ‘necessity’ and ‘possibility’. If we follow logicians in trying to explain necessity in terms making reference to possible worlds, we have presupposed the very thing we are trying to explain because to say that a statement is true in a possible world is to deny that its negation is necessary. We cannot explain possibility in terms of consistent state of affairs because a consistent state of affairs is one that is possibly true. A properly non-circular account of a modal concept will not rely on any modalities presupposing the modality in question. Quine, of course, would respond that this shows that the necessity-possibility distinction is problematic in the same sense as the analytic-synthetic distinction, but this takes us only so far. As Grice and Strawson point out: It would seem fairly clearly unreasonable to insist in general that the availability of a satisfactory explanation in the sense sketched above is a necessary condition of an expression’s

18

Kenneth Einar Himma

making sense. . . . Even if such explanations can be given in some cases, it would be pretty generally agreed that there [are] other cases in which they cannot. One might think, for example, of the group of expressions which includes ‘morally wrong’, ‘blameworthy’, ‘breach of moral rules’, etc.; or of the group which includes the propositional connectives and the words ‘true’ and ‘false’, ‘statement’, ‘fact’, ‘denial’, ‘assertion’. Few people would want to say that the expressions belonging to either of these groups were senseless on the ground that they have not been formally defined (or even on the ground that it was impossible formally to define them) except in terms of the members of the same group.²⁰

There are a couple of points worth making here. First, the above quote points at just the tip of the iceberg. An account of the semantics of universally quantified sentences faces exactly the same problem as an account of the semantics of modal sentences: it presupposes the very notion it is trying to explain. According to the standard account of the semantics of the universal quantifier, the statement (∀x)ϕ(x) is true in a model M if and only if ϕ(a) is true in M of every referring constant a in the language—‘all’ and ‘every’ being obvious synonyms (or, as Quine might put it, ‘every’ being as much in need of theoretical explanation as ‘all’). As far as I know, there is no properly non-circular account of the meaning of ‘all’. Likewise, it is not possible to give a non-circular account of the concept of ‘set’. The mathematical theory of sets defines ‘set’ as ‘a collection of entities’. But the concept of a set cannot be explained without using a term that is synonymous with ‘set’, such as ‘collection’ or ‘group’, or in terms of some operation like ‘aggregation’ that presupposes we can already make sense of the notion of a set. Giving up the notion of a set entails giving up the theoretical and philosophical foundations for most of mathematics—including number theory. This, by itself, is surely reason enough to reject Quine’s reasoning here. But things get even worse for Quine’s argument. Mathematical concepts are not the only ones with this property. Complex social concepts like ‘law’ cannot be explained in non-circular terms. To fully explain the concept of law as Hart intends, we must include reference to something that counts as a ‘state’; on Hart’s account, international law and Catholic law might be related forms of law, but they are not systems of ‘municipal law’ as Hart explains the term. But it is hard to see how one could explain the notion of a state without having the concept of law or legal system; if not identical with a municipal legal system that has the broadest reach in a particular region, a state surely includes that system as part of what constitutes it as a state. It is not clear even that the concept of bachelor can be defined in a non-circular way. Any definition of ‘bachelor’ must rely on one of the concepts of marriage, solemnization ceremony, or marriage contract. But none of these concepts can be reduced to a conjunction of atomic pieces. ‘Unmarried’, as Quine might say, is no less in need of clarification than ‘bachelor’. ²⁰ Grice and Strawson, ‘In Defense of Dogma’ (1956) 56 Philosophical Review 141, 148.

Reconsidering a Dogma

19

Second, Quine’s claim is considerably stronger than the point that these expressions are, as Grice and Strawson put it, ‘senseless’ in the absence of the right kind of definition. This, after all, is not an extreme or implausible idea: if there is no possible way (ie for any possible individual) to properly define an expression, it is not unreasonable to think that it lacks a meaning (or sense)—since a definition would specify the sense of the expression. Rather, Quine’s claim is that if there is no possible way for us to define a distinction in a non-circular way, there is nothing in the world to which the distinction corresponds. Properly construed, then, the point would be that if, for example, we human beings cannot give a noncircular definition of the distinction between universal and existential statements or between law and non-law, then there is nothing in the world marked by the associated expressions. That goes well beyond the more modest point that such expressions are ‘senseless’ or lack meaning. It would, of course, be a disaster from a practical point of view to do without these concepts. One might think we can get along just fine without modal statements; one would, of course, have to deny the impossibility of all metaphysics and much moral theorizing in the process, but naturalists who believe scientific method is the only route to knowledge are perfectly comfortable with that. But no one will be tempted to deny these other distinctions. We simply cannot understand very much about the world without concepts like ‘all’, ‘set’, and ‘law’. But it is a mistake to infer from our inability to provide a non-circular account of such concepts that there is no genuine distinction between universal and existential statements, sets and non-sets, or law and other normative systems. That limited beings like us cannot provide a non-circular definition of a concept is a point about our epistemic limitations. That we cannot define a term in a noncircular way doesn’t tell us anything about what there is in the world. It seems clear, for example, that any plausible ontology of the world will include things like sets, numbers, and laws—even if we cannot explicate these notions in a noncircular way. In fact, it would even be a mistake to infer that there are no sets in the world even if it is simply not possible to use any conceivable language to define ‘set’ in a non-circular way—and, ironically, we can look to some of Quine’s own views on language and meaning to see why. Quine argues that the semantic contents of at least some words and sentences are not fi xed entirely by some set of atomic concepts. On his view, the meanings of many words and sentences are defined holistically by reference to a web of non-atomic terms and sentences. While one can reasonably doubt that all meaning is holistic in this way, some meanings seem holistic in this sense. If one can specify or understand the meaning of, say, ‘apple’ independently of some other set of terms, one cannot specify the meanings of ‘force’, ‘mass’, and ‘energy’ this way: the specification of these terms requires a theory that links these terms to one another. There simply are no non-circular, atomistic definitions for any of these terms; they can be defined and understood only in terms of some kind of semantic web.

20

Kenneth Einar Himma

But the claim that we cannot give a non-holistic account of a concept C does not imply anything about whether there is anything in the world that satisfies the application-conditions for the concept. The explanation for why we can give only a holistic account of C might be that we simply lack sufficient intellectual ability to articulate fully the terms of our language in atomic pieces independent of one another, or it might be that any language like ours cannot be fully articulated in non-holistic terms. Either way, we cannot infer from the unavailability of a non-holistic explanation of C that C and its complement fail to mark a genuine distinction in the world. As a logical matter, a point that is ultimately about the limits of our intellects or languages cannot tell us anything about what there is in the world. Ironically, the notion of a holistic account of a concept is unclear because it has not been fully explained (and should be, on Quine’s own view, abandoned); but this much is clear: if there are concepts for which only a holistic explication is possible, those accounts can yield claims that would look very much like what an application-condition for a concept looks like on a traditional view. Whatever, for example, the proper explication of ‘bachelor’ turns out to be, it will have to yield a norm that ‘bachelor’ is properly applied only to men and hence that only men are bachelors—which continues to look like a plausible candidate for an analytic truth. And this is true even if a proper explication of the notion of bachelor is a holistic one that requires reference to closely-related terms like ‘marriage’ or ‘solemnization ceremony’. At most, Quine has shown that the concepts of analyticity, meaning, definition and necessity are holistic in that they are parts of a logically interrelated web of concepts. But, the fact that ‘analyticity’ and ‘syntheticity’ cannot be defined in a non-holistic, non-circular way does not imply there is no real distinction between analytic and synthetic truths.

III. The Unreliability of Intuitions A. The Argument from Intuition A second line of argument against TCA is directed at its reliance on ordinary intuitions about ‘possible cases’.²¹ The idea is that a conceptual analysis of, say, law is constrained by our folk intuitions about what is properly characterized as ‘law’; thus, for example, a conceptual analysis of law should cohere with the nearly universal intuition that Nazi Germany had a system of law despite the unconscionable quality of the content of its laws. Ordinary intuitions about such cases provide the foundation for a TCA of law.

²¹ F Jackson, From Metaphysics to Ethics: A Defense of Conceptual Analysis (OUP, 1998).

Reconsidering a Dogma

21

The idea that intuition should play a role in serious philosophical enquiry has been challenged for a number of reasons,²² but the most influential challenge focuses on the reliability of intuition. As is well known, intuitions have been wrong in some significant cases. Einstein’s theory of relativity presupposes a non-Euclidean geometry and arguably refutes the intuition that physical space is Euclidean. Quantum mechanics is streamlined by denying the law of the excluded middle and arguably refutes the underlying intuition.²³

B. How Much Should be Made of Intuition’s Mistakes? That intuitions have been incorrect is not surprising. First, ordinary intuitions on philosophical issues are influenced by many contingent factors.²⁴ To begin, intuitions frequently differ from one culture to another; East Asians, for example, are more likely than Westerners to characterize the Gettier cases as involving knowledge.²⁵ Additionally, what a person’s intuitions are on any particular case is influenced by how the case is presented; a person might respond in one way if the case is presented without other cases and another way if the case is presented as one of a series of cases. Intuition could not be infallible along the range of philosophical questions if the content of one’s intuitions are conditioned by such contingent matters. Second, many of these intuitions concern mind-independent aspects of reality. It is implausible to think that finite beings like us would have infallible intuitive insight into any mind-independent feature of reality; our intellectual ability is quite limited, and it would be hard to explain our having an infallible intuitive faculty to grasp any interesting class of mind-independent truths. Many moral objectivists, for example, reject an intuitionist epistemology believing that intuition lacks direct access to mind-independent moral truths. But the idea that intuition can lead us into error (even systematically), by itself, does not vitiate TCA’s reliance on intuitions. First, as George Bealer has pointed out, there are a variety of different types of intuition—moral intuitions, physical intuitions, and rational intuitions: We have a physical intuition that, when a house is undermined, it will fall. This does not count as a rational intuition, for it does not present itself as necessary: it does not seem that a house undermined must fall; plainly, it is possible for a house undermined to remain in its original position or, indeed, to rise up. By contrast, when we have a rational ²² See, eg J Hintikka, ‘The Emperor’s New Intuitions’ (1999) Journal of Philosophy 127–47. ²³ Still, as BonJour points out, the mistakes are comparatively few and far between: ‘It should also be stressed, however, that while identifiable mistakes of these kinds are clear enough and frequent enough to undeniably refute the thesis of infallibility, they are at the same time extremely rare in relation to the overall body of claims that are, if the rationalist is correct, accepted on an a priori basis.’ BonJour (1998) 111. ²⁴ J Weinberg, S Nichols, and S Stich, ‘Normativity and Epistemic Intuitions’ Philosophical Topics. ²⁵ ibid.

22

Kenneth Einar Himma

intuition—say, that if P then not not P—it presents itself as necessary: it does not seem to us that things could be otherwise; it must be that if P then not not P.²⁶

The sociological studies, as Bealer points out (209), that purport to cast doubt on the general reliability of intuition do not distinguish the different kinds of intuition. This is a critical omission because one, but by no means the only, class of intuitions that matters for conceptual analysis is rational intuition, for example, about valid logical inferences. Second, the same is true of other sources of belief we rightly regard as trustworthy. Both perception and introspection can lead to serious error, as when I perceive that the sun is a small disk or when I introspect that I am in love when I am not. Moreover, sociologists have shown that perceptual observation, the foundation of the scientific methodology naturalists believe should be adopted by philosophers, can be systematically misleading; as Ernest Sosa puts the point: [I]t is known that human perception is systematically misleading in certain conditions. What should we conclude about our faculty of perception? In conditions known to psychologists, a normal subject would be systematically misled unless aided by collateral information. . . . Absent collateral information, human perception falls far short of epistemic perfection. Moreover, the flaws are not just accidental and occasional; they are systematic and widely shared.²⁷

It would be premature to abandon perception or introspection in philosophical or scientific theory; it is no less premature, on Sosa’s view, to abandon intuition. Third, the kinds of intuition relied on by conceptual analysis—rational intuition and shared intuitions conditioned by and expressing shared linguistic practices—are far more reliable than other kinds and are certainly not systematically unreliable. Although rational intuition has made mistakes, this shows only that rational intuition is not infallible; and this will not justify abandoning rational intuition in philosophical theorizing. While Descartes took the position that one is epistemically justified in believing X if and only if one cannot be mistaken in believing X, the dominant position among epistemologists is that certainty of this kind is simply not possible. Moreover, while the relationships between our shared views (or intuitions) about concepts and our shared practices is a dynamic one of mutual adjustment, our shared intuitions about concepts ultimately define and are influenced by our shared practices for using the associated concept-terms. We can trust our shared intuition that only men are bachelors because they contribute to defining ‘bachelor’ to apply only to men and to maintaining this shared linguistic practice. ²⁶ G Bealer, ‘Intuition and the Autonomy of Philosophy’ in DePaul, Michael and Ramsey, William (eds), Rethinking Intuition: The Psychology of Intuition and Its Role in Philosophical Inquiry (Rowman & Littlefield, 1998) 207. ²⁷ E Sosa, ‘Minimal Intuition’ in DePaul, Michael and Ramsey, William (eds), Rethinking Intuition: The Psychology of Intuition and Its Role in Philosophical Inquiry (Rowman & Littlefield, 1998) 260.

Reconsidering a Dogma

23

More, then, is needed to justify rejecting a philosophical methodology that depends on intuition in favour of a methodology continuous with science. It makes a difference what kinds of intuitions we are talking about and what kind of theorizing we are doing. If we are theorizing about content associated with words and this content is defined by core elements of social practices consciously understood by competent speakers, then our shared intuitions about the meanings of those words can be trusted. If we rely on rational intuitions about valid logical inferences, then the critiques of intuition provide no reason to abandon those intuitions (especially given that every area of theorizing relies on those inferences and presupposes they are justified). That rational intuition is fallible is not enough to reject rational intuition; after all, infallibility is not to be found anywhere else in our thinking—with the possible exception of reports of the contents of certain subjective experience (eg ‘I think I am seeing red’). What is needed is a showing that a methodology based on intuition is more likely to make mistakes than its proposed replacement—which, according to naturalists, would be scientific methodology. In the case of conceptual analysis, the naturalist would have to provide a positive account of a methodology that does not rely on either form of intuition and show it is more trustworthy than the methodology of TCA, which relies on such intuitions. Such a showing is not very likely. Common opinions, of course, are notoriously unreliable but this is a different category than the intuitions that go into TCA. If rational intuition has sometimes made mistakes, so has science. Views about the nature of space were not just dictated by intuition; they were also believed on the strength of an emerging scientific methodology. The same is true of mistakes in other areas of science, including chemistry, biology, medicine, etc. Although scientific methodology frequently results in corrections to scientific views, so does putatively a priori inquiry when properly responsive to conflicting evidence. And, again, the examples of mistaken rational intuitions are comparatively few. There is reason to think that rational intuition is sufficiently reliable to ground theorizing in other areas of philosophy than conceptual analysis. As Bealer puts the point: [M]ore sustained empirical studies might uncover evidence that a subject’s [rational] intuitions can be fallible in a more holistic way; we already know that the theoretical output of logicians, mathematicians, and philosophers working in isolation can be flawed. But these negative facts pale by comparison with a positive fact, namely, the on-balance agreement of elementary concrete-case intuitions among human subjects. Indeed, the on-balance agreement among our elementary concrete-case intuitions is one of the most impressive general facts about human cognition (Bealer (1998) 213–4).

But even if it could be shown that a replacement methodology is more reliable than intuition, we still lack adequate reason to abandon intuition in philosophical theorizing because it is not possible to do so. Philosophical argument must

24

Kenneth Einar Himma

start from somewhere, and the best anyone can do is to start with premises widely believed among one’s intended audience because they are intuitively obvious. Even the challenge to intuition relies on an epistemic intuition that is assumed rather than defended—namely, that philosophers should not appeal to intuition because it is fallible and not properly grounded in theory. Here it is worth noting that naturalists continue to rely on intuitions. Naturalist Hilary Kornblith, for example, acknowledges not only that he relies on intuition in philosophical theorizing, but also that he sees no way to do philosophy without appealing to intuitions: For some philosophers, the status of appeals to intuition is easily explained. Our intuitions are a priori justified. They are a product of conceptual analysis, and little more needs to be said. As a naturalist, however, I find this explanation unacceptable rejecting, as I do, both the a priori and the very idea of conceptual analysis, as they are fundamentally conceived. This would raise no difficulties for me were it not for the fact that I too have appealed to my intuitions in the course of theory construction, and this is a part of my philosophical practice that I do not believe I can eliminate.²⁸

Whether this is inconsistent with naturalism is disputed in the literature, but is not important for our purposes. What is important is the idea that we cannot theorize without appealing to intuition. If this is correct, then we simply cannot abandon intuition in philosophical theorizing. The need to rely on intuitions is not unique to philosophy. All theorizing appears to be grounded in premises that are assumed because they are intuitively obvious. Any theory of pure mathematics, for example, starts from axioms that must be assumed. Although mathematicians occasionally study mathematical systems without regard to whether they model our world, most systems are grounded in axioms that are assumed because they are intuitively obvious. Likewise, scientific theorizing is grounded in the assumptions that (1) there is an external world; (2) it is governed by causal regularities describable by laws; and (3) past observations about causal transactions are a reliable guide to future causal transactions. There are, of course, pragmatic grounds for these assumptions; we cannot begin to justify causal assertions without them. But part of the reason for these assumptions is that they all seem intuitively obvious. The general point about intuitions, then, will not justify a demand to utterly reject intuition in philosophical (or any other kind of) theorizing. If we take the view that theorizing is not appropriately grounded in undefended commonly accepted premises, then we will not be able to get any theorizing off the ground— including philosophical theorizing that is ‘continuous with scientific theorizing’; again, even scientific theorizing is grounded in undefended assumptions about the world and about methodology. ²⁸ H Kornblith, ‘The Role of Intuition in Philosophical Inquiry: An Account with No Unnatural Ingredients’ in DePaul and Ramsey (eds), Rethinking Intuition n 27, 129; emphasis added.

Reconsidering a Dogma

25

C. The Relevance of Different Descriptions of Methodology It is easy to see why one might think the argument from intuition might cause problems for an immodest conceptual analysis claiming to uncover mind-independent timeless truths about the universe from intuitions about possible cases. If TCA requires us to be able to intuit the metaphysical, mind-independent, essences of things, it is hard to see how intuition could do that—much less do it reliably. What sort of a faculty would intuition have to be to grasp utterly mindindependent essences of things? If TCA requires an intuitive faculty that can identify the objectively true metaphysical nature of things, then that is a plausible reason to think that TCA is not a reliable way of coming to understand concepts. But TCA requires nothing like this: TCA purports to analyse the metaphysical commitments implied by our shared linguistic practices and hence starts from intuitions that are conditioned by, and define the content of, these practices. Again, concepts might be abstract objects independent of our social activities, but language is a social construct grounded in social practices: which concept a word picks out is determined by the content of those social practices. That our concepts are grounded in linguistic practices does some important justificatory work. If one is to defend trusting a class of fallible intuitions, it is not enough to point out that we do, and should, trust other classes of fallible mental states. We need a reason to think that the relevant class of intuitions, while fallible, is sufficiently reliable to ground TCA. At the risk of belabouring the point, the shared linguistic basis for our concepts provides such a reason. Insofar as our views about concepts are conditioned by the shared practices that contribute to fi xing the content of those concepts, our shared views about the content of a concept, unlike other intuitions, have a special epistemic status. It is our shared views about the application of a concept that constitute our practices governing its usage. If speakers converge in using ‘bachelor’ to apply only to men, then the shared view that only men can be bachelors constitutes one of the core elements of our linguistic practices for using ‘bachelor’—and this view is accepted by the community of speakers as a norm governing use of the term. The content of our concepts is a mind-dependent feature of reality defined by our shared practices regarding words. Our shared intuitions about the obvious features of the practices in which we participate are infallible in this sense: if the intuition is genuinely shared and hence expressed in the practice, then the content of the practice must reflect that intuition. Our shared intuitions about our shared linguistic practice have a special justificatory status: they are reliable because they fi x, within limits, the conceptual content we take our words to express. Thus, the attack on intuition cannot vitiate TCA. Intuitions about our concepts that shape or are shaped by our core practices for using the associated terms have a special reliability because the contents of our concepts are fi xed by those

Kenneth Einar Himma

26

core practices. If a particular community converges, for example, in assigning ‘law’ to all and only systems with certain features regardless of whether the content of the law is just, then the intuition that an unjust system with those features is a system of law enjoys a special epistemic status because it is the practice of that community that determines whether something counts as a system of law. Conceptual analysis, on this conception, is concerned with analysing the concepts of a particular community, and those concepts are partly fi xed by the practices that inform their use of the relevant terms. It is true, of course, that there are other elements to the methodology of TCA that rely on intuition, but these are, again, comparatively unproblematic. TCA relies, for example, on the assumption that we can justifiably make valid inferences, which might depend for its justification on rational intuition. If so, TCA is no worse off for its reliance on such an assumption than any other area of theorizing utilizing these deductive rules of inference. Indeed, it is just not clear how any sort of theorizing is possible without relying on such inferences—and implicitly on the assumption that these inferences are justified, whether by rational intuition or some other theory.

IV. The Revisability of All Claims in the Face of Recalcitrant Experience A. The Argument Stated The most influential of arguments against TCA is grounded in the claim that all statements are subject to being revised in the face of ‘recalcitrant experience’. Quine argues as follows: [I]t becomes folly to seek a boundary between synthetic statements, which hold contingently on experience, and analytic statements which hold come what may. . . . [N]o statement is immune to revision. Revision even of the logical law of the excluded middle has been proposed [to] simplify[] quantum mechanics; and what difference is there in principle between such a shift and the shift whereby Kepler superseded Ptolemy, or Einstein Newton, or Darwin Aristotle?

But if any statement can be revised in principle in the face of recalcitrant experience, then no statement can be necessarily true (or true ‘come what may’) and hence no statement can be analytically true. TCA is impossible, again, because there are no analytic statements.²⁹ ²⁹ It is worth noting here that synthetic statements are not fairly characterized as those that hold contingently on experience. Indeed, Kant initially defined them as statements in which the predicate is not contained in the subject and thought that metaphysical claims were both synthetic and a priori. While there is an issue here because analytic statements are thought to be justified a priori and necessarily true, Quine mischaracterizes it.

Reconsidering a Dogma

27

In support of Quine’s argument, philosophers have offered actual and possible empirical observations that have led us, or would lead us, to revise a putatively analytic statement. The theory of relativity caused us to reject the Euclidean view that there is only one line passing through a point off a line L parallel to L. Hilary Putnam argues we can imagine giving up the claim that all cats are animals if we discovered cats were robots controlled from Mars. Gilbert Harman points out that people do not characterize as a ‘bachelor’ either the Pope or someone who has been cohabiting with a woman for a long time.

B. Mathematical Truths, Analyticity, and A Prioricity To begin, mathematical and logical truths, even if necessarily true, are not true simply in virtue of the meanings of the terms. Intuitively, there is a significant difference between the claims that all bachelors are unmarried and that 2+2=4. Someone who denies it is a necessary truth that 2+2=4 might be confused but not necessarily about the meanings of the terms; indeed, some philosophers have thought an omnipotent God could bring it about that 2+2=5 even if we cannot understand how. Someone who thinks a book could be a bachelor is, in contrast, confused about the meanings of ‘book’ or ‘bachelor’ or both. Mathematical truths might depend on the meanings of the terms, but they also depend on other (necessary, according to the standard view) features of the world; this is not true of conceptual claims or analytic statements.³⁰ The same is thought to be true of logical truths. The truth of the law of noncontradiction (ie a sentence and its negation cannot both be true) is true in virtue of features of this world that every possible world has, and not merely in virtue of the meaning of the terms. Some postmodern philosophers deny the law of noncontradiction is a necessary truth; they are probably incorrect, but their view is surely not grounded in confusion about the meanings of the terms—as would be the case if they asserted that some books are bachelors. It is unsurprising we make mistakes about what we take to be necessary truths. Our inability to conceive of how space can be curved does not entail it is logically or metaphysically impossible for space to be curved. Our ability to imagine a teddy bear made of cotton that has thoughts and feelings (we do something like this, eg when we read Winnie the Pooh to a child) does not entail it is logically or metaphysically possible for something made entirely of cotton to be conscious. ³⁰ Th is line of reasoning goes at least as far back to Immanuel Kant: ‘We might, indeed, at first suppose that the proposition 7+5=12 is a merely analytic proposition, and follows by the principle of contradiction from the concept of a sum of 7 and 5. But if we look more closely we find that the concept of the sum of 7 and 5 contains nothing save the union of the two numbers into one, and in this no thought is being taken as to what that single number may be which combines both. The concept of 12 is by no means already thought in merely thinking this union of 7 and 5. . . . Arithmetical propositions are therefore always synthetic.’ Kant, The Critique of Pure Reason (Norman Kemp Smith, trans) (St Martin’s Press, 1965) Introduction, section V.

28

Kenneth Einar Himma

Conceivability and inconceivability are not perfectly reliable indicators of what is and is not logically or metaphysically impossible because our minds are limited in what they can conceive and fallible with respect to judgments about what is, or is not, conceivable. Still, it is important not to overstate the point. That these judgments are fallible is not tantamount to their being utterly unreliable. The latter idea would show that a claim about what we can conceive has no evidentiary force with respect to a claim about what is logically or metaphysically possible, while the former does not. Just because a doctor occasionally makes mistakes does not show that her opinion has no evidentiary value. Again, the extent to which our rational intuitions agree gives us some reason to trust them. If our intuitions about space do not entail that space cannot be curved, they certainly provided a (non-conclusive) reason to believe prior to Einstein’s theory that space cannot be curved. While this entails that the justification provided by such intuitions falls short of objective certainty, very few philosophers believe that epistemic justification precludes the possibility of mistake. In any event, the point about mathematics and logic doesn’t tell us anything about analytic truths because mathematical claims (eg about the geometry of space) and logical claims do not purport to be true wholly in virtue of the meanings of the terms or the social practices that define these meanings. An analytic truth, traditionally conceived, is true wholly in virtue of the core content of the relevant linguistic practices. That mathematical and logical views are subject to recalcitrant experience does not imply analytic claims are subject to recalcitrant experience because mathematical and logical truths are not analytic. Einstein’s theory of relativity has forced a lot of rethinking about a lot of things, but not about the content of the concept of space. We mean much the same thing by ‘space’ that earlier thinkers did but have different views about its properties. In particular, we disagree with these earlier thinkers about whether space is perfectly linear or might be curved and on how many lines there are passing through a point not on some line L that are parallel to L. These matters, however, do not appear to be purely conceptual disputes.

C. Is ‘All Cats are Animals’ Analytic? It is not entirely clear what to say about Putnam’s analysis of ‘all cats are animals’ because there are two possible responses to his thought experiment. While one could, of course, take the position that cats are not animals, one could also take the position that these Martian-controlled robots are not ‘cats’. It is crucial to note here that what we say after we discover that the things we called ‘cats’ are robots does not, by itself, decide the issue of whether ‘all cats are animals’ is, or is not, analytically true.³¹ The reason is that our response might ³¹ I am indebted to Laurence BonJour for this point.

Reconsidering a Dogma

29

involve changing the concept of cat. One could take any possible position with respect to whether these robots count as ‘cats’ and as ‘animals’ and claim, not implausibly, that it reflects some sort of change to the concept of cats. And there is nothing in TCA to preclude the idea that the content of our concepts might change in response to such cases. The content of our concepts (or our understanding) might be indeterminate with respect to issues like this that provoke some linguistic change in response. Assuming one’s response involves no such change, its content determines whether one thinks it is an analytic truth that cats are animals. If one takes the robots not to be cats, then one is assuming it is at least implicit in the core practices for using ‘cat’ that cats are animals and hence that ‘cats are animals’ is true wholly in virtue of these practices. If, however, one takes the robots to be cats, then one is not assuming that all cats are animals is implicit in the concept of cats and hence that ‘cats are animals’ is not analytically true. Although both responses are reasonably plausible, I think Putnam shows ‘cats are animals’ is not analytic because his thought experiment indicates what is central to our practices regarding ‘cat’—and not because it suggests there are no analytic truths whatsoever. While we might think that cats are necessarily animals, the idea that either of these responses are conceptually plausible shows this is not part of our ordinary conceptions about what ‘cat’ means.

D. The Analytic Truth about Bachelors The examples involving the concept of bachelor tell us something about the concept of bachelor but nothing that calls into question the distinction between analytic and synthetic truths. The fact that most people would not characterize either the Pope or a man cohabiting with a woman as a ‘bachelor’ tells us only that the concept of bachelor cannot be fully defined as ‘unmarried adult male’. While being an unmarried adult male might be a necessary condition for being a bachelor, it is not a sufficient condition. Apparently, the term ‘bachelor’ connotes that one regards oneself (or is regarded by others) as being eligible in some practical sense for marriage, and neither the Pope nor a man cohabiting with a woman is ‘eligible’ in this sense. Of course, it is not clear exactly what is involved in this sense of eligibility. It is probably not exhausted by considerations having to do with a man’s willingness to entertain relationships that he believes might lead to marriage; a gay man in the US, after all, is probably not fairly characterized as a ‘bachelor’ because a man cannot legally marry another man at this point in time—even if he would be willing to marry his partner if and where possible. But these are clearly conceptual issues and claims. The Pope is not a ‘bachelor’ because some such notion of eligibility is part of the core of what we mean when we use the term. Working out the relevant notion of eligibility will require consideration of how people understand the use of ‘bachelor.’ And one way of

30

Kenneth Einar Himma

ascertaining this is to test a person’s intuitions against a variety of possible cases (or taking a poll on such matters)—such as the case of whether a gay man who would marry another man if legally permitted counts as a ‘bachelor.’ Indeed, it is not clear what type of empirical observations explicitly concerned with the way in which people use the term ‘bachelor’ would lead one to reject that being unmarried is a necessary condition for being a bachelor. One might think, as Harman does, because some society pages characterize men soon to be divorced as ‘bachelors’, it is not a necessary condition for being a bachelor that one is unmarried (though the more plausible response is that those pages are, strictly speaking, misusing the term ‘bachelor’). But this is ultimately a point about our shared practices for using ‘bachelor’.

E. The Role of Experience in Revising or Falsifying Conceptual Claims Properly Understood Falsifying a conceptual claim might involve an empirical element, but these elements are different in a theoretically important sense from the elements involved in falsifying other claims. A set of empirical observations falsifying the claim that some car averages 25 miles per gallon does not depend—at least not primarily—on a claim about how people use the relevant terms. Everyone agrees about what counts as the relevant car and about what counts as averaging 25 miles per gallon. A set of empirical observations falsifying the claim that all unmarried males are bachelors focuses on how people use ‘bachelor’. While this is an empirical matter, it involves a different type of empirical issue than claims about how many miles per gallon a particular car averages. Thus, even if Quine is right in thinking all claims are subject to being revised in the face of experience, this does not imply there is not anything distinctive about conceptual claims or the methodology of conceptual analysis. Insofar as conceptual claims are grounded in empirical claims about the core practices for using the terms, there is an undeniably empirical element to conceptual analysis. But this element is not generally relevant in scientific methodology. Hard sciences like physics, chemistry, and biology will frequently have occasion to define terms as part of designing an experiment and hypothesis and perhaps these definitions occasionally resemble the ordinary ones (though, more likely, the terms are given operational and stipulated definitions or technical definitions shared by the scientific community). But, unlike TCA, the hard sciences are never concerned with identifying any particular set of linguistic practices or intuitions shared across the entire community of ordinary speakers—which is, of course, the starting point for TCA. While soft sciences like sociology are frequently concerned with what people believe, the rationale is not to try to make precise the content of some concept—which is the ultimate objective of TCA. Of course, the methodologies for TCA and science share some common elements. If deductive reasoning is not as prominent in scientific reasoning as it is

Reconsidering a Dogma

31

in TCA, it is also an important part of scientific theorizing; indeed, it is hard to imagine that serious theorizing on any topic is possible without making some deductive inferences. Although conceptual and scientific methodology might have some elements in common, they are fundamentally different enterprises—as anyone who is engaged in either science or metaphysics can tell you.

IV. Conceptual Analysis and Necessity Another related line of argument against TCA focuses on the apparently contingent character of putatively conceptual claims. If all claims, including putatively conceptual claims, are subject to being revised in the face of experience, then conceptual claims are not only empirical in character but contingent as well. Since the methodology of TCA falsely presupposes conceptual truths are necessary, it should be rejected in favour of the scientific methodology that is more reliable in justifying contingent claims about the world. It is true that conceptual claims are subject to being revised, but again there is nothing surprising about this. Meanings result from contingent social practices. It is no more surprising that our social practices regarding words might change than it is that any other social practice might change. Insofar as the content of our concepts supervene on our linguistic practices, what we take to be conceptually true can change as the underlying practices change. For example, while the claim that all actors are male was once conceptually true, it is now false because people stopped using ‘actress’ to refer to females and started using ‘actor’ to refer to female actors. While conceptual truths depend on contingent social practices, it does not follow that conceptual claims are contingent in the same sense that ordinary empirical claims are. The fact that ‘Ken Himma weighs 165 pounds at this moment’ expresses a true proposition depends on two different kinds of contingent fact. First, the claim expressed by the above sentence would be false if the right contingent linguistic practice were different; if, for example, ‘Ken Himma’ was used to refer to some other object, like a small poodle, the above sentence would express a claim that is false. Second, the claim expressed by the sentence would be false if another contingent fact—one having nothing to do with our contingent linguistic practices using words—were different; if, for example, I had eaten several candy bars a day over the last year without doing anything else differently, I would weigh more than 165 pounds at this moment. The first way in which contingent facts explain why the above sentence expresses a true proposition is uninteresting from a theoretical standpoint because whether any sentence expresses a true proposition depends on such contingent facts about our linguistic practices: if ‘2’ referred to 5 instead of 2, the claim expressed by sentence ‘2+2=4’ would be false. This does not tell us anything about whether, given the existing use of the relevant terms, the proposition expressed by ‘2+2=4’ is necessarily true. In contrast, it is not true that whether any sentence expresses

32

Kenneth Einar Himma

a true proposition depends on contingent facts that have nothing to do with our linguistic practices; ‘2+2=4’, if it expresses a true statement (given existing practices), it is not obvious that this statement is true only as a contingent matter. Intuitively, the truth of this proposition does not seem to have any contingent empirical facts as truth-makers. There might be circumstances in which we would give up this claim, but this is primarily an epistemological point—and not a metaphysical one. That there are possible circumstances in which we would give this claim up tells us only that we cannot be objectively certain that this claim is true: while we might feel certain, our belief that 2+2=4 could be mistaken for all we know. But this is an epistemological point and has no metaphysical implications about whether the claim is necessarily true. It might be true that we are right in thinking that 2+2=4 is necessarily true—even though we are not objectively certain of this. This is not to deny that some epistemological claims might have metaphysical implications. That a perfectly intelligent being cannot be certain about whether 2+2=4 implies there is no logically or metaphysically possible world in which a being could achieve objective certainty about this proposition—a metaphysical implication. In contrast, the claim that we cannot be certain is a claim about us and our contingent abilities and epistemic position. If the stronger claim entails ‘2+2=4’ is not a necessary truth, the weaker claim does not. Claims specifically about our contingent limitations do not have such metaphysical implications. Conceptual claims are contingent only in the uninteresting sense that the social practices regarding use of a word might have been different. The claim that every bachelor is unmarried might someday be revised, but its revision will be the result of either changes in the way we use the term or deeper insights about the use of the concepts—exactly the ones that TCA purports to provide. Unlike claims about weight subject to mind-independent features of the world (once we get beyond the relevant linguistic practices), the claim that every bachelor is unmarried will not be revisable exclusively by reference to some mind-independent feature of the world. If we someday decide that some bachelors are married, it will require a change in the linguistic practices for using the word ‘bachelor’ and hence will change a norm for using ‘bachelor’ that we once accepted and followed. This distinguishes conceptual claims from claims about how much I weigh: giving up the claim that I weigh 165 pounds at this moment requires no change in the social practices for applying the relevant concepts. This suggests that conceptual claims (at least claims that are specifically about our concepts) are conditionally necessary in the following sense. If core elements in our linguistic practices for using the word ‘bachelor’ require that the word applies only to men, then it is a conceptually necessary truth that only men are bachelors. While it is true these elements of our practices might change, the existing practices commit us to a norm that entails the conceptual impossibility of a married bachelor for as long as these conventions are accepted and practised.

Reconsidering a Dogma

33

Someone who understands and practises these conventions would be contradicting herself were she to assert that some bachelors are married. Ultimately, this is because any conceptual framework that a community of speakers employs to understand the world comes with certain metaphysical commitments. But because these commitments are contingent upon social practices that might change over time, the proposition a sentence expresses, which may change with changes in usage, is fairly characterized as necessary only for as long as we employ a particular conceptual framework. Again, that particular proposition might nonetheless be necessarily true, but it would not be expressed after a change in relevant usage by the same sentence. The reality we seek to understand through philosophical speculation is, in part, constructed by the conceptual frameworks we impose on the world. Consider, for example, how conceptual truths about bachelors depend on our practices. As long as we use ‘bachelor’, as a matter of practice, to refer to all and only unmarried men, existing and possible, eligible (in the right way) to marry, it is a necessary truth that all bachelors are unmarried men eligible to marry. If we started using ‘bachelors’ to refer to rocks, then the proposition expressed by ‘all bachelors are unmarried men’ would not be necessarily true; that proposition would assert that rocks are men—and this is false. It might be true that the proposition expressed at this time, given our current practices, by ‘all bachelors are unmarried men’ would still be a necessary truth; perhaps that proposition is a timeless, absolute metaphysical truth. But that proposition would not be expressed, if our practice is to use ‘bachelor’ to refer to rocks, by ‘all bachelors are unmarried men’. That sentence, under those practices, would express a different proposition than the proposition it picks out under our existing practices; and the latter would presumably remain true. What conceptual truths our words express depend on social practices that define them, which are contingent and revisable. This is not to suggest that all conceptual frameworks are equally respectable or to deny that the world has structure independent of our commitments that is accurately represented by our own conceptual framework. But it is to suggest that the world we have come to understand is partly the result of our own social activities and that TCA helps us to understand that world. The world we come to understand is partly constructed by our own cognitive activities and linguistic practices and cannot be fully understood without understanding the deeper commitments those activities and practices entail or presuppose. TCA is essential to being able to understand those commitments. Indeed, as Quine realized, the viability of any conceptual framework is defined in part by its utility in enabling us to understand other features of the world. As long as a particular framework enables us to make sense of the world we experience through our senses, we have a reason to employ it. If we find that a framework cannot help us make sense of the world or, worse, makes the world harder to understand, we might have a reason to modify or replace it. The property expressed by ‘blue’ is useful in understanding the sensible world; a word that expresses the

34

Kenneth Einar Himma

property of being blue or evil is not (except in some narrow philosophical contexts); and this probably explains why we do not have a word that expresses such a property. But, again, changes to our frameworks have to be understood as changes to our social practices using the words—and involve changes to linguistic norms. As long as the social practices remain constant, they entail that the relationship between the thing to which the term refers and the property that we self-consciously attempt to refer to by use of that term is necessary. Such truths, then, are conditionally necessary—necessary for as long as we use the terms in that way—though, again, the proposition expressed by ‘all bachelors are unmarried’ might be necessarily true even if we use ‘bachelor’ to refer to rocks; in that case, however, our employment of the sentence ‘all bachelors are unmarried’ would not express that proposition.

V. Summary and Afterword In this essay, I have argued that the most influential arguments for abandoning metaphysics and conceptual analysis are unsuccessful. The Quinean circle-ofterms argument, for example, does not even get off the ground because his claim we should give up the concept of meaning because of its obscurity cannot bear the weight Quine places on it. Likewise, sociological arguments about the unreliability of intuition fail because they neither distinguish various forms of intuition nor consider the possibility that some forms are more reliable than others; more importantly, they fail to consider that it is just not possible to theorize without simply assuming propositions that are likely to be widely accepted because they are intuitively obvious. None of this, however, says anything about problems involved in articulating a naturalist methodology for conceptual analysis that goes beyond the lexicography that naturalists dismiss as lacking philosophical value. One example might help to illustrate the problem. Leiter argues that sociological observation of the legal practices of judges and legislators is the proper methodology for analysing the nature of law. Of course, sociologists cannot make such observations without having some sort of working conception of the concept of law. On Leiter’s view, however, nothing more is needed than the sort of lexical definition given in a dictionary. But this will not suffice to address the deepest conceptual issues involving law, and we can look to Leiter’s views to see why. Leiter argues that ‘the leading’ sociological studies show that judges in legal systems like ours treat as ‘law’ only properly pedigreed norms that the exclusive legal positivists believe, as a conceptual matter, exhaust the content of law. On the assumption that these studies are correct, Leiter infers that the proper analysis of law is given by exclusive legal positivism, rather than inclusive legal positivism. As Leiter puts it: Supposing that these models are ultimately vindicated empirically—and not just for American courts—this would give the Replacement Naturalist reason to abandon any

Reconsidering a Dogma

35

a priori, intuitive confidence we had about the concept of law that conflicted with Hard [ie exclusive] Positivism—just as the role of non-Euclidean geometry in parts of physics has led everyone to repudiate Kant’s a priori confidence about the Euclidean structure of space. If social science cuts the causal joints of the legal world in Hard Positivist terms, the Replacement Naturalist argues, that is a compelling reason to work with that concept of law as against its competitors.³²

This argument, however, overlooks a distinction between the Euclidean claim falsified by the theory of relativity and the claim made by inclusive (or soft) positivists. Euclid’s parallel postulate was taken to be a necessary truth about space in every logically possible world; if space in the actual world was not Euclidean in character, then there is a possible world in which Euclid’s parallel postulate is false; and it follows that this postulate, given that it purports to be necessarily true, is false. Inclusive positivism claims there is a possible legal system in which morality is a condition of legality and denies the exclusive positivist claim that it is necessarily true that all law is source-based. The observation that a finite number of existing legal systems, no matter how large, do not incorporate morality as a criterion of legality cannot refute the claim that there is a conceptually possible legal system in which law incorporates morality as a criterion of legality. That the actual world conflicts with a claim it is necessarily true that the world has some feature f suffices to refute that claim; that the actual world coheres with a claim that it is necessary that the world has f does nothing to confirm the claim. Leiter could respond in Quinean fashion by denying there are any necessary truths, but there is nothing in the Quinean arguments, as we have seen, that compel us to accept that counterintuitive conclusion. Say what you will about the nature of space, it is hard to imagine that 2+2 could equal anything other than 4. But I can’t try to defend the notion of necessity here. For our purposes, the point is this: the naturalist doesn’t have anything that would count as a viable methodology for analysing concepts that goes beyond the ‘glorified lexicography’ naturalists have wrongly attributed to TCA and disparaged. It is not enough to articulate even plausible arguments for thinking that notions like analyticity or necessity are mysterious to successfully make a case for the methodological naturalism endorsed by theorists like Leiter; some sort of articulated methodology plausibly thought capable of solving the problems addressed by the methodology it is supposed to replace is needed. This has yet to be done.

³² B Leiter, ‘Naturalism in Legal Philosophy’ in E Zalta (ed), Stanford Encyclopedia of Philosophy; available at .

2 Six Paths to Vertigo-free Legal Theory Sylvie Delacroix

Legal theorists are no more prone to vertigo than the general population. A lot of them, however, dread a particular kind of vertigo, commonly striking those who, looking for some solid ground on which to base their evaluative outlook, find nothing but a congruence of seemingly random subjectivities. If there is nothing to our moral commitments but a contingent balance of interests and desires, how can the law ever stand on firm ground when it has to refer to such commitments in order to settle controversial cases? One may, of course, account for law in such a way as not to associate it with any other ‘end’ than a strictly prudential one—such as enabling the peaceful coexistence of individuals. Taking a stand on controversial moral issues, from this perspective, need not be any less contingent than the interests that contributed to shaping the moral issue in the first place. Yet such a minimalist, strictly prudential reading of law’s ambitions fails, in most cases, to echo any of our common perceptions of what a legal system is about. It is precisely because a legal system is typically perceived as more than just a disciplinary tool that we may feel the ground slipping from under our feet when confronted with the impossibility of ‘founding’ our evaluative commitments on anything but our own intersubjective experiences. No one expressed this vertiginous feeling more candidly than Montaigne. His critique of the natural law model was inaugural not only in the challenge it raised but also, most remarkably, in the strategy deployed to recoil from it. The following paragraphs offer six ways of avoiding the vertigo described above, starting with Montaigne’s ‘mystical move’.

I. Move it like Montaigne In the 16th century, Montaigne’s critique of the natural law model left him with the sour picture of ‘laws often made by fools [. . .] but always by men, vain authorities who can resolve nothing.’¹ What is it that needs to be ‘resolved’? ¹ M de Montaigne, The Complete Essays (trans MA Screech) (Penguin Classics, 1991) III 13, 1216 (emphasis mine).

Six Paths to Vertigo-free Legal Theory. Sylvie Delacroix. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Six Paths to Vertigo-free Legal Theory

37

Beyond the foolishness of legislators, the second element contributing to making the task of legislators difficult amounts to their being ‘always between birth and death’,² thus subject to finiteness, which does not agree easily with the purported permanence of their task. If men (and women) are the exclusive authors, without any other guarantor, of their own laws, they can undo what they have done. What kind of authority would these laws have, which would be at the mercy of a change of will from their subjects (and authors)?³ If these ‘authors’ are unable to bind their will to anything but ‘unstable weak opinions’,⁴ is there any hope of ever relating the purported ‘grandeur’ of the law—its being meant to morally bind us—to those hopelessly contingent beginnings? To spare his ‘lay’ contemporaries the ‘perilous’ discovery of law’s precarious origins,⁵ Montaigne enjoins them to obey the law ‘because it is the law’ without asking any further question: ‘Now laws remain respected not because they are just but because they are laws. That is the mystical basis of their authority.’⁶ If only from a terminological perspective, such a statement is exceptional. It is not only the sole occurrence of the term ‘mystical’ in the works of Montaigne, but the term ‘mystical’ as such seems to be rare at his time.⁷ Originally, the Greek term ‘mustikos’ meant ‘relating to mysteries’, referring more particularly to the religious domain accessible only to the initiated. From the time of Montaigne, this meaning extended to the vaguer idea of a mystical ‘discourse’, expressing the limit between a hidden and ineffable essence and the endless description of the visible.⁸ Drawing the line between what can be inquired into and what has to be kept at a distance, surrounded by mysteries, the ‘mystical basis’ of law’s authority lies in a ‘law of pure obedience’: ‘The first commandment which God ever gave to

² ‘We have no communication with Being; as human nature is wholly situated, for ever, between birth and death, it shows itself only as a dark shadowy appearance, an unstable weak opinion.’ (M de Montaigne, The Complete Essays (trans MA Screech) (Penguin Classics, 1991) II 12, 680). ³ See GA Cohen, ‘Reason, Humanity and the Moral Law’ in Korsgaard (ed), The Sources of Normativity (Cambridge University Press, 1996) 167–88, 167. ⁴ See n 2 above for full quotation. ⁵ ‘Laws gain their authority from actual possession and custom: it is perilous to go back to their origins; laws, like our rivers, get greater and nobler as they roll along: follow them back upstream to their sources and all you find is a tiny spring, hardly recognizable’ (M de Montaigne, The Complete Essays (trans MA Screech) (Penguin Classics, 1991) II 12, 658, emphasis mine). One way of reacting to the discovery of the ‘tiny spring’ amounts to rejecting all the pretence that is supposed to magnify the law and make it grow in stature (custom and usage), thus refusing to take part in the masquerade that consists in inflating what is—and what should remain—only a tiny spring. According to Montaigne, this attitude can only lead to trouble, as ‘those who refuse to be drawn away from the beginnings and sources fail even worse [than those dons who do not even dare to handle them] and find themselves bound to savage opinions’ (ibid I 23, 131). ⁶ ibid III 13, 1216. ⁷ For a synthesis of the history of the term ‘mystical’, see T Berns, Violence de la Loi la Renaissance: l’Originaire Duapolitique chez Machiavel et Montaigne (Paris, 2000) 386–89. ⁸ See the brilliant analysis of Michel de Certeau in M de Certeau, La Faible Mystique vol 1 (Gallimard, 1982) 132 (for the English translation, see M de Certeau, The Mystic Fable vol 1 (University of Chicago Press, 1992)).

38

Sylvie Delacroix

Man [. . .] the law of pure obedience. It was a bare and simple order, leaving man no room for knowing or arguing.’⁹ Is this ‘first and ultimate law’ surreptitiously restoring the law to the kind of ontological grounding whose inaccessible and thus illusory character Montaigne emphatically denounced? In a way similar to Rousseau’s ‘it would require Gods to give men laws’¹⁰ or Kant’s ‘all authority is from God’,¹¹ Montaigne’s ‘law of pure obedience’ can be understood as a theoretical device (a ‘legal fiction’¹²) meant to supplement the otherwise unthinkable authority of law.¹³ Standing where the theoretical discourse has to be brought to an end,¹⁴ it is above all meant to conjure the vertiginous unveiling of law’s empirical sources.

II. Acknowledge the Natural Some people would no doubt argue that Montaigne need not have run into so much trouble in the first place. His critique of the natural law model is based on a simplistic, ‘arithmetic’ understanding of natural laws which does not even start to tackle the complexities of the natural law tradition: ‘Philosophers can hardly be serious when they try to introduce certainty into Law by asserting that there are so-called Natural Laws, perpetual and immutable, whose essential characteristic consists in their being imprinted upon the human race. There are said to be three such laws; or four; some say less, some say more: a sign that the mark they bear is as dubious as all the rest.’¹⁵ If, instead of dramatizing the ‘localization’ of the ‘Good’ or the ‘truth’,¹⁶ Montaigne had seen customary variations as different, contextualized instantiations of human flourishing, he might not have had to endure quite the same ⁹ M de Montaigne, The Complete Essays (trans MA Screech) (Penguin Classics, 1991) II 12, 543 (emphasis mine). ¹⁰ J-J Rousseau, The Social Contract and Other Later Political Writings (Cambridge University Press, 1997) II 7, 69. ¹¹ ‘A law that is so holy (inviolable) that it is already a crime even to call it in doubt in a practical way, and so to suspend its effect for a moment, is thought as if it must have arisen not from men but from some highest, flawless lawgiver ; and that is what the saying “all authority is from God” means. This saying is not an assertion about the historical basis of the civil constitution; it instead sets forth an Idea as a practical principle of reason’ (I Kant, The Metaphysics of Morals (Cambridge University Press, 1991) §49 A, 130, emphasis mine). ¹² ‘even our system of Law, they say, bases the truth of its justice upon legal fictions.’ (M de Montaigne, The Complete Essays (trans MA Screech) (Penguin Classics, 1991) II 12, 603). ¹³ For a full elaboration of this argument, S Delacroix, ‘Montaigne’s Inquiry into the Sources of Normativity’ (2003) 16 Canadian Journal of Law and Jurisprudence 271–86. ¹⁴ In Force de loi, Derrida endeavours to explain Montaigne’s reference to the term ‘mystical’by reference to that founding point where ‘the discourse comes up against its limit’, enclosed in the ‘performativity’ of its enunciation: ‘Here a silence is walled up in the violent structure of the found ing act’ (J Derrida, ‘Force of Law: the Mystical Foundation of Authority’ in D Cornell and M Rosenfeld (eds), Deconstruction and the Possibility of Justice (Routledge, 1992) 3–67, 14). ¹⁵ M de Montaigne, The Complete Essays (trans MA Screech) (Penguin Classics, 1991) II 12, 653–4. ¹⁶ ‘What kind of Good can it be, which was honoured yesterday but not today and which becomes a crime when you cross a river! What kind of truth can be limited by a range of mountains,

Six Paths to Vertigo-free Legal Theory

39

vertigo. Yes, laws are made by fools, but we are lucky enough to also have ‘natural’ laws giving voice to a pre-given teleological order, showing us the way things are meant to be, the Aristotelian true ‘Nature’ of things. Because of our ‘fallen’ human nature we may fail, most of the time, to actualize our potential for form, but at least we can turn to Nature as ‘the finality governing completions’, ‘what is not a mere artifact of our practical reasoning.’¹⁷ So when we turn to it, what is Nature supposed to yield? It will yield ‘basic human goods’, those fundamental ‘per se nota’ reasons for action which we are bound to grasp if we engage in rational practical inquiry.¹⁸ This ‘path’ requires faith. It requires faith in a pre-given teleological order, in a value-laden concept of Nature that is ‘there to be experienced’.¹⁹ Beyond the conceptual divergences separating different trends in the natural law tradition, they all converge in providing ontological grounds for the moral propositions they assert. If one is too disenchanted to buy into this kind of metaphysics, if, as Williams, one finds that ‘messages sent to the Aristotelian essence centre are, as always, returned unopened to the sender’,²⁰ one may not succeed in warding off one’s vertigo following that particular path. There are four others.

III. Look in the Right Direction Here is the common sense remedy: just do not look below. Hart’s allegiance to JL Austin’s method, aimed at elucidating even the most complex concepts by reference to ‘the things people do with words’, predisposed him to keeping his cool when it came to accounting for law’s power to bind us. While law’s normative dimension is central to Hart’s theory, its explanation is never taken to require an inquiry into the context of social interaction which conditions and enables this normativity. Unlike Montaigne, whose despair came from having to trace law’s authority back to the contingent practices of ‘foolish legislators’, Hart keeps his attention focused downstream. His accounting for the difference between coercion becoming a lie for the world on the other side!’ (M de Montaigne, The Complete Essays (trans MA Screech) (Penguin Classics, 1991) II 12, 653–4). ¹⁷ Notice the difference between these two definitions of ‘nature’ given in opposite pages in the same paper (R Hittinger, ‘Natural Law and Virtue’ in Natural Law Theory: Contemporary Essays (Clarendon Press, 1992) 42–69, 42–3). ¹⁸ Note that, if ‘yielded’ by nature, those ‘basic human goods’ are not necessarily to be understood as deriving from a concept of human nature which would somehow precede human reasoning, as indeed ‘a complete theoretical account of human nature (unlike accounts of closed natures) would depend, therefore, on data provided by practical enquiry, reflection, and judgment.’ (RP George, ‘Natural Law and Human Nature’ in Natural Law Theory: Contemporary Essays (Clarendon Press, 1992) 31–41, 39). ¹⁹ As opposed to being a mere figment of the subjective state that purports to be an experience of it (J McDowell, ‘Non-cognitivism and Rule Following’ in Mind Value and Reality (Harvard University Press, 1998) 136. ²⁰ B Williams, ‘Relativism, History and the Existence of Values’ in RJ Wallace (ed), The Practice of Value (Clarendon Press, 2003) 106–18, 117.

40

Sylvie Delacroix

and obligation by reference to the ‘distinct normative attitude’²¹ typically associated with the use of the word ‘obligation’ presupposes law’s normative dimension. Its focus is on the ‘surface phenomena’ flowing from the fact that law is normative, not on what it takes for law to be normative in the first place. His late reference to a conventionalist framework to explain the emergence of the rule of recognition is made in passing, in a brief passage of his Postscript; ²² as if the study of the context of social interaction allowing and conditioning law’s normative dimension were unlikely to yield any significant insight as to the meaning and properties of law itself. As if, more importantly, any such inquiry into its conditions of possibility would inevitably grant the concept of normativity a metaphysical status it should not, and cannot have. For this is the worry which maybe keeps Hart from looking upstream (or ‘below’). His inscription within a philosophical context dominated by JL Austin (weary of abstract essences whose metaphysical status is supposedly independent of linguistic usage) comforted Hart in his reluctance to question what conditions law’s normativity. Given his scepticism as to the possibility of preserving the objectivity of values without adhering to some form of moral realism, a metaphysical option he was committed to rejecting, Hart had every reason to keep looking downstream.²³ Accounting for what enables law’s power to bind us without at any point referring to the moral concerns leading us to want law as a specific form of governance is indeed a tour de force which only the most hardened positivist will want to try.²⁴ Hart’s taking refuge in the belief that, to use Raz’s words: ‘the problems about the ontological standing of legal “things” such as law, rights and corporations [. . .] can be dissolved with the judicious application of speech-act theory’²⁵ allowed him to leave his meta-ethical concerns behind.²⁶ ²¹ This distinct normative attitude ‘consists in the standing disposition of individuals to take such patterns of conduct both as guides to their own future conduct and as standards of criticism which make legitimate demands and various forms of pressure for conformity’ (HLA Hart, The Concept of Law (2nd edn, Clarendon Press, 1994) 255). ²² ‘But the theory remains as a faithful account of conventional social rules which include, besides ordinary social customs (which may or may not be recognised as having legal force), certain important legal rules including the rule of recognition, which is in effect a form of judicial customary rule existing only if it is accepted and practised in the law-identifying and law-applying operations of the courts.’ (HLA Hart, The Concept of Law (2nd edn, Clarendon Press, 1994) 256). ²³ For more details on this upstream/downstream distinction and the way it applies to Hart’s outlook on normativity see S Delacroix, ‘Hart and Kelsen’s Conceptions of Normativity Contrasted’ (2004) 17 Ratio Iuris 501–20. ²⁴ For a discussion of this ‘hardened’ positivist endeavour, see section 5 below. ²⁵ J Raz, ‘Two Views of the Nature of the Theory of Law, a Partial Comparison’ in J Coleman (ed), Hart’s Postscript, Essays on the Postscript to the Concept of Law (Oxford University Press, 2001) 1–37, 5. Note that Raz does add that ‘by the time Hart published The concept of law many of these hopes had receded. But his faith in the benefits for legal analysis of learning the lessons of speechact theory is manifested in his way of understanding legal statements as statements from what he called the internal point of view.’ ²⁶ ‘Even when there is agreement on this point and certain rules or principles are accepted as indisputably belonging to morality, there may still be great philosophical disagreement as to their status or relation to the rest of human knowledge and experience. Are they immutable principles

Six Paths to Vertigo-free Legal Theory

41

This strategy only provided short-lived comfort, however, and soon enough Hart had to start worrying, like everybody else, about the status of moral norms in adjudication. As his persistent agnosticism²⁷ on the subject left many of his colleagues uncomfortable, some endeavoured to rescue him from this untenable position²⁸ by emphasising that a defence of the possibility of ethical objectivity need not entail any of the queer ontological commitments associated with a moral realist position. Having to go back to our own intersubjective experiences as the ‘fabric’ conditioning our claims to objectivity, the success of such ‘rescue attempts’ ultimately depends on shedding Montaigne’s vertigo (see section 6).

IV. Purify the Whole Mess Instead of simply not looking ‘down’, one may also endeavour to frame the reality lying ‘upstream’ from law’s normative dimension in such a way as to see only what one chooses to see. Kelsen’s attempt to build a ‘pure’ theory of law could be considered as just that. Ruling out any reference to social facts or moral content in the name of an extreme methodological dualism strictly separating is from ought,²⁹ Kelsen’s account of legal normativity hinges on a ‘transcendental’ presupposition: the basic norm. In order to interpret normatively the empirical phenomenon that presents itself as law, Kelsen presupposes a highest, ‘basic’ norm, which stands as the ultimate ‘reason of validity’ of all the other norms of the system. Crucially, this last and highest norm ‘must be presupposed, because it cannot be “posited”, that is to say: created, by an authority whose competence would have to rest on a still higher norm.’³⁰ which constitute part of the fabric of the Universe, not made by man, but awaiting discovery by the human intellect? Or are they expressions of changing human attitudes, choices, demands, or feelings? These are crude formulations of two extremes in moral philosophy. Between them lie many complicated and subtle variants, which philosophers have developed in the effort to elucidate the nature of morality. In what follows we shall seek to evade these philosophical difficulties.’(HLA Hart, The Concept of Law (2nd edn, Clarendon Press, 1994) 168). ²⁷ ‘Legal theory should avoid commitment to controversial philosophical theories of the general status of moral judgments and should leave open, as I do in this book, the general question of whether they have ‘objective standing’ (HLA Hart, The Concept of Law (2nd edn, Clarendon Press, 1994) 253–4). ²⁸ Leaving open the question of these moral principles’ objective standing indeed meant that it had to be seen as an open question too whether these moral principles could ever be deemed an integral part of a legal system’s validity criteria, or whether they should rather be considered as directions to the court to exercise their discretion in a certain way. ²⁹ Kelsen radicalized the trend set by Laband and Jellinek by supporting an expansive version of methodological dualism, separating the worlds of normativity and facticity by an ‘insuperable abyss’, corresponding to two independent spheres that are epistemologically unbridgeable. In opposition to Jellinek, Kelsen rejects the idea that a theory of law can comprise both a legal and a sociological perspective on its object. His aim is to secure the autonomy of law as an object of specifically legal cognition. ³⁰ H Kelsen, Pure Theory of Law (University of California Press, 1967), 195.

42

Sylvie Delacroix

Conveniently avoiding any reference to the ‘mess’ of human affairs,³¹ the presupposed basic norm is meant to enable Kelsen to explain law’s normativity ‘from within’. This explanation, however, never manages to overcome what Kelsen called ‘the problem of positivity’: for fear of falling into an idealism that he deems untenable,³² Kelsen cannot simply ignore the initial law-creating facts, ‘the social practice which ties the ultimate legal rule to social reality.’ ³³ Likewise, these initial law-creating facts cannot be logically linked to the norm they have created (the historically first constitution) as an ought cannot possibly derive from an is (according to Kelsen’s methodological dualism). While Kelsen initially finds a way out by making these initial law-creating facts the content of the basic norm,³⁴ he will nevertheless have to abandon that solution, and the problem of law-making will remain a source of tension throughout his work.

V. Avoid the Mess, be Pragmatic Another way of avoiding the trouble Montaigne ran into is to define carefully what is and what is not part of our ‘concept’ of law. Coleman for instance excels in the art of avoiding trouble. His account is structured around a clear-cut distinction between the project of accounting for what conditions the existence of law, on one hand, and that of ‘making intelligible law’s claim to make a practical

³¹ The systematic exclusion of any ideological or political consideration from the ‘pure’ theory may be understood as answering a situation of crisis that Kelsen couldn’t help but sense (and of which he will personally bear the brunt). What reasons would lead Kelsen to trust a political arbitrariness that, in the Weimar Republic, could only be perceived as more and more threatening? ³² ‘The first [idealistic theory of law] is wrong for it is undeniable that a legal order in its entirety, and an individual legal norm as well, lose their validity when they cease to be effective; and that a relation exists between the ought of the legal norm and the is of physical reality also insofar as the positive legal norm, to be valid, must be created by an act which exists in the reality of being’ (H Kelsen, Pure Theory of Law (University of California Press, 1967), 211). ³³ See J Raz, ‘The Purity of Pure Theory’ in SL Paulson and BL Paulson (eds), Normativity and Norms. Critical Perspectives on Kelsenian Themes (Clarendon Press, 1998) ch 12 237–52, 251. ³⁴ The reference to this ‘content’ of the Basic Norm (first appearing in Allgemeine Staatslehre) is most interesting as it brings to light a major tension Kelsen had to deal with. As he himself acknowledges—under the heading ‘the limitation of positivism’—‘the Basic Norm is not valid because it has been created in a certain way, but its validity is assumed by virtue of its content. It is valid, then, like a norm of natural law, apart from its merely hypothetical validity’ (H Kelsen, ‘Philosophical Foundations of Natural Law Theory and Legal Positivism’ in General Theory of Law and State (Harvard University Press, 1949) 389–46, 401). Kelsen is thus aware that the attribution of content to the Basic Norm makes it dangerously close to the natural law model, first target of his doublesided purity programme. The relatively confused characterization he suggests testifies, I think, to his efforts to get rid of this ambiguity: ‘the content of the Basic Norm—that is, the particular historical fact qualified by the Basic Norm as the original law-making fact—depends entirely upon the material to be taken as positive law, on the wealth of empirically given acts subjectively claiming to be legal acts.’ (ibid, 436). As these efforts prove unfruitful, Kelsen ultimately stops referring to this idea of ‘content of the basic norm’ altogether (in the second edition of the Pure theory of law).

Six Paths to Vertigo-free Legal Theory

43

difference in the lives of those to whom it applies’,³⁵ on the other hand. As such, these two explanatory targets should provide Coleman with the opportunity to develop a story with as much potential for drama as Montaigne’s. But Coleman adds an important qualification: the fact that ‘law can realize some attractive ideals’ is not necessarily part of our concept of law, just as the fact that a hammer can bang nails into a wall is not necessarily part of our concept of hammer as ‘after all, a hammer is the kind of thing that can be a murder weapon, a paperweight, or a commodity.’ ³⁶ Taken literally, Coleman’s argument consists in invoking the infinite number of potential uses an object may be put to to dismiss its aim as extraneous to our concept of that object. This leads me to think that Coleman has never had to explain the concept of ‘hammer’ to anybody.³⁷ He has, however, certainly had to explain the concept of law, and his explanation consists in a fairly simple story: given the alleged irrelevance of what we want law for, Coleman explains how a bunch of people, jointly committed to respecting a set of criteria determining what counts as law, manage to create an obligation to apply these criteria. Based on the concept of ‘shared cooperative activity’,³⁸ this explanation proceeds by sole reference to the structure of the practice constituted by the officials’ behaviour. As for the rest of the population, the law need not provide them with reasons for action, conclusive or otherwise: as long as it is capable of informing them of their various legal rights and duties, it is ‘working’.³⁹ For Coleman, the task of explaining legal normativity thus comes ³⁵ J Coleman, ‘Incorporationism, Conventionality and the Practical Difference Thesis’ in J Coleman (ed), Hart’s Postscript, Essays on the Postscript to the Concept of Law (Oxford University Press, 2001) 99–147, 114. ³⁶ ‘Let us grant that law does have the inherent potential to realize a variety of moral ideals that other forms of governance cannot realize, and that this distinguishes law from other forms of governance. Is this inherent potential really a part of our concept of law? [. . .] Law is the kind of thing that can realize some attractive ideals. That fact about law is not necessarily part of our concept of it. After all, a hammer is the kind of thing that can be a murder weapon, a paperweight, or a commodity [. . .] However, the fact that a thing, by its nature, has certain capacities or can be used for various projects does not entail that all or any of those capacities, ends, or projects are a part of our concept of that thing’ (J Coleman, The Practice of Principle: In Defence of a Pragmatist Approach to Legal Theory (Oxford University Press, 2001), 194). ³⁷ If I ever had to explain the concept of hammer to a person who has never come across it, I think I would describe it as a tool for applying maximum force at a localized point thanks to the use of high density material. And if, in the process of explaining it, Coleman were to interrupt by pointing to the fact that a hammer can be used as a murder weapon, I would treat this remark as crude provocation. ³⁸ Developed by Michael Bratman (M Bratman, ‘Shared Cooperative Activity’ 101 Philosophical Review 327–41), the concept of SCA is supposed to encapsulate three essential features of those practices whose core commitment is to do something together (like playing a quartet): mutual responsiveness, commitment to the joint activity and commitment to mutual support. By contrast with those models explaining the emergence of a normative practice by reference to convergent unilateral acceptances of the norms constituting the practice, the SCA model emphasizes the multilateral and interdependent character of the commitment at stake. ³⁹ Note the implausibility of a situation where the ‘non-official’ part of the population never even deliberates on law’s claim to impose some obligatory form of conduct: a still unlikely, but more plausible scenario, would involve a persistent and generalized failure on the part of law to provide non-officials with reasons for action that are deemed ‘strong enough’, and thus obligatory.

44

Sylvie Delacroix

down to accounting for the obligatory character of the rule of recognition for the officials. As for the obligatory character of legal rules for the rest of the population, Coleman seems to be happy to leave this issue to those who worry about meta-ethics: ‘Positivism seeks to show that the way in which law can give rise to duties is no more—and no less—mysterious than the way in which promises, pacts, reciprocal expectations, and so on can create duties. The ontology of the duties that inhabit this class of practices is not a special problem for legal theory, but is rather in the provenance of meta-ethics.’⁴⁰ In this respect, as in many others, Coleman follows in Hart’s footsteps. The philosophical context underlying his reluctance to delve into meta-ethical issues, however, has to differ from Hart’s, and instead of ‘seeking refuge’ in speech-act theory, Coleman invokes a rather vague picture of ‘pragmatism’,⁴¹ which seems to refer to some feeling of ‘worldliness and effectiveness’.⁴² Considered on its own terms, Coleman’s theory indeed ‘works’: his account of how a bunch of ‘officials’ jointly commit themselves to a set of validity criteria (thanks to the ‘SCA’ structure) ticks the boxes. It explains both the minimal conditions for the existence of a legal system and makes intelligible law’s claim to make a practical difference in the lives of those to whom it has to apply—ie ‘the officials’. One may of course ask why the officials commit themselves to that particular set of criteria, what their project is in so doing and whether, if at all, the rules that count as law actually have any impact on the ‘non-official’ part of the population. Significant as they may be, these questions—and the answers they may yield—are not, according to Coleman, a necessary part of our ‘concept’ of law. So yes, Coleman’s theory ‘works’, and it spares us any of the vertiginous feelings Montaigne had to put up with. The ‘success’ by reference to which we are meant to assess Coleman’s theory, however, entirely depends upon a fundamental, This scenario wouldn’t suit Coleman though, as it would re-introduce the necessity to include legitimacy considerations in one’s account of legal normativity. ⁴⁰ J Coleman, The Practice of Principle: In Defence of a Pragmatist Approach to Legal Theory (Oxford University Press, 2001), 159. ⁴¹ None of the major figures of philosophical pragmatism (from Peirce to Dewey) can however, be associated with quite such an insouciant attitude towards meta-ethics. A hasty reading of the pragmatist motto as formulated by William James (‘To develop a thought’s meaning we need therefore only determine what conduct it is fitted to produce; that conduct is for us its sole significance’ (W James, Essays in Philosopy (Harvard University Press, 1978), 124) may lead to the wrongheaded impression that pragmatists would agree with logical positivists in rejecting any kind of metaphysics. James, however, used this maxim to assess rival metaphysical views, and ultimately defended the possibility of cognitivism in ethics (this cognitivist position is notably made possible by the idea—fully developed by Dewey—that value judgments and judgments of facts are inextricably linked. For the relationship between moral cognitivism as an epistemological claim and moral realism as a metaphysical position, see section 6). ⁴² In a critique which is not specifically aimed at Coleman’s work, Brian Butler denounces ‘the very narrow picture of pragmatism offered among legal theorists—a picture of pragmatism reducible to feelings of effectiveness and worldliness offered from a perspective firmly entrenched within the institutional practice of law’ (BE Butler, ‘Law’s Image of Pragmatism—Another Legal Fiction’ (2004) 1 Contemporary Pragmatism 151–7, 153).

Six Paths to Vertigo-free Legal Theory

45

premise we are meant to accept as ‘given’: ‘suppose it has been agreed that for some reason a legal system is a desirable thing.’⁴³ Detached from any teleological consideration, the ‘SCA’ model indeed doesn’t do much in the way of accounting for what conditions the existence of law: structural conditions do not, on their own, ‘bring about’ anything.⁴⁴ As Coleman is, however, determined to assert the irrelevance of law’s programmatic element (the various ideals law may serve) to our ‘concept’ of law, he asks us to presuppose it. The link that can be drawn between this strategy and Montaigne’s move is interesting: because Montaigne believed that law’s authority relied on more than just its capacity to impose some disciplinary structure, it was crucial to be able to read some form of ‘objective moral commitments’ presiding over the apparent mess of law-creating practices. In the absence of so-called ‘natural laws’, Montaigne, however, could not ‘read’ anything but the foolishness of legislators. Hence the need to rescue law’s authority by stopping inquiries into law-creating practices at a ‘safe’ point—the law of pure obedience. Coleman, by contrast, proceeds in just the opposite way. Take away the belief that law’s authority typically partakes of some moral ambition, and you may peacefully confront the contingency presiding over law-creating practices (they only need to be structured around some interdependent, mutual expectations—see the ‘SCA model’).

VI. Love thy Anchor One of the striking presuppositions underlying the vertigo encountered by Montaigne (and many of its attempted remedies) amounts to the idea that the objectivity of our value judgments is somehow dependent upon there being some ‘facts of the matter’ backing up this claim to objectivity. This presupposition comes into play when one considers the various value choices that lie at the root of whatever moral ambition has been associated with one’s legal system. It may be that this moral ambition is open to diverse interpretations, arising from plural and sometimes divergent axiological choices on the part of a myriad of actors. If, however, one draws some kind of link between law’s moral ambition and its capacity to bind us,⁴⁵ the possibility of interpreting these axiological choices as somehow reaching beyond the actor’s contingent preferences becomes crucial. The mistake, in this context, is to assume that this need to ‘reach beyond’ one’s ⁴³ J Coleman, The Practice of Principle: In Defence of a Pragmatist Approach to Legal Theory (Oxford University Press, 2001), 92. ⁴⁴ For a more detailed critique of Coleman’s strategy, see S Delacroix, Legal Norms and Normativity: An Essay in Genealogy (Hart Publishing, 2006), 166–83. ⁴⁵ ‘There is a good which binding promises can serve or achieve, and that is why they can be binding [. . .] It does not follow that all promises bind [. . .] Is not the same true of law? [. . .] some specific (though possibly abstractly conceived) moral task is central to the law, essential to it being the type of institution it is [ie an institution that is meant to bind us].’ (Raz (2003) 10–11).

46

Sylvie Delacroix

contingent preferences calls for a ‘non-subjective’ interpretation of those axiological choices.⁴⁶ This particular understanding of objectivity as ‘non-subjective’has been criticized by Hilary Putnam, who associates it with a Platonic theory of ideas, or what he calls ‘Platonizing’: ‘What yields to “Platonizing” is yielding to the temptation to find mysterious entities which somehow guarantee or stand behind correct judgments of the reasonable and the unreasonable.’⁴⁷ In virtue of this widespread understanding, according to which ‘each and every instance of objectivity must be supported by objects’,⁴⁸ it is often assumed that ethical objectivism presupposes a realist position. This is the case, most importantly, of Mackie, who significantly influenced Hart on this point.⁴⁹ As we can neither demonstrate the existence of ‘values that are prior to and logically independent of [human activities]’,⁵⁰ nor show how we could have access to them, we must conclude, according to Mackie, that our belief in ethical objectivity is mistaken. This literal understanding of objectivity has been criticized not only by Putnam but also Bernard Williams, who argues: Many different things have been discussed as the question of objectivity, but they all tend either to come to nothing, or to come back to one issue: the proper understanding of ethical disagreement. Some philosophers have been very exercised, for instance, with the question whether moral judgments can be true or false. But work has to be done to find what, and how much, that question means. [. . .] I see no way of pursuing that question, which does not lead back to questions such as these: if an ethical disagreement arises, must one party think the other in error?⁵¹

This emphasis on the possibility of rationally solving our ethical disagreements as the key to ethical objectivity is particularly important for our purposes, as it shifts the focus from the truth or falsity of moral judgments to their being susceptible to rational argumentation. The possibility of entering such argumentative practice requires not only some belief in the existence of a ‘bridgehead of shared human concerns’,⁵² it also requires a minimum degree of ⁴⁶ ‘According to this, what is considered as belonging to the perceiving or thinking self is called “subjective” and what is considered as independent of the perceiving or thinking self is called in contrast “objective” [. . .] the word “objective” serves as an opposite of “subjective”. We should have to explain this use by thinking of it as connoting what might less confusingly be called “the nonsubjective”.’ (Wiggins (1996) 38–39). ⁴⁷ H Putnam, Ethics without Ontology (Harvard University Press, 2004) 70. ⁴⁸ H Putnam, Ethics without Ontology (Harvard University Press, 2004) 51. ⁴⁹ ‘In the summer of 1981, reacting in some excitement and incertitude to a weekly seminar John McDowell and I had been giving, Hart wrote me a letter strenuously defending John Mackie’s Ethics: Inventing Right and Wrong against something that I had been saying about objective and subjective not being the antonyms they are taken for’ (D Wiggins, ‘Objectivity in Ethics: Two Difficulties, Two Responses’ (2005) 18 Ratio 1–26, 1 n 1). ⁵⁰ JL Mackie, Ethics: Inventing Right and Wrong (Penguin Books, 1990) 30. ⁵¹ B Williams, Ethics and the Limits of Philosophy (Fontana Masterguides, 1985) 145. ⁵² B Williams, ‘Saint Just’s Illusion’ in Making History and Other Philosophical Papers 1982– 1983 (Cambridge University Press, 1995) 141.

Six Paths to Vertigo-free Legal Theory

47

confidence⁵³ in the resources of moral argument. To this extent, non-cognitivism in ethics (‘moral judgments are not truth-apt and are not beliefs’) may be deemed incompatible with ethical objectivity.⁵⁴ A non-cognitivist view is often attributed to Hare,⁵⁵ owing to his characterization of moral judgments as typically ‘not pure statements of facts but in part, for example, expressions of attitudes, prescriptions, or something else nondescriptive.’⁵⁶ If it does not describe or ‘report’ anything about the world, if it does not state any fact, then there is no element by reference to which one may assess the truth value of a moral judgment, or so runs the kind of reasoning underlying a non-cognitivist understanding of Hare. To counter this kind of argument, Hare emphasized that the prescriptivist component of moral judgments does not rule out their having some descriptive content as well.⁵⁷ Unlike statements about the shape of objects, the moral injunction to care for one’s elderly parents for instance does not state any fact about elderly parents. If it has a descriptive content on the basis of which it may be deemed true or false, it is, according to Hare, the moral principles by reference to which that moral injunction may be justified.⁵⁸ The moral principles enjoying a widespread recognition in a given society may indeed be deemed ‘moral facts’ about that society, conditioning the truth-value of moral statements in that society. These facts do not, however, have the same robustness as, say, a table’s rectangular shape. By contrast with the moral principles underlying the moral injunction to support ⁵³ Note the difference in the roles that confidence is made to play in Williams’ and McDowell’s respective theories. While ‘confidence’ is Williams’ answer to the need to learn to live with a lack of foundations for ethics (‘Confidence’ is both a social state and related to discussion, theorizing and reflection; correspondingly, these activities are themselves forms of practice which take up social space, just as in the individual they take up psychological space: B Williams, Ethics and the Limits of Philosophy (Fontana Masterguides, 1985) 170), for McDowell it is the background condition that can be threatened by Neo-Humean naturalism, which requires confidence to be grounded in facts of first nature, instead of letting formed practical reason—‘second nature’—to assess whether confidence is in order or not, ‘exploiting’ whatever materials for critical reflection are available? (J McDowell, ‘Non-cognitivism and Rule Following’ in Mind Value and Reality (Harvard University Press, 1998) 194). ⁵⁴ According to this position, a moral judgment is more appropriately characterized as a special kind of desire or concern. ⁵⁵ ‘Since I myself think both that moral statements can be true or false, and that we can know them to be true or false, I get extremely cross when people classify me as a non-cognitivist’ (RM Hare, ‘Objective Prescriptions’ in Objective Prescriptions and Other Essays (Clarendon Press, 1999) ch 1 1–18, 4). ⁵⁶ RM Hare, Moral Thinking: Its Levels, Method and Point (Oxford University Press, 1981) 208 (emphasis mine). ⁵⁷ cf distinction between ‘thick’ and ‘thin’ moral concepts. ⁵⁸ ‘If moral statements are universalizable in this way, they will inevitably acquire a descriptive function as part of their meaning; for someone who makes such a statement must think that the moral properties of the situation about which it is made supervene on its non-moral properties (the reasons for making it) in accordance with a principle applying to all identically similar situations. Thus he will be taken to hold that there are facts that justify the statement, and it thus acquires a factual, descriptive content or meaning’ (RM Hare, ‘Prescriptivism’ Objective Prescriptions and Other Essays (Clarendon Press, 1999) ch 2 20–7, 20).

48

Sylvie Delacroix

one’s elderly parents, the rectangular shape of a table may indeed be deemed a fact existing wholly independently of our attitude, judgments or beliefs. Now this contrast need not be fatal to our moral judgments’ cognitive ambitions. According to McDowell, the desire to ground the truth of moral judgments in some robust reality existing independently of its relationship with the human mind is answerable in part to what he denounces as the Modern scientific ‘absolute conception of reality’. According to this outlook, reality is exhausted by the natural world, in the sense of the world as the natural sciences are capable of revealing it to us:⁵⁹ ie a world maximally free from peculiarities introduced by our own, or any other observer’s local perspective.⁶⁰ As he emphasizes the ‘interdependence’ existing between the subject and her objects of cognition, McDowell for his part defends the cognitive ambitions of moral language on the basis of an ‘anthropocentric realism’⁶¹ which does not require postulating any queer metaphysical entities. What it does require, however, is the capacity to build the confidence necessary to being able to solve our ethical disagreements in ‘nothing more, but nothing less’ than ‘all the whirl of organism Wittgenstein calls “forms of life”.’⁶² ‘We would be protected against the vertigo if we could stop supposing that the relation to reality of some area of our thought and language needs to be contemplated from a standpoint independent of that anchoring in our human life that makes the thoughts what they are for us.’⁶³ ⁵⁹ J McDowell, ‘Non-cognitivism and Rule Following’ in Mind Value and Reality (Harvard University Press, 1998) 175. ⁶⁰ B Williams, ‘Saint Just’s Illusion in Making History and Other Philosophical Papers 1982– 1983 (Cambridge University Press, 1995). ⁶¹ This expression is used by Stélios Virvidakis to designate ‘milder’ alternatives to Platonic versions of moral realism (S Virvidakis, La Robustesse du Bien: Essai sur le Réalisme Moral (Editions Jacqueline Chambon, 1996) 22). ⁶² S Cavell, Must We Mean What We Say? (Charles Scribner, 1969) 52 (quoted in J McDowell, ‘Non-cognitivism and Rule Following’ in Mind, Value, and Reality (Harvard University Press, 1998) 198–218, 207). ⁶³ J McDowell, ‘Non-cognitivism and Rule Following’ in Mind Value and Reality (Harvard University Press, 1998) 198–218, 211.

3 Monism, Interpretivism, and Law’s Aim George Letsas*

‘It is indispensable to our political life that citizens have a duty, in principle, to obey the law because it is the law and that the government is justified, again in principle, in enforcing the law because it is the law.’ Ronald Dworkin¹

I. We often say, of particular areas of law, that they have a specific purpose or aim. We say for instance that the purpose of human rights law is to protect the dignity of individuals; or that the purpose of employment law is to protect parties with unequal bargaining power; or that the new taxation legislation aims at reducing unemployment or inflation. But does law as a whole have an aim? And is that aim, a moral one? Some legal philosophers have taken the view that, regardless of what this or that law does in a given legal system, law as such does not live up to its true nature unless its effects are morally beneficial. An immoral or wicked legal system is like a broken screwdriver which cannot tighten or loosen screws: it does not do the things it is meant to do and its purpose is defeated. But a wicked legal system is—in some degenerate sense—still a legal system just like a broken screwdriver is still a screwdriver. In this respect, law aims to be morally justified just like screwdrivers aim to be mechanically justified, so to speak. In a recent paper, John Gardner observes that the ascription of a moral aim to law is something that both John Finnis’ and Ronald Dworkin’s theories of law have in common. He writes: ‘For Finnis and Dworkin alike law aims to guide * I have benefited enormously from discussions with Nicos Stavropoulos, Riz Mokal and Emmanuel Voyiakis. ¹ R Dworkin, ‘Reply to Leslie Green’ in J Burley (ed), Dworkin and his Critics 380.

Monism, Interpretivism, and Law’s Aim. George Letsas. © Oxford University Press 2007. Published 2007 by Oxford University Press.

50

George Letsas

its subjects properly, soundly, upstandingly, well. Law aims to be justified in the guidance that it gives to those whom it aims to guide such that what they do, when they are guided by the law and make no mistakes about the law, will itself be justified.’² He contrasts this to Joseph Raz’s version of legal positivism, according to which law claims—rather than aims—to be morally justified. According to Gardner, Dworkin’s version of law’s moral aim is that of justifying state coercion. Yet in other parts of his work, Gardner argues, Dworkin flirts with the incompatible proposition that law is morally justified. This is when Dworkin writes that, unlike the case of the Nazi legal system, the rules of a flourishing legal system do justify coercion.³ But if, as Gardner correctly observes, something has an aim then it is always logically possible that it fails in it. Indeed, who has not been frustrated by a pen that has run out of ink? Hence, if law’s purpose is to justify state coercion then it cannot at the same time be true that when law exists, coercion is justified.⁴ It seems that Dworkin would have to choose between these two propositions. But if he chooses the first (that law aims to be morally justified), Gardner argues, then he presupposes the positivist claim that law is made up of norms whose status as law depends on some legal official having engaged with them.⁵ Let us, following Gardner distinguish between these three propositions: (A) Law aims to be morally justified. (B) Law claims to be morally justified. (C) Law is morally justified. Gardner’s taxonomy is useful in that it makes the differences between Razian legal positivism, Finnis’ natural law and Dworkinian interpretivism turn on the way in which each theory connects law and moral justification. Yet there is a fundamental way in which this taxonomy is misleading, which is the focus of this paper. I will argue that the differences between Dworkinian interpretivism on one hand and positivism and natural law on the other can be located not on the question of law’s connection to moral justification but rather on different types of moral justification in general. This is no surprise in my view, but seeing why requires showing the way in which it is possible to say that law is morally justified rather than aims to be. But this way is made possible only if certain choices are made at the level of normative political theory. My immediate focus will be on what Liam Murphy has called monist and dualist theories of practical principle. ² J Gardner, ‘Law’s Aim in Law’s Empire’ in S Hershovitz (ed), Exploring Law’s Empire (2006) 210. ³ Ronald Dworkin, Law’s Empire (1998) [hereinafter LE] 103–4. ⁴ Indeed the logical incompatibility of these two propositions is obvious. As Dworkin notes in his response to Gardner: ‘how can one aim to be what one necessarily is?’ in S Hershovitz (ed), Exploring Law’s Empire 307. ⁵ J Gardner, ibid 220. Gardner argues that if one believes that law has an aim then she is necessarily a positivist, but not the other way around.

Monism, Interpretivism, and Law’s Aim

51

Though I will not argue against monist theories directly, I aim to show that what Gardner takes to be an inconsistency within Dworkin’s interpretivism can be explained away under non-monist theories of practical principle. A final note is appropriate. Although Dworkin has responded to Gardner’s paper, I do not mean to offer an exegetical analysis of his response or of Law’s Empire. My aim is to identify a non-monist interpretivist position that stands on its own as a separate view even if it cannot be attributed to Dworkin.

II. John Rawls is famous for arguing that justice is a virtue of social institutions, indeed their first virtue.⁶ Theories of justice offer normative principles that apply uniquely to the workings of institutions. These principles are not the same as the ones that apply to the conduct of individuals. Take for example Rawls’s difference principle. This principle requires that institutions arrange social and economic inequalities in a way that benefits the least advantaged. But it does not require that people, in their day-to-day life, act in a way that benefits the least advantaged. It does not require me, for instance, to buy the newspaper from the small, family-run, about-to-close-down, local newsagent instead of Waitrose. On a Rawlsian account, it is a category mistake to describe a person as ‘just’ or ‘unjust’. People can be cruel, insensitive, mean, generous or caring. But only the basic structure of society can be just or unjust. Liam Murphy calls Rawls’s view ‘dualism’. He contrasts this to monism, which he defends, and which claims that ‘all fundamental normative principles that apply to the design of institutions apply also to the conduct of people.’⁷ The upshot of monism for Murphy is that in non-ideal situations a person’s responsibility to promote the aims of justice is engaged directly, rather than mediated through institutional structures. Murphy worries that dualism dilutes our personal responsibility to reduce inequalities or alleviate suffering. ‘If people can do more to promote the aims of justice’, he writes, ‘a view that refuses to extend the principles of justice to personal conduct is prima facie deficient.’⁸ Why should there be different principles for different subject matters and why does this matter? Recall that one of Rawls’s objections to utilitarianism is that it does not take seriously the distinction between persons. This is because utilitarianism takes principles of choice that are only appropriate for regulating the lives of single individuals and extends them to the domain of society as a whole, ⁶ ‘For us the primary subject of justice is the basic structure of society, or more exactly, the way in which the major social institutions distribute fundamental rights and duties and determine the division of advantages from social cooperation’, J Rawls, A Theory of Justice (1999, revised ed) 6. ⁷ LB Murphy ‘Institutions and the Demands of Justice’ (1999) 27 Philosophy and Public Aff airs 251. ⁸ ibid 279.

52

George Letsas

conflating all persons into one. Utilitarianism is a paradigmatic example of a monist theory that uses the same principles for evaluating institutions and personal conduct. What I have a moral reason to do and what institutions have a moral reason to do are governed by the same normative principle, namely whether average—or some other account of—utility is maximized. By contrast, under dualist theories, it is not the case that the same overarching principle justifies both ethical and political requirements. Part of Rawls’s objection is that morality is multilayered and that ‘the correct regulative principles for a thing depends on the nature of that thing.’⁹ In fact the correct contrast, as Nagel rightly notes, is not between monism and dualism, but between monism and the idea that there are multiple levels of moral concern, governed by separate principles.¹⁰ I shall for short call the latter view non-monism. People who hold it fear that a monistic account may blind us to morally important features of entities or practices that call for separate normative principles, just like it blinds utilitarians to the separateness of persons.¹¹ Still, what is it about institutions that, for non-monists, calls for separate normative principles? It would be wrong to think that the reason is an instrumental one, namely that institutions are in a better position to reduce inequalities or that individuals will do a poor job in promoting justice directly, say due to co-ordination problems. This would presuppose that institutions are a means to promote aims that individuals already have a duty to promote, independently of the existence of institutions. It would in other words presuppose what the non-monist view denies, namely that principles of justice regulate primarily personal conduct and only derivatively the workings of institutions. If that were true then Murphy would certainly be right that the more ineffective the institutions are in securing justice, the more people’s individual responsibility is engaged. But such instrumental justification for recognizing separate normative principles for institutions is not open to the dualist for she believes that justice is a virtue of institutions alone not because of some contingent or superficial features they possess (such as their likelihood to promote certain antecedent aims) but because of some essential feature that in itself makes a moral difference. What could these features be? I shall briefly sketch a non-monist account of justice which I am broadly in agreement with, though I will not defend it in great detail.¹² The starting point is the fact that social institutions create strong political ⁹ J Rawls, A Theory of Justice 25. ¹⁰ Nagel, ‘The Problem of Global Justice’ (2005) 33 Philosophy and Public Aff airs 122. ¹¹ That is not to say that monists are bound to disagree with dualists on all matters of practical principle. Not all differences between the various moral theories turn on whether they reject monism. For example, though dualism is incompatible with utilitarianism, not all monists are utilitarians. Murphy himself claims to reject both utilitarianism and dualism as he does not believe that one has to be a dualist in order to reject the utilitarian principle of maximizing aggregate well-being without regard to its distribution. But he rightly notes that whether our moral landscape is monist or dualist is not a metanormative or nonnormative issue. See Murphy, ‘Institutions and the Demands of Justice’ 262. ¹² I am here drawing on Thomas Nagel’s discussion in ‘The Problem of Global Justice’ 128–9.

Monism, Interpretivism, and Law’s Aim

53

links between individuals, whose character is involuntary. To begin with, we do not choose which society to be born in (how could that be possible?) nor is it usually in our power freely to emigrate or to found a new state. We are moreover subject to the coercive force of the institutions of the political community to which we belong, even when we do not agree with its laws. Finally, coercive force is exercised in the name of the entire political community, that is, in the name of us all. It is the combination of these features that makes a moral difference: justice is a virtue of institutions because the existence of social institutions involves involuntary membership in a political community that uses collective force in the name of all its members. Political communities land on people both the requirement to abide by its laws and the responsibility for the effects that these laws have on the lives of all its members. Since this is quite a heavy burden to carry, we should expect that the threshold of justifying what, through our shared institutions, we owe our fellow citizens is higher than that of justifying what we owe to people living in other states or what states owe to each other. That is not to say of course that we only have moral duties towards our fellow citizens and we owe nothing to the rest of humanity. Duties of justice do not exhaust the realm of morality, which may also contain many other moral duties such as humanitarian assistance, human rights protection, relief of poverty, etc. But it does suggest that principles of justice are meant to regulate a unique political arrangement and cannot be extended, without risking making a moral mistake, to entities or practices very different from that of the basic structure of a society. It also suggests, in my view, that duties of justice should not be seen as flowing directly from a more fundamental, pre-institutional concern for the well-being of others.¹³ If that were the case, then the difference between our duties towards our fellow citizens and our duties towards the distant needy could only be the practical one of not having the global institutions in place that would be necessary in order to enforce the latter. But duties of justice are non-reductive and distinctive, holding between people that are members of the same political community. They are, as Nagel and Dworkin have emphasized, associative duties. One has no duty to be a member of any association but once she becomes a member, she acquires moral rights and duties that she did not and could not possess antecedently. I have no duty to have children, but if I do, new moral reasons are born along with my children, that apply to me directly and that do not collapse into moral reasons located outside the context of the family. As Nagel puts it in the context of justice: ‘Everyone may have a right to live in a just society, but we do not have an obligation to live in a just society with everyone.’¹⁴

¹³ It is no accident that Liam Murphy’s monism is coupled by beneficence-based view of justice. See ‘Institutions and the Demands of Justice’ 262. ¹⁴ Nagel, ‘The Problem of Global Justice’ 132.

54

George Letsas

III. What does all this have to do with law? Here is my suggestion: the practice of law may be sensitive to a moral value whose structure is very similar to the value of justice, understood in a non-monist way. Just as theories of justice are not meant to explore how institutions can promote some pre-institutional moral aims, so theories of law are not meant to explore how law can promote some pre-legal or extra-legal moral aims. Justice is the first virtue of institutions but there are certainly more, and the question is whether law (or legality) is one of them. Recall that in a political community, we stand in a special relation with others: we are involuntarily members of a community that exercises power over the lives of its members in the name of the community, that is, in the name of all its members. This political relation creates obligations that are associative and that relate to many aspects of the life of the community and its members. They relate primarily, as we saw, to duties of justice to remove arbitrary inequalities between the members of the community. Yet there is no guarantee that the community’s political decisions will get justice right or that people will agree on what the correct conception of justice is.¹⁵ But in the meantime the community will use its coercive power to enforce its decisions upon its members. It will use its power to enforce decisions that some of its members think are morally sound and others reject. It would make little sense, on a non-monist account of law, to ask whether enforcing the community’s decisions can, in principle, be morally justified. It is part of what makes an association what it is, that the fates of its members are bound together through collective action and that—absent extreme circumstances—my community’s decisions are in some sense my decisions. Coercive enforcement of the community’s decisions is simply the means through which we make demands on each other, distributing collectively burdens and benefits under some conception of justice. Coercion is, on the non-monist account,a sufficient condition for there being associative duties. We cannot therefore judge the legitimacy of coercion by asking what reasons we would have to obey a community’s decisions in the absence of coercion. We do not decide whether fulfilling a friend’s request is justified by asking what reasons anyone who is not his friend would have to fulfil the request or, even worse, by asking what reasons anyone would have to be his friend. It is indispensable to any account of the value of friendship that it creates distinctive obligations.¹⁶ ¹⁵ This is what Waldron has called the ‘circumstances of integrity’. See J Waldron, Law and Disagreement (1999) ch 9. ¹⁶ Think of other examples: I may have several reasons to go on a three-month boat trip with other members of my sailing club, all complete strangers to me. It will certainly enhance my interests: it will improve my social skills, my sailing skills, my knowledge of other places and so on. But the moral weight of the decisions that are made, once I am on the boat with others, is not to be

Monism, Interpretivism, and Law’s Aim

55

Assuming that a community displays an egalitarian commitment to secure justice for all its members¹⁷ then its decisions are a source of genuine obligation for its citizens. This is because citizens have a responsibility to secure, through their institutions, justice amongst each other and this responsibility can only be carried out through making and enforcing political decisions. The result is that, on a non-monist account, the value of justice would naturally lead to the recognition of the twin value of legality. Our duty to secure justice entails the duty to respect the conception of justice that we have, through law, settled on as a community. As an equal member of the community I therefore have a right that coercive force is only used against me when licensed by our past political decisions. Similarly, my fellow citizens have a duty not to coerce me, through our shared institutions, unless coercion is licensed by past political decisions.¹⁸ Dworkin writes: ‘Law insists that force not be used or withheld, no matter how beneficial or noble these ends are, except as licensed or required by individual rights and responsibilities flowing from past political decisions about when collective force is justified.’¹⁹ Hence the question a non-monist would ask is not whether coercive enforcement of past political decisions is justified. This is taken for granted. If political obligation is associative then coercive enforcement of a genuine community’s decisions is prima facie always justified. Rather, the question is the following: assuming law obligates, what are the principles that regulate how past political decisions—which provide a ground for coercion—should be construed? It is that moral question that the non-monist thinks legal practice raises and legal theory is all about. For her, legality is the moral value that gives us the content of the moral requirements that flow from the community’s past political decisions. Dworkin’s favourite conception of legality, what he calls integrity, is the idea that propositions of law are true if they follow from the most coherent set of principles of justice and fairness that past political decisions presuppose by way of justification.²⁰ In sum, legal rights and duties are distinctive rights and duties of political morality, governed by principles uniquely applicable to legal practice. It is possible to say that the community exercises legitimate control over its members’ lives even when its decisions do not secure justice and without having to assess the moral found solely by reference to the demands of ordinary individual morality that existed antecedently. Political theories that try to assess the legitimacy of law by asking what reasons would individuals have in the state of nature, to create institutions that have law-making powers, make the mistake of ignoring the existence of special values of political morality. ¹⁷ See Dworkin’s discussion on the conditions for a true, as opposed to a bare community in LE 201. ¹⁸ Dworkin notes correctly that his theory of law does not depend on personifying the community. See his response in S Hershovitz 306. ¹⁹ R Dworkin, Law’s Empire 93. ²⁰ LE 166. Dworkin discuss three conceptions of legality: conventionalism, pragmatism and law as integrity. An interesting question is whether conventionalism or pragmatism are compatible with non-monism or whether integrity is the only plausible conception of legality if one takes a non-monist account.

56

George Letsas

merit of those decisions in terms of the extent to which they direct citizens to act according to reasons of individual morality that apply to them independently.

IV. Let us return to Gardner’s three propositions. Would a non-monist accept that law aims to be morally justified? Certainly not. For she believes that legal practice is sensitive to a distinctive value of political morality which tells us what moral requirements exist as a matter of law. She may agree that the community’s decisions aim to secure justice. But law as such cannot have moral aims, for to say that the law permits or requires some action is to make a substantive claim about what moral obligations or rights already exist, having already employed some conception of the political value of legality. Propositions of law are evaluative, applied statements of a particular conception of the value of legality.²¹ To assign moral aims to law would be analogous to assigning moral aims to justice. Would the non-monist also reject Gardner’s third proposition that law is always morally justified? The answer is yes, if by ‘morally justified’ we mean either that the law is always just or that the law always directs people to act in accordance with reasons of personal morality that apply to them independently. Indeed, who would ever possibly think a community cannot but get justice right? The nonmonist however would of course accept this third proposition if by ‘morally justified’ we mean justified from the point of view of legality. For by definition, if I have a legal duty to do X, then I have a reason of political morality to do X and the community has a right to use collective force against me. If I have a legal right I am—prima facie—morally justified in exercising it. If I have legal duty I am—prima facie—morally obligated to perform it. To use Gardner’s own example, If under Swedish law (properly identified) I have a right to set a 14-day deadline to my tenant then I am justified in using this deadline even if this is not the best way to treat a tenant as a matter of justice. For the non-monist, law is morally justified in the sense that the rights and duties that the value of legality grounds provide a prima facie justification for the use of coercive force. It is of course possible that by complying with a legal duty we act as a duty of justice or of personal morality would have us act; still, as far as law is concerned, this is a contingent fact (to do with the history of each legal practice) not a necessary one. For it is not the case, on the non-monist view, that we are justified in exercising our legal rights only to the extent that, by doing so, we will end up acting in the way it would have been morally right to act, had law not created the particular right. Nor that theories of law are about the conditions under which one is morally justified in relying on the legal norm, all things considered. This is ²¹ This is I believe the gist of Dworkin’s famous quote: ‘Jurisprudence is the general part of adjudication, silent prologue to every decision of law’ in LE 90.

Monism, Interpretivism, and Law’s Aim

57

the point of the distinction Dworkin draws between the grounds and the force of law. The grounds of law are to do with ‘circumstances in which particular propositions of law should be taken to be true or sound.’²² In other words, the grounds of law are to do with what moral rights and duties past political decisions have actually created. Do I have a right to use a 14-day deadline or could it be that it is against some constitutional principle or precedent to use this deadline? How should I determine what the law is on this? The force of law, by contrast, is to do with ‘the power a true proposition has to justify coercion in different sorts of exceptional circumstance.’ Dworkin’s argument is that, though legal practice does create moral rights and duties, the moral reasons that legal rights and duties provide for judges and citizens need not be conclusive (the force of law). But this, he says, is a separate question usually studied by political philosophers not legal theorists. He writes: ‘Conceptions of law, which are theories about the grounds of law, commit us to no particular or concrete claims about how citizens should behave, or judges should decide cases.’²³ Indeed, given the distinct character of the value of justice and legality, conflicts between the two values can be both possible and genuine.

V. What would a monist account of law look like? I shall briefly explore a monist account of practical principle and its implications for law so that the contrast becomes clear. For it may be no accident that theorists who hold a monist account of practical principle are legal positivists. Take Joseph Raz’s philosophy. On his view political reasons are subordinate to ordinary individual morality. He writes: ‘If there are special political reasons then their use is justified to the extent that it enables individuals to act for the ordinary reasons that apply to them.’²⁴ According to Raz’s service conception of authority, the role of political authorities is to issue directives based on reasons that apply to individuals independently (the dependence thesis). An authority is legitimate when the alleged subject is likely better to comply with reasons which apply to him if he accepts the directives and tries to follow them rather than by trying to follow the reasons that apply to him directly (the normal justification thesis). As Raz puts it, authorities ‘mediate between people and the right reasons that apply to them.’²⁵ Note how the service conception of authority follows monism. On Raz’s view, the standards for assessing the legitimacy of authorities are the same as those used for assessing the reasons that apply to the conduct of individuals. It is not just that the role ²² ²³ ²⁴ ²⁵

LE 110. ibid 112. J Raz, The Morality of Freedom (1986) 72. J Raz, Ethics in the Public Domain (1995) 214.

58

George Letsas

of authorities is to direct the conduct of their subject. It is that the legitimacy of authorities is wholly judged by principles of personal conduct. As Scott Shapiro puts it, ‘authorities work for us, not the other way around’.²⁶ Now suppose we assume that the monist is right that there are no distinctive reasons of political morality and that the only reasons that do exist are reasons of personal morality regulating individual conduct. How would he construe Gardner’s propositions? Take the question whether law has a moral aim. The monist would say that law has a moral aim only in the sense that law aims to direct persons to act in accordance with reasons of individual morality that apply to them independently. But in the light of the role authorities have in guiding people’s conduct, the monist would naturally assert that law claims (as opposed to aims) to direct persons to act in accordance with reasons of individual morality that apply to them independently. And she would certainly reject Gardner’s third proposition interpreted as follows: law always directs persons to act in accordance with reasons of individual morality that apply to them independently. The latter proposition would obviously be false for law is fallible. Being a human creation, law may fall short of the demands of individual morality. One option for the monist would be to replace (A) and (C) with the following proposition: (D) Law is morally justified only when, and to the extent that, it directs people to act in accordance with reasons of individual morality that apply to them independently. The problem with (D), however, is that it makes little sense to say that law is morally justified when its directives do not in fact create reasons that obligate the subjects. When law prescribes a conduct that persons are already under an obligation to perform, then the moral force of law is redundant: that course of action was already morally required. When the conduct that law prescribes is morally unjustified from the point of view of individual morality then we have no reason to obey the law. On the monist account, law as such never creates a first-order moral obligation. It never changes our moral commitments. It would therefore make more sense, on the monist view, to ask not whether law is morally justified (in the sense that it obligates), but what reasons could there be to follow what political authorities direct us, through the laws they enact, to do. It is clear that there is only one possible answer to the latter question: we can only have an instrumental reason to obey the law. We should obey an authority only if, by doing so, it is more likely to act in accordance with the rights of individual morality that apply to us. We arrive, in other words, at what Raz calls the service conception of authority: authorities are legitimate when they serve their subjects. They serve their subject when they mediate successfully between persons and ²⁶ S Shapiro, ‘Authority’ in J Coleman and S Shapiro, The Oxford Handbook of Jurisprudence (2002) 402.

Monism, Interpretivism, and Law’s Aim

59

the reasons that apply to them. Authorities are legitimate when they meet a test of instrumental rationality. This is what Raz calls the normal justification thesis, namely that an authority is legitimate when the alleged subject is likely better to comply with reasons which apply to him if he accepts the directives and tries to follow them rather than by trying to follow the reasons that apply to him directly. Now recall that the service conception of authority and the normal justification thesis are used, by Raz, to get to the ‘pre-emptive thesis’ and, most importantly, to his version of legal positivism, the ‘sources thesis’. The sources thesis provides an answer to the question of what is the content of our legal rights and duties, or as Dworkin puts it, to the questions of the grounds of law. But the sources thesis depends on Raz’s account of political authority which in turn depends on a monist view of practical principle. On this view, it is not possible to argue that law is morally justified in the sense that our legal rights and duties, properly identified, are moral rights and duties.

VI. I have tried in this paper to explore the implications of the debate between monist and non-monist accounts of practical principle for legal theory. Monists who deny altogether the existence of reasons of political morality are naturally led to an instrumental account of law’s legitimacy and to a positivistic account of the grounds of law. For the legal positivist, law never, by itself, obligates; it either claims or aims to obligate. But if someone accepts that morality is multi-layered and that there are intrinsic reasons of political morality, then the possibility to recognize both justice and legality as distinct values is open. By doing so, the non-monist can agree with the interpretivist that law is morally justified without being committed to the claim that law is always just or that law necessarily directs us to act as personal morality would have us act. I have of course said nothing in defence of either the non-monist position in general or of its application to law. Nor have I shown that all the differences between interpretive and non-interpetive theories of law turn on the distinction between monism and non-monism.²⁷ But some very important ones do and this is good news for all of us who believe that the fundamental questions of jurisprudence are not detached from moral and political philosophy.

²⁷ For example, can a non-monist recognize justice, but not legality, as value? Can a non-monist be a legal realist? I have not pursued these questions here.

4 Moral Evaluation and Conceptual Analysis in Jurisprudential Methodology John Oberdiek and Dennis Patterson

Analytic general jurisprudence has become increasingly attentive to its own methodology in recent years. No longer content with its traditional first-order questions revolving around the varieties, commitments, and defensibility of legal positivism, the discipline of jurisprudence has turned inward, asking the second-order question, How should one do jurisprudence?¹ This second-order methodological question is not unrelated to jurisprudence’s traditional first-order concerns, to be sure, as the role of evaluation, and indeed the role of moral theorizing more specifically, figures prominently in the methodological inquiry much as it does in the first-order debate among positivists, interpretivists, and natural law theorists about the role that morality plays or can play as criteria of legality. But the methodology debate has a cast all its own—it is not a mere proxy war between rival positivists and their mutual foes. In the first instance, second-order methodological positions on the role of moral evaluation in jurisprudence do not correspond directly to first-order positions regarding the relationship between legality and morality. Furthermore, the methodology debate focuses on one of the few planks in nearly all of the contenders’ platforms, forcing legal philosophers to justify or jettison their shared commitment to conceptual analysis. It is the purpose of this discussion to introduce the methodology debate, draw attention to the merits and shortcomings of various positions already staked out, and to contribute to the debate by, albeit briefly, defending the claims that moral evaluation has (at least) a modest role in analysing the concept of law and that conceptual analysis, or rather, many of its incarnations, is defensible and indeed inescapable in jurisprudence. ¹ See, for example, L Green, ‘General Jurisprudence: A 25th Anniversary Essay’ (2005) 25 OJLS 4, 575; J Dickson, ‘Methodology in Jurisprudence: A Critical Survey’ (2004) 10 Legal Theory 117, 117–56; J Dickson, Evaluation in Legal Theory (Oxford, 2001); as well as Jules Coleman, The Practice of Principle (New York, 2001) 179–217, which is substantially reproduced as ‘Methodology,’ in J Coleman and S Shapiro (eds), The Oxford Handbook of Jurisprudence and Philosophy of Law (New York, 2002) 311–51.

Moral Evaluation and Conceptual Analysis in Jurisprudential Methodology. John Oberdiek and Dennis Patterson. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Moral Evaluation and Conceptual Analysis

61

I. A Brief History of the Debate In his seminal 1961 book, The Concept of Law, HLA Hart advanced a powerful case for legal positivism, which he characterized as ‘descriptive jurisprudence.’² A central feature of Hart’s project was the clarification of legal practice through a perspicuous description of its principal features. Chief among these elements, in his view, was the conjunction of primary and secondary rules, where primary rules apply directly to citizens while secondary rules regard the identification, interpretation, and modification of primary rules. The keystone in Hart’s scheme was a particular secondary rule, namely, the Rule of Recognition. The Rule of Recognition was the master rule within Hart’s system because it alone provided the criteria determining whether a given rule was valid or not. Thus, by appeal to the Rule of Recognition, lawyers and non-lawyers alike could identify legal norms as legal norms and, in that way, decide what the law is on any given question.³ Ronald Dworkin, the most persistent critic of legal positivism, has consistently contested the possibility of the descriptive jurisprudence that Hart envisioned. One of Dworkin’s central claims—advanced most systematically in his magisterial 1986 book Law’s Empire⁴—is that ‘law’ is what Dworkin terms ‘an interpretive concept’. Law is an interpretive concept because to understand the concept, one needs to grasp its point or purpose.⁵ In Dworkin’s view, one of the key dimensions of law is its coerciveness, and consequently, any explication of the concept of law must be able to explain and justify this important aspect of state action.⁶ For Dworkin, justifying state coercion is a (if not the) fundamental task of jurisprudence. More importantly, it is an enterprise that is inescapably normative and not, pace Hart, merely descriptive. The second edition of The Concept of Law, published posthumously in 1994, contained a rich and controversial Postscript, wherein Hart maintained that he and Dworkin were engaged in different philosophical enterprises, and that both enterprises were legitimate methodologies for jurisprudence. In reply, Dworkin remained unpersuaded, maintaining that his was the only appropriate methodology for jurisprudence.⁷ Thus, the debate within analytic general jurisprudence has turned increasingly to methodology, asking not ‘What is law?’, but, more self-consciously, ‘How should one do philosophy of law?’ As noted above, this second-order debate is related to the older first-order debate about the nature of law in that the methodology ² HLA Hart, The Concept of Law (2nd edn, Oxford, 1994) preface, i. ³ See Hart, ibid 94–9 (describing the Rule of Recognition). ⁴ R Dworkin, Law’s Empire (Cambridge, MA, 1986). ⁵ Dworkin, ibid 190 (‘A conception of law must explain how what it takes to be law provides a general justification for the exercise of coercive power by the state. . .’). ⁶ Dworkin, ibid 94. ⁷ R Dworkin, ‘Hart’s Postscript and the Character of Political Philosophy’ (2004) 24 OJLS 1. Th is essay is reprinted in R Dworkin, Justice in Robes (Cambridge, MA, 2006) 140–186.

62

John Oberdiek and Dennis Patterson

debate began with the question of whether purely descriptive jurisprudence is possible, or whether it instead must presuppose normative and specifically moral commitments.⁸ Hart, and positivists generally, have argued that description of the law without justification of the law was possible. Dworkin, as well as John Finnis⁹ and Stephen Perry,¹⁰ have insisted that jurisprudence must engage in moral justification. Perry’s work on this topic is widely considered to be the clearest starting point of the new methodology debate, and it is worth elucidating his central claims. Perry has argued that Hart’s own jurisprudential methodology, though avowedly purely descriptive, in fact incorporated both descriptive and normative elements.¹¹ It was descriptive insofar as Hart attempted simply to describe the concept of law, but it was unavoidably normative, his own self-understanding notwithstanding, insofar as Hart sought to describe it from the perspective of those over whom the law claims authority. Hart could not carry out a purely descriptive enterprise, according to Perry, because law itself is inherently normative, in that it purports to provide its subjects with overriding normative reason to do that which they otherwise would lack reason to do. To the extent that the phenomenon to be described is normative in this way, Perry argued, its hermeneutic, internal description must also be normative. Put roughly, in order to explain why those subject to legal authority should respect that legal authority, as is necessary if one seeks to describe the concept of law from the perspective of those over whom the law claims authority, then it is necessary to explain what is respectable about that legal authority. Showing what is respectable about a legal authority in turn involves (at least partially) justifying that legal authority. Accordingly, Perry concluded that, in light of the inherent normativity of law—the object of description—Hart’s pure descriptivism could not stand. Hart’s most thorough defender on this score has been Julie Dickson, who has argued not only that Hart’s methodology is not nearly as unstable as Perry makes it appear, but that Hart’s is in fact jurisprudence’s proper methodology.¹² Dickson accepts that Hart seeks to describe the concept of law and that doing so requires taking up the internal perspective of those working with and subject to law, but she steadfastly denies Perry’s claim that describing the law from the internal perspective requires justifying law. While accepting that some form of evaluation is necessary in describing law, Dickson distinguishes between moral evaluation of the kind that marks a justification, on the one hand, and non-moral evaluation, ⁸ For a clear and concise account of the debate thus far, see Dickson, ‘Methodology in Jurisprudence: A Critical Survey’ n 1 above. ⁹ See J Finnis, Natural Law and Natural Rights (Oxford, 1984) 4. ¹⁰ See especially, SR Perry, ‘Hart’s Methodological Positivism’ in J Coleman (ed), Hart’s Postscript (New York, 2001). ¹¹ In addition to the above article, see, Perry, ‘Interpretation and Methodology in Legal Theory’ in A Marmor (ed), Law and Interpretation (Oxford, 1995); and Perry, ‘Holmes v Hart: The Bad Man in Legal Theory’ in S Burton (ed), ‘The Path of Law’ and its Influence (New York, 2000). ¹² See Dickson, Evaluation in Legal Theory n 1 above.

Moral Evaluation and Conceptual Analysis

63

on the other. It is Dickson’s view that legal theorists certainly do evaluate the data that they are trying to explicate, insofar as they are trying to explicate the significant and important data about law, and not extraneous data. Dickson moreover contends that the legal theorist’s evaluations must be attuned to what those persons who create, administer, and are subject to the law take to be significant and important about law. The systematization and clarification of the concept of law relies on just such theoretical evaluation. But Dickson is resolute in maintaining that evaluation of this kind is a far cry from the outright moral evaluation that Perry argues is necessary. Dickson’s is an evaluative-but-not-morally-evaluative, or what she terms ‘indirectly evaluative’, jurisprudence.¹³ If Perry and Dickson have each staked out clear alternative positions in the methodology debate, Brian Leiter has significantly widened it by relating the debate to important developments in metaphysics and epistemology. Traditionally, jurisprudence has been concerned with a single animating question, namely, the nature of law. Until Leiter’s foray into the methodology debate, no one in the field questioned jurisprudence’s fundamental animating premise, namely, that ‘law’ had a nature capable of explication. Adverting to WVO Quine’s famous 1951 paper, ‘Two Dogmas of Empiricism’,¹⁴ and the robust empiricism that it ushered in, Leiter maintains that the idea that concepts have a nature or necessary structure or content is mistaken, and that a priori inquiry into the concept of law—the task that jurisprudence has traditionally set for itself—is therefore fundamentally misguided. What jurisprudence must undergo, according to Leiter, is a naturalistic revolution, abandoning the traditional mission of conceptual analysis of ‘law’ and instead employing empirical methods to understand law. As Quine’s naturalized epistemology is Leiter’s model, reviewing Quine’s central claims motivates and sheds light on Leiter’s position. Since Plato, philosophers have endeavoured to develop accounts of concepts— concepts like mind, knowledge, or justice—that unpack their content in terms of necessary and sufficient conditions discernable a priori. According to the so-called ‘Classical Theory of Concepts’, most concepts are ‘structured mental representations that encode a set of necessary and sufficient conditions for their application . . .’.¹⁵ While a single theory cannot be continuously identified from Plato on, it is not an overstatement to say that something like the Classical Theory was a significant element in philosophy from antiquity to the middle of the 20th century.¹⁶ The reason for the longevity of the Classical Theory is not ¹³ Similarly, Jules Coleman argues, ‘[i]t is one thing to claim that normative considerations figure in the project of identifying the features of law the theory of the concept must account for; quite another to claim that the norms to which one must appeal are those of political morality.’ Coleman, The Practice of Principle n 1 above, 199. ¹⁴ WVO Quine ‘Two Dogmas of Empiricism’ (1951) 60 Philosophical Review 20, reprinted in his From a Logical Point of View (Cambridge, MA, 1953) 20–46. ¹⁵ S Laurence and E Margolis, ‘Concepts and Cognitive Science’ in E Margolis and S Laurence (eds), Concepts: Core Readings (Cambridge, MA, 1999) 18. ¹⁶ Frank Jackson is the most vigorous, contemporary defender of conceptual analysis. See F Jackson, From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford, 1998).

64

John Oberdiek and Dennis Patterson

hard to see: it holds the promise of tremendous explanatory power, ‘offering unified accounts of concept acquisition, categorization, epistemic justification, analytic entailment and reference determination, all of which flow directly from its basic commitments.’¹⁷ One or another version of the Classical Theory held sway until 1951, when Quine published the paper, ‘Two Dogmas of Empiricism’.¹⁸ Of the several theses Quine advanced in that landmark paper, the most important one for present purposes was his denial of what is known as the analytic/synthetic distinction.¹⁹ Analytic statements are true in virtue of the meaning of the constituent terms. The most famous example is the proposition ‘All bachelors are unmarried males’. Synthetic judgments, on the other hand, are claims of contingent fact. So, the proposition ‘it is raining outside’ is true not just in virtue of the meaning of any of the constituent terms but in virtue of facts in the real world. Quine argued, among other things,²⁰ that the analytic/synthetic distinction was indefensible because all statements are revisable, as he says, ‘in the light of experience.’ In other words, repudiating the analytic/synthetic distinction entails repudiating the distinction between necessary and contingent truths, a distinction upon which conceptual analysis depends.²¹ ¹⁷ ibid. ¹⁸ Quine, ‘Two Dogmas of Empiricism’ n 14 above, 20–46. ¹⁹ Here is the famous passage that sums up Quine’s position: The totality of our so-called knowledge or beliefs, from the most casual matters of geography and history to the profoundest laws of atomic physics or even of pure mathematics and logic, is a man-made fabric which impinges on experience only along the edges. Or, to change the figure, total science is like a field of force whose boundary conditions are experience. A conflict with experience at the periphery occasions readjustments in the interior of the field. Truth values have to be redistributed over some of our statements. Re-evaluation of some statements entails re-evaluation of others, because of their logical interconnections-the logical laws being in turn simply certain further statements of the system, certain further elements of the field. Having re-evaluated one statement we must re-evaluate some others, whether they be statements logically connected with the first or whether they be the statements of logical connections themselves. But the total field is so undetermined by its boundary conditions, experience, that there is much latitude of choice as to what statements to reevaluate in the light of any single contrary experience. No particular experiences are linked with any particular statements in the interior of the field, except indirectly through considerations of equilibrium affecting the field as a whole. If this view is right, it is misleading to speak of the empirical content of an individual statement-especially if it be a statement at all remote from the experiential periphery of the field. Furthermore it becomes folly to seek a boundary between synthetic statements, which hold contingently on experience, and analytic statements, which hold come what may. Any statement can be held true what may, if we make drastic enough adjustments elsewhere in the system. Quine, ibid 39–40. ²⁰ For a detailed explication of Quine’s arguments in ‘Two Dogmas of Empiricism’, see S Soames, Philosophical Analysis in the Twentieth Century (Princeton, 2003) 351–405. For an assessment of the long-term influence of Quine’s position, see PMS Hacker, Wittgenstein’s Place in Twentieth Century Philosophy (Oxford, 1996) 183–227. ²¹ In fact, it is doubtful that there is such a thing as ‘the’ analytic/synthetic distinction. For discussion, see PMS Hacker, ‘Passing By the Naturalistic Turn: On Quine’s Cul-De-Sac,’ (2006) 81 Philosophy 231, 233–34 (pointing out that the limited Quine’s target was Carnap’s account of the analytic/synthetic distinction and enumerating a variety of other accounts of the distinction).

Moral Evaluation and Conceptual Analysis

65

Even if Quine is right, how do his arguments matter to analytic general jurisprudence? Dworkin, recall, maintains that law’s coercive power is essential to law and that it is that power that must be justified in any explication of the concept of law. This project assumes the defensibility of some form of conceptual analysis; indeed, Dworkin couches his project in precisely those terms.²² Legal positivists, too, engage in conceptual analysis. Joseph Raz, the tradition’s leading living exponent, writes: In as much as a general theory of law is about the nature of law it strives to elucidate law’s essential features, i.e., those features which are possessed by every legal system just in virtue of its being legal, by every legislative institution in virtue of its being legislative, by every practice of legal reasoning in virtue of its being a practice of legal reasoning, and so on. A claim to necessity is in the nature of the enterprise.²³

If Quine is right,²⁴ it would seem that neither Dworkin nor Raz (to take two prominent examples) can be right about methodology in legal theory. The reason is that claims to be describing the ‘essential features’ of law depend for their efficacy on the analytic/synthetic distinction. If the analytic/synthetic distinction is untenable, as Quine maintains, then it would seem to follow that no project of conceptual analysis, whether Dworkinian or positivist, gets off the ground. This is, in broad brush, Leiter’s negative claim. More constructively, Leiter wants to fill the gap that he believes Quine creates by importing Quine’s naturalism into legal theory.²⁵ Just as science replaces philosophy in Quine’s naturalized epistemology,²⁶ Leiter maintains that social science must replace concept-driven legal philosophy as the best explanatory methodology for law. Instead of conceptual analysis, philosophers should be engaged in social scientific empirical research.²⁷ In light of his thoroughly naturalized jurisprudence, Leiter stakes out a position distinct from both Perry’s and Dickson’s. Like Dickson, Leiter believes that jurisprudence is a largely descriptive enterprise and is evaluative only to the extent that it must be responsive to the normative understanding of the law ²² Dworkin writes, ‘[e]ach conception [of law] furnishes connected answers to three questions posed by the concept. First, is the supposed link between law and coercion justified at all? Is there any point in requiring public force to be used only in ways conforming to rights and responsibilities that “flow from” past political decisions? Second, if there is such a point, what is it? Third, what reading of “flow from”—what notion of consistency with past decisions best serves it?’ Dworkin, Law’s Empire n 4 above, 94. ²³ J Raz, ‘On the Nature of Law’ (1996) 82 Archiv für Rechts- und Sozialphilosophie 1, 2. It is worth noting that Raz explicitly states, ‘I will leave the question of the kind of necessity involved unexplored.’ ibid. ²⁴ But see P Grice and PF Strawson, ‘In Defense of a Dogma’ (1956) 65 Philosophical Review 141–58, reprinted in Paul Grice, Studies in the Way of Words (Cambridge, MA, 1989) 196–212. ²⁵ See B Leiter, ‘Beyond the Hart/Dworkin Debate: The Methodology Problem in Jurisprudence’ (2003) 48 American Journal of Jurisprudence 17. ²⁶ See B Leiter, ‘Naturalism in Legal Philosophy’ in Stanford Encyclopedia of Philosophy (2002), available at . ²⁷ See Leiter, ‘Beyond the Hart/Dworkin Debate’ n 24 above, 48–52.

66

John Oberdiek and Dennis Patterson

held by those subject to law. However, Leiter also believes that all that is needed is the meta-theoretic evaluation that is part of any theorizing—no theory can be entirely ‘value free’—involving evaluative judgments of explanatory power, coherence, simplicity, and so forth. Indeed, while Leiter joins the consensus that jurisprudence should respond to the internal point of view of those subject to law, he parts with Dickson in maintaining that jurisprudence should do without any mediating evaluative judgments about the relevance of certain understandings of law.²⁸ In Leiter’s view, it is not for the legal philosopher to determine which internal understandings are sufficiently important and which are confused. It is all just indistinguishable factual data, the simple collection and systematization of which is the legal philosopher’s task. This is naturalistic jurisprudence in that it eschews any aspiration to capture law’s conceptual essence apart from what can be gleaned from empirical data on what people think law is and the role that they accord to it in their self-understanding. Leiter thus seems to accept a more modest role for evaluation in jurisprudence than even Dickson. The thumbnail sketch of the recent history and current state of play in the methodology debate is, of course, simplified. There are many more players and theses than those just discussed. At the same time, the positions staked out by Perry, Dickson, and Leiter mark clear poles in the debate, and the framework provided by their positions is therefore a useful one. As currently structured, the methodology debate within analytic jurisprudence (finally) joins wider debates in the philosophy of language, epistemology and metaphysics. In what follows, we take up the different facets of this debate under two broad headings: Part II discusses the role of moral evaluation in jurisprudence and Part III takes up the role of conceptual analysis.

II. Moral Evaluation in Jurisprudence In this section, we attempt to explore further Perry’s claim that the explication of the concept of law requires moral evaluation and that jurisprudence is in that sense normative. It is common ground that some evaluation is required to unpack the content of the concept of law; what is controversial is whether the evaluation must be moral in nature. The line to be explored here is that every legal system in fact possesses a morally attractive property, deriving from its adherence to the rule of law, and that moral premises must therefore be relied upon to understand the nature of law. The argument does not countenance an inference from legality to moral legitimacy. Laws and legal systems can of course be morally illegitimate. But one can reject the inference from legality to moral legitimacy while maintaining that law qua law has some moral value. Moral value, on this view, is not an ²⁸ This is Dickson’s interpretation of Leiter’s position, advanced while correcting what she regards as Leiter’s misinterpretation of her position. See Dickson, ‘Methodology in Jurisprudence: A Critical Survey’ n 1 above, 133–39.

Moral Evaluation and Conceptual Analysis

67

all-things-considered judgment, but a claim about one feature of law. In this respect, the moral value that we are suggesting inheres in law is defeasible, such that the moral worth of a legal system can be undermined by morally illegitimate laws, while retaining some moral value simply in virtue of its adherence to the rule of law. Our argument is brought into relief most clearly by attending to an argument dismissed by Jules Coleman.²⁹ Coleman suggests that ‘law’ functions as a ‘predicate of weak commendation’, that any account of law must explain why we valorize legal governance as we do, and that the case for so-called ‘normative jurisprudence’ might succeed if ‘the best explanation of how it is that law plays this role in our normative discourse is that law has a moral property adequate to warrant “law’s” linguistic role.’³⁰ Coleman connects the dots as follows: If law has such a property, an analysis of the concept of law should specify what this property M is—and in so doing, the analysis must appeal to moral argument in two ways: in specifying the content of M, the analysis must reveal how or why that property is morally attractive; and the overall analysis must be such that it picks out only things that have the property M.³¹

Coleman, to repeat, denies that this argument is sound. And he does so in two steps. First, he argues that the best explanation of the weak commendation role of law is the potential possessed by legal systems for good governance; second, that law has the potential or capacity to facilitate good governance does not, according to Coleman, entail that this capacity is part of the concept of law.³² In Coleman’s view, the potential to facilitate good government, that instrumental value, is a feature of law that any plausible analysis of the concept of law must account for, to be sure, but it is not a central feature, any more than the capacity to be used as a paperweight is a central feature of the concept of a hammer. Let us grant for the sake of argument that Coleman is correct that if law can only be commended in virtue of its potential, moral considerations need not be invoked to unpack the concept of law. It remains an open question, however, whether the commendation of law is elicited or warranted solely in virtue of law’s potential for facilitating good government. More specifically, the value of the rule of law does not depend on the contingent content of the law, but rather on features of law itself, and thus if it is the rule of law that (at least in part) elicits or warrants law’s commendation, and the rule of law is central to the concept of law, it follows that a case can be made for normative jurisprudence that escapes any criticism focused on law’s mere potential. Now, the first question is whether the rule of law is central to the concept of law. If it is not central, after all, then adducing the moral value inherent in the rule of law will do nothing to establish that moral argument is required to ²⁹ ³⁰ ³¹ ³²

See Coleman, The Practice of Principle n. 1 above, 190–5. ibid 191. ibid. ibid 192–4.

68

John Oberdiek and Dennis Patterson

explicate the concept of law. How central, then, is the rule of law to the concept of law? There is some reason to think that it is not central. Raz, for example, holds that ‘[t]he rule of law is a political ideal which a legal system may lack or may possess to a greater or lesser degree.’³³ If so, if a legal system can remain a legal system even in the absence of upholding the rule of law, then it would seem to follow that the rule of law cannot be a central feature of the concept of law. But this is perhaps the wrong way to read Raz. In discussing Lon Fuller’s claim that the rule of law is essential to the existence of law, Raz admits that the principles of the rule of law—that laws should be open, clear, accessible, and stable, for example—‘cannot be violated altogether by any legal system.’³⁴ This suggests, contrary to the import of his earlier claim, that there could be no legal system that failed utterly to adhere to the rule of law. It suggests, that is, that some degree of fidelity to the rule of law, some degree of formal fairness, is a necessary feature of any legal system, and so of the concept of law. Coleman, too, appears to accept this proposition. He maintains, ‘[t]o be sure, a legal system need not be effective in constraining the exercise of the ruler’s power, and may even stipulate that the law imposes no such constraints; but in so far as a ruler exercises purely arbitrary power, he or she does not govern by law.’³⁵ Coleman argues that the exercise of ‘purely arbitrary power’ cannot be lawful, which of course entails that power must be constrained in some way or to some degree in order to be lawful. And while Coleman nowhere mentions ‘the rule of law’, the idea that arbitrary power is to be curtailed just is the idea of the rule of law.³⁶ There is much more to be said on this score, but let us proceed on the assumption, which we have provided some reason to endorse, that the rule of law is central to the concept of law. The question now is whether the commendation of law is elicited or warranted by a feature of the rule of law itself or whether instead the commendation is due merely to the capacity or potential law has for good governance. Raz is again instructive. According to him, the rule of law cannot depend entirely for its value on the content of whatever laws are adopted. As Raz points out, ‘[i]f the rule of law is the rule of the good law then to explain its nature is to propound a complete social philosophy. But if so the term lacks any useful function.’³⁷ And again, ‘conformity to the rule of law is an inherent value of laws, indeed it is their most important inherent value.’³⁸ At the same time, however, Raz maintains, Clearly, the extent to which generality, clarity, prospectivity, etc. are essential to the law is minimal and is consistent with gross violations of the rule of law. But are not ³³ J Raz, ‘The Rule of Law and its Virtue’ in The Authority of Law (Oxford, 1979) 211. ³⁴ ibid 223. Summarizing his point, Raz maintains, ‘at least some of the rules of recognition and of adjudication of every system must be general and prospective.’ ³⁵ Coleman, The Practice of Principle n 1 above, 192–3. ³⁶ As Raz puts it, ‘certainly many of the more common manifestations of arbitrary power run foul of the rule of law.’ Raz n 33 above, 219. ³⁷ ibid 211. ³⁸ ibid 225.

Moral Evaluation and Conceptual Analysis

69

considerations of the kind mentioned sufficient to establish that there is necessarily at least some moral value in every legal system? I think not. The rule of law is essentially a negative value. . . . [C]onformity to it does not cause good except through avoiding evil and the evil which is avoided is evil which could only have been caused by the law itself.³⁹

The value of the rule of law is not, in Raz’s estimation, sufficient to endow any legal system with any moral value, however minimal. Raz’s argument in support of that conclusion is highly compressed and it is worth briefly unpacking it. Raz illustrates his claim about the rule of law’s purely negative value by analogizing the rule of law to honesty: just as one who cannot communicate or thus deceive deserves no credit for being honest, the fact ‘that the law cannot sanction arbitrary force or violations of freedom and dignity through total absence of generality, prospectivity, or clarity is no moral credit to the law.’⁴⁰ Indeed, Raz goes so far as to liken the conceptual inability of law to instantiate the evil of arbitrary rule to its inability to murder—and no one finds it a moral virtue of the law that it cannot itself murder anyone. Raz rejects the claim that there is any merit in refraining from doing that which is conceptually impossible. He thus concludes that insofar as the rule of law is conceptually incompatible with completely arbitrarilydirected power, the fact that law cannot be guilty of that evil is not to its credit. This argument is much too quick. One can concede that the rule of law is a negative virtue and yet find its negative virtue to be an authentic commendation feature, in Coleman’s words, and so find grounds for commending governance under law. One can see what more can be said on behalf of the rule of law’s negative virtue if one considers other forms of governance in assessing the value of the rule of law. Societies can take any number of different forms, and specifically, they need not incorporate a legal system. This is uncontroversial. We can, moreover, imagine a society without a legal system that is prosperous and just. Clearly, such a society would be superior to a society that possessed a legal system, and thus the rule of law, but which was neither prosperous nor just. This is also uncontroversial. But now compare a society without a legal system that is prosperous and just with one that has a legal system and that is also prosperous and just; when that comparison is made, it is hard to deny that the society with the rule of law is superior to the society lacking the rule of law. And this has everything to do with the rule of law’s negative virtues, enumerated by Raz. Neil MacCormick captures the point nicely in maintaining, ‘There is always something to be said for treating people with formal fairness, that is, in a rational and predictable way, setting public standards for citizens’ conduct and officials’ responses thereto.’⁴¹ The rule of law always adds something of value to the societies that possess it. And it is for this reason that if one were in a position to choose which of these two societies to live in, both of ³⁹ ibid 224. ⁴⁰ ibid 224. ⁴¹ N MacCormick, ‘Natural Law and the Separation of Law and Morals’ in Robert George (ed), Natural Law Theory (New York, 1995) 123. MacCormick goes on to claim, ‘That indeed is what we mean by the Rule of Law.’ ibid.

70

John Oberdiek and Dennis Patterson

which were prosperous and just but only one of which abided by the rule of law, one would have reason to live in the society with the rule of law, precisely because of its negative virtues. Even if a society achieves prosperity and distributes goods and resources justly, failing to do so according to the rule of law constitutes a moral demerit. That prosperity and justice should be pursued in an open, predictable, and rule-governed manner is an undeniable moral good. In this way, the moral value of the rule of law, while negative, is nonetheless additive; societies which are otherwise good and just are better if they also have a legal system formalizing norms that underwrite their prosperity and justice. None of this should be construed to suggest that societies that possess the rule of law are morally superior to any society that lacks it. There is no question that a prosperous and just society lacking the rule of law is morally better than one that possesses the rule of law but which has substantively immoral laws. Yet this consideration raises what might be thought to be a problem for the claim advanced here, that legality is inherently morally valuable. The potentially troubling thought is this: persons living in a thoroughly morally bankrupt society with the rule of law might in fact be better off without the rule of law, if that would mean that some who would otherwise be subject to morally indefensible laws would be spared from them, precisely because such exceptions would run contrary to the rule of law’s strictures concerning generality, prospectivity, and formally equal treatment. If this is true, the thought goes, then the rule of law by itself cannot always be morally valuable.⁴² Instead, the rule of law would have value only when the legal system giving effect to the rule of law was substantively just, which is to say that the rule of law would have merely potential value. It is important to recognize that this objection misses its mark. If one were living in a deeply unjust society, whose injustice was codified in law, then one’s interests, characterized along one dimension, could indeed be served if the rule of law were abandoned insofar as one would at least have a chance of escaping consistent application of substantively repellant laws. But there is more to one’s interests than what will or does happen to one. One also has an interest in having a certain moral status, that is, in being owed treatment of a certain kind. Having a certain moral status is reasonably conceived as among one’s interests because it makes one worthy of respect, regardless of whether, as a matter of fact, one is respected.⁴³ Conceived along the dimension of moral status, one’s interests would not be served, at least not entirely, by dismantling the rule of law. MacCormick is right to claim that ‘[t]here is always something to be said for treating people with formal fairness’, and the reason is that the formal fairness demanded by the rule of law expresses a view about the moral status of people, namely, that persons are ⁴² We thank William Edmundson for raising this point. ⁴³ See FM Kamm, ‘Inviolability’ in PA French, TE Uehling Jr and HK Wettstein (eds), Midwest Studies in Philosophy vol XX (University of Notre Dame Press, 1996) 165–75. For more expansive treatment of moral status so conceived, see Kamm, Morality, Mortality: Rights, Duties, and Status (New York, 1996) Pt III.

Moral Evaluation and Conceptual Analysis

71

worthy of being treated in non-arbitrary ways. The rule of law, quite apart from the content of particular laws, responds to the status of those subject to it as being owed treatment of a certain kind. In fact, nothing about the content of the laws whose application is formally fair can diminish the value of their formally fair application in accordance with the rule of law. People are of course owed much more than formally fair, non-arbitrary treatment—substance and not just form matters. But to the extent that people are owed non-arbitrary treatment, form does matter. It is true that fewer people might have substantive injustice visited upon them if a morally bankrupt society abandoned the rule of law, since some would be excepted from substantively unfair treatment, but if that came to pass, it would follow that the society no longer embraced the formal respect for persons that the rule of law entails. And that would constitute a moral loss. Perhaps that loss should be sustained in some cases—that multiple dimensions of value constitute one’s interests does not entail that trade-offs are not possible across those dimensions. In all cases, though, such a trade-off would remain a moral loss. That the rule of law’s value is defeasible does not entail that it is not always valuable. Returning to the line of argument for normative jurisprudence that Coleman hypothesizes only to reject, we conclude that he is mistaken to construe law’s commendation feature as purely instrumental, being based only on the law’s potential or capacity for good governance. Law surely has such a commendation feature, but it also has the non-instrumental commendation feature of instantiating the rule of law and all of its virtues. It is in this sense that we submit that the concept of law may be inescapably normative, requiring moral argument to explicate its content.

III. Conceptual Analysis in Jurisprudence Quine’s thesis regarding analyticity, which he described as the first of two dogmas of modern empiricism,⁴⁴ was a denial of what he referred to as ‘some fundamental cleavage’ between truths which are analytic (true in virtue of meaning) and truths which are synthetic (true in virtue of facts). Quine’s claim⁴⁵ was that there is no principled way to distinguish analyticity from mere co-extensionality. Quine’s argument was quickly met with a rebuttal by Grice and Strawson.⁴⁶ Grice and Strawson argued that if talk of synonymy is meaningless then so too must be talk about the meaning of a sentence. They pressed their point this way: Is all talk of correct and incorrect translation of sentences of one language into sentences of another language meaningless? It is hard to believe that it is. But if we do successfully ⁴⁴ Quine, ‘Two Dogmas of Empiricism’ n 14 above, 20. ⁴⁵ ibid 27–32. ⁴⁶ Grice and Strawson n 24 above.

John Oberdiek and Dennis Patterson

72

make the effort to believe it, we have still harder renunciations before us. If talk of sentence-synonymy is meaningless, then it seems that talk of sentences having a meaning at all must be meaningless too. For if it made sense to talk of a sentence having a meaning, or meaning something, then presumably it would make sense to ask ‘What does it mean?’ and if it made sense to ask ‘What does it mean?’ of a sentence, then sentence synonymy could be roughly defined as follows: Two sentences are synonymous only if the true answer to the question ‘what does it mean?’ Asked of one of them, is a true answer to the question, asked of the other.⁴⁷

Quine, it seems, thought that Grice and Strawson were right for, as Scott Soames describes, ‘four years after this criticism appeared, he published Word and Object, in which he does advocate . . . giving up meaning and translation entirely.’⁴⁸ Quine does, in fact, give up on meaning. Quinean naturalism mandates that talk of ‘meaning’ and cognate terms (eg ‘synonymy’ and ‘analyticity’) be jettisoned in favour of tabulations of correct and incorrect responses to stimuli. All talk of ‘meaning’ can be replaced with a naturalistic account given by way of behavioural stimuli. As Quine puts it: ‘The stimulus meaning of a sentence for a subject sums up his disposition to assent to or dissent from the sentence in response to the present stimulation.’⁴⁹ So, let us return to legal philosophy and ask what all of this means for the methodology debate. Leiter argues that Quine’s supposed demolition of the analytic/synthetic distinction makes way for naturalism in legal philosophy, for like philosophy generally, legal philosophy must go empirical and be absorbed into science, social science included. On the outside looking in, according to Leiter, is conceptual analysis, analytic general jurisprudence’s traditional method of inquiry: ‘. . . if analytic statements are gone, then so too is conceptual analysis: since any claim of conceptual analysis is vulnerable to the demands of a posteriori (ie empirical) theory construction, philosophy must proceed in tandem with empirical science, not as the arbiter of its claims, but as a reflective attempt at synoptic clarity about the state of empirical knowledge.’⁵⁰ Leiter’s claim depends, of course, on eliminating the distinction between conceptual and empirical questions. But what is conceptual analysis, and is it really vulnerable to Quine’s attack, and in his wake, Leiter’s? In our view, the key mistake that Quine and, by extension, Leiter make is to treat conceptual analysis as one particular type of philosophical work, and one, moreover, that depends upon a robust conception of the analytic/synthetic distinction. But there are at least four distinct, overlapping methods of conceptual analysis that are not challenged in any serious way by Quinean naturalism, and it is worth considering them each, if briefly, in order to assess whether anything more than what they offer is needed for jurisprudence as traditionally practised. ⁴⁷ ⁴⁸ ⁴⁹ ⁵⁰

ibid 201. S Soames, Philosophical Analysis in the 20th Century (Princeton, 2003) 377. WVO Quine, Word and Object (Cambridge, MA, 1960) 40. Leiter, ‘Beyond the Hart/Dworkin Debate’ n 25 above, 45.

Moral Evaluation and Conceptual Analysis

73

On one prominent conception of the enterprise, ‘conceptual analysis is the very business of addressing when and whether a story told in one vocabulary is made true by one told in some more fundamental vocabulary.’⁵¹ The point of this modest form of analysis is to show whether and how one set of terms is reducible to another, or what to say about a posit in an old theory when a new theory comes along. Frank Jackson and Philip Pettit explain this approach as follows: Take the new theory to have the truth of the matter—or at least more of it than the old theory. The question of the fate of a particular posit of the old theory is then decided by considering how the posit was conceived in the old theory and seeing whether something satisfying that conception nearly enough is to be found in the new theory. If the answer is yes . . . the old posit survives; if the answer is no, as it is for impetus, phlogiston, caloric fluid, et al., the old posit is eliminated.⁵²

Jackson suggests that this method is in fact quite close in practice—even if not in theory—to Quinean ‘paraphrasing’, where ‘[t]he objective would not be synonomy, but just approximate fulfillment of likely purposes of the original sentences . . .’.⁵³ A second type of conceptual analysis is what we might call ‘therapeutic’. In this endeavour, the philosophical task is to identify and eliminate claims or assertions that lack sense. Take the claim ‘to understand a legal text, one must interpret it.’ This claim expresses the familiar position within the Hermeneutic tradition that ‘all understanding is interpretation’, such that when we read a text, or hear an utterance, we achieve understanding by taking the text or utterance and translating it into something meaningful. We can conceptually analyse the claim ‘all understanding is interpretation’ therapeutically by noting that it leads to an infinite regress: if every ‘text’ requires interpretation before its meaning can be gleaned, then why stop at the initial text?⁵⁴ The ‘interpretation’ of the text will itself require interpretation before its meaning can be discerned, and so on infinitely. Hence, there must be something wrong with the idea that ‘all understanding is interpretation.’⁵⁵ And in facilitating interrogation of the claim, it is conceptual analysis that illuminates this problem. A third form of conceptual analysis is what might be termed ‘conceptual theorizing’. Kenneth Himma provides a good account: Conceptual analysis attempts to theorize . . . intuitions or core conventions by identifying deeper metaphysical commitments that are either implied or presupposed by these ⁵¹ Jackson, From Metaphysics to Ethics: A Defence of Conceptual Analysis n 16 above, 28. ⁵² F Jackson and P Pettit, ‘In Defence of Folk Psychology’ in F Jackson, P Pettit, and M Smith, Mind, Morality, and Explanation (Oxford, 2004) 33. ⁵³ WVO Quine, Word and Object n 49 above, § 46. See Jackson n 16 above, 45. ⁵⁴ See L Wittgenstein, Philosophical Investigations § 198 (GEM Anscombe (trans), 2nd edn, Oxford, 1967), ‘“But how can a rule shew me what I have to do at this point? Whatever I do is, on some interpretation, in accord with the rule.”—That is not what we ought to say, but rather: any interpretation still hangs in the air along with what it interprets, and cannot give it any support. Interpretations by themselves do not determine meaning.’ ⁵⁵ For detailed discussion, see Dennis Patterson, Law and Truth (New York, 1996) 71–127.

John Oberdiek and Dennis Patterson

74

intuitions, as well as more general principles that would explain these intuitions or conventions. While this might or might not be a distinctively philosophical enterprise, it goes well beyond the empirical task of identifying shared intuitions or core conventions.⁵⁶

Now, there are two ways of doing the kind of conceptual analysis Himma imagines. The difference is marked by the word ‘metaphysical’. Leaving that word in this passage, on a traditional approach, the activity is classic, reductive metaphysics—reducing one set of terms to another. The metaphysical aspect is filled in by commitment to some set of necessities that underlie the terms in questions. But there is another way to do this form of conceptual analysis, one that does not require the invocation of metaphysical commitments. In this ‘lite’ version of conceptual analysis, the task is not metaphysical reduction but the identification of presupposition through redescription. Consider, in this regard, Raz’s ‘service conception’ of authority, in which the law plays a mediating role between right reason and persons, serving those on whom legal authority bears by claiming to guide them better than they could guide themselves. Raz offers this conception of authority as a way of identifying one distinguishing feature of legal norms. Raz does not claim, nor need he, that the service conception of authority is ‘metaphysically necessary’. All he need do is claim, as he does, that this conception of authority is presupposed by ‘our’ concept of law.⁵⁷ A fourth and final version of conceptual analysis untouched by Quine’s critique is closely related to this third, ‘lite’ version, and consists, as Coleman puts it, in ‘uncovering the most salient features of the concept: those that figure most prominently in an explanation of the kind of thing it is the concept of—that are central to our understanding and appreciation of it.’⁵⁸ The goal here is not to find analytic truths, or essential features of concepts, but to discern what in the concept under consideration warrants picking out the concept as the concept that we do.⁵⁹ This form of conceptual analysis is, in many respects, continuous with Quinean naturalism—indeed Coleman attributes his approach in part to Quine.⁶⁰ Its roots notwithstanding, this method of inquiry is plausibly understood as conceptual analysis. Discerning which features of a concept warrant according the concept its particular role in our discourse must consist in analysing that concept. What this very brief overview of forms of conceptual analysis illustrates is, first, of course, that there is no one form of conceptual analysis but that, second, conceptual analysis need not take any form in practice that conflicts with Quine’s arguments in ‘Two Dogmas’. Our earlier points about the possibility of truth in virtue of meaning alone suggests that Quine’s arguments do not establish—or ⁵⁶ ⁵⁷ ⁵⁸ ⁵⁹ ⁶⁰

KE Himma, ch 1 above. See Raz, ‘On the Nature of Law’ n 23 above, 1–12. Coleman, The Practice of Principle n 1 above, 179. See also B Zipursky, ‘Pragmatic Conceptualism’ (2000) 6 Legal Theory 457. See Coleman n 1 above, at 6 and 9.

Moral Evaluation and Conceptual Analysis

75

rather, undermine—all that they claim to. The present synopsis of forms of conceptual analysis strongly suggests that, even if Quine’s arguments are sound, they do not put conceptual analysis out of business, but rather merely one particularly strong version of it. This overview implies a third point as well, and that is that it is simply not clear that one can get by without relying in some degree on one of these forms of conceptual analysis in understanding law. Indeed, Coleman has noted that Leiter himself must do a good deal of conceptual analysis even to frame his own position. As Coleman puts it, ‘the naturalized project can proceed only after a good deal of philosophical spadework.’⁶¹ Focusing more specifically on Leiter’s work on adjudication, Coleman argues that ‘[t]he very idea of an adjudicatory function operating on authoritative legal standards presupposes that we have criteria for determining what the authoritative legal standards are. Thus, the naturalist is committed as a conceptual matter to the existence of a test of legality.’⁶² Similarly, Leiter’s own commitment to exclusive legal positivism,⁶³ and his claims of its compatibility with Americanstyle legal realism, are conceptual claims.⁶⁴ The point here is that a naturalistic account of law cannot replace but in fact presupposes a conceptual understanding of law. It is just not clear, at the end of the day, whether Leiter recognizes that there are more forms of conceptual analysis than that which Quine had in his crosshairs. Nor is it clear whether Leiter recognizes that, at least in the main, little in jurisprudence depends upon Quine being mistaken. Of course, in light of his own conceptual presuppositions about law, noted above, Leiter would seem to be committed to there being at least one ‘acceptable’ form of conceptual analysis. But Leiter exegesis is not our goal.⁶⁵ What is important to establish, by our lights, is that there are many forms of conceptual analysis besides that which Quine argues is bankrupt, and that jurisprudence as traditionally conceived can, and in fact, always has, helped itself primarily to those secure conceptual resources.⁶⁶

⁶¹ Coleman n 1 above, 213. ⁶² ibid 213–4. ⁶³ B Leiter, ‘Legal Realism and Legal Positivism Reconsidered’ (January 2001) Ethics 286. ⁶⁴ See H Oberdiek, ‘Legal Realism: What Does Quine Have to Do With it?’ unpublished manuscript 26. ⁶⁵ For thorough discussion of Leiter’s work on naturalism in jurisprudence, see H Oberdiek ibid. ⁶⁶ In his new book, Naturalizing Jurisprudence (Oxford, 2007) 196, n 49, Brian Leiter comments on a draft of this article, with specific attention to the four forms of conceptual analysis we deem invulnerable to the Quinean critique of Carnap’s distinction between analytic and synthetic propositions. Leiter asserts that we ‘somewhat bizzarely’ claim that two of our four methods are impervious to the Quinean critiqiue. He further asserts that two of the methods are ‘obviously vulnerable granting the soundness of Quinean naturalism’. It is, unfortunately, impossible to assess Leiter’s claims as he provides no arguments to support them. It is perhaps worth noting for the interested reader, in addition, that in referring to the four methods by number and not substance, Leiter appears to confuse them with one another.

5 Objectivity and Value: Legal Arguments and the Fallibility of Judges Stephen Guest*

The force of objectivity lies primarily in the idea that our judgements about matters of value are liable to correction. Objectivity is an important idea. It is crucial for our understanding that we can be fallible, and thus to learning the limitations and strengths, as well as the importance of our judgements. But when we make value judgements as opposed to judgements of science it is usual to suppose that there is nothing ‘in the real world’ that provides any standard for correction. The argument here is that through the idea of imputation, we are able to see how objectivity is the result of a creative endeavour that imputes correctability to value judgements. Since the activity of stipulating only makes sense where it institutes a value, this argument for objectivity provides an abstract scheme for all significant judgements of value. The value of objectivity lies in its providing a sense in which personal judgements, when expressed in actions, are ones to which at least some other people can respond and make corrections. A further argument here is that making good moral judgements is a skill fundamentally dependent upon internal engagement between knowledge of others and oneself. It consists of seeing others from the ‘inside’ and as more than mere external reflections of ourselves; a person exercising such a skill must be capable of judging that others are ‘like me’. It also requires a sense that the other ‘can be learned from’. None of this sociology of a reflexive/self-reflective process is externally determined. The standards generated by the process are not ‘absolute’ or ‘timeless’ but rather arise from an empathetic understanding that ‘other people are like me’. It would be an affront to rationality to refuse to see important similarities between oneself and others. To understand this requires care and reflection; this is why it is important to understand moral judgement as a skill. As a way of bringing together these two ideas of imputation of objectivity and of reflective and argumentative skill, the conclusion analyses some recent comments by a prominent judge. Deciding questions of law in controversial appellate * The author gratefully acknowledges the support of the British Academy and the Leverhulme Trust in appointing him Senior Research Fellow for 2006–07 thus enabling him to undertake this research.

Objectivity and Value: Legal Arguments and the Fallibility of Judges. Stephen Guest. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Objectivity and Value

77

cases requires making judgements of moral value.¹ Lord Bingham, the senior law lord and former Chief Justice of England, has recently claimed that there are no right answers to legal questions. The argument in the last part of this paper is that he, and others who share his view, are wrong. Th is is partly because they underestimate the vitality in the assumption that their ultimate judgements might be fallible. Perhaps, too, it is because they do not sufficiently appreciate how legal argument is a different skill from rhetoric, where ‘being correct’ is of little, if any, importance. Law is argumentative, through and through, and that is important, since, apart from all else, it justifies the whole of the appellate procedure on matters of law. Given this value, there are compelling reasons for supposing that the moral arguments that judges such as Lord Bingham advance in support of their decisions are intended to establish right conclusions.

I. Imputation and Truth The nature of truth is not a central issue of moral objectivity. Truth is primarily a property of statements or propositions. Take the phenomenon of lying, which exploits the trust induced in others that speakers normally intend their statements to be expressive of their beliefs. The liar expresses as true what he does not believe.² Questions of objectivity, however, concern the correctability or otherwise of judgements. Because ordinary language is so imprecise and we often use ‘true’ as a synonym for ‘correct’ or ‘right’, the nature of truth does not pose too much of a difficulty, although, where matters of value are concerned, people become anxious about any talk of ‘truth’. Perhaps a quest for ‘truth’ sounds like a vital step in philosophical thought—that we have to ‘get at the truth’. But if we accept that moral judgements may be right, we then have to accept that statements expressing those judgements are true. We can distinguish whether an act is right from a statement or proposition that it is right, so we can distinguish ‘right’ and ‘true’. The issue is therefore not about the nature of truth and morality, but about whether moral judgements can be objective, and that question concerns whether such judgements are capable of being corrected. Nevertheless, we can look at judgement and truth in a way that draws out a creative function in making judgements of value. Stipulative truths are truths we create by imputing truth to some expressed act. We do so without directly depending on a correlation between what we stipulate and some existing feature of the world, for we can make it true that ‘A=2B’ by pretty much declaring it to be so. A=2B is not true by virtue of its describing our use of language for there may be no such use. And by relying on other stipulative definitions, which need ¹ According to all contemporary theories of law, judicial arguments in controversial cases involve making moral judgements one way or another. ² And so dogs and cats, which do not have a language, have no means by which to tell lies.

78

Stephen Guest

not coincide with language use, as to what ‘=’ and ‘2’ mean, we can deduce further true propositions, that, for instance, A divided by itself equals one, or that ‘2A=4B’. The point is not just one of mathematics.³ Stipulative definitions suggest a way of understanding morality that opens up more clearly the possibility of correction and thus the possibility of objectivity. While stipulation stands as true in a formal, although somewhat empty sense, its legislative nature makes the reasons for the stipulation significant. It does this more surely than by looking to the practical reasons that people engaged in any particular practice actually provide themselves. One reason is that it conjures up the idea of a single author whose reasons for stipulating we can challenge. Merely to suppose that there is a justification ‘in the practice’ is confusing because people have different and conflicting ideas as to what justifies a practice. Those engaged in a practice characteristically provide different reasons for their engagement. Sometimes practices can ensnare and entangle people from which they remain detached (for example, they become subject to totalitarian rule against their wishes). Mostly, however, people engage in practices but see those practices in different lights. The temptation to suppose that there is an explanation or justification for the practice ‘embedded in’ the practice somewhere is great. It represents a philosophical and creative conservatism; it builds upon human experience, put into practice, and there is always the possibility of modification, radical re-interpretation, and even abolition. It is nevertheless important to remind ourselves that justifications for a practice can be ideal. They can pre-date the practice; they can exist in books without any hope of our putting that ideal in place. Stipulation does not require that there even be a practice. Stipulative truths, as opposed to dictionary truths, do not rely on the actual practices of human beings. There is a definite but barely perceptible merger between the stipulative nature of language and the idea of proposing. Proposing, perhaps better called ‘creating’, is something that we can accomplish without using language. We make proposals for our own lives, in shaping the course of our characters and our relationships with other people, without doing so by the signifying to others of our intentions. On the other hand, creativity suggests very keenly the making of something and, given that it is natural for us to be social and to express to others our creations, it is also natural to associate the creative function with the production of language-cognate ³ When we argue by reducing something to absurdity (an argument reductio ad absurdum), we employ hypothetical stipulation in the first steps. Take the argumentative put-down ‘If pigs could fly . . .’. A formal way of putting it is: ‘Let us stipulate what you said is true. It would follow that pigs could fly, which is obviously false. And so it follows that what you said could not possibly have been true.’ The argument makes use of a stipulated truth, only to conclude that, on that assumption, a contradiction follows, showing the initial stipulation to be false. Imagining it to be true is insufficient, though. To enter the argument, we have to ‘make the assumption’ that something is true, and so it (hypothetically) becomes true by stipulation. Th is is one way of showing the intimacy of the connection between imagination and truth; you imagine it to be so, then you make it so, and in the case of reductio the argument shows the imagined state of affairs to be false.

Objectivity and Value

79

materials. Or, perhaps putting this round the other way, that it is natural to suppose that works of art, and other forms of creation, are for this reason languagecognates. There is, at any rate, a link between stipulative definition, which is a function of language, and the creative function. It is evident in the way that we naturally describe artistic works as ‘true’, or say such things as that the music ‘speaks to us’ or expresses ‘profound truths about human nature’. What we mean is that the work of art is like a statement of truth about some aspect of the world that is different from what we could present as an empirical description. It is this lack of direct connection with empirical description that makes it easier to ascribe truth through the idea of stipulation rather than through the dictionary to propositions about matters of creative value. That is to say, while it is easy to ascribe truth through stipulation it would be hard to do so through dictionary definition regarding things that have creative value. We can propose as true that Beethoven’s last string quartets are profound and rely on the idea of stipulation to provide the sense of truth, untrammelled by any requirement of empirical objectivity. This is true whether we are looking for the correct use of the word ‘profound’ or for empirical facts that support this proposition. The proposal stands or falls on the merits of the artistic judgement. We can go further. If we can connect truth to judgements of value—proposals—then that suggests the possibility of ascribing objectivity to value. Stipulation and value connect just because it makes sense to ask for the value of any stipulation. A common form of stipulation is the military code where the value of a good code—a good stipulated meaning—is its resistance to being broken. We find another, less discernible form of stipulation in the corporate logo, where an institution stipulates some symbol to be representative of what it aims to achieve. The corporation makes it so by proclaiming—stipulating—it to be so. The value of the logo is the test for whether the corporation achieves its end. In mathematics, stipulating some particular function of a symbol will have value, perhaps efficiency, or just elegance. So if a mathematician stipulates ‘A=x’ and then intends to use ‘x’ according to its conventional meaning of ‘multiplies’, then the stipulation is impractical. It is inefficient and inelegant for it does the subsequent calculations a disservice by creating ambiguity. ‘Axx’ could mean ‘A multiplied by x’ or it could mean ‘A multiplied by x multiplied by x’ and these produce different answers. The lack of value in the stipulation will merge with the truth function at some stage, for a stipulation that failed to have any value at all could be neither true nor false; the truth of a stipulation cannot be merely the truth of the fact that someone stipulated. Rather, the stipulation is that something is true. But if there is absolutely no value in saying something is true, and we assume, reasonably, that stating things to be true is stating a (non-circular) value, then it must be that both language and value are functions of stipulation. None of us seriously supposes that there is value in all things. We find some things have no value at all. Would there be any value in stipulating that ‘A=A’, that something equalled what it equalled? Consider the value, for example, in stipulating that it

80

Stephen Guest

was true that ‘A=not-A’? How could there be value supposing it to be true that something was false? The answer to that is that the extent of artistic value— creativity in general—in some sense has no bounds. To say there is truth in contradiction, as a proposition concerning of the current state of western society, or of mankind, is an underlying theme of movements such as nihilism and surrealism, and others. It has a sort of sense. Creation is an intentional act and it is difficult to see what the point would be of intentionally creating something of no value at all. We must think of a case, unlike the case of stipulating contradiction that, truly, has no value. At any rate, it cannot make sense to say that something that has no value is true. There is therefore a gap between objectivity and truth in the case of science that is not present in cases of value. No stipulated truth is true where there is no value in stipulating it thus. So a contrast emerges between empirical and value judgements. While the judgement of perception when we see the stag and calculate its distance from us is different from our statement that ‘the stag is 100 metres away’, our judgement that this is the most valuable way of seeing something or doing something, is not so different from the statement that seeing or doing it this way is true. While there is a clearly discernible gap between physically perceptive judgements and statements concerning their truth, there is no comparable gap between value judgements and the truth they express.

II. The Sociology and Objectivity of Value The above provides a solution to the much-discussed question whether moral arguments are ‘subjective’ or ‘objective’, and whether they are ‘cognitive’ or ‘real’. Stipulation properly understood provides a case for saying that value judgements can be more than just asserted trivial truths. Value judgements are true or false because they consist of proposals the truth or falsity of which turns on the value judgements that support them. ‘The right judgement is the one that is supported by the right reasons’ is not circular but explanatory: the definiendum contains more than the definiens. When working out whether a work of art is good, the problem of objectivity is no more than that of looking to the best reasons that support that judgement; the same goes for moral and legal judgements. So the invocation of stipulation provides an answer to the moral sceptic who says that value judgements cannot be right or wrong just because there is no empirical evidence in support. Our response need only be to confront him with a moral problem and point out that his response—there is no meaning in what we do when we judge something good or right—provides no reason for judging at all. Consider alternatives to the sceptic. Take the question whether abortion is always wrong. Is it just that people ‘prefer’ that it never happen or, if it is not the same thing, is it that they have a ‘subjective’ feeling that it is always wrong? Many people would say that this way of looking at morality is ‘honest’. It is common, too, for people to say that ‘for them’ abortion is always wrong but of others

Objectivity and Value

81

who disagree, that ‘it is true for them that abortion is sometimes permissible.’ The mixed intuitions so often articulated in this way are confused. When pressed, people are generally not willing to admit that their moral views are merely matters of taste, such as what they like to eat, but they will reject what they perceive to be the arrogance of someone else making that decision; in fact, this seems to be one of the determining pivots of discussion outside professional philosophical discussion.⁴ We can deal with these two widespread views—‘subjectivity’ and ‘arrogance’ with comparative ease. First, it cannot follow from the fact that something seems morally right to us, individually, that all moral views are ‘subjective’. All beliefs are subjective in this sense: no one can say that we do not believe something if in fact we do. Yet beliefs do not embody right judgements merely because someone believes those judgements to be right. Second, it does not follow that when people announce their moral views that they are being arrogant. Some people overstate their belief in themselves, true, but they could also be stating the truth. In fact, relatively few people are arrogant in this way. Maybe we need to preserve ‘arrogance’ in this context for fundamentalists, or for dictators, or preachers; but the ordinary person who ventures a view about what is morally right is not being arrogant. At any rate, the charges of ‘subjectivity’ and ‘arrogance’ do not directly concern the appropriateness of how we should think about moral questions. They are, rather, conversational tactics, avoiding the confrontational possibilities of engaging in argument about sticky moral problems. Here is a suggestion. A start to having a serious discussion about what is morally right would lie in our belief that morality is objective. That is, we should believe that what it is morally wrong or right to do can be mistaken, or wrong, or incorrect, or revisable, particularly in response to what other people say, and to our rethinking our own experience. Part of the problem with the word ‘objective’ is that it is ambiguous. It suggests (a) that a person’s ‘subjective’ understanding can be wrong, and that people should not think that morality is a matter of ‘owning’ what they believe. More important, it suggests (b) that a person’s ‘subjective’ understanding can be wrong only where there is an ‘object in the world’ by virtue of which that person’s ‘subjective’ understanding is wrong. It is rational to suppose that we might be wrong in what we think we ought to do; it makes us careful—it allows for our thinking our judgement could be different, and better—and that view encourages the making of more reflective and therefore wiser moral decisions. It would encourage us to revise our initial view, and to consider the worth of reasons in favour of that view. If these reasons are sensible, then morality is objective and the alternative, ‘pure subjectivism’ gets us nowhere. The irony is that, if we accept this simple argument, we can return the charge of arrogance to those who believe in ‘subjective’ morality. If our views ⁴ The majority of law students, in my experience, think that morality is ‘subjective’ and that to think otherwise is ‘arrogant’.

82

Stephen Guest

about matters of moral value were entirely subjective, our statements of what is morally required would be infallible, because to be right we need only report what, as it were, ‘our mind says’. In that light, in claiming objectivity for our view of what morality requires shows we recognize we could be wrong; we recognize the possibility of correction of our view. That exhibits humility, not arrogance. People like proof. It appears in scientific understanding, but not for ‘value’ judgements. Why do many scientific judgements require proof but not moral, or artistic judgements? We have a personal and overwhelming interest in stability and constancy in the outside world, in our being able to predict what will happen and so our survival depends on it. We have what we could call an internal reason, too. Our sense of our freedom depends on our having a relatively stable view of what we can do, of our own identity and the world that immediately surrounds us. With the certainty that comes with science, we can predict what we might do, or what might happen to us. So the proposal that we view science as that body of knowledge that attaches significance to certainty of prediction makes good practical sense. But stipulation implies choice, and it seems outrageous to suppose that we ‘made it true’ that science is a branch of knowledge that requires proof. But why should it be outrageous? If we make that stipulation for a branch of knowledge, we shape its contours by reference to our particular practical interests. It is not as if by stipulation we are making the world ‘something that it is not’; rather, we are bringing out the sense of what we should regard as significant, and the success of the stipulation is measured by reference to the practical consequences of supposing that ‘such and such propositions of science are true’. So we can say that there must be practical purpose involved in stipulating proof for science. What about morality? Would it be a good proposal to stipulate proof or demonstration as the truth condition of propositions of morality? That is unappealing. The only way one could even start thinking about what it would mean to prove a particular proposal about moral conduct would be by pointing to a convention of moral behaviour, or to some promulgated rule or pronouncement—what legal philosophers sometimes call a ‘social fact’. The content of such conventions, facts or pronouncements, would be relative to the beliefs of those engaged in them. But such conventions, facts and pronouncements could not exist independently of the reasons supporting them, and so we would have to prove the content of those conventions and so on; this suggests an infinite regress. In any case, morality governed by proof would dismantle morality’s hold on us. It would privilege certainty over all else. What would be the point in that? First, it is desirable that people themselves judge what is right by reference to content, and by reference to themselves and others, and not by reference to conventional behaviour. Certainty is not as clearly in our interest as prediction is in science, given the array of everyone’s actual and possible actions and the enormous variations in natural events. Second, it is desirable that people are responsible for their actions, an idea partly implicit in their having a point of view and thinking ‘for themselves’; these are important components of what it is to be free. Both these fundamental

Objectivity and Value

83

requirements of moral reasoning assume that people should make an independent check to see whether they are making a mistaken judgement. What, then, of the opposite proposal, that morality be a matter of ‘pure subjectivity’? The truth conditions for moral propositions would be merely that a person believes it to be wrong. That is clearly a non-starter because of the irresolvable conflict it would generate, for all points of view, however much the conflict, would be right (which would mean that ‘right’ would not signify much, if anything).⁵ Because of this, it is necessary to add to the idea a rider that the consequence would be that it is wrong only for that person.⁶ The question is not that this is absurd by any analysis of the way we treat morality but whether this would be a sensible idea for us to adopt. But if we did adopt this idea, it would stop argument dead: there would be no more to moral argument than the registering of a belief.⁷ Take a ‘moral scientist’ who argues that the important question is the ‘metaphysical’ status of morality’s claims to objectivity. Scientific work, the moral scientist says, has vindicated its own ‘metaphysical’ status, and so the tests of scientific objectivity should apply equally to moral judgements. He says perhaps something along the lines of ‘. . . the only tenable guide to the real and the unreal is science, and the epistemological standards we have inherited from successful scientific practice.’ And what ‘vindicates’ science’s ‘metaphysical’ status? ‘It ‘delivers the goods’, he says.⁸ The argument relies on that value, and does so at the cost of ignoring other values that we might weigh in the balance. If ‘delivering the goods’ as far as prediction is concerned is valuable, why are other values not only minimally respected but irrelevant? There is a plausible case for saying that ⁵ If religious belief includes knowledge by virtue only of strength of belief, and so not subject to correction, then it must surely be ruled out as a ground for morality. Why accept a God, belief in which entailed irresolvable conflict? Religious belief, defined this way, is different from belief in ultimate premises, such as that ‘life is precious’ or ‘people must be treated with respect’, because these are reasons that can be seriously countermanded. ‘Life is precious’ can be countermanded by reasons referring to the different ways life can be precious (eg in the case of abortion). Those countermanding sorts of reason are just not available where someone declares that their belief is ‘What they understand God to say’; that reverts to belief, and that means conflict. Is it impermissible to suggest that one could not believe in a cruel God? Or that one’s belief in God should be determined (or abandoned) according to the value that holding that belief would contribute to ours, and others, lives? Most religious people would disagree that religious belief entails conflict, in spite of the empirical evidence; but they need more of an argument than that which supports their own religious belief, namely, that it is just intensity that marks it out from other beliefs. ⁶ Evidenced in statements such as ‘I think abortion is always wrong but that’s just my opinion and that woman who is having the abortion, well, that is right for her.’ This rider is inconsistent with religious belief. ⁷ Students have increasingly accepted the ‘subjectivity of morality’ over the past two decades. Th irty years ago, it was always possible to have a class argument about the rights and wrongs of abortion, with the class divided on how certain specific cases should be resolved. Nowadays that is rare: students will say ‘it depends on your own point of view’; when they are reminded that our beliefs can be mistaken the retort usually is that it is arrogant to suppose that one person could tell another what is the right or wrong way to act. ⁸ Quoting Leiter, ‘Law and Objectivity’ in The Oxford Handbook of Jurisprudence and the Philosophy of Law (OUP, 2002) 969, 985.

84

Stephen Guest

good moral judgements have ‘delivered the goods’. It is not enough just to say that if the methods of science are applied to morality then we find there is no counterpart in the world that would predict what the morally right point of view is. That suggests that the right value in place is that of prediction. So what value is there in this argument? A grander attack on the moral scientist confronts his claim that scientific method has the value of ‘delivering the goods’. This judgement is an explicit value judgement and so is not subject to scientific method.⁹ Therefore, an account of the primacy of scientific method comes from another domain; it looks fair game for the domain of morality. To match the iconoclastic ‘delivering the goods’, we can easily say that science has failed miserably to ‘deliver the goods’. It has delivered us global warming and nuclear waste, while morality delivered us with democracy, rights and toleration. This is rhetoric, of course, because on each side ‘what is delivered’ has been selective. The main point is that the moral scientist can only be convincing on a shared supposition that science is the dominant, or only, form of knowledge. The argument about delivery certainly does not prove that the domain is dominant.

III. Moral Judgements as a Skill If morality were only ‘I think it so’, it is not clear that anyone could challenge our views, even by reference to reasons that we have previously announced that are contrary to our views. One can argue that ‘what a person thinks’ must take into account what ‘the person as a whole has said’ but this smacks of the artificial; she could always say ‘I have changed my mind.’ So if the thought alone counts, we would not be responsible for not having responded to reasons, and so we would not be responsible if our act was wrong. If morality were subjective in this sense (‘pure subjectivity’), there would be no need to give reasons at all. One would just act. One’s own view, however arbitrary, would be the right view and that would mean one is infallible. (Here is arrogance of just as grand a kind as supposing one were always objectively right.) It would follow from this view that there were no rights and wrongs to morality at all. But responding to correction, seeking advice, changing one’s mind, acknowledging one’s error, are all either actions or preliminaries to actions in the form of disposition-altering thoughts. Such attitudes of preparedness to argue, and to accept reasons for one’s action, are attributes of personal responsibility. Having these attitudes means you think that you need to justify your action, at least to ⁹ Richard Dawkins would strengthen his arguments about the paucity of religious belief by an acknowledgement of his own values. The argument against religion need not attack religious values; it is only necessary to attack the reason-less element in religious claims. See Dawkins, The God Delusion (Bantam Press, 2006).

Objectivity and Value

85

the extent that it is in conformity with reasons that you have already subjectively considered. As soon as we talk of a person’s response to reasons, the objectivity— the correctability—of judgements begins. Reasons are subject to what are established practical requirements of rationality. They must be consistent and they must incorporate some minimum requirements of factual accuracy. And so even people supposedly in the ‘subjective’ state of possessing their own private reasons can be intellectually embarrassed by other people: ‘your own reasons don’t stand up’ they can say.¹⁰ There are also important moral reasons that have special status. This is where people say that they just accept something as having ultimate or supreme value such as ‘treat people with respect’, or ‘you should never hurt anyone’, or ‘always respect life’ and leave it at that. Often people describe such reasons as their ‘ultimate premise’ or ‘fundamental belief’ and many connect that idea with their religion. It is a form of reason that is hard to attack; it usually does not rely on historical facts and is too abstract to give rise to a charge of inconsistency. If these are truly ‘one’s own reasons’ it is here that they lie, and they constitute a kind of subjectivism. In this area there is a truth lurking in the idea that moral views in a common way of speaking ‘are true for you’.¹¹ Moral judgement can be challenged not only on ground of consistency or factual weakness, but also on the ground that they fail to reflect one’s sense of what other people are like. People’s ideas of who they are do not arise, at least solely, from within themselves. This idea expresses itself most naturally in the form of a metaphor, that of correcting oneself by reference to one’s reflection, as we might see ourselves if we could somehow abstract those parts of us that are ours alone. I can be internally consistent, I can make good inferences from fact, my reasons are my own, not another person’s—I do not just report what other people think— yet the resulting moral view is still correctable if I am not ‘true’ to what I am like as a human being. Take someone (perhaps like Nietzsche) who is consistent in his belief that the best moral life is led by those who are intolerant of physical and mental weakness and lack of discipline, who is himself physically and mentally strong, and who supports institutions and creeds that promote and entrench the superman who is physically and mentally strong. Can we attack his belief— advanced as a moral viewpoint better than the beliefs of others—providing reasons that could appeal to this person, reasons that we can justifiably say to him that he is wrong? Nietzsche represents significant strands in political movements, some that became well-established. We can use racism as an example. People are consistent in their belief that certain ethnic groups deserve lesser treatment than other groups. They do not always rely on historical facts. There is a difficulty in criticizing their view as logically inconsistent because there is nothing in logic ¹⁰ Devlin’s argument that we find out what morality is by asking a cross-section of the community what they believed overlooked the argument that their beliefs in what was morally right might be rationally unsound. See Dworkin, Taking Rights Seriously (Duckworth 1970) ch 10. ¹¹ See above, n 5.

86

Stephen Guest

alone that rules out, for instance, using the predicate of ‘being Jewish’ to describe how certain groups of people should be treated.¹² We cannot bring a charge that the bare logical distinction between different races is morally ‘arbitrary’. If you stick to your belief that, say, blacks are different from whites, and that different consequences attach to that description, you are open to various charges, true, but it is not the usual or clear case of being inconsistent. To attack the racist requires more. The attack requires something that gets at the racist’s own being, to require him to compare his own self with those he thinks different from him. This requires a kind of hypothetical psychological-moral self-insight which would be got to along the following line of reasoning. We can accept, as a generalized objective truth, that people are similar in various ways. Indeed, people are similar to the extent that our senses of what is ‘normal’ or ‘natural’ borrow from this generalized fact, in a relationship of empathy. So it is ‘normal’ or ‘natural’ to feel particular degrees of physical or psychological pain in response to particular degrees of intervention. It is ‘natural’ for parents to be upset by their child’s illness or death to varying degrees. It is ‘normal’ to respond to particular works of art, or to particular foods, or to feel repulsion at particular acts, such as torture. We would fail to develop much of a sense of human ‘normality’ without being prepared to accept some obvious truths for which no great deal of empirical inquiry is necessary. There are, as Hart said, echoing Hobbes, certain ‘truisms’ about human nature which it is ‘natural’ for us to accept as a matter of ordinary and contingent fact.¹³ Now imagine that people are identical in these different sorts of way. They in fact are capable of having exactly the same responses and feelings to every experience. Each person is capable of suffering precisely the same degree of pain, of feeling the same intensity of pleasure, and is capable of the same judgement about whether an artefact is beautiful or not. Since we are alike in more ways than we generally suppose it is not difficult to imagine this further step of our being exactly alike in our capacities. We would assume, too, that everyone knows that everyone has exactly the same capacities. In this imagined situation, it is a crucial question whether, if we knew others were identical in such ways, there would be any room for correcting other people’s judgements. The knowledge that we have the same capacities would provide a metric; that is part of the hypothesis. In fact, it is not a strange idea to suppose that we are exactly alike in many ways. Some people make this mistake, in fact and fail to give any other possibility a second’s thought and, as a result, can make impossible demands on others. The untrained nurse in a children’s home, for example, or a nursing home, or a psychiatric hospital, who has to be taught that the very young, the very old and the mentally disabled do not have the same sorts of capacity as ¹² ‘Being Jewish’ is morally relevant in various contexts; Nazi reparations, for instance. ¹³ The Concept of Law (2nd edn, OUP, 1994) ch IX.

Objectivity and Value

87

he himself has. Apart from these special cases, it is an intelligible thought experiment to suppose that we have the capacity to make these judgements but differ in experience, intelligence, maturity and our level of education. People will make different judgements but even according to this hypothesis there will be a corrective possibility. Surely judgements ranging from the moral to the aesthetic could be made: ‘think how the mother would feel, contemplating abortion’; ‘imagine being burgled’; ‘that can’t hurt that much’; ‘if you work at eating olives the bitter taste becomes the best bit’; ‘that perfume is too sweet, although it smells good at first’; ‘lack of vibrato in that quartet movement works best’; ‘you can be upset without realizing it’; ‘think what it would be like to be economically dependent on a spouse’; ‘The Birth of Venus stuns at first sight, but it is not the best of Botticelli.’ Many of these sorts of statement are part of our daily experiences. They are not just appeals to internal reasons that can be attacked on grounds of inconsistency or misunderstanding of the evidence available. Rather, they are appeals to a different sort of judgement. They work by way of getting another person to see that, within them, is the possibility that they can ‘come to see’ how things are for them, by seeing how another understands that matter. It is a form of reflexive understanding that allows correction. One should hesitate to call it either ‘external’ or ‘internal’; rather, it draws upon the essential likeness between people. The argument shows that there is at least the possibility of objectivity in the hypothetical situation where people are, as we might put it, subjectively the same. The strength of the argument arises from the hypothesis that we have identical capacities, which suggests the possibility of agreement in judgement. Correction gains its sense in this imagined situation from the possibility of being nudged in the right direction by the equally capable but more experienced and perceptive. The imagined situation establishes how people of similar capacities might correct the subjective judgements of each other without merely finding internal inconsistencies. That conclusion undermines the popular belief in ‘pure subjectivity’. The essence is that a person can be brought to understand that his initial judgement was right or wrong by the judgement of someone else because there is a non-subjective basis for correction. In other words, the ‘purely subjective’ judgement itself is insufficient even when logically consistent and relying on historically adequate grounds. Let us now imagine a second hypothetical situation (‘hypothesis II’). According to this, people’s capacities are different, which seems closer to what we experience in reality. While people continue to make similar judgements (‘murder is wrong’; ‘intense cold is painful’; ‘Beethoven wrote serious music’; ‘we should respect human life’) their judgements differ in significant ways, even after the probing, pointing out, and educating. They differ, for example, on whether euthanasia is murder, or on matters concerning the difference between Beethoven’s earlier and later quartets. We assume that people have in various areas fundamentally different capacities for judgement; some are more intelligent than others, others

88

Stephen Guest

know more, some have wider experience and some are more sympathetic. So, while people in general agree that in certain circumstances murder is wrong and human life should be respected, some will make the judgement that in particular circumstances to carry out an abortion is wrong and others will make the opposite judgement. Many people take the abortion debate to be exactly like this; however much you try to persuade, cajole, argue and so on, people will remain fundamentally convinced in their different judgements. But the second hypothetical situation brings in something from the first hypothesis, which is that people have similarities that can be intersubjectively corrected. The second hypothetical situation brings out something we know, that people can be brought to see something by a person who has a superior (or, let us say, different) capacity. Or, rather, there is an argument that derives from basic similarity that brings with it the knowledge that some people can see things better. It is not a question merely of doctrinal education but closer to the idea of understanding that others can have better skills. Some people have better vision; some people can do sums faster than others can. In the case of making value judgements, people understand that there are some who are capable of making better judgements than others; sport, music and business are obvious examples. But these differences between people are understood, and often admired, for the link that they have to those who lack them. We can see ourselves, as we see others, in the possession of such skills; true, we do not have them, but we can understand, in some obvious senses, what it would be to have them. We can add a third hypothetical situation. Imagine that in this third situation, there are people whose judgements could never match. It is not even clear that these people are even capable of judgement. That thought experiment is entirely fanciful, however, because it builds into the situation the requirement that these characteristics are the characteristics of people. Yet how would we identify them as people? It is impossible to suppose that some people could routinely make judgements that were the exact opposite of ours (‘murder is desirable, even required’; ‘pain is pleasant’; ‘food is unnecessary’) on every point, and that we could still identify them as people. The closest we get in the idea is to radically different species, ants, say; even then, this assumes the mysterious assumption that ants make judgements, and any plausibility the imagined hypothesis has arises from supposing ants are like people because they are alive, and have heads and legs. This third hypothesis does not make sense and it is only included here because this possibility together with the other two exhausts the field.¹⁴ We can drop it. The philosophically important hypothetical situation is therefore ‘hypothesis II’. The first allows correction, and thus objectivity in value, but this second hypothesis allows us to consider the possibility of fundamentally incorrectable differences. In a situation where people’s capacities are different, are they ¹⁴ See for discussion about the projection involved in making judgements of humanity: Williams, Making Sense of Humanity (CUP, 1995).

Objectivity and Value

89

incorrectable? Moral experience shows that our judgements can change; that people can persuade, convince and bring us to see something as the right state of affairs, or the right way to act. Or we simply learn from the practices of other cultures, or from historical example.¹⁵ The ingredients for correction lie in the appreciation of this sort of situation.¹⁶ The correctability of value, which so many philosophers find problematic, does not arise from the scientific domain; that domain is not dominant for value. Instead, we should view correctability in the domain of value more appropriately as improvement of a skill. Some musicians are not only better than other musicians, but can bring others to see how and why it is better to do things the way they do. It is a sense of correction that demonstration by example, the development of talents, and the extension of the capacities of the lesser skilled best brings out. That sense of correction is obviously present in almost any situation of teacher and pupil. It would be silly to say there is no correct way to the play the violin; but equally silly to suppose that the skill could be taught merely by conventional instructions such as ‘no vibrato in this passage’, ‘it goes at this speed’, ‘understate, don’t overstate’. Such statements are accompaniments to demonstration. Even then, the pupil might not ‘get it’ and is taught, ‘See how I do it with the bottom half of my bow’ or ‘That is the right sound.’ This idea of correctability is not that of ‘hypothesis I’ because it does not assume that the pupil has the capacity of the teacher; on the other hand, correctability in ‘hypothesis II’ is compatible with the development and creation of capacity: the pupil can become more skilful. If we suppose that moral reasoning is more like the exercising of a skill we get nearer to appreciating the status of moral correctability, and thus to the question of whether it makes sense to say that moral reasoning is objective.¹⁷ There are obvious differences between the violinist’s skill and being moral. Nevertheless, the similarity is sufficient to suggest that there is a sense of correctability that meets the ‘hypothesis II’ case where there is neither unanimity in judgment nor radical difference. Consider an analogy between music teaching and moral reasoning. It seems wrong to say of two great musicians that, because they play a great violin concerto differently, one at least stands in need of correction. On ¹⁵ For a wonderful account of how we might learn from other cultures—here between people in France and Brazil in the 16th century—see David Wiggins’ discussion of Montaigne’s ‘On the Cannibals’ (The Complete Essays, trans and ed Screech, Allen Lane, 1991) in ‘The Objectivity of Ethics’ in Ethics: 12 Lectures on the Philosophy of Morality (Penguin Books, 2006) 325, 337–53. ¹⁶ If something is right from my point of view only because it is my point of view, then there is no sense in which a change in my point of view is a change in the rightness of my point of view. The earlier point of view was right because the criterion of rightness, under the subjective account of morality, is just that it is my point of view. My view changes, but we can say nothing about its rightness or wrongness. Since we are here not taking account of how we speak about these matters, agreeing that the matter runs deeper than that, it is not sufficient to argue that as a matter of ordinary discourse people do not talk like that. ¹⁷ Hypothesis III might be more accurate than hypotheses I and II in particular contingent circumstances. But this is not a problem for the possibility of truth that some people are mad or mentally weak and unable ever to grasp truths.

90

Stephen Guest

the other hand, some philosophers would say that the analogy works precisely to show that moral reasoning is not objective, and this is because of the ‘great musician’ point. Can we say that because Isaac Stern played the Beethoven violin concerto in the right way,¹⁸ all the other great violinists must have got it wrong? Even then, objectivity for the ‘hypothesis II’ group of points of view could still be possible in many, if not most or all cases. It is a matter of lowering our sights and considering those improvable skills to the point where it no longer makes much sense, as in the case of the great musicians, to think things could be better.¹⁹ It certainly makes a lot of sense to say that the way you played unaccompanied Bach just then, albeit perfectly in tune and with no significant deviation from time signature and expression markings was correctable. Any skill can be improved somewhere along the line and this is all that is necessary to establish the possibility of ‘hypothesis II’. Of the interpretations of the great musicians, though, we can say something special. There is a sense in which Menuhin’s and Grumiaux’s playing of Bach are as equally good ‘in their own way’, and we can leave it at that. We cannot say they are ‘correct’ as each other; in some sense, we can say they are not correctable at all. This is what we should mean when say that their interpretations, the great interpretations, are true for them. It is perfectly plausible to say that great moralists give equally great accounts of what it is to do right; the margin of importance is beyond correction. Perhaps Gandhi’s insights about pacifism and humanity were as perceptive and true as were Churchill’s warlike perceptions about Hitler’s intentions. We can say the same sort of thing about evil. It is probably not necessary to ask whether Stalin was more evil than Hitler; the two of them were evil in their own ways. Nothing is gained by saying one was more evil than the other. The musicianship analogy establishes a sense in which moral reasoning can be taught and corrected, and so is ‘objective’. This sort of correctability is at first sight different from the way in which a person’s errors of visual perception—the ‘paradigm’ point of view—can be corrected by pointing to the existence of some ‘object’ to the satisfaction of all points of view. It is interesting that language allows for the extension from this paradigm to cases such as musicianship and art generally. The fine musician shows ‘perception’ in what the music requires. Of course, there is nothing to ‘see’, but the analogy is powerful because it means that, similar to both visual and musical perception, ‘the way things are’ is significantly ¹⁸ People express it as ‘the right way’ (‘the goalkeeper judged it correctly’; ‘Brendel got it right’); the word ‘correct’ is appropriate, too. Those who are bothered have in mind ‘correctness’ modelled on arithmetic. ‘Criterial’ is a term that has been used by legal philosophers to refer to that form of correctness: it is not just doing something correctly that is at stake but that the standards of correctness are of a particular type, namely, clear and certain and, as a result, productive of an uncontroversial result to those who accept those standards. Arithmetic, but not mathematics, is ‘criterial’ in this sense. ¹⁹ See later remarks on Lord Bingham and objectivity. Are the highest appellate courts an analogous place where the different statements of the law are not correctable, the way Menuhin’s, Grumiaux’s, and Perlman’s different interpretations of the Bach Chaconne are not correctable? It seems unlikely, not the least because of the way lawyers, and the courts themselves, unphilosophically speak.

Objectivity and Value

91

dependent on a person’s point of view. Perhaps it is the wide-felt dependency on science that makes people suppose that the musician’s perception is a non-central case of a point of view. The idea of ‘the point of view’ is centrally something like ‘seeing the stag in the glen’ or ‘hearing the gurgling stream’ and that seeing something musically is a kind of secondary metaphor, not so helpful. But think of the perception of matters of value as central, so that the primary viewpoint is expressed in terms of what we bring to the world, say, through the way we play a violin concerto. Use that idea as a paradigm and go back to the ‘visual world’ experience where reference to an external object ‘demonstrates’ the correct view. Given that the sense of demonstration is present in the artistic case, where one musician demonstrates to us how the violin should be played, or a musical work performed, you can get to a point where you feel doubtful why you should even suppose that the stag is ‘external’. That is a feature of placing one domain more centrally than another. There are certainly elements of ‘bringing us to see’ in the case where someone fails to locate the stag within his field of view. In the same way, we can improve a person’s skill in seeing; we can teach him to spot things that were not significant to him before.

IV. Social Practices and Objectivity Common to the domains of science and value is the idea that a person makes judgements from ‘a point of view’. The external object limits the scope of that judgement. That external objects exist in some sense is, relative to the scientific domain, uncontroversial. It is also hardly controversial that there are no external objects, reference to which is determinative of questions of morality. For questions of value, the existence of external objects determinative of a correct point of view would undermine the importance of that point of view. Even though it is important that individual points of view be correctable, a person cannot be wrong merely because of the existence of something external to him; that would milk him of what is special about him. There are two ways to illustrate this. First, take a society in which people generally accept that it is morally permissible to keep people as slaves. It does not follow that because a practice of slavery exists in this society, slavery is right or wrong. To make the argument clearer, it would not follow from the fact that everyone accepted that murder was wrong that murder was wrong. Many believe that something different is required to explain what makes murder wrong. Murder is wrong because it causes pain, it restricts freedom, it demonstrates a lack of humanity, and it frustrates a life being lived and so on. That conclusion would justify something that is not contained in the external fact.²⁰ ²⁰ A justification for treating an external object in a particular way, say, for example, by supposing that certain biological facts about a person’s genes makes that person morally innocent in

92

Stephen Guest

It is not a particularly insightful point of logic that statements about external objects carry no moral meaning about what people ought to do; it is a matter merely of a missing premise.²¹ As a moral point, it is very important indeed. If people thought certain external objects determined the morality of our actions, that would be a retrograde step; they would lose sight of the necessity for evaluation and the rich critical power that moral discussion allows. It is the road to no morality at all. One particularly bad consequence would be political manipulation. People can create practices. What is well known as ‘political correctness’ leads precisely to that loss of richness of moral thinking. We should not equate morality merely with what a group of people, however good their intentions are, believe it to be. There may be occasions when an initial decision is made that a particular group, or a particular person, has particularly good evaluative judgement and that, to save time, to exercise trust, what those people or that person says, should be regarded as authoritative. The decision to follow a lead in this way is itself an evaluative decision and so open to re-evaluation. If it were not so open, what counts as authoritative would only amount to parroting what a group of people, or a particular person has said. We accept what the doctor says, or what the Bank of England monetary policy panel recommends, because we have good grounds for supposing that what they will say makes good sense. What other reason could there possibly be for concluding that their views hold authority in our lives?²² The second argument is relatively trivial but it needs to be stated. It is odd to suppose that there are external objects corresponding to moral statements in a way that would allow the same sort of correction of moral views as of visual perception.²³ We need not dwell on this point. Moral scientists do not themselves believe in such external objects for morality; and those who are not moral scientists obviously do not think that science constitutes the appropriate domain of knowledge for value judgements. All believe, in their daily and non-philosophically led lives that there are morally correct and incorrect ways to behave. Note the distinction we can draw between external objects and social practices. To suppose that certain external objects corresponded to moral facts or simply were moral facts does seem ‘moronic’.²⁴ It is more difficult to understand the relationship between moral facts and those human rule-characterized practices such as slavery that are in some sense external—they certainly have empirically some situation, is not a scientific justification of that person’s innocence. See Cohen, ‘Facts and Principles’ (2003) 31 Philosophy & Public Aff airs 211. ²¹ It is difficult to think Hume meant anything other than this when he said that ‘oughts’ cannot be ‘derived’ from ‘is’s’. See generally Pigden, ‘Logic and the Autonomy of Ethics’ 67 Australasian Journal of Philosophy 137. ²² See Raz, ‘Authority, Law and Morality’ in Raz, Ethics in the Public Domain (OUP, 1994) essay 9; also see The Authority of Law (OUP, 1979) ch 3. ²³ Mackie, Ethics (Penguin, 1977) ch 2. ²⁴ See Dworkin, ‘Objectivity and Truth’ (1996) 25 Phil & Pub Aff airs. See also, Himma, ‘Just Cause You’re Smarter than Me Doesn’t Give You a Right to Tell Me What to Do: Legitimate Authority and the Normal Justification Thesis’ (2007) 27 OJLS 121.

Objectivity and Value

93

discernible, characteristic features. There are significant differences between such practices and external objects. For instance, external objects are in many cases visible. Not much more than ‘the stag of particular size antlers, and particular red colour in the particular place’ is required for correction of a mistaken view of a certain object. In the case of human practices, the attitudes and beliefs of people engaged in the practice need to be sufficiently gathered and these are things that we cannot observe within the scientific domain. For example, that human beings dominate others and that those others act in a way that primarily benefits the dominant human beings is close to but does not establish a slavery relationship. How the relationship is seen must be part of it but that is not enough either. An account of the meaning of ‘dominate’ and ‘benefit’ is not, at least clearly, reducible to empirical statements. ‘Dominate’ might require, according to some accounts, the idea of what is improper, as with some military usurpations of democratic rule and, to other accounts, proper, as with some benign aristocratic rule. We can say the same for chess pieces, say, as opposed to stags. ‘Chess’ requires a nonempirical explanation in terms of what is justified in chess playing. The empirical—science-domain—aspect of human practices is therefore problematic. Hart attempted to define a social practice of rule-following in purely empirical terms. A social rule existed, he said, when there is regularity of conduct, when deviations receive ‘hostile’ reactions from some members of the group, and those members ‘accept’ a standard of conduct that ‘justifies’ those hostile reactions. It seems commonsense to say we can observe regularities in human conduct; human beings regularly eat and drink, for example. What marks the difference between mere regularity and rule-governed conduct is the presence of a standard against which we assess regular behaviour. Although Hart thought that acceptance was ‘reflective’ and expressed an ‘attitude’, and in that sense was ‘internal’, he was adamant that the question whether people accepted a standard was an empirically observable fact.²⁵ As many have pointed out, the question of what amounts to an acceptance could call up answers of a non-external kind. Acceptance on moral grounds would be one, or economic or self-interested grounds could be others. However, Hart was clear. If we understand human practices not through an analysis of how people generally speak but as proposals, then matters are different. For then, his theory of rule-following explains what people do when they offer a particular justification for assessing conduct: they are assessing the conduct by reference to an empirically what-is-accepted external fact. We could regard the identification of human practices as a matter within the scientific domain as a proposal concerning the practical advantages of seeing it that way. It is appealing to think of some practices as law in the form of an austere self-denying ordinance, one that eliminates value judgements from the identification of the practice. Hart did precisely this in The Concept of Law. In chapter 9, ²⁵ On Hart’s views about the role of acceptance in rule following, see M Moore, Educating Oneself in Public: Critical Essays in Jurisprudence (OUP Collected Papers, 2000).

94

Stephen Guest

he claimed that it is not a matter of ‘linguistic propriety’ (meaning it is not a matter of looking to a dictionary), that we choose between different ‘conceptions’ of law, but that it is a matter of considering the ‘practical merits’ of adopting either (that is, it is a proposal). He chose a wide conception of law—we should identify law as existing within the scientific domain and identifiable solely by empirical means—over a narrower conception that requires judgements of moral value for identification.²⁶ Leaving the practical merits of his theory of rule-following to one side, Hart was guilty of a confusion in his account of rules in his characterization of the ‘internal’ point of view, that is, the acceptance by some members of a group of the standard against which conduct is to be judged. His mistake was to explain the idea by reference to an idea of detached empirical acceptance. We can take this as a proposal for his theory of law (in which case the proposal stands or falls on its merits) but it cannot be a sensible proposal for rule-following in general. Let us suppose that all cases of rule-following are like Hart’s account of rule-following for law. It would mean that there would be no useful sense in which we could ‘get into the mind’ of those following any practice and that would remove any questioning about the point or role of the rule. It would be a disastrous proposal for education. Imagine teaching mathematics, or rugby, or Latin, or philosophy just by describing to people ‘what is regularly done’. Acceptance as ‘the empirical fact that people accept’ is not a useful way of expressing acceptance. To accept something, even without reason, marks out an attitude; that attitude can only be the attitude of a person, with a point of view. It shows not just your willingness to think and behave in a particular way but also shows an endorsement. Accepting a rule is not the same as ‘accepting that’ one has, say, a particular disposition; it is, rather, an announcement about one’s future conduct from which one is free to depart. Freedom to act is part of rule-following. People who follow rules must also be capable of failing to follow them. Otherwise there would be no practical use in coming to the conclusion that they were following them.²⁷ Good practical sense requires that a person who must conform to a standard should understand the reasons for that standard. Under Hart’s account, however, rule-following can only be explained in terms of detachment from such reasons. This must be wrong. So the idea of an empirical identification—an ‘external’ identification—of ‘the acceptance’ of a standard leaves something out of account. Merely understanding that others accept a rule—concentration camp guards, for example—is not the ²⁶ These arguments are clear in ch 9 of The Concept of Law, particularly 209–10. ²⁷ Freedom is implicit in ‘habitual behaviour’, too. As a matter of habit, most people sleep at night but it is habitual only because people have the freedom to sleep at other times. Compare this with just sleeping in general. It is odd to think people ‘habitually’ sleep; rather, they have to do these things. Habit merges into addiction where there is no longer freedom to deviate. And yet we are free to become addicted (the exception being the child born to an addicted mother) which is why none of us is addicted to water.

Objectivity and Value

95

same as accepting that rule. Hart wants to allow for this and calls this detached sort of understanding the ‘moderate’ external point of view.²⁸ Actually, this view is no different in substance from what his theory claims for the ‘internal’ point of view. Although Hart was right to see that there was something ‘internal’ and therefore important about rule-following, he was wrong to locate the identification of that ‘internal’ attitude in the scientific domain for the usual case. The issue is important because we frequently need to obtain a significant sense of the purpose that various practices have in people’s lives so that we can live harmoniously with them.²⁹ Just asking people what purpose the practice has for them does not advance us because we have then to evaluate the reports that we receive; we only have our own point of view. Even if we took a ‘representative’ report of the points of view of others, we would still have to evaluate that report. Moreover, in that sort of case, there is an evaluation required for the separate and important question of what counts as ‘representative’. In Hart’s theory of law, in which ‘official’ acceptance of certain rules determines legal validity, the question of who constitutes an official raises the same problem. One way of tackling this point might be to concede that evaluation is necessary in accounting for the nonempirically determinable purpose or significance in human practice, and then provide an empirical account of those purposes reported by those who are clearly engaged in the practice. We, as theorists, do not then say what we think the purpose of the practice is, but defer to what those engaged in the practice say about their practice.³⁰ We should reject this proposal. First, and obviously, there is the possibility or even likelihood of divergent accounts. We would have to evaluate what counted as one engaged participant’s view about what the purpose of the practice was over another. A detached account lacks the intellectual means for adjudicating between them. A less obvious but stronger reason is that we have to make an evaluation in order to choose as significant the views of the engaged participant. What scientific domain reason compels us to choose the particular participant? In choosing a particular person to be the determiner of what is significant, we make detachment significant, the way Hart chose for law. That might be a good proposal but then, of course, that proposal is still open for evaluative assessment. Sometimes it is the best way, sometimes it is not. Personal diaries, autobiographies, contemporary biography, reports, even accounts books, give a good account of historical events. But these could not—obviously—always amount to good accounts of history. In some cases, yes. Pepys’ diaries, or Anne Frank’s, or Captain Cook’s maritime logs are examples. We need to assess the significance of diaries and we base our selection of them, and whatever other historical materials, on ²⁸ The Concept of Law, ch 6. ²⁹ See MacCormick, Legal Rights and Legal Theory (OUP, 1978) 275 (app) for an excellent example illustrating this point. What could the Lilliputians make of Gulliver’s watch without having any clue that it was for telling the time? ³⁰ Th is seems to be Dickson’s approach in Evaluation and Legal Theory (OUP, 2001).

96

Stephen Guest

their significance; some, perhaps many, reports will have no significance at all.³¹ This is just another way of saying that decisions about what is significant are evaluative and not in the detached scientific domain. One reason for the difficulty in providing an external account of human practices is, as Hart himself perceived, assessing the significance of rules. What practical purposes are served by the idea of rules and rule following? Rule-type reasons fundamentally concern actions that are identifiable by their general class, and the general purpose of the idea of rule-following is to provide a rational constraint on arbitrary action, that sort of action justified on a whim. That constraint is internal, however, not external. To suppose that it determines your behaviour independently of what, overall, you should do suggests a spatial difference between you and the rule ‘out there’. Here the ugly head of ‘moral science’ rears up. Hart’s analysis of rules led him astray on this point, partly perhaps because he was concerned with the existence of what he called ‘social’ rules, and they are so clearly part of the idea of law. There is no good reason for restricting an account of rules in this way. Many of the appeals people make to rule-following do not refer to actual practices. It follows that there is no useful metaphorical sense that one can ‘draw down’ a rule that governs conduct from an externally situated position. Another way of illuminating this point is to say that what is special about rules is their patterned requirement of generality and uniformity rather than their ‘externality’. Take the following four moral judgements to be true and then propose a rule that would justify a judgement on the set of facts at the end: (1) It was wrong that Caleb maliciously dissolved paraquat in his wife’s lime-juice and thereby killed her. (2) It was not wrong that Mrs Krebs intentionally strangled her husband in selfdefence when he was about to crush her. (3) It was wrong that Mr Korba encouraged Steerforth to drop a marble tile from a crane into a procession, injuring a journalist. (4) It was not wrong that Dr Crum, who is insane, kicked a hen to death when he did not know what he was doing, let alone that what he was doing was wrong. Mr Klutz sees that Pok Choi is out of his mind on medication and tells him, for a joke, that he is about to be seduced by Nagel, who is standing nearby. Pok Choi throws a carburettor at Nagel. Nagel, thinking he is being attacked, trips Pok Choi up, killing him. What is the significance of saying that a decision is justified if it is ‘in accordance with a rule’? Could a person merely look at this particular set of facts and make ³¹ Imagine that we thought that a proper historical account of anything would include all blogs.

Objectivity and Value

97

an overall judgement on it, without having to refer to (1)–(4)? It is not clear, however, what that claim would amount to. We would expect a person who said that Klutz had done nothing wrong, full stop, to give a reason, or at the very least, to acknowledge that there were reasons which he could not articulate. This last is not a difficult idea; we can do all sorts of things—make judgements when we drive, we cook, run, catch a ball, do a mathematical calculation, where we are aware that explanations of how we do these things exist but we cannot say what they are. We could ask a person who made such a judgement whether, if the same facts arose in a day’s time, they would also disclose something wrong. It would be surprising if he said his judgement would be different. We could also ask this person their views on (1)–(4) and it would also be surprising if he did not have views. If he thought that (1)–(4) all disclosed moral wrongs, then we could put it to him that he was guilty of an inconsistency: how, if he thinks that foreseeing harm which it is in a person’s power to prevent, and then causing that harm, is morally wrong, could he think that Klutz had done nothing wrong? Reflection on intuitively-made judgements can bring us to alter our judgements; we can adjust them to fit logical patterns. The pull towards patterning is powerful and it is an attack on our moral judgement to show that our judgements do not form a pattern. Nevertheless, to say that we are required to judge in a consistent manner is often overstated. We think of the requirement as one of rationality, of logic. But the sort of consistency demanded here is not one of logic—quite. The two statements ‘It is morally wrong to harm another carelessly’ and ‘it is not morally wrong to harm another carelessly’ are contradictory; to borrow from the logical vocabulary of all possible worlds, the statements cannot both be true nor can they both be false. Both these statements state moral facts. They also imply rules to the effect of what people ought and ought not (or must and must not) do.³² It is appealing to see rule-following as a result of ‘creating a pattern’ to test consistency in argument; it allows us to say that a person has failed to create a sufficient reason in logic for their judgement that a particular action was immoral. That person can revise their judgement in any of the five situations, or they can contest the charge of inconsistency.³³ But consistency cannot be the only test, because it is obvious that we can behave in a way that is consistently wrong. Often just what is meant by a person being ‘internally’ inconsistent is that the reasons that they provide,

³² Normative logic is a different sort of thing since people are often subject to conflicting rules. Morality might require someone both to break the law (in an emergency, say?) and require them to comply with that law. But if the concepts (say of ‘conflict’ and ‘contradiction’) were determined by their practical purpose we might not need to assume that and assume something else, such that normative conflicts are always resolvable through, perhaps, categories such as ‘emergency’, ‘martial law’, and so on. Hart noted the differences in normative logic, about which a great deal has been written without reference to its point, by saying that although conflicting norms do not ‘contradict’ each other they ‘contradict’ in the sense that they imply the ‘logical impossibility of joint conformity’. See Hart, ‘Kelsen Visited’ in Essays in Jurisprudence and Philosophy (OUP, 1983). ³³ They thereby engage with us in Rawls’ idea of ‘reflective equilibrium’.

98

Stephen Guest

while presenting a consistent pattern, do not in some other sense ‘add up’; they fail to meet some other test, the precise nature of which needs pinning down.

V. Objectivity in Law The argument for objectivity in law can only be stronger than that for moral judgement. A case for proposing a conception of correctable moral judgement is superior to one that is not correctable; also, we learn that the source of correction is something that lies within ourselves. Whether we follow Bentham and Hart, who said that law and morality were strictly demarcated disciplines, or whether we think that legal judgements are only a type of moral judgement, it is common ground that when lawyers and judges discuss controversial cases they are arguing matters of morality. So the arguments about moral objectivity apply to legal arguments. However, there is an additional element in the case of controversial legal argument; that practical sense requires that some person or institution must make a decision. And of course, since the argument is moral, it must remain sensitive to substantive reasons. If it were only that a decision must be reached, judges would not be necessary. Nor would lawyers be necessary, because the need for a decision could in most cases be determined by the court clerk simply throwing dice with the attendant advantages of speed and economic efficiency. Rather, we expect judges to reason their way to a decision; the point of the final vote is that where there are differences after argument, and these differences cannot be reconciled, the vote is necessary. No one supposes that a court case, before which the arguments for the litigants are presented, would consist only of a voting mechanism. Therefore, there is a different factor in play in relation to legal argument that is the institutional requirement for a decision. Given the previously discussed argument from stipulation, there are good practical reasons supporting a conception of judicial decision-making in which judges should consider the reasons advanced by litigants, provide counter-reasons where necessary, produce their own reasons, and take all this business seriously, not just as a callous piece of verbal word play before tossing a coin. These two reasons, the existence of a social mechanism, plus a good reason for accepting an argumentative-type conception for it, support the view that the arguments for correctability—for objectivity— add to the arguments for the objectivity of moral judgements. Is there a right answer to the question as to whether the New Zealand All Blacks play better rugby than the English rugby team? We cannot settle the question as a matter of empirical fact, like a scientific matter. People certainly talk as though there were a right answer to the question; at times, it seems that many people can speak of nothing else. Even ignoring the phenomenology of people’s attitudes, there is an added factor, that there is a carefully designed social arrangement by which we seek answers to that question: the international rugby test matches. Of course, many people will say that, in spite of the points superiority of one

Objectivity and Value

99

team over another, they were largely undeserved. Nevertheless, the mechanism is in place. For the final arbiters of those scores, for which there are finely-tuned rules, namely, the referees and rugby judges, it seems odd to suppose that their decisions were merely ‘preferences’. Their decisions are frequently challenged, impugned, contradicted and criticized on the basis that they were incorrect. The sorts of criticisms seem derived from what people would expect from the nature of the institution of refereeing and judging. These criticisms accordingly seem to be something over and above general value judgements as to which team played better rugby, for it is usual for commentators to say that although one team won the Test match series, it was one on ‘points’ alone, not good rugby. What about a banal case? A prize is to be awarded for the child under 12 wearing the ‘nicest’ beachwear. Three judges will vote. We could say that the three judges need only vote their own personal taste and that this is not a matter of ‘judging’ at all. What if the rules require each entrant to provide reasons why their particular beachwear is the ‘nicest’? The judges have to consider those reasons. What if an entrant wears something that looks good but is unsuitable for the beach? What if another wears something rude? Even in this banal case, the logic of the judging forces through a particular way of thinking, one in which competition between reasons seems a natural way of describing it. If two of the judges outvoted the third on the ground that they were both related to the winner (one was her mother, the other her aunt) it seems difficult to say that there was judging at all. What is the best conception of judging that is at stake? Again, we can stipulate that it should be the fairest. And so, in the banal case, because it appears that an entrant would have a reasonable cause for complaint that the judges had treated her unfairly if they merely voted, the logic of judging in the case requires that the judges act on the assumption there is a right answer. Someone will object, of course, that ‘you cannot get a right answer out of nothing.’ There are two answers to this. If you think this is a reasonable objection, then consider whether you are just fi xated by a—perhaps pseudo—account of what amounts to correctability in science. Such a right answer is clearly not available here. If you think it should be available, then you need to show why making value judgements should be treated as part of the same body of knowledge as scientific knowledge. But you also need to show why you think that the judges in the banal case should not make an effort to distinguish the ‘niceness’ of one competitor’s beachwear over another’s. To the banal case we can add another, one that some might want to call the ‘mad’ case. One way we often draw a distinction between matters of pure subjectivity and other matters is by using the ideas of ‘taste’ and ‘judgement’. My liking white bread, vinegar and black pepper with oysters is not to everyone’s taste. We think that my liking for it is for me only, and that no one can say that there is a right answer to that question. That I have that taste we might say is ‘naff ’, but I cannot be mistaken; we must accept that I ‘like’ it. Whereas if I make a judgement about a finer point of rugby, that judgement invites a counter-argument; someone can say my judgement is wrong and I too can admit, if convinced by what he

100

Stephen Guest

says about the nature of scrums, say, that I was wrong. That does not appear to work with matters of taste. I can learn better tastes, but I cannot say I had made mistakes in the past about what tasted good; rather, we take it that my taste has changed. Let us stipulate that ‘taste’ means ‘non-correctable’ because it serves a purpose that such circumstances be non-correctable, and that ‘judgement’ means ‘correctable’ (again, as above, in the sense of correcting a skill). That stipulation will make some suppose this really is a ‘mad’ case. We now hold a competition, in which people are invited to say which of six different vegetables tastes ‘the best’. There is an entry fee; and the prize is substantial and so it would not be correct to say that nothing hangs on the outcome. Should we say this competition is only a sham? The judges have a duty to decide, with reasons, which entrant is right, which entrant has correctly identified which vegetable tastes best. What it is tempting to say has happened, of course, is that something that many people thought was entirely a matter of taste has become, by the requirement of a decision—the practical outcome of the judging—a matter of objective judgement.

Lord Bingham on Objectivity Some people will find the ‘mad’ case just that but will be struck by how our conception of judging changes once we focus on the practical outcome. There are numerous people who take the taste of vegetables to be of great importance; for many, a livelihood depends on it. Vegetables aside, however, what about legal judgements? Apart from the arguments for the desirable correctability of moral judgements, it seems there is an additional argument for correctability of legal judgments arising from the logic of the judicial process alone. Not all judges agree. In a recent interview,³⁴ Lord Bingham said ‘. . . when you are deciding a case you usually feel that there is a choice of answers, and one doesn’t usually regard the answer that one ends up rejecting as being completely and utterly hopeless or untenable; you just regard it as an answer that you don’t prefer.’ Later on he said, ‘To say that there is one right answer and one wrong answer is just not at all how it feels.’ Admitting there was ‘disagreement’ in the House of Lords (but not as much as in the US Supreme Court, or the High Court of Australia, he conceded), he pointed out how there are a ‘significant minority’ of cases in which one or two of the five members will dissent and says: ‘I don’t think the other three would say, “Well, they’re just wrong”. They’d say, “We don’t agree with them”.’ This ³⁴ See Susskind (ed), The Susskind Interviews: Legal Experts in Changing Times (Sweet & Maxwell, 2005). It is an interesting collection of interviews of members of the legal (and British) establishment. It is well done by a genial and intelligent interviewer. A mild criticism is that the book could have been more open in its selection of interviewees. These are an establishment bunch: Lord Woolf, Lord Irvine, Cherie Booth QC, Lord Falconer, David Pannick QC, and others, including both Lord Bingham and John Gardner, of Oxford. Why not Michael Mansfield QC, Professor Neil MacCormick of Edinburgh, Professor Sir Ian Kennedy QC, Lord Justice Sedley, and many others—as those of the sort I suggest are not radical by any means—who, I think, would have provided quite a different but equally as intellectual account of things.

Objectivity and Value

101

is illuminating. Most would agree with Lord Bingham that it would be inappropriate, at least in the vast majority of cases, to say of other judges in an appellate decision that they are ‘just wrong’; indeed, that would be wrong in any academic seminar, or committee of any intellectual standing. To use Lord Bingham’s term, the ‘rarefied’ nature of the argument—where controversial issues have been very finely honed—would make the existence of an obvious mistake on either side pretty well impossible. This suggests that his understanding of what ‘the right answer’ means is too narrow. Something can be right or wrong without there having to be a thundering ‘knock-down’ argument; what is necessary is the possibility of being able to indicate, with reasons, where you think that other person may be wrong. Does it matter? It does. People make mistakes about what they should morally do all the time; who would deny that? It is a mistaken moral view in many people’s eyes, for example, to suppose that the government should seek to prevent stem-cell research. We can advance reasons for that view and we can countermand those reasons with reasons. We can form a view about the balance of reasons. If we all thought that matters of morality were matters merely of ‘preference’ and that there was no ‘right or wrong’ of the matter, and everyone thought the same, there would be no argument about the matter and so, whether stem-cell research went ahead would be dependent on . . . well, what? According to ‘what the majority want’? Would that be morally right? No, because under the hypothesis that there is no ‘right or wrong’ of the matter ‘what the majority wants’ is neither right nor wrong. What would happen is that people would become morally lobotomized: there would be no felt need for argument. This would be a bad state of affairs indeed. On the top of these arguments is the extra argument for the existence of right answers in court. Think of the litigant who has a grievance that someone has infringed his legal rights and has, accordingly, taken an appeal on a decision against him to the House of Lords. He has spent a large amount of money in the hope of obtaining vindication of his rights. Now it must be reasonable for him to assume from the logic of the judicial institution—the opposing arguments presented by counsel both sides, the presence of the judges who will make ‘judgement’ on the question of whose view of the law is right—that there really is a right answer to the question of whether he did have that right. What would this litigant, at least, make of Lord Bingham’s view? What, he might ask, have I paid all that money for? It is instructive to consider results from the search engine on Westlaw of an internal search on the word ‘correct’ in the most recently reported House of Lords decisions where Lord Bingham was one of the judges. In the last case at the time of writing this paper, Jones v Whalley,³⁵ he says the following (emphasis added): ‘Mr Swift submitted . . . that the forms . . . stated the true legal position accurately ³⁵ [2006] 3 WLR 179.

102

Stephen Guest

and the form in the present case did not. But Mr Jones was not to be deprived of his rights by a misstatement in a police form. If the police chose to caution Mr Whalley and not to instigate a prosecution, that was their decision and could not deny to Mr Jones the exercise of a right he enjoyed at law. This is in my opinion correct.’³⁶ In the House of Lords decision immediately before Jones, Crehan v Inntrepreneur Pub Co Lord Bingham says: ‘As already explained, I do not, with respect, accept the Court of Appeal’s approach as correct.’³⁷ And in the case before that, Horton v Sadler, he says:³⁸ ‘Counsel for the appellant submitted that the effect of s.11 is to provide the defendant with a time limit defence in any proceedings brought after the expiry of the three-year period . . . Thus the question for the court under s.33 is always whether it is equitable or inequitable as between the parties to override the time bar . . . This analysis is, as I think, plainly correct.’³⁹ It is therefore clear that Lord Bingham when engaged in judging uses the idea of correct and incorrect judgements of law. Why could it possibly be the case, then, that he thinks that there is no right answer? Is there a distinction being drawn here between the ‘correct’ answer and the ‘right’ answer? Of course not: ‘2+2=4’ is correct, right, true and therefore ‘the best answer’, and ‘2+2=3’ is incorrect, wrong, and false. It is interesting to compare Lord Hoffmann’s forthrightness in this same case: ‘My Lords, it is with a reluctance verging on disbelief that one is driven to conclude that the deliberate opinions of Lord Wilberforce and Lord Diplock were quite wrong. . . . for the reasons given by my noble and learned friend Lord Bingham of Cornhill, I think that they misconstrued what is now s.33 of the Limitation Act 1980.’⁴⁰ (So Lord Hoffmann certainly thinks that the judgements of others can be ‘completely and utterly hopeless or untenable’). Finally, take Lord Scott’s statement in another very recent House of Lords case:⁴¹ ‘It was common ground that the school had not complied with the statutory requirements of a s.64 exclusion. But the assumption that the exclusion of the boys from the school was an exclusion to which s.64 applied was, in my opinion, mistaken.’ Perhaps taking sample phrases from judgements like this is unsystematic. On the other hand, the criterion of selection is appealingly straightforward, for the examples are the latest four cases reported in the House of Lords. The comments Lord Bingham made in the radio programme and reported affirmatively by him later contradict what he and some of his fellow judges in fact say. It may be that, for him, the idea of correctness is licensed by the insertion as above of the ‘in my opinion’ before the use of ‘correct’, ‘misconstrued’, and ‘mistaken’. How licensed? ³⁶ at 186, emphasis added. ³⁷ [2006] 3 WLR 148 at 154 emphasis added. ³⁸ 2006 WL 1546650. ³⁹ [20] emphasis added. ⁴⁰ [22] emphasis added. ⁴¹ A v Head Teacher and Governors of Lord Grey School [2006] 2 WLR 690, 713 emphasis added.

Objectivity and Value

103

Scientists will venture opinions and at the same time suppose them to be not only correct, but the sorts of question that are capable of being tested and found incorrect. That they are their ‘opinion’ makes no difference to that fact. One would wonder, too, why, if they believed them not to be correct, why they would venture them. Further, if something’s being an opinion makes no difference to the ‘hard truth’ of a scientist, why should it make a difference in the case of ‘soft truth’? More pertinently, because it has realistic appeal, lawyers make a great deal of money out of writing ‘opinions’ which, they believe, state the law correctly. Indeed, their skill in stating what the law is must be one major measure of their worth as a lawyer. That idea—of stating what the law is—must entail that there are correct as well as incorrect ways of stating the law.

6 Towards an Inferential Semantics in Jurisprudence Christoph Kletzer

I. Introduction Examining the ‘place of truth’ in jurisprudence has recently received some momentum by the growing interest in the discussion of the pragmatic significance and logical possibility of conceptual analysis and the meaning of epistemological naturalism.¹ This discussion originated form the belated reception of Quine’s work in jurisprudential circles and the growing insight that the latter’s criticism of the possibility of a synthetic/analytic distinction² might be relevant for jurisprudence in general and for attempts at conceptual analysis in specific. The relevance, of course, was intended to be of a destructive nature: if (i) legal positivism of a Hartian tint depends on the possibility of conceptual analysis, if (ii) conceptual analysis in turn depends on the possibility of distinguishing analytic statements from synthetic statements, and, finally, if (iii) this distinction turned out to be untenable, then legal positivism would be struck not on the usual but remote field of disputes about the relevance of morality but it would be struck in its very homeland, it would be shaken to its core. Given legal positivism’s emphatic self-restraint in relation to what questions it claims competence for, the fact that the retreat circumscribed by this self-discipline is under attack must seem devastating: without the analytic/synthetic distinction legal positivism loses the possibility to distinguish between properties that are necessarily predicated of a legal system and those that are predicated only contingently. Without this distinction between truths by virtue of meaning and truths by virtue of fact the whole Hartian enterprise seems pointless. To save positivism naturalism thus demands ¹ B Leiter, ‘Naturalism and Naturalized Jurisprudence’ in B Bix (ed), Analyzing Law. New Essays in Legal Theory (Oxford, 1998) 79. See also A Rosenberg, Philosophy of Science: A Contemporary Introduction (London, 2000) 153. ² We call analytic judgements those in which the predicate is already ‘contained’ in the subject, such as ‘All rings are round’ or ‘All bachelors are unmarried’, whereas a synthetic judgement is one in which the predicate is not already contained in the subject, but adds something new to the subject, such as: ‘This man is a bachelor’ or ‘All humans live on the Planet Earth.’

Towards an Inferential Semantics in Jurisprudence. Christoph Kletzer. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Towards an Inferential Semantics in Jurisprudence

105

that conceptual analysis is either replaced by empirical (or doctrinal) theories, or at least methodologically aligns itself with empirical (or doctrinal) theories. As important as those recent developments surely are, there is nevertheless a certain half-heartedness about them. It is without any doubt laudable that jurisprudence at last tries to relate itself with the history of epistemology—but this attempt remains belated and sporadic. It seems that jurisprudence still does not want to surrender to philosophy and for whatever reasons it digests philosophy only in homeopathic dosages. If, however, we take a closer look at Quine’s intervention and its position in the history of analytic philosophy it should become apparent that his arguments represent but one of many powerful discharges of aporetic tensions that have build up within the epistemological tradition of logical empiricism. Starting from the perception that jurisprudence still predominantly but mostly unconsciously operates under the paradigm of logical empiricism this paper attempts to argue along the following lines: (i) logical empiricism is a doctrine that stands under fundamental tensions, namely between its logicistic and its empiristic tenets; (ii) Quine tried to solve this incompatibility by redefining the ambit of logicism and by saving empiricism in its redefined form of naturalism; (iii) Wilfrid Sellars, on the other hand, resolved the incoherence of logical empiricism by attacking its core thesis of empiricism as such; (iv) insight into Quine’s arguments necessitate a move from legal positivism to legal naturalism; (v) insight into Sellars’ arguments necessitate a move from legal positivism to a non-empiricist theory of law of which so far we have only two (probably flawed) examples: Hegel’s theory of Objective Spirit and Kelsen’s Pure Theory of Law.

II. The Tensions within Logical Empiricism From early on logical empiricism was exposed to intellectual tensions, mainly by the first logical empiricists’ unwillingness to engage with the classical paradoxes of knowledge. Those tensions were built into the programme of logical empiricism from the very beginning, mainly by the first logical empiricists’ playing possum in relation to the classical paradoxes of knowledge. Put crudely, the trouble with knowledge is, that for knowledge to be knowledge it has to combine two elements that are mutually incompatible. Throughout the history of philosophy those pairs of mutually insoluble elements have received different labels: knowledge has to be necessary yet arbitrary, concrete yet universal, synthetic yet a priori, it has to be of the world and at the same time somehow intellectual, of the mind. Now, the trouble with all this is that as soon as we try to reconstruct the possibility of knowledge the two elements seem mutually insoluble, and trying to think their combination is as futile an attempt as trying to make a vinaigrette by simply shaking up oil and vinegar; but at the same time, in existing knowledge those elements are miraculously emulsified.

106

Christoph Kletzer

This simultaneous straightforwardness and incomprehensibility of knowledge has always puzzled philosophers. The logical empiricists, however, were less willing to be puzzled. Not only were synthetic a priori judgements deemed impossible and nonsensical they were also held to be unnecessary for the explanation of knowledge. Knowledge was supposed to be a straightforward, uncomplicated affair of inducing theories from observational propositions or protocol sentences. The verve of the Vienna Circle in its first years stems from this optimism which is but the obverse of the very denial of the existence of any paradox of knowledge. But the curse of the troublesome lies on epistemology and it has been rightly observed that the history of logical empiricism is the history of the successive selfdeconstruction of the fundamentals of logical empiricism.³ The very title ‘logical empiricism’ gives testament to this struggle. It is surprising that only very few authors have recognized that the unity of empiricity and logicality is as troublesome as the unity of syntheticity and the a priori. But whereas Kant started his philosophical efforts with an explicit exposition of this dilemma, logical empiricism delegated this unity of opposites to such a remote place (its own name) that it itself lost sight of it. Still, at some point it had to come to light that the programme of logical empiricism can only be saved if we either accept metaphysical presuppositions (ie accept the logicality of the empirical, the logic structure of the world) or fall back to a conventionalism of knowledge (ie accept the empiricity of logic, the sociological foundations rationality).⁴ Now, logical empiricism and its heir analytic philosophy have sheered between the hubris of denouncing and at the same time accepting hidden metaphysical presuppositions, on the one hand, and the self-hatred of reducing the highest claims of knowledge to arbitrary conventions, on the other: they have been torn between those poles until Quine with his criticism of the synthetic/analytic distinction discharged a bundle of tensions that threw epistemology around. In this criticism he attacked the logicistic, rationalistic side of logical empiricism. But just as the tension had two poles, so the attack on it had. It is usually Wilfrid Sellars who is given the honour of having attacked the second pole of the tension.⁵ He attacked what he called the myth of the given.⁶

³ KW Zeidler, Prolegomena zur Wissenschaftstheorie (Würzburg, 2000) 64 and D Wandschneider, Formale Sprache und Erfahrung. Carnap als Modellfall (Stuttgart, 1975). ⁴ ibid 72. ⁵ W Sellars, Empiricism and the Philosophy of Mind (Cambridge, MA, 1997). ⁶ ‘Whereas Quine’s “Two Dogmas” had helped to destroy the rationalist form of foundationalism by attacking the distinction between analytic and synthetic truths, [Sellars’] “Empiricism and the Philosophy of Mind” helped destroy the empiricist form of foundationalism by attacking the distinction between what is “given to the mind” and what is “added by the mind”.’ R Rorty, ‘Introduction’ in W Sellars, Empiricism and the Philosophy of Mind (Cambridge, MA, 1997) 1. One of the most concise formulations of the myth of the given can be found in J McDowell, Mind and Word (Cambridge, MA, 1994) 7: ‘The idea of the given is the idea that the space of reasons, the space of justifications or warrants, extends more widely than the conceptual sphere.’, ie that we can warrant our knowledge by reference to certain non-conceptual, ie given, truths or beliefs.

Towards an Inferential Semantics in Jurisprudence

107

In a sense Quine himself also attacked the myth of the given in challenging what he called the second of the two dogmas of empiricism,⁷ the criterion of meaning. This criticism, however, was not nearly as fundamental as Sellars’, because it stayed within the ambit of empiricism. To be sure, the problem of the criterion of meaning was on the table as early as the Schlick/Carnap/Neurath protocol-sentence controversy started and with Carnap the principal figure of empiricism departed from it. In his work The Logical Structure of the World he had still tried to free the empiristic doctrine of the concept [Begriff slehre] from its status of a merely programmatic declaration and to transform it into a theory that was logically grounded. Carnap’s insight into the empirical impossibility to define dispositional predicates, on the one hand, and the prevailing belief in the existence of purely theoretical concepts, which are only loosely and indirectly related to the realm of the observable, on the other, are the two main manifestations of the failure of this great project. The discussion about the nature of theoretical concepts, that has been going on for decades now, already takes place beyond the boundaries that empiricism has originally set up against metaphysics.⁸

It turned out that one could attack the logicism of logical empiricism without seriously putting the empiricist project into danger. After all, naturalism is but one reformulation of empiricism. Sellars, conversely, dealt a destructive blow on empiricism as such. So even though his major work is called Empiricism and the Philosophy of Mind, he therein nevertheless ‘dismantles empiricism’,⁹ and moves towards an epistemological position that is closer to a Leibnitzian rationalism than to a Humean empiricism. It is important to note at this stage that this dismantling of empiricism was not an outright rebuff of the empiricists’ paradigms, but rather the result of a successive deepening of the fundamental questions of empiricism. Whereas empiricism asked How can we distinguish between meaningful and meaningless statements by reference to experience? Sellars took the question further and showed that within the empiricist project a crucial question had been left unanswered: What precisely is it in or about experience that allows us to rely on it?

⁷ WV Quine, ‘Two Dogmas of Empiricism’ in From a Logical Point of View (2nd edn, Cambridge, MA, 1980) 20. Stegmüller identified four dogmas of empiricism, the first of which are the Quinean ones: (i) the dogma of the logical possibility of an analytic-synthetic distinction, (ii) the criterion of meaning, ie the dogma that non-logical theoretical terms of a theory can be traced back to fundamental concepts that only relate to observation, (iii) the dogma that logic suffices as a tool to explicate all fundamental, epistemologically relevant concepts, and (iv) the dogma, that the relation of the ‘theoretical language’ and the ‘observation language’ is a linguistic relation. W Stegmüller, ‘Das dritte Dogma des Empirismus. Die pragmatische Wende’ in Probleme und Resultate der Wissenschaftstheorie und Analytischen Philosophie. Band I, Erklärung—Begründung—Kausalität (2nd edn, Berlin, 1983) 1, 2. ⁸ Stegmüller, ‘Das dritte Dogma’ n 7 above, 1. ⁹ RB Brandom, ‘Study Guide by Robert Brandom’ in W Sellars, Empiricism and the Philosophy of Mind (Cambridge, MA, 1997) 119.

108

Christoph Kletzer

III. Three Grand Styles Now, another way to formulate the problem of empiricism is to point out that it locates the place of truth within the concept and takes the concept-object relation to be the archetype for all knowledge. This focus on the concept, however, is but one of three possibilities to lay out the paradigm of knowledge, and in the history of philosophy there have been strong counter-traditions to this approach, locating the place of truth in the judgment or in the syllogism. Those three possible locations of truth, ie in the concept, the judgment and the syllogism, do not stand for any easily identifiably schools within the history of philosophy but outline fundamental styles of epistemological reflection, the three grand styles of epistemology. In the following chapter I will outline these three styles in order to better locate the naturalistic turn within the more recent history of philosophy and to project further turns that are necessary in order to overcome fully the paradoxes of empiricism that still haunt jurisprudence. What is the relation of the concept, the judgement and the syllogism? At first it seems to be obvious that concepts refer to the world, judgements are composed of concepts and inherit the world relation from the concepts they combine. Finally, syllogisms are in turn composed of judgements. It appears only natural to attribute primacy to the concept as fundamental locus of meaning and truth. This understanding is so natural that it appears superfluous even to mention it. Concepts either by themselves refer to things in the world or we refer to things in the world by means of concepts. In either case, concepts have meaning and according to their meaning they can be divided into singular, particular and general terms. The concept dog, for instance, refers to something (the dog out there) and has a certain meaning (as defined by means of its closest genus and specific difference from other species). This unit of meaning, so it seems, is the fundamental building block of truth and the truth of a judgement depends on the right combination of the pre-given meanings of the concepts that make up the judgement. The truth of ‘All circles are round’ thus depends, first, on the content of the concepts ‘circle’ and ‘round’, which have to be assessed independently of the judgement and, second, on their relation expressed only subsequently in the judgement. The validity of a syllogism, in turn, depends on the truth and right combination of the judgements according to certain patterns in truth tables, ie the special distribution that the premises cannot be true without the conclusion also being true.¹⁰ ¹⁰ DJ Velleman, How to Prove It. A Structured Approach (2nd edn, Cambridge, 2006) 17. From the fact, however, that in modern propositional logic the truth of the proposition is discussed in abstraction from the concepts we should not infer that those formalizations do not rest on the paradigm of the concept. Rather conversely, the elimination of the concept from the calculus has saved it from critical scrutiny and strengthened its grip on the paradigms of knowledge modelled on reference and correspondence. In formal logic the relation of concepts to the world is seen to be trivial.

Towards an Inferential Semantics in Jurisprudence

109

In this sense we still speak of the concept of law: there is law out there in the world and we have a concept of it in our heads. Our concept of law thus has two important and distinguishable properties: it relates to the law out there and it has a meaning. Jurisprudence is the attempt to, as clearly and coherently as possible, analyse and explicate this given meaning. Why should then the doctrine of the concept not be a satisfactory candidate for laying the foundations of knowledge? Well, because contrary to appearances the concept is not basic. The concept relates to an object by representing it. However, it can represent the object only by subsuming the object under itself. Subsumption, however, is not conceptual but of the form of a judgement.¹¹ After all, the judgement is nothing but the expression of the relation of subsumption. In the judgement ‘circles are round’, for instance, the concept circle is subsumed under the more general predicate of being round. However, this function of subsumption does not only operate between concepts, it also applies to the relation of a concept to the thing it represents. If I apply a concept correctly to a thing, I subsume the concrete thing under the general concept and the concept fulfills no other function than to contain things under it. If I point at a concrete circle and say ‘circle’ I may have uttered only a concept, but the meaning of my utterance can ony be understood as a judgement: I have applied the concept ‘circle’ to the concrete circle, or, which is the same, I have subsumed the concrete circle under the concept ‘circle’. Accordingly, it is not only the case that syntactically judgements are made up of concepts, but also, more fundamentally, that logically concepts are expressions of judgements: ‘A concept thanks to its general validity has the function of a judgement. It aplies to other concepts potentialiter. The actual relation of a concept to others, as means of their cognition is judgement.¹² Insofar, ‘one of [Kant’s] cardinal innovations is the claim that the fundamental unit of awareness or cognition, the minimum graspable, is the judgement.’¹³ However, according to Dummett the ‘doctrine that the unit of significance is not the word but the sentence . . . is either truistic or non-sensical.’ If this doctrine claimed, Dummett contends, that a word being part of a sentence did not carry with it any meaning independently of the sentence, then the doctrine would simply be absurd; if, however, it merely claimed that ‘we cannot say anything by a sequence of words that stops short of being a sentence . . . then its is truistic.’¹⁴ Dummett misses the point here. The question is not whether (a) the concept which is part of a judgement or (b) the judgement of which the concept is a part is primary, but rather whether it has to be seen that the function of any concept ¹¹ B Longuenesse, Kant and the Capacity to Judge: Sensibility and Discursivity in the Transcendental Analytic of the ‘Critique of Pure Reason’ (Princeton, 1998) 50. ¹² I Kant, ‘Reflection 3045 ( 1776–79)’ in Gesamelte Schriften. Herausgegeben von der Königlich Preußischen Akademie der Wissenschaften. 29 vols (Berlin, 1902–83) XVI, 630. ¹³ RB Brandom, Articulating Reasons. An Introduction to Inferentialism (Cambridge, MA, 2000) 125. ¹⁴ M Dummett, Frege. Philosophy of Language (Cambridge, MA, 1973) 3.

110

Christoph Kletzer

at all can only be described in terms of a judgement. The question is whether the concept (the ‘word’) itself follows the logic of a judgement (the ‘sentence’). And it seems that as long as we focus our attention on the concept alone, we will be incapable of getting closer to a solution of the problem of the logic of object-relation. The doctrine of concept buries this question of object-relation by presupposing it to be unproblematic. The true logic of the relation of concept and object and thus the logic of the concept itself is expressible only in a doctrine of judgement.¹⁵ Kant thereby reversed the commonsensical hierarchy between concept and judgement: it is not the case that judgements are made up of concepts but rather conversely concepts are abbreviations of logical functions that find full expression in judgements. The unity, that is expressed in a concept, is not a natural occurrence, it is not immediately given, but the result of the action of our faculty of judgement. This is to say, that before being ‘given’ an analysable content of a concept, we have to synthesize this content and the action of synthesis is judgement. Equally, the unity of the object ‘out there’ is not simply given, but has to be produced by spontaneous acts of our understanding, it has to be synthesized. This motto of synthesis before analysis or act before word, or judgement before concept, is the core motto of transcendentalism. It is best expressed in the famous first paragraph of the transcendental deduction: ‘We cannot represent anything as conjoined in the object without having previously conjoined it ourselves . . . Analysis, which appears to be its contrary, must, nevertheless, always presuppose synthesis; for where the understanding has not previously conjoined, it cannot dissect or analyse. . . .’¹⁶ So, what do concepts do? They classify. Classification, however, presupposes rules and rules can be expressed only in the form of judgements. So, if a concept is to have a meaning at all, it has to be a judgement. But a judgement (and this justifies the slightly archaic usage of this term as opposed to the modern forms of statement or proposition), is something we do. It is not expressive of a relation of representation (word-object) but it is an act which is warranted or not. That what is done in the act of judging is synthesizing, creating unity. And it is this unity that counts as knowledge. In the concept this unity was thought to be immediately given. In the judgement it turns out that unity is not something immediately given but a mediated result. And as we will see, this unity expressed only abstractly and incompletely in the judgement will be fulfilled in its concreteness in the syllogism. All of those are intrinsically normative questions. Concepts are rules and signify how something ought to be done. The point is that it is not impossible to misapply concepts but that it would be wrong. Insofar concepts are certain commitments ¹⁵ B Longuenesse, Kant and the Capacity to Judge: Sensibility and Discursivity in the Transcendental Analytic of the ‘Critique of Pure Reason’ (Princeton, 1988) 90. ¹⁶ I Kant, Critique of Pure Reason (Cambridge, 1998) 245 (B 130).

Towards an Inferential Semantics in Jurisprudence

111

and their content is an exposition of this commitment. As Brandom expressed it, the conceptual content we started with is not reducible to something like an intentional state, but rather conversely: an intentional state has a normative significance in that it counts as undertaking a certain commitment. The content of the commitment is determined by the rules that are implicit in the concept. Conceptual content thus is nothing but the normative status of the commitments undertaken. What distinguishes the realm of meaning from the realm of facts thus is not kind of hidden substantiality or essence but a distinction between what is subject to normative assessment and what is not. ‘The key to the conceptual is to be found not by investigating a special sort of mental substance that must be manipulated in applying concepts but by investigating the special sort of authority one becomes subject to in applying concepts.’ ¹⁷ At this point we have to make explicit what the reader may have already guessed: the three styles discussed here express not only certain logical but also certain epistemological and ultimately ontological commitments. The ‘truth’ or fidelity of a concept depends on the correspondence with its object. Between concepts and their objects there is a straightforward relation of correspondence or the lack thereof. If I point at a dog and say ‘dog’ I have spoken the truth; if I say ‘car’ I have not. And just as the concept has a corresponding object, the judgement has a corresponding ‘fact’ or ‘something that is the case’. A fact, however, is something fundamentally different from a thing. Whereas the object corresponding to the concept is taken to exist independently of the concept referring to it, the fact corresponding to the judgement is not something that can be pinned down independently of the judgement. For instance, the fact that the door of my room is now open is not a thing in the world, but a kind of order of the things in the world. It is something in between world and mind. Clearly, the fact is not the same as the judgement which refers to it—this synthesis of referent and referee will only be reached in the syllogism—but equally it is not something entirely of the world as the object was. Now that we have seen that the concept has a corresponding thing, the judgement has a corresponding fact, what should the object of a syllogism be? Well, the syllogism expresses a certain relation. And the relation it expresses and the relation it refers to are identical. Insofar each of the three grand styles expresses a different ontological commitment, ie a commitment to what is real: objects, facts, or relations, respectively and accordingly each can be associated with a different epistemological/ ontological tradition: empiricism, transcendentalism (or critical idealism), and inferentialism (or speculative idealism), respectively. Now we can see that in moving from concept via judgement to the syllogism the referents are successively idealized: whereas the object that is the referent of a concept appears to be completely ¹⁷ RB Brandom, Making it Explicit. Reasoning, Representing and Discursive Commitment (Cambridge, MA, 1994) 9.

112

Christoph Kletzer

detached from the concept and to have an independent existence, the fact of the judgement is not something that can be pinned down independently of the judgement. It is already a kind of discursive reality, it is something intellectual. Insofar in the judgement the referent has moved closer to the referee, the gap between knowledge and world has been reduced. The unity of the relation that corresponds to the syllogism, finally, has completely lost the semblance of independence and immediacy. In the syllogism knowledge and its referent are the same and the myth of the given is overcome. This idealization of the referent, or, which is the same, realization of the referee, is the overcoming of the myth of the given. This overcoming is achieved in developing the three styles and not simply by the apodictic denial of the given, not by the simple assertion of a solipsistic self-relation of subjective spirit, but it has to take the path through the transfiguration of the object of the concept by demonstrating that what seems to be independence and immediacy actually already is mediated. All those references to the syllogism are still promissory notes. So let us return to our discussion of the judgement. It seems that even though the doctrine of the judgement has remedied the troubles of the doctrine of the concept in that it directly raises the issue of object-relation, there is still something missing in this account. This is so because ‘classification’ does not fully cover what is going on in a judgement. Apart from classifying something as something the judgement also expresses a further commitment, a commitment that is not expressed in the original judgement and that cannot be expressed in the form of the judgement at all. The problem with understanding judgement as mere classification is that classification is not necessarily a cognitive power, but can also be a natural occurrence.¹⁸ It is, for instance, not only humans who classify things, but animals do as well. A lion classifies certain other animals as prey by his responses to them, to be more precise, by his regular ‘differential responsive dispositions’.¹⁹ But we cannot stop here. It is not only animals that classify their environment by their responsive dispositions, but also plants and even inanimate objects. A chunk of iron classifies its environment as dry or humid by rusting or not, and a spectrophotometer classifies a colour sample as red just as a human observer does. The question clearly is ‘What are the salient differences between a measuring instrument, such as a thermometer or spectrophotometer, and an observer who non-inferentially acquires beliefs or makes claims about environing temperatures and colours?’²⁰ Does the iron chunk judge that the environment is humid? Does a thermometer issue a judgement that it has 31°C or does a spectrometer issue a ¹⁸ In order to explicate this line of thought I will closely follow the arguments Robert Brandom put forward in various texts on this issue. See, for instance, R Brandom, Making it Explicit. Representing and Discursive Committment (Cambridge, MA, 1994) 88. ¹⁹ Brandom, Explicit, n 16 above, 88. ²⁰ ibid.

Towards an Inferential Semantics in Jurisprudence

113

judgement that the sample it was presented with is red? We would not say so. So what are the differences between man and device? The difference obviously is that one ‘understands’ the content red, whereas the other does not.²¹ But what does ‘understanding’ mean in this context? Again, following the lead from Kant and Hegel, Brandom does not try to find an answer in investigations of some psychological faculty, or mental substance. Rather, the key element missing from a measuring device (the difference between a merely responsive and a conceptual classification) is the mastery of the role that those judgements play in inferences—the spectrometer ‘does not treat “that’s red” as incompatible with “that’s green”, nor as following from “that’s scarlet” and entailing “that’s coloured”.’²²

IV. The Doctrine of the Syllogism What makes a classification deserve to be called conceptual and not merely responsive is its inferential role. Put differently: the unity that was presupposed in the concept and laid out in the judgement finds its fulfilment only in the syllogism. The syllogism thus is the truth of the concept. ‘Kant was on the right track when he insisted that just as concepts are essentially (and not accidentally) items which can occur in judgements, so judgements (and therefore, indirectly concepts) are essentially (and not accidentally) items which can occur in reasonings or arguments.’²³ And even Béatrice Longuenesse, the author who in the most explicit fashion highlighted Kant’s role in the development of the doctrine of judgement, does see that also in Kant there are certain strands that point towards the doctrine of the syllogism. For instance, she writes: ‘Every concept is a rule insofar as its explication . . . can function as the major premise in a syllogism. . . .’²⁴ And later: ‘When we subordinate a concept to one that is more general, we attribute the marks pertaining to the concept of greater generality to all the objects contained under the first concept. But thereby every judgement, as concept subordination, is the potential major premise of a syllogism attributing the genus to the species and thereby the genus to all the individuals in the sphere of the species.’²⁵ But still she does not accept her own conclusions and states that ‘Kant insisted that judgment

²¹ It should not surprise us that the early Carnap denied any such a difference between the signal language of a measuring device and the protocol language of a human observer. By that he tried to simplify his theory and avoid many of the classical epistemological problems. That this strategy was not successful, ie that the troubles constantly resurfaced seems now settled. R Carnap, ‘Über Protokollsätze’ (1932/33) 3 Erkenntnis 215, 217. ²² Brandom, Explicit n 16 above, 89. ²³ W Sellars, ‘Inference and Meaning’ in Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars (Reseda, 1980) 262. ²⁴ Longuenesse, Kant n 11 above, 50. ²⁵ ibid 90.

114

Christoph Kletzer

takes precedent over concept formation and inference . . . or, rather, both of these have their common root in our capacity to form judgments.’²⁶ Brandom, on the other hand, expresses a much more favourable attitude towards the doctrine of the syllogism: ‘It remained for Hegel, however, to complete the inversion of the traditional order of semantic explanation by beginning with a concept of experience as inferential activity and discussing the making of judgements and the development of concepts entirely in terms of the roles they play in that inferential activity.’²⁷ I will take up the thread of outlining the Hegelian doctrine of the syllogism in a separate article and now focus on the relevance of the arguments developed so far for jurisprudence. The naturalistic turn could have elevated the debate from the concept to the judgement, but the trouble with it is that it scarcely took place. So at the moment it seems the debate hovers somewhere between the concept and the judgement, even though it might simply be the case that the debate is as concept-based as ever and all that the naturalistic turn has added was a gnawing discomfort with this situation. We have lost our innocent belief in conceptual analysis but as we still cling to the concept paradigm, we cannot possible do without conceptual analysis. All this should not surprise us, since leaving the doctrine of the concept would mean overcoming empiricism and jurisprudence in this respect is even more pertinacious that physics. Given this difficulty of a transition of jurisprudence from concept to judgement, it must at first seem impossible to even imagine a transition from judgement to syllogism. But there are at least three results that such a transition must have: (i) investigations into the concept of the law qua concept are futile; (ii) jurisprudence is a reflexive enterprise and (iii) the positive law can only be understood as objective spirit. (i) The first point seems clear and is currently struggled with under the heading of naturalism in jurisprudence. Whatever the outcome of this discourse it seems clear that a straightforward reliance on conceptual analysis has become impossible. Investigations that simultaneously claim (α) that the law exists out there in the world, independently of our inquiry, (β) that we can find out what belongs to the law by analysing its concept, and (γ) that our abstract conceptual analyses still apply to the law out there have turned to be at best epistemologically naïve and at worst metaphysical. (ii) But apart from this negative result, which is not our topic here, we can also produce a positive one. The point here is that in the case of the positive law the conceptual content understood as normative status gets a reflexive twist because the positive law itself is nothing but a tool of institutionally determining normative ²⁶ ibid 92–3. Her argument for this is that, on the one hand, concepts are formed in order to be combined with other concepts in judgements and also in order to subsume the appearances under it (two functions which presuppose the capacity to form judgements) and, on the other hand, that those two functions make all judgements potential major premises for syllogisms. ²⁷ Brandom, Explicit n 16 above, 92.

Towards an Inferential Semantics in Jurisprudence

115

statuses. So, if we say that the content of the concept of law is not given but is a commitment we take upon ourselves in certain judgements and syllogisms, then the question that immediately arises is: what kind of judgements and syllogism are we talking about? Whose judgements and syllogisms? Well, with the positive law we have the special case that the positive law itself issues such judgements that classifies certain phenomena as law and that the positive law itself issues such syllogism that specify the classificatory commitments undertaken in those judgements. The syllogisms operative in the legal process establish relations of unity in which the content of the concept of law itself is caught up. In pointing out that jurisprudence was the general part of adjudication Dworkin also expressed this fact.²⁸ However, he did so only in a limited sense. In limiting his statement to but one sphere of the law, to adjudication, he impaired the full reflexive relevance of his insight: it is not only judges who, being part of the legal process, meaningfully classify phenomena as being law, but so do executive and legislative organs, and, to say it frankly, everybody. Put briefly: all application of law needs a general part. Accordingly jurisprudence is not the general part of adjudication but of the positive law as such. We should pause a moment to contemplate the severity of this claim. Jurisprudence has an object of which it is a part. Or, put the other way around: the positive law has a part of which it as a whole is the object, ie in which it reflects itself. This reflexive character of the positive law, this mutual interpenetration of law and its philosophy is a major lesson we can take from inferential semantics. Positive law and its philosophic reflection do not stand firmly demarcated from each other but mutually interpenetrate each other. Now, it is from this point that Kelsen’s Pure Theory of Law starts, ie from the insight that the concept of law cannot be determined analytically and independently from the legal process but that its content is given contour by the inferential roles it plays in the legal process, further, that law and its philosophic treatment mutually interpenetrate each other, that both are reflexive objects. The purity of the Pure Theory of Law is but the obverse of the reflexivity of the positive law itself, ie of the fact that it is the positive law philosophically regulates its own creation. It is the positive law itself, and not an external epistemology or moral philosophy, that classifies phenomena as legally relevant or not. Insofar Kelsen’s Pure Theory of Law has little to do with Hart’s legal positivism which still rests on the paradigm of empiricism.²⁹ In this sense this reflexive understanding of the positive law also is a fulfilment of the inner precepts of naturalism: just as natural ²⁸ R Dworkin, Law’s Empire (Oxford, 1998) 90. ²⁹ Apart from Cotterrell no one in the Anglo-American world seems to have taken this to heart: ‘Kelsen’s legal philosophy exemplifies a sophisticated and rigorous elaboration of an avowedly conceptualist approach to normative legal theory. On the other hand, Hart’s “fresh start” in legal philosophy built from his criticism of Austin is best seen [. . .] as an attempt to adopt empiricist methods more satisfactory than those he attributes to Austin.’ R Cotterrell, The Politics of Jurisprudence. A Critical Introduction to Legal Philosophy (London, 1996) 83.

116

Christoph Kletzer

science is not in need of philosophy to be granted validity, so the positive law does not have to wait for legal or moral philosophy to be told what its own concept is. The Kelsenian expression of this naturalism, of this independence form philosophy, be it moral, sociological, or epistemological, is the basic norm. (iii) The Legal Process in its totality, that what Hegel called ‘the State’, is the universal institutionalization of classifications embedded in inferential commitments. In this practice there are implicit rules of how to make explicit the content of certain concepts and judgements. In that the positive law, the state, is similar to science and philosophy: it is objective spirit. It is at the same time an objective entity and self-relating and self-regulating unity, ie it is self-referential transparency not only subjectively ‘in our heads’, but objectively ‘in the world’. Thus, as soon as moral considerations are submitted to this self-regulating unity, they reveal themselves to be subjective spirit. So the anti-natural-law character of the Pure Theory of Law is not an ideological presupposition, or methodological intention, but a result of the reflexive character of the law, of the interpenetration of law and its science. From this understanding of the positive law as objective spirit we can deduce the whole of the Kelsenian Pure Theory of Law: (i) that law regulates its own creation, (ii) that law functions as a scheme of interpretation, (iii) the doctrine of the Fehlerkalkül,³⁰ (iv) the doctrine of impossibility to generate valid legal norms from syllogism (which is obvious given inferential semantics), (v) and the doctrine of the basic norm, which is nothing but a doctrine of the primacy of relation over the related, which is but an abbreviation of the theory of inferential semantics.

V. Conclusion The tradition of what is now called inferentialism is a recent development in theoretical philosophy that surely will have important repercussions on jurisprudence—and be its reception as belated the reception of Quine’s naturalism was. Inferentialism rests on and explicitly refers to an epistemological/ontological tradition, that has somehow been forgotten, at least in the Anglo-American world, a tradition that stretches from some esoteric comments of Plato and Aristotle via Kant to Hegel, Peirce, and Brandom and that is worthy of being investigated in its own right. Now, what concerns the relation of jurisprudence to inferentialism we can identify three surprising and noteworthy points: First, notwithstanding the difficulties and paradoxes that have emerged within the history of empiricism, the dogma of empiricism still holds sway over jurisprudence. Now the problem with this is not so much that jurisprudence somehow should not be allowed to continuously bark up the wrong tree, but rather that ³⁰ C Kletzer, ‘Kelsen’s Development of the Fehlerkalkül-Theory’ in (2005) 18 Ratio Juris 46.

Towards an Inferential Semantics in Jurisprudence

117

as soon as the empiristic dogma gives in, most of the contemporary discussions will be rendered waste paper. As long as the jurisprudential discourse does not dare at least partially to let go of the empiristic dogma, it will remain a pseudodiscourse. Second, this reluctance is all the more surprising given the fact since its earliest days modern inferentialism and pragmatism have amicably leaned towards jurisprudence and have constantly be jilted by the stubborn empiricism of the latter. In his Making it Explicit, for instance, dubbed by Habermas as a ‘landmark in theoretical philosophy’, Brandom not only cites theoretical philosophers and logicians but we can also find references to and whole chapters on Pufendorf and Grotius. Theoretical Philosophy has for quite a while now been building bridges towards jurisprudence. Jurisprudence has not yet reciprocated. Thirdly, everything stated above must appear yet more surprising given what jurisprudence has to offer to a theoretical philosophy which has discovered its own inferential essentials: a centuries-old tradition of dealing with syllogistic forms, of thinking in triads, of engaging with conceptual content as normative status and a rich literature of pondering over the nature of normativity. But we should not be too bemused since by inversion this surprise can easily be turned into a positive programme which asks of jurisprudence to overcome empiricism systematically, to engage more systematically with the currents in theoretical philosophy, and to make more easily available to theoretical philosophy its rich resources of syllogistic thought and inferential expertise.

7 An Epistemic Account of the Internal Point of View Antony Hatzistavrou

In this paper I offer an epistemic account of the internal point of view. I argue that to have the internal point of view towards law is to be epistemically guided by law and especially by the rule of recognition. I define epistemic guidance along the following lines: ‘consideration C epistemically guides A if and only if C informs A about what to do in a certain case and A conforms to C.’ Epistemic guidance is contrasted to motivational guidance: ‘C motivationally guides A if and only if C epistemically guides A and A conforms to C because of C.’¹ In other words, C is the motive which gets A to act. The thesis I argue for is that the internal point of view guides the agent only epistemically and not motivationally. Shapiro and Holton hold the opposite view.² They disagree, however, on what motivates the agent who adopts the internal point of view. For Shapiro to adopt the internal point of view is to be motivationally guided by legal rules qua legal rules. For example, on Shapiro’s account a judge adopts the internal point of view only if she conforms to the legal rules because they are legal rules. By contrast, for Holton to adopt the internal point of view is to be motivationally guided by moral considerations. The judge believes that there are moral reasons for accepting the authority of law and these moral reasons motivate both her criticism of other people’s illegal behaviour and her own conformity to law. Holton credits Hart with an internally inconsistent account of the internal point of view because Hart allows that the person who adopts the internal point of view may conform to law for non-moral reasons. My paper is divided in four parts. In the first I argue contra Shapiro that Hart accepts a variety of motives for one’s conformity which do not include the consideration that a legal rule applies. Further Hart differentiates the internal point ¹ As will become clear in the next section, I follow Shapiro’s account of epistemic and motivational guidance, see S Shapiro, ‘On Hart’s Way Out’ (1998) 4 Legal Theory 469, 489–92. ² Shapiro n 1 above and R Holton, ‘Positivism and the Internal Point of View’ (1998) 17 Law and Philosophy, 597.

An Epistemic Account of the Internal Point of View. Antony Hatzistavrou. © Oxford University Press 2007. Published 2007 by Oxford University Press.

An Epistemic Account of the Internal Point of View

119

of view from the motives for conformity with law. In the second part I argue contra Holton that the inconsistency in Hart’s account of the internal point of view arises only on Holton’s assumption that the internal point of view has motivational clout and that this assumption is ungrounded. In the third part I construct an account of the internal point of view on which it provides only epistemic guidance. In the final part I explain how on my proposed account the internal point of view is distinguished from the external point of view.

I. The Internal Point of View and Motivational Guidance: Shapiro’s Interpretation In his seminal article ‘On Hart’s Way Out’ Shapiro ascribes to Hart the view that ‘people are guided by a social rule whenever they take the internal point of view toward that rule and conform to the rule because they take the internal point of view.’³ On Shapiro’s interpretation for Hart rule-guidance amounts to motivational guidance by rules. Shapiro defines motivational guidance as follows: ‘Someone is motivationally guided by a legal rule when his or her conformity is motivated by the fact that the rule regulates the conduct in question.’⁴ Motivational guidance is contrasted to epistemic guidance. ‘A person is epistemically guided by a legal rule when the person learns of [her] legal obligations from the rule provided by those in authority and conforms to the rule.’⁵ Epistemic guidance does not require that the agent is motivated by the consideration that the rule applies to her case. She may be motivated to conform to a rule out of fear for sanctions or out of some other motive. On Shapiro’s interpretation Hart claims that in order for a developed legal system to exist it suffices that the private citizens are only epistemically guided by the primary rules of obligation. It is necessary, however, that the legal officials are motivationally guided by the secondary rules. That is, it is necessary that the legal officials conform to the secondary rules for no reason other than that they are legal rules.⁶ For example, on Shapiro’s interpretation, a judge who adopts the internal point of view will be motivated to apply a certain primary rule of obligation in examining a legal case simply because of the consideration that this rule is valid according to the rule of recognition. That is, the judge will not only learn from the rule of recognition that the primary rule of obligation is valid but she will be moved to apply it to the case simply because the rule of recognition validates the primary ³ Shapiro n 1 above, 490. ⁴ ibid. Shapiro’s reference to ‘the fact that the rule regulates the conduct in question’ as a motivational factor is a slip. He obviously refers to the relevant belief of the agent that the rule regulates conduct. ⁵ ibid. ⁶ Shapiro n 1 above, 492.

120

Antony Hatzistavrou

rule. By contrast, there is no requirement that a citizen conforms to a primary rule of obligation simply because it is a legal rule. It suffices that the citizen learns from the rule what she is required to do and conforms out of any motive. An elucidation is due concerning the notion of motivational guidance. We may distinguish between considerations which (merely) incline us to act and considerations which in addition get us to act.⁷ The distinction is most clearly drawn in cases in which an agent has more than one reason to perform a certain action and does it for one reason rather than the other. For example, I might have two reasons to open the window in a room: because it is smelly and because it is hot. But what actually gets me to open the window is my desire to disperse with the smell rather than my desire to cool down the room. The latter desire merely inclines me to open the window. For Shapiro the considerations which motivationally guide us are the ones which get us to act as opposed to the ones which merely incline us to act. This is suggested by the way he formulates the distinction between epistemic and motivational guidance. If a desire to avoid being reprimanded gets the judge to conform to the rule of recognition while a consideration that a primary rule passes the validity test of the rule of recognition merely inclines her to do so, she would be like a citizen who is merely epistemically guided by a primary rule of obligation. She would have learned from the rule of recognition what she is required to do but she would have not taken the consideration that the rule of recognition validates the primary rule as the reason why of her conformity.⁸ Shapiro has an intricate theory about what it is for a rule to get one to conform to it. When one accepts a rule one’s present self constrains one’s future self so that the latter does not believe that one has a choice other than applying the rule in cases in which one believes that the rule applies. The constraint is psychological. By accepting a rule one forms a general disposition to apply it in the future. This disposition psychologically constraints the agent and is at the centre of Shapiro’s notion of motivational guidance. So, on Shapiro’s theory, for a judge to accept a secondary rule is not simply for her to use it as premise in order to reach a conclusion about the validity of a primary rule of obligation. On Shapiro’s theory, the judge is in addition psychologically constrained in applying the secondary rule in the sense that she does not believe that she has a choice not to apply it and she is moved by the consideration that the rule applies.⁹ I have argued elsewhere against Shapiro’s account of the psychology of the rule-follower.¹⁰ The main drawback of his account is that it is committed to ⁷ I borrow this distinction from J Dancy, Ethics Without Principles (Oxford, 2004) 21. ⁸ Shapiro says about mere epistemic guidance: ‘As long as its citizens learn what is expected of them and act accordingly, the law need not concern itself with the reasons why they act accordingly.’ Shapiro n 1 above, 492 (my emphasis). ⁹ For a comprehensive exposition of Shapiros’s theory of rule-following see S Shapiro, ‘The Difference that Rules Make’ in B Bix (ed), Analysing Law: New Essays in Legal Theory (Oxford, 1998) 33. ¹⁰ A Hatzistavrou, ‘Instrumental Rules and Motivation’ (2006) 12 Legal Theory 1–31.

An Epistemic Account of the Internal Point of View

121

treating all rules (that is, both categorical and instrumental rules) as categorical. So, it cannot explain the normativity of the legal rules which are on Shapiro’s own account instrumental rules. I will not repeat my arguments here. Instead, I will focus on why Shapiro’s account fails as an interpretation of Hart’s internal point of view. In discussing the relation between legal validity and morality Hart claims: . . . the dichotomy of ‘law based merely on power’ and ‘law which is accepted as morally binding’ is not exhaustive. Not only may vast numbers be coerced by laws which they do not regard as morally binding, but it is not even true that those who accept the system voluntarily, must conceive of themselves as morally bound to do so, though the system will be most stable when they do so. In fact their allegiance to the system may be based on many different considerations: calculations of long-term interest; disinterested interest in others; an unreflecting inherited or traditional attitude; or the mere wish to do as others do. There is indeed no reason why those who accept the authority of the system should not examine their conscience and decide that, morally, they ought not to accept it, yet for a variety of reasons continue to do so.¹¹

Allegiance to the system obviously involves adopting the internal point of view towards its rules and conforming to them.¹² For example, a judge’s allegiance to the rule of recognition consists in her adopting the internal point of view towards it and using it as the ultimate source of validity of the rules she applies in court. Hart seems to claim that the judge in our example need not conform to the rule of recognition, that is, use it as the ultimate source of validity of the rules she applies in court only out of a moral motive, that is, because she believes that she has a moral duty to do so. On the contrary, she might use it as the ultimate source of validity of the rules she applies in court out of considerations of long-term selfinterest, say, because she wants to retain her salary or out of an unreflective wish to do as others do. Hart does not say that if she acts out of the latter motives she would no more adopt the internal point of view towards the rule of recognition. On the contrary, he suggests that whether one adopts the internal point of view towards certain rules is independent of the motivation one has to conform to these rules. This means that the internal point of view is not a motivational attitude. The motivation one might have in conforming to law is irrelevant to the issue of whether one has adopted the internal point of view. Shapiro could retort that when he was talking of motivational guidance he was not referring to motivational attitudes like moral motives, considerations of self-interest, or unreflective conformist wishes. He was referring to an independent type of motivational clout rules have qua rules which is neither identical with nor reduced to any type of these motivational attitudes. So, Shapiro could accept ¹¹ HLA Hart, The Concept of Law (2nd edn, Oxford, 1994) 203. ¹² It is clear from the context that ‘allegiance’ does not mean ‘moral acceptance’. I explore further the relation between the internal point of view and moral acceptance in my discussion of Holton’s interpretation in the following section.

122

Antony Hatzistavrou

that Hart does not treat the internal point of view as a motivational attitude in the sense I described in the previous paragraph. He would point out, however, that there is a type of motivational attitude associated with the motivational clout of rules qua rules, which is different in kind from the motives mentioned in the quotation from The Concept of Law above and which is an integral part of the internal point of view. Shapiro could interpret the quoted passage from The Concept of Law as follows. The motives Hart mentions (moral, self-interested, conformist considerations) are motives for adopting the internal point of view. But the internal point of view includes a motivational attitude. It includes ‘being motivated by a rule qua rule’. So, Hart’s point in the above-quoted passage is that one can adopt the attitude of being motivated by a rule qua rule for a variety of ulterior motives. This way out for Shapiro requires that Hart draws a distinction between two different types of motivation: on the one hand, ‘ordinary’ motives, like moral considerations or considerations of self-interest and, on the other, the special motive of following rules qua rules. However, we have no evidence that Hart draws such a distinction in The Concept of Law. Whenever Hart refers to an agent’s motives he seems to have only ordinary motives in mind. Shapiro would invite us to rely on a distinction which is not used in The Concept of Law. And neither is required to explain Hart’s dissociation of the internal point of view from any motives of the agent. For, as I will try to show, an epistemic account of the internal point of view achieves just that.

II. Holton’s Criticism of Hart—the Internal Point of View as Involving Moral Motivation Holton criticizes Hart for dissociating the internal point of view from moral motivation. He reasons as follows.¹³ Hart takes the internal point of view to involve a critical reflective attitude which is evinced in the agent’s finding criticism for deviation from and demands for conformity to legal rules justified. It is impossible to adopt this critical reflective attitude, ie to find law’s demands justified, without morally accepting the law. So, the internal point of view must be a moral attitude. Further, ‘[t]o say that agents accept the law is partly to make a statement about their motivating reasons: it is to say that they have a particular attitude to the law which explains certain actions.’¹⁴ This is a moral attitude. So, the internal point of view is a source of moral motivation. Hart is wrong in thinking that one can both have the internal point of view and be motivated to conform to law out of non-moral motives.

¹³ Holton n 2 above, 600–6. ¹⁴ Holton n 2 above, 604.

An Epistemic Account of the Internal Point of View

123

I find two flaws with Holton’s argument. The first concerns his premise that the judges cannot find criticism of deviations from and demands for conformity to the legally prescribed official behaviour justified unless they believe that these legal demands are morally justified.¹⁵ Holton inexplicably rules out the possibility that such criticism and demands for conformity may be justified on prudential reasons, like considerations of long-term self-interest. Matthew Kramer has given an example of a totally evil legal regime in which the long-term interest of each judge coincides with the persistence of the regime.¹⁶ Holton provides no arguments to explain why a judge cannot ‘accept’ the laws of such evil regimes and find criticism of deviations from and demands for conformity with the legally prescribed official behaviour justified on the basis that the laws serve the common long-term interests of the judges. Further, there are types of acceptance of rules which are neither moral nor prudential, for example, acceptance of mathematical rules or the rules of football. When one mathematician criticizes another for a mistake in calculation, she need not express any sense of moral disapproval or accuse the latter for failure to follow her self-interest. Similarly, a failure of a referee to observe a rule of football may not be criticized on either moral or prudential grounds. There might be nothing else to it other than the claim that a certain rule (of mathematics or of football) has not been properly followed. Holton does not explain why acceptance of the law cannot be like acceptance of the rules of mathematics or of football. The second flaw concerns his premise that to say that one accepts the law is to make a statement about one’s motivating reasons which explains one’s actions. This premise is crucial to his conclusion that Hart’s account of the internal point of view is internally inconsistent. If the internal point of view is a moral attitude which explains one’s conformity with law, Hart cannot allow that someone who adopts the internal point of view may conform to law out of non-moral motives. But Holton’s premise is ungrounded. There is no reason to suppose that to accept the authority of law is partly to have specific motivating reasons which explain the agent’s obedience. Moreover, there is no reason to suppose this even if we accept Holton’s problematic, as I have argued, thesis that to adopt the internal point of view is to believe that legal demands are morally justified. To provide a motivational explanation of one’s action is to identify the reasons which got one to act. But even if to accept the legal rules is to believe them to be morally justified, the belief that the legal rules are morally justified need not get the agent to obey. This belief may merely incline the agent to obey. It may be a consideration of self-interest which actually gets the agent to obey. For example, a judge may actually believe that the law is morally justified and thus she might ‘accept’ the law in Holton’s moral sense of acceptance but she might conform to it ¹⁵ I focus on the attitude of legal officials since for Hart only the latter must adopt the internal point of view in a developed legal system. ¹⁶ M Kramer, In Defense of Legal Positivism: Law without Trimmings (Oxford, 1999), 92–101.

124

Antony Hatzistavrou

because she wants to keep her salary and not because she finds it morally justified. From the mere fact that I believe that something is morally justified it does not follow that this belief is my dominant motivating reason, ie what will get me to act. So, even if we assume that those who accept the authority of the law believe in its moral legitimacy, we are still left with no argument that the internal point of view provides motivational guidance. I believe that the source of Holton’s confusion is his failure to take on board the distinction between reasons which incline us to act and reasons which get us to act. If those who accept the authority of law believe that they are morally required to obey it, then it is plausible to assume that they have motivating reasons which incline them to obey.¹⁷ But Holton’s claim that their acceptance of the authority of law is a motivational condition which explains their obedience requires that they have motivating reasons which get them to obey. This Holton has failed to show. To recapitulate, Holton has failed to show that the internal point of view is either a moral or a motivational attitude. So, his claim that Hart provides an internally inconsistent account of the internal point of view is not substantiated. I will not try to support further Hart’s claim that the internal point of view is not a moral attitude. In the remainder of this paper I will try to show that he treats it as motivationally inert and as providing only epistemic guidance.

III. Rules, Obligation and the Internal Point of View In The Concept of Law the internal point of view has two functions. First, it explains the difference between observance of social rules (of which legal rules are a sub-class) and habitual obedience. Only if one adopts the internal point of view towards a social rule does one properly follow a rule. Otherwise one simply exhibits habitual behaviour. Second, it is indispensable for a proper understanding of the notion of legal obligation. One sees oneself and others as having legal obligations only if one adopts the internal point of view towards those legal rules which ground legal obligations. These two functions are interrelated. For Hart statements of legal obligation imply the existence of a legal rule. More specifically, for Hart a statement of the type ‘X has a legal obligation to do A’ (a) can only be understood at the background assumption that a certain legal rule R of the type ‘All Ys ought to do A (when in C)’ exists (b) signifying that X is a Y (in C).¹⁸ I will begin with an analysis of how Hart uses the internal point of view to explain the difference between observance of social rules and habitual obedience. Hart’s primary target in the first four chapters of The Concept of Law is Austin’s ¹⁷ cf Holton n 2 above, 622. ¹⁸ Hart, Concept n 10 above, 85.

An Epistemic Account of the Internal Point of View

125

positivist analysis of law as ‘commands backed by threats issued by a sovereign who is habitually obeyed but who habitually obeys none.’ Hart objects, inter alia, to Austin’s analysis that a habit of obedience cannot explain the continuity of law. Assume that a certain population has developed through years a habit of conforming to the commands of a sovereign, Rex I. On Austin’s theory, given the existence of the general habit of obedience, the commands of Rex I are law. But assume further that Rex I dies and his son Rex II starts issuing commands backed by threats. At this point there is no general habit of obedience to the commands of Rex II. Neither is Rex II a sovereign. So, on Austin’s theory the commands of Rex II are not law; the commands of Rex II will become law and Rex II will become sovereign once a general habit of obedience to his commands is established. The obvious drawback of this implication of Austin’s account is that it fails to do justice to the means with which regimes regulate succession of legal and political power and secure the continuity of law. Hart observes that these ends are achieved by the establishment of certain rules of succession which create relevant rights or titles to law-making power. So, one needs to move from the notion of a general habit of obedience to the notion of accepted rules of succession in order to account for the phenomenon of the continuity of law. These rules must be social in the sense that they must be accepted by the majority of the citizens.¹⁹ In order for his criticism of Austin’s theory of law to be valid Hart needs to show that the notion of acceptance of a social rule is not reducible to the notion of a general habit of obedience. To this end Hart introduces in The Concept of Law for the first time the internal point of view. Acceptance of a social rule and habitual conformity are both evinced in behavioural convergence. But the similarities end here. For a group of people to share a habit nothing more than behavioural convergence is required. For a group of people to accept a social rule behavioural convergence should be accompanied by their adopting the internal point of view or more specifically by their understanding the ‘internal aspect’ of the relevant rule.²⁰ The group’s grasp of the internal aspect of rules involves three things. First, the group understands a certain pattern of behaviour as a standard of conduct. Any deviations from this pattern of behaviour are considered to be open to criticism (including self-criticism) and to constitute grounds for the exercise of pressure for conformity with the pattern. Second, the relevant criticism and exercise of pressure are considered to be justified; that is, deviation from a pattern of behaviour which functions as a standard of social behaviour is considered to be a ‘good reason’ for criticism and exercise of pressure. Finally, individual members of the group conceive the relevant pattern of behaviour as a common standard of conduct. It is what the group as a whole follows and not each for her part.²¹ ¹⁹ For Hart’s account of the problem the continuity of law creates for Austin’s theory, see, Concept n 11 above, 53–5. ²⁰ Hart, Concept n 11 above, 56. ²¹ Hart, Concept n 11 above, 57–8.

126

Antony Hatzistavrou

Let me analyse Hart’s initial account of the internal point of view. The criticism Hart refers to involves the use of a certain vocabulary for the evaluation and guidance of action. It is expressed in terms of what specific individuals ‘ought to’ do or of what actions are ‘right’ or ‘wrong’. So, the internal point of view includes mastery of certain evaluative and normative concepts which are evinced in expressions of criticism or demands for pressure addressed to specific individuals. Mastery of these concepts must include knowledge of their interrelations and of the conditions of their application. We may summarize this first account of the internal point of view as follows: (1) The person who has adopted the internal point of view follows a pattern of convergent behaviour. (2) She identifies this pattern of convergent behaviour as a standard of conduct. (3) She understands this pattern of convergent behaviour as what the group as a whole (and not each for her part) follows. (4) She has mastered a range of concepts for the evaluation and guidance of actions in accordance with this standard. (5) She uses these evaluative and normative terms in a variety of statements (of criticism, or of demands for conformity) which are addressed to specific individuals on the basis of their behaviour in relation to this standard. So, this first account of the internal point of view relates it to the mastery of a conceptual apparatus. The person who has the internal point of view possesses two ‘generic’ abilities: the ability to use the concept of a social rule and the ability to use some other evaluative and normative concepts. Hart does not tell us which of the two abilities he considers more basic. Perhaps they are best understood as interrelated. It might be impossible to master the concept of a social rule without relating it to the concepts of the rightness and wrongness of actions and vice versa. The person who adopts the internal point of view is able to construct two types of statements. The first are statements of the content of a certain social rule. For example, ‘one ought to show respect to one’s parents.’ The second are statements of evaluation of actions of specific individuals or of normative requirements addressed to specific individuals. For example, ‘John’s talking back to his father was wrong’ or ‘John ought to show more respect to his father.’ I suggest that when Hart claims that deviations from the behaviour prescribed by the rule are considered ‘good reasons’ for criticism or demands for conformity to it he means that statements of the first type have justificatory priority over statements of the second type. A person who adopts the internal point of view uses the content of a social rule as the justificatory basis of her criticisms or demands for change in behaviour. For example, she claims that you ought to respect your parents because (of the social rule that) everyone ought to respect their parents. She does not claim that everyone should respect their parents because you and

An Epistemic Account of the Internal Point of View

127

every other individual ought to respect their parents. From the internal point of view only the first because-statement introduces a valid justification. The statement of the content of the social rule justifies the statement of the existence of the relevant individual obligation. The second because-statement introduces at worst an invalid justification or at best some kind of evidence about the content of a certain rule. So, in addition to the five features I have mentioned the internal point of view comprises a particular type of deliberation. One reasons from statements of social rules to statements of evaluation of individual action or to statements of normative guidance addressed to individuals. The former are used as the justificatory basis of the latter. This means that the person adopting the internal point of view constructs or is able to construct upon reflection an argument the conclusion of which is a relevant evaluative or normative statement and the major premise a statement of a social rule.²² The deliberation is ‘valid’ if and only if the argument from the statement of a social rule to the relevant normative and evaluative statement is formally valid and the statement of the social rule is true (ie there exists a social rule with this content). We can speak of an evaluation of individual actions or a normative requirement addressed to specific individuals as being ‘valid’ in a similar sense if and only if they are conclusions of a formally valid argument the major premise of which is a statement of a rule and the latter statement is true. This conception of validity of the relevant deliberation or of the evaluative and normative statements is the seventh feature of the internal point of view.²³ Laws are for Hart a species of social rules. The internal point of view towards law shares all seven features of the internal point of view towards social rules. Further, Hart divides laws into two broad categories.²⁴ One who adopts the internal point of view towards law should be able to distinguish between these two types of legal rules. The first comprises general obligations. They are of the type ‘all Xs have an obligation (ought to) do A.’ For example, every driver has an obligation to stop at red traffic lights. The second comprises the conferment of powers to individuals. They normally specify procedures to be followed so that individuals, if they so wish, can define their legal relations with others.²⁵ Such power-conferring rules are rules which determine what counts as marriage, ²² In his initial description of the internal point of view Hart has deductive arguments in mind. For matters of simplicity I will describe the deliberation the internal point of view comprises as a species of deductive reasoning. For an analysis of legal reasoning along Hart’s lines see, N MacCormick, Legal Reasoning and Legal Theory (Oxford, 1978). ²³ This notion of ‘validity’ is not explicitly mentioned in Hart’s discussion of the internal point of view of social rules. Hart speaks only of ‘legal validity’ to characterize the relation between the rule of recognition and a legal rule established by it (Hart, Concept n 11 above, 100–10). The concept of validity of social rules I introduce in the text is the counterpart of the concept of legal validity in societies with primitive legal systems. ²⁴ See the relevant discussion in Hart, Concept n 11 above, 26–33. ²⁵ These are private powers. In developed legal systems there are also rules conferring powers in the public sphere, like legislative and judicial powers, see, Hart, Concept n 11 above, 28–32.

128

Antony Hatzistavrou

contract or will, etc. The distinction between the two types of rules is that only the former ground individual obligations. You have an obligation to stop your car at a red traffic light because there is the relevant rule. But you do not have an obligation to marry because there is a rule defining legally valid marriages. This is not to say nevertheless that the power-conferring rules are not justificatory bases of certain relevant evaluative or normative statements from the internal point of view. One can still criticize someone for failing to follow the right procedure for making a will or advise her about the relevant procedures. So, the deliberation characteristic of the internal point of view may be from statements of social rules to either statements of individual obligations or to statements of conferment of private powers. This initial account of the internal point of view is incomplete. It explains legal rules in a rather primitive type of society which is characterized by the absence of provisions about the identification of rules, the resolution of conflicts between rules and the determination of conditions of changes of rules. In developed legal systems these provisions are present. To account for the complex nature of law in a developed legal system Hart introduces a new distinction between primary rules of obligation and secondary rules. This distinction is different from that between rules of obligation and power-conferring rules.²⁶ On the latter distinction legal rules are differentiated on the basis of the different evaluative and normative statements about individual conduct they justify. On the former distinction they are differentiated on the basis of the type of individual decisions they purport to govern. Secondary rules purport to govern decisions about which laws are valid. Primary rules purport to govern decisions about which individual conduct one should follow under law. The primary rules of obligation determine specific types of behaviour as legally obligatory to citizens. The clearest examples are articles of criminal law; a law against murder designates abstention from intentional manslaughter as legally obligatory for citizens. Secondary rules are of three types. First, there is the rule of recognition determining the conditions under which a particular law (either a primary rule or one of the other two types of secondary rules to be mentioned shortly) is a member of a legal system. For example, ‘whatever the Queen in Parliament enacts is law’ is the rule of recognition in England. Second, there are rules of change determining the conditions of changes of legal rules. Third, there are rules of adjudication determining a range of legal powers to officials which ²⁶ Hart is not clear about the relation of the two distinctions. On the first distinction rules of (general) obligation are contrasted to power-conferring rules. On the second they are contrasted to secondary rules. Do power-conferring rules belong to the secondary rules? If so, the first distinction is subsumed under the second. Some remarks about power-conferring rules Hart makes (Hart, Concept n 11 above, 96) suggest that he takes them to belong to the class of the secondary rules of change. However, this view is problematic. Given that the presence of secondary rules differentiates primitive from developed legal systems, this view implies that in primitive legal systems there are no power-conferring rules. Th is is implausible. At least some power-conferring rules, the private ones (cf n 25), can exist in primitive legal systems, for example, rules determining marriages.

An Epistemic Account of the Internal Point of View

129

allow them to make authoritative determinations on possible breaks of law as well as to direct the application of sanctions for actual breaks of law.²⁷ Hart believes that law in its complete form is a union of primary and secondary rules. And given that the secondary rules are primarily designed to be followed by legal officials, the attitude of legal officials to law is important for the existence of a legal system. Hart maintains that in order for a developed legal system to exist it is necessary that the legal officials adopt the internal point of view towards the secondary legal rules. But it is not necessary that the ordinary citizens understand the secondary rules from the internal point of view. It suffices that only the legal officials do so. Further, according to Hart, it is not necessary that either the citizens or the legal officials in their private capacity adopt the internal point of view towards the primary rules of obligation; it suffices that they customarily conform to the primary rules of obligation.²⁸ All seven features of the internal point of view I have previously identified also characterize the officials’ internal point of view towards the secondary rules. First, the officials conform to the secondary rules. For example, judges apply the laws identified by the rule of recognition in court. Second, they treat them as standards of their official conduct. Third, they treat them as common standards of conduct, that is, what they all as a group recognize and not each for her part. Fourth, they have mastered a specific apparatus of evaluative and normative concepts. For example, they have the concepts of legal rights and obligation or of legally rightful and wrongful conduct. Fifth, they use this conceptual apparatus in statements of evaluation of officials’ conduct or in normative statements addressed to officials concerning their professional conduct. For example, judges of the appellate court may criticize the decisions of lower courts and issue directives about how certain laws should be applied in court. Sixth, they reason from secondary rules to primary rules and from primary rules to judgments about the obligations of specific individuals.²⁹ Finally, they treat judgments about the obligations of specific individuals as valid because they are justified on the basis of primary rules and primary rules as valid because they are justified on the basis of secondary rules and especially the rule of recognition. The following is an example of the last two features of the internal point of view. A judge finds a litigant guilty of an offence of breaking a contract and awards damages. The judge reasons as follows. She justifies her decree on the basis of a primary rule of obligation determining the obligations of parties engaging in a certain contract. For example, she argues that the litigant by exhibiting such and such behaviour did not fulfil her obligations as determined by the relevant primary rule of obligation. Further, she relies on this primary rule because it has ²⁷ Hart, Concept n 11 above, 91–9. ²⁸ Hart, Concept n 11 above, 116–7. ²⁹ They also reason from secondary rules to other general (private or public) power-conferring rules and from the latter to the conferment of powers to individuals. For matters of simplicity I mention only primary rules in the text.

130

Antony Hatzistavrou

been duly enacted by the legislature in accordance with the provisions of the rule of recognition. So, the rule of recognition is the ultimate justificatory basis of her judgment. She conceives the judgment as valid because it is justified on the basis of a valid primary rule which is valid because it is justified on the basis of the existing rule of recognition. The particular type of deliberation the judges engage in from the internal point of view is theoretical. In its complete form it has a statement of a rule of recognition as its major premise and a statement of the existence of citizens’ obligations as its conclusion. The function of the complete form of this deliberation is the identification of what the citizens (in their private capacity) have an obligation to do in a particular legal system in which the judge engaging in this theoretical deliberation operates. Insofar as a judge so deliberates in reaching her decree about citizens’ obligations and actually delivers this decree she possesses the internal point of view. The internal point of view can be understood as partly an internal psychological condition of the legal officials along functionalist lines. Its input are the arguments of a legal case presented to them. The output is a belief about the rules of primary obligation which apply to the legal case. The internal point of view is the mental process from input to output, a theoretical deliberation as described above. But the internal point of view is not just that, an epistemic psychological state of the legal officials. A condition that a legal official should satisfy to have the internal point of view is that she conforms to the secondary rules, that is, her decree is in accordance with her belief about the rules of primary obligation which apply to the legal case. So, the internal point of view corresponds to the concept of epistemic guidance. It picks out the conjunction of an epistemic state and obedience. What motivates the judge actually to adopt the conclusion of her theoretical deliberation as her decree? Hart’s answer is: a variety of motives ranging from a sense of moral obligation to self-interested considerations or a mere wish to do as other judges do. When these motives win over other possible opposing motives and thus get the judge to conform to the conclusion of her theoretical deliberation, then the judge is epistemically guided by the legal rules. What is the relation of these motives to the internal point of view? My identification of the internal point of view with epistemic guidance by the legal rules implies that they are not part of the internal point of view. Epistemic guidance by law presupposes that the agent has some motives which get her to conform to law. But these motives are not part of the concept of epistemic guidance. To give an example: a judge considers a case, applies the rule of recognition and the relevant legal rules in her deliberation, D, and reaches the conclusion, C, that the litigant is guilty of an offence and should be fined. She knows that were she to acquit the litigant she would receive a great sum of money from the litigant’s lawyers as a ‘gift’. However, she believes that it is a safer financial option for her to

An Epistemic Account of the Internal Point of View

131

remain lawful and not risk her career. So, she decides to make C her decree and thus conform to the legal rules. On my proposed interpretation, she (a) exhibited the internal point of view, since she was epistemically guided by the relevant rules, and (b) was motivated by a self-interested consideration (which is not part of the internal point of view). It is of course possible to redefine the concept of the internal point of view so as to include a condition that the agent be motivated by a disjunction of motives to conform to law. Though this would qualify the concept as motivational in some sense, it would clearly not be the sense required by either Shapiro’s or Holton’s readings. On the one hand, it would not make the internal point of view equivalent to Shapiro’s concept of motivational guidance. It would not be law qua law (in my example, the legal rules in D that justified C) which actually got the judge to conform. Rather the judge would be motivated by ‘ordinary’ motives (like the concern for her long-term financial interests in my example). Neither would this redefined account be compatible with Holton’s reading of the internal point of view. For on Holton’s reading only moral considerations can motivate from the internal point of view while on the redefined account non-moral motives (like the concern for long-term financial self-interest) can also motivate. So, I prefer to say that the internal point of view expresses the concept of epistemic guidance and is motivationally inert. I call it ‘motivationally inert’ because it does not motivate the agent to obey in any of the senses required by either Shapiro’s or Holton’s interpretations. The following objection may be raised against my proposed reading of the internal point of view. The judge who adopts the internal point of view must believe that she has a legal obligation to conform to the secondary rules. But is this belief motivationally inert? Why not suppose that the judge is motivated by the consideration that she has this legal obligation? This objection to my claim that the internal point of view is motivationally inert can take two forms. On the one hand, one can claim that there is only one sense of obligation, ie moral obligation, and that normally one does not understand a statement of moral obligation unless one is somehow motivated by it. On the other, one can allow that legal obligation might differ in kind from moral obligation but insist that one fails to understand something as ‘obligatory’ (that is, in either its moral or legal sense) unless one is somehow motivated by it. The point of either of the two forms of the objection is that one cannot understand a statement of legal obligation applied to oneself while failing to be inclined to conform to it. Obviously the point is not that one fails to understand a statement of legal obligation applied to oneself if one fails to conform to it out of a relevant sense of obligation. So, this objection utilizes the notion of an inclinatory reason as opposed to the notion of a reason which gets one to act and can be thus rejected along the lines I rejected Holton’s account in section II. But the objection can be reformulated so that it states that having the internal point of view requires that one obeys out of a sense of (moral or legal) obligation. To have

132

Antony Hatzistavrou

the internal point of view is to be inclined to conform to a statement of legal obligation applied to oneself and actually to conform because of this inclination. On this reading the internal point of view corresponds to the concept of motivational guidance by individual legal obligation (as opposed to Shapiro’s concept of motivational guidance by legal rules). Hart denies that legal obligation is a species of moral obligation. So, he could not have agreed with the first form of the objection and believed that to have the internal point of view is to be motivationally guided by a moral obligation. I suggest that the reason why he denied that legal obligations are moral obligations is that on his view legal obligations do not even provide inclinatory reasons for action. So, he would also reject the second objection on the basis that one can understand a statement of legal obligation without feeling inclined to conform to it. In rejecting the account of legal obligation offered by Austin’s theory which viewed statements of legal obligation as predictions of future sanctions Hart stresses that the notion of legal obligation cannot be understood without reference to the notion of legal rules.³⁰ The predictive theory fails to take into account the fact that deviation from the rule-governed behaviour is a reason for the application of sanctions. Statements of individual legal obligation describe considerations which the judges take into account in reaching their decisions. The statement ‘Y has a legal obligation to do A’ implies the statement ‘there is a legal rule that all X should do A’ and the statement ‘Y is an X.’ So, for example, the judge’s decree that the litigant had an obligation to pay back her mortgage to the bank implies that there is a legal rule that all those who got a valid mortgage from a bank should pay it back and that the litigant belongs to this class of people. As we have seen, for Hart, legal rules are a species of social rules. What makes appropriate the use of ‘obligation’ to describe the normative situation of one to which a legal rule applies is the fact that legal rules are social rules of great importance for which there is strong pressure for their application. There is strong pressure because legal rules are considered to be ‘necessary to the maintenance of social life or some highly prized feature of it.’³¹ Further, the use of ‘obligation’ indicates that the conduct required by the relevant legal rules normally involves sacrifice of the short-term interest of the person having the legal obligation.³² So, a statement of individual legal obligation implies not only the existence of legal rules which apply to the obligation-bearer but also that there is strong pressure for ³⁰ Hart, Concept n 11 above, 82–91. ³¹ Hart, Concept n 11 above, 87. ³² Hart does not make a distinction between short-term and long-term interest but I think it is the former he has in mind here. It is compatible with our having an obligation to do A that doing A is in our long-term interest. This is true for both moral and legal obligations. For example, if being just actually benefits us in the long-run we can still say that we have an obligation to do a just action which requires some sort of sacrifice of our short-term interest. Similarly, though it may be actually in our long-term interest to pay back the mortgage, it is still true that we have a relevant legal obligation which is a financial burden at present.

An Epistemic Account of the Internal Point of View

133

the application of these rules, that they are necessary to the preservation of social life and that the person who has a legal obligation is normally required to sacrifice her short-term interests.³³ I believe that for Hart this is how someone who adopts the internal point of view understands statements of individual legal obligation. Consequently, this is how a judge understands statements of her legal obligations to apply the secondary rules. She believes that there are certain rules of the type ‘all X ought to do A’ which have the social importance associated with legal rules and normally require sacrifice of the agent’s short-term interests and that she is an X. On this account of statements of individual legal obligation there is no requirement that the judge who understands them from the internal point of view is in any way inclined to conform to them. She simply understands them as being justified on the basis of rules which are legal and would normally require sacrifice of her short-term interests. These rules which justify statements of individual legal obligation of the judges are general rules of obligation to conform to the secondary rules of the legal system. They are not the secondary rules themselves. The secondary rules generally confer powers. They do not impose obligations upon judges. So, they cannot justify statements of individual legal obligation of judges to apply the secondary rules. The rules of obligation to apply the secondary rules of the system have an ancillary function. Hart claims that they ‘reinforce’ the secondary rules.³⁴ If my interpretation of Hart’s account of statements of individual legal obligation to apply the secondary rules of a legal system is correct, then their range of application is rather limited. Such statements are valid only when there are relevant valid rules of obligation to apply secondary rules. They cannot be properly invoked to criticize, for example, lapses of the application of secondary rules in the absence of such rules of obligation. So, a judge who fails to apply a certain secondary rule in a case not governed by a relevant rule of obligation to apply that rule cannot be properly criticized for having failed her obligation. One might still criticize her official conduct but without invoking the notion of obligation. To conclude, Hart provides an analysis of statements of individual legal obligation according to which their understanding does not require the agent to be inclined to conform to them. To understand a statement of individual legal obligation is to understand that it is justified on the basis of a (general) rule of obligation, to recognize that rule as legal (ie of the relevant social importance) and to believe that it may require the sacrifice of the agent’s short-term self-interest. Statements of individual obligations of judges to apply the secondary rules are ³³ The social pressure and the importance of the rules for the preservation of social life may be systemic in the sense that concerns the whole body of legal rules and not necessarily each individual rule. ³⁴ Hart, Concept n 11 above, 97. Hart makes this remark by reference to rules of obligation to apply the rules of adjudication but it obviously holds for rules of obligation to apply the other two types of secondary rules.

134

Antony Hatzistavrou

grounded on general rules of obligation of the judges to apply the secondary rules.

IV. The Internal Versus the External Point of View No interpretation of Hart’s internal point of view is complete unless it explains its difference from the external point of view. In this section I will try to show what their difference consists in on my proposed interpretation that the internal point of view provides purely epistemic guidance. Hart defines the external point of view as roughly ‘that of an observer who does not accept the rules of the society [s]he is observing.’³⁵ He distinguishes three types of statements that can be made from the external point of view: (1) An observer may report the convergent behaviour of members of a group as mere habitual behaviour, that is, without referring to the fact that they regard it as a social rule. (2) An observer may report in addition the hostile reactions to deviations from the convergent behaviour as similarly mere habitual behaviour, that is, without referring to the fact that they are treated by members of the group as reasons for the hostile reaction. (3) An observer may report both the fact that the convergent behaviour is treated as a social rule (and thus it is not mere habitual behaviour) and the fact that deviations from it are treated as reasons for hostile behaviour. The ‘epistemic guidance’ account of the internal point of view can easily explain the respects in which the first two types of external statements differ from internal statements. The observer who utters them neither conforms to the convergent behaviour, nor employs the relevant evaluative and normative conceptual apparatus, nor treats statements of the existence of the convergent behaviour as the justificatory basis of corresponding evaluative and normative statements. How the third type of external statements differs from internal statements is a more complex issue. The person who makes such a statement possesses the relevant evaluative and normative conceptual apparatus and uses it in order to conceptualize a pattern of convergent behaviour as a rule. Further, she is able to simulate the deliberations of the participants in the relevant social practice and thus reason from a statement of a social rule³⁶ to relevant evaluative and normative statements. Finally, in some cases the observer making this external statement may also be herself a participant and thus qua participant conform to the rule. For example, a judge who applies the laws of her country in court may make such external statements while explaining the workings of her legal system to a ³⁵ Hart, Concept n 11 above, 291. ³⁶ Though not as I explain from a statement of the content of the rule of recognition.

An Epistemic Account of the Internal Point of View

135

foreign judge. So, the third type external statements do not differ from internal statements in the ways the other two types do. Hart gives an example of the third type external statements. He claims that statements of the type ‘rule, R, is valid’ can be made from both the internal point of view of a participant and the external point of view of someone who does not necessarily accept the rule of recognition but states that others do. He describes the difference between the corresponding internal and external statements as follows: To say that a given rule is valid is to recognize it as passing all the tests provided by the rule of recognition and so as a rule of the system. We can indeed simply say that the statement that a particular rule is valid means that it satisfies all the criteria provided by the rule of recognition. This is incorrect only to the extent that it might obscure the internal character of such statements; for like the cricketers ‘Out’, these statements of validity normally apply to a particular case a rule of recognition accepted by the speaker and others, rather than expressly state that the rule is satisfied.³⁷

What Hart says here is rather obscure. I suggest that Hart intends to distinguish between external and internal statements in terms of their different illocutionary force. For example, in saying at the end of the trial that the defendant is found guilty the judge performs the illocutionary act of delivering a decree. By contrast in saying that the defendant is found guilty when analysing the case to a foreign judge she does not deliver a decree. She performs the illocutionary act of informing the foreign judge that this is the conclusion the legal rules of her country yield when correctly applied to the case in question. Similarly, in saying that R is valid when deliberating in court about a case a judge performs the illocutionary act of allowing the introduction of the rule into further deliberations to reach a decree. By contrast in saying that R is valid when analysing the case to a foreign judge she merely informs the latter that the criteria posited by the rule of recognition of the legal system of her country have been satisfied. The epistemic guidance reading of the internal point of view can explain this difference in the illocutionary force between the internal and the external statement that R is valid. When adopting the internal point of view the judge reasons as follows: (P1) Whatever the Queen enacts in Parliament is law. (P2) R has been enacted by the Queen in Parliament. (C) R is valid. When adopting the external point of view her deliberation takes the following form: (P1*) In England the judges recognize as law whatever the Queen enacts in Parliament. ³⁷ Hart, Concept n 11 above, 103.

136

Antony Hatzistavrou

(P2*) R has been enacted by the Queen in Parliament. (C*) R is valid (in England). P1 states the content of a rule of recognition. The argument from P1 to C ‘applies’ this rule of recognition.³⁸ C states the outcome of this application. By contrast P1* states the fact that a certain country’s rule of recognition has a specific content. P1* is used as a premise to the conclusion that a certain rule satisfies the criteria posited by the rule of recognition of that country. So, the difference in the illocutionary acts of saying ‘R is valid’ between a judge who intends to reach a decree and a judge who intends to explain the workings of a legal system corresponds to the difference in their respective deliberations. When the judge intends to reach a decree she follows the first type of deliberation. When she intends to explain the workings of the legal system of her country she follows the second type of deliberation. So, the difference between the third type external statements distinguished by Hart and the standard internal statements is explained by the epistemic guidance reading of the internal point of view as a difference in the deliberations of the agent. This analysis suggests a final essential feature of the internal point of view. In its complete form the starting point of the deliberation of the agent who adopts the internal point of view is a statement of the content of the rule of recognition. ‘Acceptance’ of the rule of recognition involves treating a statement of its content as the ultimate justificatory basis of statements of individual legal obligations. By contrast when the agent deliberates from a statement of the fact that a certain rule of recognition exists in a certain country, she deliberates from the external point of view. To recapitulate, the epistemic guidance reading of the internal point of view can explain the differences between the different types of the external point of view Hart distinguishes and the internal point of view. The person who adopts either of the two first types of the external point of view fails to recognize patterns of social behaviour as social rules, reason from these rules and conform to these rules. The case of the person who adopts the third type of the external point of view is different. That person may possess all the features of someone who adopts the internal point of view but fails to reason from a statement of the content of the rule of recognition to statements of individual legal obligations. Instead the starting point of her deliberation is a statement of the fact that a certain rule of recognition exists in a certain country.

³⁸ This is how I understand Hart’s claim that internal statements of the type R is valid ‘normally apply to a particular case a rule of recognition’. It is not C on its own which ‘applies’ a rule of recognition but rather the argument from P1 to C.

8 Antigone and the Nature of Law Tanja Staehler

Sophocles’s tragedy Antigone has inspired much philosophical discussion. In this essay, I will argue that this attention is well-deserved, especially in the context of philosophy of law, because the tragedy provides essential insights regarding the nature of law. Moreover, ancient Greek properly philosophical accounts of law, eg those of Plato and Aristotle, fail to disclose the nature of law with the same intensity as Sophocles’s tragedy. Most famous among the philosophical interpretations of Antigone is Hegel’s reading of the tragedy as a conflict between human and divine law. It is indeed plausible to read Antigone in terms of a conflict between different kinds of laws or different aspects of law. However, Hegel’s attempt to offer a solution to this conflict, designated by the concept of ‘reconciliation’ (German Versöhnung), is dissatisfactory. Reconciliation occurs not so much on the level of actual characters in the tragedy as, rather, on the level of Spirit as all-encompassing, historically evolving consciousness. Just as philosophers after Hegel (most prominently Kierkegaard) have often reacted to his concept of Spirit with a renewed emphasis on human existence in its individuality or singularity, there is also an emphasis in Antigone interpretations after Hegel to read the tragedy as an exploration of human nature and human individuals. In these interpretations, the focus lies with the role and the positions of individual characters in the tragedy. Two post-Hegelian interpretations shall be briefly introduced in the second part of this essay: Martin Heidegger’s discussion of the tragedy and specifically the first choral ode in his Introduction to Metaphysics,¹ and Judith Butler’s relatively recent text Antigone’s Claim.² Despite many substantial differences between these two readings, they share the thesis that Antigone is a tragedy revealing what it means to be human. Importantly for the purposes of my essay, they also exhibit the same lack of attention regarding the theme of law in Antigone.

¹ M Heidegger, Introduction to Metaphysics [hereafter: IM] (New Haven, 2000). ² J Butler, Antigone’s Claim [hereafter: AC] (New York, 2000).

Antigone and the Nature of Law. Tanja Staehler. © Oxford University Press 2007. Published 2007 by Oxford University Press.

138

Tanja Staehler

In light of these different interpretations, I shall argue, in the third part of this essay, that Antigone is a tragedy about law and its inherent conflicts. It is certainly not wrong to claim that this tragedy elucidates human nature; yet it does so to the extent that we are beings whose nature it is to be caught in the tension between different kinds of laws. In other words, the nature of the human as it emerges from this tragedy has to include an understanding of the human as a creature entangled in law, or specifically, as both creator of and subject to laws. There is an essential tension at the core of law. On the one hand, it is the nature of law that it has to be conceived as something unchangeable, like the sacred laws which Antigone invokes to justify her actions. On the other hand, laws are either created by humans or at least receive their specific formulations from humans, and in that sense, they are open to criticism and modification. But in order to take the law in its true nature, we need to treat it as fixed, at least to a certain extent. As I intend to show, approaching the law from the outset as open to changes and interpretations means not to treat it as a law.

I. Hegel’s reading of Antigone has been discussed in much detail by commentators from different schools and orientations. Providing even a brief overview or summary of the various positions would lead us too far astray here. Rather, I would like to focus on two aspects of Hegel’s reading, one of which strikes me as an essential insight, indispensable for any interpretation of Antigone, the other as a problematic claim which, although comprehensible within Hegel’s framework of thought, leads to a one-sided interpretation of Antigone. The first aspect concerns the interpretation of Antigone as a conflict of two kinds of laws, the second the Hegelian emphasis on reconciliation. Hegel chooses the term ‘divine law’ for the law which Antigone invokes, and ‘human law’ for the law given and executed by Creon. Antigone herself refers to the ‘sacred laws’³ given by the Gods, which warrants Hegel’s choice of terminology. According to Antigone, the procedures inscribed by these laws regarding the burial of the dead inevitably and by nature overrule the manmade decision, imposed by Creon, not to bury her brother, Polyneices. Yet as Polyneices attacked his own home city, Creon deems it plausible that the sacred laws would be suspended in this case. By focusing on the conflict between human and divine law, Hegel forces us to consider the positions at stake here, not some individual (and in that sense arbitrary) character traits of the main protagonists. While certain statements in the tragedy could tempt us to offer psychologizing⁴ interpretations of Antigone ³ Sophocles, Antigone [hereafter: Ant] (Oxford, 1962) 78. ⁴ By using the term ‘psychologizing,’ I am not referring to ‘psychoanalytic’ interpretations of the tragedy, of which there are many and which are obviously much more complex than the brief

Antigone and the Nature of Law

139

as a rebellious woman (Ant 97) or Creon as a stubborn man who does not want to admit of being defeated by a woman (Ant 483), Hegel keeps us focused on the basic conflict which surpasses personality traits. This is a significant accomplishment of his interpretation and provides insights into the nature of law. In his Phenomenology of Spirit, Hegel refers to and quotes from the Antigone text in the context of his well-known critique of Kantian ethics as too formal, a-historical, and empty of content. The turn to the Greeks promises to offer a model of laws as rich, concrete, and historical; yet it is not specific laws that Hegel wants to invoke, but the general feature, exemplified by Antigone, of laws that simply ‘are, and nothing more.’⁵ Rather than starting from an empty a priori principle like the Kantian categorical imperative, we can learn from the Greeks that laws are always already in existence, and that it is the nature of law to be encountered as something binding and pregiven. This character of simply being there is exemplified by the divine law whereas the human law opposes such mere existence as something which is ‘conscious of what it actually does’ (PhS, para 450). Do the divine or sacred laws necessarily involve a connection to the gods, or is it possible to give a secular explanation, an interpretation that does not invoke the gods at all? For the Greek context, such an explanation would certainly be impossible. However, modern interpreters have tried such readings, claiming that the gods are ‘simply those aspects and conditions of life which we have to accept because we cannot change them.’⁶ This general and perhaps overly simplifying explanation regarding the role of the gods in Sophocles can be specified in respect to the divine law. The divine law represents that aspect of law according to which it appears as something which we as humans have received rather than created, something we are subject to rather than author of. Authorship is, in this case, attributed to the gods. Yet it is possible to imagine certain high ethical principles (such as ‘You shall not kill another human being’ or, even less controversial, ‘You shall not kill another human being arbitrarily’) which might not by themselves refer us to divine authors and which nevertheless strike us as intuitively plausible, holding with a certain necessity rather than contingency. As I will explain in more detail below, every law, even the most contingent one, has to exhibit this aspect of law at least minimally. Accordingly, the human law exerted by Creon also shares in this character to some extent. The Chorus confirms: ‘You, being sovereign, make what laws you will/Both for the dead and those of us who live’ (Ant 213). This statement expresses various facets. First, it is in the nature of the sovereign to prescribe laws which the citizens have to accept. Second, these laws do not even cease to be valid when a citizen dies; this makes them somewhat similar to the divine law. Third, reference to character traits and gender designations given here. However, those psychoanalytic readings fall outside the scope of this essay. ⁵ GWF Hegel, Phenomenology of Spirit [hereafter: PhS] (Oxford, 1977) para 437. ⁶ HDF Kitto, Sophocles: Dramatist & Philosopher (Oxford, 1958) 1.

140

Tanja Staehler

however, it is clear to everybody that the sovereign ‘makes’ the laws, ie that they are man-made. If there is a certain circularity implied in this reasoning, then Creon’s earlier statement emphasizes this strange character of his legislation even more: ‘There is no art that teaches us to know/The temper, mind or spirit of any man/Until he has been proved by government/And lawgiving’ (Ant 175). It is by way of good and successful law-giving that a man proves himself as a good law-giver (and a good man in general). Does this not imply that the citizens would be well advised to question the laws of a given law-giver until he has proven himself worthy? The law given by Creon regarding Polyneices’s burial is the first law he gives after coming into office; does this weaken the respective law? Perhaps; yet according to the monarchical constitution, it is certainly not the task of the regular citizens to question the given laws. If anybody, the Chorus, consisting of Theban elders, might be allowed to comment on Creon’s laws and decisions, and Creon turns to them repeatedly for advice. The Chorus reflects the double nature of law well since it tries to support Creon’s legislation as long as possible, until too many divine and other signs to the contrary occur. When Creon finally decides to yield—much too late, when too much disaster has already overcome his family and the state—, he refers to the Goddess of Necessity whom he cannot battle. In that sense, he himself subordinates the human law to the divine law, without exactly referring to it as such. While the impression might arise that Sophocles regards Antigone as the (moral) victor of the conflict with Creon, it is Hegel’s contribution to present both protagonists and their respective standpoints as equally valid, such that both of them lose (or win) in the end.⁷ There is an essential conflict between the two positions which is ultimately irresolvable. It would be wrong to suspect Hegel of making a case for Creon based on the modernism of Creon’s position, reading Antigone as embodying the ancient perspective. Certainly, Creon’s arguments are closer to the standpoint of morality which Kant distinguishes from mere legality and which, according to Hegel, characterizes the modern approach to ethical and moral issues. From the standpoint of morality, it is the subject’s highest right ‘not to accept anything which I do not realize to be rational.’⁸ However, Hegel refers to the Antigone tragedy exactly because modern morality, epitomized in Kantian ethics, strikes him as one-sided, lacking the ancient emphasis on laws and customs as pregiven. These two standpoints can be read outside their historical manifestations as aspects of law in general, and Hegel also acknowledges this at times.⁹ ⁷ Robert Stern explicates quite clearly how Hegel favours neither Creon nor Antigone, and how this holds for all Hegelian interpretation, in the Phenomenology of Spirit as well as in the Lectures on Fine Arts and the Lectures on Religion. cf R. Stern, Routledge Philosophy Guidebook to Hegel and the Phenomenology of Spirit (London, 2002) ch 5. ⁸ GWF Hegel, Elements of the Philosophy of Right (Cambridge, 1991) § 132. ⁹ For example, Hegel claims that the conflict between state (‘ethical life in its spiritual universality’) and family (‘natural ethical life’) is valid for every epoch. cf GWF Hegel, Aesthetics: Lectures on Fine Arts [hereafter: Aesth] (Oxford, 1975) 1213.

Antigone and the Nature of Law

141

Furthermore, Hegel shows how both Creon and Antigone in a certain sense also support the standpoint they are fighting since their own, one-sided standpoint already contains its opposite (cf Aesth 1217). We have just seen how Creon’s standpoint indeed involves an element of taking the law as unalterable. Antigone, on the other hand, admits at certain points that there is no absolute guarantee of her interpreting the divine law correctly (even though she feels confident about her interpretation), and in that sense, she admits that she does not have unmediated or infallible access to the divine law. However, the fact that both standpoints in principle contain their opposite does not mean that Antigone and Creon acknowledge that the other person might be in the right. The reconciliation which Hegel traces in this tragedy and which, according to him, characterizes every tragedy (cf Aesth 1215), occurs essentially not on the level of the individual characters, but on the level of Spirit.¹⁰ This means that the one-sided standpoints of the individual characters are overcome in a higher standpoint which still contains elements of both, and which is usually represented by the Chorus. According to Hegel’s reading of Antigone, this higher standpoint involves elements of ancient law, rich with pregiven traditions and customs, and the modern standpoint of questioning and testing all norms before accepting them. While it seems indeed plausible to characterize each of the two standpoints in this tragedy as one-sided and as implicating its opposite, the idea of a reconciliatory standpoint has at least one major disadvantage: the conflict between the two kinds or aspects of law comes to appear temporary and inessential. Furthermore, it is dissatisfying that the two main protagonists fall out of the picture, as it were, as both of their perspectives prove untenable. At this point, the interpretations offered by Heidegger and Butler offer helpful supplements.

II. Without entering into the details of Heidegger’s project and the development of his thought, let me just mention that Heidegger’s interpretation of Antigone emerges in part III of his Introduction to Metaphysics, entitled ‘Being and Thinking’.¹¹ Heidegger’s thought is associated with the question of Being which ¹⁰ It is unfortunate and ultimately opposed to his own philosophical convictions that Hegel, in his Phenomenology of Spirit, gives the impression that Antigone herself realizes and acknowledges her error, as if reconciliation had to involve such insights on the level of the individual (cf PhS para 470). To this end, Hegel misquotes Antigone. Hegel quotes her as stating: ‘Because we suffer we acknowledge we have erred.’ Yet Antigone actually does not state that she has erred, but: ‘If this is what the gods approve, why then, When I am dead I shall discern my fault; If theirs the sin, may they endure a doom No worse than mine, so wantonly inflicted!’ (Ant 925). For a detailed discussion of this misquote, cf JG Finlayson, ‘Conflict and Reconciliation in Hegel’s Theory of the Tragic’ (1999) XXXVII Journal of the History of Philosophy 493–521. ¹¹ Heidegger refers to Antigone in other places (for example, in his lectures on Hölderlin), but his interpretation of the first choral ode is most interesting and helpful for our purposes.

142

Tanja Staehler

is intimately linked to the question, what does it mean to be human? As human beings, the question of Being is of immense importance to us; we continuously use terms derived from ‘to be’ without truly knowing what it means. In turn, posing the question of Being evokes the necessity to explore human beings since they are the ones capable of asking questions; moreover, a human being is ‘never an answer, but essentially a question’ (IM 149). The claim that a human being is a question is corroborated, according to Heidegger, in the first choral ode of Antigone. The most important lines in this respect are the beginning ones: ‘Wonders are many, yet of all Things is Man the most wonderful’ (Ant 333). Translated here as ‘wonderful’ is the Greek term ‘deinon’ that has a plethora of meanings, some of which seem to exclude others: wonderful, terrible, monstrous, extraordinary, uncanny, etc. Heidegger translates ‘deinon’ into German as ‘unheimlich,’ as uncanny or strange. Part of his interpretation serves to justify this translation, clarifying what is ‘uncanny’ about human beings. In the choral ode, the unusual nature of man is explained in several steps. Man is exposed to various elements, especially the earth and the sea, which have inexhaustible energies and confront man with their waves, circles, and cycles. Exposed in this fashion, man does not give up; he rises to the challenge, making himself at home in the midst of that which overpowers him. He hunts the animals of earth and sea and builds cities in which he can rest protected. However, there is one challenge which man cannot escape: ‘Death alone’ (Ant 361). In the face of his own mortality, man has to acknowledge his ultimate powerlessness. These topics inevitably strike Heidegger as closely related to his own philosophical project. His main emphasis in interpreting the ode, however, is not the issue of death, but man as the uncanniest creature. This uncanniness concerns the way in which man dwells in the world; it is summarized quite well in the admittedly rather enigmatic sentence: ‘The uncanniest (the human being) is what it is because from the ground up it deals with and conserves the familiar only in order to break out of it and to let what overwhelms it break in’ (IM 174). The conflicting movement which Heidegger describes in this sentence concerns our relation to Being, to the world, to that which surrounds us and becomes manifest as earth and sea. Man strives to make himself at home in the midst of that which overwhelms him, and yet, he himself has a tendency to overstep the boundaries of the familiar. In this respect, human nature is excessive. According to Heidegger, this excessive nature is best characterized by the Greek term ‘techne,’ or by the ‘violence-doing’ that comes about through techne. In this context, violence needs to be distinguished from arbitrary brutality. Heidegger wants to give expression to the way in which humans are not satisfied with becoming at home in a minimal fashion. Instead, humans reach out and form cities, constitute history, create language, and formulate principles for their actions. In the end, it becomes obvious that these ostensible creations are not wholly in our power, but overwhelm us on a higher level. As they augment

Antigone and the Nature of Law

143

and elevate strange powers which overpower them in return, humans are more uncanny than the uncanniness of Being to which they are exposed. This increase of strangeness generates the potential for disaster as we now meet the strangeness of the world on a much higher and sophisticated level; tragedy tells of these disasters. Heidegger himself declares that his commentary on Antigone is necessarily inadequate as it does not consider the entire tragedy, let alone Sophocles’s work as a whole (cf IM 159). It is thus not fair to criticize Heidegger on exactly these points which he so clearly announces. He does not explain why his interpretation remains rather limited in this respect, but it is obvious that his aim, in this text, is not an interpretation of Antigone per se. Rather, the tragedy arises in the context of several other questions like the one about the above-mentioned connection between Being and humans. Furthermore, Heidegger attempts to explain several Presocratic ideas, specifically Parmenides’s claim about the coincidence of Being and thinking. For clarifying this idea, Heidegger can turn to Antigone since early Greek thought is characterized by a much closer union between philosophy and tragedy than Plato, for example, would allow for. My objection to Heidegger’s reading thus does not concern the fact that he focuses on the first choral ode only, but the fact that his interpretation neglects the issue of law. More precisely, it seems to me that his interpretation of this tragedy could easily be extended into a pertinent discussion of law. An opening for such an interpretation arises within the context of the first choral ode, after death as the final assault has been mentioned. The next strophe mentions three kinds of law. First, it refers to the possibility of evil versus brave deeds, ie the moral law. Second, the law of the land is mentioned, ie human or legal law. Third, reference is made to the justice of the gods, ie sacred or divine law. However, Heidegger does not discuss these different kinds of law or the relations between them. Instead, he requests that we think the Greek term dike not as ‘justice’, but as German Fug, ie a fundamental kind of ‘fitting order’. If we understand dike in a moral or juridical sense, we go wrong, Heidegger claims. Why would the moral or juridical implications lead us in the wrong direction? Heidegger states that the ‘fundamental metaphysical content’ would get lost (IM 171). With the help of the distinction between different aspects of law gained above, we might interpret his warning to say that too much emphasis would be placed on law as man-made (or, in Hegel’s terminology, ‘human law’) whereas the unchangeable character of all law would be neglected. But is it not exactly the accomplishment of Antigone that it exhibits both characters of law, showing the irresolvable conflict between them? As Heidegger does comment on the very sentence in which both kinds of law identified by Hegel (and represented by Antigone and Creon respectively) are mentioned, it would be helpful if he commented exactly on the relation between nomos and dike. Furthermore, Heidegger’s interpretation of the first choral ode and of uncanniness would allow investigating the conflict between these two aspects in more

144

Tanja Staehler

depth. Such an interpretation, which Heidegger admittedly rejects, could start exactly from his notion of techne as a knowledge which guides human creations of various kinds. For the Greeks, legislation was itself a techne, and as such, fallible.¹² On the basis of Heidegger’s analysis of techne in his Antigone interpretation, it could be argued that law-giving, like any techne, is a form of violence-doing. By way of giving laws, man tries to arrange himself in and with the world, making himself feel safer and more at home. At the same time, man again oversteps the limits of the familiar. This becomes more obvious when considering the complementary aspect of law, ie law not as created, but as binding and unchangeable. In this sense, law itself has the character of being uncanny and overpowering. Lawfulness as a prearranged order is part of Being itself. The notion of dike or Fug could even be extended in this direction if Heidegger did not from the start exclude any interpretation which would move this concept in the direction of law. Heidegger explains that the uncanny has the dual character of dike and techne, of being overpowered and of doing violence in the midst of the overpowering. This dual movement resembles exactly the two conflicting characters of law: we experience law as an otherness out of our control, like the justice of the gods, and in that sense, it is part of Being’s uncanny character. In light of this overpowering element, we create laws, thereby organizing our life in cities and states. At the very moment when we arrange our (social) life in a more predictable fashion, it turns out that even our man-made laws reacquire the character of being overpowering and uncanny. Man thus raises uncanniness to a higher level, in respect to law just as in respect to Being in general. Summing up this interpretation, let me describe the situation of law as it presents itself both in light of Hegel’s and Heidegger’s insights. Hegel had alerted us to a conflict of law at the core of Antigone, a conflict between human and divine law or between law as created and law as pregiven. Humans emerge from this conflict as both author of and subject to law; yet the conflict between these two roles is not easily resolvable since the precise way in which we are subject to law implies that we cannot just change law at will; otherwise, the character of law would get lost. Taking this conflicting situation without moving on toward Hegelian reconciliation, Heidegger’s insight into man as the uncanniest being comes into play. The conflict within law presents itself as a form of uncanniness, evolving from the way in which man makes himself at home in the world only to overstep the boundaries of this home. Law is one fashion in which humans intend to make the world more controllable, yet ultimately turn it into a more complex scenario, thereby increasing uncanniness. Law-giving is thus one essential aspect of human uncanniness. Since I would like to continue this second part of the paper by briefly discussing Judith Butler’s way of reading Antigone as a new description of what it means to be human, let me start from her critique of Hegel. The aspect of her critique which strikes me as most convincing concerns a failure, on Hegel’s part, ¹² cf eg Plato, Statesman in Plato, Complete Works (Indianapolis, 1997) 294b.

Antigone and the Nature of Law

145

to clarify the relation between the laws that he identifies (cf AC 38 f). While he describes this relation as a conflict (and later, as reconciliation), it would be helpful to examine this conflict more precisely. Hegel analyses the conflict in concrete terms by attributing divine law to woman, human law to man, pointing out that ‘[n]ature, not the accident of circumstances or choice, assigns one sex to one law, the other to the other law’ (PhS para 465). A closer examination of the relation between the laws could reveal the conflict as essential rather than temporary, and the sentence just cited might give evidence in this direction. Yet, for Hegel, the relation between man and woman is itself open to reconciliation, through love and procreation. Furthermore, the relation between sister and brother is described by Hegel as a special one. It is much more balanced than the relation between husband and wife since not based on desire and the instability arising out of it. Hegel’s claims about different familial relations evoked much criticism as well as reactions of irony which do not need to concern us here.¹³ Judith Butler criticizes Hegel because he presents kinship and its laws as outside the state, yet nevertheless as its precondition (AC 29). Hegel does not question the rules and laws of kinship as such. For Butler, Antigone is a remarkable text because it facilitates a ‘productive crisis’ of the usual distinctions and terms, and especially a discussion of ‘kinship in terms of its social iterability’ (ibid). As Antigone’s own familial relations are of a rather unusual kind, especially given the fact that her father, Oedipus, is simultaneously her brother, a questioning of the familiar kinship concepts still presupposed in Hegel’s reading becomes possible. Judith Butler’s reading of Antigone makes two main contributions, in my opinion. First, she engages in depth with a spectrum of existent interpretations— most importantly, Hegel, Lacan,¹⁴ and Irigaray—, and she indeed focuses on their respective reading of Antigone rather than repeating already familiar comments about these thinkers which would then also apply to their interpretation of this tragedy. Second, she is audacious in bringing all relevant concepts into motion; her interpretation is not dependent on any fi xed frame. In fact, her reading is guided by the conviction that an entirely new concept of the human can emerge from Antigone. This approach makes it possible to focus again on the actual text of the tragedy. Butler is particularly interested in the relation between words and deeds. Antigone’s deed is announced in words beforehand; afterwards, she accepts the deed by not denying it. Deeds and words are intertwined, deeds respond to words ¹³ Interesting is Irigaray’s critique as she points out that the relation between Antigone and Creon can already not be a balanced one due to the fact that he has the instrument of law which is not at her disposal. As Antigone represents law as fi xed and pregiven, she cannot utilize it in the fashion in which Creon, representative of law as created, has the latter at his disposal. cf L Irigaray,‘The Eternal Irony of the Community’ in L Irigaray, Speculum of the Other Woman (Ithaca, 1985) 219. ¹⁴ As mentioned above, the psychoanalytic dimension of the tragedy will be bracketed in my article.

146

Tanja Staehler

and gain their author through words. In turn, norms emerge as practices, instituted through repetition. This intertwinement of words and deeds, of norms and practices enables a redefinition of kinship. Kinship ‘is what she [ie Antigone] repeats through her action’; in general, ‘it is not a form of being but a form of doing’ (AC 58). These insights allow, as Butler explains in the final pages of her text, to interpret kinship in a different, much more creative fashion. Rather than presupposing what a family or a relation between brother and sister is (as Hegel tends to do), we ourselves determine through our actions and practices what kinship can be, and how we can live our lives in various formats which are not predetermined. In the context of this admittedly very brief and incomplete summary of Butler’s reading, I would like to raise two points of criticism, the second of which emerges from Butler’s own philosophy. First, the concept of law does not play a prominent role in her interpretation of Antigone; its place is taken by the concept of a norm. This shift and Butler’s way of interpreting norms can be described in terms of our preceding discussion regarding law in its two aspects. While the second aspect of law is emphasized, namely, law as created by humans, the first aspect of law as pregiven and unchangeable gets lost almost entirely. This might seem surprising, given that Butler is clearly more interested in the figure of Antigone than in Creon. Yet by way of her non-traditional reading, she inadvertently emphasizes that very character of law which Antigone herself does not explicitly highlight. Despite Butler’s critique of Hegel, her reading is somewhat reminiscent of Hegel’s claim that both standpoints contain an element of the opposite side. While it is indeed beneficial and helpful that Butler takes no frame of thought for granted and does not employ any fixed concepts of either norms or kinship, it seems that the invoked ‘productive crisis’ loses an aspect of law which any description of how we experience law or how law is given to us would need to yield. With the loss of this aspect, there is also a loss of the conflict inherent in law, between law as created and law as binding. At certain points of her text, Butler implies that a conflict of law (or, as she would put it, norms) presupposes that laws are static—and hence fails to reach a deeper level where the nature of law and norms itself becomes questioned (cf AC 67 and 76). Yet the conflict at the centre of law presupposes exactly a dynamic or genetic reading of laws as created and pregiven, changeable and binding. If the conflict at the heart of law defies a static account, can we specify more exactly what gets lost through Butler’s emphasis? In my opinion, the character of law as binding or divine emerges, on a secular reading, to be most present in ethical laws, ie ethical principles which strike us as intuitively plausible (although more work is necessary to clarify this plausibility).¹⁵ This leads me to my second ¹⁵ In his critique of Austin, Dworkin expresses very well that the problem of clarifying this aspect of law is rather different from the need to acknowledge its existence: ‘[P]erhaps our feeling of some special authority attaching to the law is based on religious hangover or another sort of mass self-deception. But Austin does not demonstrate this, and we are entitled to insist that an analysis of our concept of law either acknowledge and explain our attitudes, or show why they are mistaken’

Antigone and the Nature of Law

147

point of criticism: Butler’s reading of Antigone refrains from discussing ethical implications. Such abstaining is particularly surprising in light of Butler’s own move to the ethical realm, especially in her writings after Antigone’s Claim. The relation between ethical and legal law will be discussed in more detail below; this part of the essay will conclude with some suggestions regarding the ways in which Judith Butler’s reflections on ethics, especially in her 2002 ‘Adorno Lectures’, could be brought into a reading of Antigone.¹⁶ In examining the foundation of ethics, Butler problematizes the usual assumption that there is a self-identical and accessible ‘ego’ which can be made responsible for practical decisions. What does it mean, for ethics, if I am not transparent to myself, not separable from social norms which penetrate me? What does it mean if I only have access to myself by way of narratives which I tell about myself, yet which are not entirely my own stories? If there is ignorance at the core of the self, can this ignorance have an ethical value or significance (cf KdeG 95)? Butler responds that such ignorance and elusiveness indeed have ethical significance since they show that we are always already related to others and permeated by otherness. This otherness can also take the shape of norms—or of law. In the case of Antigone, such inaccessibility plays out in the ways in which it is difficult to fi x the characters in their roles, and in the problem of deciding who is right and who is wrong. Who, for example, is Polyneices? Antigone’s brother or the enemy of the state? If he is both, can it be decided which identity is more important? Are the reasons given by Antigone and Creon for their deeds and decisions really their true reasons? Whenever a person tells their story, it will never be the whole story, and whatever story they tell will be told in a language which is not entirely their own—here Butler and Heidegger agree. Their agreement coincides with the first choral ode in Antigone, in which language appears as an instrument of man to master nature and fellow human beings; it is a precarious instrument which, despite pretences to the opposite, we do not ultimately have in our power. What do these considerations mean for the nature of law? What is the relation between ethics and law? In the final part of the paper, we need to retrace our steps before taking up these questions.

III. On the most general level, it was Hegel’s merit to introduce the concept of law and the conflict between human and divine law into the discussion of Antigone. However, these accomplishments lose some of their original strength when Hegel (RM Dworkin, ‘Is Law a System or Rules?’ in RM Dworkin (ed), The Philosophy of Law (Oxford, 1977) 40). ¹⁶ J Butler, Kritik der ethischen Gewalt [hereafter: KdeG] (Frankfurt, 2003).

148

Tanja Staehler

ends his interpretation on the note of reconciliation occurring on the level of Spirit which surpasses the individual. Martin Heidegger and Judith Butler moved the discussion to the level of the human, gaining important insights into human nature. At the same time, they bypass the significance of law for the tragedy. As we recapture the main steps of the interpretation developed so far, the important insights provided by these thinkers will be complemented by extensions of their interpretations provided in part II of this essay. Building on these insights and their extensions will then allow for a further discussion of significant aspects of law. Hegel conceptualizes the conflict between Antigone and Creon as a conflict between human and divine law. Although Hegel’s reading of these two standpoints as representative of modern ethical theories (especially deontological ethics) and ancient ethical theories (especially virtue ethics) is intriguing, such a reading has the disadvantage that the conflict appears tied to a particular moment in human history, ie the emergence of a modern impetus to question pregiven norms and values. Hegel would like us to recognize this tension as belonging to morality in principle, but he also thinks that higher levels of ethical life (German Sittlichkeit) can be reached where this tension no longer constitutes an irresolvable conflict. However, if the conflict is described in terms of two aspects of law, ie law as binding and pregiven versus law as created and changeable, the conflict gains an essential rather than historically situated dimension. Hegel himself points out that the divergence between family and state is valid throughout the historical epochs, and this idea becomes even more plausible when the conflict presents itself as a clash between morality and legality, or between ethical and legal realms. While the ethical dimension is present from the outset and already emerges with Hegel’s interpretation, the interpretations provided by Heidegger and Butler explicate this dimension further, especially if considering possible extensions of their positions. After both of the fundamental positions in ethical philosophy, namely, deontology and consequentialism, have revealed certain basic flaws (most manifest, for deontology, in the conflict between duties and, when employing consequentialism, the problem of intuitively unethical decisions for minorities), the response might be to refine these positions further. Another possibility concerns exploring new options for approaching ethical issues by linking them to the basic question, what does it mean to be human? According to Heidegger and Butler, this is the very question of Antigone. The question leads us into ethics, on the most general level, because being human means being entangled in ethical decisions and dilemmas, being confronted with responsibilities of a conflictual, if not aporetic or paradoxical nature. Heidegger’s discussion of man’s most uncanny nature invites an ethical interpretation, even if Heidegger does not explicitly raise this dimension. When Heidegger explains how humans engage in various forms of violence-doing and how this violence is not anything that we arbitrarily do, it becomes obvious that

Antigone and the Nature of Law

149

it is indeed part of our human nature to create complexity in our world. We make ourselves at home and overstep the boundaries of this home, only to encounter a higher and more complex level of uncanniness which raises new responsibilities and demands. As we design rules and laws, the possibility of dilemmas between the legal and the ethical sphere opens up. By explaining these complex structures as part of our human existence, Heidegger also implies that the appropriate reaction cannot be an avoidance of such violence and uncanniness, but a better understanding. Phrasing human dilemmas, with Antigone, in terms of uncanniness points to the significance of analysing moods (German Stimmungen)—a project which Heidegger undertakes in Being and Time as well as various lecture courses. Against the Kantian emphasis on neutral rationality, moods turn out to be a significant part of our existence in the world; they even have ethical implications.¹⁷ This does not mean a return to a ‘Philosophy of Moral Sentiments’ (Hutcheson, Hume, etc) but a request to take moods seriously and analyse them anew. Tragedy is one of the places which allows for a serious engagement with moods. Tragic dilemmas are often motivated by pride, fear, disappointment, etc, and we can only sense the full extent of the dilemma as the characters react with horror, shock, deep sadness, or anxiety. In its character as a pathos, ie as that which overcomes us and is out of our control, a mood also bears a similarity with law, namely, with law in its character as pregiven rather than created. Furthermore, we tend to experience this characteristic of law by way of a specific mood: awe.¹⁸ After Hegel and Heidegger have introduced the ethical dimension of law, we will continue further in this direction with the help of Judith Butler before thematizing the relationship between law and ethics more generally, thereby discovering other characteristics of law. Butler places Antigone within her family context which is designated by highly complicated kinship and incest relations as well as tragedies emerging from human ignorance. These complex kinship relations allow for a new investigation of what it means to be a brother, a daughter, etc. Fulfilling a certain kinship role mainly depends on how a person takes on this role, performs it by repeating or disrupting social norms, and implements this role into personal narratives. In her writings on ethics, Butler generalizes this insight; we are mistaken if we assume a fully transparent, fully responsible ego at the basis of a person’s actions. All we have are narratives and accounts, and even though some of those distort the actual events to a greater extent and perhaps more obviously than others, there is never one simple identity constituted. ¹⁷ This is a controversial topic in Heidegger scholarship as Heidegger himself excludes an ‘evaluative’ reading of Being and Time. Yet especially Heidegger’s reflections on the ‘Call of Conscience’ do have ethical implications if ‘ethical’ is distinguished from ‘moralizing’ and ‘evaluative’ (as I think it should be). ¹⁸ This is ultimately a Kantian insight; yet Kant was neither in a position to sufficiently analyse this mood as a mood, nor did he acknowledge that awe extends beyond the moral law (or categorical imperative) to law in a wider sense.

150

Tanja Staehler

In the context of the tragedy, it is not clear whether Polyneices is mainly Antigone’s brother or mainly an enemy of the state (as mentioned above). With his death, the situation becomes even more complicated since others have to provide narratives about him. The tragedy can be compared to a courtroom situation, with Antigone and Creon appearing as witnesses. On the first level, each of their positions is rather straightforward, yet on a different level, there are components which complicate the emerging picture. Creon, for example, states that he cannot bear to make an enemy of the state his private friend (cf Ant 187). However, can we assume his sincerity in this moment? Is he truly torn and tormented, or is he perhaps also happy to be the new ruler? Antigone, on the other hand, provides a strange statement to justify Polyneices’s significance for her—a statement so strange that it is often omitted when staging the play. She argues that Polyneices’s singularity hinges on the fact that he is irreplaceable to her as a brother, given that her parents are already dead (cf Ant 905). This turns Polyneices’s singularity into a biologically contingent one, and it also leaves us wondering whether Antigone buried him for his or for her own sake. The narratives or witness accounts which we receive are thus not fully coherent or free of contradiction. Butler stresses the ethical significance of ignorance or of not knowing the entire picture; Oedipus is the paradigm example, but not the only example in his family, of such ignorance which is relevant to actions. Already Hegel pointed out, in his critique of Kantian ethics, that being ethical for humans can mean having to act without being fully in the know. In many cases and situations, it would be unethical to stand by and not act due to an incomplete knowledge of the relevant factors (cf PhS para 642). Against Kant, Hegel states that it is the good deed, not the goodwill, which is good. At the same time, we are always already guilty when we act because we need to act without knowing with certainty whether we are doing the right thing. The role of the gods and of divine law, in Antigone, shows exactly the contrast between the omniscient gods and the finite humans. Both Antigone and Creon acknowledge that they ultimately depend on the divine judgment; they themselves cannot know the full picture. Reflecting on her deed and her punishment, Antigone states: ‘If this is what the gods approve, why then,/When I am dead I shall discern my fault;/If theirs the sin, may they endure a doom/No worse than mine, so wantonly inflicted!’ (Ant 925). Creon, in turn, admits that he cannot fight against the goddess Necessity (Greek Ananke), and that the many misfortunes inflicted on his family by the gods prove him wrong and guilty (Ant 1105). Judith Butler has thus moved us further in exploring ethical dimensions of legal conflicts, both in terms of a problematic self-identity and in terms of human ignorance or fallibility. These ideas can be linked to Antigone, even if Butler herself does not develop them fully in Antigone’s Claim; the special potential of this tragedy is thereby confirmed. On a more general level, Antigone exemplifies a conflict inherent to law which I would like to discuss under the heading of ethics versus politics. The term

Antigone and the Nature of Law

151

‘politics’ might appear out of place at first, especially if applying our modern understanding of the political. The Kantian distinction between morality and legality captures basically what this conflict between ethics and politics entails. However, I prefer using the terms ‘ethics’ and ‘politics’ since Kant’s terminology is tied in with his conception of the ethical (ie the deontological) and also makes it difficult to explore the nature of legality in depth. Since the ancient Greeks were the first, in the Western tradition, to explore the nature of ethics as well as politics, these terms seem appropriate for the context of Antigone. ‘Politics’ would then encompass the area of laws, both in the sense of specific laws and in the sense of a political constitution. At stake in Antigone are mostly specific laws, and particularly Creon’s law not to bury Polyneices. Ethics, translated into Hegel’s terminology, would correspond to the divine law, and to the character of law as binding and pregiven, whereas politics represents the human law, created and changeable. Antigone exemplifies this conflict in a particularly pressing and convincing fashion. The play is in a special position to do so, not because the conflict as such was born during Sophocles’s epoch, but because it is for the first time consciously understood. The conflict between ethical laws and legal laws arises as soon as humans organize their life in groups and establish rules for living together. However, the ancient Greeks started to thematize differences between constitutions, discussing which ones are superior to others. With the explicit juxtaposition of monarchy (as existent in Thebes at the time of Creon’s rule) and democracy (as existent in Athens at the time of Sophocles’s writing), the function of law in relation to the citizens is no longer taken for granted and becomes less stable. Not taking things for granted is a definition of philosophy which most philosophers could agree on. Philosophy begins as the world and objects in it inspire wonder and questions rather than being taken for granted. It varies widely, of course, whichdirectionthesequestionsandthecorrespondinganswerstake—alreadyforthe Greeks. But Plato and Aristotle agreed on wonder as the beginning of philosophy,¹⁹ and they also agreed that ethics, or the question as to how to live a good life, is central to philosophy. However, Platonic ethics differ substantially from Aristotle’s, and the same holds for their political theories. Sophocles, on the other hand, does not develop an ethical or a political theory as such. This makes it possible for him to thematize the conflict between sacred laws and man-made laws in an immediate fashion rather than as a clash between two complex areas of philosophical theory. Arguably, the most significant concern in the political domain is the concern for ‘inside’ versus ‘outside’, or the concern for borders. The city laws or laws of the land relate to the ‘inside’, trying to preserve and protect it. It is essential that Antigone meets Ismene outside the city walls:²⁰ ‘That’s why I’ve brought you ¹⁹ cf Plato, Theaetetus 155d and Aristotle, Metaphysics 982b. ²⁰ Although there is no immediate connection, it is still interesting to keep in mind that in the Platonic dialogues, Socrates almost never leaves the city walls of Athens—in line with his commitment

152

Tanja Staehler

here/Beyond the gates that you may hear my news/Alone’ (Ant 18, translation modified). Antigone’s interest does not just lie in meeting with Ismene alone; she wants to meet her on neutral grounds, as it were, or on grounds in which the sacred law rather than Creon’s law holds. Again, the conflict between different laws or principles, or the tension between the ethical and the political does not first emerge with the Greeks; and yet, they are the first who are able to reflect on it since an explicit awareness of the political, its essence, the differences between different cities, etc, has emerged. Antigone’s story is a manifestation of this conflict which arises here as a tragic and irresolvable one. But is the conflict truly irresolvable? Our interpretation, based on selective and critical readings of Hegel, Heidegger, and Butler, has confirmed the nature of the conflict as irresolvable. The most basic argument for an irresolvable conflict which could be given at this point, after tracing the conflict back to the tension between ethics and politics in the way explained here, stems from their divergent relations to borders. Political and legal laws inevitably have to be concerned with borders and hold for a defined territory; ethical considerations overstep and exceed such borders.²¹ For Antigone, her brother’s role in relation to the state does not matter when it comes to the sacred laws of burial. For Creon, state laws do matter more than anything else; and it is indeed his duty to uphold them.²² In conclusion, a common sense objection about the nature of the conflict needs to be addressed which might already have arisen earlier. Is it even true that we read Antigone as a tragedy about an irresolvable conflict? Does not Sophocles present the story in such a way that we take sides with Antigone, judging Creon to be arbitrary in his legislation and stubborn when it comes to making an exception? Before treating this objection, it is important to note that the possibility of exception represents another facet of the tension between ethics and politics. Political laws need to hold in an exceptionless fashion whereas ethics often requires us to respond to the singularity of a situation or the uniqueness of a request. When an exception to political law is granted, it has to be possible to make a special case, such that the exception becomes a justified exception and hence no longer truly an exception. Creon deems it impossible or at least unjustified to make such an exception for Antigone. Returning to the objection, it seems to me that the conflict is indeed constructed in the tragedy as an irresolvable one. Both Antigone and Creon exhibit to the city and its laws—, yet the dialogue Phaedrus leads Socrates outside the city, perhaps because the subject-matter discussed here (Eros) is not political by any means. ²¹ The cases of illegal refugees dying at the borders of a country because they do not have (legal) permission to enter is, in certain ways, a contemporary version of Antigone’s fate. ²² The question of human rights which exceed state boundaries leads us too far astray here. Suffice it to say that it seems inevitable for human rights to emerge more closely from certain historical traditions than from others, with the result that human rights are closer to the state law in some states than in others. Also, there are various difficulties regarding the execution of human rights which stem from their precarious legal status.

Antigone and the Nature of Law

153

certain questionable character traits which might make us doubt the purity of their motivation, but these character traits are ultimately irrelevant to their standpoints. When considering their respective positions, it is not possible to decide who is right and who is wrong. Perhaps Creon should have foreseen Antigone’s reaction and the problems that would emerge not just for Antigone’s family but, through Haemon, also for his own; yet that is a personal, not a political argument, and Creon is motivated by political concerns. If it is difficult for us to acknowledge the full extent of the irresolvable conflict, then this difficulty is based on problems to understand what exactly it means if a person is left unburied (such that their soul cannot find peace) and what it means if an enemy is buried on home grounds (thereby reintegrating him into the community in a dangerous fashion). Yet we can still imagine what these problems mean; more importantly, this dilemma elucidates another aspect of the conflict between ethics and politics. Ethics is a concern for the soul and does not seize with death. Politics is predominantly a concern for the living and their well-being. Hence the appropriate designations ‘divine law’ and ‘human law’. We have thus seen, in this essay, that there is an irresolvable tension at the heart of law which can be described in different ways. On the one hand, law appears as divine law, given to us as unchangeable, or as a higher authority which we are subject to. This aspect of law is closest to the ethical realm. On the other hand, laws are human laws, created or at least formulated and interpreted by us, changeable and open to criticism and improvement. It is this character of law which comes to the fore in the political and legal realm. Interestingly, the created and changeable dimension of law is simultaneously the one which calls for an execution of law without exceptions; yet due to its ethical relevance, law sometimes has to allow for exceptions. Creon realizes too late that Antigone’s action might to a certain extent be justified, and Antigone never explicitly acknowledges Creon’s position. Antigone represents the conflict between these two dimensions of law as an irresolvable one, inevitably yielding a tragic outcome. Or, speaking with Hegel, tragedy exhibits a ‘collision of equally justified powers’ (Aesth 1213).

IV. Postscript The questions and themes treated in this essay might appear not to be immediately related to traditional debates in philosophy of law; yet there are, in fact, several connections. To be sure, discussions in philosophy of law as they emerged from the controversy between HLA Hart and RM Dworkin have reached too complex a level to allow for a contribution in the form of a postscript to an essay on Antigone. However, I would like to mention a few links between the ideas raised in this essay and certain themes treated by RM Dworkin and others. These connections concern the ideal judge, ‘Hercules,’ on the one hand, and the problem of hard or difficult cases on the other.

154

Tanja Staehler

Since the conception of the mythical judge Hercules,²³ whom Dworkin imagines to have unlimited intellectual powers, provides a natural connection to the ancient Greeks, it appears permissible to speculate about the consequences of Antigone for Hercules. In the tragedy, the gods are the ultimate judges, and their decisions are right; the position of humans is defined exactly through their dependence on the judgment of the gods which they cannot anticipate (as Creon in particular has to realize). The interpretation developed here on the basis of Hegel’s and Judith Butler’s ideas stressed the importance, for humans, to acknowledge our fallibility and its ethical significances as we need to act without knowing all relevant circumstances. Of course, a judge’s decision differs from human ethical action as there will be less time pressure exerted on the judge. Nevertheless, the judge is dependent on human accounts which always give a partial picture, even if they tell the truth and nothing but the truth as it is known to them. In that sense, the lesson to be drawn from our interpretation of Antigone lies in the importance to point out that human judges are exactly not like Hercules. Yet there is an even more interesting lesson which can be learned from Greek thought in general. The Greeks were able to think two ideas simultaneously which nowadays strike us as contradictory: human nature was defined precisely as different from the immortal and omniscient gods, and ‘knowing yourself’ meant to abandon all hubris, acknowledging human limitation and finitude. At the same time, being human meant striving to become as close to a god as possible, ie striving to approach the ideal of being a god. This simultaneity of approximation and differentiation provides a fascinating challenge for thought. If it was possible for us nowadays to think these two ideas together, then being a human judge would mean striving to become like Hercules while acknowledging that human judges can never really be like Hercules. Related to the idea of judge Hercules is the problem of hard cases which, according to Dworkin, are in principle resolvable. The case of Antigone has been presented here as an irresolvable case, based on an irresolvable conflict between different aspects of law. Admittedly, the Greek legal situation is too different from current legal systems to cite Antigone as a true counter example to Dworkin’s claim. To say the least, our contemporary intuitions would regard the death penalty as too harsh a punishment for Antigone’s burial of her brother. The question as to whether Antigone is guilty or not seems more difficult to decide (and it is again a Greek idea to detach the question of guilt from the decision regarding the appropriate penalty). On the basis of the preceding considerations and especially the distinction between ethics and politics, I would like to maintain that Antigone is both guilty and not guilty, or that she is not guilty, according to ethical considerations, yet guilty according to state law.

²³ cf RM Dworkin, Law’s Empire (Cambridge, MA, 1986) esp chs IX and X.

Antigone and the Nature of Law

155

A number of arguments have been raised against Dworkin under the heading of ‘incommensurability’: what makes hard cases truly undecidable, according to those critics, is the impossibility to measure one competing law against the other, or the impossibility to find a common standard by which to decide which law ranks more highly.²⁴ The Antigone conflict appears to represent incommensurability on a deeper level. A discussion between two judges who debate which of two legal rights ranks more highly than the other still presupposes that both of these judges honour a specific orientation which can be designated as legal or political. It seems at least possible that, if liberated from time-constraints and truly open to the rational arguments of the other party, a ‘right’ or most plausible answer would emerge for the legal cases under debate.²⁵ Yet the conflict between ethical and legal/political concerns is a more fundamental one, and this deep incommensurability makes it impossible to see what a ‘right’ answer would even be in such cases. One essential dimension of this conflict, namely, the political concern for borders as it clashes with the ethical demand to neglect state borders, has been explained above and is illustrated very well by Antigone. If trying to resolve the conflict between legality and ethics by arguing that ethical principles always provide the ultimate guidelines for legal rules,²⁶ the problem is only shifted to a different level. Practical considerations and problems of such a suggestion aside, it turns out that ethical principles are themselves ‘hard cases’, as it were. In respect to Kantian philosophy, certain ethical dilemmas continue to be discussed. Regarding the often-raised problem of lying to an assassin in order to protect a guest in the house, Kant himself has claimed that the duty of honesty cannot be abandoned for humanitarian principles.²⁷ If there is, according to Kant, not even commensurability for comparing such duties as ‘saving life’ versus ‘honesty,’ much less resolvable issues appear likely to arise. A Hegelian emphasis on the norms and values of a certain community and epoch would resolve some of these dilemmas; but how to resolve the question of Kantian versus Hegelian ethics, and how to compare them to a number of contemporary alternatives? These problems lead us too far astray, or they lead us back to Antigone and confirm irresolvable conflict at the heart of law.

²⁴ cf, eg JL Mackie, Persons and Values (Oxford, 1985), and N MacCormick, Legal Reasoning and Legal Theory (Oxford, 1978). ²⁵ Dworkin himself admits that he can only show such a possibility for specific cases rather than proving that it would always be possible to find a solution. cf Law’s Empire 411. ²⁶ Dworkin’s distinction between principles and rules seems to presuppose that the former are ethical, the latter legal. ²⁷ cf I Kant, ‘On a Supposed Right to Lie Because of Philanthropic Concerns’ in I Kant, Grounding for the Metaphysics of Morals (Indianapolis, 1993).

9 The Moral is: States Make Laws Ross Harrison

For someone more familiar with other parts of philosophy, reading jurisprudence is an interesting experience. The din is terrifying and the pace of argument frenetic. Yet, at least if you just stumble into it in an ill-informed way as I recently have, it can all seem a bit like the war between the big enders and the little enders in Gulliver’s Travels. After all, everyone seemed to be eating much the same eggs, so what does it matter which end one opens first? Everyone seems to agree that it is rather a good thing to have law, that some laws are better than others, and that, on the whole, the law should be obeyed. There is lots of evaluation from all sides; the same wholesome fare is consumed; and there is nevertheless endless argument about what is happening. Is this wholesome goodness actually in the egg? And should it be opened at its big moral end or its small positivist end? At times, or so at least it has seemed to me, the argument resembles one of those parlour games played at Victorian house parties where the players in turn put on an amusing verbal show for the entertainment of others. Such as the game called Consequences where each player in turn supplies the next bit of the story: ‘he said’; ‘she said’; ‘he gave her’; ‘she gave him’; and so on. If I attempted to render the current jurisprudential debate as a game of consequences, I would not be sufficiently attuned to the exact nuance of the action to do the ‘he said’ or ‘she gave’ part properly. However, I think I have the penultimate line, in which the ‘consequence’ of the engagement is revealed. The consequence, I assert, was that they all argued happily ever after. But this is only the penultimate line. Before I started this paper, I was equally confident about the last line, which gives the ‘moral’ (this, unlike the preceding lines, in the present tense). Sufficiently confident, anyway, to submit it as the title of this paper. The moral, I thought, is that states make laws. (You want morals? I’ll give you morals.) And why was I confident, even though, as we shall see in the next two paragraphs, my confidence drained away as soon as I started to write? I thought that this simple idea was not only correct but might also be important in cracking open the cacophony. I thought that we might make better sense of all this jurisprudential noise if, instead of thinking directly about law, we thought more about the state. Get hold of the simple idea that laws are made by states. Then

The Moral is: States Make Laws. Ross Harrison. © Oxford University Press 2007. Published 2007 by Oxford University Press.

160

Ross Harrison

use this link between the law and the state to convert problems about law into problems about states. Typical jurisprudential problems are: What is law; Why have law; and whether or when we should obey the law? Use the simple linking thought and these can be converted to typical problems in political philosophy: What states are; Why have states; and whether or when we should obey them? It is not that these latter questions are simple or uncontentious, but they have had a greater range of thought applied to them for a longer time and so may be more susceptible of solution. That was as far as I had got when I had to supply the title. But now, as soon as I start to write the actual paper, I immediately realise that matters are not as simple as I had supposed. This supposed ‘moral’ I was arrogantly suggesting to the jurisprudes energetically playing Consequences is not as detached as I had supposed. It is not an impartial judgement supplied by an amused bystander or an ideal observer. Like most morals, it is instead a position occupied by someone engaged in the game. The observation that states make laws tends to identify law by its source. Shifting from the morality of law to the morality of states tends to supply the law’s morals from outside its intrinsic content. All sounds very little ender to me. I had thought that I had understood what should happen. I had all of two minutes imagining that I was a political philosopher Gulliver looking down on the legal Lilliputians. That was all it took, though, to realize that my simple thought was not correct and that, even if it were correct, it would be too parti pris to help resolve the argument. I realized that I already had my teeth firmly into the little end of the egg. In spite of this initial setback, I shall persevere with some shreds of my original plan. When I submitted the title, I thought that I had one problem, which I would use the paper to worry about: the basis of political obligation. However, as soon as I started, I realized that I had another problem: the relation of the law and the state. The grand strategy of converting issues in jurisprudence into issues in more general political philosophy is itself problematic. That is the first problem. Then, even if the strategy works, we still shall have the problem of solving these bits of political philosophy. That is the second problem. In this paper I shall have a go, in turn, at both. Even if I don’t manage to get terribly far with either, I hope that this will at least form a basis for discussion of them. Each problem, I think, has an independent interest, even though a particular answer to the first would give additional relevance to the second. And so to the first problem, the one I ran into as soon as I started. This is that, although the claim that states make law can be defended, it is equally plausible to claim that law makes states. So, rather than my hoped-for grand strategy of first understanding the state and hence coming to understand law, we might as easily have the strategy of moving in the opposite direction. We could as easily, that is, first understand law and hence come to understand the state. Rather than the goodness of law being explained by the goodness of states, we explain the goodness of states by the goodness of law. (Now you know what law is you can see why

The Moral is: States Make Laws

161

it is a good thing to have states; although only, of course, if they respect the rule of law.) What are states? Entities that make law. Why have states? Because they make law. When should states be obeyed? When they make laws that should be obeyed. And so on. We constantly see states making law. So why is it equally plausible to suppose that law makes states? The answer is that we need to be able to identify what is doing the making as a state, and also to be able to identify what it is making as law. The second of these certainly depends upon being able to know what law is independently of knowing what a state is, and the first may also do so. The second certainly depends on it because much of what states do is not making law. Hence we cannot identify law as merely what states do. Hence we have to be able to select, among the many things that states do, when they are making law as opposed to making other things. Hence we need an idea of law, independent of the idea of state action. Yet, with this independently achieved idea of law, we then have the possibility of using it to explain states, rather than the other way round. The possibility, but need we do so actually? Here is an argument that we do: states have a legal power, the power to make law; therefore we need a law giving things this legal power before we can recognise them as being states. If you want to know whether something is a state, consult international law, which includes the identification of states among its scope. First we have the (international) law, then a would-be state gets recognized as being one in accordance with this pre-existing law. Law makes states. However, it might be objected, international law is merely the codification of agreements between states. So, if law makes states, it is states that make the law that makes states. What is really happening here is that new states are being recognized by established states; those already in the family allow new members to enter. However, international law has by now moved beyond this point. Allowing, as it does, that some matters are jus cogens inter omnes means that at least parts of it are such that they cannot depend upon inter state agreement, since they do not allow states to derogate from them (as they could from agreements), and since they apply to states whether they have ever agreed to them or not. Admittedly, the recognition of states is not among these parts. But the fact that international law can exist independently of state construction, together with the fact that recognition of states depends upon international law shows that at least it has the potential to make states without first being made by states. The principal other base that international law rests on (in addition to positive state construction by treaty) is custom. And this merely reminds us of the many other customary sources of law. Local customs are the third standardly recognized source of law in our own jurisdiction, in addition to the operations of the (state) legislature and previous adjudication by the (state’s) courts. This is law without, or at least prior to, the state. And such a merely customary base is true of many other areas of law: the Law Merchant; religious law; and so on. And that is only custom; there are other non-state but also non-customary bases of law, such as natural law.

162

Ross Harrison

Therefore not all law is made by states (and not all that states make is law). Hence if the original idea is to work at all, it has to be put in a more limited or sophisticated form. The most useful idea, I think, is that law may be made and, indeed, frequently is so made. For once something is recognized as an artefact, a whole series of evaluative questions automatically arises. Artefacts are intentional objects; they express human purposes. Hence we can ask questions about ends: What is the purpose of this, and is it a worthy one? And we can also ask questions about means: How well does this fulfil its purpose? Once we decide that something is made (rather than just being found), we can immediately ask what it is for and also, crucially, whether it could be made better. So, I suggest, the most useful starting thought is that laws may be made. Once we realize that law is made, we can ask what it is for. It is not just there, waiting to be discovered; it is designed for some human purpose. So what is this purpose? Is it a worthy one? Also, once we see that laws are made, we can move to the question of how they might be made better. How may these purposes be better realised? Once I see something not just as a stone but as a knife, I then can understand what it is for. I may understand, even if I have only flint to work with, how it might be made better. But as long as it is only a mere lump of stone none of this can get started. To this fundamental idea that laws are made I can add the thought (as a shadow of my original idea) that laws are typically made by states. If so, it enables us to get a better hold on these questions of function, purpose, and value. If, typically, states make law, then we can ask why they are doing so, discover the purposes of law, evaluate these purposes, and ask how they might be better achieved by improvements in the law. You may think that I have twice too easily slipped in that ‘typically’ and that too much of the argument rests on it. However, if you want to know what the law is, have a look at what actually happens when people are engaged in it. One look at what a practitioner does is enough to show that it is nearly all created by the state. The great mass of law comes from primary or secondary legislation, explicit making by the state’s legislature, or arising from powers conferred by it on state officials. This is vastly increasing in quantity year on year; an awful lot of making is going on, and it is the state that is doing it. The rest, also massive in bulk, records the workings of the state’s adjudicative machinery, so that past operations of the courts can form precedent for possible future operations. Whether this is to count as making law, as merely applying it, or as something in between, it is something that is also done by state machinery. These tribunals are state tribunals; the courts and judiciary are arms of the state. No doubt the state would like to monopolize force. But this would be a hopeless ambition; every day privately operated force is exerted. Security is still the state’s primary task; but the hope here is more to control and deter the use of such force, rather than completely eliminate it. More than monopolizing force, the state more plausibly has the ambition of monopolizing justice, that is the exercise of legitimate (legal) force. Even here, as we have seen, it fails. Even so, the state

The Moral is: States Make Laws

163

tends to regard minor tribunals as existing by its permission, and may control their judgements. People disaffected by the decisions of such tribunals may appeal to the state’s courts and receive a different answer. They appeal on grounds of natural justice. But they do not appeal to nature, but, instead, to the state. So, I conclude, states typically make laws; the typical law is one made by a state; the typical application of law rests on a state’s adjudicative machinery and its power to force parties to accept its adjudication. Therefore, to conclude the first of my two topics of examination, I still think that we may gain some insight into the nature of law by appreciating that it is a purposive activity and that it is states that in the first instance have these purposes. Although I concede that this thought is too simply expressed in the title as the bare claim that states make law. By asking what the purposes of states are (why have states; could states be made better?), we get more grip on what the purpose of law is. Putting the question in this way may make it seem that I am treating the state also as an artefact. I do happen to think that this is one of the more brilliant ideas of Hobbes (claiming that the ‘commonwealth’ is an ‘artificial man’); and the title of part one of Burckhardt’s Civilisation of the Renaissance in Italy (‘The State as a Work of Art’) (English translation, London, 1891) is revealing. But I shall not push this aspect; it is enough for heuristic purposes to consider what would follow if it were true. It can, like the explicit state-making described in some original contract accounts, be treated merely hypothetically. The question I shall consider in the second half of the paper, therefore, is why have states. What are states for? Why are we bound to them, if we are? Why do they make laws? Why should we want law, and so, in wanting law, want states that will make law? And so on. These are large questions and I shall soon narrow down to considering only one aspect in detail, where again I am uncertain and so particularly interested in learning the right answer from the discussion. But, first, a bit of scene-setting; or approach on a broader canvas. The primary question of ethics has been said to be: What should I do? The matching primary question of politics would be: What should we do? And this immediately gives us a third question: How do the answers to these two questions relate together? And this, in turn, leads to many variants: What should I do when I know that others will not act similarly? What should I do so that they will? And so on. In social contexts where my individual decision interacts with other people’s decisions, many such questions arise. We may need a moral theory about the second best saying how a good man (me of course) should behave in a wicked world. (Hobbes thought that we are bound in foro interno but not in foro externo to keep our agreements, when we cannot rely on others to do so.) Among these questions is the question of why or whether we should have a state, expressing our collective decision-making, so that the question of what we should do becomes, at least in part, the question of what the state should do. And so to my stated primary question of the rest of the paper, which is why we have states. Why should we make law? Although this sounds like a plural

164

Ross Harrison

question (‘what should we do?’), I think that it is best understood as a generalization over a series of individual questions (why should I have law? why should you?). Apart from anything else, if we already think of this as an exercise in collective decision-making, we may need already to have a state, or something similar, to do the collective decision. So, I think, the first question is: Why should I have states, and other collective decision bodies? Then, later, we can ask: What should we do? What should these bodies collectively decide? Then, later still, we can return again to an individual question: What I should do with respect to these collective decisions? The sequence starts with the question of why I should have a state that makes law. Why should I? What is in it for me? It is a question about my individual reason, but the problem immediately arises of which kind of individual reason is presupposed by the question. Is it prudential, moral, or both? Is the problem why the state is good for me personally? Is the problem why I, morally, ought to have the state? Or what? There are here at least two kinds of personal reason, which, if they agree, do so only contingently. This leads to the problem that Sidgwick dubbed the ‘dualism of practical reason’. (What is good for me personally may not be what I ought to do; doing my duty may be at personal cost.) Sidgwick rather grandly thought that this would reduce the cosmos to duty to chaos. We may be able to tolerate much more conflict than he supposed, but there must be limits to the amount of conflicting rational recommendations to which we can be subject, while still inhabiting an organized cosmos of justifying reasons. This old problem arises just with the personal, individual, question of what I should do. What happens if we now add the question of what we (collectively) should do? That is, if as well as prudential and moral reason we add political reason? (And, if, disliking the collapse of ‘legal’ to ‘political’, we add legal reason; what we should do legally, as opposed to what we should do politically?) Again we may be insouciant about multiplying reasons; when we are about it, why not have aesthetic reasons as well? But, again, the more kinds of reason that there are, the more kinds of rationally irresolvable conflict that we are likely to be caught up in. An ambassador (or so one of them said) is someone who is sent abroad to lie for the good of their country. So what should you, who happen also to be an ambassador, do? If you, as a political leader, should lie for the good of your party, what should you do if you happen also to be an alcoholic? And so on; there are many related problems, such as the paradox of democracy. (What should the committed democrat who happens to be in the minority think ought to be enacted?) The natural way to respond to such threats to the rational cosmos is to suppose that one form of reason takes priority. Standardly, this is morality. Moral reasons trump other reasons, and I can never be in an irresolvable problem about what I should do when morals conflict with self-interest. Therefore, even if we add political (or legal, or aesthetic) reasons, the natural strategy would be similar: one kind takes priority and automatically wins out in case of conflict. Moral reasons trump political ones; ambassadors do not lie. Such priority does not require the

The Moral is: States Make Laws

165

reduction of all other kinds of reason to the favoured kind. It is enough that one kind trumps the others; that it is lexically prior. For example, we may suppose a framework of (moral) rights acting as side constraints on action. This is lexically prior to other reasons. But, providing we respect these side constraints, we can act for other reasons, making happy prudential play while observing these limits. Similarly, we could have moral limits to politics or aesthetics, but still leave a field on which these reasons could properly play. Or we might have prudential limits to the lot, so that some sacrifices may never be rationally demanded of us. And so on. That is one bit of background. Now let us consider another way in which I can nest one sort of reasons inside another sort, which are taken to be ultimate or superior. Take again personal reasons for action and the idea, most familiar from rule or motive utilitarianism, that my immediate reasons for action may differ from my ultimate ones. My reason for doing something is that it is a rule of common sense morality, but my reason for having this rule is that general observance of it will produce the correct moral consequence. The rule (the law) gives me reasons, but my reason for having these reasons is based in overall utility. With these bits in play, we can at last approach reasons for having states or state-made law. The one that I shall consider is not the most commonly given one, which starts with individual prudential reason. This common account starts with the observation noted above that there are often only advantages in me in behaving in a particular way if others also behave in this way. We have to co-ordinate our behaviour for us each to gain individual advantage, and in various ways states assist such co-ordination, getting what I should do and what we should do into line. The state, in this common account, gives me more reason to depend on the actions of others by sanctioning deviations from the generally agreed, or otherwise prescribed, line. Starting with prudential self-interest, it is an assurance game, with the state underwriting the assurance. However, the reason that I shall concentrate on is not a variety of this common justification for states and law, although it is compatible with it. Instead of being based on prudential reason, it is based on moral reason. The argument is that I have a moral reason for there being law-making states because it assists me in achieving my moral goals. Starting with my individual moral, rather than self-interested, reasons, I use the state as an indirect means of achieving them. Assurance, as we shall see, still is the name of the game. But the person whose actions I wish to ensure is not now someone else, but is instead myself. ‘Be good, sweet maid’, intoned the 19th-century clergyman Charles Kingsley, ‘and let who will be clever’. Setting aside the patriarchal implications that only men (or clergymen) are permitted to be clever, could there ever be anything in such a thought? After all, cleverness is normally supposed to be a means to an end rather than an end in itself. We are clever to serve some other good. And if we serve another good, is it not better to be clever in how we do it? Suppose that the good is being good. Then will we not be more good, the more clever we are about it? Therefore sweet and unsweet alike should strive to be clever.

166

Ross Harrison

But suppose that thinking about what we are trying to do in fact makes us less successful in performance. Then the really clever thing is not to try and be clever at all; as soon as we do, we end up under-performing. Consider again rule utilitarianism, the idea that we should just follow the rules and not consider out final goal of maximizing utility. The standard problem with it is why we should follow one of the rules in a particular case where we see that following it would not produce the overall goal. Surely, instead of stupidly following the stupid rule, we should instead be clever, constantly thinking about the consequences of following it blindly? Well, yes, providing our cleverness gives us the right answer. But it is possible that being clever in this way actually means that we actually do less well over time in promoting our overall goal. There are many reasons for this. If I think about it, I may pull the calculation in my own personal interest instead of my moral goal of maximizing the general interest. Conversely, if I think about it, I may become too dispassionate and lose all motivation to do it. Or the more I think, the more obscure it may become what I am really meant to do. Whatever the reason, it may be better for me not to think and just to follow the rule. I know in doing so that I shall sometimes not be doing what is best for my overall end, but since I cannot reliably identify these occasions I do better in the long run just following the rule. We have here an argument for indirect utilitarianism, or, more generally, for indirect approaches to my overall goal. The final justifying reasons are not the reasons that I should think of when I am actually engaged in action. My self, we might say, should be split. In my ultimately rational, guardian, commander part, I prescribe to myself reasons for action. Then, in my day-to-day performance part, I blindly follow these rules, merely obeying orders. However, any such spilt poses problems. It is difficult to keep my acting part separate from my deciding part, or forget about my role as a commander when I am pretending to myself that I am merely a foot soldier, obeying orders. If these orders are my orders, what stops me always going back and rethinking their correctness? How can I forget, with one part of myself, the cleverness of my other part? The split can only be made to work correctly by drastic cognitive self-manipulation. I have to make myself forget my real reasons for action; I repress my cleverness and fetishize its results as words of the absent father. With the state, however, we avoid the problem, and, in avoiding it, we give the moral problem a political solution. Suppose that, instead of my higher, commanding, self, the rules are decided on by the state, perhaps by some collective decision-making process in which I participate. It is what we think ought to happen. Collectively, we decide upon and enforce the right set of rules. Now, with the clever position occupied not with my individual but instead with collective decision-making, no drastic or implausible self-manipulation is required. I can transparently know my immediate reasons for action, and also our reasons for these being my reasons. Because making the rule is a matter of collective decision and so outside my individual control, there is no temptation for me to backslide

The Moral is: States Make Laws

167

and alter the rule. The position of ultimate moral judgement and the position of immediate action are under different control and so can be thought together without the one undermining the other. That is one dualism, or incompatibility, that can be solved by bringing the state in. Another is Sidgwick’s original problem of the dualism of my practical reason. Or, at least, it can be solved in a utilitarian context, such as Sidgwick himself inhabited. For consider classical utilitarianism, as for example in Bentham. The immediate motivation for action here need never be other than self-interested, prudential, reason (Sidgwick’s ‘rational egoism’). The question why I obey the law, for example, is solved by the sanctions it threatens for disobedience; possible punishment gives me an interest for conforming. So also, in a democracy, with the question of how I decide on how this law should be. Here it need not be supposed that I vote on any other grounds than my perception of my own interests. The democratic system is an information-gathering machine which takes as input every individual’s estimate of their own happiness and then, by means of majority decision, delivers as output the greatest happiness of the greatest number. So also with the question of how state officials should act, whether they count votes, enforce laws, or whatever. All that need drive them is guarding their pay packet and avoiding removal or censure. The whole machine works on a basis of rational egoism, of self-interested action. This may all seem rather grubby. Where did the morality go? But self-interest is not, in fact, the only reason at work in this account. Rather, the whole project is to construct a system in which these people, acting on self-interested reasons, will in fact deliver the correct moral result. They act for their own happiness but in fact deliver general happiness, the impartially considered maximum of happiness that it is their utilitarian duty to deliver. Politics, the state, puts self-interest in line with duty and so cures Sidgwick’s puzzling dualism. The self-interested individuals are placed in a structure so that, merely following their natural selfinterested courses, they will deliver the good. Sometimes the structure is literal, as in the panopticon prison (‘morals reformed . . . by a simple idea in architecture’). But it should be remembered that in Bentham’s infamous plan, it is not just the prisoners but also the managers who are under the all-seeing eye; it is publicity that renders them both honest so that they each are led by self-interest to do their duty. Therefore a reason for having state and law is not just the coercion of others so that we can rely on them. It is also the coercion of ourselves, so that we can rely on ourselves to deliver our ultimate goal. But, whether for others or for ourselves, this argument would seem only to make sense in such a utilitarian context. With consequentialist morality, only consequences count. It is irrelevant how these consequences are reached; providing that the right stuff comes out in the end, it does not matter how perverted is the plumbing. So we can use self-interested reasons, threats, manipulation, or whatever works, as long as we get the right results. This is why this argument works from such a utilitarian perspective. But

168

Ross Harrison

utilitarianism is a minority taste. What interests me, and what I hope we might discuss, is whether any analogous argument could be used from a non-utilitarian base, such as virtue ethics or an ethics of intention rather than result. The supposition that morals essentially involve the good will is the hardest case, and so the one on which I shall concentrate for the rest of the paper. For in this Kantian context, or so at least it seems, such an indirect strategy would not be available. As long as I am aiming at the good, then my will is good and I am doing just fine. But if I introduce another set of reasons aiming at something else, such as my own pleasure, then my will is no longer good and I have ceased to be moral. And it seems that this must still be so, even if my reason for producing or having these other reasons is to reach the good. These good consequences would be reached for the wrong reasons, and, since it is not what happens that counts but why, this would not be moral at all. Test this with a couple of examples. Suppose that I think that I should give a certain sum of money to a hospital so that they can provide needed health care. If I give for this reason, then I am good. However, suppose that I am too selfish to operate on these reasons although I know that I should. So I am not good. But then the state steps in. It taxes me the same amount of money as I think that I should give and uses it to support the hospital. Now (from a Kantian point of view) am I lucky or unlucky? That is, am I favoured because the state, by forcing money out of me through the tax system, has enabled me to do my duty? Or am I unlucky because what would before have been a moral act is now rendered a merely self-interested one because I am only avoiding the threatened sanctions of the tax man? Same good consequences of course, but in this morality consequences in themselves do not count; all that is good is the will. Or try another example. I might think that it was my Kantian duty to drive at a safe speed in a built-up area. But I find this rather a bore, and tend to backslide. Then one day, luckily or unluckily, speed limits and a speed camera is installed on my local stretch of road. The state, that is, declares that it will sanction noncompliance with a particular prescribed limit; installs information-gathering apparatus to discover when the limit is exceeded; and (let us suppose) will reliably act on this information. Again, am I lucky? Do I get better at my Kantian duty by now driving at the speed that I always knew that I ought? Or has my chance of performing Kantian duty irretrievably escaped because my will is now concentrated solely on avoiding heavy fines and additional points on my driving licence? In both cases the natural answer would seem to be that the advance of the state into the area is bad news from a moral point of view. Instead of the state enabling me and others to be moral, as in the consequentialist account, my opportunities to be moral get squeezed out by over-zealous state activity. (Just as it is thought that too much state interference in the economy squeezes out people’s private economic opportunities; in each case an attempt to help in fact diminishes or destroys the valuable activity.)

The Moral is: States Make Laws

169

Indeed, it is precisely from such a moral standpoint that we get philosophical anarchist arguments that there should not be a state at all. One example is Robert Paul Wolff, in his Defense of Anarchism (Harper, 1970). Writing from an explicitly Kantian point of view, Wolff takes our primary moral duty to be autonomous, to prescribe the (moral) law to ourselves. Yet the primary feature for him of authority is that others tell us what to do. Hence authority automatically undercuts autonomy, and so morality. As he puts it, a just state is as much a contradiction in terms as a round square. An earlier, and more closely analogous, example is William Godwin’s Enquiry concerning Political Justice (University of Toronto Press, 1793/194b). Godwin asks what happens if someone adds an incentive to a moral action. ‘Let us suppose’, he says, ‘positive institution to interfere, and to annex some great personal reward to the discharge of my duty.’ He then notes that ‘this immediately changes the nature of the action. Before, I preferred it for its intrinsic excellence. Now, so far as positive institution operates, I prefer it because some person has arbitrarily annexed to it a great weight of self-interest.’ The crunch, or conclusion, immediately follows: ‘But virtue, considered as a quality of an intelligent being, depends upon the disposition with which the action is accompanied. Under a positive institution then, this very action, which is intrinsically virtuous, may, so far as relates to the agent, become vicious.’ So, just as we have seen, the sanctions turn virtue into vice. For Godwin the morally required result is, again, anarchism. Once the state introduces a criminal law system, threatening sanctions to deter people from assault, rape, or murder, then abstaining from these activities ceases to be moral. So that would seem to be the answer. While the indirect argument may work with consequentialist morality, it will not work with the more common morality of intention. For it does no moral good forcing either others or ourselves to do what morality would otherwise require; it does not succeed in making us moral as we are doing it for the wrong reasons. So, or so at least it would seem, we cannot use these kind of arguments as a moral defence of the state. We cannot on this basis say that we should make states and law. This reason cannot show us how to make law, nor how to make it better. Indeed, as we have just seen, this moral base easily yields anarchism. It is not just that there is no moral reason for having the state; there is a conclusive moral reason against it. Any reasons for legislation must therefore lie elsewhere and, since other reasons are trumped in this perspective by moral reasons, any such other reasons can never prevail. But is this really so? Is it really true that a Kantian could not engage in the kind of indirect reasoning that a utilitarian can? Why can we not use law and state to manipulate ourselves into being moral? Why can a Kantian not provide legislation with an analogous moral base, as a support for good action? From now to the end, I shall run through a series of cases designed to show that a Kantian might, after all, permissibly do this and hence attempt to give the manipulation argument for the state considerably greater reach and purpose. Since examples in

170

Ross Harrison

moral philosophy currently seem largely to run on rails, I shall do it by a series of examples about the Underground, which is London’s best answer to the trolley problem. (Or at least it is when it works.) So I have a Kantian moral duty, say to deliver the Eurostar ticket I had promised to give Julia at the barrier before 2 pm today. The promise gives me a good moral reason (and hence a good reason simpliciter) to hand the ticket to her by that time at the Eurostar barrier in Waterloo Station. So far, so simple. But, being where and when I currently am, in the front University College London quadrangle at 1.30, the only way that I can do this is to take the Northern Line Tube, boarding immediately at Warren Street station. Therefore, surely, I have a good reason for immediately making for and entering this station. This seems straightforward: a good reason for the end is also a good reason for the unique means. (Providing, perhaps, that the further condition is met, as it is here, that this means is in itself a morally permissible action.) However pure a Kantian you are, surely you have to accept this. Yet, already, my moral end gives me a good reason for doing something which is in itself morally indifferent. There is nothing moral about now entering Warren Street station and yet in doing so I am acting from moral good will. Ah, but you say, what undercuts moral action in cases like the hospital or speed limit is that I there act from other, self-interested, reasons. So, change the example to make my boarding the Tube at Warren Street motivated by self-interest. In fact, there is more than one way that I can get to Waterloo on time; I could go by Tube but I could also take a cab. But I do not like the expense of cabs and I do not like talking to taxi-drivers. So self-interest propels me instead onto the Tube. I meet a smiling Julia at the barrier and hand the ticket over in time; I have fulfilled my Kantian duty. The means I chose to do it were on other, self-interested, reasons. But how can the good Kantian yet complain? The good will still got me on the Tube at Warren Street, even though I was thinking of saving money rather than my Kantian duty when I did. This is because my overall reason for doing so was as a means of fulfilling my Kantian duty. But if I can take in self-interested reasons en route to Waterloo, why can I not introduce self-interested reasons to support the hospital or keep the speed limit? Ah, but you say, the whole point and defect of such indirect strategies is because they interpose other reasons, displacing the end. But it is not quite clear to me what this objection means. So I shall split it into separate aspects. If the problem is that I am thinking of something else as well as my thought of doing good, then we would have had this problem right at the start of the sequence. There was no reason why I should have been thinking of Julia while I board at Warren Street. So, perhaps, the problem is instead the interposition of another agent. It is that my action depends upon their instructions (‘pay your taxes!’; ‘watch the camera!’) rather than my own free decision. However, if that is the problem, we can meet it by changing the example again. Here I am in UCL, having to deliver the ticket to Julia at Waterloo. But, being a newcomer, I am pretty vague about London

The Moral is: States Make Laws

171

topography. So I immediately consult my colleague, Charles in the quad. ‘I have to be at Waterloo at 2’, I say, ‘but I don’t know how to get there’. ‘No problem’, he says. ‘I’m going that way myself; come with me’. I do so, and next I know I am boarding the Tube at Warren Street. Now why is this not still acting from the moral good will, even though I am only doing it because Charles tells me? After all, I still get to Julia on time, which is my overall purpose. Indeed, taking Charles in order to do it does not seem different in principle to taking the Tube. In both cases I fasten on a particular means because I believe (in both cases correctly) that it will fulfil my moral end. So what is wrong with the tax and car cases cannot just be that I act on other’s instructions. Ah, but you say, the point is that when the state intervenes and supplies sanctions, it destroys my good will by forcing me to do something for the wrong reasons. I may not understand the Tube but all the time I am with Charles I can ask him about it and so understand at every point why I am doing what he tells me. I can also at any point decide not to take his advice and get off. But things would be different, for example, if he dragged me on to it, or threatened, Special Branch style, to shoot me if I left. We need another stop along this journey, another example change. Suppose, as before, I have to take the Tube at Warren Street to fulfil my Kantian duty. The only trouble (now) is that, since I saw a suicide at a Tube station last week, I have a deep psychological aversion to going anywhere near another one. (The form of the trolley problem that consists in an inability to get on the trolley in the first place.) Again, anxiously, I consult Charles in the quad. ‘No problem, son,’ he says, kindly taking my hand. And the next I know I am blinking myself fully awake before a smiling Julia and handing over her ticket in time, as my Kantian duty prescribes. What has happened? Well, implausible as it may be, Charles adds instant hypnotism to his other skills; and, considerably more implausibly, he is even able to practise this skill on the Northern Line. So all the way to Waterloo, I thought that I was gliding on a boat, rocking gently and half asleep, on my way to dock at Waterloo pier. Charles was forcing me. But he did deliver me to the barrier on time, and I wanted Charles to do whatever was required to achieve this; even though I also very definitely did not want to go on the Tube, and this was in fact what Charles forced me to do. Kantian good will? Why not? Yet if I could employ Charles to do this, why does my will fail if I similarly employ the state to force me, particularly if my reason for doing so is an overall good purpose? Ah but, you say, all the way on my dreamy ride I knew that I was going to Waterloo. I never forgot that I was fulfilling my overall moral project. Yet, you say, the point of the indirect strategies, either of self-or state-manipulation, is that the end disappears from cognition; and it is this that renders the will impure. Okay. We clearly need a further implausible step, or station, on this Underground argument. We have to make it that the end can only be reached by forgetting that it is the end; only by forgetting my destination can I safely arrive.

172

Ross Harrison

So here goes; final stop. The reason why I have a complete aversion to Tube stations is that the suicide I witnessed was at Waterloo. Hence I cannot even think of Waterloo without completely losing it. In these circumstances, even more anxious than before, I stagger up to Charles. ‘I have to deliver this ticket at the Eurostar barrier’, I gasp, and then add, as the quad begins to swing, ‘The trouble is that Eurostar . . .’ ‘Don’t worry, son’, Charles interrupts smoothly, taking my hand and sedating me before the destructive name has passed my mind or lips. And the next I know is that I am blinking myself awake, looking for the smiling Julia, and handing over her ticket on time. I have fulfilled my great Kantian duty; I have met her at the barrier. One final time, threading all manipulation by self and others, good will has won me through to my goal. It would, of course, have been more romantic to have met under the clock. But there is more to life than romance. Such as good will and Kantian duty. That’s where the morality is, with the fulfilled contract at the Eurostar barrier. And the moral is that contracts are made by law, and law is made by states.

10 The Attack on Liberalism Mark R Reiff

Liberalism is today under attack. This attack is being fought along two fronts, and so appears to be coming from different directions, but it is actually coming exclusively from the right. One source is Islamic fundamentalism, and the other is American neo-conservatism, which in turn unites elements of Christian fundamentalism with elements of neo-Platonic political philosophy and neo-Aristotelian moral theory. Both Islamic fundamentalism and American neo-conservatism are perfectionist views, and while perfectionist attacks on liberalism are nothing new, there is a special danger in the position that liberalism currently finds itself. Although bitterly opposed to each other and divided by utterly incompatible substantive views, these two sources of attack have many similarities at a more fundamental level of political theory. To varying extents, both reject the idea that the state should be neutral toward and tolerant of competing conceptions of the good, both oppose the strict separation of religious and political authority, both place higher priority on authority and security than on negative liberty and the rule of law, both attach greater weight to faith than to reason in their pure and practical reasoning, both view the community rather than the individual as the fundamental social unit, both take a Machiavellian approach to public discourse and debate, and both have a very different attitude toward the role that equality should play in organizing and regulating our communal life. The special danger of this attack is that each source may use the existence of the other to increase the force of its own denunciation of liberalism, and the voices of those who would defend liberalism may be drowned out in the din of the invective hurled by one of its attackers against the other. The purpose of this paper is to focus our attention on the fundamental differences between liberalism and perfectionism, no matter what its substantive content, and to defend liberalism against one of the most serious accusations that perfectionists make against it—the claim that liberalism leads to nihilism. This is an exercise in what I call politectonics—the study of how families of political theories clash when we abstract out to their basic themes and fundamental presuppositions. I will begin by identifying these basic themes and fundamental presuppositions, and describe some of the general criticisms that perfectionists

The Attack on Liberalism. Mark R Reiff. © Oxford University Press 2007. Published 2007 by Oxford University Press.

174

Mark R Reiff

raise in order to put their claim that liberalism leads to nihilism in its proper context. Next, I will try to give some content to this claim, for it is often unclear exactly what nihilism is supposed to entail. I will then examine how liberalism is supposed to lead to nihilism, for there are two different strands of this particular perfectionist criticism, each arising out of different and in fact mutually inconsistent initial assumptions. First, I will attempt to rebut each strand of this attack, ignoring for the moment the strategy of using one to rebut the other. But then I will assume that one of these strands of attack is true, and explore whether we really have as much to fear from nihilism as both the opponents of liberalism and its defenders tend to believe. Finally, I will assume that nihilism is both coherent and unattractive, but challenge the assumption that perfectionism can produce anything better, for even if liberalism leads to nihilism, perfectionism leads to fanaticism, and fanaticism presents a far greater threat to international peace and individual well-being than nihilism ever could.

I. Liberalism and Perfectionism In common speech, the word ‘liberal’ is often used as a shorthand way of referring to a set of substantive political positions that are typically associated with the moderate left. Used in this sense, the word ‘liberalism’ refers to any political theory or programme dedicated to the elucidation and promotion of these particular leftish concerns. But that is not how I will be using the word. I will be using the word ‘liberalism’ to refer to a collection of fundamental presuppositions or concepts that provide the background constraints within which a certain kind of political life can take place. Liberalism in this sense encompasses many different substantive and often incompatible doctrines of political morality. Liberal egalitarianism, libertarianism, and traditional conservatism are all views that can be accommodated within liberalism. One can be a liberal and be for or against greater redistribution of income, for or against greater government regulation of the market, for or against free trade with other nation-states, and on either side of any number of hotly contested social, domestic, and foreign policy issues of the day. What unites the adherents of these opposing substantive policy positions is that at a higher level of abstraction they all share certain concerns regarding the nature of political life, and they express these concerns as conceptions of the same fundamental presuppositions. Because these presuppositions are concepts, not conceptions, however, they will not generate specific recommendations for either action or belief without further specification of how the concept should be understood. This is why liberals can come to completely opposing positions on what should be done about so many questions regarding the organization and regulation of our social life, yet still be adherents of one of the family of political theories that fall within liberalism.

The Attack on Liberalism

175

Even at this higher level of abstraction, however, identifying all the fundamental presuppositions that unite the various views that fall within the family of liberalism is difficult and controversial. Those who consider themselves liberals do not agree on what should be included on this list, and neither do their critics. Any attempt to come up with a definitive description of the fundamental presuppositions of liberalism is accordingly bound to generate as much heat as it does light. But this does not mean that it is pointless to try to articulate these shared presuppositions. What we gain by trying to get a clearer picture of the nature of liberalism, regardless of whether this picture is totally accurate or complete, is a better understanding of why liberalism is under attack, and a richer sense of what distinguishes liberalism from the various forms of perfectionism that are attempting to displace it. For it is important to realize that what distinguishes liberalism and perfectionism is not that each generates different answers to questions about how to organize and regulate our social life—whether they do this cannot be determined until we give content to the particular liberal and perfectionist visions we wish to compare. What distinguishes liberalism and perfectionism is that each asks different questions—each begins with different moral presuppositions, and if we change the nature of these presuppositions, we change the whole nature of moral discourse, of what is subject to moral evaluation, and what moral evaluation even means. And when we begin with different questions, we are likely to find ourselves unable to find a common language for debate, sinking instead into the kind of irresolvable tectonic battle that can only arise when people are talking past one another and unable to see what the other is even on about. This, I fear, is the kind of battle in which we now find ourselves, and why it is so important that we try to illuminate the fundamental differences between liberalism and perfectionism despite the controversy that any such attempt will necessarily provoke. Before I embark on this comparison of liberalism and perfectionism, however, I want to clarify the precise nature of the claims I am about to make. While I contend that liberals and perfectionists embrace different sets of fundamental presuppositions, I am not suggesting that a society must be committed to either one set or the other in its entirety to be either liberal or perfectionist. Liberalism and perfectionism can exist in degrees, and even societies that embrace all the presuppositions of liberalism or perfectionism in theory may be more or less faithful to them in practice. The various presuppositions I will identify are therefore not intended to be a set of necessary and sufficient conditions for establishing whether a particular political community should be labelled liberal or perfectionist, but as a way of judging how liberal or how perfectionist a particular society—or the views of a particular political theorist—may be. Indeed, one of the advantages of developing a thicker description of the differences between liberalism and perfectionism than is usually attempted is that this will provide us with a way of explaining how mixed conceptions of each view differ from each other and from conceptions that are more pure.

176

Mark R Reiff

In a paper of this length, however, my discussion of each fundamental presupposition will by necessity be brief. Because these fundamental presuppositions are not a set of necessary and sufficient conditions, some liberals and perfectionists will disagree with the presuppositions I assign to them. I will not be able to highlight all these disagreements here, or set forth a full defence of my choices, or describe all the different ways each presupposition could be expressed, understood, or defended. All I will be able to do is sketch out a general framework within which to place the debate between liberalism and perfectionism, and convey a sense of the range of issues that a full defence of the attack on liberalism would have to address. I take the claim that liberalism leads to nihilism to be an important part of that attack, but only part. A full defence of liberalism would provide a moral argument in favour of each liberal presupposition, a critique of the arguments made against it, and an argument against the view that the perfectionist version of the presupposition is correct. The defence of liberalism I have mounted in this paper is accordingly only partial, not complete. I should also note that the kind of perfectionism I have in mind in everything that follows is hard perfectionism, the kind of perfectionism for which one can find theoretical underpinnings in the work of people like Plato, Aristotle, Machiavelli, Joseph de Maistre, Nietzsche, Carl Schmitt, and Leo Strauss.¹ Hard perfectionists are committed to some specific comprehensive set of substantive moral values, and believe that we should organize our political and social life in such a way as to ensure that we live in a community composed exclusively of like-minded individuals. Their perfectionism is accordingly not merely a moral perfectionism—a vision of the ideal life for a person, but a political perfectionism, a vision of an ideal society, where state power is used to ensure the creation and proliferation of a certain type of ideal person. Hard perfectionism is therefore a certain kind of teleological theory, one that defines the good by reference to particular view of communal excellence and makes achievement of this good the central goal of political life.² Hard perfectionists of one stripe or another hold a great deal of political power in the world today, but while the academic proponents of such views are well represented in intellectual history, they are currently underrepresented in the academy. As a result, there have been few contemporary attempts to engage the hard ¹ Some may find the inclusion of Machiavelli on this list surprising. I have done so because one can interpret his work not merely as the ultimate realist tract, an amoral elevation of means over ends, but as an expression of admiration for a particular set of ends, the ends exemplified by ancient Rome. See I Berlin, ‘The Originality of Machiavelli’ in Against the Current (Princeton, 2001) 25–79. In any event, I mention him not because he is necessarily a hard perfectionist himself, but because his approach to politics has had such influence on the hard perfectionists who came after him. ² My definition is similar but probably broader than that employed by Rawls, for Rawls may not have intended to include conceptions of perfection that are theologically based within his definition, but only those that are based on theories of human nature or culture. See J Rawls, A Theory of Justice (rev edn, Cambridge, MA, 1999) 22. Under my definition, there is no such limitation, and perfectionism can be (and often is) theologically based.

The Attack on Liberalism

177

perfectionists critically. Instead, most of the contemporary academic debate has focused on the development and critique of soft perfectionism, a view advocated by Joseph Raz, Steven Wall, Thomas Hurka, George Sher, the so-called communitarians, and many others. Soft perfectionists reject some but embrace other liberal presuppositions, or at least claim they do. Their commitment to the use of state power to create and to maintain their particular vision of an ideal society therefore purports to be weaker than that of hard perfectionists, but stronger than that of liberals. Whether soft perfectionism is a distinct, coherent political theory and whether it has the advantages over liberalism and hard perfectionism that soft perfectionists contend it does are important questions, but I will not attempt to explore these questions or develop or defend a critique soft perfectionism here. Nothing I will say in this paper depends on foreclosing the possibility that soft perfectionism presents a viable alternative to both hard perfectionism and liberalism. It is the views of the hard perfectionists—those who reject the liberal presuppositions that even the soft perfectionists claim to accept—that I take to be most dangerous and that this paper is intended to engage. What these liberal presuppositions are, and why perfectionists reject them, are the questions we turn to next.

A. Toleration One of the principal themes running throughout liberalism is toleration—the belief that there are a wide range of reasonable although incompatible comprehensive moral doctrines, conceptions of the good, and plans of life, and that people ought to be free to pursue whichever of these they prefer, regardless of whether the path they choose is in their best interests or to the liking of their fellow citizens. To understand what liberals mean by toleration, however, it is important to recognize that toleration is intended to be more than a pragmatic response to the problems that arise when people disagree over what morality requires. For liberals, toleration is supposed to be a moral imperative, not merely a pragmatic one, and therefore entails holding the following combination of attitudes. It requires that we view some attitude or conduct as worthy of suppression—that is, wrong in some important way and detrimental to some significant portion of society; that we believe this offensive attitude or conduct could be effectively suppressed—that is, any attempt at suppression would be neither pointless nor counterproductive; and yet believe that we should resist this impulse to suppress but not abandon it. Toleration accordingly tries to mark out some territory between a willingness to accept and a desire to suppress that is not occupied by indifference. But perfectionists argue there is no territory here to be claimed. They deny that there are non-instrumental reasons for resisting a desire to suppress that are not also reasons for accepting that to which we object. And they contend that if no such reasons actually exist, then toleration is not a distinct moral attitude that anyone could actually hold, but simply another way of saying that a plurality of

178

Mark R Reiff

incompatible sets of moral judgements are equally correct or, more nefariously, that no set of moral judgements is correct. In the former case, the liberal commitment to toleration reduces to value relativism; in the latter, to value nihilism; but in either case to positions that perfectionists vigorously reject.³ Perfectionists also deny there is even a pragmatic, consequentialist case to be made in favour of toleration. Rather than being a modus vivendi, allowing groups that hold incompatible comprehensive moral doctrines to coexist within a single community, perfectionists see toleration as a recipe for self-destruction. While no liberal contends that government’s commitment to toleration should be unlimited—as Rawls notes, ‘justice does not require that men stand idly by while others destroy the basis of their existence’, liberals do contend that in a well-ordered society, even the intolerant should be tolerated unless they pose a significant threat to the institutions of liberty.⁴ But perfectionists argue that the threat posed by the intolerant is almost never clear until it is too late, and therefore government cannot afford to employ suppression only as a last resort. Moreover, a liberal society is likely to be composed of such a diversity of views that it will be difficult for it to deal decisively with such threats even after they become apparent. A liberal society is accordingly going to be slow to react, and may never be able to muster the necessary unanimity and resolve to deal forcefully and effectively with those who threaten even its most cherished values. This makes toleration not a guarantee of social peace and stability, but an open invitation to evil, one that threatens the long-term sustainability of a community’s preferred way of life.

B. Neutrality Another principal theme of liberalism is the belief that government should remain neutral in some meaningful sense between various competing comprehensive moral doctrines, conceptions of the good, and plans of life. Indeed, many perfectionists view neutrality as the commitment from which all other liberal commitments are derived.⁵ But it is important to note that while liberalism’s commitment to neutrality complements its commitment to toleration, it is conceptually distinct from it. Toleration does not entail neutrality—government could tolerate a number of viewpoints while still promoting whatever one it deems best.⁶ Similarly, neutrality does not entail toleration. One could be neutral between various forms of life simply because one found them equally agreeable. ³ For some recent discussions of these arguments, and possible responses to them, see P Jones, ‘Making Sense of Political Toleration’ (2007) 37 British Journal of Political Science 383; G Newey, Virtue, Reason, and Toleration (Edinburgh, 1999); D Heyd (ed), Toleration: An Elusive Virtue, (Princeton, 1996); S Mendus, Toleration and the Limits of Liberalism (London, 1989). ⁴ See Rawls n 2 above, 192. ⁵ See, eg G Sher, Beyond Neutrality: Perfectionism and Politics (Cambridge, 1997). ⁶ Indeed, many soft perfectionists advocate exactly this—toleration without neutrality. See, eg V Haskar, Equality, Liberty, and Perfectionism (Oxford, 1979) 290.

The Attack on Liberalism

179

Because none of these forms of life were in any way objectionable, there would be no need for toleration. Regardless of whether neutrality is the mother of toleration or conceptually distinct from it, however, it has attracted a great deal of perfectionist ire. Perfectionists argue that in light of its commitment to neutrality, liberalism is empty. ‘There exists absolutely no liberal politics’, Carl Schmitt complains, ‘only a liberal critique of politics’.⁷ Rather than offer a positive moral programme for society, all liberalism offers is permissiveness. But if everything is permitted, perfectionists ask, how are we to decide what morality requires us to do? By failing to provide guidance on this question, liberalism leaves society free to drift into moral relativism, subjectivism, and ultimately nihilism.⁸ Liberals believe that the combination of tolerance and neutrality will produce value pluralism, not value nihilism. But values have to come from somewhere before there can be a plurality of them. While liberals rely on the family, on religious and educational institutions, and on culture to provide the necessary moral guidance to society, perfectionists contend that the requisite set of values cannot be generated by non-governmental sources alone. Perfectionists, however, would not just have government support private value-generating institutions, they would make government the arbiter of moral values, for they believe it is essential that government put forward a positive substantive moral programme if citizens are to develop the moral attitudes and virtues necessary for the creation of an ideal society. Private institutions would accordingly not be free to promote whatever values they saw fit, they would be free (indeed required) to promote only those values with which the government happened to agree. What values these would be would depend on the particular perfectionist vision the government happened to embrace—the sets of values promoted by an Islamic fundamentalist society, an American neo-conservative society, and a Marxist society would of course all differ radically. But at the tectonic level, the content of these values is not important. What is important is that perfectionists believe liberalism’s commitment to neutrality represents an abdication of government’s inherent responsibility to be the moral leader of society.

C. Liberty and Authority Yet another theme of liberalism is its concern for individual liberty. As a result of this concern, liberals oppose granting unlimited authority to government, although the degree to which government authority should be limited is very controversial within liberalism. Some liberals—libertarians and traditional conservatives, for example—advocate very strict limits on government power, while ⁷ C Schmitt, The Concept of the Political, (George Schwab (trans), Chicago, 1996) 70. ⁸ See, eg L Strauss, Natural Right and History (Chicago, 1953) 4–5; S Rosen, Nihilism (New Haven, 1969) xiii.

180

Mark R Reiff

liberal egalitarians believe that government should and indeed must be allowed to regulate a wide variety of human activities. But even those who advocate a generous role for government regulation believe that such a role has to be justified, for all liberals attach great importance to negative liberty—the extent to which our ability to do that which we have the capacity to do is free from interference by other human agents.⁹ Whether such interference is justified depends on whether the particular liberty involved is absolutely protected against interference, and if not, whether such interference generates a sufficiently large compensating moral benefit. Liberals may disagree as to the circumstances in which balancing is appropriate and the relative weights to be assigned to the various moral costs and benefits when it is, but they all agree that no interference can take place without considering these questions. Perfectionists, in contrast, are more concerned about limitations on government authority than they are about limitations on individual liberty. These concerns may be expressed in a variety of ways, but in the American context, they have surfaced as a series of aggressive arguments for a broad interpretation of the constitutional scope of executive power. First, there is the argument for the unitary executive, which claims that the President has absolute authority over all aspects of the executive branch, and that Congress is accordingly without power to create ‘independent’ administrative agencies or to require that members of these agencies testify before Congress free from centralized political control.¹⁰ Second, there is the argument for what is in effect executive supremacy, which claims that the President has unreviewable authority to take a certain kind of action, or rather to take any action that is motivated by a certain kind of reason, such as the protection of national security, a conveniently elastic justification that can seemingly be applied to an enormous range of executive actions.¹¹ Third, there is the argument for executive immunity, which seeks to increase the scope of the executive’s unreviewable authority even further by imposing limits on the oversight power of the courts and stripping them of various kinds of jurisdiction.¹² Indeed, neo-conservatives tend to view the role of the judiciary as exceedingly ⁹ See, eg I Berlin, ‘Two Concepts of Liberty’ in I Berlin, Liberty (H Hardy (ed), Oxford, 2002) 166–217. ¹⁰ For a discussion of the argument for the unitary executive, see SG Calabresi and SB Prakash, ‘The President’s Power to Execute the Laws’ ( 1994–5) 104 Yale Law Journal 541. ¹¹ For example, while a deputy assistant attorney general in the Office of the Legal Counsel, Berkeley law professor John Yoo argued that ‘in the exercise of his plenary power to use military force, the President’s decisions are for him alone and are unreviewable.’ See KJ Greenberg and JL Dratel (eds), The Torture Papers (Cambridge, 2005) 24 n 32. And indeed, President Bush has used this alleged grant of constitutional authority to ignore a huge swath of laws, including statutes prohibiting torture, secret detention, and warrantless surveillance, statutes protecting ‘whistleblowers’ and requiring the provision of various kinds of information to Congress, and even environmental and affirmative action laws, whenever he deems this necessary for the protection of national security. See C Savage, ‘Bush Challenges Hundreds of Laws: President Cites Power of His Office’ Boston Globe (30 April 2006). ¹² For a discussion of the relationship between the argument for the unitary executive and the argument for Congressional authority to enact jurisdiction-stripping legislation, see SG Calabresi

The Attack on Liberalism

181

narrow—the courts are there primarily for resolving disputes between the legislative and the executive, not to protect citizens from them both. According to Richard Posner, for example, as long as these two branches of government agree on a proposed course of action (even when both branches are controlled by the same party), the courts should generally just sit back and allow elected officials to proceed as they see fit.¹³ While neo-conservatives offer textual arguments in support of these various claims, their underlying political perfectionism is more likely to be what is driving their concerns. For perfectionists, limitations on authority are of great concern because the role of government is not merely to maintain the community’s physical integrity, but to enforce its values, and it takes far greater authority to accomplish this than to patrol the community’s streets and borders. While the enforcement of these values may be an infringement of negative liberty, this is not morally significant to perfectionists because the liberty to reject the values they wish to inculcate is not a liberty they believe people should enjoy. Rather than a conception of negative liberty, perfectionists effectively if not expressly embrace a conception of positive liberty—the view that people can and should be forced to embrace a prescribed set of values because only by doing so can they realize their full potential, and only through such self-realization can they truly be free.¹⁴ Granting government greatly expanded or even absolute authority therefore has no moral costs for perfectionists, only moral benefits, at least as long as that authority is exercised in a fashion that is consistent with the values that they believe government is obligated to promote. It is therefore not the degree of government authority that perfectionists find morally significant, but its objectives and effects.

D. Security and the Rule of Law Yet another theme of liberalism is that no one should be regarded as above the law, and that exceptions to the rule of law cannot be justified simply by asserting that such exceptions are required by national security. This does not mean, however, that liberals believe we should allow our own legal processes to be used to undermine the founding ideals of our political community. Just as there are limits to toleration, there are limits to the degree to which liberals will place compliance with formal legality over the need to defend society from outright attack.

and KH Rhodes, ‘The Structural Constitution: Unitary Executive, Plural Judiciary’ (1992) 105 Harvard Law Review 1153. ¹³ See R Posner, Not a Suicide Pact: The Constitution in a Time of National Emergency (Oxford, 2006) 10. ¹⁴ For an explication and defence of positive liberty, see, eg C Taylor, ‘What’s Wrong with Negative Liberty’ in Philosophy and the Human Sciences: Philosophical Papers Volume 2 (Cambridge, 1985) 211–29.

182

Mark R Reiff

A constitution is not a suicide pact.¹⁵ Perfectionists are accordingly not the only ones prepared to recognize ‘a non-legal “law of necessity” that would furnish a moral and political but not legal justification for acting in contravention of the Constitution [and would] trump constitutional rights in extreme situations.’¹⁶ But perfectionists are much more willing to find that this necessity test has been met, and much more likely to adopt broader exceptions to established principles of legality when it has.¹⁷ There are several reasons for this. First, as a practical matter, perfectionists tend to see themselves as surrounded by enemies, partly because identifying one’s enemies is one of the ways that a perfectionist community defines itself, and partly because perfectionist communities are more easily distinguished from each other and from liberal communities because of their embrace of a specific set of substantive values. Because perfectionists are more conscious than liberals of having enemies, security concerns are simply more salient for perfectionists than they are for liberals. Indeed, to perfectionists, liberals seem maddeningly oblivious to the prevalence of evil in the world and the threat such evil poses to their own community. And if liberals cannot even see who their enemies are, how can they be relied on to take timely and effective defensive action?¹⁸ Second, perfectionists view law as the servant of authority not its master. In a liberal society, the rule of law is sovereign, but in a perfectionist society the sovereign is ‘he who decides on the exception.’¹⁹ While there are often instrumental reasons for those in power to obey the law, the primary moral obligation of government in a perfectionist society is to maintain the physical and cultural integrity of its political community and keep it pure. The exercise of power in furtherance of such objectives is therefore by definition both morally right and necessary, even when this is technically in violation of the law. For perfectionists, maintaining security is an overpowering moral notion that is to be broadly construed, and the rule of law must give way, even when they are the only ones able to see that this is what maintaining security demands. Finally, a preoccupation with security among the masses enhances government authority, and this is instrumentally useful in a perfectionist society because greater authority means greater ability to enforce whatever perfectionist vision those in power happen to embrace. As Machiavelli advised, ‘it is best to keep men poor and on a permanent war footing, for this will be an antidote to the two great enemies of active obedience—ambition and boredom—and the ruled will ¹⁵ This remark was originally made by Justice Jackson in his dissent in Terminello v City of Chicago 337 US 1, 37 (1949). ¹⁶ Posner, Not a Suicide Pact n 13 above, 12. ¹⁷ For a critique of some of the policies of the Bush administration and an argument that the national security measures it has advocated are far broader than necessary to deal effectively with the threat we actually face, see R Dworkin, ‘The Threat to Patriotism’ New York Review of Books (29 February 2002). For a defence of these measures, see Posner n 13 above. ¹⁸ See, eg J Kekes, ‘The Prevalence of Evil’ in Against Liberalism (Ithaca, NY, 1997) 23–45. ¹⁹ C Schmitt, Political Theology (George Schwab (trans) Chicago, 2005) 5.

The Attack on Liberalism

183

then feel in constant need of great men to lead them.’²⁰ Perfectionists accordingly have incentives to emphasize if not exaggerate potential threats to a community’s security, for this gives them access to tools to suppress internal opposition that would otherwise not be available to them and thereby supports the ultimate realization of their perfectionist programme. In this case, Machiavellian methods of maintaining power and what is conceived of as furthering the good of a perfectionist society coincide.

E. Identification of the Fundamental Social Unit Another theme of liberalism is its identification of the individual as the fundamental social unit. This is not to say that liberals do not recognize the value of community, or that they do not believe that community identity is something that government should foster and support. On the contrary, liberals commonly hold both views. But liberals value community because a sense of community identity is a necessary background condition for individuals to develop their own concept of personal identity and thrive. In other words, community identity and values are cherished because they are instrumental to the formation and realization of individual identity and values. Under liberalism, the community is derivative of the individual not the other way around. It is because they view the individual as the fundamental social unit that liberals value personal autonomy so highly. From their concern for personal autonomy, liberals then derive a whole range of moral limits designed to ensure separation between the public and the private spheres of life.²¹ But while liberalism’s commitment to personal autonomy is important, it is not as foundational as many critics of liberalism believe.²² Liberals do not begin with a concern for autonomy, but with the presupposition that the individual is the fundamental social unit, and it is from this presupposition that their concern for autonomy is derived. Perfectionists, in contrast, begin with the presupposition that the community is the fundamental social unit. Consider, for example, this statement by Sayyid Qutb, a member of the Muslim Brotherhood whose writings provide the intellectual underpinning for some of the most radical elements within Islamic fundamentalism: [The] Muslim community does not denote the name of a land in which Islam resides, nor is it a people whose forefathers lived under the Islamic system at some earlier time. It is

²⁰ See Berlin, ‘The Originality of Machiavelli’ n 1 above, 61. ²¹ This distinction between the public and the private is accordingly political, not analytic. See, eg R Hale, ‘Coercion and Distribution in a Supposedly Non-coercive State’ (1923) 38 Political Science Quarterly 470. ²² See, eg S Wall, Liberalism, Perfectionism, and Restraint (Cambridge, 1998) esp pt 2.

184

Mark R Reiff

the name of a group of people whose manners, ideas and concepts, rules and regulations, values and criteria, are all derived from the Islamic source.²³

For those who hold this view, the community is not an aggregate of individuals; individuals are instantiations of their community, whose collective characteristics are repeated like fractal geometry in every one of its essentially identical individual parts. Communities do not derive their identities from the individuals that make them up; individuals derive their identities from the communities of which they are part. The individual is not prior to the community; the community is prior to the individual. There is no separation between the private and the public—the individual exists and persists only as a part of a communal being, ‘as a river that flows into the ocean does indeed persist in the midst of the waters, but without name or personal identity.’²⁴ Given the presupposition that the community is the fundamental social unit, a community can exist even if it has no members. This is how, for example, Qutb can state that the Muslim community has been extinct for centuries, ‘crushed under the weight of those false laws and customs which are not even remotely related to the Islamic teachings, and which, in spite of all this, calls itself the “world of Islam”.’ According to Qutb, the entire modern world, whether it calls itself Muslim or non-Muslim, is in a state of jahiliyyah, or ignorance of the divine guidance. It has therefore become necessary for a small ‘vanguard’ to initiate a movement of Islamic revival, to lift this veil of ignorance and bring the wayward home to their community, which can then reassert its rightful place as the leader of the world.²⁵ For perfectionists like Qutb, the primary role of morality is accordingly not the protection of the individual and his autonomy, as liberals seem to believe, but the protection of the community and the bonds that hold the community together. There is much more that needs to be said about the nature of these competing views and their respective ramifications, but I will not attempt to do so now. I will merely note that neither view of the fundamental social unit should be understood as an empirical claim. In other words, neither view is based on empirical observations about how individuals see themselves as constituted in the world, although arguments for each view often include such claims.²⁶ Each view is a presupposition, an ordering from which various moral questions can be derived, based on decisions about how one should define the realm within which morality is to operate. The decision to accept one presupposition or the other does not depend on contingent facts about human psychology and self-identification, and therefore is not subject to attack by reference to this kind of empirical evidence. The only kind of empirical evidence that might be relevant to this decision is evidence regarding ²³ S Qutb, Milestones (Beirut, 1980) 11. ²⁴ Joseph de Maistre, Study of Sovereignty, book 1, ch 10, in Oeuvres Complètes de Joseph de Maistre (Paris-Lyon, 1884–93) vol 1, 376, as translated by I Berlin. See I Berlin, ‘Joseph de Maistre and the Origins of Fascism’ in The Crooked Timber of Humanity (Princeton, 1990) 91–174, 126. ²⁵ See Qutb, Milestones n 23 above, 9–22, 101–3. ²⁶ See, eg M Sandel, Liberalism and the Limits of Justice (2nd edn, Cambridge, 1998).

The Attack on Liberalism

185

the consequences of holding each presupposition, for such evidence could be used to construct a moral argument in favour of one and against the other. Whether such an approach would be appropriate, however, is itself a controversial question, one that I will unfortunately have to leave until another time.

F. The Separation of Religious and Political Authority Another theme of liberalism is that it is hostile to the conflation of religious and political authority. The role of religion within liberalism is to provide moral, not political, guidance to the individual. Religious leaders are within their role when they provide instruction on the Word of God and on the moral values that can be found therein, but they trespass on the political if they use their religious authority to instruct the faithful how to vote or otherwise direct them to adopt controversial political positions they have not arrived at independently themselves. And while religious leaders are not disqualified from holding political office within liberalism, if elected or appointed they must not delegate the political power conferred upon them to some external authority, but must exercise their power according to their own conscience and in accordance with the wishes of all those that they have been elected or appointed to represent. Of course, the separation of religious and political authority presupposes that it is possible to separate the private from the public, a presupposition that as we have seen many perfectionists reject. Perfectionists are accordingly much more comfortable with mixing religious and political authority, as long as this is done to promote values with which they happen to agree. They are happy for religious leaders to publicly adopt controversial political positions and to use their religious authority to marshal the faithful in support of these positions, a phenomenon that is becoming more and more common in American politics and is largely responsible for bringing the neo-conservatives to power. It is also, of course, a phenomenon that has been going on for years in the Muslim world, most notably in Afghanistan and Iran. For some perfectionists, this conflation of political and religious authority is simply a matter of expediency, a tool for consolidating their political power, and not a consequence of their personal religious views. For others, however, it is conceptual, the result of a deep commitment to the role of religion in the organization and regulation of political and social life. Strauss, for example, viewed religion instrumentally, as the cement that held a community together.²⁷ Maistre, in contrast, believed there could be no society without a state, no state without sovereignty, no sovereignty without infallibility, and no infallibility without ²⁷ See S Drury, Leo Strauss and the American Right (New York, 1997) 37. This is apparently also the attitude toward religion of Karl Rove, the political mastermind behind George Bush’s rise to power. See J Moore and W Slater, The Architect: Karl Rove and the Master Plan for Absolute Power (New York, 2006).

186

Mark R Reiff

God, from whom all legitimate authority was derived.²⁸ Qutb also saw a conceptual unity between religion and politics, and believed that separating the two had created a ‘hideous schizophrenia’ that left secular society exhausted, with no resources left to guide human life,²⁹ a view that continues to be a central tenet of Islamic fundamentalism and is also prominent among many current members of the Christian right.³⁰ And while some perfectionist (eg communist) political communities are hostile to all forms of religion and religious authority, this is because they see religion as a political competitor rather than as something that can be separated from and exist independently of the direction of political life. Regardless of the form that perfectionism takes, perfectionists accordingly see political and religious authority as inextricably intermixed. If the exercise of religious authority can help move a political community toward perfection, there is no reason why it should not be used to do so. And if it cannot, it need not be tolerated, but should simply be suppressed. In either case, the only moral question presented is whether those with religious authority will support or hinder the perfectionist political agenda that those in power happen to embrace.

G. The Role of Public Discourse and Debate Part of the reason why liberals object to the conflation of religious and political authority is that such a combination distorts the role that liberals assign to public discourse and debate. Liberals believe that all members of a political community should have an opportunity to participate in political decision-making under conditions of full information, and that the purpose of public discourse and debate is to persuade others of the rightness of one’s position by resorting to arguments that one’s opponents could not reasonably reject.³¹ But this process is subverted when political action is simply directed by religious leaders rather than based on an individual’s own assessment of the strength of the arguments presented. The conflation of religious and political authority is therefore objectionable to liberals because it is often used as a tool for cutting off debate, while perfectionists view this as simply another method of ensuring that a political community conforms to their vision of the ideal. Not only do perfectionists often resort to religious authority as a tool for cutting off debate, they also embrace the noble lie—Plato’s idea that the common people are incapable of seeing the truth even when it is laid out before them and ²⁸ See Berlin, ‘Joseph de Maistre and the Origins of Fascism’ n 24 above, 134. ²⁹ See S Qutb, Islam: The Religion of the Future (Kuwait, 1992) 34–60, For useful discussions of Qutb’s concept of the relationship between religious and political life, see S Khatab, The Power of Sovereignty (New York, 2006) 7–47, 119–71, and S Khatab, The Political Thought of Sayyid Qutb (New York, 2006) 164. For the views of Shiite fundamentalists on this issue, see Ayatollah R Khomeini, Islamic Government (New York, 1979), in which the Ayatollah sets forth his conception of the unity of religion and politics and argues for outright clerical rule. ³⁰ See K Armstrong, The Battle for God: A History of Fundamentalism (New York, 2000). ³¹ See, eg TM Scanlon, What We Owe to Each Other (Cambridge, MA, 1998).

The Attack on Liberalism

187

therefore may be manipulated and led by falsehoods whenever necessary to do so. Machiavelli, for example, believed ‘to speak the truth is sensible only when one speaks to wise men’,³² and Nietzsche also embraced ‘the Platonic notion of the noble delusion’.³³ And, of course, a similar sentiment lurks behind Strauss’s theory of esoteric writing—the view that all the work of the great classical political philosophers was written in a kind of code designed to convey the truth to the enlightened few but conceal it from the masses.³⁴ While liberals consider such Machiavellian techniques a violation of the moral limits on public discourse and debate, perfectionists do not believe that the use of such methods presents a moral question. Under perfectionism, effective instruments of policy are by definition morally acceptable, not just because the ends justify the means, but because perfectionists disagree with the view that states of affairs necessarily include the means that were required to produce them. For perfectionists, in a world defined by ends, means are simply not subject to the same kind of a moral evaluation. Perfectionists also tend to reject persuasion as an objective of political interaction altogether. They criticize liberals for engaging in ‘perpetual discussion’ when what is required is decisive action.³⁵ ‘Values can only be asserted or posited by overcoming others, not by reasoning with them’, claims Allan Bloom.³⁶ Rather than view their opponents as individuals to be persuaded, perfectionists tend to view them as carriers of disease, threatening to infect the body politic. One does not deal with carriers of disease by trying to persuade them to get better; one simply does whatever is necessary to isolate and eliminate them from the body, thereby preserving the purity of the community. The only acceptable objective of political interaction for perfectionists is accordingly decisive victory, the kind of victory in which one’s opponents are not just defeated, but dominated and destroyed, no matter what the stakes. There is no need for intellectual engagement and debate, for this might lead to compromise, and compromise is not an option. The only choice is conquest or capitulation.

H. The Relative Priority of Reason and Faith in Pure and Practical Reasoning Whether it is a consequence of their insistence on strict separation between religious and political authority or a cause, liberals elevate reason over faith in their pure and practical reasoning, at least with regard to matters that are political, ³² See L Strauss, Thoughts on Machiavelli (Chicago, 1958) 34. ³³ See Strauss, Natural Right and History n 8 above, 26. ³⁴ See L Strauss, Persecution and the Art of Writing (Chicago, 1952) 22–37. For an extended discussion of Strauss’s theory of esoteric writing, see S Drury, The Political Ideas of Leo Strauss (New York, 1988), ch 2. ³⁵ See, eg Schmitt, The Concept of the Political n 7 above, 71–2, and Schmitt, Political Theology n 19 above, 63. ³⁶ A Bloom, The Closing of the American Mind (New York, 1978) 202.

188

Mark R Reiff

not personal. In part this is because of their attitude toward public discourse and debate, for the realm of reason is often the only domain in which those of different faiths can meet and find a common basis for discussion and agreement. But it is also a consequence of the overall approach liberals take to understanding their experience of the world. Liberals search for rational, natural explanations for all phenomena, social as well as natural, and even when this search ends in uncertainty they tend to resist the lure of supernatural explanations and rely on probabilistic reasoning to arrive at decisions about which action or belief is most appropriate to embrace.³⁷ Given their approach to pure and practical reasoning and their commitment to debating political issues under conditions of full information, liberals are committed to public education, and believe that public education should be secular, free from both religious influence and control. Perfectionists, in contrast, distrust liberalism’s commitment to reason and the scientific method.³⁸ ‘All intellectuals are bad’, Maistre tells us, ‘but the most dangerous are the natural scientists’.³⁹ Perfectionists see the scientific method and its corresponding exaltation of reason as a threat to authority, and therefore to stability and order. But more importantly, perfectionists see reason as claiming a unique relationship to truth that is undeserved. To Maistre, ‘reason is in reality the feeblest of instruments, incapable alike of altering the behaviour of men or explaining its causes.’⁴⁰ Faith is the great motivator, and therefore the better source of knowledge about the world. And faith is not subject to rebuttal by experience, so the mere fact that there is no evidence to support a faith-based belief, or overwhelming evidence against it, is not a decisive reason for concluding that it is not true, a view that other perfectionists also seem to hold. Indeed, Berlin’s assessment of Maistre could apply with equal force to the approach taken by many neo-conservatives today: Maistre is a dogmatic thinker whose ultimate principles and premises nothing can shake, and whose considerable ingenuity and intellectual power are devoted to making the facts fit his preconceived notions, not to developing concepts which fit newly discovered, or newly visualized, facts. He is like a lawyer arguing a brief: the conclusion is foregone—he knows he must arrive at it somehow, for he is convinced of its truth, no matter what he might learn or encounter.⁴¹

³⁷ See, eg J Waldron, ‘The Theoretical Foundations of Liberalism’ (1987) 37 The Philosophical Quarterly 127, 134–5. ³⁸ For a history of perfectionist hostility toward science and the scientific method, see S Holmes, The Anatomy of Antiliberalism, (Cambridge, MA., 1993) esp 23–5 and 247–52. ³⁹ See Berlin, ‘Joseph de Maistre and the Origins of Fascism’ n 24 above, 114–120. ⁴⁰ Berlin, ‘Joseph de Maistre and the Origins of Fascism’ n 24 above, 122. Maistre, of course, is simply part of a long tradition of religious hostility to reason that has many adherents, including among the contemporary Christian right. See generally M Goldberg, Kingdom Coming: The Rise of Christian Nationalism (New York, 2006) esp pp 86 and 127. ⁴¹ Berlin, ‘Joseph de Maistre and the Origins of Fascism’ n 24 above, 162.

The Attack on Liberalism

189

For perfectionists, faith reveals truth, or at least all important truth, and reason merely serves to confirm or to disguise it.⁴² Note, however, that relying on faith rather than reason to provide decisive reasons for action and belief does not necessarily mean relying on religious faith. For perfectionists, government is a form of religion, and political faith no different than any other.⁴³ While the faith on which perfectionists rely often has a religious source, it may also be an expression of some non-religious myth, such as the infallibility of the governing elite, or the natural superiority of their race, gender, ideas, nation, culture, or ‘the market’. In either case, one simply believes and then follows this belief until it is fulfilled. Lack of success is not a sign of error but of lack of commitment. One cannot be wrong, merely halfhearted or incompetent, an explanation that is becoming increasingly popular among neo-conservatives for what has gone wrong with the American intervention in Iraq, but which they treated with barely disguised contempt when it was offered by those attempting to defend communist perfectionism from neo-conservative attack.

I. Equality The final theme of liberalism that I wish to highlight is its commitment to equality. Liberals are often accused of being insufficiently committed to economic equality by perfectionists on the left, but overly preoccupied with economic equality by perfectionists on the right, an inconsistency that many liberals take as evidence that they have got their commitment to economic equality just right. But a commitment to economic equality is not, in fact, one of the fundamental presuppositions of liberalism. Rather, as I have argued elsewhere, the limited ability of liberal societies to produce greater economic equality is an effect of certain other presuppositions or, more precisely, the product of these presuppositions and certain inherent features of human nature.⁴⁴ At the tectonic level, the difference between liberalism and perfectionism does not lie in their respective commitments to economic equality, but in their respective commitments to moral and therefore political equality. ⁴² For example, the Bush administration has consistently been willing to disregard and even suppress scientific evidence that runs counter to what its neo-conservative members believe. See C Mooney, The Republican War on Science (New York, 2005); Union of Concerned Scientists, ‘Scientific Integrity in Policymaking: An investigation into the Bush Administration’s Misuse of Science’ (March 2004) and ‘Scientific Integrity in Policymaking: Further Investigation into the Bush Administration’s Misuse of Science’ (July 2004). And neo-conservative hostility to reason is not limited to topics within the natural sciences—it applies with equal fervour to topics within the evaluative or social sciences as well. See, for example, the discussion of the history of neo-conservative disdain for the intelligence community’s assessments of the dangers posed by the Soviet Union and Iraq in S Hersh, ‘Selective Intelligence’ The New Yorker (12 May 2003). ⁴³ See Berlin, ‘Joseph de Maistre and the Origins of Fascism’ n 24 above, 125–6 (quoting Maistre), and C Schmitt, The Crisis of Parliamentary Democracy (Ellen Kennedy (trans), Chicago, 1988) 76 (quoting Mussolini). ⁴⁴ See MR Reiff, ‘The Politics of Masochism’ (2003) 46 Inquiry 29.

190

Mark R Reiff

There are two spheres within which these commitments operate: among members of one’s own political community; and between members of one’s own political community and outsiders. Liberals believe that all people have equal intrinsic moral worth, regardless of the political community to which they happen to belong. Liberals may nevertheless give the interests of members of their own political community greater weight in their moral reasoning, but this is because they find all-things-considered reasons for doing so, not because they see the interests of outsiders as intrinsically less valuable. Perfectionists, in contrast, do not believe that all people have equal intrinsic moral worth. Perfectionists believe that anyone who is not a member of their political community—regardless of whether this community is defined along racial, religious, national, or ethnic lines—is of lesser moral worth, like an animal or artifact, or perhaps of no moral worth at all.⁴⁵ Perfectionists may find instrumental reasons for giving the interests of outsiders equal moral weight in their moral reasoning, but they see no intrinsic reason for doing so. Within their own political community, liberals believe that government must treat everyone’s interests with equal concern and respect.⁴⁶ Exactly what this means is very controversial within liberalism, but no matter how it is cashed out, perfectionists are against it—they reject the idea that the interests of every individual are to be given equal weight in deciding what is right to do.⁴⁷ Of course, given their belief that the community rather than the individual is the fundamental social unit, perfectionists believe that the interests of the community (defined as something other than simply the sum of the interests of its members) always trump the interests of the individual. But even when a conflict involves just the interests of individuals, these interests need not be treated equally. Given variations in human temperaments, talents, and abilities, some will naturally have a greater capacity to achieve perfection than others. Those who are perceived as being members of this elite group will accordingly be entitled to special privileges, for it is by encouraging the development of such individuals that the community moves itself closer to perfection.⁴⁸ There are perhaps other presuppositions that could be added to the list of those on which liberalism and perfectionism differ, and perhaps some of those I have included could be derived from others on the list and therefore should not be set forth as fundamental presuppositions in their own right. I do not contend that my list is necessarily exhaustive, or that it is reduced to only its most elemental ⁴⁵ Aristotle, for example, thought that some men were natural slaves, ‘intended by nature to be governed’. See Aristotle, The Politics (S Everson (ed), Cambridge, 1988) 1256b 22–26. ⁴⁶ See, eg R Dworkin, ‘Liberalism’ in S Hampshire (ed), Public and Private Morality (Cambridge, 1978). ⁴⁷ See, eg A MacIntyre, Whose Justice? Which Rationality? (Notre Dame, 1988) 98. ⁴⁸ While soft perfectionists often try to build egalitarian assumptions into their perfectionist vision and thereby avoid the charge of elitism, hard perfectionists do not shy away from the charge of elitism—they embrace it. See R Arneson, ‘Perfectionism and Politics’ (2000) 111 Ethics 37, 39–42.

The Attack on Liberalism

191

components. But we do now have a basis for comparison, and a sense of how liberalism and perfectionism differ at the tectonic level, and this should be all that is required to place the claim that liberalism leads to nihilism in its proper context. And now that this is done, we can move on to a more detailed examination of that claim.

II. Liberalism and Nihilism If we were to summarize the specific complaints that perfectionists level at the various liberal presuppositions we have examined, we might say that the perfectionist attack on liberalism has five general themes. First, liberalism is weak and indecisive, unable to recognize or deal effectively with evil, and therefore incapable of maintaining internal stability and order and, even more importantly, of securing the physical and moral integrity of the community from external threats. Second, liberalism is morally decadent, encouraging the belief that everything is permitted and undermining the authority of those best situated to provide moral guidance to society, obscuring rather than illuminating the one true path to creation of the ideal political community. Third, liberalism is overly individualistic, promoting relentless self-obsession and encouraging the satiation of base personal desires while making no attempt to recognize much less re-enforce the shared background assumptions and bonds of community that hold society together and form the basis of what we call the common good. Fourth, liberalism is excessively rationalistic, or at least rationalistic in the wrong way, looking to reason for direction when what is required is a faith-based act of will, but denying that reason plays a role when it comes to ordinary moral decision-making. And finally, liberalism is incoherent, for the ends it purports to embrace are either individually impossible to fulfill or inconsistent with one another and therefore impossible to fulfill as a set, raising social expectations that cannot be satisfied and ensuring that society will be plagued by internal conflict and division. Exactly how the claim that liberalism leads to nihilism maps onto these various criticisms is unclear. Depending on what nihilism turns out to be, the claim that liberalism leads to nihilism could simply be a shorthand way of saying that liberalism has one or more of these problems in the extreme, or it could be a separate and independent criticism, a claim that liberalism produces a society that is not merely weak, decadent, overly individualistic, rationalistic in the wrong way, and plagued by internal conflict, but one in which morality plays no role at all. In either case, however, the claim that liberalism leads to nihilism is a fundamental part of the perfectionist attack, for it is what gives that attack its force—it lends urgency to all the other perfectionist complaints, and gives us reason to reject liberalism outright rather than try to tinker with it and overcome whatever lesser shortcomings its perfectionist critics may have accurately identified.

192

Mark R Reiff

To get a sense of the seriousness that perfectionists attach to the charge that liberalism leads to nihilism, consider the following. For Nietzsche, nihilism is ‘the will to nothingness, the great nausea’ that leaves man ‘suicidal’.⁴⁹ For Strauss, nihilism is ‘the contemporary rejection of natural right’, the view that ‘every preference, however evil, base, or insane, has to be judged before the tribunal of reason to be as legitimate as any other preference’, a view that represents ‘the second crisis of modernity—the crisis of our time’.⁵⁰ For Qutb, nihilism is a ‘disease’, an absence of vital values that has left mankind ‘on the brink of the precipice’, in danger of ‘complete annihilation’.⁵¹ And similarly grim descriptions of the danger that nihilism poses to contemporary society can be found throughout the writings and other pronouncements of the followers of both Strauss and Qutb.⁵² So while any defence of liberalism would need to go beyond the claim that liberalism leads to nihilism before it could be complete, responding to the charge that liberalism leads to nihilism is the place where any defence of liberalism must start.

A. What is Nihilism? The first thing we must do before we can defend liberalism against the charge that it leads to nihilism is get clearer about exactly what nihilism is supposed to be, for those who make the charge are often not very clear about this themselves. First, by associating liberalism with nihilism, perfectionists could be claiming that liberalism leads to the belief that life is meaningless, that it has no purpose or value. We might call this the existential conception of nihilism, since it is popular with that group of philosophers commonly referred to as existentialists.⁵³ The reason why this might be the conception of nihilism that perfectionists have in mind is that if one believed that life had no value, one might not be very concerned with preserving it, and this seems to be what many contemporary perfectionists ⁴⁹ F Nietzsche, On the Genealogy of Morality (Keith Ansell-Pearson (ed), Cambridge, 1994) 71, 94, and 127. ⁵⁰ Strauss, Natural Right and History n 8 above, 5, 42, and 253. ⁵¹ Qutb, Milestones n23 above, 7. While Qutb does not use the word ‘nihilism’ when he refers to the absence of vital values in modern society, it is apparent from his writings that some form of nihilism is what he has in mind, and not merely jahiliyyah, or ignorance of the divine guidance. ⁵² See, eg I Kristol, Neoconservatisim: The Autobiography of an Idea (New York, 1995) 101, 103 (describing ‘self-destructive nihilism’ as the ‘enemy’, the one ‘authentic and permanent possibility that any society had to guard against’); A Bloom, The Closing of the American Mind n 36 above, 155–7 (describing nihilism as ‘the soul’s basement’, characterized by ‘a chaos of instincts or passions’ and ‘existential despair’). Indeed, just as the world is steeped in jahiliyyah and nihilism to the followers of Qutb, for Straussians, ‘the world is full of nihilists’. A Norton, Leo Strauss and the Politics of American Empire (New Haven, 2004) 120. See also Drury, Leo Strauss and the American Right n 27 above, 7–9, 19 and C Zuchart and M Zuchart, The Truth about Leo Strauss: Political Philosophy and American Democracy (Chicago, 2006) esp 236–9. ⁵³ See, eg A Camus, The Rebel (A Bower (trans), London, 1953); J-P Sartre, Nausea (L Alexander (trans), New York, 1964). Nietzsche also sometimes refers to nihilism as the belief that life is meaningless or valueless. See B Reginster, The Affirmation of Life: Nietzsche on Overcoming Nihilism (Cambridge, 2006) 21–53, esp 23–33.

The Attack on Liberalism

193

are thinking when they blame liberalism for the cavalcade of horrors that characterized the first half of the 20th century. And believing that life had no value would certainly explain why modern man might be depressed and ‘suicidal’, as Nietzsche claims. There is a problem, however, if we assume that the existential conception of nihilism is what perfectionists are referring to when they claim that liberalism leads to nihilism. To believe that life has no value, one must have a theory of value, for otherwise one would have no basis for judgement. But this is precisely where liberalism is supposed to have gone wrong—it lacks a sufficiently substantive theory of value, which is why perfectionism is supposed to be an improvement upon it. So if nihilism is the belief that life has no value, it is hard to see how liberalism could lead to this. And if liberalism does contain a sufficiently substantive theory of value, it is hard to see why liberals would be especially likely to become nihilists, for there is no reason to believe that their theory of value would necessarily be one that life could not fulfill. The existentialist conception of nihilism accordingly cannot be what perfectionists have in mind, for it cannot be causally tied to some inherent feature of liberalism. Alternatively, the claim that liberalism leads to nihilism could be an epistemological claim about the ability of people in a liberal society to access or otherwise determine what their values are. On this interpretation, nihilism is not the belief that life has no value, but a form of malaise, a sense that life does have value combined with the frustrating feeling that the source and nature of this value is beyond our ability to comprehend. As C Wright Mills observed, when people know what they value and do not feel that what they value is under threat, they experience well-being. When they know what they value but do feel under threat, they experience fear, and in extreme cases panic. When they do not feel under threat but do not know what they value, they experience apathy. And when they do not know what they value yet still feel under threat, they experience anxiety.⁵⁴ On the epistemological conception, a nihilistic society would accordingly be steeped in either apathy or anxiety, depending on whether it perceived itself as under threat. There is some reason to think this is what perfectionists might have in mind when they speak of nihilism because they do claim that liberalism’s lack of an overriding substantive theory of value leaves a political community alternatively indifferent and indecisive, paralyzed by internal conflicts, and characterized by a great deal of unproductive existential angst. But the crisis in which modern liberal society finds itself is not merely a psychological crisis, according to most perfectionists, it is a moral and political one, one that is directly caused by conceptual errors and inconsistencies, not indirectly caused by epistemological difficulties. The problem is not that people in a liberal society are unable to figure out what their values are, but that such people are led to believe there are no values to be found. It is moral leadership that people require, perfectionists contend, ⁵⁴ See CW Mills, The Sociological Imagination (Oxford, 1959) 17–8.

194

Mark R Reiff

not simply a better education in the mechanics of moral reasoning. Apathy and anxiety may accordingly be symptoms of the crisis of modernity, but they are not the cause. While it comes closer to describing what perfectionists think ails liberalism, the epistemological conception of nihilism still fails to capture precisely what perfectionists claim has gone wrong. A more promising alternative is to interpret the charge that liberalism leads to nihilism as a claim that liberalism leads to the belief that there are no objective values, meaning that liberalism suggests that values are not ‘real’, not ‘part of the fabric of the world’, and that evaluative statements accordingly cannot be either true or false. Understood in this way, the claim that liberalism leads to nihilism becomes a claim that liberalism encourages a certain metaphysical view about the ontological status of moral judgements, and value nihilism reduces to value subjectivism. There is some reason to believe that this is what perfectionists have in mind when they claim that liberalism leads to nihilism because there is textual support for this view in many of their writings. Nietzsche, for example, claimed that moral values are ‘false projections’ onto a world that is empty of them, and that when we make moral judgements, we simply ‘fashion something that had not been there before.’⁵⁵ And Leo Strauss expressly referred to nihilism as the belief that there are no objective values.⁵⁶ It is not at all clear, however, that what these and other perfectionists identify as the consequences of moral nihilism are consequences that moral subjectivism alone could possibly entail. Indeed, it is not at all clear that embracing moral subjectivism would necessarily have any effect on the way people function in or experience the world. For example, while moral judgements cannot be true or false in the same way that judgements about physical facts can be true or false if moral judgements are subjective, they can still be true or false in a meaningful sense—they can be more or less well-supported by reasons.⁵⁷ Subjective judgements can therefore still be the subject of disagreement, for people can disagree as to what counts as a reason for action or belief, the weight to be assigned to the reasons that do count, and how conflicting reasons are to be balanced against one another. Subjective judgements can also function like objective judgements in every other way—they can form the basis of a conception of the good, allow us to distinguish right from wrong, and enable us to implement a coherent plan of life. There is accordingly no reason to believe that embracing moral subjectivism would change the way anyone actually behaves, or leave people feeling that they are disabled from subjecting others to moral criticism.⁵⁸ ⁵⁵ See F Nietzsche, The Will to Power (Walter Kaufman (ed), New York, 1967), s 12(B); F Nietzsche, The Gay Science (Bernard Williams (ed), Cambridge, 2001), s 301. ⁵⁶ See eg Strauss, Natural Right and History n 8 above, 37. ⁵⁷ For further discussion of this view, see Scanlon, What We Owe to Each Other n 31 above. ⁵⁸ Th is, for example, is Hare’s view. See RM Hare, ‘Nothing Matters’ in Hare, Applications of Moral Philosophy (London, 1972) 32–47, esp 40–41. For a recent discussion of Hare’s view, see D Parfit, ‘Normativity’ in R Shafer-Landau (ed), Oxford Studies in Metaethics: Volume I

The Attack on Liberalism

195

But perhaps there is a deeper problem that moral subjectivism creates. While it may be true that subjective moral judgements can be based on reasons at the applied level, one must have a moral theory to determine what counts as a reason for what, and moral subjectivism seems to imply that the selection of a moral theory at this higher level of abstraction cannot be based on reason, but must be based on unfettered self-interested desire alone. And according to some perfectionists, if the selection of a moral theory cannot be justified by reason, then its applications cannot either. Roberto Unger, for example, argues that what is wrong with liberalism is that it ultimately reduces to a ‘morality of desire’ that ‘provides no standards for preferring some desires to others’. Having accepted the ‘moral impotence of reason’ at this higher level, liberals are accordingly committed to ‘absolute moral scepticism’, for a moral theory that cannot be selected by reason cannot justifiably be considered a moral theory at all.⁵⁹ There are two problems, however, with associating nihilism with this view. First, the same criticism could be levelled against perfectionism. Once one has embraced a particular perfectionist view, one has a basis for using reason to make moral judgements. But on what is the selection of a particular perfectionist view to be based? Whatever conditions characterize this pre-moral choice situation and whatever criteria for choice apply, they apply with equal force to the position of the objectivist and the subjectivist. Remember, the disagreement between the objectivist and the subjectivist is not about the content of morality, but about its ontological status—that is, whether moral truths exist independently of our ability to recognize them. There is nothing about this disagreement that gives the objectivist more tools of reason in the pre-moral choice situation than the subjectivist. So if reason is unavailable to the subjectivist at this level, it is unavailable to the objectivist too. If this means subjective moral judgements cannot meaningfully be said to be true or false, or right or wrong, then objective moral judgements have these problems too. No doubt this is why many perfectionists expressly deny that reason is available in the pre-moral choice situation, and criticize liberals for thinking that it is. Indeed, rather than deplore the unavailability of reason at this level, these perfectionists celebrate it. Schmitt, for example, argues that the unavailability of reason at this level is precisely why dictatorship as a form of government is justified— political leadership requires the ability to make decisions without resort to reason, decisions that are absolute, irrevocable, and ‘created out of nothingness’.⁶⁰ And as we have already seen, a similar contempt for the idea that there is a role (New York, 2007) 325–80, 325–30. While Parfit disagrees with Hare’s claim that the difference between moral objectivism and subjectivism is trivial, he agrees that embracing one view rather than the other would not have any discernable effect on the way people actually behave or experience the world. ⁵⁹ R Mangabeira Unger, Knowledge & Politics (New York, 1975) 52. Strauss, of course, held a similar view. See Strauss, Natural Right and History n 8 above, 42, 49, 62–3, and 74. ⁶⁰ See Schmitt, Political Theology n 19 above, 66.

196

Mark R Reiff

for reason in the selection of moral theories can be found throughout the work of Maistre.⁶¹ So even if moral subjectivism did disassociate morality and reason at this higher level of abstraction, this cannot be what these perfectionists have in mind when they claim that liberalism leads to nihilism. Second, it is not at all clear that moral subjectivism does disassociate morality and reason at this higher level of abstraction. Reason has a role to play even here, for requirements of coherence, consistency, and universalizability impose limits on the moral theories we may select. People in the pre-moral position are therefore not subject to the unrestrained forces of raw self-interested desire, but to the rational forces of enlightened self-interest, and there is no reason to believe that these forces will produce mostly unacceptable versions of morality. Indeed, the entire Kantian project, both in its original and its various contemporary forms, is devoted to developing a moral theory using the facilities of reason alone. And if the requirements of coherence, consistency, and universalizability do not eliminate all but one moral theory from consideration, this merely means that as far as reason is concerned, a plurality of moral theories are equally defensible, a result that supports the liberal pluralist approach to political morality and suggests that the anti-pluralist perfectionist approach is incorrect. There is nevertheless one further possibility to consider. Perhaps the problem with moral subjectivism is not that it leads people to believe that nothing matters, but that it leads them to believe that nothing matters very much. In other words, what moral subjectivism does is undermine our commitment to our values. Subjectivism is therefore problematic because it weakens the force of moral judgement, and a society that is only weakly committed to its values is open to moral drift and ultimately to capture by charismatic moral zealots whose commitment to their own values is, in contrast, fierce. But where are these zealots supposed to come from? If they can be produced by a liberal society too, then perfectionists have to explain why the commitment of most but not all people in a liberal society is likely to be weak, for there is no problem if many people have a strong commitment to their values or if no one does. And why should we assume that those who are fiercely committed to their values are likely to be committed to values we should objectively reject? Unless there is some reason to believe that these values are most likely to be pernicious, there is no cause for concern. And finally, even if there were cause for concern, why is the solution to turn everyone into zealots rather than to limit the influence of the zealots who manage to insert themselves into our midst? The solution perfectionists propose simply assumes that perfectionism is correct. If a subjectivist conception of moral nihilism is not sufficient to show that nihilism would have the profoundly negative effects that perfectionists associate with it, there is only one possible candidate left. When perfectionists charge that liberalism leads to nihilism, they must mean that liberalism leads to the belief ⁶¹ See Berlin, ‘Joseph de Maistre and the Origins of Fascism’ n 24 above.

The Attack on Liberalism

197

that that there are no values, where ‘values’ is understood to include both subjective and objective values. Not only does this conception of nihilism directly incorporate the central complaint that perfectionists make against liberalism and capture the tone in which this claim is usually conveyed, there also appears to be a plausible causal mechanism to explain how liberalism might lead to this kind of nihilism. In fact there appear to be two. I will say more in a moment about whether these purported causal mechanisms are as plausible as they seem, but before we get to this, there are two other problems that I want to address.

B. Is Nihilism Coherent? First, is the belief that there are no values coherent? What would it be like to hold such a view? Could anyone actually do so? If you believed that there were no values, what criteria would you use for making decisions? Suppose, for example, you were a nihilist and hungry. Would you get up and get something to eat? Or would you, like one of the characters in Beckett’s Endgame, simply have a detached, bemused attitude toward your hunger, unable to determine whether this feeling was something you wanted to encourage or suppress?⁶² In answering these questions, I will first assume that the unitary theory of value is correct. Under the unitary theory of value, there are not moral, prudential, and aesthetic values, there are simply values, or to put it differently, all values are moral values in some meaningful sense. If nihilism is understood as the belief that there are no values, subjective or objective, and the unitary theory of value is correct, then if one were a nihilist one could not have values of any kind. But it is not clear how we could eliminate all forms of value from an agent yet still leave the agent with the ability to engage in practical reasoning. Indeed, as Rawls notes, without at least a thin theory of the good, it is not clear how one could evaluate options or make rational choices about anything.⁶³ It is not even clear how one could think. To see why, consider the following. The desire-based account of value says that something has value because we desire it. But if I have no values, the desire-based account of value will not do, for if I have desires I must have values and thus I could have neither if I am a nihilist and the desire-based account of value is true. And with neither desires nor values, what source of reasons for action would remain? Without some explanation for this, it does not appear that the nihilist would be able to respond intelligently to the world or function in it in any meaningful way. The nihilist could only act or react on instinct, out of habit, or perhaps simply on some randomized basis irrespective of stimulus. On this conception, the nihilist would be reduced not merely to the functionality of an animal, but to something less than this. He would be a primitive organism, incapable of forming ⁶² See S Beckett, Endgame (New York, 1958) 36. ⁶³ See Rawls, A Theory of Justice n 2 above, s 60, 347–50.

198

Mark R Reiff

ends or rationally selecting means to deal with his environment, and therefore at the mercy of the principles of evolutionary biology, which would alone determine whether his instinctual genetic programming was sufficient for him to survive. The result is no different if we assume that the buck-passing account of value rather than the desire-based account is true. Under the buck-passing account, having value simply means having properties that give us reason to value whatever has those properties.⁶⁴ If a nihilist has no values, then either reasons to feel otherwise do not exist, or if they do exist, the nihilist is unable to see them and therefore cannot act on them in any event. Under the buck-passing account, the nihilist could only have reasons for action that arise out of properties that did not provide reasons to attach value or disvalue to some action. But it is not clear that any such reasons would exist. To believe that there were, one would have to believe in the buck-passing account of value but the desire-based account of reasons, under which desires can provide reasons for action. It is not clear, however, that anyone could coherently hold such a combination of views. Sorting this out would be a complex matter, too complex to fully explore here, but examples of contemporary theorists who embrace this combination of views are certainly difficult to find. Those theorists who accept the buck-passing account of value all seem to embrace a value-based account of reasons, under which desires do not provide reasons for action except perhaps in some trivial or exceptional cases, and they reject the idea that a functional set of reasons for action could be provided by desires alone.⁶⁵ In any event under either the desire-based account of value or the buck-passing account, once one has removed values, it is not clear that there is anything left. At least it is not clear that whatever is left is enough to leave the nihilist equipped to engage in practical reasoning. And if that is true, then nihilism is not a belief about the world that anyone could actually maintain. But suppose we reject the unitary theory of value, and accept that there is something that we can characterize as the moral point of view, where this is taken to mean something more than simply the view that one theory of morality is superior to whatever other theories are on offer. If we accept the existence of a specifically moral point of view, value can be moral, prudential, or aesthetic. Under this view, the moral nihilist is not without values, he is simply without moral values. Such an individual has no conception of the greater good, but he does have a ⁶⁴ See, eg Scanlon, What We Owe to Each Other n 32 above, 95–8. ⁶⁵ See, eg J Raz, ‘Incommensurability and Agency’, reprinted in Engaging Reason: On the Theory of Value and Action (Oxford, 1999) 46–66; Scanlon, What We Owe to Each Other n 31 above; D Parfit, ‘Rationality and Reasons’ in D Egonsson, J Josefsson, B Peterson, and T RonnowRasmussen (eds), Exploring Practical Philosophy: From Action to Values (Burlington, Vt. 2001) 17–39. For an argument in favour of a ‘hybrid’ view under which all values and some desires can provide reasons for action, see R Chang, ‘Can Desires Provide Reasons for Action?’ in R Jay Wallace, P Pettit, S Scheffler, and M Smith (eds), Reason and Value: Themes from the Moral Philosophy of Joseph Raz (Oxford, 2004) 56–90. Even Chang, however, does not contend that desires alone can supply the full panoply of reasons necessary for practical reasoning; she merely suggests that some reasons for action can be desire-based rather than value-based and that neither account of reasons is entirely correct.

The Attack on Liberalism

199

conception of his personal good, of what would make his own life go better, and of aesthetics, of what constitutes beauty. The moral nihilist is therefore simply someone whose actions are driven by his unrestrained self-interest, or by the pursuit of beauty, or by some combination of the two. Such a nihilist can still have desires, as long as these are limited to desires for prudential or aesthetic goods, and he can still have reasons for action, for these can be provided either by these desires or by properties that make whatever has them good for him personally or aesthetically good. While the claim that there is a moral point of view is controversial, it is at least possible that such a point of view exists.⁶⁶ And if we allow for non-moral value, a conception of nihilism that denies both objective and subjective moral value would be coherent. But if perfectionists rely on the existence of a moral point of view to make their conception of nihilism coherent, they have another problem. If they allow the moral nihilist to have prudential and aesthetic values, it is once again no longer clear that nihilism would be bad for us. At least it is no longer clear that the consequences of embracing nihilism would be as bad as most perfectionists seem to claim.

C. Is Nihilism Bad for Us? To argue that moral nihilism is bad for us even if it does not constitute total nihilism, perfectionists have to claim that there is something profoundly morally disturbing in a society were everyone is dedicated to the pursuit of non-moral values. But if self-interest is a non-moral value, then this is a difficult argument to make, for many theorists argue that pursuing one’s enlightened self-interest is actually what morality requires, or at least that pursuing one’s enlightened self-interest and doing what morality requires would often and perhaps even always produce the same kind of behaviour.⁶⁷ Of course, versions of morality that reduce to pursuing one’s enlightened self-interest might be overly individualistic, and this is indeed a criticism that perfectionists raise against liberalism, but this a very different category of complaint than the charge that liberalism leads to nihilism. Nihilism, after all, is supposed to bring us to the brink of a precipice. If nihilism is to be the ultimate horror that perfectionists make it out to be, they must do more than show that it would produce a society that is overly individualistic; they must show that it would produce some sort of moral and political catastrophe. One way to do this, I suppose, would be to claim that the moral nihilist might not be driven by his perception of his self-interest, but by his conception of aesthetic value. It is not inconceivable that someone motivated purely by a ⁶⁶ For a sceptical discussion of the existence of a moral point of view, see J Raz, ‘On the Moral Point of View’ reprinted in Engaging Reason n 65 above, 247–72. ⁶⁷ See, eg A Smith, An Inquiry into the Nature and Causes of the Wealth of Nations (London, 1950).

200

Mark R Reiff

conception of aesthetic value might be capable of some rather frightening conduct indeed. The emperor Heliogabalus, for example, is rumoured to have had people slaughtered because he thought that red blood on green grass looked beautiful.⁶⁸ And then there is the Joker, the character played by Jack Nicholson in the film Batman, who engaged in a series of violent and horrific murders because of what he saw as their aesthetic value. Of course, one can deny that such acts would have the aesthetic value that Heliogabalus and the Joker supposedly attached to them, but if someone were to believe that murder was a beautiful act and did not have any moral values that would prohibit such conduct or any concerns about how committing such acts might affect his own life or life prospects, he might indeed try to make his idiosyncratic aesthetic vision a reality. At the very least, a world full of Jokers would be a world in which it would be exceedingly dangerous to live. But there is no reason to believe that moral nihilists are likely to be Jokers. First, there is no explanation for why being driven solely by aesthetic value would necessarily or even regularly lead to such abhorrent conduct. Indeed, on some interpretations, Nietzsche saw the pursuit of aesthetic value as the way forward for overcoming nihilism, not as the reason why nihilism was bad for us. So even if moral nihilism did leave people inclined to engage solely in the pursuit of aesthetic value, it is not clear why perfectionists would necessarily be overly concerned about this. Second, to desire to engage solely in the pursuit of aesthetic value, moral nihilists would also have to lose their prudential values, and therefore no longer be motivated by the pursuit of what they saw as their self-interest, long-term or otherwise, for prudential values would often if not always restrain those otherwise inclined to pursue some murderous aesthetic vision. So unless perfectionists can explain why moral nihilists are also likely to be prudential nihilists but not aesthetic nihilists, it is not clear that nihilism would bring us anywhere near the brink of the precipice on which liberal society supposedly is teetering. Indeed, it is not even clear that total nihilism would do so, even though this would leave us unable to engage in practical reasoning. Like moral nihilists, total nihilists could not deliberately do evil, for they would have no basis for determining what evil was. But total nihilists could not even deliberately act selfishly, for they would have no basis for determining what their self-interest was, and they could not deliberately seek out aesthetic value either. Indeed, it is hard to see how they could do anything ‘deliberately’ at all. Of course, the common good that evolutionary biology alone would produce might be very different from the common good of most perfectionist visions of the ideal, but it is not clear that this cost would outweigh the benefit of eliminating the possibility that a society or elements within it might deliberately turn to evil. Total nihilism is bad only

⁶⁸ See RM Hare, Freedom and Reason (Oxford, 1963) 161.

The Attack on Liberalism

201

if instinctive or otherwise non-evaluative choices are more often evil than good, but it is not obvious that this would necessarily be true.

D. Two Roads to Nihilism Let us assume, however, that nihilism is a belief that people could actually hold and that the consequences of this would be as unattractive as perfectionists like to claim. Even so, the claim that liberalism leads to nihilism still has problems, for it is unclear how liberalism is supposed to lead to nihilism, defined as the belief that there are no values. Indeed, there are two branches to this particular perfectionist criticism, and each branch offers a very different explanation for how nihilism is reached. The first branch has its roots in Nietzsche. Nietzsche argued that nihilism was the natural result of the relentless pursuit of truth, the search for natural, rational explanations of all phenomena.⁶⁹ Those who follow this path must inevitably come to the conclusion that there are no values, for indeed there are no values. Nietzsche took this simply as a fact, and although he implied it was to be mourned rather than applauded, he did not believe this process could be reversed, only overcome. Nihilism was simply the existential state in which modern liberal society found itself, a fact on the ground that had to be dealt with rather than ignored. Exactly how Nietzsche thought nihilism could be overcome is a matter of some controversy, but while the details of exactly what he might have had in mind remain a bit obscure, it is his diagnosis of the cause of nihilism that is of interest to us here, not the content of what he thought would be the cure. Following in the Nietzschean tradition, one strand of modern perfectionism argues that it is liberalism’s obsession with reason and the pursuit of scientific truth that leads to nihilism. This leads to nihilism because the belief that there are no values is true—there are indeed no values—and if one pursues the truth relentlessly enough it is only a matter of time before one will discover that this is so. But this is not a truth that is good for the masses to know. Not in the sense of good versus evil, but in the Nietzschean pre-moral self-referential sense of good versus bad. It is not good for the community, for to have a community one must have culture, and to have culture one must have values. What matters is not whether these values are true, but whether they are sufficiently comprehensive and well-defined to give a community a particular cultural identity. Only then can a community exist, and only if he is embedded in and supported by a community can an individual achieve anything of substance. The paradox of modern life is accordingly that to overcome nihilism we must believe in values we know not to be true.⁷⁰ This is hard enough for the most sophisticated among us; for the masses, it is probably impossible. Without the purportedly absolute ⁶⁹ See Nietzsche, On the Genealogy of Morality n 49 above. ⁷⁰ See Bloom, The Closing of the American Mind n 36 above, 202–3.

202

Mark R Reiff

objective values that perfectionism provides, there is no hope of surviving the crisis that liberalism has created. For the Nietzscheans, truth is the enemy of moral certainty, but moral certainty is the cement that holds a community together. We must accordingly defeat nihilism, not because it is false, but because its defeat is necessary. The second branch of the claim that liberalism leads to nihilism has it roots in Hegel, and specifically in Hegel’s criticism of Kant’s theory of freedom. For Kant, individual autonomy was central, and so Kant argued that government should restrict freedom only when necessary to ensure that all people have equal freedom. But Hegel argued that Kant’s notion of equal freedom was empty.⁷¹ All freedoms do not have equal importance, and therefore one needs a conception of the good or a sense of the shared background assumptions and practices of a particular community before one may begin to conceive of what freedom means. And this conception of the good or sense of shared background assumptions is something that only perfectionism can provide. The Hegelian branch of perfectionist criticism accordingly focuses on liberalism’s commitment to neutrality and its corresponding unwillingness to endorse any specific set of substantive values, and claims that like Kant’s theory of equal freedom, liberalism is empty. Because it is empty, liberalism must inevitably lead those who embrace it to believe there are no values. But unlike the Nietzscheans, the Hegelians see the belief that there are no values as false—there are indeed values, although those within this branch of perfectionism disagree about where these values are to be found. Some believe they are to be found in the Word of God, others that they can be derived by the light of reason alone, or found in the philosophy of the ancient Greeks, or that they simply are the values that all or most members of a community happen to share. But the fact that these critics disagree as to where the requisite set of values is to be found is not important. What is important is that they all agree that such values do exist, and that what is wrong with liberalism is that it leads people to believe the contrary.⁷² Nihilism, of course, cannot be both true and false. Perfectionists are accordingly divided when it comes to explaining how liberalism leads to nihilism. It is merely because both groups arrive at the same conclusion that the fact that they do so for inconsistent reasons has remained unnoticed. Once we bring this into focus, however, it becomes apparent that one branch of this perfectionist criticism can be used to undermine the other. If nihilism is true—if there are indeed no values—then the Hegelian branch of criticism loses its force, since this depends on proving that nihilism is false. In contrast, if nihilism is false—if there are indeed values, but liberalism falsely suggests there are not—then the Nietzschean branch loses its force, for there would be no reason to believe that a ⁷¹ See GWF Hegel, Elements of the Philosophy of Right (AW Wood (ed), Cambridge, 1991) s 135. ⁷² See M Lilla, ‘The Closing of the Straussian Mind’ The New York Review of Books (4 November 2004) (characterizing Strauss’s view in Natural Right and History).

The Attack on Liberalism

203

relentless pursuit of the truth would lead to nihilism. On the contrary, it should lead to a form of perfectionism, to the discovery of the one true set of objective values, whatever these might be. But there are also independent reasons to challenge both branches of criticism. If there are no values, as the Nietzschean-inspired critics contend, then there are no values, regardless of whether we believe in perfectionism or liberalism. All perfectionists can do is offer a set of values that falsely claim objective status, making perfectionism a form of error theory, which seems inconsistent with the very nature of perfectionism itself.⁷³ But many liberals claim there are objective values, and they do so without embracing a perfectionist agenda—in other words, without committing to any particular set of values or even to the belief that only one true set could exist, for values could be plural yet objective.⁷⁴ Defending perfectionism against this claim places the Nietzscheans in an awkward position, for they must argue that pluralism is false while maintaining that objectivism is true even though they believe that it is not. The Hegelian-inspired claim that liberalism leads to nihilism runs into serious problems too. Recall that this branch of perfectionism claims that liberalism is empty because it does not endorse a specific comprehensive set of substantive values. But how does the lack of such an endorsement lead to the belief that there are no values? To think that it might flies in the face of everything we know about human nature. The belief that there are no values seems to be something that people are naturally designed to resist, for if this were not the case such a belief would be far more common than it is. Only the most extreme set of experiences could possibly produce such a belief, and even then, people who witness or are subjected to the most abominable acts often survive with their values intact. To think that people could be led to the belief that there are no values because this might be an implication of liberalism’s lack of endorsement of a specific comprehensive set of values is to assign such an implication a degree of persuasive power that could not reasonably be expected to obtain. But more importantly, it is unclear why anyone would think that liberalism’s failure to endorse a comprehensive set of values might be understood as implying that there are no such values to endorse. A suitcase may be empty, but that does not lead anyone to believe that there is nothing that could be placed inside it. Liberalism is like that suitcase—it is a container for the substantive values of multiple comprehensive conceptions of the good, a way of providing for the nondestructive interaction of competing sets of belief. The fact that it does not come pre-stocked with a full set of substantive values provides no reason to believe that such substantive values do not exist. Indeed, if there were no further substantive values for liberalism to contain, there would be no reason for such a container to exist. Contrary to what its Hegelian-inspired critics contend, the fact that the ⁷³ For more on error theory, see JL Mackie, Ethics: Inventing Right and Wrong (London, 1977). ⁷⁴ See, eg WA Galston, The Practice of Liberal Pluralism (Cambridge, 2005).

204

Mark R Reiff

values endorsed by liberalism are limited to the realm of the political actually implies that further substantive values do exist, not that they do not. But even if liberalism’s lack of endorsement for a specific set of substantive values did imply that there were no such values, this would not constitute an argument for perfectionism, for the same criticism could be levelled against perfectionism with equal force. While those who embrace the fundamental presuppositions of perfectionism all embrace some comprehensive moral doctrine, there is nothing in these fundamental presuppositions that specifies which particular comprehensive moral doctrine this must be. At the tectonic level, perfectionism is empty too, which is why perfectionism can take on many forms. So if this is a problem for liberalism, it must also be a problem for perfectionism. In either kind of society, people still have to select a comprehensive moral doctrine, and pressure to select a particular comprehensive moral doctrine arises only after the beliefs of the particular society are given content. True, in a perfectionist society, this pressure comes directly from the government, but in a liberal society there will be pressure too—not only from relatives and friends, but also from teachers, religious leaders, and public figures. Even if these people do not try to impose their values on us, they will nevertheless have great influence on us as long as they are people we respect. Some people may choose nihilism in either kind of society despite this pressure to conform, but there is no reason to believe that people are more likely to be nihilists in one kind of society than the other. So whichever branch of perfectionist criticism we embrace, its central claim—that liberalism leads to nihilism—is highly questionable at best. But even assuming that liberalism does lead to nihilism and that nihilism is as unattractive as perfectionists make it out to be, this does not amount to an indictment of liberalism. It is at best only the first part of an indictment. It is a consequentialist argument against liberalism, and like all consequentialist arguments, it has moral force only if the consequences of the proposed alternative— embracing perfectionism—would be better. As we shall see in the next section, however, there is substantial reason to believe that in the long term—and very often in the short and medium term too—this will not be true. Whatever the consequences of embracing liberalism, the consequences of embracing perfectionism are likely to be worse.

III. Perfectionism and Fanaticism Perfectionists tend to make their case against liberalism agonistically. In other words, they illustrate liberalism’s supposed defects, and then simply claim that perfectionism is preferable because it avoids these problems. Of course, there is nothing wrong with making a case agonistically in politics or philosophy. Indeed, resort to this form of argumentation is common among liberals too. So it is not surprising that liberals tend to respond to perfectionist attacks simply by

The Attack on Liberalism

205

arguing that liberalism does not suffer from the defects that perfectionists have supposedly identified, and if it does, that perfectionism suffers from these same defects too. Given their similar approach to defining the issues between them, the battle between liberalism and perfectionism is almost always fought exclusively on liberalism’s turf. Perfectionists make a positive rather than an agonistic case for perfectionism only rarely, and liberals never seem to wonder whether perfectionism might have problems of its own, problems that are unrelated to its inability to capture what liberals believe is attractive about liberalism. The claim that liberalism leads to nihilism is no exception to this pattern and practice of debate. Both perfectionists and liberals have tended to assume that if liberalism leads to nihilism, the case against liberalism has been won. The focus of both camps has accordingly been on this claim, rather than on whether perfectionism suffers from some even more distasteful defect. The result is that perfectionists have enjoyed a strategic advantage in their battle with liberalism, for they have been able to avoid any examination of the consequences that embracing perfectionism might bring and any comparison between these consequences and the consequences of embracing liberalism. If we do make such a head-to-head comparison, however, the perfectionist attack on liberalism looks much more dubious indeed.

A. The Instability of Perfectionist Communities One characteristic of perfectionism that is often not acknowledged is that it is inherently unstable. All perfectionist communities have a vision of an ideal society, a society composed of those and only those who embrace a specific comprehensive set of substantive values. Often, perfectionist movements start by specifying these values very generally, but regardless of the specificity with which they begin, as they mature and attract more and more adherents they will naturally come under increasing pressure to define themselves more finely. After all, the whole point of perfectionism is to strive toward an ideal, and the closer one gets to this ideal, the more detailed one’s perfectionist vision can become. But the more specific a perfectionist vision becomes, the more people it will exclude, for fewer and fewer people will be able to meet its ever more demanding requirements. Perfectionist communities accordingly have a tendency to implode, like a collapsing star, until ultimately only a handful of extremists are left to fight it out among themselves for the mantle of being the only ‘true’ or ‘pure’ members of the community. The implosive forces at work within perfectionism can be found in perfectionist movements on both the left and the right. Marxist movements, for example, have historically been plagued by factionalism, often splitting over trivial matters of policy, shedding members, and directing as much of their energy toward internal purification as toward external expansion and success. Compounding this problem is the fact that like all perfectionist movements, Marxist groups are

206

Mark R Reiff

more concerned about obedience than they are about acceptance, and therefore tend to eschew the kind of deep intellectual engagement that would be necessary to overcome scepticism in favour of simple indoctrination. Among the obedient, the degree of commitment to perfectionist values is accordingly likely to be thin. Indeed, the collapse of communism in Eastern Europe in 1989 can be seen as the ultimate example of this self-defeating weakness in perfectionism’s internal architecture. Instead of amassing greater and greater numbers of true believers as these societies matured, they eventually came to be populated by almost no true believers at all. Similar examples of these implosive forces can be seen in Islamic fundamentalism. What began as a movement claiming to speak on behalf of the entire Muslim political community is now divided between Sunni and Shiite factions and to some extent between Arab and non-Arab factions as well.⁷⁵ And while still relatively united, American neo-conservatism is also beginning to splinter and divide, with various former champions accusing one another of betraying the core values of the movement.⁷⁶ Without the built-in constraints for dealing with the disagreements over values and policy that are bound to arise as a perfectionist movement develops and matures and encounters new and more complex problems, constraints that only liberalism can provide, the movement has more and more difficulty maintaining its internal cohesion. Even without pressure from its external enemies, such a movement will inevitably begin to devour itself as it pursues greater purity and perfection. Exacerbating this problem is the tendency for perfectionist communities to be hijacked by Machiavellians who use the pursuit of perfection and the excuse of purification that perfectionism provides as a route to personal power. Stalin is perhaps the most conspicuous example of this, but history is replete with others. The disintegration of the Balkans can be attributed to this, as people like Slobodan Milosevic used a nationalistic brand of perfectionism as a way of capturing the imagination of the masses and manipulating the democratic process to bring them personal power. Joseph McCarthy and his henchmen in the United States were arguably as much Machiavellians as true believers, and depending on how cynical one wants to be about the motives of those involved, some American neo-conservatives might be put into this category as well. In any event, there is no denying that it is far more difficult to sort Machiavellians from perfectionists than it is to sort Machiavellians from liberals. But internal instability is not the only problem that makes perfectionist communities less attractive than liberal ones. If it were, we could simply isolate any ⁷⁵ See, eg E Wong, ‘Sectarian Divide: On Web, a Sunni-Shiite Split on Hezbollah’ New York Times (22 July 2006) A8. For a history of the sectarian struggle between Sunni and Shiite fundamentalism, see V Nasr, The Shia Revival: How Conflicts within Islam Will Shape the Future (New York, 2006). ⁷⁶ See, eg F Fukuyama, America at the Crossroads: Democracy, Power, and the Neoconservative Legacy (New Haven, 2006).

The Attack on Liberalism

207

particularly unattractive brand of perfectionism and until it eventually burns itself out. Unfortunately, in addition to the implosive forces they tend to face, perfectionist movements also have a marked tendency toward violence.

B. The Perfectionist Affinity for Violence To understand why perfectionist political communities have a tendency toward violence, it is important to distinguish perfectionism as a set of fundamental presuppositions that govern the structure of political life from perfectionism as a specific comprehensive set of substantive moral values. I do not contend that the mere embrace of a specific set of substantive values makes anyone prone to violence. Everyone, even those who live in a liberal society, will embrace some set of substantive values. It is the embrace of perfectionism as a political structure that leads to violence, not the nature of the specific set of values that a particular perfectionist community happens to adopt. The reason for this has to do with the way that perfectionist political communities define themselves. The driving force behind perfectionist politics is the friend/enemy distinction, which means that perfectionist political communities tend to define themselves negatively, by defining what they are not.⁷⁷ Note that I am not suggesting that perfectionist political communities lack a positive moral programme. On the contrary, putting forth a positive moral programme is an essential part of any perfectionist agenda. I am merely suggesting that their positive moral programme is predominantly reactionary, drawing most of its specificity and organizing force from that which it rejects rather than from some underlying set of independently derived free-standing principles. Indeed, as a matter of practical if not conceptual necessity, perfectionist communities need enemies, for without the cohesive force and the compelling sense of mission that facing a threat from a supposedly powerful enemy provides, a perfectionist movement is unlikely to be able to capture the imagination of the populace and acquire and maintain any real political power. Hitting on the right enemy—one that can not only be portrayed as powerful but that can also be made to seem sufficiently foreign and unfathomable that any natural inclination toward empathy can be easily suppressed—is accordingly the key to any perfectionist movement’s success. Islamic fundamentalists, for example, were largely marginalized until they switched their attention from the near enemy (the corrupt and oppressive and largely secular governments in their own countries) to the far (America, the great Satan).⁷⁸ And American neo-conservatives nearly became irrelevant after the Soviet Union and its ‘evil empire’ inconveniently self-destructed, reclaiming ⁷⁷ See Schmitt, The Concept of the Political n 7 above. Schmitt’s view of the friend/enemy distinction as the driving force of politics was also held by Leo Strauss, and through him passed on to many of his neo-conservative disciples. See Drury, Leo Strauss and the American Right n 27 above, 23. ⁷⁸ See FA Georges, The Far Enemy: Why Jihad Went Global (Cambridge, 2005).

208

Mark R Reiff

their former influence and power only after 9/11 when they redefined themselves as the enemy of terrorism and the inscrutable ‘axis of evil’. The problem is that built into the friend/enemy distinction is the constant threat and inherent possibility of violent conflict.⁷⁹ Enemies are naturally viewed with suspicion, so a perfectionist community will tend to interpret the ambiguous actions of its enemies as aggressive and act pre-emptively before such enemies have an opportunity to grow stronger. Indeed, the friend/enemy distinction often becomes a self-fulfilling prophecy, for the mere act of designating some one or group as an enemy is an act of vilification, and vilification is a provocative act, thereby increasing the chances that those so designated will behave aggressively and requiring the designating community to respond in kind. This, for example, appears to be at least part of the dynamic that is currently driving relations between the United States and Iran. But the tendency toward violence is not merely the unintended consequence of the perfectionists’ division of the world into friends and enemies. It is also a necessary consequence of the conceptual commitment to a perfectionist political outlook. To embrace perfectionism as a political morality is to commit oneself to the creation and maintenance of a specific culture. Cultures, however, do not exist in isolation. The very idea of culture implies a multiplicity of them. Unless they are able to remain isolated from one another, cultures will necessarily compete for influence both between and within political communities. Indeed, in the modern age, technological advances have effectively made it impossible to prevent one culture from interacting with another. But without the mechanism for peaceful interaction that liberalism provides, without the ability to accommodate competing conceptions of the good, cultural interaction is equivalent to an attack by one cultural community on the other. This, no doubt, is what Allan Bloom means when he says, ‘liberal democracies do not fight wars with one another, cultures fight wars.’⁸⁰ Cultures that are allowed to interact cannot help but be influenced by one another. Absent suppression of competitors, a perfectionist culture accordingly cannot hope to be kept pure. And suppression, of course, usually requires or provokes violence, and often both. This is why violence is so often mythologized and celebrated by those who have a perfectionist point of view. ‘Whatever value human life has does not come from reason; it emerges from a state of war between those inspired by great mythical images to join battle’, says Schmitt.⁸¹ ‘The whole earth, perpetually steeped in blood, is nothing but a vast altar upon which all that is living must be sacrificed without end, without measure, without pause, until the consummation of things, until evil is extinct, until the death of death’, says Maistre.⁸² And it would ⁷⁹ Schmitt’s The Concept of the Political is laced with statements to this effect, but see especially 32–7. ⁸⁰ Bloom, The Closing of the American Mind n 36 above, 202 (emphasis added). ⁸¹ Schmitt, Crisis in Parliamentary Democracy n 43 above, 71. ⁸² Berlin, ‘Joseph de Maistre and the Origins of Fascism’ n 24 above, 111 (quoting Maistre).

The Attack on Liberalism

209

be a mistake to dismiss such statements as mere perfectionist hyperbole. For perfectionists, war is the natural means by which competing cultural myths come to have purchase in the world. Indeed, for perfectionists, having a culture means engaging in war with other cultures.⁸³ The jihad against the West is accordingly not just a matter of cultural self-defence, brought on by contingent circumstances that might have developed in some other way, it is a conceptual necessity, a perfectionist act of creation for the radical Islamic fundamentalist community. The War in Iraq, in turn, is not just a (misguided) response to this jihad, it is a perfectionist act of creation of a different sort, one designed to both express and sustain America’s new neo-conservative identity. In addition to the conceptual connection between perfectionism and violence, perfectionists also often have powerful instrumental reasons to resort to violence. Perfectionist communities tend to thrive under conditions of low-intensity armed conflict, for this justifies the community’s internal emphasis on security and its subjugation of the rule of the law, increases the willingness of the masses to cede authority to the governing elite, and makes the community more cohesive and narrowly focused on protecting its ideals. And while the complete destruction of its enemies can have a destabilizing effect on a perfectionist community by removing this incentive for community cohesion, it is difficult to keep such conflicts under control. Chances are that the tension between a perfectionist community and its enemies will eventually erupt into all out war. Perfectionist political communities are also a danger to their own members as well as to outsiders. Internal enemies must be purged if the community is to remain pure, and while ex-communication or expulsion is sometimes enough to accomplish this, calls for suppression of internal opposition will often take a more ugly, violent, turn. As Stalin said, ‘Death solves all problems: no man, no problem’, a view that has adherents among perfectionists groups of all sorts, including at least some on the contemporary Christian right.⁸⁴ Violent suppression may even be used against the innocent, such as in show trials, as another way of promoting internal unity and cohesion. Indeed, given their view that the community rather than the individual is the fundamental social unit, sacrificing the innocent individual for the good of the community does not present the same sort of moral problem for perfectionists that it would for liberals. Under liberalism, the individual is often inviolable, even when the good of one is outweighed by the good of many. Under perfectionism, however, the individual and the community do not exist on the same moral plane. The sacrifice of an individual in such a situation has the same relative moral significance as the sacrifice of an animal or a work of art. What all this means is that even if nihilism were coherent, even if liberalism necessarily or often leads to it, and even if nihilism were as unattractive as ⁸³ See Bloom, The Closing of the American Mind n 36 above, 202. ⁸⁴ See Goldberg, Kingdom Coming n 40 above, 160.

210

Mark R Reiff

perfectionists make it out to be, this would not constitute an argument against liberalism. Whatever the consequences of embracing liberalism might be, the consequences of embracing perfectionism are even worse. Indeed, it was fanaticism, not nihilism, which was responsible for the parade of 20th-century horrors that perfectionists typically try to blame on liberalism. The Nazis were perfectionists, not nihilists, and even if it were true that there was some inherent defect within liberalism that made the rise of Hitler possible, one cannot claim that liberalism is more responsible for the horrors generated by the Nazi brand of perfectionism than perfectionism itself. Whatever the defects in liberalism, it can produce a tolerable quality of life even when its implementation is only rough and imperfect; by definition, perfectionism cannot. In any head-to-head consequentialist comparison between liberalism and perfectionism, it is liberalism not perfectionism that comes out on top.⁸⁵ ⁸⁵ Earlier versions of this paper were presented at the Cambridge Forum for Legal and Political Philosophy, the University College London Colloquium on Law and Philosophy, and the Manchester Centre for Political Theory. I would like to thank all those in attendance for their comments and suggestions. I would especially like to thank Peter Jones and Jonathan Quong for providing me with extensive and very helpful written comments on earlier versions of the paper, and Sorin Baiasu, Adrian Blau, Geoff rey Brennan, Kimberley Brownlee, Harriet Davidson, Ross Harrison, Christoph Kletzer, Matthew Kramer, John O’Neill, John Oberdiek, Hillel Steiner, and Stephen de Wijze, for their many helpful comments and suggestions and for much stimulating discussion, all of which led to substantial improvements to the paper.

11 Moral Reflections on the Responsibilities of Soldiers: The Clue to Devising a Legal Definition of Terrorism¹ Robert Morris

In 1984 Schmid and Jongman, in their extensive work, Political Terrorism, identified 109 different definitions of terrorism and then, in a later edition, claimed ‘the search for an adequate definition of terrorism is still on.’² I hope this paper will not just add to the list. Now, there are two different kinds of definitions of terrorism which can be given—conceptual definitions and legal definitions. A conceptual definition needs to be morally accurate, devised from an objective viewpoint and it must highlight the important distinctions between terrorists and non-terrorists. I would suggest, however, that a conceptual definition can use disputable terms. A legal definition, however, is quite different. It is of key importance that a legal definition is clearly defined but one must also realise that a legal definition will always be devised from the perspective of a particular state or group of states. This does not mean that the two kinds of definition have no relation to each other. If our legal definition of terrorism is not to be arbitrary and if it is to be justifiable then we would like our conceptual definition of terrorism to inform our legal definition of terrorism. Indeed, as I will go on to explain, in the ideal situation there would be one universally accepted conceptual definition of terrorism which informed every state’s legal definition of terrorism. What I hope to do in this paper is show that the answer to a question in applied ethics surrounding the responsibilities of soldiers provides us with the means of devising a conceptual definition of terrorism which all states can agree upon and a morally acceptable means of legally distinguishing between terrorists and non-terrorists. ¹ This paper came out of research I conducted at King’s College London on the Just Cause Principle. I would like to thank Barrie Paskins and the AHRC for enabling me to do this research. I would also like to thank Antony Duff, Mark Reiff, Kimberley Brownlee and Sandra Marshall for their comments and objections. ² AP Schmid and AJ Jongmann, Political Terrorism (2nd edn, New Brunswick, 1988) 1.

Moral Reflections on the Responsibilities of Soldiers: The Clue to Devising a Legal Definition of Terrorism. Robert Morris. © Oxford University Press 2007. Published 2007 by Oxford University Press.

212

Robert Morris

This paper is split into four parts. First, building on the work of Schmid I will propose a conceptual definition of terrorism which I think is both morally accurate and also politically acceptable to states. In the second part of the paper I will lay out the moral problem that is often thought to be fatal to the kind of definition I have proposed. In the third part I hope to solve that problem by examining the nature of the responsibilities of soldiers. And then in the last part I hope to show how the understanding of terrorism that I have offered here can help us to improve legal definitions of terrorism. Before, I move on, however, I should clarify that in talking about terrorism, I am only focusing on the kind of violent, terrifying activities which are committed by non-state groups for political reasons. These acts may be designed to coerce or intimidate governments, or they may be acts of revenge.³ This is the kind of activity which states aim to legislate against and this is the kind of activity which receives so much attention in modern discussions. Whilst the term terrorism originally meant ‘Government by intimidation’⁴ I think a strong case can be made for supposing that the terms terrorism and terrorist should now only be applied to non-state groups.

I. A Politically Acceptable Conceptual Definition of Terrorism The main challenge that we face in devising a conceptual definition of terrorism is that we need to distinguish between a terrorist and a freedom-fighter. We are all familiar with the claim one man’s terrorist is another man’s freedom-fighter. Following Jenny Teichman, I will refer to this as the ‘dismal slogan’.⁵ The dismal slogan truly is dismal not just in the extent to which it has been uncritically accepted by so many but also insofar as it trivialises important moral issues. Violent civil disobedience is not always impermissible but that obviously does not entail that there is a moral equality between any two acts of violent civil disobedience or that disagreement about the permissibility of a particular act of violent civil disobedience is purely a matter of taste. For the remainder of this paper I will contrast terrorists with freedom-fighters who may be simply defined here as those non-state actors whose use of force is permissible. Now, it is of great importance that we refute the dismal slogan and distinguish between freedom-fighters and terrorists if our conceptual definition of terrorism is to inform coherent and justifiable laws which aim to criminalize terrorist organizations which operate wholly or partly in foreign countries. But we must refute the dismal slogan in such a way that the distinction drawn between terrorists and freedom-fighters is not only morally acceptable but also politically ³ I owe this point to Mark Reiff. ⁴ Oxford English Dictionary 820. Emphasis added. ⁵ J Teichman, ‘How to Define Terrorism’ (1989) 64 Philosophy 507.

Moral Reflections on the Responsibilities of Soldiers

213

acceptable. If a conceptual definition of terrorism is to have more than philosophical interest it must be acceptable to states. It is thus of great importance to have an accurate and politically acceptable definition of terrorism and it is this challenge which I think requires analysis from both lawyers and philosophers. So let us now try to devise a politically acceptable definition of terrorism. Until the international community is made up solely of legitimate states which respect the fundamental rights of the various political communities around the globe, there will be no universal agreement on which particular actions are acts of terrorism. But some signs of consensus are visible. In 1992 A Schmid suggested to the then UN Crime Branch that the existing consensus on what constitutes a war crime is a profitable starting point.⁶ If the core of war crimes—deliberate attacks on civilians and hostage taking—is extended to peacetime, we could simply define acts of terrorism as actions which occur during peacetime which are such that, if they were committed during wartime by uniformed soldiers they would be considered to be war crimes. Let us call this the Schmidian definition.⁷ I think that Schmid is right that most terrorist acts are peacetime equivalents of war crimes. Violent acts such as the events of 9/11, the London, Madrid, Bali, Delhi and Amman bombings, which are normally characterized as terrorist attacks, would have been war crimes if they had been perpetrated by soldiers during a conflict. The deliberate targeting of civilians is a cruel and despicable act and is proscribed under international conventions such as the Rome Statute of the International Criminal Court. However, we cannot simply define terrorist acts in this manner. The Schmidian definition, as it stands is unacceptable as a definition of terrorism for two main reasons: First, states will not accept this definition since they will claim that attacks on political and military sites perpetrated by non-state groups must be considered to be terrorist actions. This makes the Schmidian definition politically unacceptable. Second, the Schmidian definition intuitively seems wrong. If the philosophy department of King’s College London banded together, irate at the British Government’s lack of action in Darfur and set off a bomb at Sandhurst, we would tend to think of this action as a terrorist attack even though if this act were perpetrated by soldiers during wartime it would not be a war crime. Thus, this definition is also not morally acceptable. To fill the gap I think we should extend Schmid’s definition using the notion of a crime against peace. The International Military Tribunal at Nuremberg ⁶ See . ⁷ This is a slight modification of the definition that Schmid proposed to the then UN crime branch. Schmid himself made no reference to whether the soldiers were uniformed. Given that perfidy is often considered to be a war crime it is important that the definition refers to uniformed soldiers.

214

Robert Morris

introduced the term claiming that a crime against peace consisted in the ‘planning, preparation, initiation, or waging of a war of aggression, or a war in violation of international treaties, agreements, or assurances, or participation in a common plan or conspiracy for the accomplishment of any of the foregoing.’⁸ So crimes against peace are the preparation and performance of actions which begin unjust and illegal wars. The use of force is, in general, not permitted in the international system. A crime against peace is a violation of that norm. I think that by using this term we can devise a politically acceptable conceptual definition. A terrorist activity is a violent terrifying act committed by a nonstate actor for political purposes during peacetime which is either: • of such a nature that if it were committed during wartime by uniformed soldiers it would be considered to be a war crime (war crime terrorists) or • of such a nature that if it were committed by a state actor for those reasons and with that level of support it would be considered to be a crime against peace (CAP terrorists) I believe this is a good definition because it captures the fact that terrorism lies, conceptually, in between wartime and peacetime. So we can now distinguish between terrorists and freedom-fighters. Freedomfighters are not war crime terrorists because they do not commit peacetime equivalents of war crimes. Nor are freedom-fighters CAP terrorists. States may go to war without it being a crime against peace if they are exercising their political community’s right of self-defence or its right of political self-determination. So a group which uses violent strikes against military and political targets is a freedom-fighter and not a CAP terrorist if it has the support of a political community (which is large enough to have collective rights) and is fighting against an oppressive regime because that regime is attacking the political community which supports the group or that regime is denying the political community which supports the group its right to political self-determination. This line of argument is profoundly Lockean. This argument accepts Locke’s contention that violent civil disobedience is only permitted in cases where the government is effectively in a state of war with its own people. As Locke claimed, [W]henever the legislators endeavour to take away and destroy the property of the people, or to reduce them to slavery under arbitrary power, they put themselves into a state of war with the people, who are thereupon absolved from any farther obedience.⁹

⁸ ‘Trial of the Major War Criminals before the International Military Tribunal, Nuremberg’ vol XXII (IMT Secretariat, Nuremberg, 1948) 413–4 cited in A Roberts and R Guelff, Documents on the Laws of War (3rd edn, New York, 2003) 177. ⁹ J Locke, ‘Two Treatise of Government’, bk II, ch 19, para 222 in P Laslett (ed), Locke: Two Treatise of Government (student edn, Cambridge, 1988) 412.

Moral Reflections on the Responsibilities of Soldiers

215

II. The Moral Problem Whilst I think this definition, properly understood, would prove to be politically acceptable, people have been reluctant to accept this definition since it appears to imply that there is no moral difference between a CAP terrorist and a soldier who is fighting in a war of aggression. This is problematic for two main reasons: First, many intuitively feel that there is an important moral difference between the actions of terrorists and the actions of soldiers obeying orders. Second, states are often reluctant to ascribe prisoner-of-war status to captured terrorists and want to prosecute individual terrorists under the criminal law. In contrast, captured enemy soldiers are supposed to be treated extremely well under prisoner-of-war conventions and individual soldiers are not considered to be criminally responsible for their participation in crimes against peace. So this conceptual definition can only be considered to be morally and politically acceptable if it can be shown that there is a moral difference between a CAP terrorist and a soldier fighting in a war which is a crime against peace. Now, from a different perspective, this same moral problem has recently entertained a number of applied ethicists working in the philosophy of war. Philosophers such as David Rodin and Jeff McMahan have argued that soldiers should be considered to be morally responsible if they participate in wars which are crimes against peace.¹⁰ Now for ease of exposition I will refer to soldiers fighting in a war which is a crime against peace as unjust soldiers and I will refer to soldiers fighting in a war which is not a crime against peace as just soldiers. Rodin and McMahan reject a traditional tenant of Just War theory. Traditional Just War theory has held that leaders are morally responsible if they initiate wars which are crimes against peace whereas individual soldiers are not morally responsible if they participate in a crime against peace. Traditional Just War theory claims that there is a moral equality between just soldiers and unjust soldiers—I will call this the moral equality thesis. McMahan and Rodin’s objection to the moral equality thesis employs the moral and legal distinction between a justification and an excuse. Excuses accept that the act was impermissible but serve to explain why the agent should not be blamed for his participation in the act. Justifications, however, serve to explain why a prima facie impermissible act is not, in fact, an impermissible act. McMahan and Rodin argue that the moral equality thesis must claim that unjust soldiers are justified in participating in a war which is a crime against peace if it is to be claimed that unjust soldiers are morally equal with just soldiers. But, McMahan and Rodin argue, this cannot be true because the just soldiers have done nothing to forfeit their rights to life. If a man unjustly and viciously attacks ¹⁰ See J McMahan ‘The Ethics of Killing in War’ (2004) 114 Ethics No 4 and D. Rodin, War and Self-Defense (1st edn, New York, 2002) 163–73.

216

Robert Morris

a woman, the woman is allowed to use force to defend herself if necessary. In this case the man is not permitted to use force against the woman because she has a right not to have force used against her. In contrast, the woman is justified in using force against the man so the man must have forfeited his right not to have force used against himself. There is a moral asymmetry between the man and the woman. The fact that the man has forfeited his right not to have force used against him can be explained by the fact that he has viciously and unjustly attacked the woman. So then, A can only be justified in killing B if B has forfeited his right to life. But the just soldiers have done nothing to forfeit their rights to life, so they cannot be justifiably killed by the unjust soldiers. So the unjust soldiers cannot be justified in killing just soldiers. Furthermore, it might seem that the only things which reduce the blameworthiness of unjust soldiers are excuses. We might appeal to ignorance or duress but such pleas are at best excuses rather than justifications. Moreover, not all unjust soldiers will be fully exculpated by such excuses since soldiers tend to fight for a plethora of reasons—some good, some bad. So on this basis McMahan and Rodin reject the moral equality thesis. So the moral problem with the definition I have proposed is: What is the moral diff erence between an unjust soldier and a CAP terrorist? There seem good reasons for supposing that both should be considered to be morally responsible for crimes against peace.

III. Solution to the Moral Problem In what follows I will argue that Traditional Just War theory is right and the moral equality thesis is true. I will further contend that the reason why the moral equality thesis is true also provides us with an explanation of why there is a moral difference between unjust soldiers and CAP terrorists. McMahan and Rodin’s argument turns on its employment of the excuse/justification distinction. This distinction, however, is too rigid to deal satisfactorily with events which consist of more than one agent and more than one act. I would suggest that there are in fact two kinds of case in which A is not blameworthy for his material involvement in the performance of an impermissible act. There are cases where X may be imputed to A but A is not blameworthy because some factor prevented A from acting as a fully-informed rational agent. I think these are excuse cases. But there are also cases where it makes more sense to claim that whilst X is a crime and A was materially involved in bringing X about, X is not A’s crime. I think these cases should be known as cases of exoneration. The central difference between cases of exoneration and excuse cases is that in cases of exoneration it may be quite acceptable for A, in full knowledge that X was a crime, to claim I did nothing wrong. In excuse cases, it would be inappropriate for A to make such a claim.

Moral Reflections on the Responsibilities of Soldiers

217

To illustrate this difference consider the following two cases: (1) Kevin is a member of a group which kidnaps the daughters of rich business tycoons. Kevin’s role in this group is limited to standing guard over the kidnapped individual to make sure that she does not escape. Kevin hates being a part of this group and only became a part of the group and follows orders because the group leader has threatened Kevin that if Kevin does not do what he says then Kevin and his wife and children will be shot. (2) Peter is a state prison guard and is directly responsible for ensuring that prisoners do not escape. Peter’s role is currently confined to making sure that a particular prisoner does not escape. This prisoner is in fact an innocent man who has been found guilty of a heinous crime by a legitimate court and has been sentenced to life imprisonment. It is clear that both Kevin and Peter are materially involved in grave injustices. It would also be strange to blame either Kevin or Peter. Yet we would only think Peter could reasonably claim I have done nothing wrong. Some might think that Peter is justified in imprisoning the innocent man because that is his job, but surely a prison guard is morally justified in imprisoning an inmate, at least in part, because the inmate’s crime makes incarceration a morally appropriate form of punishment. The fact that Peter is just doing his job hardly makes the imprisonment of a man who still has his right to liberty justified. Now whilst Kevin is, at best excused, for his participation in the kidnappings, Peter’s situation is quite different. Peter is not justified in imprisoning the innocent man, precisely because the man is innocent and still retains his right to liberty, but neither would Peter consider that he has, even blamelessly, contravened some duty. Thus, Peter’s case is neither an excusing case nor a justificatory case. It makes better sense to claim that the fact that Peter’s profession is morally worthwhile together with the fact that Peter fulfilled his professional duties exonerates Peter from the injustice that he participated in. He is not excused from the crime because he was just doing his job, nor is he justified because he was just doing his job. The fact that he was doing his job means that the injustice was simply not his crime. The salient difference between the cases of Kevin the Kidnapper and Peter the Prison Guard is that Peter is part of a morally worthy collective whereas Kevin is part of an immoral collective. The institution Peter is a part of is morally worthy since it has a morally worthy function within society—it is good for a society to have institutions which uphold the law. The collective Kevin is a part of, however, has no morally worthy function or role within society; the end of the collective Kevin is a part of (ie gaining ransom money) is morally reprehensible. In terms of establishing culpability and complicity within a collective or an institution, it is of utmost importance first to establish whether or not the collective or institution has a morally worthy role or function within society. Even if a crime is committed

218

Robert Morris

through a collective or an institution, so long as the institution has a morally worthy function within society, then those who fulfil the subordinate roles are not inculpated in the crime. This is not the case with institutions or collectives which lack moral worth. In the cases above, both Kevin and Peter occupy lowly positions within the collectives, but it is the fact that the group of kidnappers is a morally despicable collective which inculpates Kevin in the crime. Thus, a method of establishing the moral equality of soldiers becomes apparent. If it can be shown that the military has a morally worthy function in society then one can argue that so long as unjust soldiers do not transgress their professional duties they should be exonerated from crimes against peace. Now, a brief word on soldiers’ professional duties. Soldiers do not transgress their professional duties so long as they obey all their orders except those which are manifestly unjust. In international law, an order is manifestly unjust if it is clearly a war crime. That a soldier must disobey manifestly unjust orders is the most important professional duty that faces soldiers and provides the limit on what a soldier may do in the heat of battle. An individual does not act as a soldier when he obeys manifestly unjust orders—he acts as a war criminal or perhaps as a desperate man.¹¹ Similarly, in refusing to obey a manifestly unjust order given by a superior officer, a soldier is not engaging in mutiny. As Walzer claims, ‘It is best understood as an appeal up the chain of command over a superior officer to the superiors of that superior officer.’¹² Not obeying manifestly unjust orders thus provides a moral limit to the individual in warfare and provides the conceptual limit on our idea of what a soldier really is. Now whilst we may accept that there is a moral equality between war criminals and war crime terrorists the important issue here is whether there is a moral asymmetry between unjust soldiers who do not obey manifestly unjust orders and CAP terrorists. So let us assume that the unjust soldiers do not obey any manifestly unjust orders. Now we turn to consider whether the military is a morally worthy institution. The principal reasons for considering the military to be a morally worthy institution are that the military deters attacks against the political community and enables states to exercise, in a disciplined and ordered way, their right of selfdefence and (I would also suggest) their right to intervene to prevent crimes against humanity. In an international society which is often unstable and where crimes against peace and humanity are frequent, states need to deter external attacks and sometimes need to exercise their rights to use force in an ordered, disciplined and restrained manner. It is in this light that we see the morally ¹¹ Walzer discusses a case where a soldier knows he will be executed if he does not obey the manifestly unjust order (M Walzer, Just and Unjust Wars (3rd edn, New York, 2000) 314). The desperate situation that the soldier is faced with will be a mitigating factor in assessing his responsibility. It is still the case that the soldier should not carry out the order, but he is to a great extent excused. Disobeying orders in such situations are not supererogatory acts, since one ought to disobey the orders. Rather, doing one’s duty in such a case is an heroic act. ¹² M Walzer, ‘Two Kinds of Military Responsibility’ in M Walzer, Arguing About War (1st edn, Harrisonburg, 2004) 27.

Moral Reflections on the Responsibilities of Soldiers

219

worthy function of the military. In an unstable international society the military is there to uphold the basic rights of communities. This is not to say that the military is always used by governments to fulfil this role but nevertheless this is part of the function of the military. So, if we accept that the military is a morally worthy institution then, as long as soldiers do not obey orders which are clearly war crimes, then they are not inculpated in any crimes against peace which are engaged in by their army. Thus, there is a moral equality between just and unjust soldiers because neither are responsible for participating in a war which lacks a just cause. Rodin and McMahan would both reject this line of argument for two very different reasons. Arguments which have drawn parallels between the military and the penal system have been offered before and have failed to convince either Rodin or McMahan. I shall examine each of their responses in turn. Rodin would reject the argument offered here on the grounds that the military should not be considered to be a morally worthy institution. Rodin has argued that an institution can only be considered morally worthy if it is reliable in reaching and exercising just decisions. Rodin then claims that since a war can only be just on one side, but can be unjust on both sides, a soldier should judge that it is more probable that the military institution, of which he is a part, will be involved in an unjust war than a just war. The soldier should thus conclude that the military is not reliable in reaching and exercising just decisions. Rodin’s argument, however, is mistaken in two respects. First, it is highly questionable whether we should accept Vitoria’s claim that a war can only be just on one side. A war may in fact involve more than one conflict and separate conflicts may have or lack just causes. Second, this argument adopts the wrong calculus. We should not simply focus on all the wars that have been fought, we must also consider all of the wars which have been deterred by the presence of the military. One does not need to be a card-carrying realist to believe that the United Kingdom’s political independence and security from external attack is partly due to the existence of the United Kingdom’s armed forces. McMahan, however, would object to this argument from a different standpoint. McMahan has claimed that there must be a limit on what an institution can demand of us without affecting our culpability. McMahan would claim that whilst individuals might be exonerated from certain crimes by belonging to a morally worthy institution the fact that soldiers are ordered to do such terrible things imposes upon them further duties to make sure that they are not partaking in heinous crimes. McMahan has claimed that there is an important disanalogy between individuals working in the penal system and the military for this very reason. Whilst McMahan’s line of argument has more force if we are restricted to the excuse/justification model, it seems remarkably ad hoc if we accept the category of exoneration. Why are soldiers and not prison guards or policemen the only ones with these duties to deliberate about the justice of their orders? Policemen may

220

Robert Morris

be ordered to kill in certain situations and prison guards are ordered to imprison individuals whose guilt is highly questionable. McMahan’s argument might be stronger if deliberating about the justice of wars was a straightforward measure. However, it is clear that this is not true. It is not true because even if soldiers all had great moral senses, soldiers have a limited epistemic access to the true reasons for going to war for reasons of national security. As Vitoria claimed: ‘A prince is not able, and ought not always to render reasons for the war to his subjects, and if the subjects cannot serve in the war except they are satisfied of its justice, the state would fall into grave peril.’¹³ The military is an incredibly important institution and its members are ordered to undertake extreme measures at great risk. That soldiers are ordered to do terrible things must be understood in the incredibly dangerous environment in which the military must operate. McMahan’s argument fails to recognize the moral worth of the military and thus fails to inculpate soldiers in crimes against peace. So we can now claim that there is a moral equality between just and unjust soldiers on the grounds that neither are responsible for committing crimes against peace. However, CAP terrorists are individually responsible for their participation in such crimes. Such organizations are not morally worthy collectives since their use of force does not have the proper relationship to the rights of a political community. CAP terrorists do not fight on behalf of a political community to defend its members nor do they defend political communities’ rights of political self-determination. CAP terrorist organizations thus lack a morally worthy role in international society. So we have, I think, an accurate, politically acceptable conceptual definition of terrorism which refutes the dismal slogan by distinguishing between terrorists and freedom-fighters. Furthermore, given what we now know about the responsibilities of soldiers we can distinguish between CAP terrorists and unjust soldiers.

IV. The Clues to Devising a Legal Definition of Terrorism Now, the politically acceptable conceptual definition is hardly going to end the discussion. The definition of aggression is highly contested and states and individuals will disagree about which non-state groups are engaged in crimes against peace. But this definition is useful in two respects: First, it frames the discussion which would ensue. By distinguishing between war crime terrorists and CAP terrorists, a great deal of confusion and duplicity can be avoided. Second, it gives us tangible clues as to what should be included in a legal definition of terrorism. ¹³ F Vitoria, ‘On the Laws of War’ reprinted in A Pagden and J Lawrance (eds), Vitoria: Political Writings (1st edn, Cambridge, 1991) 313.

Moral Reflections on the Responsibilities of Soldiers

221

The Terrorism Act 2000 provides a legal definition of terrorism which now informs nearly every piece of counter-terrorism legislation in UK law. Part I s 1 of the Terrorism Act 2000 gives the following definition of terrorism:¹⁴ Terrorism: interpretation (1) In this Act ‘terrorism’ means the use or threat of action where— (a) the action falls within subsection (2), (b) the use or threat is designed to influence the government or to intimidate the public or a section of the public, and (c) the use or threat is made for the purpose of advancing a political, religious or ideological cause. (2) Action falls within this subsection if it— (a) involves serious violence against a person, (b) involves serious damage to property, (c) endangers a person’s life, other than that of the person committing the action, (d) creates a serious risk to the health or safety of the public or a section of the public, or (e) is designed seriously to interfere with or seriously to disrupt an electronic system. (3) The use or threat of action falling within subsection (2) which involves the use of firearms or explosives is terrorism whether or not subsection (1)(b) is satisfied. (4) In this section— (a) ‘action’ includes action outside the United Kingdom, (b) a reference to any person or to property is a reference to any person, or to property, wherever situated, (c) a reference to the public includes a reference to the public of a country other than the United Kingdom, and (d) ‘the government’ means the government of the United Kingdom, of a Part of the United Kingdom or of a country other than the United Kingdom.¹⁵

This is an incredibly broad legal definition of terrorism. This definition applies to all violence against persons and property for religious, ideological or political purposes which is designed to influence any government in the world or intimidate the public of any country. It is troubling to reflect on this definition when one realizes that this is the definition which is used by the Home Secretary to decide which groups should be proscribed organizations and this is the definition which is also used to determine whether people have committed a criminal offence by directly or indirectly encouraging terrorism. The definition used in the Terrorism Act 2000 is clearly unacceptable since it makes no attempt to distinguish between freedom-fighters and terrorists. Charles Clarke, the then Home Secretary, supported the definition of terrorism used in ¹⁴ While it is true that the Terrorism Act 2000 merely seems to offer an ‘interpretation’ of terrorism and does not specifically create a criminal offence of terrorism, this ‘interpretation’ effectively serves as a legal definition of terrorism for UK law. ¹⁵ .

222

Robert Morris

the Terrorism Act 2000 in claiming that he could not ‘think of a state of affairs in the world today where violence would be justified as a means of bringing about change.’¹⁶ I’m not sure many would share his optimism. So how should we improve the legal definition in light of our discussion? First, the politically-acceptable conceptual definition given above gives a morally acceptable principle which the Home Secretary should use in deciding which organizations should be proscribed and under what conditions a previously proscribed organization should be de-proscribed. The second improvement that this conceptual definition suggests is that a legal distinction should be drawn between war crime terrorists and CAP terrorists. In law this could be drawn by using internationally accepted war crimes legislation. Following the 1998 Rome Statute of the International Criminal Court,¹⁷ preparing or instigating the following activities should be characterized as war crime terrorist activities if they were perpetrated by non-state actors for political reasons: (i) Intentionally directing attacks against the civilian population. (ii) Intentionally directing attacks against buildings dedicated to religion, education, art, science or charitable purposes, historic monuments, hospitals and places where the sick and wounded are collected (iii) Torture or inhuman treatment. (iv) Hijacking or the taking of hostages. (v) Intentionally launching an attack in the knowledge that such an attack will incidentally cause large-scale loss of life or injury to civilians. In contrast CAP terrorist acts would be, given the important exception below, the preparing or instigating politically motivated attacks on political and military sites which do not fall under sections (i)–(v). If we recognize in law the distinction between war crime terrorists and CAP terrorists then a method of legally distinguishing between freedom-fighters and terrorists becomes apparent. To distinguish between freedom-fighters and terrorists, a defence to CAP terrorist charges should be available (similar to that found in Article 187A(8) of the Greek Penal Code)¹⁸ which excludes groups who are engaged in violent strikes against political and military targets abroad, where that group is acting on behalf of, and with the support of, a political community ¹⁶ Q21, evidence of the Rt Hon Charles Clarke MP, Secretary of State for the Home Department, to the House of Commons Home Affairs Committee inquiry on the draft Terrorism Bill (HC 515, 11 October 2005). ¹⁷ ‘Rome Statute of the International Criminal Court’, cited A Roberts and R Guelff, Documents on the Laws of War (3rd edn, New York, 2003) 671–97. ¹⁸ ‘Article 187 A(8) of the Greek Penal Code provides that an act shall not constitute a terrorist act: if it is manifested as an effort for the establishment of a democratic government or for the safeguarding or restoration thereof or as an activity in favour or freedom.’ See E Metcalfe, ‘The Definition of “Terrorism” in UK Law: JUSTICE’s Submission to the Review by Lord Carlile of Berriew QC’ (2006) 13–4 .

Moral Reflections on the Responsibilities of Soldiers

223

which is exercising its right of self-defence or political self-determination against an oppressive regime. A defence could be constructed along the following lines: An attack or threat of an attack on political or military sites will not be considered to be an act of CAP terrorism, if, and only if: (a) the attack takes place outside the United Kingdom, (b) and the attack is committed for a purpose wholly connected with the affairs of a country other than the United Kingdom, (c) and does not involve harm or the threat of harm to representatives of the United Kingdom, (d) and the action or threat of an action is undertaken in defence of a political community or a political community’s right of political self-determination. By recognising the two different types of terrorism, our understanding of the problem of terrorism is greatly improved. The problem of defining war crimes terrorism mirrors the problems we face in defining war crimes, and the problem in defining CAP terrorism mirrors the problems we face in defining international aggression. Terrorism and crimes of war are intimately connected and it is for this reason that we must reflect on the differences and similarities between terrorists and soldiers. So, even if people disagree about whether we should adopt some of the suggestions I have made for the UK definition of terrorism, I hope that I have at least shown that thinking about the moral responsibilities of soldiers provides a good insight into the legal problems that surround the definition of terrorism.

12 Criminal Responsibility and Public Reason RA Duff and SE Marshall¹

This paper begins, in section I, with a conception of the criminal trial as the forum in which criminal responsibility is formally and publicly actualized, and with an explication of the idea of responsibility as answerability on which that conception depends. Section II then provides an account of criminal responsibility as a matter of being answerable to our fellow citizens for the public wrongs that we commit, a ‘public’ wrong being a wrong that properly concerns all citizens simply in virtue of their shared membership of the polity. Section III discusses the kind of public reason on which such answerability depends—the common ‘frame of reference within which we are to . . . justify our attitudes, judgements, and actions’,² and without which we could not thus call each other to account: we explore the character and implications of that common ‘frame of reference’ in the context of a liberal democracy and its law, and show how answerability is integral to the very idea of public reason.

I. Being Called to Answer The criminal trial is concerned with the question of whether the defendant can be proved to have committed a criminal offence that he is alleged to have committed. However, a ‘rational reconstruction’ of the criminal trial as it exists in modern liberal legal systems, which seeks to uncover the aims and values internal to this practice,³ will show it to be something other and more than an inquiry which takes the defendant and his conduct as its object, and which aims to determine ¹ This paper is partly based on work done during an AHRC-funded project, The Trial on Trial, in which we worked with Lindsay Farmer and Victor Tadros (see RA Duff, L Farmer, SE Marshall, and V Tadros, The Trial on Trial III: Towards a Normative Theory of the Criminal Trial (Oxford, 2007)); grateful thanks are due to our co-authors, though we cannot blame them for the deficiencies in this paper, and to the AHRC. ² G Postema, ‘Public Practical Reason: An Archeology’ (1995) 12 Social Philosophy & Policy 43, 56. ³ On ‘rational reconstruction’, see DN MacCormick, ‘Reconstruction after Deconstruction: A Response to CLS’ (1990) 10 OJLS 539.

Criminal Responsibility and Public Reason. RA Duff and SE Marshall. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Criminal Responsibility and Public Reason

225

whether he did what he is said to have done and what should now be done to him. It is a process through which a defendant is called to answer, and to which his participation is therefore crucial. He comes to trial not merely as the object of an inquiry, but as a subject, an agent who is called (although he cannot of course be forced) to take part in a rational process of proof and argument: hence the importance attached, in ‘adversarial’ systems of trial, to the defendant’s presence at the trial;⁴ to his entering a plea in response to the indictment;⁵ and to his ‘fitness to plead’—to whether he is capable of understanding the charge and the trial, and of playing his part in it.⁶ We should distinguish two ways in which the defendant is called to answer at the trial. First, she is called to make an answer to the charge. Her answer might be ‘Guilty’: she might admit her guilt, thus admitting that she must answer for the wrong that she committed. This is not to say that defendants who plead ‘Guilty’ typically do so because they recognize and are willing to confess their guilt; such pleas are, all too often, simply the outcome of a process of tactical bargaining of which the defendant was an uncomprehending spectator.⁷ It is rather to say that that is what a plea of ‘Guilty’ purports to mean: it is a performative which constitutes a formal and public confession (a performance which can, like all performances, be sincere or insincere). Her answer to the charge might instead, however, be a plea of ‘Not Guilty’. Given the presumption of innocence, she need formally say no more than that, unless and until the prosecution proves that she committed the offence charged—although she will in fact need to say more than that, if she is to avoid conviction, once the prosecution adduces evidence that would suffice to prove her guilt if it was not rebutted; ‘not guilty’ can therefore be read, not as a declaration of innocence, but as a denial of provable guilt, or a non-assertoric challenge to the prosecution to prove guilt. The indictment addresses the defendant as an agent who is responsible in the sense that she can, and should, respond in one of these ways to the charge.

⁴ For trials on indictment, a defendant who does not appear for trial can be arrested and brought to trial (Bail Act 1976, s 7; see J Sprack, Emmins on Criminal Procedure (9th edn, Oxford, 2002) 300). For summary trials, the defendant’s presence is not required (Magistrates’ Courts Act 1980, ss. 11–13; see Sprack, above 148–52). ⁵ A defendant who refuses to plead is no longer subjected to the ‘peine forte et dure’ (being pressed by heavy weights) to induce a plea—that was abolished in 1772 (see A Mckenzie, ‘ “Th is Death Some Strong and Stout Hearted Man Doth Choose”: The Practice of Peine Forte et Dure in 17th and 18th Century England’ (2005) 23 Law and History Review, ; Singer v US 380 US 24 (1965); the practice seems to have originated in a misreading of the ‘prison forte et dure’ that an earlier statute provided for those ‘mute of malice’). Until 1827, a refusal to plead was treated as a plea of ‘Guilty’, but now a plea of ‘Not Guilty’ is entered on behalf of a defendant who will not plead. ⁶ See Sprack, n. 4 above, 252–3; RA Duff, Trials and Punishments (Cambridge, 1986) 29–35, 119–23. The conception of the criminal trial that we sketch here is developed and defended in detail in Duff et al n 1 above. ⁷ See, eg J Baldwin and M McConville, Negotiated Justice (London, 1977); MD Dubber and MG Kelman, American Criminal Law: Cases, Statutes, and Comments (New York, 2005) 92–101.

226

RA Duff and SE Marshall

Second, if the prosecution proves, or the defendant formally admits, that she committed the offence charged, there is something—the commission of the offence—for which she must now answer in court: such proof, or admission, establishes her criminal responsibility for that offence. A key point to notice here is that this is not yet to establish guilt or criminal liability: a defendant who admits that she committed, or who is proved to have committed, an offence can avoid conviction by proffering a legally recognized defence. She might adduce evidence that, for instance, she acted under duress, in which case the formal burden shifts back to the prosecution, to disprove that defence. In offering such a defence, the defendant implicitly admits criminal responsibility: she admits that she must answer for her commission of the offence—and offers an answer that, she claims, serves to exculpate her. This distinction between the two stages or elements of a criminal trial— between offence, as something that the prosecution must prove ab initio, and defence as something as to which the defendant bears at least an evidential burden—exemplifies the crucial distinction between responsibility, understood as answerability, and liability, understood as liability to conviction and punishment in law, or to moral blame outside the law.⁸ To admit that I am responsible for Φ-ing, or to be held responsible for Φ-ing by others, is to admit that I must answer, or to be called to answer, for Φ-ing. Such admissions or holdings of responsibility are for actions or states of affairs that are in some way untoward: they admit or claim that I acted as I had good reason not to act, and therefore have something to explain.⁹ Now liability presupposes responsibility: I am liable, either to blame or to criminal conviction and punishment, only for that for which I am responsible.¹⁰ Responsibility does not, however, entail liability: I can admit responsibility (criminal and moral) for an action, but avert liability by offering a defence. I might offer a justification: I deliberately broke your window, but that was the only way to help your grandfather, who had collapsed inside the locked house. Or I might offer an excuse: I committed perjury, but only under a kind of duress that, whilst it did not perhaps justify my action, was frightening enough to render me non-culpable for giving in. In both cases, I admit responsibility for my action; I recognize it as something for which ⁸ We assume here that (despite the scepticism exemplified by G Williams, ‘Offences and Defences’ (1982) 2 Legal Studies 233), the distinction between offences and defences is substantive, and not merely terminological or pragmatic: see, eg G Fletcher, Rethinking Criminal Law (Boston, 1978) 552–79, 683–758; K Campbell, ‘Offence and Defence’ in IH Dennis (ed), Criminal Law and Justice (London, 1987) 73; V Tadros, Criminal Responsibility (Oxford, 2005) ch 4. We cannot deal here with the rather different logic of ‘civil’ liability, either within or outside the law, understood as liability to pay the costs of harm that has been caused. ⁹ We are also of course responsible for the good and the right that we do, as well as for the untoward: but we cannot explore the significant asymmetries between these two types of responsibility here. ¹⁰ Criminal liability might be imposed without moral responsibility, as when I am held vicariously liable for the actions of others; but it still presupposes criminal responsibility: I am vicariously liable only if the law holds me vicariously responsible.

Criminal Responsibility and Public Reason

227

I must answer. But I offer an exculpatory answer that, if acceptable, blocks the transition from responsibility to liability, thus averting moral blame or criminal conviction and punishment. If we understand responsibility as answerability,¹¹ we must also attend to its relational aspects. (There are of course non-relational ideas of responsibility. We talk of responsible agency and, in commendatory terms, of responsible parents or teachers or citizens. But such ideas can be explained in terms of relational responsibility. Responsible agency is a matter of having the capacities that are necessary if one is to be answerable for one’s actions: hence the attraction of accounts of responsibility as reasons-responsiveness;¹² a responsible agent is one who is capable of responding to the reasons that bear on her actions, and of answering for her actions by reference to those reasons and to her reasons for acting thus. A responsible parent or teacher is one who takes her prospective responsibilities seriously, and discharges them conscientiously: she pays due attention to, and takes due care for, the matters for which she is responsible, and is therefore well placed to answer for her actions, ie to accept retrospective responsibility in relation to those matters.)¹³ It is commonplace to point out that responsibility involves a dyadic relationship: an agent is responsible for something. But the relational conception of responsibility that concerns us here is not merely dyadic, but triadic: I am responsible for X, to S—to some person or body who has the standing to call me to answer for X. Furthermore, I am responsible for X to S as Φ—in virtue of satisfying some normatively laden description, which makes me responsible (both prospectively and retrospectively) for X to S. The ‘for X’ dimension of this triadic relationship is familiar: I can be held responsible for a wide range of ‘objects of responsibility’, including actions, omissions, thoughts, feelings, and states of affairs. One important question that we cannot pursue here concerns the general conditions of responsibility: if I am to be properly held responsible for X, whatever X might be, must X have been in some sense within my control; must it be true that I knew or could have known that and how I could affect X?¹⁴ We would simply note here that such general conditions set only weak constraints on the objects of responsibility. The control condition is a condition of responsibility, but is satisfied so long as the occurrence or non-occurrence of X depended substantially on what I did. The epistemic condition is not even a general condition of responsibility: I can be, and in moral ¹¹ See JR Lucas, Responsibility (Oxford, 1993); G Watson, ‘Reasons and Responsibility’ (2001) 111 Ethics 374; contrast V Tadros, Criminal Responsibility (Oxford, 2005) 24–31. ¹² See, eg RJ Wallace, Responsibility and the Moral Sentiments (Cambridge, MA, 1994); JM Fischer and M Ravizza, Responsibility and Control (Cambridge, 1998). ¹³ Compare HLA Hart, Punishment and Responsibility (Oxford, 1968) 212–4, on ‘roleresponsibility’. ¹⁴ On control, see D Husak, ‘Does Criminal Liability Require an Act?’ in RA Duff (ed), Philosophy and the Criminal Law (Cambridge, 1998) 60; also Fischer and Ravizza n 12 above. On the epistemic condition, see J. Feinberg Harm to Self (Oxford,1986) 269–315.

228

RA Duff and SE Marshall

matters I often am, held responsible for bringing X about when I did not realize and could not reasonably have been expected to realize that my action might have that result; such epistemic conditions are often conditions of liability rather than responsibility.¹⁵ Our main concern now, however, is with the ways in which the objects and directions of responsibility differ between different contexts: how I can be responsible as Φ for X, but not for Y (although the control and epistemic condition are satisfied in relation to both X and Y ); or responsible to S for X, but not for Y; or responsible for X to S, but not to T. To illustrate. We can specify my responsibilities as a university teacher. That will involve in part an account of my prospective responsibilities: it is my responsibility to keep up to date with my subject, plan courses that contribute suitably to the curriculum, prepare and conduct my classes in pedagogically effective ways, mark students’ work, and so on. We can disagree about what falls, or should fall, within my responsibilities as a teacher, but any account will set some limits to them: as a teacher in a secular institution, for instance, it is no part of my responsibility to attend church services; as a philosophy teacher it is not my responsibility to play football on Saturday for a local amateur team—though I may have such responsibilities as a member of a church or as a member of that local football team. To say that these are my prospective responsibilities as a teacher is to say that I may be called, and must be ready, to answer as a teacher for my conduct in relation to these matters—for my discharge of (or my failure to discharge) these responsibilities. A full specification of my responsibilities must also specify the people or the bodies to whom I am responsible as a teacher—and by implication those to whom I am not responsible as a teacher. I am responsible as a teacher to my students, to my colleagues, to my employer: they have the right to call me to account for the ways in which I discharge or fail to discharge my pedagogical responsibilities; they can, for instance, call me to answer for failing to turn up to a class, or for giving an ill-prepared lecture. But, first, I am not responsible as a teacher to a passing stranger, or to my aunt, or to my fellow footballers: they have no standing to call me to account for missing the class or for giving a bad lecture; if they challenged me about it, or demanded that I answer to them for it, I could reply that it was not their business. Second, I am not responsible to my students, my colleagues (qua academic colleagues) or my employer for my conduct as a member of a church or football team. There will be someone to whom I am responsible, as a member of the church, for my religious beliefs and practices: depending on the character of my religion, it might be my priest, fellow members of my congregation, or only God. There will be people to whom I am responsible as a member of the football team—other members, our supporters if we have any: depending on the kind of team it is, I might be responsible to them not just for how I play (and whether ¹⁵ See further RA Duff, ‘Who Is Responsible, for What, to Whom?’ (2005) 2 Ohio State Journal of Criminal Law 441, 452–6.

Criminal Responsibility and Public Reason

229

I turn up), but for keeping fit, and taking part in training. But I am not responsible to my fellow footballers for my performance as a teacher, or to my academic colleagues for my performance as a footballer. The responsibility-laden descriptions (teacher, parent, friend, colleague, member of this church or of that team) that help to structure our lives determine the content and the direction of our responsibilities: they determine what we are responsible for, to whom (or what). Such determinations extend our responsibilities: as a member of a team I have responsibilities that I would not otherwise have, and am responsible to people to whom I would not otherwise be answerable. But they also limit our responsibilities: there are matters within my control and knowledge, even matters that bear on my students’ welfare, for which I might deny that I am responsible as a teacher (it is not my responsibility, I might say, to offer them advice on their sexual relationships, or to ensure that they attend classes); and there is only a limited range of people or bodies to whom I am responsible as a teacher.¹⁶ Claims of the form ‘A is responsible for X’ are therefore incomplete: they must be filled out by specifying as what, and to whom, A is responsible for X. That specification might not need to be explicit, if it is obvious from the context; but it must be available.¹⁷ Such claims can therefore only be made within a normgoverned practice, whose secondary norms govern the ways in which participants are held responsible for their conduct in relation to its primary norms; and they require a language in terms of which participants can be called to answer and can answer—a normative language which they can speak in the first person as their own. (We will see the importance of this point in section III).

II. Criminal Responsibility and Citizenship If we turn now from responsibility in general to criminal responsibility in particular, and ask as regards for what we are in fact criminally responsible, under existing positive law, the simplest answer is territorial: we are criminally responsible under a state’s laws as agents who act within that state’s geographical territory. The basic jurisdiction of English criminal law is defined by the principle of ¹⁶ We may disagree about what I am responsible for, and to whom, as a teacher: but nothing in our argument depends on claiming that the content or direction of our responsibilities can be uncontroversially specified. ¹⁷ Contrast J Gardner, ‘The Mark of Responsibility’ (2003) 23 OJLS 157, 165; in response, see RA Duff, ‘Answering for Crime’ (2005) 106 Proceedings of the Aristotelian Society 85, 88–9. We can distinguish a strong from a weak thesis in what we say. The strong thesis is that responsibility is always to some person or body; if we cannot say to whom or what A is responsible for X, we cannot say that A is responsible for X. The weaker thesis is that if A is responsible for X, we can always properly ask to whom A is responsible, and can usually expect an answer; but in some (abnormal) cases there might be no answer—not even ‘only to herself ’. Whatever the merits of the strong thesis, what follows depends only on the weaker thesis.

230

RA Duff and SE Marshall

territoriality, and similar provisions are found in other legal systems.¹⁸ But that principle cannot help us determine as what and to whom we should be criminally responsible. It might, subject to certain qualifications, capture the extension of criminal responsibility, but not in a normatively illuminating way: ‘acting within a specified geographical area’ does not have the normative significance that an answer to the ‘as what’ question requires. We are to be answerable under a system of criminal law: but a system of law, and a state within which a system could exist, require something more than a collection of people who happen to live or act in the same geographical area; without a lot more than that, there would indeed be nothing recognizable as a human society, let alone a political or legal system.¹⁹ That ‘more’ requires an idea of community: not necessarily a richly normative communitarian idea of the sort that eschews liberal individualism (and that worries liberal individualists), but a metaphysical idea of the sort that even liberal individualists must presuppose;²⁰ an idea of people living together (as distinct from merely beside each other) in a society defined by a set of shared values and understandings which might be implicit, inchoate or disputed,²¹ but without which society, politics and law would be impossible. A minimalist idea of political and legal community is provided by the various versions of legal positivism, and offers a slightly more substantive—though still normatively inadequate—answer to the ‘as what and to whom’ question. According to classical legal positivism, of the Austinian-Benthamite variety, the relevant community consists of those who habitually obey an identifiable sovereign, and its criminal law consists of those authoritative commands from the sovereign that are backed up by the threat of sanctions. Members of that community are therefore criminally responsible to the sovereign as her (or its) subjects: their prospective responsibility is to obey those commands, and the sovereign is the person or body with the standing to call them to account for alleged acts of disobedience to those commands. This is no doubt how the criminal law is (reasonably) understood by many who appear in our courts: as a set of peremptory demands, which they are punished for disobeying. But it is not how criminal law should address the members of what aspires to be a liberal democracy: it should not speak to them in the alien voice ¹⁸ See M Hirst, Jurisdiction and the Ambit of the Criminal Law (Oxford, 2003) ch 1; U Ebert, Strafrecht Allgemeiner Teil 2nd ed, (Heidelberg, 1994) 8; WR LaFave, Criminal Law (3rd edn, St Paul, Minn, 2000) 123–43;. The exceptions and qualifications to this general principle need not concern us here, although we mention some below. ¹⁹ See P Winch, ‘Nature and Convention’ (1960) 60 Proceedings of the Aristotelian Society 23. ²⁰ On this metaphysical-normative distinction, see C Taylor, ‘Cross-Purposes: The LiberalCommunitarian Debate’ in N Rosenblum (ed), Liberalism and the Moral Life (Cambridge, MA, 1989) 159; S Mulhall and A Swift, Liberals and Communitarians (Oxford, 1992) chs 5–6. ²¹ And which might have as much to do with decision procedures as with the content of those decisions, if proceduralist versions of liberalism are plausible: see, eg J Waldron, Law and Disagreement (Oxford, 1999); D Archard, ‘Political Reasonableness’ (2005) 35 Canadian Journal of Philosophy 1.

Criminal Responsibility and Public Reason

231

of a sovereign who claims authority over them, and whose commands give them content-independent reasons for action (as if the criminal wrongfulness of murder or of rape consisted in disobedience to the law that prohibits them). We are subject to the law: we are, supposedly, bound by its demands, and answerable in its courts for our crimes. But the law by which we are bound should not be one imposed on us by a sovereign: it should be a ‘common law’—a law which is our law, which speaks to us in our collective voice in terms of the values by which we define ourselves as a polity; a law by which we bind ourselves.²² The beginning of a better answer to our ‘As what, to whom’ question can be found if we turn from positivism to legal moralism. Moore offers a simple version of legal moralism: the criminal law’s function is to do ‘retributive justice’ by punishing ‘all and only those who are morally culpable in the doing of some morally wrongful act.’²³ The familiar objection to legal moralism is that it is overinclusive. It implies that we have good reason to criminalize every kind of moral wrongdoing, even if other factors often tip the balance against doing so:²⁴ but we surely have no reason to criminalize, for instance, the cruelly hurtful ending of a love affair, or the persistent and arrogant dismissal of a colleague’s ideas, wrongful though such conduct is.²⁵ However, there is a core of truth in legal moralism: the criminal law is properly concerned with moral wrongs. ‘Moral wrongdoing’ is the start of an answer to the question ‘For what should we be criminally responsible?’; the criminal law should aim to identify and declare the public wrongfulness of certain kinds of moral wrongdoing, and provide a suitable public response to them.²⁶ The point is not merely that wrongfulness is a necessary condition of criminalization: many who reject legal moralism accept such a constraint, as a condition of deserved punishment.²⁷ The point is that conduct that is properly criminalizable should be criminalized because it is wrongful: we should

²² On the classical idea of the common law as the law of the political community, see G Postema, Bentham and the Common Law Tradition (Oxford, 1986) chs 1–2; for a contemporary democratic development of that idea, see R Cotterrell, Law’s Community (Oxford, 1995) ch 11. ²³ MS Moore, Placing Blame: A General Theory of the Criminal Law (Oxford, 1997) 33–5. ²⁴ See Moore n 23 above, chs 1, 5, 18; also Moore, Act and Crime: The Philosophy of Action and its Implications for the Criminal Law (Oxford, 1993) chs 2–3. For Moore, only wrongdoing is (normally) criminalizable. ²⁵ It might be objected that legal moralism is also under-inclusive, because it cannot justify mala prohibita. This objection can be met, by showing how conduct that is not wrongful prior to its legal regulation can become wrongful as a breach of justified legal regulation: see SP Green, ‘Why it’s a Crime to Tear the Tag off a Mattress: Over-Criminalization and the Moral Content of Regulatory Offenses’ (1997) 46 Emory Law Journal 1533; RA Duff, ‘Crime, Prohibition and Punishment’ (2002) 19 Journal of Applied Philosophy 97; for criticism see D Husak, ‘Malum Prohibitum and Retributivism’ in RA Duff and SP Green (eds), Defining Crimes (Oxford, 2005) 65. ²⁶ In saying this we do not commit ourselves to Moore’s particular brand of metaphysical realism, which portrays criminal law as a ‘functional kind’. ²⁷ See, eg D Husak n 25 above; ‘Limitations on Criminalization and the General Part of the Criminal Law’ in S Shute and AP Simester (eds), Criminal Law Theory: Doctrines of the General Part (Oxford, 2002), 13.

232

RA Duff and SE Marshall

criminalize it in order to ensure that it attracts the kind of formal response, of condemnation and punishment, that it deserves.²⁸ The central error in Moore’s legal moralism is what generates the familiar objection to it: it makes any and every kind of moral wrongdoing in principle the business of the criminal law. What grounds that objection is the answer to the ‘as what, to whom’ question that simple versions of legal moralism imply. It suggests that we are criminally (just as we are morally) responsible simply as moral agents, since it is as moral agents that we commit culpable moral wrongs; and this also suggests that we are criminally (as we are morally) answerable to other moral agents for such wrongdoing. Now it is not even true of moral wrongdoing as such that we are responsible simply as moral agents, to all other moral agents: although many serious moral wrongs might be the business of all moral agents as such, many wrongs are the proper business only of the smaller moral communities within which they are committed. It is even more obvious that we should not be criminally responsible simply as moral agents, to other moral agents. The familiar liberal argument for this is that some types of moral wrongdoing, such as my bad behaviour towards my lover or my colleague (wrongful though it is), is not the criminal law’s concern: I must answer for it to my friends or colleagues, to my conscience or to God; but such wrongdoing ‘is, in brief and crude terms, not the law’s business’.²⁹ A less familiar argument appeals to jurisdiction. If we are criminally responsible simply as moral agents, if the function of criminal law is to provide for the retributive punishment of ‘all and only those who are morally culpable in the doing of some morally wrongful act’, then English law has reason to criminalize theft committed not merely in England, or by or against English citizens, but anywhere by or against anyone: to make it a crime under English law for a Polish citizen to steal from a fellow Pole in Poland, for instance, and to give English courts the authority to try, convict and sentence such a thief should they get the opportunity to do so. A Polish thief would, however, reasonably object that his wrongdoing, culpable though it is, is not the business of English law or English courts: he is answerable for it in Poland, but not in England. This is a second way in which simple legal moralism is radically over-inclusive: it pays no attention to, and provides no normative basis for, such jurisdictional limits to the criminal law of a nation state. A legal moralist might argue that a system of criminal law has, in principle, a legitimate claim to jurisdiction over culpable moral wrongdoing wherever and by and against whomever it is committed; but that there are in practice good reasons to limit that claim in the ways that the jurisdictional claims of national ²⁸ See Duff n 25 above, and Punishment, Communication, and Community (New York, 2001) 56–68. ²⁹ The Wolfenden Committee’s trenchant words on consensual sexual activity between adult homosexuals (Sir John Wolfenden, Report of the Committee on Homosexual Off ences and Prostitution (Cmnd 247; London, 1957) para 61).

Criminal Responsibility and Public Reason

233

systems are typically limited. There are, first, pragmatic reasons for maintaining a division of legislative, adjudicative and punitive labour: since we have a structure of nation states, it makes practical sense for each of them to exercise criminal jurisdiction over wrongs within its reach—those committed within its territorial boundaries. Second, states should respect other states’ sovereignty, and therefore should generally leave the prosecution of wrongdoers to the state in whose territory they commit their wrongs.³⁰ I have some reason to intervene if I see a child misbehaving—to tell the child off, and to visit on her misconduct the censure it deserves; but out of respect for the child’s parents’ authority, I should normally leave it to them to deal with the child. So too, out of respect for sovereignty, states should normally leave it to other states to deal with crimes committed by their citizens within their territory.³¹ On this view, a national legislature which takes its responsibilities to punish wrongdoing seriously will begin with a provisional claim to universal jurisdiction—it should criminalize, and seek to punish, moral wrongdoing wherever, by and against whomever, it is committed. However, as it began to move from the question of what it had in principle good reason to do to the question of what it should all things considered do, it would see better reasons to limit its jurisdictional claims. But that is an implausibly imperialistic view of the responsibilities of a national legislature. When I learn about a stranger’s moral misconduct towards her friend or her parents, I do not think it my business to intervene, to call her to answer for what she has done: the logic of my view is not that I have some reason to call her to account, since she and I are both moral agents, but better reason not to interfere; it is ab initio that her wrongdoing is not my business. Analogously, legislatures should not begin with the idea that they have good reason to criminalize all moral wrongdoing, and then see reasons to limit their jurisdictional ambitions; they should, rather, begin with the idea that only a certain range of wrongdoings is even in principle their business. How can we begin to identify that range? The obvious answer, for any liberal democrat, is that we should replace ‘moral agent’ by ‘citizen’ as the description under which we should be criminally responsible—to say that we are responsible as citizens, to our fellow citizens. In a liberal democracy the legislature, and the whole apparatus of the state, supposedly speak and act in the name, and on behalf, of the citizens who make up the polity: they are not the organs of a sovereign whose subjects we are; they are the institutional manifestations and instruments of our shared political lives—of the civic enterprise in which we are collectively engaged (just as the institutional structures of authority in a properly functioning university are manifestations or instruments of the shared academic enterprise in which the university’s members ³⁰ See Hirst n 18 above, 10–11, on the ‘unwarranted usurpation of [another state’s] sovereignty’. ³¹ The legal moralist might also cite the ways in which nation states do sometimes claim a universal jurisdiction over certain types of wrongdoing: see at nn 34–6 below.

234

RA Duff and SE Marshall

are collectively engaged). That is what makes the law, and in particular the criminal law, a common law—a law that is our law as citizens: its voice should not be the voice of a sovereign who demands our obedience as subjects, but our own collective civic voice; a voice in which we speak to ourselves, as citizens, about the shared values and goals by which our civic enterprise as a polity is constituted. That is also to say, however, that it is a voice in which we speak to ourselves, rather than to the whole world of moral agents at large. The criminal law speaks of wrongs: wrongs that are public in that they concern the whole political community, and require a public response of condemnation and punishment.³² The wrongs that are paradigmatically public in this sense are those committed within a polity by its own members: our criminal law underpins our lives as citizens; its proper concern is with the wrongs that we commit within that civic enterprise. More precisely, its primary purpose is to specify those wrongs for which citizens must answer to the polity (to their fellow citizens) through its courts: its interest is not in moral wrongdoing as such, but in kinds of wrong that properly concern the members of the polity in their role as members, for which they therefore have the right (indeed duty) to call the perpetrator to account. As against classical positivists, the liberal republican’s argument is that we are criminally responsible not as subjects, but as citizens; not to a separate sovereign, but to ourselves, ie to each other. As against ambitious legal moralists, the argument is that the criminal law is the law of a particular polity, and is therefore concerned with what is internal to that polity—primarily, that is, with the wrongs committed between citizens of that polity. That is why criminal cases are properly labelled not (as in England) ‘R v D’, but (as in American states) ‘People v D’, or ‘Commonwealth v D’ (or even ‘State v D’, so long as the state is presented as the agent of the people themselves, and as calling D to answer in the name and on behalf of the people). This is not to say that only citizens should be criminally responsible under a polity’s law: the criminal law of a decent polity covers visitors and temporary residents as well as citizens. Such visitors should, as guests, be accorded many of the protections of citizenship, as well as being expected to accept many of its duties and responsibilities. In particular, they should be both bound and protected by the polity’s law, including its criminal law. If they commit what the local law defines as a public wrong, they must answer for that wrong to the polity whose law it is—just as someone who commits such a wrong against them must answer for it as he would for wrongs committed against fellow citizens. To say this is not to revive a territorial account of jurisdiction; what makes normative sense of jurisdiction is still the law’s character as the law of a particular polity, whose members are its primary addressees. But given such a polity, whose members are ³² On crimes as ‘public’ wrongs, see classically Sir William Blackstone, Commentaries on the Laws of England (London, 1765–69) IV ch 1; on this sense of ‘public’, see SE Marshall and RA Duff, ‘Criminalization and Sharing Wrongs’ (1998) 11 Canadian Journal of Law & Jurisprudence 7.

Criminal Responsibility and Public Reason

235

responsible to each other for what their law identifies as public wrongs, its laws can also bind and protect visitors to the polity and its territory. (We would emphasize that in thus relativizing responsibility to particular polities, we are not relativizing wrongfulness. We can be as non-relativistically universalistic as we like about morality, and insist that what is wrongful in England is also wrongful in America, or Poland, or Korea. The point here concerns not what is or is not wrong, but to whom we must answer for the wrongs we commit: that in a liberal democracy we are criminally responsible to our fellow citizens. Different systems of criminal law also differ, of course, in their content: what is a criminal wrong in one system might not be criminal in another system. Such differences need not—though they sometimes do—reflect different underlying views of what is wrong: they might reflect different views about which wrongs are ‘public’, or about the interpretation and definition of such public wrongs; or different systems of regulation, generating different mala prohibita. For English criminal law not to claim jurisdiction over a theft committed by a Polish citizen in Poland is not for it to imply that such a theft is not a wrong; it is, rather, to remain properly silent about something that is not its business.) By grounding the criminal law’s jurisdiction in citizenship rather than in territory, we can also explain why a polity might claim jurisdiction over wrongs committed by, or against, its citizens abroad.³³ In claiming jurisdiction it claims that anyone who wrongs its citizens must answer to the polity, and that any member of the polity is responsible to the polity for such public wrongs as he commits anywhere. Such claims rest on conceptions of the scope of the bonds of citizenship that may be disputed: but there is nothing puzzling in a claim that I must answer to my fellow citizens for wrongs that I commit abroad, since I do not leave my status as a citizen behind when I go abroad; or in a claim that as citizens we have a proper interest in any wrongs done to our fellow citizens, and the standing to call the wrongdoer to answer for them. This is not to say that a polity should claim criminal jurisdiction over all crimes by or against its citizens wherever they are committed anywhere: one could instead argue that we are not answerable to our fellow citizens for wrongs we commit abroad, and that whilst our fellow citizens should have some collective concern for wrongs that we suffer abroad, that concern does not properly extend to calling the wrongdoers to account. The point here is simply that such claims to extra-territorial jurisdiction make straightforward sense if we take citizenship to be the basis of criminal responsibility, whereas if we start with the territorial criterion of jurisdiction they will seem much more puzzling. Polities do also claim criminal jurisdiction of a more universal kind, over certain crimes committed anywhere, by anyone, against anyone.³⁴ We cannot ³³ For different provisions of this kind, see eg the French Code Penal, art 113.6–7; the German Strafgesetzbuch, s 7.1–2; the English Sex Offenders Act 1997, s 7; Hirst n 18 above, 5. ³⁴ See, eg Criminal Justice Act 1988, s 134, giving effect to the UN Convention against Torture; see R v Bow Street Metropolitan Stipendiary Magistrate ex parte Pinochet Ugarte [2000] 1 AC 147. For other examples, see A Cassese, International Criminal Law (Oxford, 2003) 277–322.

236

RA Duff and SE Marshall

discuss such claims, and the extent to which they should survive the creation of the International Criminal Court,³⁵ here: we would simply suggest, first, that they are plausible only in relation to the most terrible of crimes; and second, that they still depend on a claim that there is a community to which the perpetrator must answer, in whose name the court is acting—though it is not clear whether we should see that community as the polity within which the crimes were committed, on whose behalf another state’s courts now claim to act, or as ‘humanity’ itself.³⁶ If the claim that in a liberal democracy we are, paradigmatically, criminally responsible as citizens, to our fellow citizens, is to be more than a rhetorical slogan, it must of course be fleshed out by an account of democratic citizenship, and of the civic enterprise in which we are engaged qua citizens: what is it, normatively speaking, to be a citizen, and to recognize others as one’s fellow citizens? We cannot attempt that task, which is the task of explicating a theory of liberal democracy, here;³⁷ we must instead say something about the kind of ‘public reason’ to which such an account of criminal responsibility must be able to appeal.

III. The Sphere of Public Reason As we noted above, we understand crimes as ‘public’ wrongs in the sense that wrongs which are to be criminalized should be wrongs which properly concern all members of the polity, in virtue simply of their citizenship: they are public not in the sense that they directly impinge on or harm ‘the public’, but in the sense that they are the business of ‘the public’, ie of the citizen body as a whole. We have also argued elsewhere that wrongs which are in this sense public are wrongs that are shared by all citizens. As citizens of a polity, we should recognize our shared civic identity and the shared values that structure it, and recognize each other as fellow citizens to whom we owe concern and respect; an attack on one of our number is then a attack on ‘us’—on the shared values, and the conception of a common civic good, by which we define ourselves as a polity.³⁸ It is also worth emphasizing, given the readiness with which the rhetoric of crime is used to portray offenders as outsiders or enemies, that when citizens commit crimes against ³⁵ Rome Statute of the International Criminal Court, arts 1, 5; see Cassese n 34 above, 340– 405 (and 327–40 on the earlier ad hoc international tribunals). ³⁶ See A Altman and CH Wellman, ‘A Defense of International Criminal Law’ (2004) 115 Ethics 35; RA Duff, ‘Can We Punish the Perpetrators of Atrocities?’ in T Cushman and T Brudholm (eds), The Religious in Responses to Mass Atrocities (forthcoming). ³⁷ See R Dagger, Civic Virtues: Rights, Citizenship, and Republican Liberalism (Oxford, 1997) for one account of the kind of republican conception to which we would appeal. ³⁸ See Marshall and Duff n 32 above. We should emphasize that although our account is therefore in an obvious sense ‘communitarian’, it is a firmly liberal communitarianism: the civic identity that we share as citizens is only one of the identities that structure our lives—and a limited one at that; the values that we share as citizens are limited in their scope and their claims. See further, Duff, Punishment n 28 above, ch 2.

Criminal Responsibility and Public Reason

237

citizens, we must recognize both victim and offender as fellows: the wrong that we share is a wrong committed by one of us against one of us. As citizens, we therefore have two kinds of appropriate interest in the criminal law. First, we have an interest in its content: not just because it impinges on our lives, by constraining our own conduct and by protecting against the conduct of others, but because it purports to express values that are our values as members of the polity. Second, we have an interest in its application: not just because we have an interest in seeing it effectively enforced, but because someone who commits a wrong in which we share is answerable to us for that wrong. If we take seriously the values to which we are supposedly committed, and which the criminal law (our criminal law) supposedly expresses, we cannot remain collectively indifferent to their violation: a response is called for. One kind of response focuses on the victim (when there is one): we respond to her not simply as someone who has been harmed (as we might respond to someone who has suffered a natural calamity), but as someone who has been wronged;³⁹ hence the importance of various kinds of ‘victim support’, and the legitimate basis for the idea that a special kind of compensation might be appropriate for criminal injuries.⁴⁰ But we must also focus on the wrongdoer. This is partly a matter of what we owe to the victim—to try to bring the person who wronged him to account (this is a further aspect of the difference between harms and wrongs);⁴¹ but we owe this to the wrongdoer as well, as a fellow citizen. If we are to treat him with the respect due to him as a responsible citizen, we must be ready to hold him responsible for his actions—for the wrongs that he commits. We do not just ignore what he has done, as if it was either insignificant or not our business; nor do we just coerce him, either by way of revenge or to prevent further crimes by him or by others: we call him to answer for what he has done. We can see the practice of jury trials (infrequent though they in fact are), and other kinds of lay participation in the criminal process (lay magistrates or judges),⁴² as expressions of this public interest in calling wrongdoers to account: the offender must answer to us, to his fellow citizens, for the public wrong that ³⁹ Advocates of ‘restorative justice’ who urge us to focus on harm and its repair rather than on wrongs and their condemnation fail to do justice to this point: see, eg L Hulsman, ‘Critical Criminology and the Concept of Crime’ (1986) 10 Contemporary Crises 63; L Walgrave, ‘Imposing Retribution Instead of Inflicting Pain’ in A von Hirsch et al (eds), Restorative Justice and Criminal Justice (Oxford, 2003) 61. ⁴⁰ Which is not to defend the particular provisions for financial compensation for injuries caused by crime that we have in England, on which see AJ Ashworth, ‘Punishment and Compensation: Victims, Offenders and the State’ (1986) 6 OJLS 86. ⁴¹ A further question, which we cannot pursue here, concerns what (if anything) the victim owes us: what responsibilities victims should have, as citizens, to take part in the process of calling offenders to account; see SE Marshall, ‘Victims of Crime: Their Station and its Duties’ in M Matravers (ed), Managing Modernity: Politics and the Culture of Control (London, 2004) 104. ⁴² See, generally, T Hörnle, ‘Democratic Accountability and Lay Participation in Criminal Trials’ in RA Duff, L Farmer, SE Marshall and V Tadros (eds), The Trial on Trial II: Judgement and Calling to Account (Oxford, 2006) 135. On jury service as a civic duty see, SJ Clark, ‘The Courage of Our Convictions’ (1999) 97 Michigan Law Review 2381.

238

RA Duff and SE Marshall

he committed; it is therefore appropriate that we, his fellow citizens, be involved in his trial. That first interest, in the law’s content, is reflected in the democratic claim that if we ‘are the subjects of the law, and if the law is to bind [us] as free men and women, [we] must also be its makers.’⁴³ This is not the place for a discussion of the nature of democracy, or what renders law properly democratic, ie made in an appropriate sense by ‘the people’. All we would say here is that if the law is to be, as it is supposed to be, an institution that protects rights and serves the ends of justice, we cannot plausibly understand the idea of democratic law in terms of any simple majoritarian conception of democracy, since majorities are not of their nature oriented towards justice or the protection of rights. Participatory and deliberative conceptions of democracy are more promising for our purposes. The legitimacy of the law can then be grounded not merely in its majoritarian credentials, but in the process of public deliberation and debate from which it emerged—a process that can itself be legitimate (and thus can ground the law’s legitimacy) only if it is structured by a due respect for the standing of all the citizens who are to be participants in it. Such a respect, for citizens as responsible agents who can take part in such an enterprise of self-government, will constrain the possible content of the criminal law in ways that simple majoritarianism cannot, since a collective and mutually respectful deliberation must generate laws that themselves express that respect and protect the rights that it recognizes. Our main concern here is with the second kind of interest that citizens have in the law—in its application, and in particular in the calling to account of those who commit wrongs that have been legitimately declared to be public. What we want to explore, however, is the way in which these two interests are closely connected. If we see ourselves, democratically, as fellow participants in a civic enterprise of self-government, of making laws for ourselves, we must also see ourselves as answerable to each other; and whilst such answerability need not, as a matter of logic, take the form of criminal responsibility, it underpins the claim that we are criminally responsible to each other as citizens, and the importance that we have been attaching to responsibility as answerability. To make clearer sense of these remarks, we can begin with the idea (familiar in various versions, whose differences need not concern us here, from the work of philosophers such as Wittgenstein and Habermas) that reason or reasoning is itself essentially a public matter. Our mode of relating to the environment that includes other human subjects is not to regard them merely as parameters of our individual choice and decision, but as participants in a common social life. Thus, for this, common rules of thought are necessary. . . . [H]uman thinking is essentially a social ability, depending on social interaction, on a first-person plural orientation toward the world, and of rules of thought which can be common property among us. The public dimension of thought demands that we transcend our ⁴³ M Walzer, ‘Philosophy and Democracy’ (1981) 9 Political Theory 379, 383.

Criminal Responsibility and Public Reason

239

immediate subjective responses without abandoning them. . . . From this common project emerges a common voice or point of view which enables each subject to judge his or her own attempts at rule-following as successes or mistakes, as correct or incorrect.⁴⁴

Thinking (and acting) thus presuppose ‘community’ in the sense of participation in a set of common practices—in a common life; such participation requires a shared language, and, as Wittgenstein pointed out, a shared language requires ‘agreement not only in definitions but also . . . in judgements’.⁴⁵ However diversely multi-cultural a society might be, however deeply its inhabitants disagree about values and conceptions of the good, it can count as a society (and its inhabitants can count as disagreeing, rather than talking past each other in different languages) only if there is at some level agreement in judgements—including judgements of value. Those judgements might be both general and imprecise, leaving room for widespread disagreement about their interpretation and implications. There might be no one set in which there is agreement throughout the society— there might instead be a series of overlapping agreements,⁴⁶ making possible a series of overlapping conversations. But the possibility of deliberation, and of either agreement or disagreement, depends on the existence of such an underlying agreement in judgements. Participation in such practices of reasoning and deliberation provides us with an essential ‘frame of reference within which we are to . . . justify our attitudes, judgements, and actions.’⁴⁷ The important point to notice here is that we must be able to articulate them, and be ready to defend them, not just to ourselves, but to others: ‘the “discipline” of reason holds us to public argument and reflective thought.’⁴⁸ This is true of any mode of thought and judgement—that to engage in it is to engage in a social practice within which I am answerable to standards of correctness and incorrectness that are external to me, and answerable to other participants in the practice in the light of those standards. Just what those standard are, just how and under what conditions I must answer to others, and (most crucially) just who those ‘others’ are will of course depend on the character of the practice in question; all we are so far claiming is that some kind of answerability, to some set of fellow participants, is a feature of any mode of public reasoning— and thus of reason itself. That answerability takes on a distinctively significant character in the context of moral thought and reasoning. If we are to take our own moral agency seriously, ⁴⁴ G Postema n 2 above, 53, drawing on P Pettit, The Common Mind (New York, 1993) esp chs 2, 4. ⁴⁵ L Wittgenstein, Philosophical Investigations (trans GEM Anscombe; Oxford, 1963) para 242. ⁴⁶ Or, one might say, an overlapping consensus: see J Rawls, ‘The Idea of an Overlapping Consensus’ (1987) 7 OJLS 1; Political Liberalism (New York, 1993). ⁴⁷ Postema n 2 above, 56. ⁴⁸ Postema n. 44 above, 43, drawing on I Kant, Critique of Pure Reason (trans N Kemp Smith; New York, 1965) 593. Compare also Kant’s comment—‘How much, and how correctly, would we think if we did not think, as it were, in common with others with whom we mutually communicate!’ (‘What is Orientation in Thinking’, as quoted by Postema, n. 44 above, 44).

240

RA Duff and SE Marshall

we must of course think for ourselves about what we may, should or must do (or think, or feel): the first person perspective is crucial, and (for any liberal moralist) must be expressible as an authentic first person singular perspective; what matters in the end is what I can determine as being right. But to think morally is still to think as one of a ‘we’, not as a solitary individual, and as one who must be ready to recognize and to answer to others as participants in this practice: hence the force of Korsgaard’s comment that ‘[t]o say that you have a [normative] reason is to say something relational, something that implies the existence of another . . . . It announces that you have a claim on that other, or acknowledges her claim on you’;⁴⁹ for claims must be, if necessary, justified by those who make them and to those on whom they are made. A crucial moral question is, of course, just who those ‘others’ are, on whom we have claims and whom we recognize as having claims on us, and appeals to the necessarily public character of reason cannot by themselves settle that question: someone whose moral concerns are unduly limited in their scope, who does not see himself as having any responsibility towards some group of (as he sees them) outsiders, is morally deficient in his vision and his sympathies, but does not thereby show himself to be rationally defective; for he still operates within a public realm of reason, even if that realm is morally too narrow. But the central point is, again, that to think and act morally is to participate in a social practice (or more plausibly, one or more of a range of social practices) within which we must see ourselves as responsible, and answerable, to the other participants.⁵⁰ A similar point applies when we turn to the public reason that is involved in democratic political life. To engage in practical political thought, including deliberation about the proper content and application of the criminal law, is to participate in a public practice; it is to see oneself as a member of a ‘we’ who are collectively engaged in this enterprise, and to whom one is answerable for one’s judgements and decisions. The public character of reason does not itself dictate the scope of that ‘we’: just as a central moral question concerns the extent and membership of the moral community (or communities), a central question for political theory and political practice concerns the membership of the political community by whom political decisions are to be made, and to whom they must be justified.⁵¹ However, if we add a basic democratic principle to the picture, we can say that the ‘we’ must consist in all the citizens; it is we, as citizens, who must

⁴⁹ CM Korsgaard, ‘The Reasons We Can Share: An Attack on the Distinction between AgentRelative and Agent-Neutral Values’ (1993) 10 Social Philosophy & Policy 24, 51; and see Postema n 46 above, 53–85. ⁵⁰ This should not be taken to imply, unacceptably, that only those who can make claims, or demand justifications, or call us to answer, have moral claims on our concern. ⁵¹ Note too that the group to whom decisions must be justified could be wider than that by whom decisions are to be made: a non-democratic sovereign body should still recognize that it must justify its demands to all those on whom they are supposedly binding.

Criminal Responsibility and Public Reason

241

collectively determine our laws.⁵² Furthermore, central to such a process of collective (self-)determination is a ‘basic premise of reciprocity’—that ‘citizens owe one another justifications for the institutions, laws and public policies that collectively bind them’:⁵³ to recognize others as fellow citizens in a democracy is to recognize them as fellow participants in an enterprise of rational deliberation and action in which I must answer to them, as they must answer to me, for my judgements, claims and actions insofar as they fall within that enterprise. This is not yet to say, of course, that the citizens of any such democracy must, or would inevitably if rational, institute a system of criminal trials through which they call each other to answer for such public wrongs as they commit. Indeed, it is not even yet to say that they would institute a criminal law at all: they could collectively opt for an abolitionist approach. We have already suggested, however, that if they take wrongdoing seriously, as something distinct from causing harm, they will need to determine a category of public wrongs, which concern the whole polity as such; and that an appropriate response to the commission of such wrongs is to seek to call the wrongdoer to public account—which is just what criminal trials, on our account, aim to do. The argument we have sketched in this section, about the public character of reason and the way in which that is exemplified in a deliberative democracy, does not by itself warrant the conclusion that a liberal democracy should maintain a system of criminal trials whose central purpose is to call citizens to answer for the public wrongs that they commit. But it does provide a broader and deeper context within which that conception of the criminal trial can find a plausible home: for we can now see the trial, as a process of calling to account, as just one of the various ways in which, as participants in the wide range of practices of reason that structure our lives, we hold each other responsible. One concluding remark is worth making. Responsibility is typically reciprocal: if I am to claim the right or the standing to call you to account or to answer for what you have done, I must accept that you likewise have the right to call me to account; if you are answerable to me, I am answerable to you, though not necessarily in just the same terms or for just the same things. Criminal trials are often seen, and all too often actually operate, as one-way processes of calling to account: the defendant is called to answer to or before the court, but there is no serious suggestion that the court, or the citizens in whose name the court acts, must answer to him. If the trial is to be legitimate, however, as a process in which citizens of a democracy are to be expected to take part, responsibility must be reciprocal and shared: all the participants have their particular responsibilities, but we must also recognize a collective responsibility to the defendant. There are ⁵² Any adequate democratic theory must also, of course, have something to say about the conditions of citizenship—about how citizenship is to be decided, claimed or granted; this is another issue that we cannot pursue here. ⁵³ A Gutmann and D Thompson, ‘Why Deliberative Democracy is Different’ (2000) 17 Social Philosophy & Policy, 161, 167.

242

RA Duff and SE Marshall

three dimensions to this responsibility. First, he is entitled to a trial, and indeed a criminal process more generally, that treats him with due humanity and respect (hence the importance of the procedural rights enshrined in the US constitution and in the European convention); if that entitlement is not met, he has a claim against us, as the fellow citizens who subjected him to this process. Second, he is entitled to demand a justification for the law that he is to be condemned and punished for violating: the justification that we owe each other for the ‘laws . . . that collectively bind’⁵⁴ us is especially owed to those to whom those laws are to be coercively applied. We might legitimately refuse to engage in such a process of challenge to, and justification of, the law in the courtroom itself; for reasons to do with the appropriate division of legislative and adjudicative labour, we might argue that it should take place in a different forum—in the political forum of legislative deliberation: but we must then ensure that there is such a forum in which his challenge can be heard and met.⁵⁵ Third, if we are now to hold him responsible, as a fellow citizen whom we have the right and the standing thus to call to account under the law that binds us all, we must also be ready to answer to him for any other ways in which we have collectively failed to treat him with the respect and concern due to him as a citizen—for instance for any kinds of unjust exclusion from the rights and benefits of citizenship that he may have suffered.⁵⁶ By seeing responsibility, in particular criminal responsibility, as a relational enterprise of calling each other to answer, we can understand the criminal trial as (in normative aspiration even if all too often lamentably not in actual fact) a process through which citizens are called to answer to their fellows for the public wrongs that they commit; and we can see too that we must ourselves answer to those whom we would thus call to account. ⁵⁴ See at n 53 above. ⁵⁵ See further, Duff, Farmer, Marshall and Tadros, n 1 above, ch 5.5. ⁵⁶ On the significance and implications of this point, see further, Duff, Punishment n 28 above, chs 3.1, 5. 2–3.

13 The Educative Function of Law Brian Burge-Hendrix

Classical philosophers invariably paid considerable attention to the connection between law and education. They held that one of the primary functions of a legal system is to make its subjects aware of the set of norms that structured social and political life, and to persuade them to follow those norms. This educative function of law was thought to be indispensable for the state’s existence: it was not merely a jurisprudential thesis. Works of political philosophy that proposed substantive reforms of existing political constitutions, or presented ideal political constitutions always gave careful attention to the pedagogical effects of those proposals. Consider, to give just one example, Plato’s Republic and the enormous amount of space given to discussions of appropriate educational regimes and their effects on the health of the state, both in regard to the well-being of individual citizens and the political relations between them. Furthermore, as a matter of practical politics the successful inculturation of the young into a stable political society ruled by law was a common focus for legislative action. It is not an exaggeration to say that for the Ancients, law and education were wholly conjoint. Today, however, it is rare to find any jurisprudential analysis, consideration, or even awareness of the law-education connection. This is due in part to a division of labour: philosophy of law and philosophy of education are separate sub-disciplines. However, regardless of the modern arrangement of institutional arrangements in the academy, the educative function of law is too important to be ignored any longer by legal philosophers. If it is the case that classical philosophers were correct in attributing an educative function to law, and we initially and for the sake of argument give them the benefit of the doubt, then we are faced with something of a mystery in the history of jurisprudence: given the prominence of law’s educative function in classical philosophy and classical jurisprudence in particular, why do we no longer speak of it? Perhaps our current lack of a concern with the law-education connection marks a significant change in our understanding of the nature of law, or a change in the nature of law itself, or both. In this paper I offer an explanation for that jurisprudential mystery while arguing for the philosophical importance of law’s educative function. We ought, it

The Educative Function of Law. Brian Burge-Hendrix. © Oxford University Press 2007. Published 2007 by Oxford University Press.

244

Brian Burge-Hendrix

seems to me, to be concerned with that function not only because it merits a place in our philosophical understanding of law, but also because it demands consideration in many other academic disciplines and sub-disciplines concerned with legal systems and their effects, including political philosophy, moral philosophy, legal sociology, and legal anthropology. By focusing our attention on law’s educative function we shall see how the philosophical analysis of law has implications for the entire scope of inquiry into legal systems and their effects.

I. Three Theses About Law A. The Impossible Education Thesis It may be that philosophical attention is no longer paid to law’s educational function because, in fact, law is not capable of doing any such thing. Let us call this claim the Impossible Education Thesis: No legal system can educate its subjects. If the Impossible Education Thesis is true, then the claim that it is necessary for law to play an educative role is unsound, and classical philosophers are simply mistaken on that point. However, we cannot as a matter of course just assume that classical philosophy incorrectly asserts a law-education connection. And it happens to be the case that the Impossible Education Thesis shows itself to be false in the face of a fairly simple argument that shows that all legal systems do in fact educate their subjects and, what is more, cannot help doing so. Insofar as it is always possible to learn about one’s environment so as to plan future action, there can be no question that law, as a feature of our social environment, has at the very least an inadvertent educative effect. This is most easily understood by analogy with our physical environment: one learns, for instance, that certain types of clouds warn of a rainstorm, and that stepping off a tall ladder into thin air leads to a fall. This sort of prudential knowledge is admittedly very basic—so basic that we give it little thought at all once we gain a certain degree of familiarity with our physical environment. Yet despite our habituation to this fact, it is nonetheless true that we develop a kind of knowledge through experience of the effects of particular entities in our environment. Despite being the source of some aspects of prudential knowledge, however, it does not follow that rain clouds or gravity are educators. That is to say, these features of our environment are not agents that deliberately act so as to educate us: the education we gain from our contact with them is inadvertent. Just as with our physical environment, so too can we gain prudential knowledge of our social environment. And, again, such knowledge does not require any deliberate attempt to educate us. For example, you might learn not to tease your sister while she is carrying a baseball bat, and this knowledge is real even if your sister, immediately prior to the moment of your illumination, had been carrying the bat with the intention of playing baseball rather than from an intention

The Educative Function of Law

245

to teach you not to tease her. Any effective legal system is educational in this very basic way, and so the Impossible Education Thesis is clearly false: the social consequences of a minimally effective legal system cannot help but produce prudential knowledge in those who encounter it. Yet the educative function attributed to law by classical philosophers is not merely an inadvertent one. There is a difference between simply learning from something and being taught by it—we can distinguish the inadvertent educational effects of law from the idea that a legal system is an educator whose effects are deliberate and intentional. So not only is the Impossible Education Thesis false, it cannot be modified so as to make sense of the classical philosophical view of law by simply asserting that in classical philosophy law’s educative function is an inadvertent one.

B. The Possible Education Thesis Law, according to Plato and Aristotle and Cicero, is necessarily an educator, and yet our jurisprudential accounts of law rarely discuss a law-education connection. One possible explanation for this discrepancy is that we do in fact attribute an educational function to law, but do so in such a way that it is transparent to us. Perhaps, for instance, the notion of publicly ascertainable, peremptory content-independent reasons for action implies a law-education connection even if it does not explicitly mention one. We shall consider that possibility later, but first let us consider a more methodologically radical explanation for the discrepancy between classical and modern legal theory, an explanation that will further define the conceptual terrain on which this issue rests. It could be the case that classical philosophers were correct in thinking of law as an educator, and that we are also correct in not thinking of law in that way. If this is so, then we should avow the Possible Education Thesis: A legal system may, but need not be an educator. The Possible Education Thesis would allow for an expedient and seemingly charitable explanation of the change between classical and modern legal theory insofar as the thesis does not force us to assume that the classical philosophers are grossly mistaken. Instead, we could accept that it happens to be the case that in the classical world, law did indeed have the educative function attributed to it, but it is also the case that present-day legal systems do not evince that function. Asserting the Possible Education Thesis, however, presents a serious obstacle to a commonly shared goal of modern legal philosophers, namely the development of a truly general jurisprudence. If we invoke the Possible Education Thesis to account for the difference between classical and modern views of law’s educational function while simultaneously giving credence to classical philosophers’ estimation of the significance of that function, we commit ourselves to the view that its presence or absence marks a significant difference in the structure of legal systems. It is both easy and dangerous to underestimate the significance of such a difference. Should it be real, then it would be

246

Brian Burge-Hendrix

a mistake to think of present-day legal systems as the same type of creature as those things we are accustomed to call ‘legal systems’ in the classical world. It would then become impossible to develop a general jurisprudence which encompass both types of institution, and that is a result which most contemporary legal philosophers are probably not willing to accept. By abandoning the view that both classical and modern legal systems are instances of an ‘institution [which], in spite of many variations in different cultures and in different times, has taken the same general form and structure’,¹ we forsake the possibility of general jurisprudence as it is currently conceived. There is a compromise available that would mitigate our disavowal of general jurisprudence as we know it. The compromise emphasizes historical change as a form of changing social contexts by suggesting that the general form and structure of a legal system is relative to its socio-historical context. This view would have some support insofar as most legal theorists recognize that there are close relations between a particular society and its legal system.² In brief, we could argue that classical Greek and Roman societies imposed particular requirements on Greek and Roman legal systems such that those systems had to fulfil an educational function, while modern legal systems do not, by virtue of their own peculiar social context, need to fulfil that function.³ Unfortunately, the compromise solution is also unacceptable, and again for methodological reasons. Relativizing jurisprudence along historical lines would lead to a considerable increase in the complexity of our jurisprudential theory. We already employ trans-historical concepts in our contemporary legal theory, and those concepts are explanatorily powerful despite being relatively simple. The concept of the ‘sovereign’, for instance, can be applied to a representative parliament or to a Roman dictator, and the concept of an educative function of law is equally powerful in large part because it can be applied to different societies in different times and places. An account of law that renders its educative function contingent on to socio-historical circumstances would be hard-pressed to explain why that function is not ¹ HLA Hart, The Concept of Law (2nd edn, Clarendon Press, 1994) 240. ² Joseph Raz, for instance, remarks that ‘the law is an aspect of a political system, be it a state, a church, a nomadic tribe, or any other. Both its existence and its identity are bound up with the existence and identity of the political system of which it is a part’ (J Raz, The Concept of a Legal System (Oxford University Press, 1980) 251). However, Raz, like many other legal philosophers, is concerned with the necessary features of legal systems. ³ One can envisage a claim like the following: contemporary societies impose a mandatory educational regime that carries out what used to be the educative function of law. However, there are three obvious problems with this claim. First, not all contemporary societies impose such a regime or even make available widespread public education, yet some of those societies at least appear to have a legal system. Second, where such regimes exist, the fact that attendance in such a system is made a legal requirement still evinces a law-education connection, though perhaps a more complex one in light of the institutional structures. Third, the legal requirement to educate one’s children was not an uncommon requirement in ancient societies, and it is not at all clear that an education lacking what we now consider to be a suitable ‘civic education’ sufficed to fulfi l that requirement. In any event, the more telling objection to the modified version of general jurisprudence is a methodological one, as can be seen in the following paragraphs.

The Educative Function of Law

247

properly applied to all societies ruled by law despite the apparent ease with which we can apply an explanatory concept that applies to all of them. Thus a proponent of the historical relativization of general jurisprudence is faced with a very difficult meta-theoretical challenge indeed, and any version of general jurisprudence attempting to meet that challenge must defend a much more detailed taxonomy of social contexts and applicable explanatory concepts than that required by a general jurisprudence which accepts the existence of law’s ‘general forms and structures’ throughout history.⁴ The Possible Education Thesis is simply untenable in contemporary legal philosophy.

C. The Necessary Education Thesis The Impossible Education Thesis is untenable because law cannot help but be at least an inadvertent educator, and because we have as of yet no good reason to doubt the classical account of law whereby law is necessarily an educator. The Possible Education Thesis is tenable only if law is merely a possible educator in virtue of the fact that it is a contingent matter whether a particular legal system educates its subjects. For that to be true, we would have to discount the prominence given to the law-education connection in classical legal theory, or abandon our aim to develop a general jurisprudence, or modify the very idea of general jurisprudence so as to render it into something utterly different from its present form. If classical philosophers were not wholly mistaken regarding what they considered one of law’s primary features, and if we continue to regard both ancient and modern legal systems as tokens of the same type, then we must do one of two things: (1) explicate contemporary legal theory to reveal the currently implicit educative function of law, or (2) borrow or create, and then apply a conception to that function. In either event, if we are to defend the possibility of general jurisprudence as it is currently understood without assuming that the Ancients were utter fools as regards their legal philosophy, then we must adopt the Necessary Education Thesis: All legal systems act as educators. The Necessary Education Thesis holds that every legal system is an educator. On its own, however, it does not commit us to the view that law’s educative function is properly considered to be something requiring philosophical description in a general jurisprudence. We could accept the Necessary Education Thesis, including its concomitant claim that a legal system deliberately attempts to educate its subjects, while denying the thesis a role in a philosophical account of law rather than some other type of inquiry, such as legal sociology. We might cite ⁴ It is also possible to reconcile the Possible Education Thesis with the aim of general jurisprudence by describing the relevant change of social contexts as simply a cultural difference rather than an historical difference. In fact, the socio-historical approach can be seen as a subset of a the cultural-difference approach. However, this too would face the same challenge.

248

Brian Burge-Hendrix

Joseph Raz’s warning against confusing the contingent characteristics of a legal system with its necessary characteristics: Since a legal theory must be true of all legal systems the identifying features by which it characterizes them must of necessity be very general and abstract. It must fasten only on those features of legal systems which they must possess regardless of the special circumstances of the societies in which they are in force. This is the difference between legal philosophy and sociology of law. The latter is concerned with the contingent and with the particular, the former with the necessary and the universal. . . .⁵

In this regard, Raz’s view of the nature of legal theory reflects the considered consensus of most all legal philosophers today. It is not, I hasten to add, a specifically positivist position. Even Ronald Dworkin claims to present a general jurisprudence. Nor does this view of legal theory require us to determine whether law is necessarily linked to morality: it leaves open the possibility that there is an inherent connection between law and morality (though of course Raz’s developed legal theory, like those of other exclusive positivists, denies that a legal rule can rely directly on moral considerations). The fallback position described in the previous section, where it was suggested that law’s educative function may be contingent or necessary depending on its socio-historical context, may limit the scope of the Necessary Education Thesis such that it applies to law’s socially contingent character rather than its normatively necessary character. If that scope limitation is appropriate, then we can still give some credence to the classical legal philosopher’s claim that law must be an educator. We could say that modern philosophy maintains an explicit distinction between philosophical and sociological theories, and that classical philosophy does not maintain that distinction. We need not argue that ancient and modern legal systems are radically different; we need only argue that the educative function of law is a contingent social, and thus non-philosophical, characteristic. Although applying the distinction between legal philosophy and legal sociology would enable us to retain the Necessary Education Thesis, thereby giving some credence to the classical philosophical concern with law’s educative function, and justify our modern unconcern with law’s educative function, this strategy is still untenable insofar as it is disingenuous. It is certainly foolish, for example, to suggest that Plato failed to distinguish the contingent from the necessary, or that he failed to understand that philosophy is concerned with what is necessary rather than what is merely contingent. Plato was not a sociologist, and he was well aware of the contingent character of many of law’s social characteristics. Nor could we plausibly assert that there was insufficient diversity among the legal systems of which he was aware to allow him to develop a philosophy of law, rather than a sociology of it. ⁵ J Raz, The Authority of Law (Clarendon Press, 1979) 104–5.

The Educative Function of Law

249

Even Raz’s legal theory implicitly posits a necessary educative function of law, despite his distinction between legal philosophy and legal sociology. What remains to be shown, however, is whether the educative function of law is philosophically significant, and that is the question to which we now turn.

II. Education and Normativity A. Law as a Normative Fact The educative function of law is tied to a legal system’s capacity to instantiate and effectuate a normative system. Insofar as every legal system must establish a public set of shared norms that legally ‘trump’ private norms, even and perhaps especially those private norms held by individuals who have deliberated on and rejected the contrary public norms, every legal system must effectively communicate its legal norms to its subjects. The first step in that process is the specification of particular norms as legal norms. For that reason, Joseph Raz suggests it is ‘an essential part of the function of law in society’ to mark particular norms as members of the set of legally mandatory public norms.⁶ It is important to note, however, that there is a difference between ‘marking’ a particular norm and effectively establishing that norm. To function normatively—to have an effect as a reason for action in the fashion of legal reasons—putative legal norms must not merely be marked but also gainsay the backing of a supposedly legitimate authority (regardless of whether that authority is actually legitimate). The simple existence of a law as a social fact backed by state power provides only for the possibility of its being an object lesson in self-directed prudential education. But knowing that a law exists is not like knowing that clouds bring rain, or that it is dangerous to anger sisters carrying baseball bats. To know that a law exists is to know that the norm that has been marked as a legal one is part of a set of such norms, and that the set makes a claim to authority which distinguishes it from things whose effects are merely inadvertent. To establish a law is to convince its subjects of the existence of a legal system, and this requires an understanding on the part of its legal subjects that goes beyond simple prudential knowledge of cause and effect relationships. The existence of a legal system ensures that its effects are seen not merely as actual consequences, but also as normative consequences. For a particular law to have normative consequences, it must be at least grasped as the product of a the larger structure which gives it normative weight, ie the legal system. Only by being understood in that way can it be seen (though not necessarily accepted) as ‘a view binding on all members [of society] notwithstanding their disagreement from it.’⁷ ⁶ Raz, The Authority of Law n 4 above, 51.

⁷ ibid.

250

Brian Burge-Hendrix

In brief, we must conclude: (1) that the establishment of legal norms as properly normative rather than merely as a kind of brute social fact demands a relatively sophisticated understanding on the part of those subject to the norm; (2) that this understanding is not merely prudential; hence (3) legal systems must fulfil a deliberate educative role if they are to appear as and be understood as something more than the social equivalent of gravity.

B. Law as a Persuasive Educator On the account we have developed, the existence of a legal system is dependent on its subjects’ ability to identify it as a normative system, that is, as ‘a view binding on all members [of society] notwithstanding their disagreement from it.’⁸ One outstanding merit of this view is the conceptual continuity it provides between classical and modern legal theory. Let us consider an example. In Plato’s Crito, Socrates tells his human interlocutor, Crito, that he will decide whether or not it is morally correct to flee Athens and save his life only on the basis of principles, and not on the basis of the threat—the brute social fact—presented by his impending execution. Socrates asserts, first, that one must never do wrong, even if one is being wronged by others.⁹ He goes on to assert that one ought to fulfil all one’s rightful agreements.¹⁰ It is clear that Socrates is not relating to the law merely in a prudential fashion (though he acknowledges that such a perspective is possible).¹¹ Rather, Socrates aims to determine the right course of action, an aim that clearly involves deliberation about the norms he considers to be relevant.¹² It is noteworthy that Socrates, in his elenctic discussion with Crito, invokes a fictional person in the guise of the laws and constitution of Athens: the Laws. And the Laws, when confronting Socrates, suggest that by taking flight and thus disobeying the verdict of the Court, Socrates would be attacking the Laws and thereby threatening the existence of the City.¹³ In other words, Socrates identifies disobedience with a threat to the normative structure that makes complex social life possible.¹⁴ The Laws go on to claim that ‘all our orders are in the form ⁸ ibid. ⁹ Plato, Crito 49 B 9–10. ¹⁰ Plato, Crito 49 E 6–9. ¹¹ The mere fact that the people have the power to put others to death is no more relevant to correct decision-making, according to Socrates, than the fact that they can cause fear by conjuring ‘fresh hordes of bogies to terrify our childish minds’. (Plato, Crito 46 C 7). ¹² This point holds regardless of whether the kind of deliberation Socrates is engaged in is practical or moral. For a discussion of the distinction, see A Hatzistavrou, ‘Socrates’ Deliberative Authoritarianism’ (2005) 29 Oxford Studies in Ancient Philosophy n 45 at 100–1. ¹³ Plato, Crito 50 B 4–5. ¹⁴ Socrates, however, was always willing to obey the dictates of the legal authorities. In the Apology, he recollects how, during the rule of the Thirty, he was ordered to assist in the execution of Leon of Salamis, but in order to adhere to his principle that it is always wrong to commit an injustice, Socrates decided to disobey the order and simply went home, despite the potential consequences of that act of disobedience. (Plato, Apology 32 C 4–E 1) That act of disobedience prevented

The Educative Function of Law

251

of proposals, not of savage commands, and we give him [the citizen] the choice of either persuading us or doing what we say.’¹⁵ The ‘persuade or obey’ doctrine is the subject of considerable controversy, though we need not resolve that controversy here. What is striking is the logical consequence of that claim, and what the claim reveals about Socrates’ (and perhaps Plato’s) legal theory. The Laws do not assert that they merit blind obedience; rather, they assert that they merit at least the possibility of obedience even in the face of disagreement.¹⁶ Obedience is not always appropriate, however, and there is always the chance of sincere disagreement on the part of a legal subject.¹⁷ The crucial point that the doctrine of persuasion requires that the subject understand a particular law to be a norm belonging to a larger normative framework, and that this relationship is intentional and meant for her to understand.¹⁸ In both Raz’s legal theory and Socrates’ legal theory, the legal subject’s experience of law cannot be merely prudential (in the sense I described in section I), for understanding that a law exists is conjoint with understanding that the law is being asserted as one norm within a larger normative system where that system has not merely the happenstance character of something that can produce an effect in the form of a brute social fact, eg to restrict freedom or inflict pain or cause death, but is instead something that intentionally communicates a norm, whether we call that norm a ‘rule’ or a ‘reason for action’ or use some other term of art. Moreover, this communicative feature is best described as educative because it does not simply aim to create a habit of obedience, nor merely inculcate a blind faith in shared social norms, but always vividly and unmistakably asserts its norms as part of a larger social project that one is enjoined to obey and align oneself with.¹⁹

Socrates from unjustly harming another; if Socrates had fled to save himself from his own education, he would only be suffering a harm himself. ¹⁵ Plato, Crito 52 A 1–3. ¹⁶ Some philosophers, of course, have argued that Socrates’ considered stance towards law eschews a strong obligation to obey entirely. Hatzistavrou suggests that Popper, Klosko, Vlastos, and Gulley consider Socrates ‘to advocate individual moral autonomy and rule out subordination to the authority of another person in moral decision-making.’ Hatzistavrou himself, however, rejects the anti-authoritarian reading. He holds that ‘a demand for moral autonomy is not a basic tenet of Socratic ethics’ and asserts that ‘for Socrates it is legitimate to subordinate oneself to someone else’s judgment in moral decision-making.’ (Hatzistavrou n 12 above, 76). ¹⁷ See n 14 above. ¹⁸ This holds regardless of whether the subject believes or even wonders whether one should disobey laws one sincerely disagrees with. It is the fact of disagreement that is primary in the proof of law’s deliberately educative function, not obedience in the face of disagreement or disobedience as a result of disagreement. The question of obedience or disobedience is separable from the potential for disagreement, and the latter is logically prior to the former. ¹⁹ It should be clear, too, that the educational project embarked upon by a particular legal system need not be morally good; in fact it may be manifestly evil. We ought not to equate ‘education’ with any degree of virtue beyond that of simple deliberation over the claims asserted by the supposed social authority, and it may be the case that a better education into virtue or right action is best achieved through other means. It may even be the case that the educative project entailed by the existence of a legal system is counter-productive to a superior education by some other means.

252

Brian Burge-Hendrix

III. Education and Personal Autonomy Why is the educative function of law an implicit rather than an explicit feature of contemporary jurisprudence? To answer that question we must return to the history of legal theory and take note of what can only be called a regression in our understanding of law. The command theory of law’, Gerald Postema observes, ‘is widely associated with classical positivist jurisprudence.’²⁰ Despite that association, however, the command theory actually provided a common vocabulary and set of concepts for philosophical discussion and debate whose terms were applied not only by Hobbes, Bentham, and Austin, but also by medieval and early modern natural lawyers as well as 17th and 18th-century experts on common law.²¹ And, as Postema suggests, ‘it is reasonable to assume that the headwaters of this intellectual stream were classical’, given that we can find traces of it in Heraclitus and Aristotle, and its explicit appearance in the Digest of Justinian.²² Present-day jurisprudence, sometimes to its detriment, has not wholly abandoned the assumptions and conceptual underpinnings of the command theory. One of that theory’s core elements—‘a conception of the authority and distinctive normativity of law’—remains largely intact. Both ‘classical’ and present-day positivism portray individuals’ law-guided actions as the result of ‘deliberation carried on elsewhere’. Command theorists, including classical positivists, suggested that the end-result of ‘deliberation carried on elsewhere’ was to be found in a social superior whose will, expressed in the form of a command, takes the form of law. Contemporary positivists, however, no longer focus on the human commander and the commanded—the social superior and the inferior—but instead focus on an abstract individual, someone who reflects upon impersonal legal rules, thence to be guided by them, or not.²³ Contemporary positivism offers descriptions of legal situations that relate individuals to rules and reasons rather than to the will of other, superior individuals. Hence legal authority, on this more recent view, is effected simply through rules and reasons given to legal subjects, rather than through marks of the will of a social superior: Law’s normativity is understood entirely on the model of an individual’s guiding herself by rules. On the account preferred by Hart and Raz, this is understood in terms of the content-independent and peremptory reasons one adopts. On a recent account proposed by Scott Shapiro, it is understood in terms of certain psychologically constraining dispositions one develops. On these representative views, law guides when officials or citizens guide themselves by law in the way they describe. . . . Law is still identified with the execution of deliberation carried on elsewhere. Law claims to guide citizen/official action, ²⁰ GJ Postema, ‘Law as Command: The Model of Command in Modern Jurisprudence’ (2001) 11 Philosophical Issues 470. ²¹ Postema, ‘Law as Command’ n 20 above, 472. ²² Postema, ‘Law as Command’ n 20 above, 473. ²³ Postema, ‘Law as Command’ n 20 above, 492.

The Educative Function of Law

253

to make a difference to their practical reasoning by displacing individual deliberative judgment from them onto a surrogate.²⁴

As Postema rightly notes, the current positivist conception of authority is compelling. It is methodologically compelling because it allows for a fine-grained and general description of the efficacy of laws where that theoretical description need not ponder the content of a particular law nor its substantive moral merit. The positivist conception of legal authority meshes well with theories of practical reason, such as we see in Joseph Raz’s work, and makes it possible to seek consilience between theories of law and theories of practical reason. Moreover, despite its use of a conception of ‘deliberation carried on elsewhere’ it is a striking feature of the dominant form of contemporary positivist jurisprudence that, even given its assertion of the content-independent character of legal reasons, it allows us to highlight the practical and moral tension between personal deliberation and supposedly authoritative legal norms.²⁵ Raz, for instance, claims that every legal system must claim legitimate authority, but quickly notes that in the instance there may be no system whose claim to legitimate authority is actually sound. The positivist conception of authority is also compelling because it is consilient with moral theories that promote the doctrine of fundamental moral equality, ie theories which claim that no individual has greater intrinsic moral worth than any other individual. Given the widespread acceptance in modern liberal democracies of the idea of fundamental moral equality, the current positivist conception of authority is also descriptively and politically compelling. Its descriptive project does not rely on a type of political relationship that is perceived, in modern liberal democracies at least, to be an anachronism: the inherent political superiority of one individual over another. Hence the currently dominant positivist conception of authority both fits with the ideological character of modern liberal democracies, thus allowing for a more accurate description of those systems, and is also capable of lending its support to substantive moral/political arguments in favour of individual equality.²⁶ Notwithstanding these laudable theoretical and practical characteristics of contemporary positivism, it is the case that positivists have made conceptual room for personal autonomy at the expense of our ability to understand law’s ²⁴ Postema, ‘Law as Command’ n 20 above, 495. ²⁵ Not all positivists hold to this doctrine. Inclusive positivists, such as Wilfrid Waluchow, suggest that at least some legal norms establishing criteria for legal validity rely for their effectiveness on our evaluation of their moral content, while incorporationists, such as Jules Coleman, go so far as to suggest that the moral content of a norm may be sufficient to grant it legal validity even if it lacks a distinct social source. ²⁶ Of course, all of the seemingly compelling features of the positivist conception of authority I have just listed are controversial, and each is subject to methodological, moral, and political counter-arguments. I have addressed the methodological issues at length in B Burge-Hendrix, Epistemic Uncertainty & Legal Theory (PhD Thesis, McMaster University, 2004). Ronald Dworkin has proffered many well-known attacks on the supposed inability of the positivist methodology to account for the role of moral reasoning in judicial decision-making. And throughout its history legal positivism has been attacked as a politically degenerate and dangerous way of understanding law.

254

Brian Burge-Hendrix

educative function. Positivism’s philosophical inheritance from the command theory of law is not unproblematic, and problems central to the command theory, problems which Jeremy Bentham responded to and attempted unsuccessfully to address, raise questions for contemporary positivism as well. Near the end of his review of Bentham’s struggle with the command theory, Postema remarks that ‘[l]aw at its best, on Bentham’s view, addresses and engages the deliberative reason (“understanding”) of law subjects.’²⁷ According to this ideal, law has a communicative role to play, and that role is not simply the presentation and enforcement of rules or reasons for action. The communicative aspects of law, I have argued, are most clearly evinced in its necessary educative function. My arguments are not dependent on an ideal of law at its best, but rather depend on a thoroughgoing understanding of law as it is. It would be a mistake to abandon legal positivism and ignore its theoretical merits simply because it obscures law’s educative function. But we must make ourselves aware of the fact that positivism inherited from the command theory its tendency to de-emphasize the communicative, indeed the educative, features of legal systems. For the most part, the role of personal autonomy is something that merits considerable pride of place in any legal theory that aims to elucidate law as it exists today. This is not to say, however, that we should uncritically accept the picture of personal autonomy we so often find in current legal philosophy. There are good reasons to suspect that autonomy is also the product of an educative process, and that we have failed to distinguish between the ideal of absolute personal autonomy and the reality of the social relationships that make the development of autonomy possible.²⁸ Most importantly, we must explicate our current conception of education. As it stands, our deployment of that conception in legal theory—when it does make a rare appearance—emphasizes the ideal rather than the reality and overlooks difficult questions regarding the political necessity of inculturation, or, to use a pejorative term, indoctrination. Political philosophers have grappled with that issue for some time when debating the role of civic education in a liberal society. It is time for us, as legal philosophers, to come to terms with law’s educative function and to try to explain it as thoroughly as we have tried to explain the many other necessary characteristic of that most important of social institutions: the legal system. ²⁷ Postema, ‘Law as Command’ n 20 above, 494. ²⁸ We should take notice, for instance, of Susan Sherwin’s conception of ‘relational autonomy’ and Nkiruka Ahiazu’s conception of ‘constitutive authority’. Sherwin examines the conception of personal autonomy prevalent in medical ethics and argues that it is inaccurate insofar as the actual exercise of autonomy often relies upon the existence of supportive relationships, and suggests that we view personal autonomy not merely as freedom from undue interference, but as a co-operative effort leading to effective personal decision-making. Ahiazu argues that Raz’s service conception of authority overlooks the fact that in some communities the actual conception of what it is to be an individual human being is in tension with the idea of personal autonomy as freedom from interference, and suggests that in at least some societies the legal system must favour personal autonomy in principle while simultaneously upholding the educative social relationships that enable individuals in that society to develop into full moral agents.

14 Protest and Punishment: The Dialogue between Civil Disobedients and the Law Kimberley Brownlee

Introduction A parallel may be drawn between the communicative aspects of lawful punishment and the communicative aspects of civil disobedience.¹ In punishing an offender, the state aims to engage the offender in a moral dialogue, communicating to her both its condemnation of her action and its desire for repentance and reformation of her conduct.² Both the state’s backward-looking aim of communicating condemnation and its forward-looking aim of achieving a reformation in conduct are mirrored in a civil disobedient’s efforts to engage in a moral dialogue with authorities and society. In civilly disobeying the law, a person typically aims to communicate to policymakers, authorities, and society her condemnation and disavowal of a particular law, policy, or institutional practice (be it governmental or non-governmental) as well as her desire for a lasting change in policy.³ I have argued elsewhere that the confrontation between civil disobedience and lawful punishment not only allows for a direct comparison of their respective justification, but also presses the claim that authorities and disobedients each endeavour to engage the other in a moral dialogue.⁴ In this paper, I consider more fully the nature of the moral dialogue ¹ I shall refrain from defining ‘civil disobedience’ and adopt instead a paradigm case approach, which highlights certain key features exemplified in typical or paradigm cases of civil disobedience. These key features include a conscientious conviction and the communication of condemnation against a given policy through breach of law. ² cf J Feinberg, ‘The Expressive Function of Punishment’ in A Duff and D Garland (eds), A Reader on Punishment (Oxford, 1994), 77–9. ³ Civil disobedience undertaken in opposition to the decisions of non-governmental agencies such as banks, trade unions, and private universities often reflects a larger challenge to the legal system that permits those decisions to be taken. It would be a mistake to hold that the policies and practices of non-governmental institutions are somehow not matters of law and thereby not the proper objects of civilly disobedient protest. In condemning such policies, civil disobedients challenge, amongst other things, the legal framework that accepts these policies and practices as lawful. ⁴ K Brownlee, ‘The Communicative Aspects of Civil Disobedience and Lawful Punishment’ (2006 online) in Criminal Law and Philosophy .

Protest and Punishment: The Dialogue between Civil Disobedients and the Law. Kimberley Brownlee. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Kimberley Brownlee

256

which purportedly occurs between disobedients and the law as well as the kinds of responses that a genuine dialogue would permit from each party. The justified response by each to the other cannot be determined, I argue, solely on the basis of deserved censure. To be justified in communicating censure through punishment or through civil disobedience, authorities and dissenters respectively must not simply have good grounds for condemnation and a legitimate motivation, but also accommodate concerns of charity or good will, restoration, and the prevention of further wrongdoing. I conclude by examining briefly some ways in which the analogy between lawful punishment and civil disobedience might break down.

I. A Communicative Theory A. Civil Disobedience My characterization of civil disobedience and lawful punishment as communicative practices requires some explication.⁵ Concerning civil disobedience, in paradigm cases, a disobedient seeks to communicate not only her disavowal and condemnation of a certain law or policy as well as her desire for recognition by authorities and the relevant majority that a lasting change in policy is required, but also her personal dissociation from both the policy in question and the authority that enacted it.⁶ When the policy in question is a non-governmental policy by a private company or institution, a disobedient’s condemnation includes a challenge to the legal framework that accepts this policy as lawful; and as such she dissociates herself not only from the institution that generated the policy, but also from the government that permits it. The dissociative aspect of civil disobedience requires teasing out, as it does not figure prominently in many accounts of civil disobedience. And, while the communication of protest does figure in most accounts of this practice, it too requires explication since thinkers disagree over the form this communication takes. According to John Rawls, for example, the communicative character of civil disobedience, by definition, involves publicity and fair warning to authorities of the intended action. The fidelity to law which distinguishes civil disobedience from other types of illegal protest necessitates, says Rawls, that disobedients commit their offences openly and with fair notice.⁷ In a similar spirit, Hugo Bedau suggests that often it is essential to a dissenter’s purpose that both the government and the public know what she intends to do.⁸ One may question, however, not only the claim that civil disobedients show fidelity to the law (Gandhi is the obvious counterexample), but also the claim ⁵ My discussion is intended to apply to the context of an imperfectly liberal democracy. ⁶ cf K Brownlee, ‘Features of a Paradigm Case of Civil Disobedience’ (2004) 10 Res Publica 337–51 for a fuller discussion of paradigmatic civil disobedience. ⁷ J Rawls, A Theory of Justice (Cambridge, MA, 1971) 366. ⁸ H Bedau, ‘On Civil Disobedience’ (1961) 58 The Journal of Philosophy 655.

Protest and Punishment

257

that such fidelity is and must be reflected in publicity and fair notice. Although forewarning authorities sometimes may be prudent or may be essential to a dissenter’s strategy, this is not always the case since, as Brian Smart observes, publicizing the intention to breach the law provides both political opponents and legal authorities with the opportunity to abort disobedients’ efforts to communicate.⁹ Consequently, sometimes unannounced or (initially) covert disobedience better serves disobedients’ communicative aims than does disobedience undertaken publicly and with fair notice. To succeed, for example, in releasing animals from research laboratories or vandalizing military property, disobedients must avoid publicity of the kind that Rawls defends. Such acts of civil disobedience nonetheless may be regarded as ‘open’ when followed soon after by an acknowledgment of the act and the reasons for acting.¹⁰ In making that acknowledgment, disobedients (whose actions meet other constraints I shall discuss) may claim plausibly to be willing to deal fairly with authorities.¹¹ A different conception of the communicative aspect of civil disobedience characterizes it in terms of ‘expression’ rather than publicity and fair notice. Joseph Raz defines civil disobedience as a ‘politically motivated breach of law designed either to contribute directly to a change of a law or of a public policy or to express one’s protest against, and dissociation from, a law or a public policy.’¹² As a concept, however, expression not only lacks the other-directedness of communication, but also lacks its emphasis on understanding and rational engagement by the ‘speaker’ with her ‘hearer’. In engaging in civil disobedience, a person aims not simply to express her objection (which she could do alone and in private), but to lead her audience to understand and hopefully to accept her reasons for both condemning a given policy and dissociating herself from its authors. Concerning dissociation, often civil disobedients are represented as asserting their right to participate in the decision-making process rather than as honouring a moral conviction which prompts them to make a communicative act of dissociation from a policy and its creators. Paradigmatically, however, civil disobedience reflects principled opposition based upon deeply and conscientiously held commitments, and it is this principled objection which leads dissenters communicatively to dissociate themselves where possible from particular policies.¹³ Their actions demonstrate their moral consistency. As Antony Duff notes in his discussion on punishment, when one judges a certain decision or policy to be wrong one must not only avoid that type of decision oneself and judge the conduct of others who pursue such decisions as being wrong, but also, in some contexts, ⁹ B Smart, ‘Defining Civil Disobedience’ in H Bedau (ed), Civil Disobedience in Focus (London, 1991) 206. ¹⁰ cf J Raz, The Authority of Law (Oxford, 1979). ¹¹ Note that a willingness to deal fairly with authorities should not be confused with fidelity to law. I explicate this difference below. ¹² Raz, Authority n 9 above, 263. ¹³ Brownlee n 5 above, 337–51.

258

Kimberley Brownlee

make known one’s judgment of that decision: to remain silent and to let an action pass without challenge can cast doubt on the sincerity of one’s conviction that this action is wrong.¹⁴ Avenues for conscientiously motivated disobedients to communicate their judgment will vary depending upon the object of their judgment. When breaching the law or disrespecting the policy they oppose is either impossible or indefensible then they may have to resort to indirect disobedience, a strategy which lacks the explicit act of dissociation exemplified in direct civil disobedience. The person who uses indirect civil disobedience breaches a law which ceteris paribus she does not oppose in order to challenge the law or policy that she does oppose. Her dissociation from the object of her condemnation is moral rather than legal in character.¹⁵ Through her communication, she affirms that these policy-makers do not speak for her when enacting this law or implementing this policy, and that were it possible or defensible for her directly to distance herself from this policy (or to intrude upon its parameters) she would do so.¹⁶ The centrality of dissociation to both direct and indirect civil disobedience highlights a difference, beyond legal status, between these practices and legal protest. Although legal protest, like civil disobedience, typically aims to convey to others certain commitments and values, nevertheless legal protest lacks the conscientious and communicative dissociation which distinguishes civil disobedience from ordinary offences. Either the legal protester does not believe that her cause is sufficiently weighty to warrant a breach of law in its defence or she does believe this but for various reasons does not act upon that belief. While her reasons not to offend may be compelling (eg concern for personal security or others’ well-being or a commitment to legal obligation), the point remains that she does not dissociate herself from the policy she opposes. And, in some cases, that can cast doubt on the sincerity of her judgment that the policy is wrong. A further reason to favour civil disobedience over legal protest as a vehicle for communicating opposition is its effectiveness. As Bertrand Russell observes, typically it is difficult to make the most salient facts in a dispute known through conventional channels of participation.¹⁷ The people who control mainstream media tend to give defenders of unpopular views limited space to advance their causes. Given, however, the sensational news value of illegal protest, engaging in civil ¹⁴ cf A Duff, Punishment, Communication, and Community (Oxford, 2001) 28. ¹⁵ The distinction often drawn between direct civil disobedience and indirect civil disobedience may be less clear-cut than one might think. For example, would refusing to pay taxes that support the military be an act of indirect or direct civil disobedience against foreign policy? Although this act presumably would be classified as indirect civil disobedience, nevertheless a part of one’s taxes, in this case, would have gone directly to support the policy one opposes. ¹⁶ The language of dissociation suggests a refusal to follow a direct order or an abstention from a social practice. But, often dissociating oneself means interfering with or intruding upon the policy one opposes. For example, James Meredith dissociated himself from the admissions policy of the University of Mississippi in the early 1960s when he demanded to be admitted as a student. James Meredith became in 1962 the first African American to attend the University of Mississippi. ¹⁷ B Russell, Autobiography (London, 1998) 635.

Protest and Punishment

259

disobedience often leads to wide dissemination of a position where legal protest might not do.¹⁸ But the effectiveness of a communicative strategy must be weighed against its credentials as a contribution to a moral dialogue. Although the civil disobedient often may be better placed than the legal protester is to ensure that her view is heard, her communication only counts as a genuine contribution to a moral dialogue when it respects the constraints upon such contributions. Forcibly conveying one’s view is incompatible with the parameters of genuine dialogue when it combines with a failure to take into account the contribution of other parties to the dialogue. The civil disobedient’s desire to communicate effectively, therefore, must be tempered by her awareness that she must eschew modes of communication at odds with her persuasive aims. I shall pursue this issue further in the next section. For now, let us consider the communicative conception of punishment.

B. Punishment Whereas there is general agreement amongst philosophers that civil disobedience is a communicative practice (though philosophers disagree over how that feature is to be understood), there is less agreement that punishment is an essentially communicative practice. Traditional theories of punishment highlight either its preventative aims or its retributive aims as central both to its definition and to its justification while disregarding the appropriateness of characterizing punishment in terms of moral communication. I shall reserve my comments on justification for the next two sections, highlighting here some reasons to regard lawful punishment as an essentially communicative practice. As noted above, communication is an other-directed activity that emphasizes, first, engagement at a rational level by the ‘speaker’ with the ‘hearer’, and second, understanding by the hearer of what has been conveyed to her. Communication does not necessarily involve a reciprocal rational engagement, as Duff suggests it does, because the hearer need not be an active participant in the communication.¹⁹ There is no reciprocal engagement, for example, between a newscaster and her viewing audience or between a billboard and its observers, but there is communication when the ‘hearers’ understand the information or position conveyed. That said, successful communication usually requires awareness by the speaker of her hearer and the kinds of communicative measures he is likely to understand. Communication means, at a minimum, that the hearer comprehends the content conveyed, and typically (but not necessarily) that he makes some response in return. But ‘communication’ is not synonymous with ‘dialogue’ since only the latter requires that both ¹⁸ John Stuart Mill makes a similar point with regard to dissent in general. Sometimes, says Mill, the only way to make a view heard is to allow, or even to invite, society to ridicule and sensationalize it as intemperate and irrational; J Mill, On Liberty (London, 1869). Admittedly, the success of this strategy of self-sensationalization depends partly upon the character of the society in which it is employed; but we should not rule it out as a mode of communication. ¹⁹ Duff, Punishment n 12 above, 79.

260

Kimberley Brownlee

parties be speakers as well as hearers. Arguing that lawful punishment is communicative is easier than arguing that punishment is an attempt to engage the offender in a moral dialogue, because the latter requires the state to be responsive, at least in principle, to the response made by the offender. I will say more about the nature of dialogue, particularly moral dialogue, in the next section. Punishment can be plausibly presented in terms of a rational (if not obviously reciprocal) engagement with the offender in the first instance and with other hearers in the second. In imposing a burden on an offender for an offence, the state seeks to bring her to understand certain things such as the reasons she has to follow the law, the censure rightly elicited by her offending action, and the expectation that she will reform her conduct. Like civil disobedience, lawful punishment involves a performative act of condemnation, dissociation and disavowal of certain conduct. The state demonstrates through isolating, ostracizing, and burdening a person for her action that she did not act in its name or in keeping with its values when she carried out this offence. In a liberal polity, constraints of publicity and openness apply to lawful punishment, greater perhaps than those upon civil disobedience, to make sense of the claim that the state wishes to deal fairly with its citizens. Given the burdensome nature of punishment, openness and fair-warning are important to treating those liable to the pain of punishment as rational agents with equal moral status. While fair-warning is not always possible, striving to provide it is essential for a state that purports to respect citizens’ rights and autonomy. In my comments on civil disobedience, I suggested two reasons that favour civil disobedience over legal protest as a vehicle for communicating with policymakers and society. The first related to the moral consistency and sincerity of the disobedient’s action as communication of condemnation, the second to the relative effectiveness of disseminating a view (though that might not translate into a greater effectiveness in persuading people of the view). By no means do these reasons amount to a justification for civil disobedience; rather, they highlight advantages of civil disobedience over lawful protest as a communicative practice. The question is whether similar reasons favour lawful punishment over alternative vehicles that the state might use to communicate censure for breach of law. Concerning effectiveness, the state does not have the same difficulties that disobedients have in airing its views at least to its primary hearer. Therefore, only when the state’s efforts to communicate to persons other than the offender would be unsuccessful through non-punitive measures would there be a reason relating to communicative effectiveness or dissemination to punish (persuasiveness is a separate issue). Concerning moral consistency, the issue is whether punishment is the appropriate way for a state to dissociate itself from the conduct of its citizens. Would a refusal to punish reflect negatively on the sincerity of the state’s judgment that an action was wrong? Whether it did would depend upon the nature of both the offence and the proposed punishment. Less serious offences invite less condemnation and may make other forms of communication more appropriate than the burdens of punishment: in charging a person with an offence, for

Protest and Punishment

261

example, the state makes known its dissociation from her conduct. Some thinkers maintain, however, that penal hard treatment is to be understood as an essential aspect of the enterprise of moral communication itself. Duff, for one, argues that it would be a mistake not to recognize the contribution that punishment can make to reparation and restoration: A crime involves not merely (nor always) material damage to a victim’s interests, for which material compensation might be appropriate, but a wrong done by the offender to the victim. An appropriate ‘restoration’ or ‘reconciliation’ between them must then involve a proper recognition of that wrong and of its implications . . . A suitable programme of ‘restoration’ must thus involve the offender in recognising, repenting, and apologising for the wrong he has done.²⁰

Duff maintains that this process of restoration involves the same central elements as punishment understood as the government’s effort to engage in a moral dialogue with the offender. Punishment, on this picture, not only communicates both disapprobation of the offender’s action and a desire for repentance and reformation on her part, but also gives the offender an opportunity to communicate her repentance by accepting the punishment, apologizing, and making reparation where possible. But the fact that punishment can make a contribution to reparation and restoration, does not mean that it always must be part of that process or that it would play that role better than other kinds of responses. The question for our discussion is whether a programme of punishment is appropriate when the state engages with civil disobedients. With this understanding of the communicative character of lawful punishment and civil disobedience, we may now consider what criteria must be met for the parties to these practices to engage in a genuine moral dialogue. After outlining in section II the conditions for moral dialogue, I consider in section III the conditions for a justified communication of censure by authorities and dissenters through their respective vehicles of punishment and civil disobedience.

II. Moral Dialogue A. Requirements for Moral Dialogue Before analysing the features of a moral dialogue, we should consider the key features of a dialogue. First, as noted above, a dialogue is a specific form of communication in which both (or several) parties to the exchange play the dual roles of speaker and hearer. Second, the term ‘dialogue’ implies a more extensive exchange amongst these parties than a simple call-and-response scenario in which one party communicates, another responds and the exchange ends. A dialogue involves a ²⁰ A Duff, ‘Punishment’ in H LaFollette (ed), Oxford Handbook of Practical Ethics (Oxford, 2003) 348.

262

Kimberley Brownlee

sustained reciprocal exchange in which the parties attend to each other’s contributions and modify their responses in light of those contributions: a dialogue is neither a monologue nor a quarrel. Third, there are connotations of fairness and equality in the notion of a dialogue. The parties are represented as equals in some important sense; moreover they are represented as willing (if not enthusiastic) participants to the exchange. The notion of a dialogue suggests that the parties respect each other as rational beings capable of understanding and responding to the reasons and arguments offered by others. Fourth, while a dialogue need not lead to any resolution or agreement amongst the parties, engaging in a dialogue implies a certain progress or development in mutual understanding. With this conception of dialogue, we may now ask what makes a given dialogue a moral dialogue. First, such a dialogue either has a moral issue as its topic or has as its topic an issue which has some moral implication. Second, in many cases, such a dialogue will involve some moral disagreement, be it about the nature or status of morality, the nature or status of a given moral principle, the application of a certain moral principle, the moral evaluation of a given act or type of conduct in light of a particular moral principle or value, and so on. In the case of the purported dialogue between disobedients and authorities, the parameters of discussion may include disagreement on several of the above fronts. It cannot be assumed that authorities and disobedients engage each other on a single point of disagreement or indeed the same points of disagreement. A potential implication of this complexity in their exchange is that the discussion often may be, pardoning the expression, ‘a dialogue of the deaf’, that is, an exchange in which none of the parties makes allowances for the others’ positions or even recognizes a common focus of debate. However, as we shall see, this possibility is rendered unlikely by the constraints placed upon those who may legitimately claim to be engaged in either civil disobedience or lawful punishment understood as attempts to conduct a moral dialogue.

B. Constraints upon Civil Disobedients and Law Enforcers The forward-looking aspect of civil disobedience has two parts. The first is to lead the authors of the policy in question to reform that policy. The second is to lead these policy-makers and society generally to internalize the reasons behind the disobedient’s condemnation and disavowal of that policy so that no similar policy will be implemented in future. As I have argued elsewhere, when a civil disobedient has these kinds of aims, this places certain restrictions upon the modes of disobedient communication that she reasonably may use to realize those aims.²¹ I wish to note here, which I have not before, that the key constraint is non-consequential in nature. Since, in paradigm cases, a civil disobedient aims to persuade her hearers of the merits of her cause, this constrains how she may act ²¹ Brownlee n 3 above.

Protest and Punishment

263

in defence of that cause. To aim to coerce authorities and society rather than to persuade them of her view would be to treat her hearers as less than fully autonomous beings with whom she could engage rationally. If she is to claim legitimately that she endeavours to engage in moral dialogue, her modes of communication must aim to respect the autonomy of her hearers as rational beings capable of responding to the reasons she believes she has to challenge current policy. The constraints imposed by her persuasive aims also have a consequential aspect. Since, as a strategy, coercion and intimidation are likely to turn policymakers and society against a position, a disobedient who sincerely aims to have a long-term impact upon people’s views has reason not to try to force policy-makers to adopt her position, but rather rationally to persuade them of the merits of her view and the flaws in the law or policy she opposes. In short, to be serious in her aim to bring about a lasting change in policy, she must recognize the importance of engaging policy-makers and community members in a moral dialogue.²² Too radical a protest could obscure the moral force of her objection. Both the nonconsequential reason of respect and the consequential reason of succeeding in her persuasive project are reasons to prefer civil disobedience as a communicative strategy to more radical disobedience.²³ The constraints upon civil disobedients arising from their conscientious, forward-looking aims have certain limits since disobedients sometimes may find it necessary or unavoidable to employ limited coercion in order to get their issue onto the table. Only when they have an ear may they undertake meaningfully to persuade authorities and society of their view. Their actions also may have a coercive aspect irrespective of their intentions since many kinds of civil disobedience—illegal boycotts, illegal strikes, refusals to pay taxes, draft dodging, roadblocks, sit-ins—make it difficult for a system to function and thus can have a potent effect upon leaders’ decisions. However, modest, incidental coercion does not muffle disobedients’ moral plea in the way that radical protest can do, and so its use can be consistent with both the persuasive aims of civilly disobedient communication and a respect for their hearers. The backward-looking aim of civil disobedience to communicate condemnation and dissociation from a certain policy also places a non-consequential constraint of respect for her hearers upon the disobedient’s mode of communication since only censure that is in some way proportionate to the wrongness of the policy is potentially justified as part of a moral dialogue. In the next section, I argue that the censure by disobedients that is justified is sometimes less than that which a policy or practice warrants. The same holds true for the censure communicated by the state through punishment: sometimes the justified censure is less than the censure the offender deserves. ²² Brownlee n 5 above, 337–51. ²³ By ‘radical disobedience’, I mean extreme forms of dissent—militant action, coercive violence, terrorization—which lack the conscientious communication and persuasive aims exemplified in paradigmatic civil disobedience.

264

Kimberley Brownlee

Concerning the forward-looking aims of punishment, the constraints on mode of communication that apply to paradigmatic civil disobedients also apply to authorities who sincerely aim to have a lasting effect upon an offender’s conduct and views. The non-consequential and consequential reasons that disobedients have to avoid overly radical responses to laws or policies are mirrored in the reasons that authorities have to avoid overly harsh punishments. When authorities aim to communicate censure in a way that will lead the offender to appreciate the moral reasons behind this condemnation, they must recognize that they have reason to be modest in the imposition of punishment. In other words, when the state is properly sensitive to the moral values it wishes to inculcate in the offender, this places certain constraints on how it may treat that person in its efforts to bring her to change her attitudes and conduct. First of all, excessive or inappropriate punishments would fail to respect the offender as a rational agent with whom authorities may endeavour to engage in moral dialogue. Second, as Andrew von Hirsh observes, to perform the function of normative communication, moderation in punishment is required because too harsh threats (and penalties) from the law would drown out the moral appeal.²⁴ In previous writing, I highlighted von Hirsch’s ‘drowning out’ metaphor for communication through punishment and applied it to civil disobedience as well, saying that since disobedients aim to persuade, they have reason not to be overly radical in their communication. The moral appeal of the communication of censure is lost if it is drowned out by preventative measures. I wish to stress, though, that it is not just that the appeal would be drowned out if the measures are too extreme. The appeal rests on treating the other as an interlocutor with whom one may engage in a rational and moral discussion. Respecting the other as a moral being with rights and dignity constrains the kinds of actions one may employ to communicate one’s condemnation to her. With this conception of both moral dialogue and the constraints upon authorities and disobedients who may claim to engage in such a dialogue, we may dismiss the worry that their exchange will be a ‘dialogue of the deaf’. To claim legitimately that they aim to persuade the other by appealing to reasons which justify their attempt to persuade, they must engage with that other in a respectful manner at a rational level in keeping with people’s status as autonomous beings. We may now consider what particular responses by these parties would be permitted within these parameters.

III. Justified Censure The constraints discussed above are formal constraints of consistency that derive from conceiving of punishment and civil disobedience as communicative ²⁴ A von Hirsch, ‘Proportionate Sentences: A Desert Perspective’ in A Ashworth and A von Hirsch (eds), Principled Sentencing: Readings on Theory and Policy (2nd edn Oxford, 1998) 171.

Protest and Punishment

265

practices whose practitioners seek to engage their hearers in a moral dialogue. The substantive question we must now address is: Which particular acts of punishment or civil disobedience are admissible within this moral dialogue? In this section, I consider four factors that bear upon the justification of a given act of punishment or act of civil disobedience. These factors are deserved censure, prevention of wrongdoing, charity or good will, and restoration. As I noted above, for both disobedients and authorities, deserved censure is central to the justification of their response. For the civil disobedient to be justified, for example, she must have objective reasons for her challenge deriving from a respect for the value of considerations that are objectively valuable. Various considerations could generate objective reasons to condemn a policy such as justice, transparency, trust, security, social welfare, rights, integrity, democracy, stability, autonomy, equality, privacy, and so on. When a law or policy offends these considerations in a significant way, this opens up a space in which a person could be justified not to respect it or another law in order to demonstrate her objection. Similarly, to be justified in punishing, the state’s objection to a person’s conduct must be well-founded. There must be objective reasons for regarding this action as wrongdoing. Having a well-founded objection does not license just any kind of burdensome response that fits the description of ‘civil disobedience or ‘punishment’. As noted above, the severity of the censure must be in some way proportional in both cardinal and ordinal terms to the seriousness of the wrongdoing at issue. For punishment, for example, the severity of the response must not only be proportional to that meted out for other offences of comparable seriousness and reflect the relative reprehensibility of the offence (ordinal proportionality),²⁵ but also be proportional to the absolute seriousness of the offence (cardinal proportionality).²⁶ The seriousness of an offence is determined by both the harm it does or risks and the offender’s degree of culpability; the greater the harm or culpability, the greater the justification for punishment. For the civil disobedient to be justified too, her communication of censure cannot be excessive relative to the wrongness of the policy at issue. The wrongness of a policy could turn on a variety of factors, including its perniciousness, impact, the intentions of the authors, and the relative harm. If, however, the disobedient’s censure goes unheard (in the way that punishment sometimes goes unheeded), she may be justified in resorting to more extreme or prolonged action to gain a hearing for her view. ²⁵ A Ashworth, ‘Desert’ in A Ashworth and A von Hirsch (eds), Principled Sentencing: Readings on Theory and Policy (2nd edn Oxford, 1998) 143. ²⁶ For example, as Tasioulas notes, focusing only on ordinal proportionality would allow one to punish an act of less relative seriousness overly harshly in absolute terms provided that the punishment was less severe relative to the punishment for the more serious offence. Borrowing an example from Tasioulas, since murder is a far graver crime than littering, one cannot impose the same sanction for both. But ordinal proportionality would allow a year’s imprisonment for littering provided that the penalty for murder is something like life imprisonment or death. cf J Tasioulas, ‘Punishment and Repentance’ (2006) 81 Philosophy, 292.

266

Kimberley Brownlee

Determining which acts of civil disobedience would be proportionate responses to a given objectionable policy is not a simple matter: attention must be given not only to the seriousness of the objectionable policy, but also to the current political climate (eg national emergency), the ability of the government to be responsive if it chose, and so on. Determining the proportionate response by the state to civil disobedience is also not a simple matter. It requires establishing both how serious (in absolute terms) a given act of civil disobedience is and which offences that act of disobedience most resembles. Should a given act of civil disobedience be compared against other politically motivated offences or against other offences classified as, say, mischief or property damage (which may not have been politically motivated)? Assessing the seriousness of the offence requires analysing both the nature of the harm caused or risked by civil disobedience, which typically has no specific victim, and the disobedients’ culpability, which differs from that of ordinary offenders. Since disobedients engage with the state at the level of deeply held conviction, there is reason to treat them more like conscientious objectors than ordinary offenders. While desert and proportionality are central to justified censure, they are not everything. In previous writing on this topic, I over-emphasized desert as the factor that judges should focus upon when they decide whether punishment is appropriate for civil disobedience. I wish to stress here that attention must also be given to prevention of wrongdoing, charity or good will, and restoration. Concerning prevention, Andrew von Hirsch argues in his discussion on punishment that censure itself cannot justify the use of hard treatment; added to the communication of censure must be the aim of deterrence. Only the aim of prevention, says von Hirsch, can justify hard treatment as opposed to mere formal or symbolic punishment. If our parallel between punishment and civil disobedience holds here, then, on this view, the communication of censure against a bad policy would be insufficient by itself to justify the use of civil disobedience; added to the censure would have to be the aim of preventing the development of similar policies. This constraint seems too stringent for civil disobedience since typically people are regarded as most justified in protesting through unconventional channels when the state is least responsive and when success seems unlikely; these dissenters need not aim to bring about a meaningful change in policy to be justified in communicating their opposition. And, according to John Tasioulas, the introduction of deterrence into von Hirsch’s account of justified punishment threatens to render his account incoherent because what is added to censure (namely the prevention of crime) threatens to undermine the communicative character of punishment.²⁷ Thus, deterrence cannot operate as an independent justifying aim for either practice because, first, the case for censuring would be defeated when prevention of further wrongdoing would not result, and second, deterrence would permit or even require modes of censure at odds with treating hearers as rational agents. Our discussion of the constraints imposed by a genuine moral dialogue showed ²⁷ Tasioulas n 24 above, 279–322.

Protest and Punishment

267

that, for both civil disobedience and lawful punishment, there are restrictions upon the means that may be used in pursuit of preventing further wrongdoing. Excessive measures fail to respect hearers as rational agents. This does not mean that all consequential reasoning is ruled out for each practice. Prevention of further wrongdoing can and does figure prominently in the decision by dissenters and the government to engage in these practices. Prevention is part of the aim of communicating censure and of seeking through moral dialogue to effect a change in attitudes and values. A disobedient, for example, must consider the appropriateness of her particular strategy (for instance vandalism, sitins, trespassing, illegal boycotts, illegal strikes) since the specific action that she performs and the manner in which she performs it will determine its effectiveness as the vehicle through which to communicate censure. A person may have reasons for engaging in one form of disobedience, but choose to engage in another form that is not supported by these reasons. For example, she may have an undefeated reason to participate in a road block because this action is well suited to her political concerns and is one that her government and society understand and respond well to or because this action has a public impact that does not greatly harm the interests of others; but she has no undefeated reason, say, to trespass on government property or to engage in strategic violence. In taking the latter actions, she is guilty of a certain error of judgment about which actions are supported by reasons that admittedly apply. Given her error, the best she could claim is that her conduct is morally excused, as she had reason to believe that she had reason to undertake that particular act of civil disobedience. When, by contrast, her civilly disobedient action is supported by undefeated reasons that apply to her situation then, provided she acts for those reasons, her choice of action may be justified. Similarly, not all consequential reasoning is ruled out of the justification of punishment. When deciding on the appropriate punishment, the state must consider how the punishment will be received, that is to say, what form of punishment will most effectively communicate the particular condemnation that the state seeks to convey.²⁸ And when choosing between two punishments that are equally defensible on a desert basis, the state must consider their respective benefits, including their deterrence benefits, to determine which is preferable. Adducing such consequential considerations as part of the justification for punishment reaffirms the parallel between this practice and civil disobedience since instrumental reasons to use the most appropriate and effective modes of communication also bear on the justification of an act of civil disobedience. In addition to consequential evaluation, which plays a role, though not an independent role, in the determination of justified censure, there are two other factors that both judges and disobedients must accommodate to be fully justified in the censure they communicate, namely, charity (mercy) or good will and restoration. Using Tasioulas’s definition, mercy by the state to an offender ²⁸ I thank Antony Duff for highlighting this point.

268

Kimberley Brownlee

is a charitable concern for the well-being of that offender as a potential recipient of deserved punishment. When the law has reason to be merciful toward an offender, this can reduce the justified punishment in relation to the deserved punishment. Repentance is one factor, says Tasioulas, which may give the law reason to be merciful. Since a key aim of punishment, on the communicative theory, is to lead the offender to repent her action and to reform her conduct, when the offender demonstrates prior to punishment that she does repent and has reformed, this gives the law reason to show mercy and to impose a lesser punishment on her than that which she deserves according to justice.²⁹ On this nuanced communicative theory, the notion of moral dialogue is filled out in a meaningful way that takes account of the offender’s contribution to that dialogue. As such, this account offers the most plausible conception of punishment as an attempt to engage in genuine moral dialogue. There may be grounds other than repentance for the state to show mercy toward a given offender, some of which are particularly relevant for civil disobedients. Given the sincerity of conviction and moral consistency that distinguish the paradigmatic civil disobedient from both ordinary offenders and radical protesters, the state has reason to appreciate the onerousness for her of not interfering with nor effectively challenging (through non-radical means) a policy she finds objectionable. Concern for her well-being as a conscientiously motivated offender may give the state reason to be merciful toward her whether her cause is well-founded or not. One might argue that when civil disobedience is well-founded, there are no grounds for mercy, as there is nothing for these offenders to repent. Judges must appreciate (particularly when they sympathize and even share dissenters’ opinions) that civil disobedience paradigmatically is disobedience grounded upon deeply and conscientiously held values and commitments which, quite rightly, make it very hard for their holders either to follow laws which contravene those values or to refrain from communicating in effective ways their objections to laws and policies that they cannot breach directly. Given this, sometimes it would be inappropriate to demand repentance from offenders. Demanding repentance from such disobedients would attack values which the state and society themselves should advocate. However, not all instances of civil disobedience are well-founded. And sometimes, even when disobedients’ cause is well-founded, their choice of civilly disobedient action may invite censure. In other words, although disobedients’ defence of genuine values itself does not warrant punishment, the mode of civilly disobedient communication they employ may deserve censure if, for example, it has significant negative effects for individuals or society (some of which might ²⁹ J Tasioulas, ‘Mercy’ (2003) 103 Proceedings of the Aristotelian Society, 101. This position discounts the suggestion that full punishment must be imposed for the offender to demonstrate the sincerity of her repentance. For a discussion of mercy and repentance see Tasioulas, 2003; Tasioulas, 2006.

Protest and Punishment

269

not be foreseeable). Censure of such offenders may be justified to the extent that they intentionally, knowingly or recklessly brought about harmful consequences through their chosen civilly disobedient action. How much weight these considerations should be given will depend upon the facts of the particular case. The language of mercy and charity is less relevant for civil disobedients than for law-enforcers given the disparity in their coercive power. However, the related language of good will is relevant to civil disobedience. Disobedients have reason to deal fairly with authorities, as this demonstrates their respect for authorities and policy-makers as rational persons capable of responding to the arguments disobedients make against current policies or practices. Good will, however, should not be confused with fidelity to law. Since civil disobedience may be part of a revolutionary project (as in the cases of African protests against the Apartheid South African government or Gandhian resistance in India under British rule), it need not exemplify a fidelity for the political system in which it is undertaken. That said, the constraints of moral dialogue require civil disobedients to treat their interlocutors as rational persons and thus to engage with them on terms of equality and fairness. Although good will and respect constrain disobedients’ actions, these demands do not require that disobedients soften their efforts when they begin to succeed. There are various other ways that disobedients can deal openly and fairly with authorities than by lessening their dissent when the state seems responsive. Co-ordinating with law-enforcers where possible, giving fair warning of their intended actions when doing so will not undermine their efforts to communicate, explaining their reasons for action, behaving well when arrested, attending to the contribution of authorities to this dialogue, and so on, are all ways to show good will. In the same way that the state must recognize that it deals with rational human beings when it punishes them as offenders, so too citizens must recognize that they deal with rational human beings when they challenge the practices or policies of their government and society. When fully deserved censure would be at odds with good will then the justified acts of civil disobedience may be less forceful than those which the policy warrants. Turning to restoration and reconciliation, I suggested above that, although punishment may contribute to the process of restoration and reconciliation, it might not always be the appropriate response particularly to civil disobedience. Although punishment can serve the various communicative aims of a society, it is not the only response that can do so. That said, sometimes the only forms of permissible communication for the state are those involving punishment. However, these forms of communication have the potential to be less severe or impersonal than traditional forms of punishment are. The inappropriateness of imposing customary sanctions on much civil disobedience does not leave the state speechless because there are other responses that the state can make to paradigmatic civil disobedience, even when its particular form is morally problematic.³⁰ ³⁰ During the 2005 British Columbia Teachers Federation’s (BCTF) illegal strike, BC Supreme Court (trial court) Justice Brenda Brown chose not to adopt the customary strategy of imposing

270

Kimberley Brownlee

As John Gardner observes, there are many types of normative consequences for breach of law apart from liability to punishment, such as a duty to show regret, to apologize, to make restitution, and to provide reparation. These actions may be required of disobedients even when whatever wrong they do is also justified. All legal systems, continues Gardner, presuppose that even fully justified wrongdoing has at least one normative consequence: it makes it the offender’s job to offer up what justification she can as a responsible agent who answers for her own wrongs. In the case of justified civil disobedience, this answering comes almost automatically. Through their effort to engage in a moral dialogue with authority, such disobedients offer justification for their actions in the form of an account of their values and their reasons for disobeying the law. When authorities attend to these reasons and when disobedients attend to the reasons the state gives not to resort to certain methods, then their exchange may truly be called a moral dialogue.

IV. The Contrast between Protest and Punishment Let me conclude by noting some ways in which the parallel that I have drawn between civil disobedience and lawful punishment might break down. First, these practices differ in their directness. Whereas there is space within an account of justified civil disobedience for indirect disobedience, there is no such space within the communicative theory of punishment for what one might call ‘indirect punishment’. Indirect punishment could involve either punishing a person other than the offender in order to punish the offender or punishing the offender for some action other than the offence for which she should be punished. In either case, the punishment would be disproportionate and at odds with a predominantly desert theoretic approach which requires that only the guilty be punished and only to the extent that they deserve (and sometimes less than they deserve when considerations of mercy, restoration, or likely benefits reduce the justified punishment). Indirect civil disobedience, by contrast, is not necessarily disproportionate on the communicative theory since sometimes it is a more appropriate and less harmful vehicle

heavy fines on the union, but chose rather to freeze the union’s strike fund, preventing the union from paying its members their daily strike pay. In this precedent-setting decision, Justice Brown argued that the BCTF acts through its members to commit the contempt of the court order that required them to return to work and that the BCTF is using its assets to facilitate the continuing breach of that order in part by paying teachers. Her response to the union was to impose a punishment of sorts in denying members their strike pay, but not the traditional punishment of a fine that the union had expected (and that it received a week later after it continued to act in contempt of the court order). In making this judgment, Brown stepped around the union and addressed teachers directly. Many teachers took the judgment as a challenge to them to show the strength of their convictions, that is, to show whether they would continue the strike without their strike pay. See BC Public School Employers Association v BC Teachers Federation [2005] BCSC 1446.

Protest and Punishment

271

for communicating censure than direct disobedience is. Also, sometimes it better communicates the disobedient’s objection than direct disobedience would do. Second, there is a difference in how the practitioners of punishment and civil disobedience conceive of their own activities. Whereas the state (in a democratic society) may with some legitimacy represent itself as the voice of the people, disobedients may not do this unless the minority for which they speak is in fact the majority. However, they may claim to speak for the interests of the community when their communicative condemnation and dissociation serves the genuine ideals and values of that society. Third, a related difference pertains to the attitude that each adopts toward their activity. It is often argued that the state should adopt an attitude of regret when it is required to punish its citizens. No comparable attitude of regret may be expected from civil disobedients for the mere act of dissociating themselves from a law or policy they oppose. They have reason to regret any harm they cause through their action or duties they breach by their action, but not to regret the mere act of breaching the law.³¹ Fourth, disobedients and authorities also differ in their expectations of how the target audience will respond to the communication. The disobedient, presumably, does not expect repentance as such from government officials or policymakers in the way that the state expects repentance from an offender. However, when her challenge is well-founded and her dissent persuasive, she may reasonably expect an acknowledgement of wrongdoing by the government or the institution. Moreover, she may reasonably expect reformation in policy if not in officials’ outlooks. Certainly, there are other points of contrast between the communicative features of punishment and those of civil disobedience not noted here. While interesting, these warps in the parallel I have identified do not diminish the usefulness of this parallel both for explicating the key features of these two practices and for analysing their respective justifiability as contributions to a genuine moral dialogue. ³¹ I have argued elsewhere that there is no good reason to hold that there is a general moral obligation to follow the law.

15 Apology and Reparation in a Multicultural State Christopher Bennett

In this paper I raise a problem for a particular version of the communicative theory of punishment, one that gives central place to the notions of apology and reparation. I argue that a condition on the viability of such a theory for a pluralistic or multicultural state is that citizens of such a state are able to reason together about questions of appropriate emotion: that such questions, in other words, can be resolved by the public reason of that state. I reach this conclusion as follows. First, I argue that communicative theories of punishment should give a central place to judgements about the expression of emotion. Second, I suggest that judgements about the expression of emotion, though not subjective, are based on ‘thick’ concepts or ‘ways of seeing’ that are embedded in particular local cultures. This seems plausibly true, at any rate, of the emotions that are invoked in justifying our strongest version of the communicative theory of punishment. Third, I argue that when we think about how the communicative theory might apply to state institutions, we ought to accept what has been called the ‘liberal principle of legitimacy’: that the exercise of the coercive power of the state ought to take place on principles that can be justified to all. But then, fourth, we ought not to adopt a theory of punishment that rests on thick concepts that are compelling only within some of the cultures represented in a multicultural state. The communicative theory of punishment will be applicable to a pluralistic state, on the terms of this argument, only if the citizens of such a state have the means by which to resolve questions about appropriate expression of emotion—the means by which, that is, they could come to a conclusion the validity of which all could in principle recognize—or that they could together recognize to be justified.

I. Let us start with the thesis that judgements about emotions are not merely subjective. Janice confesses to feeling jealous that her partner would rather spend

Apology and Reparation in a Multicultural State. Christopher Bennett. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Apology and Reparation in a Multicultural State

273

a weekend at a conference than at home with her. Janine tells her friend that her jealousy is quite unreasonable, that she is lucky to have a partner who is so devoted to her, and that it is just part of the job of an academic to have to go to conferences that are sometimes on at weekends. Here Janine makes a judgement about what attitude or emotion is reasonable or appropriate to the situation, and she puts her point in a way that suggests she takes herself to be in the right about what she is saying. She assumes, that is, that there is such a thing as correctness and incorrectness in such judgements. Another example: Bill sees a parent in a supermarket shouting angrily at a young child who is running amok and making a mess; and he says to him ‘Hang on. You may be angry, but you don’t need to be that angry.’ Here Bill might quite agree that anger of some sort directed towards the child (or at any rate, anger that the child is doing these things) is quite appropriate. What disturbs him is the level of the parent’s anger at what is really a rather trivial case of naughtiness—and (what is not an entirely separate question) the way the person expresses it. These judgements about which emotions are appropriate to particular situations, what level is reasonable and how they ought to be expressed, are quite familiar and show that we do not take such matters to be entirely subjective: we take it that they can be the subject-matter of rational criticism and argument (that is, one can evaluate and criticize someone’s reasons for being angry or jealous, or their reasons for being this angry or so angry that they would do that). Of course, as in any other area of rational argument, there can be hard and controversial cases; but on the other hand there can be perfectly clear cases. It is not the case that all disagreements about appropriate emotion will be rationally resolvable, though many will. This is hopefully all we will need for our argument about the communicative theory of punishment. Arguments about appropriate emotion, like other sorts of argument, are rationally resolvable in many cases because we share an understanding of some paradigm cases in which such emotions would be appropriate or inappropriate, and can extend our understanding to cover new cases. It is paradigmatic, for example, that grief is appropriate over the irrevocable loss of something or someone that, like a friend, is both irreplaceable and of great importance to one and one’s sense of the value of one’s own life. This is not the only kind of thing about which one can feel grief, but it is clear that grief is appropriate in such cases. Then there are cases in which grief seems unintelligible. For instance, imagine feeling grief because the sun rose, as predicted, at 7.27 am. It is hard to imagine how this fact could be a possible object of grief, since it would be hard to understand what it would mean—or be like—to experience grief over the sun rising at its regular time. Grief is unintelligible in such cases because the object (the sun rising) seems to bear no meaningful or explicable relation to paradigm intelligible cases of grief. Of course, in this case we can make the example intelligible by telling a story about how the sun rising at 7.27 signifies something that does have some meaningful connection to a paradigm case of grief. Maybe, for instance, the sun

274

Christopher Bennett

is rising on a day on which one’s friend is to be executed by a tribe of sun-worshippers, and the chief of the tribe had made it known that he would have treated any irregularity in the sun rising as grounds for a reprieve. In that case it would make sense to feel grief over the sun rising at its appointed time. But this is because we now have a different understanding of what the grief is really about. We now have an understanding of how the sun rising can be an intelligible object of grief. Because we share an understanding of these paradigm cases of the intelligibility of an emotion, we have a sense of the ‘shape’ of the emotion. We have some sense of the sort of thing that ties these paradigm cases together, such that we can go on to identify new cases that the emotion ‘fits’. This shared sense is what allows Janine and Janice to have a rational disagreement about whether Janice’s jealousy is justified. They have to determine to what extent her case really is one that fits within the overall shape of the emotion. Thus Janine implicitly accepts that there are cases in which jealousy would be appropriate—cases for instance, in which one’s partner’s absence were a real sign of lack of devotion—and in doing so she appeals to a shared sense of the shape of the emotion. Judgements about emotion are therefore not merely subjective. But they do rely on a shared sense of the contours of a certain emotional response, a shared sense of the sorts of things towards which it makes sense or is appropriate to have this sort of response. And for this reason it could be said that judgements about appropriate emotion are tied to particular cultures or ways of life: they are part of the inheritance that people living together and sharing a certain way of life share. For the shape of an emotion is given by a shared sense that certain things matter. Grief and jealousy take their shape from the fact that things matter to us in particular ways. But different cultures have different priorities; they have different world-views and different understandings of the moral order, different understandings of their place in the scheme of things. Maybe we should expect, therefore, significant variation in emotional responses across cultures. At any rate we cannot simply assume that how it will make sense to feel about some event in one culture will also make sense in another. And it follows from this that we also cannot assume that judgements about appropriate emotion can easily cross cultural boundaries; or that questions about which emotion is appropriate can be rationally resolved by members of very different cultures.

II. Judgements about appropriate emotion are not subjective in the sense of having no criterion of correctness beyond the individual subject; but it may be true that they are tied to particular cultures. How is this point relevant to the discussion of punishment? It becomes relevant once we come to think about the communicative, expressive or symbolic dimensions of punishment. One of the central problems that a communicative theory has to solve is why the moral message to

Apology and Reparation in a Multicultural State

275

the offender (or to others) has to be communicated through punishment rather than some non-punitive form of symbol—or whether the justification for ‘hard treatment’ comes from some other purpose, such as deterrence.¹ And one way in which defenders of the communicative theory have tried to justify hard treatment is to claim that, because what has to be communicated is condemnation, it is symbolically inadequate to do this without punishment.² The argument for this claim might go something as follows. In informal or non-legal contexts, the expression of condemnation is an expression of certain emotions (outrage, indignation, resentment, etc). Condemning in a purely emotionally neutral manner would not be full condemnation: it would undercut its own message since it would suggest that the offence was to be condemned but was not the object of the relevant emotions. If the state is in the business of expressing condemnation then it has to find appropriate symbols through which to do so. These symbols can be found by looking at the characteristic expression of the emotions that are appropriate in cases of wrongdoing. What we need to find is therefore some institutional expression of the appropriate reactions to wrongdoing, such as indignation, blame, outrage, and so on. This suggests that we might look at punishment as an institutional expression of indignation or outrage at an offence (these emotions expressed on behalf of the people as a whole). This already suggests that the communicative theory will give a central place to judgements about appropriate emotion. But there is a problem with understanding punishment as the expression of emotions like outrage and indignation. For these are emotions the characteristic expression of which tends towards some aggressive response towards the offender: attacking or at any rate taking hold of the offender; raising our voices; slamming doors, etc. In the normal run of things we have to be very careful when we express outrage. Emotions run high, as they say, and usually we take it that it is inappropriate to get so outraged with someone that we cause them physical pain or discomfort. And yet this is precisely what happens in punishment. In punishment the effect is to impose on the offender something that consists in hard treatment, or at any rate is a deprivation of some of the things they value. But if we do not take ourselves to be justified (in our considered moments) in doing violence to the offender in non-legal cases, why should we be justified in the legal case? This question seems hard to answer without undermining the idea that the punishment really is simply the institutional expression of the emotion, with the same basic logic of appropriateness. Someone who wishes to defend the idea that punishment is or ought to be the expression of outrage therefore has ¹ See, eg J Feinberg, ‘The Expressive Function of Punishment’ in J Feinberg, Doing and Deserving (Princeton, 1970) 95–118; A von Hirsch, Censure and Sanctions (Oxford, 1993). ² See, eg H Moberly, The Ethics of Punishment (London, 1968) esp ch 8; I Primoratz, ‘Punishment as Language’ (1989) 64 Philosophy 187–205; J Kleinig, ‘Punishment and Moral Seriousness’ (1991) 25 Israel Law Review 401–21; RA Duff, ‘Punishment, Communication and Community’ in Matravers (ed), Punishment and Political Theory (Oxford, 1999) 48–68.

276

Christopher Bennett

the burden of explaining why such a violent response is justified in the non-legal case (or at any rate that it can be in the case of very serious crimes). However, I will take it that it is far from clear that such a violent response ever is justified. Perhaps all that the expression of outrage would justify us in doing to the offender would be to tell them forcefully that what they did was wrong: I am not sure that it would justify actually imposing hard treatment. Therefore it is not clear that punishment would be justified if it were understood as an appropriate expression of emotions like indignation and outrage. The theory that the state ought to express collective condemnation of crimes relies on being able to find an adequate institutional form for condemnation to take. Communicative theorists of punishment share the intuition that punishment has this function, but we have seen that it can be difficult to find emotions that would appropriately be expressed through the infliction of punishment. However, what follows is a suggestion for solving the problem. Rather than looking at how one might express indignation or outrage, one might turn the question around and think instead about expressing condemnation by conveying a view of how the offender herself ought to react to what she has done. The idea on this theory of punishment would be that the state should require offenders to undertake proportionate penance or reparation for their offence, reparation that would express the appropriate attitude of being sorry.³ Proportionate penance consists in the amends that they would be motivated to make were they really sorry. By requiring offenders to undertake the reparation they would undertake were they genuinely sorry the state would be expressing its condemnation of their offence in a symbolically appropriate way because it would be conveying a message about how sorry the off ender ought to be for what they have done. In order to see how this works we need a different sort of story about appropriate emotion and its appropriate expression. We need to see what being genuinely sorry consists in and why it is best expressed in apology and reparation.

III. The basic idea is that in social life we engage in relations with others that exert certain normative demands on us. Being what we are, however, we do not always meet these demands, or treat one another as we ought given the relationships we have: nothing is more familiar than this fact of occasional disrespect. Infractions such as these would threaten the existence of our social relationships—given that it seems that these relationships are constituted by the existence and mutual ³ This idea is based on an aspect of the penitential theory of punishment put forward by RA Duff in Trials and Punishments (Cambridge, 1986); and RA Duff, ‘Punishment, Communication and Community’ (Oxford, 2001). Where it differs from Duff ’s account is that the punishment on this view symbolizes the appropriate penitential attitude (and thereby expresses condemnation in a symbolically adequate way), whereas for Duff punishment ought also to be instrumental in evoking the penitential attitude.

Apology and Reparation in a Multicultural State

277

recognition of these normative demands and ties—were there not a way of restoring the relationships in the event of infraction. Perhaps it seems hard to imagine a form of social life in which there were not such normatively demanding relationships and in which there were not a procedure or ritual for restoring relationships in the wake of infraction. In our culture that mechanism is apology.⁴ I will briefly describe how apology works before turning to the question of criminal justice. The reason wrongdoing damages and threatens to destroy relationships is because of the attitude (towards the victim or towards a set of rules and values) that it expresses. When Jemima steals Jim’s car, for instance, she treats the car as hers for the taking; she treats the fact that it is his car as irrelevant to her taking it; she treats the normative demand that one asks before taking as having no claim on her. This threatens her relationship with Jim, but also raises a question about her place in social life more generally, since it raises a question about her attitude towards the basic rules and values that structure our interaction. Participation in ‘ordinary, adult, interpersonal relationships,’ as PF Strawson points out, rests on the expectation that participants will show one another a certain degree of good will or respect.⁵ Otherwise put, such relationships are based on trust rather than surveillance, supervision or coercion. When the expected good will is not forthcoming the nature of the relationship—for instance, whether trust still exists between us—is put into question. Apology, as is often claimed, can restore trust to these relationships. But in order to see how it does so we have to look again at what it expresses. Apology works to restore relationships because it shows that the person feels sorry. But what is involved in feeling sorry and why does it have such an important role in restoring relationships? This may seem mysterious until we see that we can replace the idea of feeling sorry with the more informative term repentance. Apology works because and insofar as it expresses a remorseful recognition that the initial action was wrong; a repudiation of the attitude expressed in the wrong; and a renewed commitment to the rules and values that structure the relationship.⁶ Words are cheap and have no magic power in themselves. Saying sorry works because and insofar as it is an expression of a new attitude, a change of heart, in which one sees the error of one’s ways, takes responsibility for it and rejects it, and determines to change for the future. The story of apology is therefore a story of redemption.⁷ Transgression damages social relationships because these relationships rest on the assumption of some ⁴ See E Goff man, Relations in Public (London, 1971) ch 4 (on apologies as ‘remedial interchange’); N Tavuchis, Mea Culpa: Towards a Sociology of Reconciliation; A Bottoms, ‘Some Sociological Reflections on Restorative Justice’ in A von Hirsch, J Roberts, AE Bottoms, K Roach and M Schiff (eds), Restorative Justice and Criminal Justice: Competing or Reconcilable Paradigms? (Oxford, 2003) 79–114. ⁵ PF Strawson, ‘Freedom and Resentment’ in G Watson (ed), Free Will (Oxford, 1982). ⁶ See, eg Duff, Punishment, Communication and Community 94–5. ⁷ For the story given here, see for instance writers as diverse as Goff man, Relations in Public; Duff Punishment, Communication and Community; and R Swinburne, Responsibility and Atonement (Oxford, 1985).

278

Christopher Bennett

sort of trust in shared commitment, some sort of mutual respect or recognition; and wrongdoing expresses an attitude contrary to respect and trust. Repentance and apology allow such relationships to be restored because in demonstrating a new attitude a person demonstrates their trustworthiness and allows themselves to be included once again in the community of people with whom one can form trusting relationships. How does reparation enter the picture? The crucial thing to restoring relationships is that the person’s actions express an adequate understanding of the significance of what they have done, that they now have a firm understanding of the importance of the values they violated and that they can now be trusted to behave as they ought. As I have said, words are cheap and when a person has done some serious wrong a merely verbal apology will not be enough to show that they fully recognize the significance of what they have done. There seems to be a notion of proportionality at work here: in order to express one’s remorse at having done some minor wrong an apology might be enough, but in order to express one’s remorse at having done something more serious a more serious undertaking is required. Remorse over serious wrongs can therefore only be expressed through undertaking some sort of penance or atonement. I believe that this represents a promising way in which to understand the communicative theory of punishment. In imposing a certain amount of reparation the punishing authority symbolizes what the offender’s attitude to the offence ought to be, since this is an attitude that would be expressed in the willing undertaking of that amount of reparation. By imposing a certain amount of reparation on the offender the state communicates to the wrongdoer the idea that she has committed a wrong of a certain degree of seriousness. It condemns the offender by communicating the idea that she should have a repentant attitude towards that wrong that should express itself in her undertaking that amount of willing reparation. The idea of imposed reparation is therefore a way in which we can explain why hard treatment has to be imposed on the offender if our business is to express the proper condemnation of her for her action.

IV. What the preceding account of apology aims to do is to explain and vindicate—to articulate—our sense that feeling sorry, saying sorry and making amends can be called for in cases of wrongdoing. I have tried to explain how apology restores relationships damaged by wrongdoing. However, in trying to explain this I have had to resort to language (repentance, redemption, penance, atonement) that is usually at home in religious contexts: specifically in the Christian or Roman Catholic tradition. I do not think, however, that this undermines my account as an explanation of the rituals of our social life, since it might simply be that this vocabulary is necessary to give a full articulation of these social forms and

Apology and Reparation in a Multicultural State

279

of why we find them important. However, this fact does need to be explained, and it might be said that the explanation is not hard to find: this is that our sense that these reactions are appropriate was forged in—and has been passed down to us from—a culture in which the Roman Catholic understanding of sin, self and religious community was pervasive. The thought would be that my analysis exposes the religious heritage of reactions and responses that we act out without a second thought, that are part of our second nature. If this argument is plausible then it suggests that the efficacy of arguments about appropriate emotion, in this sphere at least, may be culturally local. In explaining the workings of apology as I have done I have clearly appealed to a shared sense that the demands of apology are compelling in order to suggest that the shape of these demands is best explained by a story about repentance and atonement. I stand by this story as the best understanding of our (considered) intuitions, our paradigm cases. But the line of thought we are considering here is that this resort to religious terms to articulate our sense that apology and reparation is ‘called for’ makes the cultural provenance of these norms clear. Only those whose cultural education contains elements of Catholic or post-Catholic understandings of sin, self and community are likely to find these arguments about why apology and reparation are necessary in the event of wrongdoing at all plausible. Let us return to the communicative theory of punishment. The idea was that we were looking for an institution that would communicate, in a symbolically adequate way, the appropriate condemnation of an offender for an offence. My claim was that we could find such adequate symbols in the practice of apology. We have a sense that the seriousness of a wrong one has committed ought to condition how sorry one feels, the degree of one’s guilt. And if one feels sorry about doing some very serious wrong one will feel motivated to undertake more onerous amends than one will if one feels sorry about a minor wrong. Therefore in imposing amends of a certain amount on an offender, the state would communicate a certain degree of condemnation, for it would thereby symbolize how sorry the offender ought to feel, or how serious she should understand her crime to be. This account finds the appropriate symbolism of punishment in the appropriate emotions of feeling sorry for what one has done. But now we have raised a question about the cultural specificity of these symbols. The problem is that these symbols will seem compelling only to those who share the practice of apology on which they are based. And, if we accept the line of thought entertained in the previous paragraph (and we will consider some reasons for rejecting it below), the practice of apology, although it is second nature to those engaged in it, looks to be the heritage of a certain culture rather than part of a universal moral consciousness. This conclusion is relevant to our present inquiry because here we are interested in the question of whether apology and reparation—at least in the shape of the communicative theory of punishment—should be made a cornerstone of criminal justice in a modern liberal pluralistic state. And the problem raised by this paper is that we ought to reject the use of any such ideas about what is ‘called

Christopher Bennett

280

for’ in the wake of wrongdoing as the basis of a public policy in a multicultural state since they are drawn from a culture that is only one of those represented among citizens. In the following section we go on to look at the reasons for this rejection.

V. Political liberals take it that the actions of the state ought only to be justified in terms of a ‘minimal moral conception’ rather than our ‘thick’ ideas about virtue and the good.⁸ One can reach this conclusion on the basis of two assumptions. First of all, a certain conception of political legitimacy, according to which coercive state action has to be justified in terms that all citizens are in some sense capable of accepting.⁹ The guiding intuition here is that the power of the state should belong to all rather than to a particularly powerful group; the body of citizens should be collectively self-governing rather than one group ruling another. However if the power of the state is to belong to all citizens, if it is to express a common political will rather than becoming a weapon of one sectional interest against another, then it has to be guided by principles that all can affirm. State action has to be action that citizens can understand as action they take together. It therefore has to be justified to citizens in terms that they are not excluded from accepting. If the state is to be more than a matter of the majority culture ruling over minorities then the principles on which the state’s coercive power is exercised have to be neutral in the sense that they can in principle be accepted by all the various cultural groups that make up the population. The basis of the justification for political action has to be found in an ‘overlapping consensus’ to which all parties can in principle subscribe.¹⁰ However, secondly, political arguments based on thick terms are, as we have already seen, unlikely to be acceptable to the variety of citizens in a multicultural state. Thick terms are taken to be rooted in particular cultures, a certain shared form of life. As Michael Walzer describes them, thick terms are ‘richly referential, culturally resonant, locked into a locally established symbolic system or network of meanings.’¹¹ The implication is that the import of thick terms can only be grasped by those who see the world in a certain culturally-specific way.¹² In modern conditions, for these political liberals, the state is not a moral community, sharing a thick moral vocabulary: rather a legitimate political community has to be forged from diverse constituents using less controversial materials. ⁸ ⁹ ¹⁰ ¹¹ ¹²

See, eg C Larmore, Patterns of Moral Complexity (Cambridge, 1990). See, eg J Rawls, Political Liberalism (New York, 1993) 137. The term ‘overlapping consensus’ is drawn from Rawls, Political Liberalism. M Walzer, Of Thick and Thin: Morality at Home and Abroad (London, 1994) xi n 1. cf A MacIntyre, After Virtue (London, 1985) chs 2–3.

Apology and Reparation in a Multicultural State

281

The conclusion of this argument would be that, even were it the case that the story about apology that I have given articulates something present in the cultural consciousness of the majority or indigenous culture of a state like the United Kingdom it would not be acceptable to use this as the basis of a public institution like punishment when the state is made up of many other minority groups who ought to be included as full citizens and whose cultural understandings ought not to be excluded by the public institutions that we share. Basing the system of punishment on a basically Christian understanding of guilt and redemption, it might be argued, would be like keeping shops shut on Sunday because that is traditionally the Christian day of rest. It would only be legitimate to shut shops for that reason if it were legitimate for those with a Christian heritage to rule over those who do not. But those guided by the aspiration to genuinely collective selfgovernance should reject this idea. And hence this raises a problem for the communicative theory I have sketched. For the communicative theory embodies a view of what an offender ought to feel—that they ought to feel sorry—and that they ought to express this feeling sorry in certain types of reparative action; that this is what would lead to redemption or satisfactory reconciliation with the rest of the community. And the account I gave makes use of concepts such as repentance and atonement that are distinctive of the Christian religious tradition.¹³ Of course, I have argued that these terms do not need a particularly religious theoretical background in order to be compelling: I have claimed that these religious-sounding concepts are deeply embedded in our profane social relations. But the problem is that, as I said before, this might be taken to show the (at least partly) religious heritage of our conception of the roles and relationships of social relations. It shows that in explaining our sense that something like apology is called for we need to appeal to ideas that are not universal rational requirements: they are part of the deeply embedded ethical outlook of certain societies, certain forms of life. The story about the importance of apology is compelling to us—or so I would argue—but the ‘us’ to whom it is so compelling is culturally and historically located. My argument makes use of the resonances and symbols characteristic of thick and culturally local concepts (concepts articulated within a certain tradition that lose their meaning when removed from that context).¹⁴ If this is true it is not clear that such concepts can be justifiable to all in a multicultural society. It might be argued that in drawing up a theory of punishment that will allow for legitimate ¹³ One might deny that they are concepts belonging exclusively to a certain tradition, of course. See, eg A Etzioni and D Carney (eds), Repentance: A Comparative Perspective (Lanham, 1997). I consider this objection later on. ¹⁴ Indeed Duff specifically claims that thick moral concepts have a central role in mediating between the formal language of law and legal procedure and the moral understandings of citizens, mediation that is essential to the legitimacy of law. Thus for the law to justify its demands on citizens in terms that they can reasonably be expected to affirm, it cannot dispense with thick moral concepts. However, if this were true legitimacy could only be achieved in communities that share the relevant thick concepts. See Duff, Punishment, Communication and Community 190–1.

282

Christopher Bennett

state action in conditions of ‘deep pluralism’ we need rather to work with thinner materials than the rich social meanings deployed here. At this point I want to dismiss quickly one argument that might seem to be capable of coming to our rescue here. This is an argument that is often made by restorative justice theorists and it appeals to the alleged universality of restorative traditions. A number of writers argue that all societies have something resembling a practice of restorative justice to deal with conflict.¹⁵ Indeed Braithwaite has written: ‘Restorative justice has been the dominant model of criminal justice throughout most of human history for all the world’s peoples.’¹⁶ If this was true, would it mean that we could deny that the idea of using apology and reparation as a basis for criminal justice must rest on certain thick concepts that are peculiar to particular cultures? If so we could see how procedures based on apology and reparation could be justifiable to all in a multicultural society: they would form a common vocabulary between cultures that could form part of the overlapping consensus. But is it true that our account of apology and reparation is not culturally local? It may well be plausible to argue that all societies have something resembling apology. Societies are based on co-operation; co-operation requires rules; rules can be broken. Thus it is hard to imagine a society that lacked any notion of ‘wrongdoer’. Similarly, it is hard to imagine a society in which the action taken against that offender is not in some way negative. And finally, it might be said that all societies must have some way of bringing the offender back into the community, signalling that the period of disapproval is over. Thus it might be argued that some structure of crime-as-rupture/restoration-as-healing is to be found in all societies, simply because of the need for societies to exist as continuing entities bound by shared rules and normative relationships.¹⁷ We might, indeed, seek to go further. For instance, we have stated that two of the central concepts on my account of apology, repentance and atonement, belong to the Christian tradition. But in a collection of essays edited by Etzioni and Carney, a number of writers trace the influence of the idea of repentance in Christianity, Judaism, Islam, Buddhism and Hinduism.¹⁸ If these writers are to be believed, it is not only the basic structure of rupture and restoration, but even ideas specifically of repentance are not just the province of Christian and postChristian societies.

¹⁵ J Braithwaite, Crime, Shame and Reintegration (Cambridge, 1989) 58–9; EG M Weitekamp ‘The History of Restorative Justice’ in G Johnstone (ed), A Restorative Justice Reader: Texts, Sources, Context (Cullompton, 2003) 111–24; and other readings in part B of Johnstone, A Restorative Justice Reader. ¹⁶ J Braithwaite, ‘Restorative Justice: Assessing Optimistic and Pessimistic Accounts’ in M Tonry (ed), Crime and Justice: A Review of Research vol 23 (Chicago, 1999) 2. ¹⁷ J Skorupski, ‘Freedom, Morality and Recognition: Some Theses of Kant and Hegel’ in J Skorupski, Ethical Explorations (Oxford, 1999) 179–84. ¹⁸ Etzioni and Carney (eds), Repentance 13 above.

Apology and Reparation in a Multicultural State

283

However, I do not think that this is enough to get around the problem posed by my original argument. For while there may be something in the way of ideas of apology and reparation in many or most cultures, I do not think we would want our state to adopt any version of the idea of reparation that did not include certain important elements that are no doubt culturally local. For instance, most of us would take it that an adequate conception of restorative justice would have to be underpinned by a conception of responsibility that requires not only actus reus but also mens rea, though this distinction seems not to be a cultural universal. It is central for us that a person cannot be asked to make amends for what was done by someone else, or what they did in (non-culpable) ignorance or by (nonnegligent) accident.¹⁹ This is in contrast to many cultures for which an agent can or must be punished for evils committed accidentally, in ignorance, or by her ancestors. To take another example, it would surely be unacceptable if our practices of punishment or restorative justice were not structured by certain ideas about what it is reasonable to ask an offender to do by way of restorative action. We should reject, for instance, the idea that penance can consist in certain kinds of self-mortification, or physical self-punishment, or that any sort of corporal punishment could make amends, however willingly undertaken. Perhaps it is part of our secularism that we tend to think that the response to crime from the offender has to be in some way constructive and orientated towards the victim. Yet again this is not a cultural universal. These examples show that, although there may be some universality to restorative traditions, it is clear that the version we would wish to endorse is one detailed variation. Therefore we cannot straightforwardly argue that some form of reparative justice could be acceptable to all in a multicultural society on the grounds that all cultures have such a notion.

VI. If the argument developed in this paper is convincing then may be we should reject any idea that apology and reparation could play a role in the criminal justice system of a modern pluralistic state. However, there is another possibility. Because it is ubiquitous and because it is second nature it seems that the practice of apology is something that could usefully be harnessed by a criminal justice system that seeks to achieve legitimacy in the eyes of the people. In other words, if my analysis of the phenomenon of our damaging and restoring relationships is correct then it would help citizens to see that justice was being done if the criminal justice system reflected their own sense of what is ‘called for’. This idea that ¹⁹ The need for the terms in brackets in this sentence show the complexity of the conception of intentional action on which this distinction is based. Nevertheless I think that most of us would accept the reality of some such distinction. But the more complex it becomes, the more culturally local it is.

284

Christopher Bennett

our informal social relations already contain important practices for constructively dealing with wrongdoing, practices that aim at reconciliation rather than retribution, and that are humanistic rather than merely administrative or bureaucratic, is one of the motivations of the restorative justice movement.²⁰ Some writers on restorative justice claim that formal legalistic processes of trial and punishment at best distort and at worst discard entirely the informal means at our disposal for dealing with offences; and that the way forward for criminal justice is to become more informal, to do away with the cumbersome bureaucracy of criminal justice and instead to bring in more community involvement, involvement that would hopefully aim at apologetic reconciliation—and reparation for victims—rather than punishment and its resultant problems. The reason why adopting an informalist restorative justice approach might be helpful at this juncture is that it might allow the strongly intuitive practice of apology and reparation to play a role in criminal justice after all. The problem for the communicative theory of punishment arises because on that theory the state, in punishing, is understood as issuing condemnation on behalf of the political community as a whole. Punishment is a state action, and in acting thus the state employs and thereby endorses a certain set of symbols that, as we have seen, are likely to have a specific cultural provenance. This is why there is a problem about how, in conditions of multiculturalism or deep pluralism, the political community as a whole can endorse this set of symbols. However, the restorative justice theorists in whom I am interested argue that criminal justice ought to be done in a more informal way. On their view the state ought to allow victim and offender to bring their own cultural preconceptions—however differing—with them to a meeting in which they resolve how to deal with the aftermath of an offence. The state ought not to dictate in advance what response the offender ought to make to the victim: that is a matter for the meeting to decide. This is because the notion of restorative justice emphasizes the involvement of the ‘stakeholders’ to a crime (those who are most directly involved with or affected by it: say, the victim and offender; close relations or friends of these two; members of the local community); these stakeholders are to take responsibility for dealing with the offence and deciding what has to be done as a result of it. It is a strong theme of the restorative justice literature that when the state takes the responsibility for dealing with crime away from citizens it ‘disempowers’ them and that its clumsy interventions are rarely adequate to the nuances of particular situations. It is better if citizens are allowed to take responsibility for these matters themselves. This feeds into a wider political agenda concerning active citizens and active communities.²¹ For our purposes, however, the crucial aspect is that, if ²⁰ See, eg N Christie, ‘Conflicts as Property’ (1977) 17 British Journal of Criminology 1–15; Bottoms, ‘Sociological Reflections on Restorative Justice’; M. Wright, Justice for Victims and Off enders: A Restorative Response to Crime (2nd edn, Winchester, 1996). ²¹ This is made explicit, eg in Braithwaite’s ‘republican’ account of criminal and restorative justice. See, eg J Braithwaite and P Pettit, Not Just Deserts (Oxford, 1990).

Apology and Reparation in a Multicultural State

285

citizens are left to negotiate the specific nature and levels of reparation for themselves there is no need for the state to take any stand on how the offender should (ideally) react. Therefore the state in restorative justice need do nothing that in that way favours one culture over another. Thus one of the leading theorists of restorative justice, John Braithwaite, writes that the state has a duty to set and presumably enforce the ground rules for meetings between the stakeholders to a crime, but these ground rules are simply to ensure fair discussion. The state ought not to do anything to second-guess the result of the discussion or to set the nature or level of reparation to be agreed by the meeting. He is explicit that this allows the state’s involvement in the meeting to be neutral as to the various cultural groups that might be represented in such a meeting.²² The design of restorative justice institutions can be rather minimalist. A conference, for example, can be defined as a strategy for who is invited and a small number of procedural rules about advising the defendant of a right to leave and take their chances in court, speaking in turn, and so on. Even ‘speaking in turn’ may be too Eurocentric to be a minimal requirement as in some cultures it shows polite engagement to finish another person’s sentence or speak at the same time. Perhaps the ideal is undominated speech. The ideal is certainly not to be culturally prescriptive: to allow participants to begin and end a conference with a prayer if that is their wish, to include noisy babies if they wish or exclude them if they wish, to allow Samoan offender to kneel at the feet of victims and First Nations Canadians to wash the feet of victims . . .²³

If the state gives up any role in dictating what response it is appropriate for the offender to make as a result of their crime, if it passes this responsibility over to citizens, then there is no need for it to take any stand on the problematic issue of what reactions or emotions are appropriate in such situations. There would therefore be no question as to how such a stand can be reconciled with the requirement that a liberal state act on behalf of the community as a whole rather than the majority or most powerful groups within it. But should it give up this role? The problem is that if it gives up the role of dictating what response it is appropriate for the offender to make then, by the arguments I have given previously, it also loses the ability to issue symbolically adequate condemnation of the crime on behalf of the political community as a whole. And the proponents of the communicative theory of punishment are right in thinking that this would be a serious loss to a political community. The central insight of the communicative theory is that there are certain wrongful acts that, by virtue of their public nature or their inherent seriousness, merit the condemnation of the public as a whole. This is the category of ‘public wrongs.’²⁴ We might say that they are wrongs through ²² See also the aspiration to ‘culturally plural justice’ in Braithwaite, ‘Restorative Justice’ in M Tonry (ed), The Oxford Handbook of Crime and Punishment (Oxford, 1998) 323–44. ²³ Braithwaite, ‘Restorative Justice: Assessing Optimistic and Pessimistic Accounts’ 85–6. ²⁴ See, eg Duff, Punishment, Communication and Community 60–3.

286

Christopher Bennett

the commission of which an offender damages their relationship, not just with a particular victim, but with the community as a whole, and which therefore call for an apology to the community as a whole. The proponent of the communicative theory will claim that the idea of a restorative justice that effectively devolves decisions about sentencing to citizens cannot express an authoritative collective view about the wrongness of such acts: it cannot provide a forum in which citizens together express their condemnation of these crimes. Yet it is precisely this sort of collective condemnation that these crimes call for; and the state that fails to make such official denunciations is failing in an important responsibility towards the victims of such crimes. Now a proponent of restorative justice might respond that her view could have a place for collective condemnation being issued against an offender. This could be the moment at which the offender is officially pronounced guilty—and this theorist might argue that this is enough to satisfy the need for the offender to be condemned by the public as a whole.²⁵ However, the problem with this suggestion is that it would mean that the public condemnation of a thief would have the same symbolism as the public condemnation of a murderer: both would be simply pronounced guilty of a crime. Of course, it might also be verbally expressed that the murderer is guilty of a very serious crime whereas the thief is guilty of a more middling one. But this will not satisfy someone who is swayed by the view canvassed earlier, that condemnation is not really condemnation unless it acknowledges the action as an appropriate object of the relevant emotions. On the communicative theory I am considering it is the nature and level of the sentence that is essential to expressing the requisite condemnation because this is what signifies how sorry the offender ought to be. The restorative justice view, even if it can accommodate some form of official condemnation, cannot condemn properly since it says nothing about how sorry the offender ought to feel. So we are returned to the crux of the matter. The problem is that if we agree with the proponents of the communicative theory that there are some public wrongs that have to be condemned as such, and that the symbolically adequate way to condemn them is to require the offender to make such reparations as would adequately express his repentance, then it seems that we can only condemn public wrongs if we are able to take a stand on what level of reparation is appropriate to them. However, if judgements about the appropriateness of repentance and so on are not likely to be part of public reason then it seems that such symbolically adequate collective condemnation is both morally required and—in conditions of deep pluralism—morally inappropriate. An institution on which some official condemnation is made—for instance, at the end of a trial—but where the precise nature and level of reparation is left to stakeholders (perhaps ²⁵ Though we should note that some proponents of restorative justice, such as Nils Christie, reject the idea that the dealings of a restorative justice meeting should not be limited by the prior definition of its participants into the roles of ‘victim’ and ‘offender’. For a contrast, see eg H Zehr, The Little Book of Restorative Justice (Intercourse, PA, 2002) 9.

Apology and Reparation in a Multicultural State

287

within reasonable, though wide, limits), may be an attractive second best. But it is clear—if one shares the motivations of the communicative theory—that it is only a second best. In this paper I have attempted to raise a problem for the communicative theory of punishment, particularly for the variety defended by RA Duff. However, it is this variety of the theory that I take to be otherwise the most promising; and I hope that there is a way of solving the problem that will leave the theory more or less intact. There is clearly much more to say on this issue than I have put forward here. But the basic problem is that the communicative theory sets itself the task of issuing symbolically adequate condemnation of public wrongs on behalf of the political community as a whole. What I have tried to point out is a tension, in multicultural conditions, between on the one hand making the symbols adequate; and on the other hand acting on behalf of all the communities represented rather than just the majority or most politically vocal one. In order to resolve this tension the state would need to be able to call on symbols for condemnation that could be taken to belong to the political community as a whole. A condition on the success of the communicative theory, therefore, is that in multicultural societies we are able to reason together about which emotions are appropriate in the event of wrongdoing, and hence which symbols would properly express the necessary condemnation.

16 Contracts, Promises, and the Demands of Moral Agency Emmanuel Voyiakis*

If we set out to justify our contract law to a moral agent, what kind of story should we be aiming to provide? Or, to sharpen the question, if, as moral agents, we cast a critical eye on our contract law, what should we take as its proper moral measure? For instance, should we test contract law for its effectiveness in protecting the moral rights of promisees and enforcing the moral duties of promisors? Should we test contract law for its ability to maximize the efficient use of society’s resources? Should we test it for its ability to bring about just distributions? Or should we test it for its success in making us more autonomous? Picking a test from this far-from-exhaustive list is important because each test will tend to show different parts of contract law as morally sound or morally wanting. Those who find it proper to test contract law for its ability to safeguard the moral right of promisees to performance will take issue with the fact that contract law rarely requires specific performance and grants compensation instead. Those who think that contract law should enhance individual autonomy will disagree; for them, contract law should only intervene to protect certain autonomy-enhancing moral practices—such as the practice of promising—against erosion in the hands of those who violate its rules. Those who believe that contract law should maximize the efficient use of society’s resources are also comfortable with the primacy of compensation. Their critical gaze is more likely to focus on rules that tie parties to their bargain (eg punitive damages) even when performance turns out to be overall more costly than breach. Those who see contract law as an instrument of distributive justice will be troubled by the fact that contract rules on duress and unconscionability provide only minimal protection against the use of unequal bargaining power, and so on. Since these different views claim the same justificatory space, moral agents intent on scrutinizing contract law face a problem of perspective at the start of their inquiry. My aim in this paper is to discuss some aspects of how general * I am grateful to Peter Jaffey, George Letsas, and participants to the UCL Law and Philosophy Colloquium for fruitful discussions of parts of the paper.

Contracts, Promises, and the Demands of Moral Agency. Emmanuel Voyiakis. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Contracts, Promises, and the Demands of Moral Agency

289

moral theory, and our intuitions about the demands of moral agency and responsibility in particular, could help moral agents respond to this problem. A straightforward way of bringing moral theory to bear on the matter would be to pick a perspective for justifying contract law right at the level of global moral theory. To give a rather crude example, a thorough-going consequentialist would waste no time in proclaiming that the proper moral test for contract law is whether it produces overall better states of affairs. If the consequentialist principle has global moral reach, as its advocates certainly think it does, it should without more ado apply to the case of contract law as well. By the same token, staunch libertarians—and some disillusioned anarchists—would quickly embrace the idea that contract law makes moral sense if and only if it provides adequate protection for the interpersonal (or ‘pre-institutional’) moral rights of parties to voluntary transactions and so on. More generally, those drawn to this ‘top-down’ way of looking at things will tend to think that the choice between different perspectives on the justification of contract law should depend on the outcome of more general debates about the respective merits of global moral theories. They will thus tend to move quickly from questions about the justification of contract law to general questions about the nature of all moral justification, revisiting the former only when they have answered the latter to their satisfaction. My discussion will assume without much argument that moral theory does not come into the story in quite that way. Global moral theories need to earn their influence on the justification of contract law—as on any moral question—before the tribunal of our first intuitions about what contract law is and what its rules are for. More to the point, there is no guarantee that any given moral theory will pass this test unscathed. Just as our intuitions about promising tend to alert us to some more pervasive weaknesses in moral consequentialism, it may well turn out that our intuitions about contract law provide fresh support for some global moral theories and point to important flaws in others. I propose to concentrate instead on the more nuanced idea, recently developed by Seana Shiffrin, that we need to measure contract law not so much against general moral theories but against some general truths about moral agency and responsibility. Whatever else it does, this line of thought goes, our contract law should at least be consistent with our considered intuitions about how moral agents ought to conduct themselves and when they may be held morally responsible for their actions. One extension of that point is that the rules of our contract law should not make it difficult for moral agents to behave as they ought to, or that contract law should not contradict our basic ideas about the conditions of moral responsibility. This approach does not take the correctness of any general moral theory for granted, but it retains a certain ‘top-down’ character in the sense that it begins from certain general moral truths and then works its way towards the proper justification of contract law.¹ ¹ Jules Coleman calls this way of proceeding ‘ideal’ theorizing, which he contrasts with ‘practicebased’ and ‘middle-level’ theorizing, see J Coleman, Risks and Wrongs (1992) 8–9.

290

Emmanuel Voyiakis

My paper falls into three parts. Part I describes a couple of general parameters of the task of justifying contract law by appealing to conceptions of moral agency. Part II discusses Seana Shiffrin’s recent account of that question and part III concludes by drawing out some implications of the points in the preceding parts for contract theory. The paper puts forward two claims. My first claim is that we have good reason to be morally sceptical towards accounts of contract law that show the bulk of what we uncontroversially take to be its rules and doctrines to be unjustified. The more an account shows contract law and the demands of moral agency to diverge, the more reason we have to suspect that the account measures contract law against inappropriate moral standards. My second and broader claim is that, for all its surface appeal, the quest to accommodate contract law within a general conception of moral agency and responsibility underdetermines a host of important disagreements about which aspects of our agency contract law speaks most pertinently to.

I. Consider the basic idea that contract law, indeed all law, ought not to make it difficult for us to be decent moral persons.² The idea sounds attractive in that it seems to set a plausible and workable moral test for our contract law to meet. Moreover, it seems to follow quite naturally from the fact that law makes demands on us, demands that can sometimes stand at odds with our moral duties towards each other. At the same time the attempt to align contract law with our intuitions about moral agency and responsibility carries two considerable risks. I draw attention to them as threshold issues in order to give more shape and bite to my understanding of the challenges facing moral agents intent on discovering whether their contract law has moral merit. The first risk is that by projecting a certain conception of moral agency and responsibility onto contract law, we may fail to explore the possibility that contract law displays properties that require us to extend or modify our working understanding of those concepts. One important way in which this possibility could materialize would be for us to find that contract law speaks to a distinctly political aspect of our agency, as opposed to the aspect or aspects of our agency that are engaged in interpersonal moral affairs.³ We could hold, for example, that contract ² Seana Shiff rin develops her account of the justification of contract law from a similar starting point, n 8 below. Shiff rin provides a more general account of the accommodation of law with moral agency and its demands in S Shiff rin, ‘Egalitarianism, Choice-Sensitivity and Accommodation’ in RJ Wallace et al (eds), Reason and Value: Themes from the Moral Philosophy of Joseph Raz (2004) 270 and 295ff. See also S Smith, Contract Theory (2004) 106. ³ It seems to me that John Rawls held a view similar in spirit. Considering the value of autonomy in an agent’s moral life, he writes: ‘it is not the parties [to the original position] but citizens of

Contracts, Promises, and the Demands of Moral Agency

291

law is not there to tell us how we ought to behave towards other people, but to tell us whether our political institutions are justified in using their enforcement powers in the context of a voluntary undertaking. It seems to me that, were we to endorse it, this conception would affect not just what we take to be the proper moral test for contract law, but also our view of the structure and demands of our moral agency, insofar as it would imply that being a moral agent involves being subject to a plurality of mutually irreducible moral concerns.⁴ That conception may, of course, turn out to be false. The point is that even if we were unmoved by its pluralist approach to moral agency and responsibility, we could not begin our inquiry into the moral merits of contract law from a starting point that rules it out of contention. For that reason, I believe that an attractive approach to the problem of finding a justification for our contract law needs to have a more open-ended character, in the sense that it leaves our conceptions of both contract law and moral agency/responsibility liable to revision in the light of each other. In particular, an attractive justification for contract law should allow that our contract rules might display moral properties that fit best with a pluralist conception of agency and responsibility. What follows from this commitment, I think, is that a plausible justification of contract law should be one that we find morally attractive for the sort of thing that we normally take contract law to be. This idea entails that the search for an appealing moral justification of contract law is subject to the following weak constraint of fit. If a moral principle shows the bulk of what we regard as settled contract rules to amount to a comprehensive moral failure, then we have some reason to suspect that this principle may be missing some important moral properties of our contract law and therefore that it might set an inappropriate moral test for its rules.⁵ This is a genuine constraint because it limits the range of plausible justifications to those that have an adequate affinity with what we uncontroversially take the subject-matter of contract law to be. But the constraint is weak for two reasons. First, our suspicions against principles that show contract law to be in especially bad moral shape might prove unwarranted. It might turn out that contract law emerges as largely unjustified even under the moral principles that fit it best. Second, the constraint cannot exclude anything but the most implausible views as contenders for the justification of contract law without begging the question. We may be able to dismiss out of hand the view that contract law should

a well-ordered society in their public life who are fully autonomous . . . Here I stress that full autonomy is achieved by citizens: it is a political and not an ethical value’, J Rawls, Political Liberalism (1994) 77. ⁴ Samuel Scheffler has brought to the fore the moral pluralism that underlies the Rawlsian conception of a distinctly political liberalism in S Scheffler, ‘Egalitarian Liberalism as Moral Pluralism’ (2005) 79 Proceedings of the Aristotelian Society—Supplement 238–48. ⁵ The nature of the practice restricts the kinds of principle that can be applied to justify or criticize it, would be a Rawlsian way of putting it.

292

Emmanuel Voyiakis

be measured against the moral standards of friendship, but serious argument is required to disqualify any more substantial view. My second threshold point arises from the following concern. The task of bringing contract law in line with our conception of what we ought to do as moral agents can be taken to imply that the accommodation will be successful when contract law and morality converge in their substantive demands on us. However, it seems to me that this way of thinking fails to make space for a plausible idea, namely that our contract law can be justified in the eyes of a moral agent in virtue of the procedural fact that it reflects the past political decisions of our community. According to this view of legality, the rules of contract law are morally justified not so much because they have this or that content, but because—or to the extent that—they are the product of a morally sound procedure of political decision-making.⁶ Now, this view can hardly be the whole story: even if our laws bind us morally because they flow from the past political decisions of our community, this does not preclude us from asking whether these are the laws that we ought to have. To put the point differently, it seems quite reasonable to allow that contract laws in our communities are morally binding to the extent that our laws generally are, but to query whether they are the laws that we would endorse as moral agents if we were designing our contract law from scratch. And yet this manoeuvre is too quick in one respect. If our contract law can draw justification from the moral merits of the decision-making process that has laid it down, we can no longer treat the task of justifying contract law to a moral agent simply as the task of finding out what contract rules would appeal to moral agents with a free hand in designing it. We should rather allow that, under certain conditions, it might be possible to justify our contract law to fellow moral agents who could reasonably reject the political decisions that have shaped it. For example, although moral agents might not endorse the requirement of consideration if they were designing contract law from scratch, one could justify this requirement towards them on the ground that it derives from their community’s past political decisions about the enforcement of voluntary undertakings. No doubt, there will be cases where the community’s decisions carry no moral weight, eg when the community’s practices are deeply unjust or corrupt. But in any less dramatic context the expectation should be that the moral merits of a community’s contract law are not exhausted in the merits of its substantive contract rules, as seen from the perspective of moral agents designing the law from ⁶ I think that this formulation is flexible enough to accommodate both instrumental and noninstrumental conceptions of the value of legality. For example, it is consistent with Joseph Raz’s view that the past political decisions of our community have moral authority over us only when they make it more likely for us to comply with the reasons that apply to us independently, see J Raz, The Morality of Freedom (1988) chs 3–4 at 38ff. As far as our laws meet this condition in the main and overall, they have legitimate moral authority, despite the fact that a certain branch of the law may diverge from what morality requires.

Contracts, Promises, and the Demands of Moral Agency

293

scratch. The point can be extended in terms of moral agency. A plausible conception of ourselves as moral agents must be taken to include not just our role as persons interacting with one another but also our role as citizens in a political community, the decisions of which make a moral difference for everyone and have genuine justificatory power.⁷

II. ‘Contract law tells people what to do and how to conduct themselves in the context of voluntary undertakings. It allows them to break their contracts, but it orders them to pay compensation or damages on breach. It requires the other party to mitigate their damages. It requires people to refrain from making fraudulent or negligent misrepresentations and so on.’ Suppose that we are drawn to this way of looking at contract law. When we come to assess the moral merits of particular contract rules, it is natural to think along the following lines. Contract law may require people to do certain things and to conduct themselves in a certain way in their voluntary undertakings, but morality has its own rules about the same matters. They are the moral norms about promising, giving one’s word, leading people to form expectations about one’s future actions, etc (for the sake of simplicity I will henceforth refer to the morality of promising only). If people who engage in voluntary undertakings are subject to both sets of rules, it is important to ask whether the demands that contract law makes on people are compatible with the demands of morality. And the greater the divergence between contract law and the morality of promising, the more pressing the question will become whether a moral agent eager to comply with contract law can still do the morally right thing. Seana Shiffrin has recently deployed and developed this line of thought. She describes her general standpoint as follows: [I start] from the premise that law must be made compatible with the conditions for moral agency to flourish—both because of the intrinsic importance of moral agency to the person and because a just political and legal culture depends upon a social culture in which moral agency thrives. The content and normative justifications for a legal practice —at least one that is pervasive and involves simultaneous participation in a moral relationship or practice—, should be capable of being known and accepted by a self-conscious moral agent.⁸

⁷ Thomas Nagel has this point in mind, I think, when he says that the force of political obligations derives from ‘a special involvement of agency or the will that is inseparable from membership in a political society’, T Nagel, ‘The Problem of Global Justice’ (2005) 32 Philosophy & Public Aff airs 113, 128. ⁸ S Shiff rin, ‘The Divergence of Contract and Promise’ (2007) 120 Harvard Law Review 709, 712.

294

Emmanuel Voyiakis

The idea that contract law should be justified in the eyes of decent moral agents serves two purposes in Shiffrin’s account. First, it explains why we should compare the rules of contract law against the moral rules of promising, rather than another set of moral principles, eg principles of distributive justice. The key claim here is that the legal practice of contracting and the moral practice of promising are sufficiently analogous in the normative sense, so that an individual who makes or receives a voluntary undertaking participates in both practices at the same time. Second, and more importantly, this normative premise raises the stakes of the comparison between contract law and the morality of promising in the following way. Normally, we would think that if contract law fails the test of morality, the only thing that follows is that moral agents have less of a reason to conform to it, or, less dramatically, that any actions by our institutions pursuant to this body of law lack a proper justification. In other words, we would think that the only issue at stake is the moral legitimacy of contract law in the eyes of agents subject to it. Shiffrin claims that the stakes are higher. She argues that when contract law diverges too much from the morality of promising, this undermines not just the law’s legitimacy but also the capacity of individual agents to lead a morally decent life, in the sense that it contributes to a legal and social culture that a morally decent person would find hard to live in. Asking how close the requirements of contract law are to those of morality is thus important in discovering not just whether the moral agent should confirm to the law, but also whether the moral agent can navigate both legal and moral requirements when they diverge substantively and thus whether the law’s structure and justification are compatible with the conditions that sustain moral agency. Against this background, Shiffrin sets out the moral challenge that contract rules need to meet. She says: [C]ontract law would run parallel to morality if contract law rendered the same assessments of permissibility and non-permissibility as rendered by the moral perspective, except that it replaced legal permissibility for moral permissibility and used its distinctive tools and techniques to express and reflect those judgements . . . [I]f contract law ran parallel to morality, then contract law would—as the norms of promises do—require that promisors keep their promises as opposed to paying off their promisees. The only difference is that it would require this as a legal, and not merely a moral, matter.⁹

Then, having surveyed several aspects of the content and structure of AngloAmerican contract law, Shiffrin offers a stark, if unsurprising, diagnosis: The moral rules of promise typically require that one keep a unilateral promise, even if nothing is received in exchange. By contrast, contract law only regards as enforceable promises that are exchanged for something or on which the promisee has reasonably relied to her detriment. When breach occurs, the legal doctrine of mitigation, unlike morality, places the burden on the promisee to make positive efforts to find alternative ⁹ S Shiff rin n 8 above, 722.

Contracts, Promises, and the Demands of Moral Agency

295

providers instead of presumptively locating that burden fully on the breaching promisor. Morality classifies intentional promissory breach as a wrong that, in addition to requiring compensation, may merit punitive reactions, albeit sometimes minor ones; these may include proportionate expressions of reprobation, distrust and self-inflicted reproofs, such as guilt. Contract law’s stance on the wrongfulness of promissory breach is equivocal at best, manifested most clearly by its general prohibition of punitive damages.¹⁰

On that basis she concludes, alarmingly, that some aspects of U.S. contract law not only fail to support the morally decent person, but contribute to a legal and social culture that is difficult for the morally decent person to accept. Indeed, U.S. contract law may sometimes make it harder for the morally decent person to behave decently. (emphasis added)¹¹

To sum up: Shiffrin’s account of the moral merits of contract law is based on the following two claims. First, that contracting and promising are sufficiently analogous in the normative sense, so that agents who make and receive voluntary undertakings participate in both practices at the same time. Second, that the greater the divergence between the rules of contracting and those of promising, the harder it becomes for individuals to do the morally right thing. I will leave aside the second of those claims for one moment, because its plausibility seems to depend on whether it is actually the case that contract law and promising are sufficiently analogous in the normative sense. In particular, the second claim envisages a moral ‘dumping-down’ scenario: it suggests that wide divergences between law and morality can present a serious threat to one’s moral integrity. Shiffrin’s particular worry is that, if contracting and promising require people to behave in divergent ways, say the practice of contracting allows breach of an agreement on payment of damages whereas the practice of promising forbids it, people might gradually begin to doubt whether promises create even a moral right to performance. Now, there are various ways of trying to dispel this worry, one being that if Shiffrin’s ‘dumping-down’ scenario did indeed pose a serious threat to our moral integrity, that threat would have materialized by now, given that our contract law has almost never in its history enforced specific performance or imposed punitive sanctions on contract-breakers. But for the ‘dumpingdown’ scenario to get off the ground at all, it must be the case that contracting and promising are indeed analogous practices in the required sense. Unless it is true ¹⁰ S Shiff rin n 8 above, 709–10. I am sceptical about one aspect of Shiff rin’s claim here. It seems to me that whether one who fails in one’s promissory duties should be subject to some form of reproach by others is a moral question about blame and blameworthiness, the conditions of which include a lot more besides a breach of one’s promissory duties. On this point, see TM Scanlon, What We Owe To Each Other (1998) 267; ibid, ‘Blame’, paper presented in the 2006 NYU Colloquium on Legal, Political and Social Philosophy, available at . In fact, unless certain other conditions are met, the sole fact that you have breached your promissory duties towards me does not license any kind of punitive action against you, either by me or by others. ¹¹ S Shiff rin n 8 above, 710.

296

Emmanuel Voyiakis

that people who engage in voluntary undertakings participate in both practices at the same time, the notion that moral agents may be torn between the demands of contract law and the morality of promising in ways detrimental to their integrity does not get any traction. One way of casting doubt on Shiffrin’s claim about the ‘parallelism’ between contracts and promises would be to say that promising and contracting are not analogous but very different practices. One might agree, for example, with Dori Kimel that the equation of contracting to promising misunderstands the value and function of both practices.¹² Kimel claims that whereas promising develops its full normative significance in the context of personal relationships and attachments, contracting is designed for the very different purpose of allowing the parties to achieve a collaborative result without having to form a personal attachment.¹³ Or one might agree with utilitarian theorists that the purpose of contract law is to facilitate the efficient exchange of goods, under some attractive conception of efficiency.¹⁴ Those attracted to either of those arguments and their several variations will be tempted to think that the morality of promising cannot be the arbiter of the law of contract, and that their divergence is not morally or ethically problematic, because each practice has a different point and serves a different value. Kimel and the utilitarians might be right to say that contracts are not promises, but the temptation to use their arguments against Shiffrin’s first claim should be resisted. Shiffrin does not claim that contracting and promising have the same point and serve the same value. She is rather concerned to find out whether moral agents could endorse whatever point the practice of contracting might have and still lead a morally decent life. To put the point differently, even if we allow that contracts are not promises, the question remains whether agents who take their moral duties towards each other seriously could reasonably accept the putative values that underlie the practice of contracting. Efficient exchange, or the value of ‘detachment’, or the value of averting tangible harm to others might actually be the values that best account for the legal practice of contracting. Shiffrin’s concern is external to that issue because it raises the very different question of whether agents intent on doing the morally right thing would be comfortable with a practice that encourages individuals to take a utility-maximizing or ‘detached’ or merely harm-avoiding attitude towards their contracting partners. The point can be generalized. Anyone who thinks that contract law requires or permits people to act towards each other in ways that promote certain interests will, at some point, have to address Shiffrin’s questions about the compatibility of pursuing of those interests, on the one hand, and leading a morally decent life, on the other. Saying that contracts are not promises does not sidestep questions about the relationship between the practices of contracting and promising; it sets them up. ¹² D Kimel, From Promise to Contract: Towards a Liberal Theory of Contract (2003). ¹³ Kimel calls this the value of ‘detachment’, ibid at 78ff. ¹⁴ See M Trebilcock, The Limits of Freedom of Contract (1993) 7–10; C Goetz, R Scott, ‘Enforcing Promises: an Examination of the Basis of Contract’ (1980) 89 Yale Law Journal 89.

Contracts, Promises, and the Demands of Moral Agency

297

To put the question to Shiffrin’s claim that contracting and promising are parallel practices, we have to look to another kind of account of contract law and its justification. I want to introduce it as a genuine alternative to the accounts of contract law that Shiffrin has in mind, in order to support a conditional assertion: if the alternative account is right, Shiffrin’s claim of parallelism between contract and promise comes out as mistaken and her criticisms of contract law fail to get enough traction. That alternative view could be quickly summarized as follows. In sharp contrast to the moral rules of promising, contract rules do not make any demands or impose any requirements on how people should behave towards each other. Their purpose, rather, is to specify the circumstances in which the community’s institutions are justified in using their enforcement powers in the context of voluntary undertakings between individuals. So whereas the morality of promising engages the aspect of our moral agency that concerns our interaction with other persons, contract law engages the aspect of our moral agency that concerns our attitude towards institutional action. A simple—if slightly inaccurate—way of putting the matter would be to say: while the morality of promising concerns the actions of individuals, contract law concerns the actions of institutions. On its face, this story about the point and purpose of contract rules does not run against the idea that contract law may be intimately related to the morality of promising. It may be that—to the libertarian’s delight—the exercise of enforcement powers by political institutions in the context of a voluntary undertaking is legitimate only to the extent that it is used to vindicate the pre-institutional moral rights that people have towards each other. At the same time, the alternative story reveals an important sense in which seeing contract law as concerning the actions of institutions rather than individuals does indeed change the character of the debates about the justification of contract law. The point is that that new story requires us to account for the moral merits of contract law by staking and defending general theses about the proper purpose and limits of institutional action in a political community. For instance, we now find ourselves having to address the claim that the use of enforcement in the context of a voluntary undertaking can be legitimate even when no moral rights of the promisee have been violated; or the stronger claim that the standards of political legitimacy applicable to institutional action have very little to do with the interpersonal moral rights between promisor and promisee. Such claims may or may not necessarily resonate with us, but at least they suggest that if contract law concerns the actions of institutions, it falls to be justified or criticized within a more general theory of the moral principles applicable to institutional action, ie within a general theory of political morality. This is also the sense in which the alternative story is incompatible with Shiffrin’s claim that when people make and receive voluntary undertakings they are participating in both a moral and a legal practice. In the story’s terms, the legal practice of contracting is first and foremost an institutional practice; it is not so much that people make contracts, but that contracts are the voluntary

298

Emmanuel Voyiakis

undertakings that institutions can legitimately enforce. Similarly, although people can undertake promissory commitments, it would be wrong in the terms of this alternative story to say that they make a promise legally enforceable. Whether the commitment can be legitimately enforced is not a matter of what people have willed or what moral commitments they have undertaken towards each other, but of what our institutions are morally required to do according to the principles that regulate their conduct. So even if the view that contract law concerns the actions of institutions rather than to individuals is compatible with a strong connection between contracts and promises, it is not obviously compatible with the claim that contracting and promising are parallel practices in the strong sense that Shiffrin has in mind. I think that this alternative view finds enough support in our rough intuitions about the nature of contract rules to merit serious consideration. Contract law may not require people to keep to their contracts, but neither does it allow people to breach their contracts. To put it bluntly, contract law says neither ‘you are not permitted to breach your contracts’ nor ‘you are permitted to breach your contracts’. In fact, rather than require people to behave one way or another, contract law only specifies certain enforcement measures that institutions are entitled to take in case people decide to breach their contractual obligations. Similarly, no rule of contract law says ‘do not make misrepresentations’ or ‘do not exercise undue influence on the other party’ or ‘you ought to mitigate your damage’ and so on. The familiar contract rules on misrepresentation, undue influence and mitigation simply specify in which cases the community’s institutions will use their enforcement powers and to what effect. Of course, the point should not be overstated because other intuitions clearly fit better with the view that contract law speaks to individuals. We talk of ‘contractual obligations’, implying that people have obligations to behave towards each other in certain ways and that these obligations flow from their contract. We talk, for example, about the seller’s duties to disclose some information or the buyer’s right to request it. These phrases, deeply embedded in contract discourse, encourage a conception of a contract as essentially a relationship between individuals, a relationship to which institutions play an external facilitating role. Although first-order intuitions do not clinch the matter, however, they do suggest that it is very much an open issue how one should arbitrate the stand-off between the view (shared by Shiffrin and the theorists she targets) that contract law concerns the actions of individuals, and therefore that contracting is indeed a practice parallel to promising, and the opposite view that contract law concerns the actions of institutions. In the rest of this part, I want to suggest that at least we have good reason to hope that Shiffrin’s story about the parallelism between contract and promise is mistaken and, to that extent, that we have good reason to pursue the alternative story as far as it can go. That reason is that if our contract law is best seen as speaking to individuals, it will most certainly come out as a comprehensive, rather than just a partial or even extensive, moral failure. And as I noted

Contracts, Promises, and the Demands of Moral Agency

299

in part I, this is a conclusion that we cannot accept unless we have looked at and rejected any account that promises to show contract law in a better moral light. I will argue this point by bringing into discussion Shiffrin’s second claim about the problematic effects of a wide divergence between contract law and the morality of promising. I will contend that, judging from the character of the solutions that she sees to this problem, Shiffrin has underestimated the moral difficulties that the truth of parallelism would give rise to. Suppose that Shiffrin is right that contracting and promising are parallel practices in the sense that those engaging in voluntary undertakings are participating in both practices at the same time. Also suppose that our contract law is interpreted or modified so that it runs parallel to the morality of promising in every way required to meet Shiffrin’s concerns. Suppose, for example, that our modified contract law now requires people to perform their contracts and specifies punitive sanctions for failure to do so; that it orders specific performance at first instance; that it relaxes the rules on mitigation and so on. Then take the morally decent person of this scenario to have all the first order intuitions that all of us (save for act-utilitarians) have about keeping promises, namely that a promise commits one to perform; that breach of a promise is a wrong; that the prospect of a greater benefit is generally not a good reason for not performing and so on.¹⁵ Now, would a person of this moral profile find it easier to live in a community where contract law mirrors the requirements of the morality of promising? Or, to pose the question in Shiffrin’s terms, would our modified contract law make it easier for that ‘morally decent person to remain morally decent’? Shiffrin suggests so, but I think that this is a mistake. What makes the person of this scenario morally decent is not the fact that he performs the actions that his promise has committed him to but the fact that he does so on the basis of certain reasons, namely reasons provided by the morality of promising.¹⁶ A person that acted on the basis of reasons provided by the law of contract would hardly fit the same profile, just as a person who refrained from killing others because the law attaches severe penalties to homicide would not fit the profile of a morally decent person.¹⁷ This is so because, even with the modifications that would satisfy Shiffrin, the law would still require people to behave in a certain way on pain ¹⁵ I am not sure what that person would think about mitigation, but—contrary to Shiffrin—I am inclined to the view that morality does require victims of a wrongful breach of a promise to try to lessen the extent of their damage, if they can do so without incurring substantial transaction costs. ¹⁶ A trickier question is whether keeping a promise for fear of legal enforcement constitutes an immoral act. John Gardner argues that acts that merely conform to moral reasons, instead of being motivated by them, may lack moral worth but are still instances of morally right action, J Gardner, ‘Prohibiting Immoralities’ available at . Gardner’s argument cannot help Shiffrin’s thesis at this point. His concern is whether a person who keeps a promise because the law attaches adverse consequences to breach may be said not just to be of short of moral virtue but also to be committing a moral wrong. In any event, Shiffrin seems to disagree with Gardner on this point, see S Shiffrin, n 8 above, 718. ¹⁷ I take this example from Joseph Raz, see J Raz, ‘The Obligation to Obey: Revision and Tradition’ (1994) Ethics in the Public Domain 341.

300

Emmanuel Voyiakis

of enforcement measures being taken against them. But just as moral agents could not endorse a practice that required them to break a promise in order to maximize utility, they would not endorse a practice that required them to keep a promise in order to avoid the adverse consequences that the law attaches to breach. They would think that an agent who kept to a contract for that reason would not just be acting on ‘one moral thought too many’ but on a morally wrong thought.¹⁸ By extension, they would think that a practice that required people to act on such reasons would not be supportive of decent moral agency. The upshot is that, if the claim of parallelism between contracting and promising is correct, contract law would fail in the eyes of morally decent persons not just because a few of its rules diverge from the morality of promising, but because the threat of enforcement is a typical feature of contract rules and, therefore, a typical feature of the reasons that law provides people with. To that extent, decent moral agents who agreed with the claim of parallelism would have grounds to regard our contract law as a comprehensively rather than just partially defective moral compass. Note that it is not open to Shiffrin to say—along with philosophers such as Joseph Raz and Tim Scanlon—that the threat of enforcement would be acceptable to a decent moral agent as an efficient way of dealing with indecent or morally weak people who may be easily tempted to break their promises.¹⁹ Shiffrin’s distinctive claim is not that our present contract law lets bad people get away, but that it is liable to corrupt the moral integrity of decent people, who would normally not break their promissory commitments. The problem for Shiffrin is exactly that it is unclear how, even with the modifications that she suggests, our contract law could fail to corrupt the decent folk. For even if our contract law required everyone to keep their contracts, it would do so by threatening coercive action, and it is not clear how persons who took themselves to be responding to such a threat could still be regarded as morally decent agents. Now, this last phrase might suggest that I have overloaded the profile of a decent moral agent. Although it may be true that, generally speaking, morally decent folk do not need to consult the law of contract in order to find out how they ought to behave towards others, there is no denying that even such people do sometimes drop their moral guard and commit moral mistakes. These people would therefore have some reason to endorse a system that authorized community institutions ¹⁸ Pace Gardner, I am inclined to think that the rule ‘keep your promise because that is the morally right thing to do’ and the rule ‘keep your promise or face certain legal consequences’ do not even prescribe the same action, my intuition being that, at least in the moral domain, actions are individuated by the kinds of reason that motivate the agent. But the more modest claim of the text will do for my present purposes. ¹⁹ See J Raz, ‘Promises in Morality and Law’ (1981–2) 95 Harvard Law Review 916 (Raz thinks that the aim of contract law should be to protect the practice of promising against erosion and debasement); TM Scanlon, ‘Promises and Contracts’ in P Benson (ed), The Theory of Contract Law (2001) 86, reprinted in TM Scanlon, The Difficulty of Tolerance (2003) (Scanlon argues that contract rules that failed to provide protection against damage caused by breach of promissory obligations would be reasonably rejected).

Contracts, Promises, and the Demands of Moral Agency

301

to force them to do what morality requires, on the ground that such enforcement (or its threat) could help nudge them back onto the morally right track. Responsiveness to coercive threats may not be a typical characteristic of the morally decent person, but it can function as an effective reminder that one’s actions have been morally out of line and as a means for better moral self-awareness. Sound as this argument might be, it fails, in my view, to back the high stakes of Shiffrin’s second claim about the significance for people’s moral integrity of contract law running parallel to the morality of promising. Even if we grant that the aim of our contract law is to help occasional offenders back onto the morally right track, it is still unclear why we should pursue that aim by enforcing the content of the offenders’ moral obligation, eg by enforcing performance. Would milder measures, eg an order of compensation, not be enough to alert those persons to their moral mistake and to incite them to morally correct action? They might or they might not, some will say; a lot would depend on the moral psychology of the persons involved and the surrounding temptations. I agree. My point is that the further one requires contract law to go in order to nudge people back onto the morally right track, the less sense it makes to describe those people as morally decent persons in the first place. Similar thoughts should frustrate another attempt to show contract law in a more favourable light to agents who are keen to remain morally decent in their dealings with each other. This attempt trades on the idea that sometimes what one morally ought to do will be unclear and consulting the law may well be a good source of information about how the moral question should be resolved, since the law is undoubtedly a rich repository of moral thought.²⁰ The very last point is surely correct. The issue, however, is not whether we can derive moral insights from the law, but rather which moral questions these insights bear on. This takes us back to the original stand off. Does contract law give us moral insights into how we ought to behave towards each other, or does it give us moral insights into the rather different question of when state enforcement in the context of voluntary undertakings is justified? Unless we resolve this issue, we cannot know how to interpret contract law as a source of moral reasons or how to make use of it as a repository of moral insight, so the objection does not get any traction. I think that the above points can be generalized. If we take our contract law to provide moral agents with a response to the question ‘what ought I to do in my effort to remain a decent moral agent?’, we will almost certainly find that it fails to live up to even the least demanding of our moral standards, not because it requires different behaviour of moral agents, but because its requirements typically include the threat of coercion, a responsiveness to which finds little place in our conception of morally decent character. The fact that, should Shiffrin’s claim of parallelism turn out to be correct, our contract law will appear as a comprehensive failure in the eyes of morally decent ²⁰ The argument has been made by Patrick Atiyah, P Atiyah, Promises, Morals and the Law (1982).

302

Emmanuel Voyiakis

persons does not entail that Shiffrin’s account is mistaken. Contract law may just be in much worse moral shape than Shiffrin estimates. It does, however, provide us with enough incentive to pursue the alternative possibility that contract law creates requirements for our political institutions rather than for individuals. If our contract law indeed comes out as largely justified under the moral principles that apply to actions by our political institutions, we will have good reason to regard those principles as giving a better account of the moral nature of contract law and, on that basis, to reject Shiffrin’s suggestion that contracting and promising are parallel practices. Then, we will be able to say that contract law is not there to tell us how people ought to behave towards each other, but when the community’s institutions are justified in using their enforcement powers in the context of voluntary undertakings between individuals. However, the proof of that justificatory pudding will be in the eating.

III. The paper began by asking whether moral agents intent on passing judgement on contract law could expect much help in their inquiry from general theories of agency and responsibility. The main point I have been pressing is that, for all its appeal, the task of accommodating contract law with moral agency does not by itself provide us with a stable standpoint from which to survey the moral merits of our law. The reason is that quite a lot turns on what we understand the point of contract law to be and, more importantly, which aspects of moral agency we take it to speak most pertinently to. The sheer appeal to moral agency and accommodation underdetermines these questions. Those who believe that contract rules tell us how we ought to behave towards others see those rules as speaking to the aspect of our moral agency that is concerned with our interpersonal moral duties. Those who believe that contract law concerns the legitimacy of the use of state enforcement in the context of voluntary undertakings disagree. They think that contract law speaks to a very different aspect of our moral agency, namely our proper attitude towards our community’s political institutions. In the former view, contract law and moral agency are successfully accommodated when they issue compatible directives about how people ought to behave in their dealings with one another. In the latter view, the successful accommodation of contract law and moral agency depends on whether contract rules can be adequately justified by reference to the principles of political morality that apply to the conduct of our institutions and are acceptable to reasonable moral agents. My point has been that the two views stand apart not simply on their assessment of the moral merits of contract law but primarily on their conception of the structure of moral agency and its demands. I may not have offered a full argument as to which view of moral agency we should prefer, but I have given reasons why we should hope that the first,

Contracts, Promises, and the Demands of Moral Agency

303

which Shiffrin advocates, turns out to be an inappropriate yardstick for measuring the moral merits of our contract law. If I am right in this, those who believe that contract law speaks to institutions have every reason to be optimistic about their ability to offer an account that shows contract rules to be a coherent and justified whole.

17 Number and Government Claire Grant¹

I. Good government needs number. That intuition might be attributed to Jeremy Bentham, for it is a leit motif of his legal and political thought. The importance of number to the writings of Bentham is incontrovertible. His axioms are invariably expressed as quantitative propositions. This is evident in the most famous of them all, the principle of utility, which he formulated as the ‘greatest happiness of the greatest number’, but is conspicuous throughout his writings. Moreover, Bentham does attend to data from empirical enquiry, where he can. We may instance his plea for the foundation of a statistical-gathering society to make compilations of facts, for factual information about actual states of affairs would support rational and informed decision-making. Bentham was also an early advocate of official criminal statistics which, he enthused, would furnish ‘a measure of excellent use in furnishing data for the legislator to go to work upon.’² Now Bentham did not have access to a comprehensive set of statistics by which his analysis might be informed, for his time just preceded the ‘avalanche of numbers’ of which Hacking (1990) has written, but he did see the promise of statistics. There are hence few actual numbers cited in Bentham, but we do see what we might call a kind of balance-sheet thinking throughout. Bentham’s thinking in numbers pursued a trichotomous advantage. First, the general idea of a metric was commended by his commitment to reason, witness his admonition that ‘the law-giver should be no more impassioned than the geometrician. They are both solving problems by sober calculation’ (J Bentham, Deontology, or the Science of Morality (Longman, Rees, Orme, Brown, Green and Longman, 1834) 19). Number promised a way of preventing the arbitrary sentiments of ¹ I am grateful to Jo Wolff for stimulating discussions about crime and disadvantage, to Philip Schofield for helpful suggestions on an earlier draft, to Christopher Bennett for useful comments on the paper at Cumberland Lodge, and to Melissa Lane and discussants at King’s College, Cambridge. ² J Bentham, The Works of Jeremy Bentham (ed, J Bowring) (Simpkin, Marshall & Co, 1843) IV 29.

Number and Government. Claire Grant. © Oxford University Press 2007. Published 2007 by Oxford University Press.

Number and Government

305

antipathy and sympathy, or irrational self-interest, from interceding as a ground for action. But reason need not be numeric of course. It might proceed by definition or taxonomy. More fundamentally, Bentham espoused an enumerative method because it offered a systematic reasoning process by which the most rational course of action might be elucidated, given his views on the task and the end of government. Bentham thought that rationality was predominantly based on a calculus of self-interest. ‘All men calculate’, he asserted (J Bentham, An Introduction to the Principles of Morals & Legislation (Clarendon Press, 1789/1996) 174). It was the job of government to act upon that calculus so as to bring self-interested action in line with utility. He thought that people routinely deliberate practically under an informal numeric logic, valuing possible actions through a tally of their projected hedonic consequences. In fact, Bentham surmised, ‘we perform many arithmetical calculations almost without knowing it’ (J Bentham, The Theory of Legislation (trans CK Ogden) (Kegan Paul, Trench, Trubner & Co, 1931) 32). Of course practical reasoning can be erroneous, and Bentham certainly thought of people as normally fallible. While rationality requires the maximization of utility, individuals might not always be able to discern the optimal course of action, or indeed resolve to act on it. Given that the end of government was maximal utility, a method of working out what institutional arrangements would best further this maximization was required, of indicating the institutions and practices that would most efficiently guide practical reasoners toward acting on correct utility estimates. This method would need to elucidate the value of pains and pleasures as a function of their quantity, and indicate the resulting measure of happiness likely to result from each of the available courses of action. The third reason for Bentham’s thinking in numbers is that he was, as will be shown, a committed egalitarian, and egalitarians need number. That assertion is quite contentious. Hart (HLA Hart, Essays on Bentham: Studies in Jurisprudence & Political Theory (Clarendon Press, 1982) 24), for example, remarked on the ‘grand scale’ of Bentham’s cost-benefit method, and held that this conducted to grossly inegalitarian results. This paper aims at an exposition of Bentham’s argument, and a qualified defence of it against objections about number and about egalitarian distribution. Bentham held that the principal aim of government was to guard against pains, and especially those inflicted by people upon other people.³ In this paper, attention will be directed to a particular calculus of pains, namely that elaborated in Bentham’s copious writings on crime and its reduction. Bentham wrote extensively on the crime problem, and integrally to his most important treatises. We might see his thinking in numbers about crime as a testing ground for the merits of practical calculus, as well as for examination of his egalitarian credentials. ³ This negative utilitarianism about government was for at least two reasons, first a regard for the anti-paternalist tenet that the care of her pleasures should be reserved to the individual, and second the recognition that pains and pleasures are asymmetric. Pains spread and magnify more rapidly and sensibly than pleasures, so under the principle of utility the reduction of pain is a priority.

306

Claire Grant

Before commencing upon this task it behoves to deal with those who would give number short shrift. A critic might dismissively say that any kind of numeric evaluation of hedons and dolors is inane, and even vulgar. This objector might press the point, saying that human misery cannot be grasped in its depth and subtlety by an enumerative, aggregative logic. All this measuring and summing amounts to no more than a rationalistic reduction of human sentiment and sensation. Human feeling cannot, or should not, be subjected to an analysis in mathematical terms. There are some things that you just cannot do by numbers, for example, making romantic choices. Likewise there is no unit of suffering, and so no way of envisaging distress through numbers. This distaste is observed in Margery Fry’s (M Fry, ‘Bentham and English Penal Reform’ in GW Keeton and G Schwarzenberger (eds), Jeremy Bentham and the Law (Stevens & Sons, 1948) 55) disapproving remark that Bentham occupied ‘a purely Euclidean universe, with axioms unqualified and unquestioned, filled with people who could be classified, and whose actions could, it was supposed, be understood and predicted in accordance with a few elementary rules.’ Bentham studied only man in the abstract, she said, and not men as they actually are. This species of objection indicts Bentham’s approach for its formulistic character, which even admits of learning by rote. The critic would no doubt be singularly aggravated to learn that Bentham set aside his game of pushpin in order to pen the following mnemonic doggerel on the seven elements of the felicific calculus: Intense, long, certain, speedy, fruitful, pure— Such marks in pleasures and in pains endure. Such pleasures seek if private be thy end: If it be public, wide let them extend Such pains avoid, whichever be thy view: If pains must come, let them extend to few (J Bentham, An Introduction to the Principles of Morals & Legislation (Clarendon Press, 1789/1996) 38).

If we subscribe to utilitarian thinking, our objector contends, we end up like Gradgrind, stuck on the numbers game, mechanically following the calculus, imaginatively enervated and so unable to empathize with the plight of the suffering or to do anything about it. A similarly adverse response comes with Marx’s (K Marx, Capital: A Critique of Political Economy (Lawrence & Wishart, 1867/1954) 571) infamous shopkeeper jibe, that Bentham deduces the nature of man from the principle of utility, and ends up ‘stupidly’ mistaking the English shopkeeper for the normal man. One response to these objections simply replies that Benthamic government must entail a mathematical method because it deals with quantities, but that would be question-begging. Rather, it is necessary to show why numbers should count. In this respect, an appeal to novelty will not do, for example, a plea that we conventionally represent distance, weight and temperature by numeric scales, and are just not accustomed to thinking of pain

Number and Government

307

and pleasure in that way. This may be true but still fails to show why good government needs number. Of course we may concede that there are genuine difficulties in representing values like pain and pleasure numerically. This concession has led some critics to concur with Bentham as to the desirability of a felicific calculus but to consider the project, or at least the ambitious claim to precision, hopelessly utopian. Instead of opposing mathematizing in principle, an interlocutor might argue that it suffers practical difficulties. But the recognition of difficulties does not discredit the method. In fine, the general objections to Benthamic calculus boil down to two, namely that his account of motivation is perverse, and that he advocates measurement of what cannot be measured. To dispel these objections, we need to show why we need numbers, and what they can do, and accordingly to examine Bentham’s argument on law, number and government.

II. Bentham held that whenever people deliberate over their self-interest, which is often, they are engaging in a sort of numerical thinking. Under those two ‘sovereign masters’ of pleasure and pain as we all are, that govern us ‘in all we do, in all we say, in all we think’, our thoughts are focused on augmenting our experience of the former and diminishing that of the latter (J Bentham, An Introduction to the Principles of Morals & Legislation (Clarendon Press, 1789/1996) 11). People, in short, are expected utility maximizers. When pleasure and pain are at stake, ‘who is there that does not calculate? Men calculate, some with less exactness, indeed, some with more: but all men calculate’ (J Bentham, An Introduction to the Principles of Morals & Legislation (Clarendon Press, 1789/1996) 173–4). According to Bentham there is (numeric) method even in madness. Because people routinely calculate, their calculations may be guided. It is the task of the legislator to so arrange the stakes that the numbers stack up on the side of utility, so that people are guided to choose the course of action that best accords with utility. Accordingly, in Theory of Legislation Bentham presents an argument about utility, government and law. Bentham subscribes to a value monism according to which the only proper object, and intrinsic good, is the general happiness (that is, the happiness of the greatest number of the members of a given political community). Value ultimately depends upon utility. Now according to Bentham, security is foremost among the subsidiary objects conducive to the general happiness. This is the case because the security of individuals is a general prerequisite for the attainment of happiness. Security consists, primarily, in security for person, property, reputation and condition in life. It can only be produced through law, and is a precondition for the possibility of other important goods, namely abundance, subsistence and equality. Bentham perceives security to be constantly threatened, requiring the legislator to undertake continual vigilance and defensive action,

308

Claire Grant

indeed the entire object of law is the care of security. Law permits individuals to ‘submit to the yoke of foresight’, working in return for labour’s rewards because law promises to them the products of their labour, guaranteeing to preserve to the worker the rewards of her industry. Enlarging upon this point, Bentham says that in order to comprehend the full scope of the principle of security we must realize that, unlike other animals, a human’s capacity to experience pleasure or suffering is not limited to the present; rather we are susceptible of experiencing pleasure and pain by anticipation. Accordingly to give her security it is insufficient to protect a person against an actual loss; she must also be guaranteed, where possible, against future losses. Moreover, the idea of her security must be prolonged in her mind, as far as her imagination can extend. This human disposition to anticipate, to look forward, Bentham calls ‘expectation’. The expectation of the future permits us to plan ahead, and gives to the discrete moments and events in a life a sense of forming a continuous whole. The principle of security comprehends the maintenance of all these hopes, and directs that events should be conformed to the expectations which the laws have generated. An injury to the sentiment of security produces a distinct evil, which Bentham names the pain of disappointed expectation. Insecurity, then, causes misery and ill-being, and undermines prosperity, for fear stifles hope. For Bentham, a utilitarian system of law is an important means of procuring to citizens a relatively stable, predictable and ordered set of relations within which to go about their affairs. Now a utilitarian conception of maximizing the general happiness does not conduct automatically to an argument for government through law. Consider Godwin’s (W Godwin, An Enquiry Concerning Political Justice vol II (University of Toronto Press, 1793/1946)) argument. Godwin believed that government is deleterious to liberty and happiness because it fosters dependence and ignorance, and he subsequently rejects co-operative and rule-governed practices like law. He does, however, concede that a minimal sort of government is necessary, for the limited end of criminal justice, that is, ‘the prevention of the hostile attack of one member of the society, upon the person or property of another’ (Godwin above, 322). Godwin recognizes the claim that punishment, as incapacitation, is necessary as a temporary expedient, in the interim preceding the advent of that utopia in which there will be no need, and hence no justification, for law. Godwin disdained the very idea of laws as indispensable reasons for action. Bentham, however, did not share Godwin’s faith in the proposal that radical changes in social and political arrangements should be undertaken permitting people to live harmoniously, within a system of communal property-ownership, and by the exercise of private judgment alone, unaided by governmental intervention. Bentham envisaged a large area wherein people would act without legal intervention, but insisted that law was needed to restrain harm to others. The intervention of the legislator, according to him, was necessary in order to approach more closely the greatest happiness of the greatest number. Bentham’s theory of government owes to his theory of action, which entails a complex analysis of motivation, will and understanding. A thorough rendition

Number and Government

309

of it cannot be attempted here. However, a few indications are required, because Bentham’s view of government through law was premised on the likelihood that people will act according to their own perceived self-interest, seeking to experience pleasure and avoid pain. Harrison (R Harrison, Bentham (Routledge & Kegan Paul, 1983)) clarifies Bentham’s argument about law and self-preference in the following terms: the work of the legislator consists in constructing the fabric of felicity. The legislator’s task concerns the co-ordination of self-interest with the public interest, which must be brought about artificially, by means of superadded pains in particular. With the end of promoting happiness, the legislator acts upon peoples’ interests, so as to bring those interests in line with their duty artificially. This is particularly done by creating threats of punishment. The design of institutions works on the basis that people likely will act in their own interests, but that even if they do, they will still in fact act so as to bring about the desired consequences. All this was unappealing to Godwin, who denounced the idea that coercion could persuade rationally. Bentham (J Bentham, Of Laws in General (ed HLA Hart) (Athlone Press, 1970)) however sought to reconcile the reasongiving (‘directive’) and coercive (‘sanctional’ or ‘incitative’) elements of law. The reason-giving quality of criminal law rests on a principle of non-maleficence, that is, refraining from harming others.⁴ Bentham contended that if an exposition of reasons for each law was published alongside the law codes, people might come through rational persuasion to be law-abiding, from a concurrence of the will, rather than being obedient through blind fear. To elucidate the kinds of actions that the law should seek to guide, and the precise legal means to be employed, Bentham turned to the felicific calculus. Bentham’s (J Bentham, An Introduction to the Principles of Morals & Legislation (Clarendon Press, 1789/1996) 38–41; 1931 J Bentham, The Theory of Legislation (trans CK Ogden) (Kegan Paul, Trench, Trubner & Co, 1931) 31–2) calculus is a device for estimating the value of pleasure and pain as a function of their quantity. This provides that the measure of any particular pleasure or pain will vary under certain circumstances. First we have the four circumstances of its intensity, duration, certainty, and propinquity (proximity). The value of a pain or pleasure depends on its intensity and duration. For instance, an intense pain is worse than a less intense one, and a pain of a given intensity is worse in accordance with the longer it lasts. Concerning certainty, when attempting to calculate the value of a given action in advance of actually having performed it, an element of uncertainty as to what the actual consequences will be applies. This is a general feature of choosing. Some outcomes, however, are more probable than others, and so we might build in a ratio of chances for and against the pain or pleasure’s coming ⁴ See ‘Promulgation of the Laws and of the Reasons Thereof ’ (J Bentham, The Works of Jeremy Bentham (ed, J Bowring) (Simpkin, Marshall & Co, 1843) I, 162), where Bentham explains that the justificatory reasons in penal law should be drawn from the nature of the evil of offences, and the remedies of which they are susceptible. In ‘Specimen of a Penal Code’ (Bentham above, I, 164–8) he provides the example of a bodily harm offence, for which he supplies ‘a commentary of reasons upon this law.’

310

Claire Grant

about. Furthermore, pains and pleasures may have other pains and pleasures as their consequences, so we must take into account two additional circumstances, namely those of fecundity (productiveness) and purity. A fecund pain is one likely to be followed by other pains. A pure pain is one not likely to produce pleasures. Finally, if the calculation is to be made concerning a collection of individuals, we must also consider the extent of persons affected. Now for the balance sheet: to estimate the general tendency of any act, sum up all the values of all the pleasures on one side, and all of the pains on the other. If the resulting balance bears on the side of pleasure, the act tends to the good, if on the side of pain, it tends to the bad. We should further note that in Bentham the felicific calculus relies on two assumptions, viz, that the painfulness or pleasurableness of experience is quantifiable, and that the quantities so obtained can be added across people. Now we come to crimes, and their disvalue according to the felicific calculus. In his account of the nature of crime, we see Bentham working through this algorithm.

III. The felicific calculus permits Bentham to formulate crime as a distinctive category of harmful acts, because of the marked surfeit of pain over pleasure that they tend to produce. Bentham’s (J Bentham, Of Laws in General (ed HLA Hart) (Athlone Press, 1970) 209–10) argument to that effect is as follows: the distinguishing marks of crimes are thought to be threefold. First, crimes are thought to be distinguished by the magnitude of the mischief that they occasion. Second, crimes are thought to be marked out by their odiousness, that is, by the amount of disapprobation attached by the community to them. Finally, crimes are thought to be distinguished by the fact that a punishment is attached to them by the law. Bentham comments that only the first of these marks of crime, which concerns the gravity of the mischief occasioned, is reliable. It is also the only acceptable basis for the other two; that is, the gravity of the offence should be the only cause for its being odious and hence punishable. Odiousness alone, or as the primary criterion, will not do; rather, there must be harm. Only acts that harm, or that pose a genuine danger of harm, should be prohibited. Moreover, these should only be acts that harm others. Bentham calls offences against oneself ‘reflective offences’, and says that these ought to be beyond the sphere of the legislator. The imposition of restraints upon the individual for her own welfare may be the business of the educator, or the parent, but not of the legislator. Furthermore, criminal harms are mostly positive, rather than negative. Positive duties to one’s neighbour there may be, but these are a matter for private ethics and not for the criminal law. Criminal harms are things that are done, by and large, and not things that are allowed to happen.⁵ ⁵ A special case, concerning which Bentham does support the criminalization of a negative offence, is omissions to rescue those in grave mortal danger (J Bentham, An Introduction to the

Number and Government

311

This brings us on to whom crime harms. If crimes are harms to others, who is it that crime harms? Bentham discerns different kinds of ‘evil’ in a criminal act, or rather two kinds of ill consequences of a crime, namely the primary and the secondary (J Bentham, An Introduction to the Principles of Morals & Legislation (Clarendon Press, 1789/1996) 143–7; J Bentham, The Theory of Legislation (trans CK Ogden) (Kegan Paul, Trench, Trubner & Co, 1931) 49–53). The distinction may be made apparent when we consider a case of robbery. The robbery causes pain to the person whose money is taken. That is the primary evil. Moreover, the robbery alarms other owners of money, and it suggests to would-be robbers the facility of robbery, and so may weaken their disinclination to rob. These latter two are secondary evils. The harmfulness of crime thus comprehends actual pain, and the expectation or objective risk of pain. The primary, or first-order evil, is that sustained by an assignable individual or assemblage of individuals. It includes the ‘primitive evil’, that sustained by the direct victim, as well as the ‘derivative evil’ sustained by those associated with her, whether by interest or sympathy. The secondary evil originates from the primary, but may extend over the entire community. This secondary evil consists in two bad things, namely alarm and danger. When the report of a crime circulates alarm, which may mount from disquiet to terror, is occasioned by one’s apprehension of suffering the same evil seen to have befallen the victim; each person sees that she could be made to suffer in the same way, and is alarmed. Danger is the chance of pain. In the case of crime, the danger is the actual probability of other similar crimes occurring. Bentham holds that the commission of an unpunished crime reduces deterrence. A bad act, by its deleterious example, paves the way for other bad acts and so makes it more probable that similar crimes will occur. Knowing this to be the case is part of the alarm. The alarm occasioned by attacks on person and property is emphasized by Bentham as a distinctive evil, a significantly adverse consequence of crime. It can far outstrip the actual losses and injuries sustained by victims because the alarm is communicable. For instance, take an attack on property. This occasions four kinds of consequential evil, namely the evil of non-possession, the pain of losing, the fear of losing, and the deadening of industry. Concerning the latter three: the knowledge of attacks upon property induces a fear of the loss of one’s possessions. An anxiety about one’s possessions is created, and at a certain point of insecurity the fear of loss diminishes the enjoyment of one’s possessions. Moreover, the concern to preserve them condemns the owner to ‘a thousand sad and painful precautions, always liable to fail’ (J Bentham, The Theory of Legislation (trans CK Ogden) (Kegan Paul, Trench, Trubner & Co, 1931) 116). Notably then, Bentham does not say that the more individuals implement security measures, the safer and so more content they will be. In fact he suggests that private expenditures against Principles of Morals & Legislation (Clarendon Press, 1789/1996) 293). See also J Bentham, The Works of Jeremy Bentham (ed, J Bowring) (Simpkin, Marshall & Co, 1843) I, 164 on abstaining from helping one in danger.

312

Claire Grant

crime are burdensome and ineffectual. This is readily imagined. Perceived threats acquire a quality of urgency, and work to demand of agents a special degree of vigilance, over and above ordinary regard for property and person. Much of this effort will be wasted, or even counterproductive. Other bad consequences of crime include a third-order evil. This is a certain kind of demoralization, a declivitous process in which people who hear about crimes come to despair of enjoying the fruits of their labour and so are discouraged from industry. The debilitating effects of fear and demoralization extend well beyond the actual individual who has been victimized. Indeed they may spread across the entire body of the community, exercising a persistent and deleterious effect far and wide. We may surmise by dint of Bentham’s felicific algorithm that crimes ought to be prevented, but by what means? Again this is a matter for the felicific calculus, which may guide in the choice of legislation and penal regimes.

IV. In Bentham the ideal response to crime is one whereby the outlay hazarded is convertible to profit, primarily in prevented offences, and secondarily in making up for the costs imposed on victims and on society at large by crimes. He sees that all means of crime reduction are evils, because they all involve pain, and defeats the prima facie presumption against the infliction of pain by indicating ways in which the dose of pain is necessary, and ideally beneficial. Bentham always keeps in mind the notion of an economical punishment, viz, one that produces the desired crime-reductive effect with the investment of the least possible cost of suffering. Punishments are too expensive when they produce more good than evil, or when it is possible to obtain the same good by means of a lesser punishment. According to Bentham’s penal economics, we must calculate the quantity of the evil in each case, comparing the evil of the offence and the evil of the law. Notably, an estimation of the profit of punishments has reference to three discrete parties, namely the public, the party injured and the offender. Bentham examines the merits of different kinds of crime reduction, in broadly felicific-calculatory terms. His argument is for less expenditure on private precautions and more on publicly-funded and co-ordinated measures. The direct means of crime prevention is criminal legislation, which prohibits certain conduct that it defines as criminal offences. Bentham emphasizes that the pain suffered by the victim of the crime outweighs any pleasure that it might occasion the criminal. Moreover when we add in the pain of alarm suffered by the general public, and the steps taken by private individuals to protect themselves, the pleasure to the criminal is yet further outweighed. The upshot is the overwhelming net pain caused by actual and potential offences. On the basis of an operation something like: P=H–G the net pain (P), which is the difference between the summed harm (H) and the gain (G) to the offender, appears clearly, and so the action in question may be prohibited. Criminal laws are hence to be rules of conduct that maximize

Number and Government

313

utility; if the actions are proscribed and people do refrain from doing them, on balance we will all be better off. The indirect means of crime prevention considered by Bentham include justifiable ways of arresting crimes. Bentham (J Bentham, The Works of Jeremy Bentham (ed, J Bowring) (Simpkin, Marshall & Co, 1843) I, 367–8) insisted that the resources and powers of the police must be fortified in order to deal with an offence as soon as it ‘announces itself’ or is in the process of being committed. Bentham’s indirect means also include ways of operating on people physically and morally to dispose them to be law-abiding. These latter include the prophylactic methods of education, religion, the fostering of a culture of benevolence, and the provision of poor schools.⁶ Elsewhere he advocates poor relief as prophylactic. In the Essays on the Poor Laws (J Bentham, Writings on the Poor Laws vol I (Clarendon Press, 2001) 10) he distinguishes indigence, the lack of subsistence due to an inability to labour sufficiently to have the necessaries to sustain life, from poverty, which is always relative to subsistence. Indigence must be relieved: on grounds of public security, relief must be afforded. This is the case because for the indigent punishment is less likely than certain death. In circumstances of indigence, where the stark choice is one of ‘steal or starve’, criminal conduct is not the less eligible option. Of course no dissuasion may be efficacious if the expectation of sanction is inferior to the utility of crime.⁷ Here the ordinary directivesanctional force of law cannot hope to succeed, so the relief of indigence must be co-ordinated, and indeed arranged through law. Bentham (J Bentham, Rationale of Punishment (Robert Heward, 1830) 19–27) quite straightforwardly relates punishment to security, by rendering deterrence the primary aim of punishment. Because it may effect general deterrence, punishment is justifiable as the lesser evil. Now, following his own utilitarian penal economics, Bentham must ensure that it really is a lesser evil. Punishment must therefore have the property of frugality, that is, it should not be gratuitous: no superfluous pain must be caused the offender. Moreover the ideal of frugality entails no surplus of pain inflicted on the offender, and that the pain produces a good for another. The first rule in the measure of punishments, in Bentham’s view, provides that the pain of the punishment must outweigh the profit of the crime. It must also be the minimum required to achieve deterrence. In this respect, Bentham differed markedly from Paley, who also emphasized the secondary evil occasioned by crime but employed this in an argument for deterrence by highly severe penalties.⁸ Moreover the end of punishment, that is security, must ⁶ Bentham professed a lifelong aversion for religion. Consequently, the appearance of religion in this list is probably attributable to an unwarranted recension of Dumont. ⁷ The potential of prophylaxis was taken further by Chadwick, who advanced various proposals for preventive police, the aim of which was ‘to get at the removable antecedents of crime’ (E Chadwick, ‘Preventive Police’ (1829) 1 London Review 28). ⁸ Paley (W Paley, The Principles of Moral and Political Philosophy (R Faulder, 1785) 537), discussing the comparative malignancy of crimes, says: ‘Injuries effected by terror and violence, are those which it is the first and chief concern of legal government to repress; because their extent is unlimited; because no private precaution can

314

Claire Grant

not be better met by less painful alternatives. He (J Bentham, An Introduction to the Principles of Morals & Legislation (Clarendon Press, 1789/1996) 158–64) considers these conditions when talking about cases in which punishment is unjustified. Punishment is unjustified, he states, when it is ‘unprofitable’, which is whenever the expected evil generated by the punishment outweighs the expected good to be produced. Punishment is unjustified also when it is ‘needless’, which is the case if its aims can be achieved by other non-punitive means, for example, where there is a certainty of adequate compensation. In addition to general deterrence, Bentham recognized two other ends of punishment, namely incapacitation and rehabilitation. In incapacitation, there is crime reduction because the inmate has had the ability to offend removed. Bentham (J Bentham, Rationale of Punishment (Robert Heward, 1830) 21) proposed incapacitation for crimes that inspire great alarm and manifest a very dangerous disposition on the part of the offender, making it necessary to take the power of reoffending from her. He (above, 327–51) disdained, however, the practice of transportation to the Australian colonies, as only ostensibly incapacitative. It is easy to understand, he said, the approving t