Power, luck and freedom: Collected essays 9781526104557

Table of contents :
Front Matter
Contents
List of figures
Series editor’s foreword
Acknowledgements
Introduction: power, luck and freedom
Part I Power
Why should we care about the definition of power?
Agency and structure: interpreting power relationships
Rational choice and community power structures
Power, capability and ableness: the fallacy of the vehicle fallacy
Part II Luck
Resources, power and systematic luck: a response to Barry
Shaping future luck
Luck, equality and responsibility
Luck and leadership
Part III Freedom
Choice: its increase and its value
The value of choice in public policy
Republican freedom, rights and the coalition problem
The construction of rights
Social choice and the grammar of rights and freedoms
References
Index

Citation preview

w

o

D

e

it

es

d

K

te

ec ll o C

www.manchesteruniversitypress.co.uk

sa ys h

d

in

g

lu P fr c o w k ee a e d n r, o d m

Keith Dowding is Professor of Political Science at the Australian National University, Canberra

Power, luck and freedom

Featuring a substantial introduction written specially for the volume, Power, luck and freedom serves as an excellent overview of the work of a key thinker on power. It will support undergraduate and graduate work in political science, political sociology and political theory.

C o l l e c t e d e s s ay s

The power of agents, notably leaders, is partially based on our perception of their power. In strategic settings luck plays a role in outcomes, but these outcomes then feed into the ability and reputation of actors, enhancing or damaging their power. Perceptions feed into reality which then feeds back into perceptions. Dowding shows how luck is related to responsibility – reducing some types of luck in outcomes will, paradoxically, also reduce responsibility. He integrates this account into our understanding of the value of choice and freedom, demonstrating that collective action provides problems for republican accounts of freedom from domination. Arguing that choice is valued instrumentally, he then shows that the ‘liberal paradox’ of Amartya Sen, while demonstrating the need to balance welfare and rights, does not provide the fundamental evaluational problems Sen claims.

DOWD I N G

This book presents thirteen essays by Keith Dowding, one of the world’s foremost authorities on political and social power. Ranging across a number of related topics – including luck, choice, freedom and rights – Dowding criticises static explanatory models of power, emphasising the need for a more dynamic approach. The alternative he proposes is a ‘rational choice’ resourcist model, which takes into account the crucial factors of reputation and luck.

Social and Political Power

Power, luck and freedom

SOCIAL AND POLITICAL POWER Power is one of the most fundamental concepts in social science. Yet, despite the undisputed centrality of power to social and political life, few have agreed on exactly what it is or how it manifests itself. Social and Political Power is a book series which provides a forum for this absolutely central, and much debated, social phenomenon. The series is theoretical, in both a social scientific and normative sense, yet also empirical in its orientation. Theoretically it is oriented towards the Anglo-American tradition, including Dahl and Lukes, as well as to the Continental perspectives, influenced either by Foucault and Bourdieu, or by Arendt and the Frankfurt School. Empirically, the series provides an intellectual forum for power research from the disciplines of sociology, political science and the other social sciences, and also for policy-oriented analysis.

Power, luck and freedom Collected essays

Keith Dowding

Manchester University Press

Copyright © Keith Dowding 2017 The right of Keith Dowding to be identified as the author of this work has been asserted by him in accordance with the Copyright, Designs and Patents Act 1988. Published by Manchester University Press Altrincham Street, Manchester M1 7JA www.manchesteruniversitypress.co.uk British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data applied for

ISBN  978 1 5261 0728 2  hardback ISBN  978 1 5261 0456 4   paperback First published 2017 The publisher has no responsibility for the persistence or accuracy of URLs for any external or third-party internet websites referred to in this book, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate. Typeset by Servis Filmsetting Ltd, Stockport, Cheshire

Contents

List of figures page vi Series editor’s foreword vii Acknowledgementsxi Introduction: power, luck and freedom

1

Part I  Power

27

       

29 47 68 82

1  2  3  4 

Why should we care about the definition of power? Agency and structure: interpreting power relationships Rational choice and community power structures Power, capability and ableness: the fallacy of the vehicle fallacy

Part II  Luck

103

       

105 121 134 155

5  6  7  8 

Resources, power and systematic luck: a response to Barry Shaping future luck Luck, equality and responsibility Luck and leadership

Part III  Freedom

171

  9  10  11  12  13 

173 188 205 230 256

Choice: its increase and its value The value of choice in public policy Republican freedom, rights and the coalition problem The construction of rights Social choice and the grammar of rights and freedoms

References275 Index296

Figures

  3.1 Simple Prisoners’ Dilemma page 69   3.2 Luck and power 72   3.3 Relationship between power and ability74   6.1 Three voters, three options127   6.2 Public and private preference (simplified from Kuran (1995: 36))132   8.1 Spatial representation of leader reputation164   8.2 Authority of leader167 10.1 Cardinality and diversity191 13.1 Sen’s Liberal Paradox as a toy game258 13.2 Shirt game: rights as strategies260 13.3 Conformist non-conformist shirt game260 13.4 Hohfeld’s categories263

Series editor’s foreword

Series editor’s foreword

Bertrand Russell once argued that power is to social science what energy is to physics (Russell 1937: 10). While power is one of the most important concepts in the social sciences, it is also one of the most complex and elusive to research. Weber’s analysis of power and authority (1947 and 1978) is one of the first social scientific discussions of power and it influenced the US power debates, which developed post-Second World War. In these debates Dahl’s careful analysis stands out for its clarity in providing us with a conceptual vocabulary of power (Dahl 1957 and 1968). This includes an agencybased, exercise and decision-making definition of power; conceptualized in terms of powerful actors (A) making subordinates (B) do something that they would not otherwise do. This exercise of power is distinct from resources (that may or may not be exercised) and it provides power-holders with power of specific scope. However, while providing a new set of conceptual tools to analyse about power relations, Dahl’s work was subject to sustained critique from Bachrach and Baratz and others, who argued that power is also exercised through structural biases that are not necessarily reducible to overt decision-making (Bachrach and Baratz 1962). Lukes followed this critique with his theorization of the third-dimension of power (Lukes 1974), which concerns the mobilization of belief and ideology to legitimize power relations of domination. The three-dimensional model was applied in a richly textured empirical study of Appalachian mining communities (Gaventa 1982). Overall, as the three-dimensional power debates develop, the focus shifts from actions of the dominating actor A to the counter-intuitive and fascinating phenomenon that subordinate actors B often appear to actively acquiesce or participate in their own domination. In a qualified critique of Lukes, Scott argued that appearances are often deceptive (Scott 1990). The relationship between public and private discourse renders the working of three-dimensional power more complex

viii

Series editor’s foreword

than any simplistic images of the oppressed willingly participating in their own domination, or internalizing false-consciousness. In turn, Scott’s work has inspired an on-going power-literature on the complexities of resistance versus acquiescence. In the 1980s, under the influence of the translation of Foucault’s work (for instance, Foucault 1979 and 1982), the Anglophone power debates shifted towards more epistemic and ontological analysis, which resonated with the shift of emphasis from the powerful to the conditions of the oppressed. This gave rise to fascinating work on the relation between power and discourse; power and truth; the way power influences the ontological formation of social subjects through discipline; and how governmentality has changed systemic power relations (including Clegg 1989, Dean 2010, Laclau 2005, Flyvbjerg 1997 and Hayward 2000). However, in critique, many have argued that neo-Foucauldians tend to lose sight of the significance of individual agency (Lukes 2005). Bridging the intellectual divide between those following the Dahl-Lukes trajectory and the neo-Foucauldians, another important thread to the power debates comes from Giddens (1984) and Bourdieu’s (1989) conceptualizations of structure as a verb. This way of thinking provides us with conceptual tools for making sense of how agents both structure and are structured by relations of power. In international relations, the shift from agency toward systemic, epistemic and ontological perceptions of power took the form of a gradual move from realist focus on resources to a more idealist emphasis upon soft power (Nye 1992 and 2012). Similarly, in rational choice theory, there emerged an emphasis upon the systemic situatedness of strategic choices. This gave rise to Keith Dowding’s critique of Brian Barry, ‘Resources, Power and Systemic Luck’ which appears in this collection as Chapter 5 (pp. 105–20). The effect of interrogating the social contexts and social ontology of agents caused many theorists to re-evaluate the nature of power normatively, moving away from the automatic equation between power and domination to a perception of power as a condition of possibility for agency, thus freedom (Morriss 2002 and 2009). Thus, freedom and power move from being opposing categories to being mutually constitutive. Associated with this normative re-evaluation, power theorists distinguish between power-to, power-with and power-over (Allen 1998 and 1999; Pansardi 2012). To begin with power-over was considered a normative negative, suggesting oppression, while power-to and power-with were the positives. However, recently some theorists argue power-over can also have emancipatory, as well as the more obvious dominating, aspects (Haugaard 2012). This book series seeks to build upon these rich traditions of power



Series editor’s forewordix

analysis, which currently make the study of social and political power one of the most vibrant fields in the social and political sciences. It also builds upon the success of the Journal of Political Power, which provides an important forum for article analysis, while this book series facilitates longer works on social and political power. The book series is open to any of the multiplicity of traditions of power analysis, and it welcomes research that is theoretically oriented, as well as empirical research on power or practitioner-oriented applications.

Mark Haugaard National University of Ireland, Galway, Ireland References

Allen, A. (1998) ‘Rethinking power’, Hyptia, vol. 13. Allen, A. (1999) The Power of Feminist Theory: Domination, Resistance, Solidarity, (Boulder, Colorado: Westview Press). Bachrach, P. and Baratz Morton S. (1962) ‘The two faces of power’, American Political Science Review, vol. 56.4: 947–52. Bourdieu, P. (1989) ‘Social space and symbolic power’, Sociological Theory (7)1: 14-25. Clegg, S. (1989) Frameworks of Power (London: Sage). Dahl, R. A. (1957) ‘The concept of power’, Behavioural Science, Vol. 2(3): 201–15. Dahl, R. A. (1968) ’Power’ in David L. Shills (1968) International Encyclopedia of the Social Sciences (12): 405–15 (New York, Macmillan). Dean, M. (2010) Governmentality: Power and Rule in Modern Society (Second edition) (London: Sage). Flyvbjerg, B (1998) Rationality and Power: Democracy in Practice (Chicago: The University of Chicago Press). Foucault, M. (1982) ‘The subject and power’ in H. L. Dreyfus and P. Rabinow Michel Foucault: Beyond Structuralism and Hermeneutics (London: Harvester Wheatsheaf, London). Foucault, M. (1979) Discipline and Punish: The Birth of the Prison. (Harmondsworth: Penguin). Giddens, A. (1984) The Constitution of Society (Cambridge: Polity Press). Gaventa, J. (1982) Power and Powerlessness: Quiescence and Rebellion in an Appalachian Valley (Oxford: Clarendon Press). Hayward C. (2000) De-facing Power (Cambridge: Cambridge University Press). Haugaard, M. (2012) ‘Rethinking the four dimensions of power’, Journal of Political Power 5(1): 35–54 Laclau, E. (2005) On Populist Reason (London: Verso). Lukes, S. (1974) Power: A Radical View (London: Macmillan). Lukes, S. (2005) Power: A Radical View, 2nd edn (Basingstoke: Palgrave Macmillan). Morriss, P. (2002) Power: A Philosophical Analysis, 2nd edn (Manchester: Manchester University Press).

x

Series editor’s foreword

Morriss, P. (2009) ‘Power and liberalism’ in (ed.) S. Clegg and M. Haugaard The Sage Handbook of Power (London: Sage). Nye, J. S. (1990) ‘Soft power’, Foreign Policy (80): 153–72. Nye J. S. (2011) ‘Power and foreign policy’, Journal of Political Power, 4(1): 9–24. Pensardi, P. (2012) ‘Power to and power over: two distinct concepts?’, Journal of Political Power, 5(1): 73–89. Russell, B. (1938) Power: A New Social Analysis (London: George Allen & Unwin). Scott, J. C. (1990) Domination and the Arts of Resistance: Hidden Transcripts (New Haven: Yale University Press). Weber, M. (1947) The Theory of Social and Economic Organization, ed. T. Parsons (New York: New York Free Press). Weber Max (1978) Economy and Society Vols 1 and 2, ed. G. Roth and C. Wittich (Berkeley: University of California Press).

Acknowledgements

Acknowledgements

I would like to thank Mark Haugaard for suggesting that I collect these essays for his new series for Manchester University Press and my editor Tony Mason for his advice and for guiding me and it through editorial board processes. I would like to thank Anne Gelling who turned the original versions into the precise text for the book, and also for more than merely copy-editing the introduction. I also thank Will Bosworth for comments on the introduction and general conversation over the issues concerned in this book over the last couple of years. I acknowledge below the publications in which these papers first appeared thanking who originally commented on the papers. I would like to give additional thanks to Patrick Dunleavy, Peter John, Desmond King, Helen Margetts and Martin van Hees, my co-authors on some of the chapters, for permission to reproduce our joint work in a collection of my essays. I also thank my department, the School of Politics and International Relations at the Australian National University, for funding the rekeying through their research grant awards. Chapter 1 was originally published as ‘Why should we care about the definition of power?’, Journal of Political Power 5 (1) (2012): 119–35. Copyright © 2012 Taylor & Francis. Available online: www.tandfonline. com. DOI: 10.1080/2158379X.2012.661917. Chapter 2 was originally published as ‘Agency and structure: interpreting power relationships’, Journal of Power 1 (1) (2008): 21–36. Copyright © 2008 Taylor & Francis. Available online: www.tandfonline.com. DOI: 10.1080/17540290801943380. I should like to thank Claus Offe for our conversations during the time of writing this essay, which affected its form and content in a manner not entirely within knowledge of the author.

xii Acknowledgements

Chapter 3 was originally published as Keith Dowding, Patrick Dunleavy, Desmond King and Helen Margetts, ‘Rational choice and community power structures’, Political Studies 43 (2) (1995): 265–77. The research reported here forms part of the UK Economic and Social Research Council’s Local Governance Research Programme, supported by Grant No. L311253008. I was Hallsworth Fellow at Manchester University during the redrafting of this paper. Chapter 4 was originally published as ‘Power, capability and ableness: the fallacy of the vehicle fallacy’, Contemporary Political Theory 7 (2008): 238–58. Copyright © 2008 Macmillan Publishers Ltd. Reproduced with permission. A version of this paper was first presented at the PSA conference in Leicester 2003. I thank the participants, particularly Patrick Dunleavy and Peter Morriss, for their comments. I also thank an anonymous referee and Matthew Braham and Martin van Hees for many discussions of related issues. Chapter 5 was originally published as ‘Resources, power and systematic luck: a response to Barry’, Politics, Philosophy and Economics 3 (2003): 305–22. Chapter 6 was originally published as ‘Shaping future luck’, Journal of Conflict Processes and Change 4 (1999): 1–11. Chapter 7 was originally published as ‘Luck, equality and responsibility’, Critical Review of International Social and Political Philosophy 13 (2010): 71–92. This paper was presented at the ‘Choice Group’, London School of Economics, November 2005; the British Academy, December 2005; and the Political Ideas conference, St Catherine’s College, Oxford, January 2006. I would like to thank participants for their comments, especially Chris Armstrong, Brian Barry, Richard Bradley, Alex Brown, Ian HampshireMonk, David Owen, Adam Swift, Simon Tormey, Alex Voorhoevre and Andrew Williams (though I admit I have not made as many changes to the text as their comments deserve). I would also like to thank Martin van Hees and Christian List for the many conversations we have had about luck prior to the writing of this paper. An earlier, briefer version of Chapter 8 was given at the ‘Public Leadership and Beyond’ conference, November 2007, at the Australian National University and published in Paul ‘t Hart and John Uhr (eds), Public Leadership: Perspectives and Practices (Canberra: ANU Press, 2008). This version was given at the ‘Why Leaders Can’t Lead’ workshop, June 2012, at the US Studies Centre, University of Sydney. I thank participants at both conferences for comments. I also thank Michael Dalvean, Anne Gelling, Pete Hatemi, Andy Hindmoor, André Kaiser, Paul ‘t Hart and Ken Shepsle for specific comments.



Acknowledgementsxiii

Chapter 9 was originally published as ‘Choice: its increase and its value’, British Journal of Political Science 22 (1992): 301–14. Copyright © 1992 Cambridge University Press. Reprinted with permission. I would like to thank Bob Goodin for his suggestions and insights. Chapter 10 was originally published as Keith Dowding and Peter John, ‘The value of choice in public policy’, Public Administration 87 (2008): 219–33. The research for this paper was supported by ESRC grant no. RES-15325-0056 ‘Public Services: Exit and Voice as a Means of Enhancing Service Delivery’. We thank referees for their comments. Chapter 11 was originally published as ‘Republican freedom, rights and the coalition problem’, Politics, Philosophy and Economics 10 (3) (2011): 301–22. This paper was presented at the University of Queensland and the Australian National University. I would like to thank the audience for their contributions. I would also like to thank Ian Carter, Mhairi Cowden, Anne Gelling, Bob Goodin, Matthew Kramer and Peter Morriss for their comments on the paper and Martin van Hees for discussions on freedom over many years. Chapter 12 was originally published as Keith Dowding and Martin van Hees, ‘The construction of rights’, American Political Science Review 97 (2) (2003): 281–93. Copyright © 2003 Cambridge University Press. Reprinted with permission. We would like to thank Cécile Fabre, Ruth Kinna, Matt Kramer, Anna Pilatova, three anonymous referees and participants at the Analysis of Measurement of Freedom conference in Palermo, Italy, September 2001, the 2002 meeting of the Dutch Political Science Association and the Economic Decisions Conference in Pamplona, Spain, June 2002, for their comments on earlier versions of this paper. Chapter 13 was originally published as ‘Social choice and the grammar of rights and freedoms’, Political Studies 52 (1) (2004): 144–61. The first draft of this paper was written whilst the author was a Visiting Fellow at the Social and Political Theory Group, Research School of the Social Sciences, ANU. I would like to thank the RSSS for offering an ideal community in which to conduct research. It was subsequently presented at the ECPR Joint Sessions of Workshop, Edinburgh 2003. I would like to thank all participants at the ‘New Approaches to Rights, Freedom and Power’ workshop for their comments, particularly Sebastiano Bavetta, Matt Kramer and Serena Olseretti. I also thank Ian Carter, Martin van Hees, Bruno Verbeek and three anonymous referees.

Introduction: power, luck and freedom

Conceptual analysis There are many different accounts of the nature of social or political power, and almost as many of the concept of freedom. I have published two books and many articles on power, and quite a few articles on freedom. By no means all of these articles are published in this collection. This introduction does not provide an overview of each article; the reader can read them for themselves. But I am going to discuss some of the major issues they raise and chronicle the role each essay plays in the more general account of freedom and power I have tried to develop over the years. I begin with why we should care about developing concepts of power and freedom. Both are concepts that many claim are essentially contestable (Connolly 1983; Lukes 2005; Haugaard 2010). I have always taken that to mean that such concepts are necessarily contestable. We might ask wherein lies that necessity. For Gallie (1956), who invented the phrase ‘essentially contestable’, it seems to lie in the fact that any account of such terms needs to look back to the original exemplar; but historically different trajectories have emanated from that original exemplar and we have no way of claiming which is correct. If the original conceptualization was not perfectly coherent and consistent, there might be multiple ways of bringing about coherence and consistency, none of which can claim historical, or logical, superiority over the others. Add in the fact that we might also be critical of the original exemplar for missing aspects of the concept that the original author and/or we would think important, and perhaps also new knowledge of the way the world works, and we will, of necessity, never be able to agree on which extension of the original exemplar truly is that concept, because logically there is more than one way of extending it consistently with that first use. For scientific concepts, however, the original connection to first usage is less important. That is not to say that scientific concepts, which are types,

2

Power, luck and freedom

have no causal connection to their first usage. The meaning of ‘hydrogen’ is tied to the token hydrogen that served as the chemist’s evidence when it was first identified, so words used to denote modern scientific concepts do not bear some referential causal connection to aspects of earlier uses of those same words. But that history is less important. We can trace a history of the use of the term ‘atom’ – but the referential correlation to past use is not important to judgements of the correct meaning and reference of current usage. Where, for example, science goes beyond the original term H2O for water, what matters is the nature of the chemical bonds and what it means for our understanding of water, rather than some historical connection to the first use of the term ‘water’ – which might have included H2O as its essential constituent, but also included many other impurities. As scientists, we do not care about the historical antecedents. We only care to the extent we are historians. So whilst scientific language is connected to historical antecedents, scientists need not care too much about what those antecedents are. Scientific concepts are not essentially contestable, since there are criteria for judging applications, namely the systematic role they play in a predictive scientific theory or model. For the religions, or traditions more generally, there are no such criteria beyond claims made about historical antecedents. It seems to me that social science concepts occupy a linguistic space between concepts in the natural sciences and concepts referring to bodies of belief or ways of behaving within religions and traditions. Some social scientific concepts play the same role as natural scientific ones. Others come close to being the types of proper names that denote traditions such as Christianity or Republicanism. Some terms in the social sciences are technical ones. In game theory ‘veto player’ and ‘agenda setter’ are examples of terms whose reference applies to theoretical agents in systems. They are explanatory in the same way that concepts are in the sciences. The error sometimes made in the social sciences is to try to find a token biological individual in an actual political system who corresponds one-to-one with a veto player in a theoretical model. Scientists applying theoretical models of gas expansion or explosion do not try to identify particular gas molecules that trigger such events. They merely develop predictions based on the characteristics of the system as a whole. Social science has theoretical models based on the structural features of systems where the number of veto players varies, which generate predictions over policy stability. The predictions hold at the system level, not for the individual behaviour of actual token agents at any given time (Ross 2005; 2014). Other terms in the social sciences are much closer to religions or traditions. Obvious examples are ‘isms’ such as ‘Marxism’, that clearly have a historical basis and, in this case, a figure, Karl Marx, for whom historical



Introduction3

reference must be given. Marxism, like Christianity, attracted important later disciples, such as Lenin, Trotsky and Gramsci, or even G. A. Cohen and Jon Elster, all of whom interpret and may explicitly depart from the original texts. Discussion of the ‘true’ path is directly analogous to such debates in religious dispute. However, some self-identified Marxists are completely uninterested in such discussion. They happily take on the ‘ism’ for what they identify within it, rather than because they wish to follow it (this is a line I take for myself, self-identifying with many ‘rival’ isms: see Dowding 2016: ch. 2). Terms like ‘democracy’ also have some original exemplar, by which they might be thought to be essentially contestable. However, they also bear a descriptive technical meaning, causing problems of a different nature: one that does not lead to essential contestability in the same manner. Democracy (originally) means rule by the people. This leads to the question ‘who are the people?’. We can answer in different ways, but then we can see, given those responses, different versions of democracy. Then we can apply what Chalmers (2011) calls the subscript strategy. Each version can be demarcated by a subscript and there need be no confusion. Only sectarians need believe that a single subscripted democracy is the ‘true democracy’, just as only sectarians need believe that only one Christian sect is true Christianity. I adopt the subscript strategy in Chapter 5 with regard to the concept of power, as I have previously done (Dowding 1991: ch. 8) with regard to the concept of interests, in order to avoid some claims of essential contestability of these concepts. Democracy is an interesting case in another sense. For once we have decided who the people are, we need to decide by what mechanisms they can rule. Representative democracy is one general form; but then we ask what is being represented. If we want to answer the will of the people, we find via Arrow’s theorem (Arrow 1951/63) that the meaning of the will of the people has no definite referent. Rather, the meaning is tied to the decision mechanism itself, so different decision mechanisms will (at best) give different referents to what that will consists of. And, further, any decision mechanism is manipulable (Gibbard 1973; Satterthwaite 1975), so we might have to judge representativeness by the probability of and the moral nature of that manipulation (Riker 1982; Dowding 2006a; Dowding and van Hees 2008a). Thus again, whilst we can have clear definitions of democracy, none will track directly to the original referent, and there need be no single definition that is obviously superior to all others. (Again, that is not to say that some might not be obviously superior to some others.) These ramifications do not, however, as I see it, mean that ‘democracy’ is essentially contestable. Or rather it is only essentially contestable if what is being contested is the definition which best represents the spirit of the

4

Power, luck and freedom

original definition. And that might have no determinate answer. Instead we have competing notions of decision mechanisms and dispute over which is preferable, usually based on issues concerning the normative weight to be placed on different Arrovian conditions or types of manipulability. So the different definitions are rival in ‘bestness’ by some criteria, but not conceptually rival. In other words, it is not the concept that is essentially contested, but which of the different concepts is best applied scientifically (to explain what is happening) or normatively (for our judgements of socially just society). When it comes to terms such as power and more particularly freedom, the referent is more problematic, since ‘bestness’ is itself implicated in defining the concept. The subscript strategy can still be usefully adopted, but the debate over the use of the term is in part over the normative elements contained within that definition. And to the extent that those normative elements follow from the tradition (the language) in which one writes, the referent might be open to some necessary dispute. Writers who prefer to start from first principles, so to speak (and I guess I am one of them), are less concerned with historical precedents; hence they are less moved by the claims of essential contestability as Gallie develops them. We tend not to care very much about how far our conceptualization fits with previous ones. We care about logical consistency and the utility of the term in its application to the normative arguments we are developing. And these are developed within general normative theories (liberal egalitarianism in my case), with a close eye on their explanatory value in helping us understand the social and political processes of the actual world (again, at least in my case). Of course, one cannot completely ignore what others have said about a concept, and indeed the very fact that we each speak in a language shared by others means that in fact we never can, really, start from first principles. Indeed, most analytic conceptual analysis begins by critiquing some versions of a concept on the grounds of inconsistency, before drawing out a new version that avoids that inconsistency. That is what is analytic about analytic philosophy. (It is not always logical inconsistency, of course: it often also involves appeals to our moral intuitions or judgements. More on that below.) Writers in this tradition might be aware that agreement over concepts is unlikely, but do not see that it is necessary, since their very method presumes that the debate is about reaching agreement lest persisting logical inconsistency. There are a number of ways in which there can be different accounts of power and freedom without, necessarily, any substantial disagreement existing. Some disagreements can be merely verbal (Chalmers 2011). There might be many different ways of conceptually partitioning reality,



Introduction5

without any substantive disagreement about the way the world is. The subscript strategy helps deal with that. For example, it can help us in the dispute between defenders of the republican and the negative versions of liberty. As I discuss in Chapter 11, Carter (2008) suggests that referentially there is little that differs between his and Pettit’s accounts of liberty. They will agree in all cases (with an important caveat) over when liberty is gained and lost. We might assume therefore that there is nothing of substance that differs between Carter and Pettit, or at least nothing that has to be packed into a debate about the meaning of the word ‘freedom’. Using the subscript strategy, some of what Pettit believes is important when maximizing freedomP, something missing from freedomC, can be handled by Carter using freedomC and rightsC. They might both broadly agree about what should happen over a range of cases, or on what the good society looks like (and in fact I think they do), but want to describe that normative situation by different conceptual partitioning of the world. There might still be some further substantial disagreement between them about what should happen in some cases (their moral intuitions or judgements do marginally differ), but it needs to be demonstrated that these differences can only emerge because of conceptual disagreement. I believe that some disagreements do hinge on the way we look at concepts, but this fact needs to be demonstrated and not merely assumed (again, more on that below). The fact that we can partition reality in multiple ways without there being substantial disagreement does not mean all partitions are equal. We have criteria for preferring some partitions to others. Empirically, partitions that are more predictive (which I take to mean identifying real patterns in the universe) are preferable (Dowding 2016). I believe that moral and political theory also needs to be constrained or anchored by such real patterns. Simpler theories, all things equal, are preferable to more complex ones. We should not overpopulate our moral language with too many concepts. On the other hand, we do not want to ignore important distinctions. One simple rule of conceptual analysis is: make distinctions when they perform some role, and ignore them when they have no work to do. In social science generally, not only in moral and political philosophy, we need to remember what the theorist was trying to do with their concepts. We might want to extend a concept, or make some finer distinctions, because our question is rather different. This does not mean the concept might not have been fine for the job its originator was doing (see Dowding 2016: ch. 8, for the criteria of good conceptual analysis). It is also usually held that our concepts should match up somehow to our moral language. If moral philosophers want to affect public policy, then that is indeed the case. However, I am becoming less convinced that this is a necessary requirement of moral and political theory. I am not persuaded

6

Power, luck and freedom

it matters whether non-philosophers understand the technical aspects of some debate. If those technical details matter to public policy, then they need to be explained, though not necessarily in full technical detail, any more than scientists need to explain the technical aspects of what they do as it bears on society. (Scientists are not very good at engaging with public policy, though, so perhaps we should not follow their practice too closely.) The important point, however, is that whilst conceptual debate might at times be merely verbal, we might have reasons for preferring one verbalization to another, one of which may be how well it helps us engage with public policy. Some debates in moral and political philosophy I consider to be petty squabbles that are not worth spending much time on. In power analysis there has been some debate over whether the most important (or even the only proper use) of ‘power’ is ‘power to’ or ‘power over’. It seems fairly clear to me that ‘power over’ is a subset of ‘power to’. Whilst, normatively speaking, it might be the more important, it is not worth trying to demand that ‘power’ in social analysis should only mean ‘power over’. We can, of course, use the subscript strategy here, but the important point is that we are interested in the power of agents, even when that power is not straightforwardly analysable in terms of ‘power over’. If someone only wants to analyse society in terms of the power of some over others, I am content to let them get on with it. Some of my dispute with Brian Barry is of this nature (Chapter 5). He wants to define power in a certain manner (Barry 2002). OK, he can do that. But he then tries to use that definition to critique some of the things I say. That is not acceptable. Of course some of the things I say about power in society do not follow from his definition. However, they do follow from mine! That dispute is merely verbal. In Chapter 5, I defend my conception, and what follows from it. I think those implications are substantial, and provide a good way (better than others) to analyse the power structure and to examine the roles and responsibilities different groups have in the injustices and problems we face. The important issue is whether the conceptual demarcations I make – my way of partitioning reality – helps us in our explanatory and normative quests. I believe it helps us see why the collective action problem is an important aspect of power in society (see below). Importantly, part of what underlies my dispute with Barry (and others: Lukes and Haglund 2005; Hindmoor and McGeechan 2013) is a question of whether a particular concept, ‘systematic luck’, helps us here. Part of the dispute is ideological, or at least normative. Barry believes that calling capitalists ‘systematically lucky’ lets them off the hook for their responsibility for social injustice. I do not think so. First, because there is nothing in my argument that means they are not also powerful and can bear the relevant



Introduction7

responsibilities for that power and how they use it. Second, to the extent they are lucky due to their creating their own luck they do bear responsibility (which is the point of Hindmoor and McGeechan 2013). In Chapter 6, I make this point that agents can shape future luck (in that sense it was a response to Hindmoor and McGeechan avant la lettre). More pertinently, however, I think the distinction between power and systematic luck is important precisely because it allows us to see how the power structure operates and thus what we might need to change in order for it to operate differently. One can change injustice in capitalist society without lining the capitalists up against the wall (though sometimes that might help). Indeed, seeking blame often gets in the way of reforming a system. We can see this in a small way all the time, when court cases or judicial reviews pin the blame for some tragedy on a particular biological agent who made a mistake. We would be better off recognizing the distribution of mistakes across any class of agents and designing systems to reduce the occurrence or the importance of those mistakes. Furthermore, when it comes to blame we need to look at the incentives different systems create for people. The logic of finance capital, for example, leads finance capitalists to behave in certain sorts of ways. Rather than blaming them for their behaviour, we would be better off concentrating upon setting up incentives for them to behave in ways we consider morally superior. That is, design systems to reduce the occurrence of such agent-types. Nevertheless, the normative baggage that terms carry is important, as I discuss in Chapter 1. There I point out that ‘systematic luck’ is not so different from terms such as ‘hegemony’ and ‘soft power’, but these are associated with rather different normative implications. I argue there that we should make concepts as ‘non-normative as possible’ (which is one reason why I like the term ‘systematic luck’: its technical nature, developed from Barry’s concept of ‘luck’, carries less, though not, of course, no normative baggage). All terms are normative to some degree, but by defining them as non-normatively as possible, we open up debate between rival moral or political theories. If every concept is defined only in terms that make sense by the theory in which it is encapsulated, debate between those who hold rival theories of what is good or right is more problematic. This is just to say that it is easier to communicate if we all speak the same language, rather than having to translate every term of one theory into a different set of terms in another. Throughout these chapters my accounts of power, freedom and luck are designed to be as non-normative as possible. I will briefly run through a set of issues that tends to demarcate accounts of power. These are discussed at more length in the chapters in Part I. I do so, however, simply to set up an argument that is missing from all the chapters on the nature of power in this book. And that is that the concept

8

Power, luck and freedom

of power I define is static. I think any static account will always disappoint, because when we witness power it is dynamic. There is an important role for comparative statics that I will defend, but then I will discuss why power is dynamic, and why ‘luck’ is so important in that dynamic analysis. That will take us to Part II of this book. I will then run briefly through aspects of luck and how this bears on responsibility. Our notion of responsibility is closely linked, I argue in Chapter 7, to the range of our choices. Choice is also connected with freedom, the subject of Part III. My introduction then ties up the elements of choice, power, luck and responsibility with my discussion of freedom and rights. The concept of power There are many different ways of analysing the concept of power (see Dowding 2011). I am not going to try to discuss any, let alone all, of them. I shall instead discuss some of the lines of disputation. I shall start, however, by explaining why one particularly well-known account of power is not discussed in this book. Steven Lukes wrote a little book on power that has probably attracted as many citations as any book on the subject (Lukes 2005 (first edition 1973)). I felt it was horribly mistaken in two regards. First, it placed normative dispute at the heart of the power debate. For reasons mentioned above, I think this is the wrong way to go about conceptual analysis. Rather than celebrating normative dispute, one should try to define concepts as nonnormatively as possible and then examine the normative dispute. Second, I thought the entire argument about the different dimensions of power and what we could conclude from their failures was based upon ignoring the collective action problem (popularized earlier by Mancur Olson (1971)). In short, often one does not need to impute social power or ‘power over’ to an agent to explain others’ lack of outcome power or ‘power to’. Because of the collective action problem, people can be powerless all on their own. Once we have understood that, then we can examine how agents can make mobilization more difficult for others. That argument was given in my first book on power (Dowding 1991). I have only once, by invitation, returned to it again (Dowding 2006b). The interesting aspects of Lukes’s argument, in my view, are the structural aspects to which l now turn to (and see Chapter 2). The first important distinction in power debates is whether power is best seen as a predicate of agents or of structures. Clearly, there is a power structure – nobody denies that – but should we see power as all-pervasive or should we concentrate upon power as held and exercised by agents? For me, one problem with seeing power as all-pervasive, like Foucault



Introduction9

(1980, 1982) or Lukes (2005, chs. 4 and 5) is that it means power loses its critical facility. It simply becomes a term that denotes the fact that we respond to our environment. To be sure, the state of the world creates incentives for us to act in one manner or another. I am interested in how it does so, how we respond and how we can alter the incentives. I had two arguments against viewing power as structural (Dowding 1991). The redundancy argument concludes that adding the predicate ‘power’ to structure is redundant. We can say everything we want about environmental, institutional and cultural influences without adding ‘power’ to our claim. Second, the conceptual argument concludes that power is a dispositional concept. Agents have power given their relative resources, but can choose not to wield it. Having power and exercising power are different things. Structures are not like that. They cannot choose not to influence our behaviour. What is important about power is that holding it and exercising it are two different things. Peter Morriss (2002) explains carefully the exercise fallacy in his outstanding book on power. In Chapter 2, I make some important concessions to structural accounts from my initial claims. Structures affect us not simply by giving us incentives to behave one way or another given our preferences, but also by helping create and form those preferences. This is what I call deep structure. However, the important aspect of structural versus agential accounts of power depends on the questions that are being asked. Where the dependent variable, or what is being explained, is a token event – such as a policy outcome or an institution – then the independent variables will include agents as well as structural factors. However, when it is the behaviour of or effects upon a set of people that is being analysed, the independent variables will be structural. It is the question posed that determines whether the answer is given in terms of one side or other of the agency–structure divide. Structures are denoted by description. When we are theoretically discussing the power structure, their type denotes the agents. The agents are ‘capitalists’, ‘the prime minister’, ‘immigrants’, ‘bureaucrats’, ‘the firm’, and so on. To perform their theoretical role, they rigidly designate across different systems and then we can consider these descriptions within theoretical models, especially analytic or formal models, to be theoretical terms, akin to the theoretical terms in the natural sciences. And, as in the natural sciences, the theoretical models’ usefulness derives from how well they predict the patterns we find in society. The agents in the models, both in their description and in the detail of their behaviour, need not correspond directly to any biological individual or token collective agent in any actual society. Just as theoretical terms such as atom, or molecule in natural science do not correspond one-to-one but describe statistical properties of

10

Power, luck and freedom

such items in the actual world. So the utility of theoretical concepts comes from how the system gives incentives for these types of agents to behave. Actual firms do not all optimize their behaviour to maximize profit, but the capitalist system operates, as a capitalist system, to the extent that, as types, firms do so (Ross 2005, 2014). In my account of power (summarized in Chapters 3, 5 and 6; and see Dowding 1991, 1996), we examine the resources available to different types of agents in different sets of institutional and social circumstances. The resources agents have vary across systems, and the best way of identifying systems is by those resources. So we compare the powers of prime ministers in different countries based on the institutional resources they have. We compare the powers of prime ministers and presidents across the different sets of resources and expectations that they have. In all these cases, power is a function of relative resources. And these resources in these models are institutional rather than personal. The method is one of comparative statics that give the general outline of the relative powers of different agent-types. I think that is the right way to go for comparative static analysis. But it is inherently unsatisfactory in another sense. When we see power in action, it is not static. It is dynamic. I argue that we estimate the power of agents by their relative resources, and I break these resources down into five general categories (see Chapters 3, 5 and 6); but for particular analyses we might want to concentrate upon just a subset, and only a few within that subset. When comparing prime ministerial power across systems or time, for example, we usually concentrate upon prime ministers’ institutional resources (Dowding 2013). More generally, we can measure resources in different ways, depending on the precise system or element type-comparisons we are making. However, when we come to measure two actual token prime ministers, we look beyond the merely institutional to their personal characteristics (Dowding 2013). Those personal characteristics are formed through deep structure, including how others have treated them and perceive them now. This is the topic of Chapter 8. The obvious problem with the comparative statics of the first four sets of resources in my account is that actual token agents do not always use those resources equally effectively. In comparative statics across systems, that might not matter – effectiveness gets lost in the residual – however, the reaction of others to that effectiveness can multiply the effects. This matters. I tried to build that in conceptually in comparative statics by packing it into a fifth category of ‘reputation’. We can see this structurally. Sometimes the preferences of agent-types are satisfied within the system in which they operate more often than their resources (under my first four categories) dictate. There may be two reasons for this.



Introduction11

First, an agent can have a power reputation that goes beyond the relative resources that agent could in fact command. Second, that agent might simply be lucky. In Dowding (1991) I explained the success of agricultural interests in the UK by such systematic luck. The luck is systematic, since satisfying these preferences was built into the guiding principles of public policy. I also suggested capitalists are systematically lucky in that sense, though here they are lucky since their preferences are satisfied as a byproduct of the preferences of rulers. The governments of the world consider the banks too important to our way of life to fail (Bell and Hindmoor 2015). That knowledge gives finance capital the ability to behave as it does. It is systematically lucky, and that luck gives it greater powers in other ways. Having said that, my fifth category of reputation probably tries to hide too many sins, even for comparative statics, but the real problem for definitions of power – and not only my own – are their comparative statics status. In order to move beyond comparative statics to a dynamic account, we need to look at the interaction of relative resources, perceptions and preferences (or interests). Dynamic accounts and the role of luck The most overt exercises of power can be seen in competitive battles. Here we can often see the marshalling of resources – indeed, battles often begin and end with a display of resources. In the animal kingdom males strut to show their strength, and often what might be costly bruising battles are avoided as the smaller specimens slink off. Similar strutting occurs in the social and political world, most obviously in leadership battles, as challengers estimate their likely support before deciding whether to enter the fray. Yet perceptions can be seriously misleading. What appears to be muscle might be fluffed feathers, and in the subtle linguistic world what is cheap talk is not always obvious. How do agents in the political world increase their resources? Political battles are about coalitions. If we think about the Hobbesian metaphor of a war of all against all, where even the strongest can be stabbed in the back, having an ally at your back provides protection. We are familiar with the idea that political parties are coalitions of interests, governments are often coalitions of parties and electoral support is a coalition of voters. But dictatorships are equally the product of coalitions. That is the logic of political survival (Bueno de Mesquita et al. 2003). Dictators rely upon their supporters being bought off with side-payments; long-lasting dictators depend on shifting coalitions of support, marshalled to ensure the supporters compete with each other so that they do not challenge the dictator’s overall control. In fact, as I point out in Chapter 11, we can model human freedom or outcome power as a coalition. What

12

Power, luck and freedom

each person is able to do is what a coalition of others allows that person to do, since the rest of society could always stop them from doing anything. Mapping the exercise of power is a dynamic game particularly when you consider that coalition formation is a collective act, and as such suffers from collective action problems. In theory, any set of people could form a coalition. However, the basis of most coalitions is shared interests or preferences. The coalition of voters is really a tacit coalition of those whose shared preferences, given the menu of alternatives on offer, lead them to vote for the same party or policy. The coalition of society that allows you to go about freely is a tacit coalition of preferences with no immediate interest in trying to stop you. I use this fact against Pettit’s conception of republican freedom in Chapter 11, since he wants to define the outcome power or freedom of people as nondomination, where that concept is institutionally defined outside of the preferences, or arbitrary will, of people. But it cannot be so. Our freedom is always tacitly based on the consent of others. Often that consent is easily forthcoming, but where prejudice exists it might not be. Where the preferences of others are bound together historically for example, they might naturally form a coalition against someone’s freedom. Those preferences thus matter, something I will return to below. Power in terms of resources in a ‘preference-free’ manner tells us the comparative statics of the resource holding of agents. Power indices provide measures of coalition power of agents given their relative voting weights. That is all they do (Dowding 2000). When normalized, they give the relative power in terms of weights; otherwise they give probabilities that each voter will be a member of a winning coalition assuming equiprobable distributions of voting yea or nay (Felsenthal and Machover 1998). Those who rail against the power indices for providing misleading estimates are railing against the utility of such preference-free comparative statics. And often such preference-free comparative static analysis does not bear on the questions in which we are interested. If we are interested in the likelihood of some agent getting what they want, given the preferences everyone has, then the power index is the wrong tool. We can try to fix the power-index approach by attaching some probabilities to the distributions, but for a much clearer representation and explanation of likely outcomes, we need to shift to dynamic game theory (Garrett and Tsebelis 1999). Once we move from game forms (a more general approach, of which normalized power indices can be considered a subset) to game theory, we can start to model likely outcomes. We can start to model actual power struggles. Such models will, of course, be modelling types of agents, moreover types of agents who are defined by their resources and their preferences. These models are meant to represent actual social situations,



Introduction13

however, so the resources and preferences are drawn from those associated with actual agents of those general types. When we analyse the actual processes within which token individuals operate, we see complexities that are created by agents’ perceptions of each other and structures around them. This is where luck becomes deeply embedded into an account of power. In Chapter 6, I discuss the deliberate shaping of future luck through the preference shaping of other agents. Through persuasion, inducements and changing the structural position of other agents, one can shift their preferences and thus make them more likely to be coalition partners in the future. Such coalitions might be explicit, wherein these others deliberately support one’s aims, or simply tacit, where they are less likely to oppose one in the future. This idea of preference shaping has taken hold in recent years in international relations, where the concept of ‘soft power’ has taken off (Nye 1990, 2004). Soft power is coming to mean two rather different things. One meaning is using persuasion or offers to change other countries’ preferences. The second is changing their structural conditions, for example through trade. The concept is not very analytic. In my terms, the first is simply a form of power where the resources are used to conditionally change others’ positions in a positive manner. Probably only realist international relations scholars who assume power is about the military think the adjective ‘soft’ is appropriate for this use of resources. The second meaning, to shift one’s future luck by creating new potential coalitions of like-minded countries, is conceptually very different. In my terms, this is unconditionally changing the behaviour of others in order to change one’s future luck. Creating tacit coalitions is an important way in which luck can add both to the probability of getting what you want and, by changing your reputation, to your power resources. The inherently unsatisfactory nature of static conceptions of power is because in the exercise of power the contingencies of strategic interaction, of perceptions, of error and of chance come into play. These factors iterate and interact in ways that change the power of agents, through how agents are perceived both by others and by themselves. Recall the exercise fallacy. It is a mistake to identify the power of an agent with that agent’s exercise of power. Power is a dispositional, a capability that agents can, but do not have to, exercise. We cannot simply judge an agent’s power through what she does or does not achieve. She may be holding back. Indeed, that is the reason why we look to the resources she can command, in order to judge her power. But of course others have resources too, and how useful her resources are depends upon how others use their resources. That is not only true in conflict situations where one agent’s power might be matched by another, but also in cooperative situations. All we have achieved as

14

Power, luck and freedom

humans has been done in collective action and social cooperation. Our judgements of resources must therefore be relative ones. In Chapter 8, I use the idea of luck to explain how it is that some leaders appear to be more powerful than others but, more importantly, to argue that because they appear to be more powerful, they are more powerful: because how they appear to others affects how those others treat them. The results of that interaction feed into the next one, both in terms of how the agent views herself and how others view her. It creates a new reality of their power, and how that power is viewed, to feed into subsequent interactions. The perceptual argument comes in two forms. First, I base it upon my earlier accounts of luck (Chapter 7), which I will discuss below. Second, I base it upon the perception of whether a leader gets what they want despite resistance. If you are always a member of a winning coalition, you might seem to wield little power. You get what you want because others (say a majority) want it too. If you are pivotal in some committee, so always get what you want, but sometimes by siding with the right and sometimes with the left, then you might appear to be inconsistent. If the left and the right are vociferous and warring, then shifting from one side to the other makes you look weak, even if you are actually getting what you want all the time. That perception can make you weak, since when you need to fight others will resist, thinking you are someone who weakly compromises or gives way when facing resistance. On the other hand, if, as a leader, you are sometimes in the minority and fight to get what you want on those occasions, and sometimes because of your multiple resources you win, you can be seen to be strong. That helps in future battles. But the strong might get what they want less often than the weak. That is the luck of the weak in this example, but that luck also further weakens them. In Chapter 7, I explain more carefully how we need to handle luck and then develop its importance in another way. Whereas in Barry’s sense luck is defined as getting what you want without trying, and then applied at the type level to groups of people, particularly in the sense of systematic luck, I apply it here more to individuals, in the sense of what we might term ‘personal identity luck’. This is a key chapter, for it pivots the structural and the agential, especially over the issues of freedom and responsibility. Luck, choice and responsibility The point of proposing an individualist analysis of the problem of structural power is to separate two issues in power debates: the epistemological stance of power theories from the characterization of the distribution of power. Individualist theories were associated with pluralist accounts and



Introduction15

broadly just distributions – power being more widely shared – and structural accounts with unjust distributions. I have tried to show that one can describe a view that is individualist but in a manner that illuminates the structural features of society that lead to unjust distributions. In doing so, I hope to connect up macro-analysis of power distribution with micro- or individualist accounts of freedom and responsibility. As I have suggested, one of the problems I see with accounts of structural power is they make every influence upon individuals one of domination or power leaving no room for responsibility, let alone for degrees of responsibility. I want to maintain that everyone bears some responsibility for the conditions of society, though some bear more than others. Those who bear more responsibility are those who hold more power. We cannot simply blame anyone who gains for the losses of others or those who are privileged for the unfairness of social outcomes, since some of the gains are due to luck. We can blame people for not doing their bit to bring about greater social justice, but it is hard to assert such blame if the reason people have for not acting is that such acts would be pointless. If nobody contributes to some collective action that is the interests of all, then all are equally to blame. Many might be willing to conditionally contribute if others do, and feel that on those grounds blame should accordingly be conditional. Sometimes all that is needed is someone to coordinate, and perhaps persuade and cajole others. Some people are more fitted or able to do those roles and if they fail in these vital activities, then they must shoulder greater blame. It might seem unfair that those who could do most to change situations should also bear the greatest blame if action does not take place. But usually those with greater powers are those who are better educated, more self-assured and wealthier. These characteristics go with the abilities that enable some to facilitate coordinative or leadership activities. So any unfairness is only apparent, or so it appears to me. It further seems to me that there is a relationship between choice, luck and responsibility that bears on questions of powers and freedoms. We note here a tension. I have suggested that it might be better to move attention away from blaming people for mistakes and instead accept a distribution of mistakes and then design systems to mitigate their effect. Here I am suggesting that we can distribute blame. However, we can affect the likelihood of mistakes by the incentives we provide for people to be careful. We can affect behaviour by the systems we put in place. A few people have made quite a splash in public policy circles by suggesting people can be nudged into more favourable behaviour (Sunstein and Thaler 2008). If we place all the emphasis on structural or environmental effects on behaviour, measured by the distribution of behaviour across types, we leave no room for responsibility. The environment then explains differences in the

16

Power, luck and freedom

distribution of types of behaviour and we ask no questions about the distribution within agent-types. We can then explain differences within types of agents by the personal characteristics of those individuals. If we do this by genotype, we would be placing all the explanation on the environment (that is, the interaction of genes and the individual’s personal history). However, we like to assume free will and responsibility, but we need to assign that given the incentives people face. The argument of Chapter 7 examines this tension within the political context of a just distribution of resources. The argument of Chapter 6 is complex, and is directed more at an elucidation of luck egalitarianism – what luck egalitarianism really means – than the distribution of power in society. Obviously the two are connected. Importantly, however, it re-examines the notion of luck and treats it somewhat differently from how it is treated in the earlier chapters on power. To see the relationship between the two general ways in which I use luck, we need to take a closer look at luck. Growing out of the collective action problem is the issue that the power of a group should not be confused with the power of the individuals it contains. In voting contexts, a group has the voting power of all its votes added together. That is true whatever those votes together mean for what the group can attain. In large groups, each individual agent’s voting power might be small. Each individual vote is just one among thousands or millions. Some of the power-index paradoxes of power come about when a particular agent’s vote becomes far more important than its weight would suggest. If a particular agent’s vote is pivotal – the one vote required to change a decision – then that agent seems to take on far more power. However, we have to be careful when analysing this. Because that agent’s power, as measured by his voting resources, remains the same – viz, one vote – he takes on more importance in the decision because of the structure of the voting game given the way all the others are going to vote. If the individual agent is indifferent as to which way the vote goes, he can sell his vote for some side-payments. His power in this regard has come about from the structure of the situation, a structure outside his control. What is in his control is how he casts his vote. I maintain that this extra power comes about through luck; in the context of the power indexes, he has no control over how others are going to vote. This means his voting power, as measured by the power-index approach, has not altered. His extra power has come about through the luck of his situation. We can see this if we think about him having a strong preference for which way the vote goes. He is going to cast his vote that way, along with the other agents doing likewise. In that sense, he has no more power than they. Since I take power to be dispositional, I take it that, in the power-index approach to measuring power, it is only the weight of his vote that counts.



Introduction17

When extended more generally to collective acts, each individual brings to the collective act the resources they have. That is their contribution to the collective act. The dispositional power of the group is the sum, or rather the interaction, of those resources. Of course, often in collective action each person’s power resource only counts if enough others also contribute. A single consumer refusing to buy a mass-produced product on political grounds will have no effect on the company selling the product, since that one single non-sale will be lost in the noise of all the other effects on buying and selling of that product. The company will only notice falling sales if a substantial number of consumers boycotts the product, and they will notice those falls in sales only up to the number of boycotters. Often in collective actions we need to coordinate activities in order that the contribution of each agent will be useful to the outcome. I utilized Brian Barry’s concept of luck in this account. Barry (1991a) developed the notion of luck to critique certain claims in the literature about paradoxes that could be generated from the power-index approach to measuring voting power, notably the paradox of quarrelling members. Barry points out that the probability of getting what you want is not the same as one’s power. An agent might get what they want even if they have no control over an outcome, if someone else provides what they want. We can see that ‘outcome power’ here is similar to ‘freedom’. Even if I get what I want, but have no control over what I get, I can hardly to be said to be free to choose to get what I want. Barry defines success as power plus luck or, more strictly in power-index terms, luck = success − decisiveness. Decisiveness here is the difference one makes to an outcome in voting terms; it is the probability that one’s vote will be crucial to some outcome (or, in normalized measures, a score between 0 and 1: that is, a measure of the contribution to some outcome). He also says luck is getting what you want without trying. I took on these definitions, extending them a little in ways of which, it turned out, Barry did not approve. Morriss (2002: p. xxxvii) points out that, in the voting examples such as I have been discussing, the agent does vote, so it is not true he gets what he wants without doing anything. But we can take the casual definition to be just that, a casual definition. Luck in power index terms is the difference between actually getting what you want, given everyone else’s preferences, and the probability of getting what you want if everyone else randomizes their preferences. I’ll call this Barry luck. In Chapter 7, I return to the idea of luck but utilize a rather different account of it. I do so because I am discussing a rather different literature. In my 1991 book I distinguish the kind of luck used there from what I call ‘personal identity (PI) luck’, the luck of being who you are. I thought of PI luck here as being the luck of the individual resources you have, such as

18

Power, luck and freedom

personal strength in body and mind, the skills you have, but also the social situation you are born into. Someone born into a rich family has more resources from birth than someone born into a poor family. Someone who belongs to a superior caste has more power and freedom than someone born into an inferior caste. Chapter 7 interrogates this notion of PI luck, and to some extent closes the gap between it and Barry luck. Chapter 7 probes luck egalitarianism. Luck egalitarianism maintains that any inequality in outcomes (success) in society should be due to the effort that people put in and should not be dependent upon their luck. The basic idea is that justified inequality only includes inequalities for which people are responsible. If you gamble and lose, then you have to take responsibility for the luck involved. If someone gambled and won, they will be better off than you. But no one has any complaint. More particularly, however, if one person works hard and another lazes around, then the lazy should not complain that they end up with less than the diligent. Luck egalitarians thus take on the values of the ideological right, that people have to take responsibility for their lives, but remove all privilege and fortune that makes life easier for some than others. The ‘luck’ in the luck egalitarian account cannot simply be assumed to be PI luck. As part of their argument, luck egalitarians make a distinction between option luck and brute luck. Option luck is the outcome of a gamble explicitly entered into. Option luck can be insured against, and for that reason if you lose through option luck when you chose not to insure, that is your responsibility. Brute luck concerns outcomes that cannot be insured against. So, given that you cannot insure against who you are, PI luck is a form of brute luck. In Chapter 7, largely in order to help develop my argument, I distinguish between PI luck and other forms of brute luck that I call ‘contingent brute luck’. Contingent brute luck is important in my argument. Egalitarians, indeed philosophers in general, seem to take ‘talent’ as a given. They tend to think of talent as something that someone is born with, which is then, with effort, utilized well or less well. However, I argue that a great deal of how well such talent (itself identifiable with PI luck) develops is due to contingent brute luck. The higher one’s brute luck per erg of effort, the more diligent one might become because of the rewards. Furthermore, both observers and the person herself might confuse that brute luck with her talent, but because of that confusion, she becomes more talented than she would have done without that degree of brute luck. That is the iteration I use when discussing the power of token leaders. The major contribution of Chapter 7 to the debate on egalitarianism is to interrogate the notion of luck that operates across these three definitions. Barry luck is the difference between two probabilities of getting what you want (they make no cardinal utility calculations). So luck here is a type of



Introduction19

probability where the prize it awards is taken as given. However, I argue that generally speaking when we judge how lucky people are, the probability of their getting what they gained is less important than how much they gain. Someone winning a small prize in a lottery with poor odds of winning would not ordinarily be considered as lucky as someone who won a big prize in a lottery with better odds of winning. This idea of luck is then put together with the effects of contingent brute luck on effort and talent. I argue that the greater effort, and the greater belief in individual talent, will be determined, by and large, by the difference in rewards from that differential contingent brute luck and PI luck. In other words, I reduce responsibility down to what I call the ‘reward structure’. If one wants to reduce contingent brute luck and PI luck, then one needs to equalize rewards. However, in doing so, one also reduces responsibility. If all rewards are the same, there will be no luck, but there will be no responsibility either. I was never too sure whether Chapter 7 was a critique or an exposition of luck egalitarianism. Certainly it exposes a problem. Full egalitarianism, as the right maintains, would mean no personal responsibility for achievements, for every effort and every talent would be equally rewarded, leaving little by way of incentive to be diligent. On the other hand, if the reward structure does reward diligence, it will, if my account is plausible, also reward a certain degree of contingent brute luck and PI luck. That does not show there cannot be some optimal reward structure, though it does help much in deciding how we measure optimality here. Responsibility to freedom If we are to make room for responsibility over and above the incentives that predict a distribution of behaviour over types, then we are making room for freedom, at least freedom in the sense of free will. One of the ironies of Gerry Cohen’s critique of the incentive argument in Rawls is that his egalitarian ethos does not seem to require structure to incentivize behaviour (Cohen 2008: ch. 1) – strange for someone who was, at least once (Cohen 1976), a structural Marxist.1 The incentive–responsibility issue is also at the heart of debate over freedom. And it is how we handle it that determines our attitudes, and might mean that freedom might be considered essentially, that is necessarily, contestable, to the extent that it cannot be resolved empirically. In MacCallum’s (1972) famous analysis, all accounts of freedom can be analysed by taking the general account that an individual i is free from constraints c to do a. All differences in analyses depend on what we consider relevant descriptions of i, c and a. Without considering how correct that is

20

Power, luck and freedom

(and I consider it broadly correct), how we handle individuals in terms of how we handle free will, what types of constraints (physical, legal, material) and what we consider the proper scope of actions that are the subject of political or social freedom will help us demarcate different accounts of the meaning of ‘freedom’. The aspects I have been describing in the previous section primarily concern the first two. How we view the agent and how we see the constraints (and opportunities) affecting choice. The first two chapters in Part III of this book draw attention to the third element in MacCallum’s formula, the things we do or gain through our choices. I will consider that aspect, before returning to how it bears on the first two. Some aspects of how we consider freedom in these regards concerns our attitudes to what is valuable about freedom. For some people, the value of freedom inheres largely in what we can do with it. Freedom has instrumental value. Free societies are better than unfree ones because people can do more of the things they want. Few would disagree with that claim, but some also want to claim that freedom would be valuable even if that were not true. Even if a dictator could satisfy people’s wants better than they could be satisfied in a free society, such a society would be less valuable overall. Wherein that independent or intrinsic value of freedom resides is a major issue. There can be an instrumental and an independent answer to that question too. The intrinsic value of freedom might come about because part of what we value about what we attain is contained in how we reach those ends. A life is valued more because we have chosen it than if it had been chosen for us. Even if we regret the choices we make, our regret will cause us less anguish at having missed better opportunities than any anger we might feel for never having had those opportunities in the first place. This is sometimes called process utility, and we might try to value ends by including process utility within their overall utility. Here the intrinsic value of freedom is instrumentally derived from the utility of our choices, given they are chosen. For some, however, the very fact of choice brings value somehow independently of all these aspects: an intrinsic justification of freedom’s intrinsic value. I find this claim puzzling, and Chapter 9 is designed to throw doubt upon it. It argues that there can be no independent value in choice because optimal choice is not maximal. That is, maximizing the number of alternatives is not the optimal opportunity set. One can have too much choice, in the sense of too many alternatives to choose from. I think it follows that the value of choice must come about through process utility, and that derives from the fact that we see ourselves as choosing creatures. Chapter 10 provides more evidence to bolster the claims of Chapter 9. Seeing ourselves as choosing creatures is how I want to handle the issue of free will. Harking back to Chapter 7, if we do not have significantly



Introduction21

different alternatives, we will not seem to have any choice, would have no incentive to try to choose wisely, and so bear little responsibility for what happens. We would not feel free. Another aspect of free societies, at least in the market guise, is that they seem to provide a greater number and variety of alternatives. (This would then constitute a separate instrumental justification of freedom in society.) Choice on my account therefore has two separate relationships to freedom. First, part of the value of freedom comes about because we make choices. The utility we gain from alternatives is affected by the manner in which they arrive. (In his work Amartya Sen considers this fact to be a critique of standard economic interpretations of behaviour and of utilitarianism or welfarism; I do not: see below.) Choosing alternatives can be an important aspect of how alternatives affect our utility function. Second, choice is important to us, because we are choosing creatures. Consciously making choices is what we see as our free will. Choice therefore connects freedom of the will and social freedom. But freedom and choice are different things. Some argue that we are free even if we have no choice (Carter 1999; Kramer 2003). That is one way of viewing freedom, though I can see no value in such a freedom. In that sense, choice gives value to freedom. Does it matter whether we think no-choice situations are still free ones? Perhaps not. Does it matter whether they have any value? Here, it turns out, it does matter when we try to provide a measure of freedom, which we need to do if we are going to make the claim that some societies have more freedom than others. I touch on this issue very briefly in Chapter 10 (see Pattanaik and Xu 1998; Dowding and van Hees 2009). How we value items in an opportunity set will obviously predict the sorts of choices we make; or, rather, at least by the methods of revealed preference, the choices we make predict the valuation of the items (which can then predict the sorts of choices we are likely to make in the future). If there are no differences in valuation, then we can make no prediction either way, for there will be no choices, at best just picks. This is where, in Chapter 7, I suggest we cannot assign responsibility and we might think there is no freedom, or perhaps more correctly no value to freedom. If we define freedom simply in terms of the (relevant) constraints upon action, then we still have freedom even if there is no value to it. That is the line of the negative libertarians such as Carter and Kramer. In this account the preferences of the agent whose freedom we are considering are not relevant to an assessment of her freedom. In Dowding and van Hees (2007) we tried to argue that in strategic situations the agent’s preferences will counterfactually affect the probability of their being constrained, and they will be affected by the preferences of others, and hence will be a part

22

Power, luck and freedom

of the measure of that constraint. In Chapter 11 I argue, against Pettit’s version of republican freedom, that we cannot have a purely institutional account of freedom in which the preferences (or in his terms ‘arbitrary will’) of others can be discounted. I look at the supposed distinction between negative and republican freedom in terms of the nature of ‘resilience’ or robustness of freedom in Chapter 11. Republicans hold that freedom is not freedom from constraint as in negative accounts, but freedom from domination, where the latter is couched in terms of being subject to the arbitrary will of others. We need institutions that do not allow such domination. I argue that any such account cannot get away from the will or preferences of others for reasons already alluded to. Whilst we can compare across systems by people’s resources in comparative statics, when we consider the actual freedom of a person in a dynamic setting, the preferences of others matter. I argue that republican freedom is really no more robust than negative freedom, or only in the sense that republican freedom tracks liberal rights (Hohfeldian claims) as opposed to liberal freedoms. To the extent that outcome power is one’s addition to the power of a winning coalition, freedom is the luck of being a member of a winning coalition and is therefore robust only to the extent of others’ preferences. These preferences might be structurally suggested, thus giving some people more systematic luck in their membership of winning coalitions. Nevertheless, actual freedom must be measured to some extent by the structure of preferences in a society. We might change coalitions of preferences through preference change, importantly increasing freedom. Such structured changes can be achieved through law and through convention. To some extent, our freedoms are constituted not only by what is lawful (thus giving others incentives to respect freedom and rights), but also by what we come to accept as moral and rightful through convention. In Chapter 12, co-authored with Martin van Hees, we try to cash this notion out in terms of the existence of freedom and of rights in two three-fold categories: the existence of rights as tokens – actual things a person can do – and as type – the set of tokens available to types of people in a given society. Within each of these two categories, there are three types of existence of rights: by moral theory, by law and materially. So we can say, by theory M, right R exists; however, in Society S, these rights are not recognized by the law nor by the people, so no one can exercise them. A constitution might recognize that right R, yet because the law is not enforced, still people cannot exercise it. Finally, right R is recognized most of the time by most people, so it does materially exist; at this point the type–token distinction matters. For whilst it makes sense to say that a person’s right to life has



Introduction23

been materially violated because they have been murdered, it does not follow that ‘the right to life’ has been materially violated in that society simply because there is a murder rate. We suggest, however, that if the murder rate is high enough, then it does make sense to suggest that the right to life is materially violated in that society. This approach recognizes that preferences do matter to the existence of rights (and freedoms), and thus to their constitution (their definition). We can say that rights exist to the extent that we recognize they exist – by a moral theory, by law and most importantly by the manner in which people behave. Only if the rest of society recognizes their duties not to interfere in your use of your car do you truly own it. Only by others’ recognition of my freedom or liberty to behave in a given manner given their liberty to behave in a given manner do I have that liberty. I do not lose a given liberty if I cannot exercise it on a given occasion if the reason is that someone else has exercised her liberty, making it impossible for me to exercise mine. For example, I do not lose the liberty to pick wild blackberries because when I get to the local blackberry bushes I find someone else got there first and picked all the ripe ones. Our behaviour in these regards thus underpins the institutional powers, freedom and rights we have in society. We recognize individuals’ freedom to act and responsibilities for those actions within those institutional forms. Finally, Chapter 13, relies upon these material understandings of rights and freedoms. It returns to the issue of the value of freedom in terms of the utility we gain from alternatives. Amartya Sen has delivered the most sustained attack upon utilitarianism and welfarism as the grounds of morality. His impossibility of the Paretian liberal demonstrates that giving two people rights over a given choice will not necessarily deliver a Paretopreferred outcome – that is, an outcome both prefer (Sen 1970; 1982a). To some extent, that result is trivial. The collective action problem is one where Pareto-inferior outcomes derive from choices freely made. There is nothing puzzling there once the structure of collective action is specified. In Sen’s famous theorem, however, the Pareto-inferior outcomes take on greater importance because he uses it to argue that we should not, morally, prefer the Pareto-preferred outcome. We should prefer the alternative, since that guarantees rights. What is odd about the conclusion is that we, as liberals, should prefer something that, as welfarists, we do not prefer. Obviously, to make sense of this paradox, there must either be two notions of preference here or, in reality, two different things that are being preferred, despite a description that makes those different things seem identical. Sen believes the former is at play; I believe the latter. Sen has two different notions of preference: preference for the outcomes and preference for derivation of the outcomes.

24

Power, luck and freedom

So in his theorem Pareto efficiency is derived from preferences over the outcomes, but he argues we can only guarantee those outcomes in ways that would detract from their value by their derivation. We can only guarantee it by not guaranteeing the individual’s rights. Sen’s interpretation gives rights or freedom a value independent of valuation in terms of the utilities of regular welfare judgements. I see the problem another way. I think we can regard the outcomes in terms of their derivation, such that if they can only be derived in ways that detract from their value, then the outcome is not Pareto preferred. Indeed, I see that as true by definition: by the methods of economics. The ‘contractual solution’ to Sen’s paradox I discuss in Chapter 13 allow individuals to keep their rights by contracting not to exercise them. I argue that my way of seeing the situation is to be preferred because that is how we see rights in rights theory. We need to distinguish Hohfeldian claims from Hohfeldian liberties, and understand that holding each of these is different from exercising each of them. When we do so, Sen’s critique of welfarism does not hold up. Rights and welfare might need to be traded off at times, but the practical manner of those trade-offs render them relatively uncontroversial for those who accept plural values. Hence I do not observe the fundamental clash of valuation that Sen sees in his theorem. A note on the essays I have chosen to leave the essays much as they were originally written, with only some minor cutting and editing, particularly for consistency. I have updated the references where that is likely to be helpful to the reader. I have changed references to my own work from the original publications to the relevant chapters in this book, and occasionally added a crossreference to the chapter where a point is expanded. I have removed two paragraphs from Chapter 5 that reproduced material from Chapter 4 and replaced them with a cross-reference. I have made a couple of changes to terminology. In Chapter 3 ‘systematic luck’ was originally called ‘systematic advantage’, a term imposed at the time of publication. I was always unhappy with it, since luck can be bad as well as good (some are systematically unlucky), and have now changed it to be consistent with the rest of the book. In Chapter 9, the original refers to ‘choice sets’. In social choice this means something else (the alternative chosen), so I have changed it to ‘opportunity sets’, consistent with social choice and my later use. Finally, some of the figures have been redrawn, so are not precisely the same as the originals, and all have now been given titles. Essentially, though, the essays have been reproduced in their original form, with all their faults and problems.



Introduction25

Note 1 I am being a little unfair on Cohen here, since his argument relies upon the merits and interpretations of ideal theory, both for his motivation and that of moral philosophy. As he was wont to say to me when I tried to draw out broader implications from any aspect of his argument: ‘my paper is not about that’.

Part I

Power

1 Why should we care about the definition of power?

Introduction Power is a contested concept; of that there is no doubt. There are many dimensions in which it is contested. Some see it as a property of agents, whether these agents are individuals or collective entities such as firms, governments or political parties (Simon 1953; Dahl 1957; Weber 1978; Dowding 1991; Morriss 2002). Others insist that power is a property of systems or structures (Foucault 1980; Ward 1987; Clegg 1989; Hindess 1989). Some argue that power is ubiquitous and obscure and therefore cannot be measured (Lukes 2005); others that we can measure it (Dowding 1991). Some see power as necessarily conflictual. It always involves the clash of interests or plans and what one gains another loses (Dahl 1957; Weber 1978; Mann 1986, 1993; Lukes 2005): the power game is always zero-sum. Others believe that power is consensual: social power is about people working together to accomplish aims that could not be accomplished alone, hence increasing their collective power (Arendt 1972, 2004; Parsons 1963; Barnes 1988); in other words, power games are positive-sum. Others use definitions of power that encompass conflictual and consensual elements (Foucault 1980; Giddens 1984; Clegg 1989; Dowding 1991, 1996; Haugaard 1997; Morriss 2002). Scholars also debate whether power should be defined in terms of ‘power over’ – where the ‘real’ meaning of power is the power of one agent over another (or of the system over agents) – or in terms of ‘power to’ – where power is the power to attain what one aims for and ‘power over’ is simply a subset, a means by which some agents attain their goals (Pansardi 2011). Other concepts of power are, in themselves, general theories about the nature of relationships between people. Foucault’s notion of biopolitics or biopower is one example (Rose 2007; Foucault 2008). In addition to these more general accounts, there are related concepts that are used in a variety of contexts: authority, autonomy, domination,

30 Power

freedom, hegemony, influence, legitimation and manipulation, each of which might be as contestable as the concept of power itself. There are also subsets of some of the accounts of power that are used in specific contexts or within certain treatments of power relations. Such types of power include power as control (Sen 1982b) P-Power and I-Power (Felsenthal and Machover 1998), soft power (Nye 2004), voting power, agenda and veto power (Tsebelis 2002); or more general accounts of power that are used in certain explanatory frameworks such as ‘balance of power theory’ (Morgenthau 1949; Little 2007); bureaucratic power, hierarchy, pressure politics, power of business; and so on. It is not my purpose to review any of these concepts (for a comprehensive account of many of the issues, see the entries in Dowding 2011). Rather, I ask more general questions. Why do we have so many different concepts of power? Are they actually competing concepts or simply diverse uses of a general term contextualized in different settings? If the different uses are not actually competing, why is there so much disputation over the term? Why should we care about the definition of power? Certainly there is a general underlying notion of social or political power. There is something that holds these diverse uses together. Haugaard (2010) suggests power is a ‘family resemblance’ concept. It is also clear that many uses are not competing. Some specific concepts of power are contextualized to a specific purpose – the most obvious examples derive from the more formal approaches, in terms such as I-Power and P-Power, agenda and veto power. I-Power is the degree of a voter’s ability, under a given decision rule, to influence the outcome of a voting division (Machover and Felsenthal 2011a). P-Power is the voter’s expected relative share, under a given decision rule, of the constant sum reward of any given outcome from a voting division (Felsenthal and Machover 1998; Machover and Felsenthal 2011b). In the context of decision theory, agenda power is the power to order the items for consideration under any set of divisions under different decision rules, whilst veto power is the power to stop any outcome by one’s vote. Some other usages are defined in terms of what they are contrasted with and are developed to make a specific point or to be used within a given explanatory framework. Joseph Nye’s concept of ‘soft power’, for example, is contrasted with ‘hard power’ and designed to make a specific policy-relevant point in international relations. (I discuss this example in a little more detail below.) However, some of the ways in which power is defined directly compete with each other. Indeed, some are designed to do so since the conceptual dispute is part and parcel of the theoretical or normative debate. Those who claim that ‘power over’ or ‘zero-sum’ uses constitute the real meaning of power are claiming that usages of power in terms of ‘power to’ or in a ‘positive-sum’ manner are illegitimate in social



Why should we care about the definition of power? 31

discourse (for example, Barry 2002, 2003). Similarly, to proclaim that power relationships necessarily involve a clash of interests (usually an important element for those who see ‘power over’ as the only legitimate use of the concept of social or political power) is also to bring into the concept a strongly moralized component. It is these strongly moralized components of some uses of the term power that lead it to be proclaimed ‘essentially contestable’. What is essentially contested, of course, is different views about the way society ought to be governed. If these moral or ideological viewpoints are built into our very concepts, by which we also try to describe and explain the social world, then those descriptions and explanations will be equally moralized or ideological. It is for this reason I suggest below that all of our concepts should be kept as ‘non-normative as possible’. I will argue that to deliberately moralize concepts by bringing normative commitments into their basic definition is an ideological move that tries to win dispute by sleight of hand. Our basic concepts should be kept as non-normative as possible to ensure that we can be as clear as possible wherein lies our normative disputation. In truth, of course, the crudest attempts to colonize behaviour by inappropriately normative language fail, sometimes in spectacular ways. The propensity of dictatorial communist regimes to place ‘democratic’ in the title of their countries is one obvious example. Other well-meaning attempts to shift attitudes can also fail. To label children with learning disabilities as ‘special’ can be seen to fail when ‘special’ becomes a playground term of abuse, replacing ‘moron’ and other older, less deliberately moralized, descriptive terms.1 I will suggest that moralized definitions of power or social scientific concepts are normative or ideological ruses that should be kept out of analytic conceptual dispute. I do not claim that it is possible to define power in a completely non-normative manner, nor that any given definition might tend to support one ideological viewpoint over another. I do claim that some ways of looking at the world are superior to others, and if these support some viewpoints over others so much worse for the latter perspective. To deny such a claim is to deny that disputes in the social sciences can ever have any status beyond ritualized abuse. It is to give the status of social science that of the old logical positivist ‘boo–hurrah’ theory of morality and politics where all we ever do in moral and political debate is express our preferences. Rather, I argue that all definitions of power have normative implications, but those implications need to be kept in their proper place. The normative issue partly explains the degree of disputation over the concept of ‘power’. (The profession of academia and its ‘publish or perish’ imperative explains the bulk of the rest. The residual is what we should be doing.)

32 Power

The last question – why should we care about the definition of power? – is the harder one to answer. We have different concepts of power, since people have different normative commitments and political views about the world. Power is a key concept in social analysis and people want to stake out their own territory using a concept that is friendly towards their own views. There is so much disputation because there are different normative commitments and political viewpoints and these views matter to people. We care about the definition since we care about the disputation. However, that is not enough. If we leave the issue to the idea that people have different normative commitments and political viewpoints, hence the disputation, then we leave the concept of power not just contested, but essentially so. Essentially so, because there are different normative commitments and political views and that is all that can be said. Again, we are in danger of seeing such dispute as simply expressions of moral preference; and if we try to examine those different expressions as scientific observers we should then not care about the definition of power: let it be contested. But we do care. I shall argue that power is a theoretical concept that can play a different role in different theories or mechanisms that seek to explain aspects of human society. ‘Power’ can sensibly have plural meanings. ‘Power’ might be used in different ways, but in that sense those meanings are not rival. It does not follow from such pluralism, however, that all definitions are equally good. Some ways of analysing power might be more useful than others. Some definitions might be preferable to others, since the theories or mechanisms in which they are utilized are superior in some manner. In other words, even when concepts are non-rival in normative contexts, we might choose to use some rather than others, since some are explanatorily preferable. Nevertheless, we can define power differently without there being any fundamental empirical or normative conflict that is implicated in the definition (Dowding 1991: ch. 8). Such pluralism does not suggest that ‘anything goes’ in power studies; certainly, I think, some ways of looking at power are preferable to others. Nevertheless, non-rival but different accounts of power can be used in different contexts, depending upon the research question. Thus, rather than seeing ‘power’ as a battleground concept for which one version needs to triumph (a zero-sum ‘power game’ version of academic debate), we can accept different uses in different contexts. Power debate can become a positive game where we allow different conceptions to be utilized in very different contexts and for very different theoretical questions. Even so, as I have shown, there are some fundamental disputes where the notion of power is contested and where some academics (apparently) believe that it is very important that we see ‘power over’ as the ‘real meaning’ of power rather than as a subset, albeit an important if not the



Why should we care about the definition of power? 33

key morally important subset, of ‘power to’ (Barry 2002). This issue is deeper than mere dispute over the concept of power, for it involves dispute over conceptual analysis in empirical and normative theorizing. Here I want to argue that we should keep our basic concepts, such as power, freedom and equality as ‘non-normative as possible’. All of these concepts (indeed I will suggest all concepts period) are normative to some extent. However, some concepts are ‘more normative’ than others. Even so, I will argue, despite differing in degrees of normativity, each concept should be defined in as non-normative a manner as possible. Conceptual construction Our language designates patterns in the universe, whether they are the nominal denoting of similar objects or processes or more theoretical concepts that enable the generation of hypotheses. We pattern the universe in this manner to enable us to make predictions (Dennett 1998). Patterning the world allows us to see how our environment will affect us and how it will respond to our actions upon it. We see patterns in our fellow creatures and adopt the intentional stance about animals at least in part because it enables predictability (Dennett 1987). Predictions are thus built into our very language and understanding of the world. It is for this reason that explanation always entails prediction in this sense. Explanation is about patterning the universe through descriptive or causal inferences. It follows, however, that prediction and therefore explanation are in themselves normative. We want predictability because it is useful to us and such utility is normative – we want to be able to predict. We want to be able to explain. And we want to predict and explain those elements of our environment that are of most interest to us. (At its most basic we need predictability to be able to survive as an animal and reproduce as a species.) All concepts thus have an element of normativity and the better they enable predictions, the better the concept. Providing better predictions is the first test of the utility of any theoretical term. Science refines our concepts as scientific study discovers new aspects about the world. Terms such as ‘cat’ or ‘water’ might not seem theoretical; nevertheless, they are theorized to the extent that we might shift some animal from the class of ‘cats’ to some other related mammalian class if we were to discover morphological features that suggest re-categorization. Water has an obvious folk meaning, but chemical and atomic analysis enables us to understand it theoretically and relate it via its base elements to other aspects of the world, and to further understand features about it that affect its texture and taste. For example, natural water is ‘harder’, depending upon the quantity of dissolved minerals. ‘Soft’ water (generally

34 Power

treated to make it so) has only one positively charged ion through its sodium content. The minerals in water give it a characteristic flavour, hence some natural mineral waters are desired for their taste and purported health benefits. Science tells us much more about ‘water’ than our folk understanding, but its referent remains the same; we simply understand more about it. For that reason, such terms are often considered ‘natural kinds’ which might be thought of as terms for which science is authoritative or the identity of which is fixed by reality. The terms are rigid designators (Kripke 1980). Such an understanding might be based upon distinctions we make which denote real patterns (Dennett 1998). Patterns are real to the extent that they constrain prediction and hence conceptual, theoretical and descriptive analysis. Concepts such as ‘centre of gravity’ or ‘gene’ are more theoretical still. In the first case, ‘centre of gravity’ has no physical form, but as a theoretical concept can be given a location that enables predictions about the attraction of bodies. In the second, the term ‘gene’ is used in somewhat different ways in different contexts, but again enables prediction. We identify a given gene or allele at a specific locus by the effects it has. An allele can be regarded as having a certain selection coefficient relative to another at the same locus at a given moment. The coefficient is a number that can be treated algebraically and inferences drawn from one locus to other loci. In that sense, we identify genes by their statistical effects in terms of their function, location and relationship to other genes. So across subjects, we note differences in a specific locus on a chromosome and see statistical differences in the phenotype. Not all who have the gene will display all those characteristics associated with it, and some of those without the gene might well display some of them. In context, though, those with the gene are more likely to display those behavioural qualities. In other words, genes are identified by their predictive capacities and those predictions are always relative to environmental conditions. Some real concepts are entirely theoretical. ‘The centre of gravity’ is the centre of the mass of any object in the presence of a gravitational field. There is nothing to be seen there to mark the centre of gravity, but it can be perceived in the sense that enables predictions about the relative movement of large bodies. That is the reality of the centre of gravity. Some patterns are not real: seeing patterns in the rolling of fair dice can be misleading, as the gambler’s fallacy makes clear. (Though we might also note that finding a pattern in the rolling of a die might be the only evidence a gambler might have that the die is not fair.) Such terms might not count as ‘natural kinds’, because we might be able to make equally good predictions using different theoretical terms, though we note there might be creatures that have no use for our concepts



Why should we care about the definition of power? 35

which cover natural kinds, since they do not impinge upon their lives. So we might equally well say that theoretical terms constitute natural kinds. In other words, the concept of ‘natural kinds’ is itself relative to ourselves as perceivers of the world. What can we say about terms such as ‘freedom’ and ‘coercion’? Certainly they are more moralized than ‘gene’ or ‘centre of gravity’. Both ‘freedom’ and ‘coercion’ are theoretical terms, but their greater normativity arises, at least in our times, because they have – in the first case positive and in the second negative – ethical connotations. We should note that both terms also have some empirical or predictive qualities. Without getting too far into a conceptual defence of a specific notion of either term, we might expect a greater diversity of behaviour in a country with more freedom than in one with less. We might expect people who are coerced into some activity to be less cheerful than those doing the activity of their own accord. Such predictions are not irrelevant to the positive and negative connotations of ‘freedom’ and ‘coercion’ respectively. Greater diversity of behaviour within a community reveals different sets of preferences or interests that can be expressed in a freer society, whilst happiness is something that people tend to desire. I certainly do not claim that these (and other predictions) associated with magnitudes of these other terms in different societies exhaust the ethical connotations of freedom and coercion. For example, too much freedom can be associated with ­unhappiness – or, at least, people seem happiest when their lives are structured to some degree and structure might be thought to reduce freedom. There is a great deal of debate about the value of freedom as a concept in itself apart from other values that it might instrumentally bring about (Carter 1995, 2004; Dowding and van Hees 2009). Indeed, the positive ethical connotation of the term ‘freedom’ might not hold for all people or societies and might be thought to be a relatively recent concern. We should also note that the ethical connotations of such terms do not trump all considerations. If we use these terms in as non-normative a way as possible, we leave open the possibility that we can justify redu­ cing the freedom of a person or people, or can justify coercing a person or a people. From this point on, when I use the term ‘non-normative’ or ‘non-­moralized’ I mean ‘as non-normative or non-moralized as possible’. So we might conclude that terms such as ‘freedom’ and ‘coercion’ are more normative than terms such as ‘gene’ or ‘centre of gravity’, let alone ‘cat’ or ‘stone’. However, non-normative terms such as ‘gene’ or ‘cat’ might take on great normative significance. In a society that worshipped cats, for example, the term ‘cat’ would have ethical connotations exceeding those of one which merely saw harm in the maltreatment of animals, let alone one that did not consider animals worthy of any moral respect. Here then ‘cat’ takes on moral significance within a given moral theory. However,

36 Power

by keeping our basic concepts as non-normative as possible, ‘cat’ can be defined in terms that enable identification without its moral significance. The moral significance is something added that supervenes upon the physical attributes of cats. Indeed, the specific physical attributes would be likely to have important ethical significance. Any theory that gave great religious significance to cats would need to explain why cats were thought to have this property. This property might be god-given, but we would need to specify how it attached to cats. For example, if originally the cats so specified were a particular variety, the theory would need to specify how that variety was defined and how to handle any mongrel elements. Or, if all cats were specified, it might need to demarcate between domestic and feral ones, or any cross-breed, and so on. When it comes to religious significance, such distinctions have sustained many a monastery and quite a few armed conflicts. Or, if we see harm in maltreating cats, we need to state what maltreatment consists of and why it is wrong to inflict that upon cats. Such accounts of maltreatment of cats would then have to justify why similar treatment of other animals is not also unjustified, and so on. Again, we can see in the burgeoning account of maltreatment of animals how laws initially governing subsets of the animal world get broadened as the implications and consistency requirements of moral theory become recognized. And it is the reasons, and how those reasons develop over time, that explain why the class of animal maltreatment has expanded over time. We also note that ‘cat’ is problematic as a ‘natural kind’, since species morph into one another and decisions need to be taken both contemporarily and historically over lineages. Our social scientific concepts are thus theoretically normative, yet share some features of natural kind terms. We ought to think that a non-normative definition of power holds good across different settings and different possible worlds. Some concepts are primarily normative, and thus must stand at the top of the hierarchy. Terms such as ‘good’, ‘just’, ‘right’, ‘fair’, ‘benevolent’, ‘kind’ or conversely ‘bad’, ‘unjust’, ‘wrong’, ‘unfair’, ‘malevolent’ and ‘cruel’ must all sit near the top of the moralized hierarchy, since their intension clearly contains normative implications. It might seem odd to say that these terms should be defined as ‘non-normatively as possible’, but I will maintain that claim, even though ‘as possible’ with such terms might be as normative as one can get. The extension of such terms is then given by the lower-order concepts on which the normativity supervenes in the manner that Carter (2011) describes. In terms of a hierarchy of concepts relative to their normativity, ‘power’ stands as an interesting case. Power clearly plays a predictive role, but also has an ambiguous normative reference. When we say, of one agent A



Why should we care about the definition of power? 37

relative to another agent B, that the former is more powerful (with respect to some domain X), we are making a prediction that within that domain X agent A will achieve her ends more often than will B. Or if A and B come into conflict over outcomes within domain X, then agent A will prevail more often than agent B. Thus ‘power’ in the social sciences, as in the natural sciences, is clearly a predictive concept to be used within explanatory contexts. However, it is also a term that can carry strong, though ambiguous, normative connotations. Not perhaps as strong as either ‘freedom’ or ‘coercion’, since whether we see positive or negative connotations attached to the term ‘power’ will depend upon the context. Indeed, in that sense, power is a somewhat morally neutral concept. Consider one of the contexts in which power is most often discussed in political science: the power structure of the state. All states and communities must have a power structure. Thus the ‘power structure’ as a concept must be normatively neutral. Of course, in the context in which it is generally used, the power structure is considered critically, since it is applied to situations where power is considered to be illegitimately unequal or where certain actors have diffuse and hidden powers. But, we can note, it is the nature of the power structure and not ‘the power structure’ itself that must be considered illegitimate. Power can be seen in a normatively positive light. We might see advantages in increasing the power of citizens against the bureaucratic state, though of course that entails reducing the power of the state. In zero-sum or conflictual contexts, power must be a neutral term since it merely attaches to agents in a symmetrical manner. If we think some agents should be made less powerful, we are automatically assuming that others must be made more powerful. Power also carries positive-sum connotations. Clearly people can act together in concert and such collective action increases the power of all of them. Indeed, often agents’ lack of power in some regard is a feature of their failure to overcome collective action problems (Dowding 1991). In both zero-sum and positive-sum situations, whether we want those agents to gain or lose power will depend upon moral considerations about those agents. It is for these reasons that I have attempted to provide an account of power that is as non-normative as possible (see Chapter 5; Dowding 1991, 1996). It is not my intention to offer a long defence of that account here, but rather to defend defining power as non-normatively as possible. Why defining power matters We can, of course, define our concepts how we wish. We can see any concept formally defined or informally used by a social theorist as a technical term that might only take on that specific meaning in the context

38 Power

in which it is used. The term ‘veto power’ or, as it is used more normally the concept of a ‘veto player’, in game theory is such a usage. Formally, a veto player is any agent who can stop some policy proposal from succeeding. Tsebelis (2002) uses the concepts of institutional and partisan veto players as an explanation of policy stability under different constitutional regimes. In prime ministerial systems, the prime minister must control parliament, or at least be able to gain enough support to govern. Being elected separately from the legislature, presidents do not always secure the support of the legislature. This means that presidential systems have a greater number of veto players, which means they tend to have lower policy volatility. The number of parties within coalitions also affects the number of veto players, as does who controls the parties, and so on. ‘Veto player’ is a theoretical term which can cover various different constitutional, legal or de facto powers. Its utility derives from its predictive power and its role revealing sub-surface structural features of political systems. We can note, again, that the notion of veto players and veto power is as non-normative as possible. People’s attitudes towards policy stability and the role of checks and balances within constitutional regimes might vary considerably. However, even though concepts can be used in a ‘technical sense’ for the purposes of specific arguments, we need to be careful. Any term that is used in our ordinary language is likely to carry normative implications that cannot simply be rendered irrelevant by some formal definition. For example, defining a concept such as ‘human emancipation’ as non-­ normatively possible is, by my account, to be recommended, but it is bound to be morally further up the hierarchy than concepts such as ‘centre of gravity’. In modern western analytic philosophy, the term ‘social justice’, often simply rendered as ‘justice’, is usually at the summit of the hierarchy. All other concepts are defined in relation to and underpin the analyst’s concept of justice. (See Kukathas 2016, for an attempt to knock justice off this pinnacle.) When I claim that we should make our concepts as non-normative as possible, I am opposing those who want to moralize concepts so that actions they consider justified are only described in language with positive normative connotations. For example, Mansbridge et al. (2010) attempt to distinguish appropriate from inappropriate deliberative processes. Their argument involves the notion of ‘coercive power’ that they see as an actual or potentially causal relation between an outcome and the interests of two agents. For them, coercive power always involves conflicting interests. On their account coercive power is always wrong. However, it makes perfect sense to suggest that coercing people can sometimes be the right thing to do. For example, I sometimes coerce my children into certain actions with the force of my speech acts and implied threats or sanctions. My coercion here



Why should we care about the definition of power? 39

is in their best interests (and they admit so even as they are being coerced). For that reason, I think it better to define an agent’s coercive power in terms of that agent being able to change the behaviour of another agent under certain conditions, leaving aside what the interests of the agents are. Coercion is the use of force, threat of sanctions or manipulation of the conditions in which an agent operates, whether or not the intentions of the coercer are benign or malign. An agent A might manipulate evidence that leads another agent B to believe something (let us say something that is true) and do so for the best of reasons, and we might consider such manipulation as justified in the circumstances. That does not make it any less manipulative, however. These arguments follow from the account of power that is as non-normative as possible which I have defended elsewhere (Dowding 1991, 1996). That account disengages the concept of interests from that of power and includes manipulation within it. I recommend that account not because the clash of interests is not a vital aspect of power struggles, but rather because keeping the concepts of power and interests separate gives greater analytical purchase on difficult moral and political issues. My account of power is not the only possible one, and other uses of the term are perfectly legitimate, but they should, by my account, be as non-normative as possible. We can best illustrate how terms take on normative resonance by considering an example. I want to compare and contrast three terms: ‘hegemony’, ‘systematic luck’ and ‘soft power’. The three concepts are not, I think, identical, but in certain contexts they have much the same extension: that is, what we might predict using any of them is very similar. Their normative intension varies greatly, however, and that makes for a nice contrast. The original meaning of hegemony is, of course, one associated with a form of imperialism, where the dominant or leader state rules other states by implied power rather than direct force. Often this power is economic as much as military. Hegemony implies both conflict and consent. Hegemonic peace exists in international systems, according to some accounts, where a dominant power is able to impose its orthodoxy and ideology upon other states. Thus the modern liberal economic order is one of US hegemonic influence. Hegemonic wars occur where two states vie for dominance, not only as the major military but also the major economic and ideological forces within the international system. Within hegemonic systems, lesser states consent to the order, because they cannot militarily challenge it, but also because, given the political and economic stability it provides, they can gain some advantages. In other words, whilst there is a conflict of interests between the hegemon which dominates and the lesser states which are exploited through the system, the latter consent to a relationship that brings some advantage to them.

40 Power

Lesser states gain through a stable order rather than a warring one, but the benefits of such arrangements are greater for the hegemon, which continually cements its power. Hegemony’s modern usage in political theory owes much to Gramsci (1971), who shifted the application towards relationships between classes. In its Gramscian context, hegemony is again a term that covers a degree of both conflict and consent. In his consideration of revolution in the Prison Notebooks, Gramsci argues that the Bolshevik road to revolution is not the only one, and that where the state and civil society are closely interwoven, revolutionary wars or wars of manoeuvre are inappropriate. He suggests that a war of position is likely to occur where the conflict of interests is less apparent. In order to take control, socialists need to transform the attitudes and ideas of people. The interweaving of state and society in modern life implies a shared set of values that tends to hide class conflict. In a war of position that conflict needs to be brought to the surface; and the existing hegemony needs to be replaced by another. Related to these concepts are notions of cultural imperialism or hegemony where ideas, ways of life, what are considered morally correct attitudes, spread from the hegemon to other classes, states or cultures, as its products – books, films, music, information through the internet, social media or other forms of ­communication – dominate thought worldwide. The term ‘hegemony’ tends to carry negative normative connotations. It contains the idea of one agent (class, state, culture) dominating and exploiting others. It is true that it is often argued that hegemonic power systems bring order and stability, and therefore advantages to all. And Gramsci expresses the idea that revolutionary change with implied advantages involves replacing one hegemony by another, thus using the term as non-normatively as possible. Nevertheless, by and large, to describe the USA as hegemonic is an implied criticism, as is describing western liberal values as hegemonic or culturally imperialistic. I introduced the concept of systematic luck (Dowding 1991). The term is an extension of the notion of luck, where an agent gets what they want without trying. An agent might be powerful, but if they do not need to act to get what they want, they can be thought of as lucky. In its original context, both luck and systematic luck were supposed to be residual categories useful for explaining how some agents, groups or classes, seem to gain advantage without having to act. I tried to distinguish luck from power derived from reputation – that is where one agent B does something for another agent A because he fears A – and suggested that luck for A is where B acts to benefit himself in ways that also benefit A. Systematic luck is the idea that the very structure of society can benefit some at the expense of others without the first having to act.



Why should we care about the definition of power? 41

Initially I was trying to explain why farmers are a privileged group in liberal democracies, gaining benefits (subsidies and price control) from governments beyond those of other industries which, by measures such as voting power, economic strength and interpenetration of major parties, would seem to be more powerful. I suggested it was simply that governments saw defensive advantages in having a strong agricultural sector and this, combined with inertia and past institutional and constitutional legislative actions, privileged farmers. I mentioned too that venture capitalists and developers also seem to enjoy such advantages. The story there is that governments see economic growth as needing such capitalists and so operate to their advantage without these capitalists even having to act. In neither case did I claim farmers or capitalists had no power, but suggested that they often get what they want without having to try very hard. In some ways, systematic luck is a similar concept to hegemony. It involves the idea of conflict and consent. It involves the idea that one group gains advantages over another because the latter can also see advantages in acting in ways that help the former. To the extent that local growth is good for an urban community, that community (and its representative government) welcomes development, and, given competition between communities, local politicians might bend over backwards to provide incentives for developers to come to their community. Developers hardly need to lobby or threaten at all. Like the notion of hegemony, systematic luck could be applied to ideas or culture. Some groups might be systematically advantaged because all share some level of belief in a set of ideas – such as that economic growth is a good for all or that a strong agricultural base is important for any polity. The belief that to allow major banks to fail would be disastrous for all meant that governments bailed out the very institutions that created the global financial crisis with their irresponsible financial practices and have remained ineffective in curbing the personal financial rewards such bankers have continued to take. Those who have to pay the price are the public. Of course, the public did gain, or thought they gained, through the high growth that was maintained through the unsustainable economic practices, but it was not they who caused the crisis. It is the belief that major banks cannot be allowed to fail that, in this case, makes the bankers systematically lucky. Whilst bankers could argue that they had to be saved in order to stop economic meltdown bringing disaster to all, they were without genuine threats, since they were not in a position to disinvest and invest elsewhere. An agent can affect their future luck (see Chapter 6; Hindmoor and McGeechan 2013). The idea of structure-induced equilibrium (Shepsle 1979; Shepsle and Weingast 1981) is that legislative rules and institutions

42 Power

are set up in the interests of agents at a given time which then institutes their continued dominance. Thus agents can influence their systematic luck and future power by manipulating institutions. Changing the way in which we aggregate preferences by changing electoral systems, gerrymandering constituencies, devolution or shifting responsibilities to and from different tiers of government are all well-known means by which institutions can be manipulated to change the structure of future coalitions, ensuring that the relative levels of luck of different social groups alters. One may also alter outcomes not by changing the institutions themselves but by changing the composition within them: zoning laws, for example, can encourage certain types of household into some urban areas, cementing a natural support base for a given party. There can also be manipulation of strategic behaviour by giving ­information and encouraging people to behave strategically for their interests in order to promote your own. Kuran (1995) deals with preference change through the manipulation of preference falsification in closed and ­authoritarian societies. Preference falsification is the expression of false preferences motivated by the fear of social disapproval which may be incurred by expressing privately held views. This may lead everyone to falsify their preferences, although they in fact share private beliefs. Or agents can change other preferences by giving new information about the world. Knowledge is power, and control over information is an important power resource. But structurally altering individuals’ preferences is a means by which actors may try to ensure their future coalition partners and systematic luck. Tying households’ preferences over the economy to the mortgage rate by encouraging homeownership means that the interest of capital encompasses the interests of everyone. Households with large mortgages have a much greater interest in the state of the financial markets than they do even over the rate of pay rises, since a 1 per cent drop in the mortgage rate will bring greater financial rewards than a 5 per cent pay increase for many people. By tying a set of interests to your own, you can gain coalitional advantages by creating, in a systematic manner, your own future luck. Another way in which actors can try to make themselves systematically lucky is to change the preferences of other actors. Persuasion or winning rational arguments is one way, but the use of unconditional incentives is the most powerful. For the West, changing the preferences of peoples and their governments by encouraging the development of capitalism and liberal democracy and trying to limit the influence of anti-liberal forces will be of much greater long-term benefit than any other form of influence or use of power. In this sense, the ability to use one’s power to create future systematic luck is the best form of power to have. These ways of creating future systematic luck have much in common with notions of hegemony.



Why should we care about the definition of power? 43

My account of systematic luck has brought furious criticism (Barry 2002, 2003; Lukes and Haglund 2005; Hindmoor and McGeechan 2013), at least some of which seems to be directed at the thought that by calling capitalists systematically lucky rather than systematically powerful I am somehow letting them off the hook (see Chapter 5). In other words, the criticism is normative or connotative rather than analytic. (Morriss (2002) and Hindmoor and McGeechan (2013) do have some analytic criticisms related to whether it is really possible to empirically distinguish systematic luck from systemic power.) What is important here, then, is not the analysis as such, but the normative connotations of my choice of words. Perhaps the same analysis but using the term ‘hegemonic power’ would not have drawn criticism from these writers (though may have attracted it from some others). Joseph Nye’s (1990, 2004) concept of ‘soft power’ has attracted a great deal of attention in recent years. It is also a similar concept to that of hegemony or systematic luck. The basic idea of soft power is that organizations have capacities to indirectly influence the behaviour of other organizations. Used largely in the context of international relations, it was developed to counter arguments that the power of organizations, notably governments, was confined to ‘hard power’, such as military or economic force, threats and legal action. Soft power is based on culture, shared values, persuasion or requests that can be seen as reasonable. By using your own values and culture, advocating your policies and political institutions, you are able to attract (or resist) other agents. Nye suggests that sometimes a nation with greater hard power resources, such as military or economic might, might fail against another that utilizes its soft power. He elaborates: In international relations if one or more partners are convinced their point of view is correct they can both cooperate on the basis of negotiated agreements, and even participate in the process of finding consensus. In such situations the concept of ‘soft power’ is really working. This step forward towards the acceptance or the perceived attractiveness of a proposition brings with it the idea of appropriating that proposition as one’s own, a notion that is essential to ‘soft power’ as it somehow obliges partners to take action. The real weakness of ‘soft power’ is that it is non-binding and therefore often leads to inaction. (quoted in Calmy-Rey et al. 2007: 17–18)

The role that soft power plays in Nye’s analysis is to highlight an important aspect of social relations: that of consensus and agreement, and the role ‘power’ plays in influencing others through shared values. Part of the concept is unashamedly normative. It has been used as much by international organizations and through diplomatic channels as through academic ones. The idea is that often it is better and easier to use shared

44 Power

values than to try to get one’s way through bullying. Similar ideas abound in the concept of regime theory, though here power is not always seen as so amenable. After all, the values that might be shared might be those of a dominant group or ideology. We can see that these concepts of hegemony, systematic luck and soft power are all trying to conceptually capture similar processes. Hegemony is sometimes used explicitly as an ideological concept. It is based around an organization, state or class dominating other organizations, states or classes through the power of ideas, economic relations, culture, and so on with an explicitly implied criticism of the social or international order. However, in Gramsci’s original use (though less often in those utilizing Gramsci), the term is not used exclusively morally negatively. A new hegemonic order is implied in Gramsci’s work, with that new order being preferable to previous ones. Systematic luck is an attempt to describe why certain groups or classes (though the term could be applied to organizations or states) seem to dominate without explicitly using resources at their command (developers) or how certain groups might seem to get more than one might predict given the resource base they have (farmers). It was designed specifically to be not normative but technical, following Barry’s (1991a) concept of luck, though the term clearly was viewed normatively by some critics. Soft power is used to suggest that states might use the power of ideas and culture to persuade other states to follow their interests. It also seems to have been designed to make a normative appeal. The use of the term ‘soft’ to ­contrast with ‘hard’ power – the use of military forces or economic ­sanctions – is to suggest it is normatively ‘nice’. Yet the conceptual descriptions underlying hegemony and soft power seem identical, or at least very similar. The concept of hegemonic ideas is similar to soft power, yet the normative implications are contrary. Where soft power is seen in a positive light, the power of hegemony in thought is seen as domineering. In this way we can see both how different concepts (in the sense of different words) might have broadly similar descriptive content whilst having contrary normative implications. To analyse any particular exchange in these terms would require going to more basic concepts: what the shared values are; how they are communicated; who compromised over what; how values were taken on by one party. What we make of this normatively would then depend partly on that analysis and partly on our normative predispositions to such an exchange. These might determine whether we label the processes analysed as a soft power exchange or as ideological hegemony at work. Whilst it is impossible to develop purely descriptive concepts in the (social) sciences – all concepts carry some normative baggage; some carry more than others – we should still strive to conceptualize in a manner that



Why should we care about the definition of power? 45

is as non-normative as possible. It is always worthwhile breaking down the concept into its constituent components to see precisely where the normative elements reside. That is the reason why I think it best to utilize concepts which are as non-normative as possible. Essential contestability To the extent that the terms hegemony, soft power and systematic luck have the same extension, they can be thought to have the same meaning. To the extent that they differ in their normative connotations, they have a different meaning. Which term is chosen when analysing the same phenomenon reveals much about the intentions and normative commitments of the user. Power is one of the concepts often claimed to be essentially contestable (Connolly 1983; Lukes 2005). The fact that power can be defined differently in different contexts should not make it essentially contestable. We have plural uses for the same word, but that is often the case, even when those uses share many features in common. Rather, what would make it essentially contested are the normative implications that attend those different uses. Lukes (2005: 62–4), for example, suggests that power is essentially contested because people can agree about the empirical facts but disagree about where power lies. But that can only be the case if they disagree about the definition of power; yet, as we have seen, plural uses do not entail essential contestability (see also Dowding 1991: ch. 8; Haugaard 2010). It is because of the normative implications of using the term power, rather than elements that might be used in its definition, such as ‘resources’, ‘intentions’, ‘persuasion’, ‘force’, and so on. What is essentially contestable is who gets to use the loaded term. The criteria by which we judge evidence are biased through our language use. There is hidden or unrealized bias in the perlocutionary force of language and we can battle over key terms to shift that bias to our side. In that sense our discourse constrains us. But it cannot completely constrain, or new ideas and discoveries could not be made. Whatever biases exist can be set aside or overcome. Whilst the definition of ‘power’ matters, it does not matter that much. What is more important is that the term that is adopted in a specific context does the job it is intended to do, and no more. We can persist with plural uses of the notion of political power without implying in any sense that we cannot judge that some definitions are better than others or that some are completely confused and contradictory and should be set aside, whilst others have only specific technical applications. My entire argument has rested upon the claim that there is an underlying reality that constrains our language use. We can pattern the universe in diverse ways, and specifically we can pattern it differently with diverse

46 Power

theories and theoretical constructs. Specifically, plural patterns are possible where they do not conflict. Where they do conflict, it is because they entail different expectations or predictions about the world. Indeed, the test that there is such conflict is constituted of those predictions. These patterns we find in the evidence are real and it is their reality, their predictive capability, that constitutes the constraints upon our conceptual analysis. Without any such constraints there can be no way to stop the slide to radical relativism, where the definition of power really does not matter; or, rather, it only matters in the sense of a naked power struggle. These constraints are normative in the sense that prediction is itself ­normative – it is something we desire. Those who do not so desire prediction, or truth under a different guise, might play academic games, but these games we can ignore. And there is an undeniable hypocrisy in playing such games academically without living one’s life in that manner. All lives are lived constrained by prediction or expectations. Other stronger normative commitments are not so constrained, however. We can agree over all the facts about the power structure, but still disagree about how legitimate it is. I have argued that in order to bring out that form of normative disagreement, we need to make our concepts as non-normative as possible so as not to conceal that normative disagreement within a conceptual one. To try to hide one’s normative commitments by the conceptual mixing of the extensionality of a term and its normative connotations is an ideological sleight of hand. Such tricks might work for a while, but I am convinced that careful analysis will always reveal the trickster’s games in the end. And we can all be exposed as tricksters if we are not careful. Note 1 The terms ‘moron’, ‘imbecile’ and ‘idiot’ were originally applied to people who were assessed in terms of IQ (or forerunner intelligence tests such as the Binet scale) – where an idiot has an IQ of 0–25, an imbecile 26–50 and a moron 51–70. Psychologists saw these terms as simply descriptive. Some would argue that the IQ test is not in reality objective. More pertinently, since we tend to value intelligence positively, to give a class of people a name such as ‘imbecile’ or ‘moron’ can never be simply descriptive. It has to be evaluative as well. Psychologists rarely use such terms now and no longer claim them to be simple descriptors, partly for these reasons, but largely because they came to be used by the general population as terms of simple abuse. Nevertheless, they were not defined deliberately by psychologists for specific normative reasons. The term ‘special’ to cover all learning-disabled children was. Unfortunately, such attempts are not likely to succeed in the long run. We cannot quickly change attitudes through simplistic strategic discourse.

2 Agency and structure: interpreting power relationships

Agents or structures? There are many different accounts of the nature of social or political power in society.1 One of the deepest divides between those accounts is whether ‘power’ is predicated on agents or on structures. Within each segment of the divide reside other divisions, notably over what constitutes an agent and what is the nature of the structure of society. These divisions are deeply implicated within the broader divide I have identified. Some accounts of agents are themselves deeply structural; and some accounts of structure implicate agents as the holders of power. In this section I will begin by examining how agents are defined with relation to power holding, then examine different definitions of structure and how these affect the nature of power holding. Agents According to the basic idea of methodological individualism in social theory, the only agents are individuals where ‘individuals’ means biological human beings. Methodological individualists have, in the main, been happy to state that whilst we often write of other agents – such as firms, political parties, pressure groups, and so on – their actions can always be decomposed into the actions of biological human agents. Thus we might explain a political party adopting a specific programme and election manifesto in order to appeal to the median voter and so try to win the plurality vote and the election; the actual process of the political party adopting that programme can be explained in terms of the specific actions of members of that party. And those members may not all have adopted that programme in order to appeal to the median voter. Some might have supported it because they thought it was the ‘right one’. Some might have supported that programme because they realized during the debates that it was the best they could hope for, despite in their view it not being the ideal policy.

48 Power

Others might have supported it because they thought it would command most support in the country and they wanted victory (their way of expressing the academic notion of ‘appealing to the median voter’); others still opposed the programme and, at best, they will only give it lukewarm support during the election campaign. In this manner, the methodological individualist will argue that we can explain the actions of the political party as an agent in terms of the actions of each of its members that led the party to that programme.2 Part of that story will include not only the beliefs and desires of the members by which we explain their actions, but also the rules and institutions of the party that ‘structurally suggest’ what actions its members might take in producing that programme. That is, the rules of the organization will help determine the strategies of the agents. The rules will also, of course, structure the powers of different agents. One party might be highly democratic and its programme adopted following a series of votes involving all the members of the party; another might be hierarchical and its programme adopted by the leader supported by a small coterie of advisers. Nevertheless, the core idea of methodological individualism is that only agents act. Collective agents can be decomposed into biological human beings (together with institutional rules) and thus all social theory can be reduced to the actions of human agents. The methodological individualist need not deny that there are other agents that are causally important in social affairs. A prime minister might fall because of some natural disaster. Foot and mouth is a viral disease affecting cloven-hoofed bovids, notably cattle and sheep. The virus can, of course, be seen as an agent that was causally important in a prime minister’s fall. Without the outbreak of the disease, he would have continued as prime minister for many years. However, for the individualist, as far as social theory is concerned, foot and mouth is an agent exogenous to social explanation. What is important for social explanation of the prime minister’s fall is his (and his government’s) reaction to the disease (Boin et al. 2005). All sorts of natural disasters, such as AIDS, floods or earthquakes, might have massive economic and political effects, but they are exogenous to social explanation. In social explanation we explain the fall of governments due to natural disaster not in terms of the agency of disaster itself, but rather in terms of the state responses to that disaster: state responses which can be decomposed into the actions of biological human beings.3 Of course, we might have a completely structured answer as to why some prime minister failed if there were no actions whatsoever that could have saved her. We might consider this to be a form of determinism, where events conspire to bring down a prime minister and there is nothing that she, or anyone else, could have done to stop that fall. We note here that



Agency and structure49

this form of determinism is fully compatible with free will or autonomy, in the sense that the agent can be in full command of their faculties. Their beliefs and desires have not been determined for them. However, given the events surrounding the agent, that agent is unable to stop the process that leads to her fall. There was nothing she could do. Even this ‘fully structured’ explanation of the fall of a prime minister is compatible with a strong form of individualist explanation, since the fall is still the result of events caused by something exogenous to the social reality, but those events still followed the actions of others in society. The ‘prime minister’s fall’ is an event that gets its meaning in a social context (of prime ministers, of leaders, of failure) that requires social action by agents. The prime minister’s fall might be due to the action of some virus (in cattle, or in the prime minister herself – she resigns because of illness), but the social explanation is couched in terms that are reducible to (sets of) individuals as agents. The idea of ‘structural suggestion’ here is that, given the biological individuals’ interests, the environment they face – both the social or institutional rules and the interests of other people – will structure their behaviour. For social scientists, the methods of revealed preference theory should allow us to rationally reconstruct individuals’ interests from their behaviour given the strategic situation that we can model. What might distinguish the commitments of most methodological individualists from those of structuralists is the provenance of interests. Whereas structuralists (as we shall see below) tend to assume that individuals’ interests are determined by the structure; individualists tend to assume that individuals’ interests are not so determined – at most they are ‘structurally suggested’. We might assume that there are some basic or fundamental underlying interests – such as the desire for warmth, food, status, and so on – common to all people (and if caused, then caused by our biology) and how we promote those interests will be suggested to us by our place in society. Exogenous interests are those so suggested, endogenous ones are those we choose more independently, such as a preference for detective over science fiction, for example (Dowding 1991; Dunleavy 1991). We note briefly here that structure is not completely absent even in the most individualist of methodological individualist explanation, because the structure of an organization will affect the outcome no matter what the preferences of the actors are. That is, a party composed of the same individuals might reach different policy programmes depending upon the institutional rules governing how policy is adopted. This will happen both because different preference-aggregation mechanisms produce different outcomes and because different mechanisms will affect the strategies actors adopt in order to promote their favoured outcome. However, in terms of the power structure, methodological individualists will predicate

50 Power

power to the individual agents. The power of the political party (to win an election) will be a result of its strategies in relation to the strategies of other parties; the power of its members with regard to policy formation within the party will be a result of the strategies they can adopt to help promote the outcomes they want; and their power to win office is based on their ability to adopt the best programme that furthers that aim. A further manner in which mainstream ‘methodologically individualist’ explanation is structural is the fact that in most models of social and political processes biological humans are not the agents. Rather, agents are composed of social roles such as ‘leaders’, ‘bureaucrats’, ‘consumers’, ‘voters’, ‘middle-class professionals’, and so on. As we see in the next section, one of the major ways in which ‘structural’ explanations diverge from individualist ones is in the provenance of individuals’ interests. Structuralists understand individual interests as being caused by social structure, whereas individualists tend to see agents as reacting to structure in defining their interests. However, in most social explanation, agents are seen as social roles and those roles are largely defined by structure in terms of expectations of the role and the interests of the agents occupying those roles. Thus in a second way virtually all individualist explanation has structure deeply implicated within its very form. Still, we can maintain that individualists see power as predicated on individuals; and we can mention many analyses of power that view power in this manner. Weber (1978) defines power in terms of an agent’s ability to get what he wants despite resistance. Dahl (1957, 1968) sees power as the ability of one agent to bring about changes in the activities of another agent. Oppenheim (1981: ch. 2) likewise defines social power in terms of how one agent (person) affects the activities of another agent (person). Such examples could be multiplied. The importance of individualism in all these analyses is the assumption that only agents act; and power is a causal notion that is applied to the causing agent. Now structure, as we have seen, cannot be irrelevant to social explanation; and it cannot be causally irrelevant. Two political parties might adopt different policies, not because the individual agents have different views over policy, but rather because of the institutional rules by which those views are aggregated. And, unlike natural disasters, such mechanistic structuring cannot be considered exogenous to social explanation. After all, we might well want to explain socially why each party adopted different institutional rules. However, therein lies the rub. For the individualist, the explanation of the social rules will have an individualist explanation, even as those rules affect specific outcomes today. For the individualist, structures are not irrelevant to social explanation, but they will also, ultimately, have individualist explanation. We return to this point, and its converse, repeatedly.



Agency and structure51

Structures Most simply, a structure is the relationship between variables. Take a set of red bricks and put them together one way, and one gets a bungalow; place them another way and get a two-storey house: same bricks, different structure. However, the two-story house will also require a staircase that will not be needed in the bungalow. Breezeblocks are a different type of brick, and could be placed in relatively different relationships to build a bungalow or a two-storey house, though these houses would not be identical to the red-brick ones. We might make the houses look identical, perhaps by putting red bricks as fascia around breezeblocks, and using plasterboard and skimming inside to make both outside and inside identical. The houses might appear identical to one another, but there will be qualitative differences, noted perhaps in heating bills or during renovation. Society could be viewed the same way, with people taking the role of the bricks and social structure the form of the relationships between people. Thus genetically identical people might be formed into different social structures – one hierarchical and one egalitarian, for example. The different social structures which place genetically identical people into different roles vis-à-vis one another will form a different system of power. A structuralist would examine the relationships between the variables ‘people’ and thereby describe the power structure. A structuralist here would predicate the term ‘power’ to the structure, for it is the structure that both constrains and enables the options of agents. It is those relational factors which determine the outcomes that result. The individualist would interpret those relational factors as properties of the agents. Whilst those properties are given to people by those different structural relationships, the individualist sees them as properties of individuals. Using the language that I have used in the past, each agent’s resource would be dependent upon the structure of relations, and those agents’ resources would depend upon the way in which we could both identify and provide some measure of their power (see Chapter 5; Dowding 1991, 1996; Morriss 2002). We can see in this representation of the agent–structure divide a commonality that distinguishes one from the other only by where the emphasis is laid. If we concentrate upon the set of relationships that constrain and enable, then we privilege structure; if we concentrate upon how those relationships constrain or enable the agents to do what they want – that is, we concentrate attention upon individual agency – we examine the power of individuals. The structural forces of one account constitute the full set of resources of the other account. In Dowding (1991) I argued for a resource account that puts ‘power’ firmly into the agency camp, using a ‘redundancy’ and a ‘conceptual’ argument against making power a predicate of structure. The redundancy

52 Power

argument claims that we can always remove ‘power’ from sentences where it is used in terms of structural power and produce sentences with the same empirical implications. The use of the term ‘power’ here is redundant. I also claimed that conceptually ‘power’ is closely aligned to cause and thus to agency. Since agency is naturally aligned to agents, the natural use of the term ‘power’ is predicated to agents and not to structure. At the same time, however, I believed my rational choice account of power crossed the structure–agency divide. As many others have also claimed their account crosses the agent–structure divide (see below), I still believe that ‘power’, at least in the English language, is more naturally used as a predicate of agents than of structure. However, my two arguments only really apply to this first cut at the agency–structure relationship. When we look at a deeper account of structure, the two arguments lose some of their bite. The analogy of bricks and buildings with people and structures can only be taken so far. Whilst bricks might be affected by the structure of relationships in which they are bound – loading factors would vary across the structure, perhaps causing some bricks to crack or fail – we would expect human agents to be affected to a much greater extent. And people are affected in two ways. One is ‘of the moment’. The situation in which they find themselves might affect their very being. For example, a sportsman’s confidence might be affected by his opponent’s demeanour. That confidence leads to greater effort and eventual victory. Here we might say one person’s action is (a part of) another’s structure. Seeing the lack of confidence in one’s opponent brings confidence and that creates a new person for oneself: a newly confident person. And that, of course, leads to the second way in which structure affects people; indeed, structure largely makes them who they are. To continue the sport analogy: our sportsman’s victory brings greater confidence still and renewed effort in training, so making him a champion. A sequence of luck can turn one person into a winner with a winner’s temperament and another into a ‘choker’ who fails at the last hurdle. Both genuinely have the qualities with which they are labelled (‘winner’ and ‘choker’), but these have become part of their identity because of events that occurred: events due to the situation in which they found themselves and perhaps a smattering of chance (see Chapters 7 and 8). In other words, history makes the person. It alone does not make the person, perhaps, but in the way agents themselves are structured by their environment and that environment over time, history makes them. It is in this manner that structuralism goes deep inside agents. I will call this ‘deep structuralism’; it is deep structuralism that forms the strongest basis for my account of structural power.4 In most individualist analyses of power, an agent’s power is measured by what she could achieve given the circumstances in which she finds herself.



Agency and structure53

The individual might ‘bring something’ herself to the achievement – skills and talents provided by her genes, say – but even here what a person’s genes can provide is determined by the environment.5 A breezeblock brings something different from a red brick to the structure into which it is built – the breezeblock might withstand forces differently, might provide different thermal properties – and in the same way genetically different people might respond differently to social structures. For example, some people might behave more responsibly than others in an egalitarian society; some might be more prone than others to abuse their positions within hierarchical systems, and so on. However, our expectation is that structural relationships have bigger effects upon human agents than they do upon physical entities such as bricks. To a much greater extent, human agents are constructed by their relationships to one another. For example, both the original formulation of the authoritarian personality (Adorno et al. 1950) and its more recent version in the work of Altemayer (1996) argue that submission to authority and the expectation of how to dominate when in authority are established by how one has been treated within a household during one’s early life. The argument here is that we are partially built by our environment, a view few would dispute – though many would dispute the analysis of the authoritarian personality (for a good review and critique, see Martin 2001). There are two ways, then, in which individual human agents and structure might work together to explain power in a society. One way is to assume that human agents are fixed in some manner – we take their desires and beliefs (which together form their preferences) as given – and then examine how they behave given the environment in which they operate. Formal results in social and rational choice theory demonstrate that identical preferences lead to different outcomes given different aggregation rules. It follows that we can expect to see different outcomes from hierarchy than from bargained decision-making, different results from different electoral systems, and so on. Now it must be recognized, even at this first level, that examination of these fixed preferences needs to take into account the structural or institutional relationships that exist. Individual agents can be expected to respond to different institutional relationships with the different strategies available to them: that is, they respond strategically. Any institutional form can be manipulated in one way or another. Different forms of manipulation will work best in different circumstances (Gibbard 1973; Satterthwaite 1975) and different tactics of manipulation will be adopted given the possibilities open to both the powerful and the weak (Goodin 1980). However, we can imagine here that people have underlying preferences, and these together with the institutional form lead to their behaviour.

54 Power

At this level of analysis, we should acknowledge that different structures have different effects upon behaviour. Some institutions are constructed of expectations about ways of behaving. Norms are constructed about how we should respond to one another in different social contexts. Different societies and cultures generate different sets of expectations. In game-­ theoretic terms, different equilibriums have been reached by agents playing essentially the same game, and complex games often have multiple equilibriums. Equilibrium strategies are maintained by rational players, since it is in each player’s best interest to maintain that equilibrium strategy. Some equilibriums will be more stable than others, since out-of-equilibrium action or exogenous shocks might lead rational players either to play new strategies or to return quickly to the original equilibrium strategy. Games which are straightforward assurance or coordination games need little explanation of why particular strategies are adopted. Other cooperative games, where free-riding is possible, might require more policing either by the players themselves or external agents. Games where conflict is more endemic are likely to require greater policing still. The point remains, however, that different institutions are more or less stable, and thus more or less enduring. Some structures are more enduring than others. Those which do not require agents to have ‘power over’ others are more enduring than those which do require such power usage, since structures that do not invoke ‘power over’ can endure with each agent merely using their ‘power to’ get what they want, especially in mutually beneficial ways. The greater amount of ‘power over’ needed to keep a structure in place, the less enduring or stable it is likely to be, for it is more prone to suffer when exogenous shocks shake the power structure or when losing groups collectively organize to change the power structure. Game-theoretic analyses of institutional forms can explain a given equilibrium and can explain why some institutions (those with more stable equilibriums) endure longer. However, why one equilibrium solution has been adopted rather than another will not be revealed by the game form itself, but requires particularistic historical explanation (Dowding 1994). One set of equilibrium strategies might have been adopted historically by pure chance, perhaps, or might have served some interests better than others. Most obviously, rules governing property rights can secure the rights of those who already hold large estates, or can be rewritten to spread property over a wider community. Which interest prevails is likely to depend upon specific historical locations. Acemoglu and Robinson (2006) explain the growth of democracy based upon the specific needs of securing the continued interests of the governing elites. The rich elite controls the state in non-democratic societies and the poor can only contest power through revolution, especially when the opportunity cost is low – that is,



Agency and structure55

during very hard times. The elite can buy off the poor through redistribution, but can only credibly do so through democratizing. Where society is relatively equal redistribution will be limited; where it is highly skewed towards the rich, redistributive policies will make them much worse off, encouraging them to counter-revolt through elite coup after a period of democratizing. For Acemoglu and Robinson this explains the essential difference between Latin America and northern Europe. The latter had a more equal society at the crucial democratization stage. The Acemoglu and Robinson story shares features with many other historical accounts that examine large differences in the fabric of social structures. They posit a key variable within a social structure, in their case relative equality, to explain historically diverse courses of history. Such path dependencies can reinforce existing structures or they can lead to big changes, as in the democratization process. The essential elements of such stories are that individual actors respond to the incentives provided by their environment in diverse ways, leading to completely different sets of outcomes. We can explain what is going on in the histories by the relationship between agent and structure seen as this first level. Agents respond to their environment and their actions might reinforce existing structures or, at crucial or critical junctures, lead to changes in those structures. Such changes might be planned by agents – the transformation of the Republic of Turkey by Mustafa Kemal Atatürk is a particularly dramatic example – or they might come about through the unintended consequences of a series of agents’ actions.6 A second level of relationship between agents and structure exists, however. Not only do structures create incentives for agents to behave differently given some underlying set of preferences, but structures can be thought to delve deeply into agents themselves to create those underlying sets of preferences. This I refer to as deep structure. The deep structure not only provides incentives for agents to act in certain manners given their objectives, it creates those objectives for them. One way of making this point is that it is not that individuals have created their own propositional attitudes towards the world, but that those attitudes are created by the structures. So it is not that I believe x, desire y and, given structure Z, I will do α to achieve y: the individualist agency story. Rather, structure Z creates in me my belief x and desire y. In this sense, structure becomes the agency as it creates the neurological activities that lead biological human beings to act. The ‘agents’ have become mere transmission agents for the s­ tructure. Thus in its most radical form deep structuralism takes agency from individuals and gives it to structures. At the level of deep structuralism, the issue of structure and agency is in fact one of free will or autonomy.

56 Power

Now, it is not clear to me whether there are any, or at least many, social theories that are deep structuralist in this sense. To be sure, the issue of autonomy is broached and discussed at the edges of the agency–­ structuralist debate. Theorists accuse others of not allowing room for agency or autonomy, but few explicitly assert that human beings have no autonomy or free will. Few proclaim deep structuralism explicitly. Most authors in this field tend to claim that they are transcending the agency– structure divide or providing a solution to the agency–structure problem. I turn to the latter in the next section. Beyond agent–structure Most writers who have considered the subject argue that social analysis needs to bridge the agency–structure divide. Often it is rather difficult to see precisely how they differ in their attempts to do so (despite their critical comments about others’ attempts). In many cases the major differences seem to be in terminology rather than substance; the rhetoric of their arguments differ more than the analytics. Talcott Parsons sees individuals working within a physical environment to which they responded; and within a social environment which they both help to create (and recreate) and also to which they responded. In The Social System (1952) he examines the pattern of relations in terms of the systems theory that was prevalent at that time. The essential element of the relationship between structure and agency is of role development. Individuals create understandings of their role in society by internalizing standards that they have learned by seeing other structured behaviour. The structure has a two-fold ‘binding in’: first, an internalizing of the standards of conformity that has instrumental or expressive value to the person; and, second, through the structuring of the reactions of others to the agent’s conformity with the standards expected of his role. Value patterns are thus created by the reactions of people: first, in how they see their own role and, second, in how others view that role and how they view the agent’s view of that role. The latter also reacts to the agent’s playing of the role through negative and positive sanctions which then modify the agent’s future playing of the role. The attempt here is to see how structure (expectations about roles) affects agency, whilst also allowing agents to interpret their roles in ways that change the structure, especially given how others react to those interpretations. On my reading, Pierre Bourdieu (1977) attempts to reconcile the concepts of structure and agency though his own further concepts of ‘field’, ‘habitus’ and ‘capital’. The agent is socialized into a set of fields. A field defines a set of roles and relationships within given sets of social domains.



Agency and structure57

These fields are relatively autonomous social spaces which socialize humans into roles. These roles and relationships change or evolve over time and their worth or capital is defined by the prestige, the monetary rewards and the status afforded to each role within a field. Thus different activities might be differentially rewarded within different fields, and actors within those fields would develop different value systems based upon those rewards. Nevertheless, each field is envisaged as a competitive environment for the different types of worth or capital. We can view these fields as professions, such as ‘academia’, or as more generalized social spaces in which agents compete. The agent then learns her role and the relationships that role has within the field by internalizing a set of expectations and normative understandings of her role. Her internalized understanding of what is expected of her and how her role fits with others in the field develop over time and forms her ‘habitus’. Habitus is the system of dispositions, or enduring ways of looking at the world, that an agent develops in response to the objective conditions in which she finds herself (the structure of the world around her). The objective structure thus enters into the subjective awareness of the agent, who then processes that structure into her interpretation of the expectations of her role. In this manner, objective structure enters into the very agency of a person, but the person is still the agent, so to speak. The habitus becomes important as the dispositions imbued create the actions of the agents and these, of course, constitute the field and give it meaning. So habitus creates the structures of the field, and the field then mediates between habitus and practice. Critics of Bourdieu contend that habitus leaves no room for individual agency, since it seems to determine the whole of an agent’s disposition to act in one way or another. Defenders suggest that Bourdieu leaves room within the fields of expectation for individuals to interpret what is expected of them. I have taken the latter line, as it seems to be how the agency–­ structure divide can be bridged. However, Bourdieu does not write much about the autonomous side of agency; and in that sense he is usually regarded as a deep structuralist. Indeed, he suggests that agency cannot be directly observed but only be experienced in practice or through habitus. That is not so far away from revealed preference theory, where we interpret an agent’s preferences (beliefs and desire) to reconstruct her agency. ‘Internalists’ in the theory of mind try to track what is ‘really’ going on in the mind that is not observable. All ‘externalists’ have are interpretations of actions where the subject has no more natural authority over the correct interpretation than the observer (Dowding 2002, 2008a). In some circumstances the subject might be the best interpreter of her agency; in others the observer might be better placed. What matters is how much information the interpreters have to go on.

58 Power

I do not consider Bourdieu’s methodological comments on practice here, other than to suggest they allow room for the observer to interpret action structurally by using the world around agents in order to understand the agent’s behaviour. In that sense, the process seems hermeneutic and could be construed in either an internalist or externalist sense. One aspect of Bourdieu’s analysis important for considerations of power is his notion of ‘symbolic violence’. ‘Violence’ is a rather dark and powerful word in this context, and we might lighten it in English by replacing it with the term ‘force’. Some of the elements of the values agents see within roles are what Bourdieu calls ‘symbolic capital’. These are such things as prestige or honour that are conferred upon people. For Bourdieu, these elements of symbolic capital (or specific types of resources) are crucial sources of power. Any use of symbolic power that causes another person to change their actions is, according to Bourdieu, an exercise of symbolic force. So where someone acts according to duty or expectation, they are acting through symbolic force. A woman clothing herself modestly in order to comply with what she believes her family desires acts under the symbolic force of her family. So symbolic force is the imposition of ways of thinking about the world and one’s place it in. At one level, symbolic power seems to be a predicate of agents. In our example, the woman acts because of the symbolic force of her family – what she believes they want her to do as communicated by their actions (even though they do not explicitly tell her what to wear). However, power might also be seen as a predicate of structure, since the imposition of ways of thought seems to be the incorporation of structures (other desires and actions) into the thought and behaviour of the agent. It is embedded into our very ways of thinking. Anthony Giddens argues that all approaches to the agency–structure problem deal with it inadequately because they fail to transcend the ‘dualism’ of agency and structure. They give precedence to either one or the other when dealing with any aspect of society. He claims to replace this dualism with a ‘duality’ which manages to truly transcend the conflict. He asserts his approach is neither agent- nor structure-centred, because it does not merely recognize the interface between agency and structure but produces a social ontology where structure is both the medium through which action is made possible and is reproduced through social practice itself. He claims others see structure as merely constraining, but his notion is that it also enables action. His account he labels ‘structuration’, which is simply intended to convey the idea that social structures create social action even as social action creates those structures (Giddens 1979, 1984). For Giddens, structure is the rules and resources that enable people to act and interact. The modality is the way in which structure is



Agency and structure59

translated into action: that is, through people, whilst interaction constitutes the activity of people within the structure. He discusses three types of structures: ­signification (the meaning and interpretation we place on our behaviour); legitimation (the moral order of ethical norms, and so on); and domination (the exercise of power through the control of resources). The three work together as we place meaning upon some actions given how they are legitimized and underpinned by the control of resources. Power is central to Giddens’s account, since all social relations involve some kind of power, and this power encompasses both conflictual and, more importantly, consensual power as collective action. Whether Giddens’s approach to the agency–structure divide transcends it to any greater extent than others is a moot question (Archer 1984, 1995; Bryant and Jary 1996). It is not my intention to provide a full account and critique of anyone’s attempts to reconcile agency and structure, though a few comments on Giddens’s claims are appropriate for my argument. Giddens claims that his approach is the true ontology of society where both aspects – structure and agency – are brought together into one entity. In fact, he claims to bring three ‘dualisms’ into three ‘dualities’, namely the dualisms of (a) free will and determinism, (b) subject (people) and object (society) and (c) static analysis from dynamic analysis. Now, it might well be that the basic ontology of the universe is some seamless flow, but Giddens fails to explain how we explain that flow without breaking up the social world into manageable pieces. If we cannot distinguish the explicandum from the explicans or, in the language of statistics, the dependent from the independent variables, then we cannot manage to explain any outcome, event or institution, since we cannot distinguish it from that which is meant to explain it. Furthermore, whilst we all acknowledge that dynamic models are generally superior to static ones, it is often necessary to solve the static in order to get a handle on the dynamic. (And comparative statics can be very revealing.) In fact, I suggest that, despite his grand claims, Giddens does not offer much more of an analysis of the agency–structure ‘dualism’ than those who went before. For his illustrations of the dualism of free will and determinism, he examines the habits of routine action (constraint) and the opportunities allowed by structural facilitation. Not only do the illustrations not transcend the dualism, they provide no insight into when one occurs and the other does not (Archer 1984: 459); and they certainly do not provide any insights into free will and determinism.7 Similarly, Giddens’s claim that each action is part of the social structure is, in one sense, trivially true; but, in another, unenlightening if not false. We can define the structure of any society as the conglomeration of all actions within that society; but to contrast two different structures we

60 Power

need to examine the forms of them that are different – an egalitarian ethos versus a competitive one, for example. But one would not expect every action in the egalitarian society to be non-competitive, nor every actor in the competitive society never to display egalitarian tendencies. In other words, some aspects of structure are more important than others and they are more important for explanatory purposes. By concentrating upon the ‘flux of the universe’ Giddens loses sight of the fact that social theory has to be provided in the logical form of explanation. This is no more apparent in his claim for dynamic over static models. He seems unaware that even dynamic models have independent and dependent variables that need to be distinguished conceptually and (often) temporally. In reality, I claim, everything that makes sense in Giddens’s approach to the agent–structure question takes us no further than his predecessors. Yes, agents are implicated in structure and structure in agents, but in social explanation we prise them apart in order to be able to explain their relative influence on the outcomes, events or institutions in which we are interested. This failure to understand the basic logical form of explanation is particularly pertinent to the account of power. Power in one form enters as the ability of people to take the opportunities open to them, enabled by structure. However, this certainly predicates power to people, yet does not seem to allow structure enough to say about the relative capacities of people to take their opportunities cooperatively. Giddens’s account of domination does not here analyse the power inequalities that exist through the inequality of resources that most social structures engender in one form or another. Contrary to Giddens’s claims, structuration neither overcomes the dualism of structure and agency, nor provides greater explanatory power than earlier theories. All approaches to the structure–agency divide note that people act, but they act as constrained and enabled by others. Their views, interests, beliefs and desires are formed through their interactions with others, and with ideas learned from people in the past. We gain ideas from the past, modify them with others, form habitual patterns of behaviour, which we sometimes break out of. The roles in which we find ourselves, sometimes thrust upon us, at other times partly chosen, give us specific interests, direct our beliefs and actions which affect others, reverberating into the future, creating structural forces affecting others living now and in the future. These patterns of behaviour are not entirely accidental. We can explain many conventions by their enhancing and enforcing mutual interests. Quite why one society has one convention and another an entirely different one that solves the same collective action or coordination problem might be simply fortuitous. However, it is the mutual benefits of some sets of conventions over others that lead societies to converge on those sets. In the metaphor of



Agency and structure61

Parikh (2002), such conventions act as social software, enabling societies to run more smoothly. And humans can examine and rewrite that software, just as they can with computer software, to help society run more smoothly. And, as with computer software, their changes can unwittingly introduce bugs that cause breakdown under specific conditions. As part of that social software, people take on specific roles which further structure their interests, strategies and interactions with others. In all of these attempts to examine the interrelationship of structure and agency, theorists have taken very similar lines. The major differences are in language used – itself dependent upon these theorists’ backgrounds. Anthropologists tend to read different books from sociologists, sociologists tend to have a different background from political scientists, who in turn differ from economists. The tools each use differ also. The technical nature of game theory allows for a close analysis of the mutual benefits of conventions and show why equilibrium strategies are adopted even in areas of conflict. Sociologists and anthropologists might be more interested in the provenance of the preferences more often taken for granted by game theorists and economists. Whereas the former concentrate more upon mutual benefits, the latter query whether those benefits are so mutual, and are more prone to critically examine whether the conventions and equilibriums benefit all equally. Delving more into history, we can see that conventions often benefit one set of people over others, so entrenching privileged interests. Such historical accounts lead us to ask whether structured benefits in this manner are formed through power structures that inevitably favour some groups over others. Structure and agency are deeply entwined, but if some explanations seem to privilege structure over agency, then that is because for that explanation we look to the structure to explain the nature of the agency. If other explanations seem to privilege agency over structure, then that is because we look to an agent or set of agents to explain structural change. However, the fact that specific explanations seem to privilege one over the other does not mean we have to believe that the social theory from which that particular model or explanation was derived always privileges one over the other. Particularly so, as I have argued, because individualist models generally have structure built deeply into their conception of the agent (by social role) and structural explanations generally leave room for agent discretion. And whilst I have argued that power is best seen as a predicate of agents, that is only because it seems a more natural use of English. Agent-centred and structured accounts can be translated from one to other through the medium of agents’ resources, which are themselves structurally defined. Nevertheless, preferring the language of structure to agency or vice versa might reveal some value commitments, as I discuss in the next section.

62 Power

The valuational element of power It has long been argued that power is an essentially contested concept and any attempt to define it will necessarily reveal specific value commitments (Lukes 2005). As an empirical concept that could, in theory, be measured across sets of people or compared across societies, I believe that it need not be so contested. To be sure, there might be different accounts of power, but if they have empirical application they should be translatable one to another (Dowding 1991: ch. 8, 2006c). However, there is little doubt that the line one takes on agency and structure reveals something of one’s interests and values. I have tried to argue that the agency–structure divide has been bridged, though in specific explanations one or the other might appear privileged. The divide is no longer a puzzle; indeed, once the relationship between agency and structure is laid out, one wonders why it was ever considered to be so.8 Now that is not to deny that there are some minor conflicts. I have tied my own flag to the mast to say that power is best predicated (at least in English) to agents; nevertheless, I see the type of explanation that we offer in social science as structural as much as agential. This is so because the interests of social agents in social explanation are often defined almost exclusively in structural terms; and can be seen as proceeding from or caused by structures. I do not exclude structural accounts of power as such – they are not meaningless – but argue that they can be translated into my preferred form of power analysis without loss. In other words, when we predicate power to individuals, we can see the source of that power within structures; and further recognize the structural influence upon individuals’ interests. Similarly, when we predicate power to structures, we can still leave room for human agency. But why might we sometimes want to concentrate upon power as structural and sometimes as agential? Whilst these are not incompatible views, each does suggest different value commitments. Mainstream political science has almost left ‘power’ behind.9 When discussed, it is largely in terms of the ‘power of the president’ or ‘power of the prime minister’ or discussion of the relative power of countries (at the supranational level) or between organizations within a state or across states.10 The more structural or social issues that once transfixed the ‘community power debates’ have been left behind in network, regime or other similar analyses. It would be incorrect to suggest that issues of structural power are not implicated in these terms, often explicitly so (see Stone 1980, for example),11 nor that there is not a vibrant sub-community of those within political science whose central topic is power (I am one of them). But mainstream political science has left power studies behind (as Jordan and Richardson (1987) suggested, in order, perhaps to concentrate upon



Agency and structure63

topics more easily graspable). One reason might be that mainstream political science examines policy outcomes at a fairly detailed level. In order to explain those outcomes, it brings in all the resources and events that agents have to show how an institution or policy outcome formed. Both agents and institutions (or structure) play a part in this, but the concentration is upon the specific processes that lead to policy outcomes. Or where political science examines institutions as dependent variables, it does so again at a concrete level of how one specific institutional form was chosen rather than another. Given that mainstream political science tends to view power as a predicate of agents, and agential power is essentially about what agents could do, and not simply about what they actually do, power does not become a focus of study or a term within the analysis. Analysing the power structure at a given point in time is to analyse a game form; not the game itself. That is, it is to examine the possible strategies and outcomes, and not those that are deployed given the specific interests in play at any one moment (Dowding 2008b). Thus the agential and particularistic focus takes us away from explicit power considerations. Or, to put the point another way: once one has analysed an outcome through the interplay of agential resources, there is no need to mention power at all. It has become redundant; it has been analysed away (Moe 2005; Dowding 2008b). Thus positive or empirical analysis tends to analyse situations without mentioning power at all. Power is still normatively important, in terms of what could have been or what should be the case. We can say precisely the same for economics – in fact here power is even further analysed away (see Dowding 2008b for further argument on this issue). Anthropology and sociology, on the other hand, are still involved in broader accounts of power. Sociology has less of a focus upon specific outcomes (and where it does, what I have said about political science tends to apply). These more general concerns of the effects of culture, socialization, socio-economic class, and so on upon behaviour examine the structural side of the agency–structure relationship. The questions posed in sociology are more likely to be concerned with the effects of social processes on human behaviour, with the agency as the dependent variable and the structures the independent variable. Thus these studies tend to concentrate more upon the structural side. Dividing the nature of the social enquiry into ‘political science/­ economics’ and ‘anthropology/sociology’ as I have done might be misleading. The analytic point is that where the dependent variable is an event such as a policy outcome or an institution, then the independent variables will include agents as well as structural factors. Where the dependent variable is behaviour of a set or class of people, then the independent variables

64 Power

will be largely composed of structural variables. We thus see the nature of the question posed directing the concentration upon one or other side of the agency–structure divide. However, the fact that interest is framed in this manner does not entail that researchers cannot recognize the truths contained in seeing structure and agency as a duality or structuration bridging the agency–structure divide. Both agential and structural variables are important in social explanation. Too often, when theoretical debates become embroiled in agency–structure debate, writers choose examples from one general tradition rather than another to illustrate their argument, thus making the division appear greater than perhaps it is. Conclusion I argue here that the ways in which most individualists conduct research leads them to see agents as being highly structured and that most structuralists leave some room for human agency. I suggest that many authors have ‘solved’ the agent–structure problem in very similar ways, and that the differences between these solutions are more rhetorical than analytical. Minor differences need not distract us from the fact that the problem is solved. That is not to say that there are not real differences between accounts of power that are tied to individuals and those that are tied to structure. I have reiterated my claim that ‘power’ is best seen as a predicate of individuals, since that is its more natural form in the English language. Accounts rendered in a structural form can be translated into an agency form without loss, I maintain; and a similar translation the other way is possible. However, that is not to say that the choice to concentrate upon the structural over the agent-based (or vice versa) conception of power is valueless. The specific research question that one is posing might naturally lead to a concentration upon either structure or agency. In that, rather weak, sense there is an evaluative element to seeing power in agency or structural terms. The translatability of one form to the other, however, suggests that this does not entail any essential contestability in the concept of power. There is an important caveat to my argument about the translatability of agent-based and structure-based accounts of power. I distinguished between a structure seen as merely affecting the strategies of agents given their interests and structure affecting the very nature of those interests. The latter form of structuralism I described as ‘deep structuralism’. There can be shallower and deeper forms of structuralism. One that leaves no room for agent autonomy is the deepest form. I suggested that structural accounts are deeper when structure is seen to affect action not simply through the strategies that actors develop in order to achieve their aims,



Agency and structure65

but also when it affects those aims themselves. Structure could affect those aims at an ever deeper level. Explaining an individual’s action using the intentional stance (Dennett 1987), we give the agent’s ‘reasons for action’. Indeed, that is how we interpret our own actions. Now, our reasons for action might, to a greater or lesser extent, be formed through customs, habits, ethical rules, and so on. These elements structure our reasons for action. If we fully explain a person’s reasons for action in terms of these structures, then we leave no room for agent autonomy or free will. Even when there appears to be an opportunity for choice, the choice itself will be determined by the relative structural weighting in the person’s mind. It has been my argument that, in fact, few if any writers have explained structure in quite this deep sense: ‘deep’, because the structure goes all the way into the mind of the subject. One reason, perhaps, that structure is not normally defended as providing all the explanation is that it is not clear how it can provide agency – how it can cause agents to act, down to the smallest components of their actions. It is hard to see how we can use structure in explanations of individual human action that are fine-grained enough to predict every aspect of a person’s actions. This fact suggests that even if all of our actions are fully determined by the interaction of our genes and our environment – there is no free will or autonomy as classically understood – then we would still need to leave room for individualized ‘reasons for action’ where we cannot perceive how they are structurally determined. It is for this reason, I suspect, that even the deepest of structuralists wants to leave room for human agency. Not necessarily because, ontologically speaking, there is such agency – but simply because we do not have the tools to completely specify structural explanations of human behaviour down to the smallest detail. Giving agents’ ‘reasons for action’ is the best explanation, and provides the most efficient predictions (Dennett 1987, 2003). The fact that we need to leave such room for human agency at the individualized level does not mean that we cannot adequately explain the large differences between the types of outcomes we see at the societal level purely in terms of structural features of societies. Many properties of societies can be expressed only at the societal level: the degree of inequality, the class nature of social power, the family- or network-oriented nature of political power, and so on are all social properties. They affect the amount of power, freedom and welfare of each individual in society, to be sure, but we do not explain those individual properties by simply agglomerating them. Rather, we do so by categorizing them in a structured manner. Our interpretation of power relationships will largely be determined by social categories as differentially structured in assorted societies. Nevertheless, we leave room for biological individuals to have different qualities, and thus power resources,

66 Power

even though they are members of specific classes or sets of people we have categorically structured. There is no agency–structure puzzle here: simply different types of questions that need to be addressed using different tools and focus. Notes  1 I make no distinction between social and political power in the following. Whether or not one considers political power to be a subset of or identical to social power depends upon the breadth of one’s definition of what constitutes ‘the political’. I use the term ‘power’ here in the social sense.  2 And some will further argue that Arrow’s theorem shows that there is no sense in which a collective being has a collective will (see Riker 1982). For a modern view on how a collective being can have a common will (despite Arrow 1951/63), see Pettit (1993a).  3 ‘Natural’ disasters might have human causes, of course. Social structures might engender famine, for example (Drèze and Sen 1999). Floods might be natural disasters because of where people have chosen to locate or government decisions about flood defences. My point, however, is to the extent that we can locate an event with important social consequences beyond human action, it is, according to the individualist, exogenous to social explanation. To the extent that it can be located within human action, then the methodological individualist can point to individual action as the locus of social explanation.  4 I argue in Chapter 7 that how we identify and measure ‘responsibility’ is directly related to how we measure luck. If we were to remove all luck from the determination of character, we would also remove all responsibility. See pp. 146–8 below.  5 This is so even for physical aspects of phenotype. Newborn rats are blind and, without a source of light to ‘switch on’ the relevant genes, they remain blind.  6 Systematic attempts to examine how individuals can dramatically change political and social situations often rely not only upon the skill and perception of the agents, but also the opportunities they grasp at a crucial moment in history. Riker (1982, 1986) calls such agents ‘herestheticians’. In his account, heresthetic politicians forge new coalitions of interests to create different political cleavages, and then redirect the political or social situation on to a new path. But heresthetic politicians can only utilize the opportunities provided by such moments. Similarly, the application of ‘punctuated equilibrium’ to policy formation suggests that periods of stability where policy problems partly caused by the institutionalized solutions to old problems lead to ‘crisis’, which enables more radical policy development. Equilibrium is punctuated by rapid change (Baumgartner and Jones 1993; Jones et al. 1998, 2003).  7 I say above that an outcome might be fully determined in the sense that a person can do nothing about stopping something happening, but that in itself



 8

 9

10 11

Agency and structure67 does not bear on the free will issue. If I cannot break a habit, then perhaps I have no autonomy; but if I can choose or not to take opportunities, perhaps I do. The puzzle might be: is there free will? Or, in another form, what does agent autonomy mean? But the interrelationship between agents and their environment, including the social environment of past and present actions of others and how this affects an agent’s beliefs, desires and actions, is not a puzzle. In other words, we might not have a clear way of deciding between deep structure and any approach which combines agency and structures, but we should have no problems in understanding the latter combinations – even if they are not all identical. Jeffrey Isaac, who has published one of the best ‘realist’ accounts of power (Isaac 1987), suggested at a panel on Dahl’s concept of power fifty years on, that ‘power studies’ were now irrelevant to political science (Isaac 2006). His claim was certainly hyperbolic, but does contain some truth. Or, at a more detailed level, in terms of the ‘voting power’ of such bodies. It is also the case that power is still discussed more explicitly in urban studies than at the national and supranational level.

3 Rational choice and community power structures Keith Dowding, Patrick Dunleavy, Desmond King and Helen Margetts

The community power debate had a focused interest in discovering who the major power holders were, and how far power was dispersed, in different communities (Hunter 1953; Dahl 1961; Bachrach and Baratz 1970; Polsby 1980; Domhoff 1983). Political theorists, rightly criticizing the simple-minded use of power, formulated a conception of power that moved ever further away from empirical demonstration. Consequently, urban scholars became less interested in studying the power of observable power holders (Foucault 1980; Peterson 1981; Connolly 1983; Stone 1989a; Lukes 2005). Enlarging the scope of power has exculpated observable power holders from responsibility for their actions, and pushed the state into the analytic background. It is time to return power to centre stage in urban analysis. Institutional rational choice analysis enables us to understand the problems of the first and second dimensions of power, which initially triggered dissatisfaction, without disappearing into a metaphysical wilderness. Power and social explanation A fundamental cause of the sterility of the classical community power debates was ignorance of the problems of collective action.1 This can be illustrated in three areas. Decisional studies Decisional studies were flawed because most pluralist writers did not understand that studying human action without independently verifying actors’ beliefs or desires prevents the analyst understanding their interests. Consider the behaviour of a group of actors choosing a course of action D, causing social outcome P. If each had followed course of action C, social outcome R would have resulted. Can we conclude from their choice D that



Rational choice and community power structures 69

Aj Cooperate Defect

Cooperate R,R

S,T

Ai

Defect T,S P,P 3.1  Simple Prisoners’ Dilemma

they prefer P to R? Such a conclusion is premature, for we need to understand something of their beliefs and desires to comprehend the structure of their decision situation. Consider the matrix in Figure 3.1. If the preferences of the two actors i and j are {T>R>P>S}, then they are in a Prisoners’ Dilemma. Here we would expect the actors (in a pure oneshot PD) to choose to defect and end up with outcome P, even though both prefer R to P, attainable had they both chosen cooperation. Alternatively, if they ordered the preferences {T>R>S>P} or {P>T>R>S}, then they would also both choose to defect. In this case the decisionists’ assumption that lack of mobilization on certain issues demonstrated a lack of interest would be true. But this needs to be demonstrated by independent assessment of the beliefs, desires and, therefore, interests of the groups under analysis. The decisionist studies of community power ignored the collective action problem and therefore were unable to demonstrate the accuracy of their conclusions about the structure of power in the communities they studied. One must determine the structure of the choice situation in order to fully appreciate the nature of decisions that are being made. The structure is the interrelations of the social actors, their preferences and their resources. We have to model this situation, and in doing so may need to begin by making assumptions about beliefs and interests. Later, these assumptions are modified as more is learned about the individuals or groups concerned. The blame fallacy The blame fallacy was committed by neo-elitist writers in criticizing pluralist research. This mistake was the opposite of the pluralist misreading of collective action situations. The fact that individuals or groups may not reveal their interests through action and not collectively act to promote their interests does not mean that another group or person is acting against them. In Prisoners’ Dilemma and Chicken games it is the structure of the situation which often leads to mobilization difficulties, and not the actions of other powerful groups. The blame fallacy occurred because the critics of

70 Power

decisional studies did not understand the distinction between two different sorts of power – outcome power and social power. ‘Outcome power’ is the ability of an actor to bring about, or help to bring about, outcomes. ‘Social power’ is the ability of an actor deliberately to change the incentive structure of another actor or actors to bring about, or help bring about outcomes. Groups may lack outcome power without other social actors wielding social power. Of course, the structures of power in society have often been created to increase mobilization costs. In such a case social power has been wielded in order to reduce the outcome power of others. When groups do mobilize, other groups with contrary interest often mobilize in response: indeed, this mobilization pattern can be mapped through time (Walker 1983). The blame fallacy is inherent in the expanded concept of power associated with Lukes and Foucault, who characteristically see social arrangements as suffused by power. For example, summarizing Foucault’s view, Philp (1992: 159) argues that ‘Struggle is always necessary to avoid domination, yet it cannot guarantee liberation since power is an inherent feature of social relations – we cannot act without affecting the conditions under which others act’. The problem for such a shorthand is exactly the implicit claim that every way in which we affect the conditions under which others act is a social power relation, rather than an interaction or interdependency effect. Power and luck Power and luck are linked concepts, but not in the one-to-one way that neo-elitist writers suggested when they elevated the question ‘who benefits?’ into a critical test for ascribing power. The fact that groups get the outcomes they desire does not demonstrate that they have power, let alone that they are wielding it. Groups may get the outcomes they desire merely through being ‘lucky’ (Barry 1989). ‘Luck’ here means ‘getting what you want without trying’, not random chance. Modern egalitarians use the term ‘advantage’ to mean (roughly) ‘the opportunity for welfare’ (Cohen 1989). Groups may be ‘advantaged’ (lucky) in the sense that, through no deliberate action on their part, they are in a position to gain benefits or welfare for themselves. By ‘luck’ we mean a situation where one person or group benefits through the actions of another group or person A, or their interests are served by the actions of A, but not because A (primarily) intended to achieve this effect. In this sense, free-riders are lucky if the collective good they desire is provided by others. The clear danger in reading back from ‘who benefits?’ to power is that we may misascribe power. A great variety of social groups and individuals may get what they want without trying, because others will generate outcomes meeting

72 Power

Lucky   Not lucky Powerful 1 2 Not powerful

3

4

3.2  Luck and power

Careful examination of the interrelationships between power and luck should allow power ascriptions to be made in a more authoritative and grounded way than in earlier studies. Consider the simplified matrix in Figure 3.2. On any given issue at one time, we should be able to locate any given group or actor in terms of these combinations. An elitist position is that there are groups made up of a minority of people located in boxes 1 and 2 (usually in box 1) over the vast majority of issues, or those which are most important. Meanwhile, groups covering the majority of people fall into boxes 3 and 4. There is considerable common interest and so the majority of people get what they want most of the time, placing them in box 3. But when they do not get what they want without trying, they are not powerful, and so end up in box 4 on those issues. Demonstrating that large groups of people have their interests served does not in itself show any degree of power equality or pluralism. To test for the existence of either pluralism or elitism, it is necessary to concentrate upon the issues over which groups end up in boxes 1 or 4. A pluralist must maintain that there is no systematic bias which leads some groups to end up in box 4 and others in box 1. If this account seems too individualist and conjunctural, we should add that in our view, rational choice entails neither of these propositions. Systematic inequalities of power will be based upon resources which need to be determined for all actors, and will often themselves depend upon social location. Modelling the power relationship between social actors as a bargaining game means only accepting that the resultant outcome depends upon the actors’ bargaining resources and the strategies they adopt, given the structure of the game in which they are involved. It is consistent with this approach to look at five main types of resources. First, the information set of the actors. Second, their legitimate authority. Third and fourth, the unconditional and conditional incentives actors are able to create in others’ structural conditions. Unconditional incentives occur when one actor changes the decision structure of others by making changes in the environment. Conditional incentives occur when one actor threatens or makes an offer to another in order to bring about behavioural changes. Material resources generally underlie both conditional and unconditional incentives. In all four cases, empirical examination may



Rational choice and community power structures 73

reveal the resources of actors. Discovering the resources of actors does not demonstrate that their potential powers are utilized, but it does allow modelling such potentials as bargaining games. The fifth and final resource for actors is their ‘reputation’ (see below). Against the criticism that our approach is too conjunctural, consider its implications for the best-known previous applications of rational choice methods to political power: the literature on power indices (Riker 1969; Shapley and Shubik 1969). These indices concentrate on measuring the ‘pivotality’ of voters in a committee or other voting forum – the abstract ability of voters under all possible combinations of preferences to get what they want. But they take no account of the distinctions between power and luck. The index is the probability that a voter is pivotal under all possible voting combinations. But any given instance that one is pivotal depends as much upon luck as power: Happening to be pivotal or being lucky that one is pivotal brings power, but these circumstances are not in themselves a form of power … The greater power of pivotal voters, for example small parties in hung parliaments, comes about not because they happen to vote one way or another but because their actions change because of the realization that they are pivotal. (Dowding 1991)

Against the claim that a rational choice approach is inherently individualist, and hence cannot take account of structural influences on the distribution of social power and outcome power, luck is likely to be consistently patterned. Luck is systematic if it attaches to certain locations within the institutional and social structure and not to others. This form of luck is non-random, and may be methodically predicted if the structures of society can be formulated. In a democracy under capitalist relations of production, capitalists are systematically lucky, since economic growth – which is positively correlated with re-election – is in the interests of the government. Policies which promote growth are therefore likely to be pursued by government whether or not capitalists wield social power over government. This does not entail the conclusion that capitalists are not powerful as well as lucky, nor that at times they deploy their power, merely that they do not need to wield that power all the time. It also does not follow that merely because economic growth policies are in the interests of capitalists, they are not also in the interests of others. Even recent elitist accounts of community power equate the fact that local political elites pursue growth with the power of growth coalitions (Molotch 1976, 1979; Domhoff 1983, 1986; Logan and Molotch 1984, 1987). Growth coalitions are often lucky, and when they do wield outcome power, they may not be acting against the interests of the community as perceived by that community.

74 Power

Identifying the power structure therefore requires consideration of many factors. A social actor may get desired outcomes because of luck, systematic luck, outcome power or social power. Serious problems arise if analysts search for the power of groups which are, in fact, systematically lucky. This equation of power with benefit is understandable but erroneous, leading radicals to seek non-existent evidence. Their inability to empirically discover the logically undiscoverable led some to scorn empirical work altogether. A social actor may fail to secure desired outcomes because of a lack of power, because she has the power but has not utilized it, or has chosen not to, or could not mobilize the resources. In this last case, attempting to find a group wielding social power over a set of unmobilized people will be the search for a chimera. Such searches led to the study of non-events and non-decisions which, not surprisingly, were never found (Bachrach and Baratz 1970; Lukes 1974; Gaventa 1980). Our account emphasizes that luck, systematic luck, outcome power and social power can all be empirically investigated. To do so requires assumptions about individual and group interests, developing independent means of verifying those assumptions and identifying preferences. The next stage is to generate models of the policy process entailing predictions of the expected distribution of power and luck given the resource holdings of groups and the unintended consequences of social action. These models will develop the networks of interactions between elites and their interaction with other organized and unorganized groups. Power and social action What is the relationship between power and the ability to solve collective action problems? The key developments in this argument can be represented by Figure 3.3. The community power debate was mostly premised upon a Weberian background conception of power as inherently conflictual, well illustrated in the definitions used by Dahl (1957). Power was equated with social Conflict

POWER

Possible actions

Variable sum

ENVIRONMENT

3.3  Relationship between power and ability



Rational choice and community power structures 75

power, and the distinction between outcome power and social power was little appreciated. Second- and third-face accounts of power conceptualized power struggles as zero-sum games. By contrast, Parsonian pluralist sociology viewed power as a medium for securing social action, akin to money or economic exchange. In this view the application of power is generally not associated with conflicts, for it enables society to resolve otherwise intractable collective action problems. Echoes of this position occur at points in the community power debate, notably in Banfield’s (1961) study of Chicago. He concludes that interest group pluralism combined with a concentration of institutional power in the hands of the city mayor and state governor combines a ‘social choice mechanism’ and central decision making in a mixed decision choice system approximating the logically preferable outcome. More recently, Niklaus Luhmann (1979) has restated the same view of power relations as a positive-sum game: power in his view is like trust, an essential part of the fabric of stable social relations. It is extensively separated into distinct, complex social subsystems which are ‘autopoietic’, ending on internal lines of development and interacting with each other only in constrained ways. From a rational choice perspective, both the conflictual and the actionmedium conceptions of power are over-polarized and inadequate. Social relations involving power, much as bargaining games between actors, cannot be argued a priori to have either zero-sum or positive-sum outcomes across the board. Instead, they are variable sum, depending on many detailed situational effects, in particular the distribution of power and luck and the structure of the decision situation. Some results produced by the exercise of outcome and social power will be positive sum, some zero sum and others negative sum. Negative-sum outcomes figured hardly at all in previous debates, except that their avoidance was dimly re-expressed as a positive-sum use of power. Methods in power research Elaborating theory without also improving the methods of empirical research leads to ungrounded and over-theorized analysis. A key contribution of institutional rational choice theory has been to develop new ways of investigating issues in urban power research (Peterson 1981; Gurr and King 1987; Schneider 1989). We discuss three new developments: the rehabilitation of reputational power; the convergence between public choice and substantive conceptions of socialized provision offered in more radical approaches; and the possibilities inherent in rational choice methods to open up ‘regime analysis’ for understanding urban politics.

76 Power

The rehabilitation of reputational power No concept was apparently more discredited in the old community power debates. Pluralists waxed lyrical in their scorn for the early elite studies’ methods. As used by Hunter, the reputational method involved compiling a list of potential influentials, some of whom were then asked to rank each other’s ability to influence the outcome on a series of hypothetical issues (Hunter 1953). A smaller subset was then interviewed, continuing the ranking process. The characteristics of the final nominated subset (usually around 20 people) were then described. Pluralists objected that the whole process proceeded on a ‘what if …’ basis, and the fact that interviewees gave some response in the terms put to them was no indication that the scenarios envisaged were real: When an interviewer asks ‘who are the top power leaders in this community?’ he [sic] suggests by his question that there are top power leaders. Probably few respondents are thoughtful enough or cantankerous enough, to challenge the hidden premises of such questions. (Banfield 1961: 8)

Similarly, the fact that respondents agreed with each other on power rankings was irrelevant: ‘agreement may mean only that the informants share a common mythology’ (Banfield 1961: 8). Other critics emphasized the arbitrariness of the filtering process, the number of ‘influentials’ for whom nominations were sought, and the size of the subsets distinguished for more intensive investigation or further reduction in the scope of the study. Seen from a rational choice perspective, these pluralist criticisms seem exaggerated and misplaced. The behavioural fervour which informed decisional studies’ contempt for mere reputations has receded. In a wide range of rational-choice-influenced work, ‘reputations’ are identified as a key power resource for actors engaged in strategic interactions with others. Reputation is a central concept in modern bargaining theory (Roberts 1985; Sutton 1986; Kreps 1990a); in explaining the power and actions of large companies, typified by the Chain Store Paradox (Selton 1976); and it is now a standard variable in the economic approach (Rasmusen 1989; Milgrom and Roberts 1992; Ordeshook 1992). Whereas it was once argued that reputation hardly constituted ‘scientific’ evaluation of power, now we can recognize the concept as one of the most important elements in developing a bargaining strategy where knowledge is asymmetrically distributed and the players in a game lack perfect information. Rational choice theory has demonstrated the importance of players cultivating reputations for stubbornness, not backing down in conflicts and not responding to threats/offers from others; and in collaborative situations developing a reputation for cooperative behaviour (Axelrod 1984; Raub and Weesie



Rational choice and community power structures 77

1990). Thus Hunter’s conception of reputation not only reveals what others think about given individuals’ power resources, but is itself a source of those individuals’ power. The identification of reputation is also much easier than in Hunter’s day, due to changes in research technologies. The explosion of text databases promises a revolution in the methods of the previously ‘soft’ social sciences, paralleling the advent of mass survey research in political sociology after the Second World War. Where print media sources are stored on databases (as are all the ‘quality’ daily newspapers in the UK, along with the London evening newspaper), it is possible to extract an objectively defined set of reputedly influential policymakers in any functional area, or across all areas. The scale of any actor’s reputation can be mapped quantitatively in the scope and character of press coverage. Of course, this coverage does not constitute the full measure of each actor’s reputation, and Hunter-style analysis can be generated through semi-structured interviews in order to discover important actors screened from media purview. Nevertheless the media have an extensive coverage of certain types of issue, and the fact of being reported brings a certain reputation. This reputation is itself a power resource. The agent doing the filtering here is no longer the researcher but the mass media system as a whole. It is difficult to envisage any credible political scientist arguing that media reputations for influence cannot constitute important power resources in their own right. This kind of research permits a more comprehensive investigation of agenda setting and the mobilization of bias against other social groups and actors. The issues involved in establishing actors’ reputations are no longer hypothetical scenarios, but extant policy conflicts. Once a sample of policy influentials is obtained, their interrelationships can be determined by cross-checking media references, by surveying them and by undertaking a network analysis on a scale and with a speed which would have been inconceivable without the aid of modern relational databases. The wheel has turned full circle on ‘reputational power’: it is theoretically legitimate and important. Socialized provision Socialized provision has been an important focus in rational choice theory since Samuelson first promoted ‘public goods’ as a foundational concept for welfare economics. The public–private goods boundary has been variously drawn since then in terms of indivisibility, non-excludability, nonrivalness and coercive consumption. But the empirical articulation of the distinctions between fully private, quasi-public and fully public goods has been vestigial. Students of externality effects have also made clear that what count as pure public goods (or positive externalities in general) for

78 Power

one group or subset of society are often ‘public bads’ for other groups (Hirschman 1970). Although economists have continued to deploy the ‘public goods’ notion in analyses of public choice processes, political scientists have generally concluded that ‘Very few of the goals or goods that [interest] groups seek can accurately be described as pure public goods … [they] need to be collective only in the sense that they are collectively provided’ (Hardin 1982: 19) Perhaps the best way of articulating a position without advocating a particular approach to collective goods is to revitalize the distinctions between production, provision and consumption (Ostrom et al. 1961). Some goods, under current technologies, are most efficiently provided by the state, regardless of their status on the collective–private good continuum. This says nothing about efficient production, which may be either public or private for such goods, nor about the possibilities of private provision under different technologies. Many commentators do not seem to understand that technical innovation or changed economic circumstances can make private provision economically profitable where previously only public provision was feasible (Malkin and Wildavsky 1991). Not only is the public–private good distinction ‘fuzzy’, it is also space- and time-specific. Applied work has demonstrated that most government policies produce goods which are not jointly consumed but perfectly excludable (‘toll goods’) or are private goods which society for some reason or other wishes to subsidize (‘worthy goods’) (Savas 1987). Neo-classical economists have traditionally neglected the substantive nature of goods; to do more would involve them in searching for the origins of preferences. The assumption of fixed preferences has been firmly maintained to preserve economics as a set of perfectly generic models, and from a conviction that the division of labour between disciplines legitimates an arbitrary cut-off point at some stage. So mainstream economists have no theory to explain which goods or services succeed in reaching ‘worthy’ status and which do not. From the viewpoint of institutional public choice (whose push is towards endogenizing preference formation within the political process), this position is intellectually unsatisfactory. More substantive accounts of the origins of ‘worthy’ goods have been offered by radical urban theories, especially in Castells’s emphasis on the specificity of ‘collective consumption’ processes (Castells 1980; Otnes 1988). Stripped of its neo-Marxist premises, updated to allow for the impacts of privatization and fitted into a production–provision– consumption model, Castells’s position can be rephrased to mean the claim that urban politics revolves around those kinds of final appropriations of products by people (consumption processes) where the state intervenes in some way, either to socialize some of the costs of provision or to organize



Rational choice and community power structures 79

provision via governmental means. Socialized consumption thus includes full collective provision, services which are both organized by government and subsidized (for example, state schooling, public health care, highways, urban planning and public housing in some countries); quasi-collective consumption, services partly privately and partly collectively provided, that is, either organized by government or subsidized by the state, but not both (for example, cultural provision, municipal transport, social housing); and quasi-individualized consumption, where the state subsidizes (a form of collective provision) the acquisition of privately owned goods (for example, homeownership subsidies and tax breaks for company cars in some countries). At a certain stage of industrialization, the logic of capitalist urban development pushes the state to prioritize these ways of expanding consumption capacity – because it socializes costs of production, reduces labour costs, stimulates complementary consumption and aids in legitimating capitalist social relations. At the same time, in capitalist societies there are strong pressures to limit the scope of non-market interventions, to preserve traditional forms of socializing provision, to foster market ideologies, and to avoid the fiscal overloading of the state. These conflicting tendencies dominated the post-war development of advanced industrial states. But similar state interventions also underlie the rapid economic growth of some newly industrialized states such as Hong Kong and Singapore (Castells et al. 1990). In advanced capitalism, state interventions to underpin particular forms of consumption still cluster around these ‘urban’ issues for structural reasons, and not because of some kind of collective psychosis or whimsy, as new right views seem to maintain. Linking the public choice analysis of public goods/collective action problems with the analysis of socialized consumption in radical urban theory presents critical opportunities to explore the changing social designation of worthy goods. Three levels of influence can be examined: structural pressures for state intervention; the play of power around the public–private interface (in the form of sectoral conflicts); and the dynamics of collective action problems themselves. Regime theory Regime theory, together with analyses concentrating attention on urban growth machines, has begun to dominate urban political studies (Elkin 1987; Logan and Molotch 1987; Stone 1989a, 1989b). Regime theorists construct a structural analysis which they believe goes against the grain of the individualist methods of behaviouralism and rational choice. Stone writes: ‘Structural constraints are real, but they are mediated through the political arrangements that enable a prevailing coalition to govern a

80 Power

community … settings vary and what community actors make of these settings varies as understandings change from one time and governing coalition to another’ (1987: 16). He thinks political analysis is about the ends which actors have in the roles they play within the fabric of society. Rational choice methods are an ideal tool for examining these roles and the dynamics of change across and within coalitions. Where Peterson (1981) argues that the type of policy causes the type of participation, Stone believes both that type of regime shapes policy and that type of policy shapes the preferences of the actors in the regime. These two positions may seem to be irreconcilable as the two sides of the structure–individual divide in social explanation. However, both elements can be modelled by the techniques described above. Rational choice, as we have emphasized, is a structural as much as an individualist method. Stone gives the example of the Atlanta business elite, driving a white administration unable to expand the city boundaries to secure a white majority coalition. To maintain their interests within the city, they learned to cooperate with black political leaders. On the assumption that whilst political leaders do not have a preference for working with black political leaders, Stone claims that ‘Regime analysis instructs us that policy innovation is not about individuals and their preferences’ (1989a: 160). However, rarely do personal preferences enter into rational choice modelling of political processes. Rational choice looks at the roles that people play in society; the game here is business–political relations, where preferences are driven by the needs of business. Regime theory was developed to explain how ruling coalitions of different political complexion pursue some policies which are similar. In an early attempt to explain this phenomenon in urban politics, Stone created the concept of systemic power (Stone 1980). This concept was intended to bridge the pluralist–elitist gap, by explaining how formal political equality and inequality through social stratification modified electoral coalitions once they became governing coalitions. The systems of power created incentives for governors to change their behaviour in order to achieve some of their aims. The ‘system’ can be modelled by the relationship between actors denoted by their preferences and resources. It is not the system that has power, but the actors bargaining to achieve their aims as completely as possible (Dowding 1991). Stone’s creation of regime politics correctly shifts analysis towards individual actors, whilst recognizing that the roles which actors adopt are what determine the interests and resources which lead them to bargain. Our analysis suggests that we should concentrate upon these interests and resources and upon recognizing that our models denote the structures through which these actors operate.



Rational choice and community power structures 81

Conclusion Writing more than two decades ago, Brian Barry commented on how disappointing ‘power’ research has tended to be in expanding our predictive capacity: We tend to feel that if we could understand the distribution of power and the lines and amounts of conflict … we would have the key to the explanation of political events. But if we had an analysis that was sufficient to enable us to talk confidently about the distribution of power we would already be in a position to explain political events. It would not be that because we could talk about power we could explain political events, but rather that because we could explain political events we could talk about power. (1974: 189)

We have shown why the cause of Barry’s unease lies rooted in the difficulties of determining interests, distinguishing power effects from situational factors, separating the impacts of power and luck, and ascribing power in social or outcome power terms. There are no shortcuts in power research, and those which have been taken over the years have led only to ungrounded or non-authoritative attributions of power. We have also demonstrated the need for power research to be refocused on medium-range theory, retreating from conceptions of power and contexts for power research which are inherently unempirical, whilst still remaining sensitive to the important ways in which power resources and the social incidence of luck are systematically structured. By illuminating the theoretical issues involved, by expanding the range of methods applicable to power research, and by offering new insights into the empirical transformation of power structures, rational choice theory can help create a new research agenda in urban politics. Notes 1 This section draws principally on Dowding (1991), where its core arguments are fully explicated. 2 Westminster’s poll tax was £195 in the financial year 1990–1, and £36 in each of the following two financial years; Wandsworth’s was £149 in 1990–1, and zero in the two subsequent financial years.

4 Power, capability and ableness: the fallacy of the vehicle fallacy

Power Power is surely one of the most important concepts in political science and political theory. Political science might be seen as the attempt to understand how political institutions work to produce the outcomes they do. Part of that understanding must involve how actors use those institutions in order to produce the outcomes they judge best. If our most basic understanding of power is Weberian – an actor’s capacity to get what she wants – then seeing how actors use institutions to get what they want is to analyse their power. Political philosophy can be seen as the attempt to design institutions that produce, overall, the best outcomes: to maximize our power as a community and to distribute powers to individuals in ways which seem normatively appropriate. Strangely, perhaps, individual power is not so often viewed as the central concept in discussions of social justice. Debate over what the central distributive concept should be usually centres around ‘utility’, ‘freedom’, ‘resources’, ‘capabilities’ or ‘opportunities for advantage’, to name but a few. One reason power is not often seen as a central normative concept is surely that there seems something wrong with thinking that individuals want, as a matter of justice, power. They may want many things we think appropriate, such as happiness, freedom or material wealth (though we might be wary of them desiring too much of the last), but not power. That smacks too much of an unhealthy ego. Power should be useful to get things that are appropriate, but it is not something to be desired itself. One of the candidates listed above is Sen’s suggestion for the appropriate candidate for equal distribution – capabilities. Capabilities are the abilities to achieve ‘functionings’ or achievements: ‘doings and beings’ (for example, Sen 1987a, 1987b, 1993: 12). A capability for Sen is the actual ability of someone to achieve outcomes (Dowding 2006c). In that sense, individual capabilities can be seen as individual powers – though, as Peter Morriss



Power, capability and ableness83

(2002: xxiii) points out, Sen rarely if ever uses the term ‘power’. Morriss also complains that Sen does not use the ordinary (dictionary) meaning of capability. Capabilities are latent abilities – abilities that a person could develop – not abilities they already have. Morriss further suggests that Sen should have used the term ‘power’, ‘ability’ or ‘ableness’, an archaic word that Morriss rehabilitated in the first edition of his Power: A Philosophical Analysis. Morriss seems to think that what we should be trying to distribute according to some just-distribution rule are ablenesses – a type of power – which are equivalent to Sen’s capabilities. This type of power cannot be further analysed. It is not reducible to anything else. Here I shall concentrate upon Morriss’s significant discussion of the vehicle fallacy and power-as-ableness in order to reflect how we might view power or capabilities as an object of justice. In part this question is one of measurement. If we believe that power or capabilities should be distributed in some manner – equally, for example – as a matter of justice, then how we measure that concept is as important as what it is. Pertinent here are two issues raised by Morriss’s analysis. First, is ableness reducible to anything else? Second, is ableness the important sense of power, as Morriss claims? The first argument is the point of Morriss’s account of the vehicle fallacy. I will argue that we can analyse power as ableness by reducing it to agents’ resources. Morriss thinks not and labels this the vehicle fallacy. To the extent that power as ableness can be reduced to agents’ resources reflects back upon Sen’s capabilities as an alternative to Dworkin’s (2000) resourcist account of egalitarian justice. I will then argue that whilst power-as-ableness is a useful concept, it is not the most important type of power. Power-as-ability is the underlying concept; power-as-ableness is what one can do with one’s power given one’s luck. This is not to deny that power-as-ableness is important for certain types of questions, but it is a mistake to think that power-as-ableness is the basic category. In that sense I also distinguish power and capabilities. The vehicle fallacy Peter Morriss’s important book Power: A Philosophical Analysis was r­ e-issued in 2002 (Morriss 1987, 2002); he introduces the concepts of the exercise and vehicle fallacies into modern debates about the nature of power. He bases them on an argument of Anthony Kenny’s.1 In the introduction to the second edition, Morriss says that many people have followed him in abhorring these fallacies and ‘nobody (to my knowledge) has tried to argue that these are not fallacious’ (2002: xiii). I have no quibble with the exercise fallacy. To confuse the dispositional power of an object or an actor with the exercise of that power is a mistake.

84 Power

It is not a mistake, of course, to use the evidence provided by the exercise of power to examine the properties or resources of the actor to try to understand the basis or foundations of that actor’s power. The resource account of power suggests that we can understand actors’ power by looking at their resources. Morriss does not disagree with this sentiment. He believes that examining and comparing resources is important indirect evidence of actors’ powers. He also believes that other (non-resource-based) indirect evidence is useful. However, these other categories all track actors’ resources, hence all evidence of power is resource-based. Importantly, though, Morris claims it is a mistake to identify an actor’s power with his resources. This is an ontological claim. The resource-based account of power claims that once all actors’ resources have been measured, we have a complete account of the power structure. Morriss disagrees. This would be to identify power with resources. The advantage of such a reduction is that it allows us to measure a dispositional concept – power – in terms of a non-dispositional one. This reduction from the dispositional to the non-dispositional will not allow us to produce point predictions. Even if we could accurately measure all actors’ resources we still could not predict precisely what the outcomes of their interactions would be. First, we would have to feed in their preferences. Second, even if we fed in their preferences, many interactions might have the form of multi-equilibrium n-player games with mixed strategies. We might be able to predict outcomes across probability distributions, but we would not be able to produce point predictions. Some of us might find this reassuring. In other words, seeing social life in terms of noncooperative game theory will ensure that the resourcist route will be predictive, but only in marginalist as opposed to deterministic terms. Morriss resolutely avoids considering non-cooperative game theory in his book. I believe that non-cooperative game theory needs to be at the heart of the concept of power as well as its analysis. This is so because what counts as an actor’s resources can only be understood in terms of how others view them. This will determine their actions towards him, his reactions to them, their reactions to his reactions, and so on. What others believe, and what the actor believes, are also predicated upon each other’s actions. Morriss (2002: 139) agrees: ‘we cannot observe resources directly. We have to infer that things are resources by examining other people’s reactions to them; one cannot simply measure resources since the worth of a resource is determined by the effects it produces.’ But he does not treat them in this way. Rather, he seems to equate resources with the physical attributes of actors. How this happens in his account is subtle, since his argument is complex and nuanced, but we will see how this works when I discuss his ability–ableness distinction. However, according



Power, capability and ableness85

to Morriss, to reduce an actor’s power to her resources would be to commit the vehicle fallacy. The vehicle fallacy as explained by Morriss does not allow the reduction of power to properties of the objects with that power. Morriss has two takes at this argument. One cites Anthony Kenny, from whom Morriss derives the idea of the exercise and vehicle fallacies, and the other Willard van Orman Quine, an arch reductionist. Kenny is quoted with approval, Quine is attacked. Kenny writes: Consider the capacity of whisky to intoxicate. The possession of this capacity is clearly distinct from its exercise: the whisky possesses the capacity when it is standing harmlessly in the bottle, but it only begins to exercise it after being imbibed. The vehicle of this capacity to intoxicate is the alcohol that the whisky contains: it is the ingredient in virtue of which the whisky has the power to intoxicate. The vehicle of a power need not be a substantial ingredient like alcohol which can be physically separated from the possessor of the power … The connection between the power and its vehicle may be a necessary or a contingent one. It is a contingent matter, discovered by experiment, that alcohol is the vehicle of intoxication. (1975: 10, quoted in Morriss 2002: 14–15)

The necessity or contingency of the relationship is not something that need unduly worry us here.2 The relationship between intoxication and the drinking of alcohol is well understood. Some people are much more prone to intoxication (and alcoholism) than others. The bundle of genes associated with alcoholism and susceptibility to intoxication is beginning to be identified, though few researchers (and no modern evolutionists) would doubt that one’s environment (or ‘nurture’) may also play a part (NIAAA 1992, 2000a, 2000b). Other factors, including age, may also affect susceptibility to intoxication. In one study of rats, degrees of intoxication were clearly age-related. The explanation is ­probably that young brains are built to learn, whereas old brains are more set in their ways, hence well-trodden neural pathways may continue to function despite the influence of alcohol (Holtz 1997).3 When people appear intoxicated may depend not only on how much they have drunk, but on how much they are used to drinking, how much they have eaten, their state of mind at the time they drink, and so on. Alcohol intoxicates, but it does so under all sorts of conditions. The degree of intoxication is affected by aspects of the drinker as well as the strength of the alcohol. So the power of whisky to intoxicate depends upon all sorts of factors, and these are analysable: we can judge the probability of a given bottle of whisky intoxicating different sorts of people under different conditions. This is the way we would analyse the ‘disposition’ of whisky for intoxication.

86 Power

The mistake of Kenny and Morriss is to think that because the power of the object (in this case alcohol) is affected not only by properties of the object, but also by the properties of other objects, we cannot reduce its power to that of its own properties. The thinking seems to be: the properties of an object only have power relative to the properties of other objects, hence the power of the object cannot be reduced to its properties. But this is false, since our judgement about the power of an object is also calculated relative to the environment. We calculate powers with scope restrictions. How we calculate those powers and how we calculate the scope restrictions are fully determined by the properties of all the relevant objects. In other words, we might calculate for those who drink a given bottle of whisky that it has the power to kill a child, stupefy a teenager and slow the reactions of the most hardened drinker. And we can compare its strength to another bottle of spirits. The comparative power of another bottle of spirits can be given simply by its alcohol content (and impurities which can affect speed of uptake of alcohol and short- and long-term damage to the brain affecting, for example, hangovers). Of course, other things affect how people react to a bottle of whisky, but we can analyse its power relative to other bottles simply by its properties. And, moreover, its total effects can be measured by reducing all the other elements to their component parts. The explanation of a bottle of whisky’s power to intoxicate can be reduced to more basic elements. We can do likewise for the social scientific concept of power. Morriss argues against this view by taking an example from Quine, who suggests that when we say sugar is soluble we are making a claim about its sub-visible structure. Morriss suggests that there is no reason why dispositionals must refer to sub-visible structures. In fact, he does not show there is ‘no reason’ why dispositionals should not refer to sub-visible structures; rather, what he argues is that we can understand sentences containing dispositions without further analysis. This is not the same as suggesting there is no reason why dispositions should refer to sub-visible structures, and certainly not that dispositions may be analysed and examined by looking at other properties of objects. It expresses merely that we do not need to know everything in order to understand a sentence of a natural language. Brian Barry (2002: 161) likewise suggests that the claim that sugar is soluble does not mean that it has a structure suitable for dissolving. It is true, of course, that the intension (the meaning) of a statement ‘sugar is soluble’ is not ‘sugar has a structure suitable for dissolving’; rather, it means something like ‘when you put sugar in liquid it becomes part of the liquid’. However, the statement ‘sugar is soluble’ has the extension (the reference) that Quine attributes to it. And Morriss (2002: 18) is correct that we can ‘talk perfectly sensibly’ about the dispositional properties of objects without understanding the causes of the dispositions the objects have. But I take



Power, capability and ableness87

it that when we do social science we are interested in more than simply talking sensibly about the world. We want to understand and analyse the world. Knowing how the word ‘power’ can be used sensibly is one thing, but understanding through analysing the power structure of society is another, much more important, thing to do. The extension rather than intension of ‘power’ is important when doing empirical political science. Outside of the social sciences it is obvious that it is the extension that is important, because causation is extensional. One event does not cause another because of the meanings we happen to attach to the words that describe those events.4 In the social sciences it is less obvious that it is the extension that is important, since people’s actions are based on what they think, so the intensional context is important. However, the intensional context can be captured if we include belief and desire within the extensional description. In other words, I think the best way of analysing the power structure is to examine what sorts of properties (or resources) people have (or can command) given what they want, or may want, to achieve. It is the properties that make sugar soluble, not the dictionary definition of solubility that is of interest to those trying to explain why some objects dissolve in liquid more readily than others. Similarly, it is the properties that lead some people to get what they want more often than others manage, and the properties that lead them to want what they want, rather than the dictionary definition of power, that is of interest to the social scientist. At times Morriss’s version of the vehicle fallacy seems to consist of the argument that simply because two things have the same dispositional properties, they do not have to have the same (structural) properties. Of course not, but it does follow that whatever (structural) properties the objects have must include some which give them the disposition in question. One might just deny this. To some extent, that is precisely what Kenny (1975) does. It is a standard move in the philosophy of mind where, it is claimed, certain properties simply are not further analysable.5 In this sense, the term ‘dispositional’ tends to mean something different to philosophers than it traditionally does to social scientists. A disposition for a philosopher usually refers to a causal regularity that cannot be mechanistically decomposed. In that sense dispositions are not necessarily unobservable, but are contrasted with ‘occurrent events’. Thus individuals’ beliefs and desires are not objects, and they are not ‘occurrent events’, so they have to be dispositions. But this just means that, under certain conditions q, an individual (or object) i will do (or display) x. Dispositional properties in the social sciences are thought to be unobservables that must be inferred to have causal properties. In which of these senses is a dispositional property of actors? If one takes the first line, then one may resist the reduction of power to resources,

88 Power

but only at a cost: the cost of forcing the analysis of power of actors to their beliefs and desires. Morriss will not, I think, want to pay that price, though below I suggest that the issue deserves very close attention, and the relevance of beliefs and desires to power ascription is more closely bound than many (including myself in earlier work) have recognized. Morriss’s important distinction between ‘ability’ and ‘ableness’ is particularly pertinent here and will also be discussed below. It is in the second sense that I have taken it that power is a dispositional property of actors. I take it that the causal properties of the actor to get what she wants are constituted of her resources (that which enables her to get some of the things she wants) and the wants themselves. Her causal properties per se include all those things that lead her to cause whatever she causes, and could cause. In these contexts, however, the power of an actor is usually conjoined to her desires, intentions or wants (more formally, her preferences), so that power does not simply become the infinitive of ‘can’ and be too closely associated with causal powers more broadly (see Chapter 5). What makes power dispositional is the fact that the resources that specify the ‘can’, given the environment, are only switched on under certain circumstances. ‘Fragility’ is a dispositional property of fine china, which is demonstrated when the china is dropped onto a hard floor. But it is a structural property of china that can be analysed without the china actually being dropped. What ‘switches on’ the power for actors are their preferences (beliefs and desires). But again, that dispositional property does not need ‘switching on’ in order to be analysed. In the same way that ‘fragility’ can be analysed without dropping the china, the power of actors can be analysed without their doing anything. At least, by and large, it can. (I will say more on this below.) Power in the second sense of disposition is measured by factors in principle observable, but often unobserved – actors’ resources: though of course the resources may only be exercised under certain conditions. If we can observe all resources, we have reduced the disposition to its (structural) properties and we have completely analysed power. In that sense, a resource-based approach to power should be able to analyse the power of all actors to the extent that all resources can be observed. However, the strategic possibilities which open up when actors do not have complete information about the resources of others, which we can safely assume is always the case, means that such a determination of the power structure will remain beyond our means. It does not follow, however, that a resource-based approach, theorized within a strategic framework, is not the best way to analyse the power structure. Morriss thinks this is false. Because of his belief in the ‘vehicle fallacy’, he does not think that a resource-based approach to studying power is the



Power, capability and ableness89

right procedure. He does not deny that strategic considerations enter into power relations, but thinks they do not enter in the definition of power. He would be right, but for the fact that it is strategic considerations, based on incomplete information, that stops us from fully specifying power in terms of actors’ resources.6 Thus strategic considerations cannot be kept out of the meaning of power any more than they can be kept out of its analysis.7 First let us consider the nature of resources. Resources In my own work (see Chapter 3; Dowding 1991, 1996) I used the term resources to cover five categories in which individuals may use their power. The first four were based on categories used by John Harsanyi (1976a, 1976b) and the fifth the addition of reputation. All are resolutely game-theoretical. The first two are information and legitimacy. The relative advantage that information gives allows some to get what they want more easily. Similarly, others may follow the lead, suggestions or orders of others for content-independent reasons if they recognize the authority or legitimacy of the first to so lead. The third and fourth categories are the abilities of an actor to change the incentive structures of others: through unconditional incentives and through conditional ones. Unconditionally changing incentive structures is to change the choice situation of others. This means that others’ preferences have not changed, but the conditions under which they make choices have altered. For example, a government may try to change the eating habits of citizens. It could try to get people to eat more healthily by persuading them that fried food is bad for them or it could tax cooking oil. The first uses information, the second unconditional incentives. The first will change the preferences of the citizens (to the extent it works), the other ‘preference shapes’ – it does not change the preferences, but it changes the choice conditions and so behaviour. Conditional incentives are simply the threats and offers, or combinations thereof that may lead people to behave differently. The fifth resource is reputation. Here people may respond differently to an actor because they believe he has other resources (that he may not in fact have) or because they may predict his actions based on his (or others’ similarly situated) past behaviour. With the exception of information, none of these sets of resources explicitly mentions the many attributes of people that might be thought to be resources – money, charisma, a standing army – and so on. Any such list of physical attributes one could produce would almost certainly be incomplete, for many things might be used as a resource in this sense. Rather, the five sets of resources provide

90 Power

an organizing principle around which we may study how actors respond to each other in different environments. Morriss sees actors’ powers, correctly, as their property. Their powers are their abilities to get things done. He says: A person’s abilities are a property of the person, not of the environment. Like all dispositionals, they carry explicit or implicit references to the conditions in which they apply – and assert nothing about what happens when those conditions do not apply, nor about the likelihood of those conditions occurring. (Morriss 2002: 80)

The problem with the confidence with which this claim is asserted is that some properties of people only become properties under certain conditions. It is difficult to denote them without the environmental conditions applying. Take the ‘gene for reading’. A gene was once thought to be a ‘chunk of DNA’. But as we have learned more about genes, it becomes clear that complexities occur. A ‘gene for’ something is best viewed as an abstract truth about inheritance, and the physical attributes of the gene may be distributed around in the physical attributes of DNA. One might think that a ‘gene for reading’ will include any part of the DNA structure that, if missing, will lead to reading abnormalities. But this would be too quick. Genetic defects that cause, say, SIDS (‘cot death’) will mean that the baby will not be able to learn to read in its life. But the fragment of DNA which ensures healthy sleeping is not normally thought of as part of the DNA which constitutes the ‘gene for reading’ – otherwise the DNA for healthy lungs, heart, and so on would also be part of that gene. More pertinently still, the ‘gene for reading’ was present before people could ever read. It may have been useful for tracking, say, and was selected when hunting was important. Later, as humans started using symbols and then writing was invented, the same abilities made possible by these genes allowed most people to develop the ability to read. Of course, the ability to read entails more than having this gene. It also entails having been exposed to and been taught how to interpret the symbols. There is no doubt that the ability to read is a property of the person, but the ability to read only makes sense in an environment (a culture) where reading exists, even though people who have never lived in such an environment have the requisite genes and so could have developed the ability had they lived in such an environment. The ability to read is not, of course, a property of the environment, but the phrase in parentheses above is important. The environment in this case is a culture that includes writing and reading. Abilities (at least some of them) are then intimately connected to the envir­ onments in which people who have them live. Counterfactuals concerned



Power, capability and ableness91

with such abilities must take into account the mechanisms by which societies (the environment within which lives our individual with his ability to read) develop such abilities. Ability and ableness Morriss makes a distinction between ability and ableness. Whilst both might be measures of an actor’s power, he believes that the best measure of an actor’s power is ableness. I have criticized Morriss’s notion of ableness on several occasions. He suggests that my criticisms are false for two reasons, one elementary and one instructive (Morriss 2002: xxxv–xxxvi). My elementary error is misunderstanding how to counterfactualize ableness; my instructive one is where I introduce the concept of luck. The two are closely related. Let us first consider my elementary error of misunderstanding how to counterfactualize ableness. I do not, of course, think I misunderstand the counterfactuality; rather, my argument is that the counteractualization of ableness in which Morriss indulges is inappropriate. It comes into the class of intuition pumps that go wrong because some relevant factors are omitted (see, for example, Dennett 1984). The example used by both Morriss and myself is whether or not all the judges on a supreme court have equal powers or whether one may have more power than the others. Now, in a real supreme court it would be unlikely that all had precisely equal powers. Some will be more persuasive, aggressive or charismatic than others, or simply have more authority in some sense, and hence be able to determine the decision of the court more often than other members.8 But this is not what is at issue. What matters is whether any member has more power based entirely upon one resource: their vote. If each member has one vote, is there any sense in which one member might have more power than the others? Morriss thinks yes; I think no. Imagine there are two blocs of four voters (one conservative, one liberal) plus one maverick justice (MJ) who sometimes votes with the liberals and sometimes with the conservatives. Morriss says MV has more ‘ableness power’. I say: MJ only appears to have more power because she happens to have been denoted in the example as MJ: that is, she has a different preference structure from the others. Her different preference structure does not, in fact, give her greater power over the results since the two blocs have their reasons for voting together and she has her reasons for voting as she does. She only appears more powerful because the preferences of the other two blocs are taken as given; hers is not. (Dowding 1991: 60)

92 Power

Morriss says ‘this last sentence is just wrong. When investigating the power-as-ableness of an actor, we take the actions of all the others as fixed, and examine the outcomes that can be brought about by the actor whose power is being investigated’ (2002: xxxvi). I am not too sure what is ‘just wrong’ about the final sentence. In my sentence the preferences of the eight justices are taken as ‘given’ and in Morriss’s the ‘actions’ of the others are ‘fixed’, which, given that I use the term ‘preference’ as used in revealed preference theory (the standard axioms of rational choice), means they seem to be identical. Of course, if the actions of eight actors are fixed to a tie, then the ninth will determine the result. That is not in question. What is in question is whether this is a measure of the actors’ voting power. It is not. The counterfactual is inappropriate to this question. Imagine a member of one of the blocs, called BJ. When we consider BJ’s power, keeping all the other (non-MJ) votes fixed, sometimes the vote is 5–3 and sometimes 4–4, and ‘we can conclude that BJ is only sometimes able to win the vote. MJ therefore does have more power than any BJ’ (Morriss 2002: xxxvi). No, we cannot. BJ can no more ‘win the vote’ than can the other four voting his way. As I point out (Dowding 1991: 59–61), being pivotal in this sense, if it is known that one is pivotal, can bring other powers. One is open to side-payments, one is in a position to get other things one wants (such as bribes). But being pivotal in this sense does not bring any more power to achieve the outcome under vote than the one vote which each of the justices brings to the court. Morriss misunderstands this counterfactual argument since he follows a poor intuition from an under-specified intuition pump. He says he looks at the ‘bribe value’ of MJ and BJ to some other actor. He suggests that if the CIA wants to bribe a justice, MJ will be better for them than any BJ. The others’ votes are predictable: they will split the court, and MJ can be purchased to get the results the CIA wants. Not necessarily. Despite being a maverick, MJ may have strong views on each of the cases brought before her. She may not be prepared to sell her vote to the CIA. She may be honourable, and whilst everyone may have their price, hers may be more than any given BJ. Indeed, the CIA may find they can buy two liberals for each maverick. This is precisely my point. Each justice has their reasons for voting one way or another, and it is those that give us predictions about their behaviour. But their power is to get what they want given their resources (and the environment they inhabit), and in the voting index about which we are both writing the relevant resource is their one vote each. If we want to look further at their bribe value, we need to look at their reasons for voting as they do. It is simply not true that MJ is more easily bribable than the other eight justices. Just because she is less ideological does not mean her



Power, capability and ableness93

preferences are weaker, and hence she is more bribable. The CIA may well get what it wants by bribing one of the liberal justices (assuming the CIA is right wing) rather than the MJ. Then it can ensure a five-way right-wing vote, whichever way MJ votes. (Hence the CIA will win 5–4 some of the time and 6–3 some of the time.) Assuming strength of preference is identical among all the justices, the bribe value of the actors is identical – hence Morriss’s argument demonstrates that in this scenario all actors have the same voting power. But why should we not do the counterfactual analysis that Morriss wants in this example? Why not hold the environment, including the votes of the other justices, constant, and then look at what someone can do? What is wrong with ableness here? The answer is that holding other things constant in this way ignores the fact that parts of the environment are reactive. In many contexts what a BJ will do may depend on what an MJ will do. Many, if not most, power situations are strategic. That is not to say that one way of working out what an MJ should do to get what she wants is to imagine what the others would do; but she should work out what they are likely to do under several scenarios, scenarios that include her actions varying. Obviously this takes us into game theory, and without a game-theoretic analysis of such interactions our concept as well as our analysis of power will be poorer. It is Morriss (2002: 91) who seems to make the elementary error: Your ableness to win the competition would not be lessened because many other people had this ableness too. When we assess powers person-by-person like this, we take the actions of others as given and counterfactualize the preferences and actions only of the actor whose power we are considering (and, of course, the preferences of any others that we predict will be altered if the actor under consideration changes her preferences).

How can we hold all others’ actions as given, if we are allowed to vary their preferences (as we surely must) with the preferences of the actor under consideration? (That is not to say, however, that keeping others’ preferences and actions constant whilst we vary those of the individual under consideration is always illegitimate, but in many strategic contexts it would be. The only way one can find out how to play a strategic game is by working out what all the other players will do depending upon one does. Here we look at all the logically possible scenarios, but it is the feasible set that matters for analysis.)9 It might be thought that Morriss and I are simply looking at the issue from two different perspectives. He is looking at the issue as one of the power of coalitions, and I am looking at the issue as one of the power of

94 Power

individuals. Imagine the committee of nine is not a court but a parliament. There are two parties composed of four MPs and one party of one MP. In this case it is quite natural to claim that each of these three parties has the same power. In either the Shapley–Shubik or Banzhaf power indices each party has the power of one-third. Here surely the member of the party composed of one member (MJ) has more power than any member of the parties composed of four. We might think that each member of the party blocs has a quarter of the power of 1/3: that is, 1/12. But that is not correct. MJ may have the power of 1/3 seen thus as a party bloc. But MJ’s power disappears if one member of either bloc votes against his bloc. Each member of each bloc has the same power as MJ. They only appear to have less power than MJ because we are considering the bloc as necessarily voting together. In my terms, MJ is lucky, not powerful. It is this that Morriss calls my instructive mistake. Following Barry, I suggest that luck is getting what you want without trying or, as Morriss correctly amends it, one is lucky to the extent that, given one’s preferences, one can get more than one might expect. MJ is lucky in that she is always decisive. As I point out, that luck might have been based on someone else’s power (a president who appointed her who shares precisely her ‘maverick’ preferences) – but for her it is luck. Morriss seems to want to equate ‘ableness power’ and luck when he says ‘[w]here Dowding goes wrong is in thinking that luck (in his sense) and power are incompatible’ (Morriss 2002: xxxvii). I make explicit the point that an actor may be both powerful and lucky (for example, Dowding 1991: 83; see also Chapter 5); that is, someone may get more than their effort would suggest, but they could have put more effort in. But this is not what Morriss means here. He agrees that MV is lucky; but why, beyond the fact that she gets what she wants all the time because she is pivotal? In that sense, luck and ableness seem to be the same thing. Similarly, my extension of ‘luck’ to ‘systematic luck’, where the structures of society lead some to be lucky due to their positions in society (which also shape their preferences), Morriss equates with his concept of ‘passive power’. Are Morriss and Dowding having an uninteresting spat over the term to use for a concept? I think not: at least not, given other ways in which Morriss uses the notion of ableness and his critique of realist approaches to power. For Morriss the distinction between ability and ableness is vital, for he holds that power-as-ableness is the important sense of the term. (For me, luck is much less important than power. Indeed, it is a residual used to explain why some people get more of what they want than we might expect given their resources – or given how they have spent those resources.)10 Morriss (2002: 83) suggests that ‘an ability-sentence contains (explicitly or implicitly) counterfactual descriptive and manifestation conditionals’,



Power, capability and ableness95

whereas an ‘ableness-sentence cannot contain counterfactual descriptive or manifestations conditionals’. The former is descriptive under which the conditional holds (for example, the conditions under which sugar will dissolve in water) and the latter denotes the conditionals which hold when those conditions obtain (when the water and sugar actually come together). He goes on: ‘For this reason, social and political power is usually a sort of ableness and not an ability; in social philosophy we are not usually interested in what people could do if they had resources that in fact they do not have’ (Morriss 2002: 83). There are, I think, two things wrong with this claim. First, it is not clear that ableness sentences do not contain counterfactual descriptive or manifestation conditionals. The sentence ‘I cannot read it today (my glasses are broken)’ implicitly contains the descriptive conditional ‘if my glasses were not broken, I could read it today’. The importance of the counterfactual conditions is that they specify the resources that I require in order to be able to read. In this sense, ableness is what we are interested in, for it is the resources that I require in order to be able to read. The only way I can see a distinction between ableness and ability is some physical feature (the gene) of an individual that implies that an ability could be there. But, as we saw above, genes cannot simply be treated in that way, for the phenotypes (us and our behaviour) are a relationship between our genes and the environment, and an environment, moreover, we have partly fashioned. Reading is something we have fashioned, given our genetic inheritance to be sure, but one cannot read off ableness from anything simply physical about a person, such as a fragment of their DNA. Imagine, for example, that Homo erectus had the fragments of DNA that are identified as ‘the gene for reading’. Would we say Homo erectus had the ableness to read, 300,000 years before writing was invented? And this leads to the second problem. Typically, when analysing the power structure, we are interested in the resources that people do not have, since we are explicitly or implicitly comparing the powers of classes or types of people. One of the reasons that Morriss does not get to the heart of the difficulties of counterfactual analysis in the main part of his book is that he only considers somewhat contrived examples of what people might do if the CIA bribed them or threatened them. The real problems in social science concern not such token individuals, but types of individuals. Typically, the sorts of questions we ask in the social sciences are: what is the US president able to do when another party controls the legislature? How does his power differ from that of the French president under the same circumstances? Or, can institutions be designed to counteract the power of developers in local politics? Is the march of globalization inevitable? A lot more factors enter into these considerations than whether someone will do their best, or not, to try to

96 Power

help the CIA (see Morriss 2002: 77–9). And many of these factors involve the mechanisms that may (or may not) exist within the structures of the legislatures and the political systems of the respective countries (in the presidents example), as well as considering the actions and reactions of the multitude of actors involved (in all the examples). Here, more control is required than that allowed by Morriss’s approach to counterfactuals. Morriss’s arguments are focused almost exclusively on a token individual’s power. However, my focus is almost exclusively on examining the power structure of societies and thus the power of types of actors. Thus part of an actor’s ableness – his skill – ‘is, of course, itself a dispositional term’ (Morriss 2002: 138). From my perspective, however, we can assume that the unobservable elements such as ‘skill’ within a type are normally distributed across the set of actors within the type, and therefore need not worry too much about this aspect of their resources. We may assume their luck is normally distributed also. It is only when we come to compare across types that we become interested in their relative resources and luck.11 For someone writing a book on a theoretical concept like power, Morriss has a very inductivist approach to understanding and truth. Time and again he suggests that we can only discover actors’ true powers (and the dispositions of objects such as sugar) by experiment. Theoretically, however, once we understand why salt dissolves in water, we should be able to predict that substances with similar ‘sub-visible’ structures, such as sugar, will also dissolve in water. Demonstration in science and the social sciences is not always through experiment or empirical observation.12 Morriss’s atheoretical inductivist approach leads him to discount realism and underlies his approach to the ability–ableness distinction. Anti-realism Part of Morriss’s attack on the vehicle fallacy is his dismissal of realist approaches to power. In this context a realist approach is not merely a claim about the truth-values of particular power claims, but of mechanisms that underlie power ascriptions. Ted Benton (1988: 492) calls Morriss’s approach one of power-attribution that can function as an explanationsketch or explanation-attribution. In Morriss’s inductivist approach one cannot go further in causal claims than Hume’s constant conjunction. The idea underlying realism is that, to take the fragile china example, there is something structural in the nature of fine china that makes it fragile. This entails that when dropped it will (usually) break. Morriss argues against this view of underlying structures, entailing that china is fragile by considering ‘finkish dispositionals’.



Power, capability and ableness97

We can follow an intuition pump of C. B. Martin (1994) discussed by Morriss. Let us say that the fragility of china is related to temperature. The warmer the china, the less breakable it is. Martin suggests that if the china is warmed, then dropped and does not break, it is not fragile.13 But imagine, with Martin, that between being dropped and hitting the ground ‘a divine agent’ intervenes to make it break. In Martin’s example, this divine agent ‘makes it fragile’ again – perhaps by cooling it. However, it might do it by any sort of means: speeding it up just before it hits the ground or zapping it with subatomic waves that cause it to break up as it lands. Is this ‘finkish disposition’ a real disposition to fragility? According to Morriss, we should sometimes ascribe fragility under these finkish circumstances and sometimes not. We should not if we are doing a scientific analysis of the china, but we should if we are labelling this special china (with its ‘divine agent’ mysteriously attached) for sale to the public. The problem with these absurd intuition pumps is that their authors do not take them seriously enough. If one is to entertain the notion that a ‘divine agent’ can intervene between dropping and breaking, we must consider how we, as scientists, would analyse that condition and how we would then consider it in our scientific analysis of china. I suggest that we would categorize the fragility of objects in terms of their properties – all their properties – and relate these to the conditions under which they causally lead (under some probability distribution) to certain outcomes. In the case of the china, we would measure the structural properties of the china, and relate this to the probability of its breaking when dropped under different gravitational forces, temperature, humidity and other conditions. Given the fact that some types of china break under conditions that our theory does not allow, we reanalyse in appropriate ways, discover the divine agent who intervenes, note the type of china that the divine agent is attached to, and modify our laws accordingly. In Martin’s example as I have further specified it – by suggesting ways in which our divine agent intervenes – we do not have to modify our account of fragility, but rather the conditions under which the special china is dropped. These always alter in some manner caused by the divine agent to ensure the china breaks. On the other hand, we might find that the divine agent has a new way of breaking china and then we might have two rather different mechanisms for the same outcome. Generally speaking, in natural laws, such mechanisms are often related, with more general accounts underlying both. However, prior to the discovery of such underlying accounts, we can easily work with different mechanisms. As pointed out earlier, the fact that different structural properties may underlie the same dispositional properties does not mean that those structural properties do not fully explain the dispositions. (To repeat, they do so under certain

98 Power

conditions, but these conditions can be specified as part of the dispositional property.) So even under the finkish circumstances we can still reduce the disposition to natural properties (taking the intuition pump seriously) and can discover real properties and mechanisms.14 Morriss makes strange use of the finkish disposition example. He suggests that the powers of actors might indicate different things: ‘It may be that each of two leaders has the power to do the same specific thing but the explanations of their powers could be very different: the power of one derives from the social structure, whilst the other’s would come from the leader’s personality’ (2002: xxx). Yes. But why is this statement a criticism of realism? Actors may have very similar powers based on very different resources, which is precisely why it is of interest to carry out the reduction of explaining their powers in terms of their resources rather than leaving them as explanation-indicators. Where Morriss and some of the realists he criticizes may both go wrong is insisting on a strong ontological distinction between an actor and the environment such that we have either individualist or structural explanations of power in society. As I have tried to explain elsewhere, the two are not contraries. Only individuals act, but their preferences are shaped and their actions enabled and constrained by their resources and those of others around them (Dowding 1991, 1994, 1996, 2001a; Dowding and Hindmoor 1997). One person’s action is (part of) another’s structure deep down into their preferences (see Chapter 2). Again, I repeat, in the social sciences most explanation is given in terms of types – leaders, dictators, capitalists, and so on. We are interested in the types of resources that these actors have, and the distributions of these resources relative to other distributions.15 This does not imply, of course, that we cannot use these types when we investigate the powers of token actors. But our knowledge of their resources will be informed by what we have discovered, both theoretically and empirically, from the resources of others in similar and often dissimilar situations. What is the fallacy of the vehicle fallacy? The vehicle fallacy is itself fallacious. It is ‘ignoratio elenchi’ – supposedly refuting one thing, whilst in fact arguing against something else. Kenny and Morriss argue that one cannot reduce the dispositional concept of ‘power’ to the properties of the object to which the power is attributed. But one can. What one cannot do is analyse the causal effects of those properties without also taking into account the properties of other objects. That is what they should have been demonstrating. But that is true of all features of all objects. Indeed, any unary property of an object can only be defined in terms of binary properties – the relations of that object (and its properties)



Power, capability and ableness99

to other objects (and their properties) (Dowding 1991: 15). Hence we cannot identify an object without taking into account other objects; nor can we measure its physical properties, causal properties, powers, or whatever. Not understanding this logical feature of our descriptions of the universe causes almost as much confusion in social scientific theoretical discourse as misunderstanding the importance of the type–token distinction in explanatory contexts (see Dowding 1994, 2001a, 2016). Power undoubtedly is a resource of actors, but we can only analyse that resource in relationship to the resources of other actors. How much we can buy with the money in our possession in part depends upon how much money is in the possession of others. The power of money is relative to its distribution. This means that strategic considerations must enter into the very essence of the concept of power. How much power people have, what resources they have, may in part be composed of their luck. We may have no control over what resources we have, or what resources others have. Nor indeed over our preferences or those of others (though we may have some power in both cases). For that reason, I think the analysis of both power and luck is important, but we must start that analysis by looking at actors’ resources and the strategic situations in which they find themselves. I began by suggesting (following the lead of Morriss) that what Sen means by the term ‘capabilities’ is actually a form of individual power, ability or ableness. And, for Sen, crudely speaking, justice requires that we maximize functionings (what people do with their capabilities) and equalize capabilities.16 To the extent that capabilities are power, and to the extent that I am right that we can best measure individual power in terms of resources, then Sen’s capabilities account of what egalitarian justice requires is not so distinct from Dworkin’s (2000) resource account that Sen criticizes in building up his capabilities approach. To be sure, my account of what individual resources are is broader than Dworkin’s, and in that broader account might lie the difference between Sen and Dworkin. Nevertheless, I believe that working through the measurement of power, or capabilities, will lead us to examine individual resources, and, in that sense, Sen’s capabilities account at base, is resourcist. Notes  1 All references are to the second (2002) edition of Morriss’s book, first published in 1987. The second edition is identical apart from minor typographical changes and a new long introduction.  2 I would say it is a ‘naturally necessary’ feature of alcohol. Alcohol was developed and drunk because it helped destroy germs. It also has the pleasant

100 Power effects of intoxication. Thus whisky and its alcohol content are not simply contingent features of our social world. They were developed for a purpose and they fulfilled a function, so their presence and why we drink whisky are not entirely contingent. Nor are people’s reactions to its effects contingent. People of European heritage are less susceptible to the effects of alcohol than those from Asia or the Americas, since in these regions other techniques were used to rid drinking water of germs – heating and using herbs, for example (Diamond 1998).  3 Alcoholics, on the other hand, may be strongly affected by the first drink, as the alcohol immediately sets their brain on its ‘alcohol-directed path’.  4 Though what we recognize as a ‘cause’ depends upon what causal question we are asking. And so what we define as ‘cause’ and what as ‘background conditions’ attach to the meaning of the words by which we describe those events.  5 Such as qualia.  6 See Dowding (2006c) for a related argument about capabilities, and Chapter 7 for an extension of the strategic problems over what to equalize for luck egalitarianism.  7 Braham and Holler (2005) argue that the concept of power is contained with the game form though it is exercised through the game. This is a subtly different account from Morriss’s. Here the game-form will specify the power of the player given how the game will be played once preferences are assigned. (A game form is the rules of a game without the assignation of preferences.) But how the preferences are assigned is, in my terms, luck, since that will specify which coalitions are likely to form. An agent’s power is then partly a matter of luck and partly a matter of the game form. We can identify resources with the rules of the game (the game form).  8 Studies of voting patterns on supreme courts in US states have shown that where the votes are taken in order of seniority, decisions go with the first voters more often than when votes are taken in reverse seniority or random order.  9 Dowding and van Hees (2007) consider the strategic interrelationship between preferences and the concept of freedom. 10 It is not an ad hoc device used to ‘save’ any purported explanation. Rather, luck – especially in its systematic form – is analysable in terms of the convergence or divergence of actors’ preferences. And, I think, other structural features of society can help explain such convergence or divergences. I show in Chapter 7 that if outcomes are equalized (by whatever that equalization is measured) luck disappears. 11 For more on the importance of type and token in the measurement of rights and freedoms, see Chapter 12; in relationship to different measures of welfare, Dowding (2008c); and in the social sciences more generally, Dowding (2008a). 12 Of course, in the sugar example empirical demonstration may well be easier than theoretical demonstration! But this is not so for all cases in the sciences or social sciences.



Power, capability and ableness101

13 To begin with this just seems like a mistake. As Morris recognizes, dispositions need to be defined in terms of the conditions under which they hold. An object is still fragile under the conditions in which it easily breaks. We consider fragility of ordinary objects under the temperatures in which we handle them. But scientifically fragility would be defined relative to temperature. 14 If the divine agent only sometimes intervenes, then we either analyse the conditions under which he does – perhaps it always happens on wet Tuesdays – or recognize that the intervention is random under some probability distribution. 15 This is what we are interested in when considering the power structure. In mainstream political science, ‘power’ as a concept has almost disappeared as much is concerned with explaining specific outcomes rather than the relative outcomes (Dowding 2008b), though as the literature on the power of different institutions of the EU demonstrates, power does re-enter when those relative comparisons are made. 16 Sen denies that he has a ‘theory of justice’ and would, I think, say this aphoristic representation of his argument is simplistic; nevertheless, the upshot of his considerations seems to me to boil down to this claim.

Part II

Luck

5 Resources, power and systematic luck: a response to Barry

Introduction Brian Barry presents some ‘puzzles’ about political power, and defends the idea that capitalists are powerful (Barry 2002). I doubt that anyone would dispute that capitalists are powerful; it is my fault that Barry feels the need to argue the case. I have argued that capitalists are ‘systematically lucky’ (Dowding 1991, 1996). I did not deny they were also powerful, but did claim that capitalists were lucky in the sense that governments would often do what capitalists want without them having to intervene in the policy process. I think the reason Barry feels the need to argue that capitalists are only powerful, rather than powerful as well as systematically lucky, is ideological, or at least normative. He feels that saying capitalists are systematically lucky somehow lets them off the hook. If capitalists are systematically lucky when governments do what they want them to do, then capitalists are not so responsible for how the world runs. If we want to hold them responsible, we must deny they are ever lucky, and claim it is their power that is decisive. This implication should not be taken from my argument. In the next section, I briefly recount the basics of my resource account of power, and in the subsequent sections I contrast this account with Barry’s thoughtful attack on the resource account. The idea of ‘luck’ in the context of studying power was introduced by Barry himself when he suggested that ‘getting what you want without trying’ is a form of luck rather than power (Barry 1991a). He introduced the concept in the context of some of the ‘paradoxes of power’ found when examining power through the ‘power index’ approach (for example, Brams 1975). Systematic luck suggests that people may get what they want without trying systematically. Others will do what the lucky want without the lucky needing to intervene specifically to ensure this happens. Systematic luck is tied to the social location rather than to the person.

106 Luck

Of course, the power of people is often tied to social location as well. A president, for example, gains many powers on assuming office. When we think about the power structure in society, we usually map it in terms of social types – presidents, capitalists, top bureaucrats, the middle class, and so on. We often also look at individual people, and the abilities and skills they may have developed during their own personal history. I am going to defend the idea of ‘systematic luck’ and the fact that capitalists may be systematically lucky. In fact, I think the idea of systematic luck has a better application to other groups, and I discuss at much greater length the case of British farmers in the post-war years, though Barry does not mention them (Dowding 1991: 152–7). My aim here is much broader, however. ‘Systematic luck’ is not an important concept in itself, but is significant only as part of the approach to empirically studying power in society. I want to defend the ‘resource approach’ to mapping the power structure that is explicitly the main target of Barry’s article. Some preliminaries The notion of power that I discuss here is tied to actors. Only actors can have power. Structures cannot have power. If you want to claim that structures have power, then, I suppose, you can. But whatever power they have is not the power that I discuss. You can call yours ‘powers’ and I will call mine ‘powera’. The power that I discuss is something that actors can choose not to use. A structure cannot do that. I call this the conceptual argument. But I believe that using the term ‘power’ when we are talking about structures is a mistake. Anything that one can say using the term ‘structural power’ one can say just using the term ‘structure’: this is the redundancy argument. The word ‘power’ in the phrase ‘structural power’ is redundant (Dowding 1991). When examining concepts and distinctions in the social sciences, I think we always need to look at why the concept was developed in the first place. All too often, concepts are developed for a specific purpose, but are then used rather differently. Criticisms are often directed at the concept or distinction as they are later used and the original context is lost from sight. Examples of this process include Clarence Stone’s (1989a) concept of ‘regime’ (see Dowding 2001b). This was initially developed to explain why politicians who apparently had divergent ideologies and divergent electoral bases of support end up producing convergent policies – especially over development issues. Later uses of the term ‘regime’ in urban politics take it far from this original use. We therefore need to keep in mind the reasons I originally introduced the term ‘systematic luck’ into considerations of the power structure. It was intended, first, to defend a structural



Resources, power and systematic luck107

account of power without falling into complete or radical determinism; and, second, to ensure a place for the empirical analysis of power structures within a theorized model. In my work on power I was concerned to do two things: to critique what had become the dominant empirical use of the term power, and to critique those who believed that power was such an essentially messy concept it could not be empirically studied. To the first end, I defended behavioural (or revealed preference) analysis that utilizes a theory of action. The valid criticism of behavioural (or revealed preference) analysis is that simplistic conclusions are often drawn from mass or individual action because that action has not been modelled theoretically. One cannot conclude that because a group of individuals do not promote some end x they have no interest in promoting x. They may simply face a collective action problem. On the other hand, simply because some group does not act to promote some outcome in their interests does not show there is any one or any thing which acts to stop them. Acts here is a key term. Something stops them – in my example, the collective action problem. But collective action problems do not act to stop them. If one does not think that the use of the phrase acts here is important in a discussion of the power structure, that if someone is powerless it must be because something else is powerful, you will find yourself committed to radical determinism (though I suspect many who might claim this would not be aware it commits them thus). I have no argument that radical determinism is false. Indeed, I think that any sort of formal analysis leads us down a deterministic path (see below), but like most people I play the game as though determinism is false, and my use of the terms ‘luck’ and ‘systematic luck’ are terms enabling me to play that game. In other words, it seems to me, that playing the non-fully determined game must leave room for someone not getting what they could have got and wanted despite the fact that no one and no thing stopped them. That which did not stop them includes the fact that the person made a calculation to the effect that the effort of trying to get what they wanted was more costly than the expected gain from that activity: a probabilistic calculation. Given that probabilistic calculation, some would say that we have no need to introduce luck (Felsenthal and Machover 1998: esp. ch. 3). Outcome power can be measured probabilistically – at least it can in terms of the decision theory underlying part of my conceptual analysis of power. That, indeed, may be a better way to proceed than the course I took. However, it does bring costs when doing empirical analysis – especially empirical analysis using qualitative rather than strictly quantitative data. The statement ‘developers have more outcome power than the local citizenry because the probability that they will get more of what they want is greater than the probability that the local citizenry get what they want’

108 Luck

may be quite clear, but it does not elucidate to the same extent why developers are more likely to get what they want. Nor does it suggest different strategies to change the relevant probabilities. My approach (following Barry) is to create a distinction between the probability of success (and thus power seen as a success concept) and power. In doing so, power is seen as ‘decisiveness’ – understood in a simple voting game as the difference one can make in ensuring that one gets what one wants, but more generally the ‘difference’ one can make to some outcome (itself analysable probabilistically). The residual component left over is luck and is composed of the difference between probability of one being decisive (the difference one can make to the likelihood of the outcome occurring) and the likelihood itself.1 When one attempts to empirically cash power in this sense, one is led, I believe, to analysing the power of an actor in terms of the resources they command. We end up by making the calculation the actor does herself when acting. For analyses of types of individuals, this involves us considering the institutional resources they can bring to bear on some problem. The actor’s resources are a proxy for her power. But this proxy is comparative. The difference my resources can make to the probability of some outcome occurring depends upon the resources others can bring to it as well. Those others not only include those who may oppose what I want, but also those whose interests are convergent with my own and with whom I may form a coalition. There is a danger in seeing the concept of ‘resource’ taking on anything that could be brought to bear, until resource plays the role of the environment in which decisions take place, as well as those parts of the environment controlled by the actor. I think we must restrict them to those that are in the actor’s control. For example, how much influence my money can buy depends upon how much money others have and how they spend it. But your money should not be considered my resource, only my money. In other words, my resources need to be quantified relative to others, but this relation is both to how others could use their resources and to how they actually do use them. There may be no best way of slicing up the social world. Different approaches to the concept of political power may all be equally acceptable. There are no natural kinds in power talk. Barry carefully distinguishes between different types of power, and distinguishes power from influence, from persuasion and from legitimate authority. He illustrates these by appeal to linguistic practice. For example he suggests we do not attempt to influence hot dog sellers by offering them $1.25 for a hot dog, but we would if we suggest they might please more customers by using a different brand of mustard. I have to admit I am not too concerned with such linguistic quibbles. When empirically studying power in society, I am only



Resources, power and systematic luck109

interested in making distinctions if they help me to understand a causal story and help explain what is going on in the social world. Different ways of cutting up the world may be equally as good, but if they are equally as good, then the concepts used in one approach must be translatable to the other. This does not entail one-to-one correspondence from concept to concept within each language, but it does entail that everything conceptually captured in one can be captured in the other. Any residual terms left over after the translation denote that either one language is superior to the other with regard to explaining some part of the world or that these leftover terms are otiose. As we shall see, Barry introduces a host of definitions of power that differ greatly from the definitions I offered. In itself this should not matter, but I will argue that Barry’s definitions produce just as many ‘linguistic oddities’ as my own and, more pertinently, are not as useful in making important causal and therefore normative distinctions. Barry’s definitions and the case of consumers and voters Barry makes various distinctions within the scope of power. His most general definition is ‘power is the ability to bring about desired states of the world by acting’. He refers to this as power as ‘ability’. He contrasts this with ‘the possession of the means of bringing about desired outcomes’ (Barry 2002: 160). Now, one might be confused by this contrast. What is one’s ability if not the possession of the means of bringing about desired outcomes? If I have the ability to win the Wimbledon tennis championship, then surely this is because I am in possession of the means to win that tennis championship. Indeed, Barry claims there is a ‘logical gap between having the means and having the ability’ (2002: 161). I am not sure wherein this logical gap consists. There is certainly a logical gap between having the ability to do something and actually doing it; but that same gap surely exists between having the means to do something and actually doing it. The clue to what Barry might mean is his reference to Peter Morriss and the quoting of Quine’s claim that when we say sugar is soluble we are saying that is has a structure suitable for dissolving. Barry says ‘No doubt it does, but that is not what we are saying’ (2002: 161). What he means is that the intension (the meaning) of our statement ‘sugar is soluble’ is not ‘sugar has a structure suitable for dissolving’; rather, we mean something like ‘when you put sugar in liquid it becomes part of the liquid’.2 Nevertheless, as I argue in Chapter 4, when we do social science it is the extensional account that is important and we best analyse the powers of agents in terms of their resources. The vehicle fallacy is itself fallacious.

110 Luck

Barry’s idea is that by distinguishing ability to bring something about from having the means to do so brings specific advantages when we restrict analysis of power to ‘social power’. He defines social power as ‘the ability to bring about desired states of the world by acting in such a way as to overcome the resistance of others’ and says ‘this is intended to capture … the idea of a power relationship as one that incorporates a clash of wills’ (Barry 2002: 161).3 He further claims that it is in this sense of power that voters have power over those elected and that consumers have power over producers, [and so] we also have to say that those who own or control capital have power over government. Conversely, the reasons that can be given (and have been given) for denying that owners of capital have power over governments would be equally good reasons for denying that voters have power over governments and that consumers have power over producers. (2002: 156)4

Now, of course, if capitalists, as I argue, can be powerful as well as lucky, then consumers and voters can be powerful as well as lucky. I will argue that when a voter gets the government she voted for (and possibly wanted), she gets it more through luck than social power. When a consumer buys a product with the precise specifications he wants, he gets it more through luck than social power. As for capitalists, when governments pursue what a capitalist wants, then she may also get it more through luck than social power, but the social power of the capitalist in that regard far exceeds that of consumers and voters, and capitalists use that social power far more often. I have twisted what Barry is claiming. I have turned attention to individual voters, consumers and capitalists. Barry uses the terms as collective nouns and it is to the collective that he seems to apply the term ‘power’. Each individual consumer does not have the power to force a firm to change its product specification. All each consumer can do is buy a rival product. Each consumer has the power of their own purse to spend on the products of their choosing. (Below I define this as outcome power. They have the power to bring about the outcome of getting the product on offer they can afford.) Of course, firms must respond to market signals, and that is sometimes referred to as ‘consumer power’. But ‘power’ is usually put in quotation marks to show its metaphorical use. In what sense is this ‘power’ social power in Barry’s sense? Does the consumer, when he switches products, really do it in order to get a firm to change the specification of its products? Surely each switches purely to get a better buy, which, if many others do likewise, may, as a by-product, lead a firm to so change its product specifications. Where is the ‘clash of wills’ in this case? The firm just wants to make profits, and will change its



Resources, power and systematic luck111

product specifications in any way that will suit that aim. Consumers want products that best suit their needs, and so will buy the products that best suit. Self-interest and the invisible hand may lead to both those desires being satisfied. Why is this social power in Barry’s sense? Linguistically, we are more likely to talk about the influence of consumer spending on boardroom decisions and this is so because managers will look at the spending in the aggregate, survey evidence, and so on. This evidence will influence them in their deliberations about product specifications. They are not fighting against those to whom they are trying to sell goods. There is no clash of wills. Consumers may have social power in Barry’s sense. They may lobby government to get new regulations that may be opposed by firms. But such lobbying is only ever effective if the consumers are organized collectively. That is a very different case, as we see below. Voters do vote in order to try to influence the formation of government. But we all know that the probability of any voter being decisive is minimal, just as minimal as one consumer switching products being decisive in a firm changing is product specification (Dunleavy and Margetts 1995). Barry once argued persuasively that success = luck + decisiveness, so each voter getting the government she wants is more lucky than powerful (Barry 1991a). Of course, we talk about the power of voters to remove governments and governments must respond to evidence over what voters seem to want if they wish to win elections. But, again, where is the clash of wills here? Parties want votes in order to help form governments and voters can make demands in order to get those votes but, contrary to Barry’s claim, this does not seem to be social power. There may be a clash of will between parties, or between different sections of the electorate, but this is not what his analogy is about. In fact, rather than showing that if consumers and voters have social power, capitalists must be socially powerful, the case for capitalists having social power seems much more secure. An important capitalist concern is much more likely to be decisive over some government decision than any voter, or even a set of voters. In order to force firms to produce different types of products or government certain types of policies, consumers and voters (or citizens) need to act collectively. There are far fewer capitalist enterprises and so collective action is much easier for them; indeed large capitalist enterprises may be privileged, in Olson’s sense, and able to act on their own to change government policy (Olson 1971).5 But might firms get what they want from government without using social power? Might they be systematically lucky? And what advantages, if any, does saying they might be lucky as well as powerful bring to social analysis?

112 Luck

The resource account Barry’s main target is a resource-based account of power, where actors’ powers are measured by the resources they can call upon.6 In order to defend the resource approach, I need to make some conceptual distinctions. First, power must be distinguished from cause. Power could simply be the infinitive form of cause; indeed in physics it is. If something has the power to do x, then under the right conditions x obtains. This can be used to examine the power of people, but we usually think of power in politics not simply as the causal properties of individuals, but rather that subset which is deliberate. Hence, in Barry’s definitions, the use of intentional expressions such as ‘desired states of the world’. Given that ‘power’ is still the infinitive, the power of an actor is what she could do, if she wished or tried. Most things we do, we could do deliberately,7 though many of these, even under the right conditions, we can only bring about probabilistically. I might hit the bullseye first shot, but my true ability to hit the bullseye will be over some probability distribution. In all situations, however, this probability will have to be measured under certain conditions that include what others are doing. Barry defines power as the ability to get what you want despite resistance. But what are the conditions of this resistance? Could it be any resistance that others might put up? If so, then no actor has any power at all, since anyone doing anything could be stopped by a coalition of other people. Every dictator remains in power because of coalitions he forges and keeps together somehow. So how, under Barry’s conception, are we to restrict the activities that others might do when measuring an actor’s power? We could restrict what others could do by restricting their preferences in some regard, and the most obvious choice would be to restrict their preferences to what they actually desire. So an actor’s power is what she could get given what others (actually) want. I suggest this will not give a very satisfactory account of the power structure, since each person’s powers rise and wane with the preferences of others, with the paradoxical conclusion that if everyone else wants exactly what you want, then you are (socially) powerless, since there is no resistance (Morriss 2002: 236, n. 49). It is for this reason, I think, that we are better advised to shift our attention to the actual resources that all actors command, and carry out our counterfactual analyses by allowing preferences to shift. In that way we can see what coalitions might form, and then bring in preferences to predict which will actually form. Bluntly, I do not think that Barry’s preferred formulation makes a lot of sense linguistically (how we ordinarily think about power in the social sciences) and is hopeless for empirical research. It is for this reason that we must turn to a resource account.



Resources, power and systematic luck113

In order to develop the resource account, I made a distinction between ‘outcome power’ – the ability to bring about what one wants – and ‘social power’ – the ability to bring about what one wants by deliberately changing the incentive structure of others.8 The latter is obviously a subset of the former, whilst the use of a second intentional concept ‘deliberately’ in the latter’s definition ensures that the concept of social power does not collapse into that of outcome power. Note that my definition of social power is both broader and narrower than Barry’s. It is broader since it carries no necessary connection with clash of wills or the resistance of others and is narrower since clash of wills does not entail ‘social power’ in my terms. For example, if the son goes to the concert despite his father’s express prohibition, the son has exercised outcome power without social power in my terms, but has exercised social power in Barry’s. Nor is Dowding ‘social power’ a subset of Barry ‘social power’, since Barry excludes from the analysis the case where an actor changes the incentives of others so they behave differently, but where those manipulated are unaware of this fact (there is no clash of wills). When governments change laws they ‘preference shape’ – they give people incentives to behave differently (without necessarily changing their wants or desires).9 When the Thatcher government made public housing cheaply available for private purchase, it shaped the preferences of the public with regard to subsidizing public housing, and made the interest (and hence mortgage) rate more important as it affected more people. This was one of many acts carried out by that government in order to try to secure a constituency of interest and lock themselves into power (Sanders 1993a, 1993b, 1996). That was an act of social power over all those affected by it and would have been so even if no one opposed it. For Barry, it was only an exercise of social power over only those who opposed the policy. I think this a distinct analytic loss in Barry’s account. I develop a resource-based account of social power, based upon but extending that of Harsanyi (1976a, 1976b). There are four types of resources in Harsanyi’s account: information, legitimacy, the ability to conditionally change the incentive structure of others and the ability to unconditionally change the incentive structure of others. Information and legitimacy are fairly straightforward. The ability to conditionally change the incentive structure of others includes threats, offers and their combination, whilst the ability to unconditionally change the incentive structure of others is to change features of the world that lead others to behave ­differently – such as preference shaping. Note that this is social power only if these activities are used deliberately to change others’ incentive structures. To some extent everything we do changes other incentives structures. Each of us is (part of) others’ structure.

114 Luck

In Harsanyi’s models, the strength of A’s power over B is measured by the strength of B’s incentive to yield to A’s influence over some action x. He modifies this for the n-person case by considering coalitions. He wants to measure the power of the coalition to get its preferred policy adopted against opposition from rival coalitions, and the power of each individual within the coalition to affect the form of the preferred policy. The model thus tries to combine the power of a group in relation to other groups with the power of the individuals within a group in relation to each other. Of course, individuals within different coalitions may bargain across coalition lines, which is what makes solutions unstable. One coalition may try to entice an individual from another if this will strengthen the former coalition. That individual will be enticed only if she can get that coalition to adopt a policy closer to the one she prefers than the policy of the coalition she leaves. Harsanyi wants to try to quantify the individual power of each member of a coalition along with that of a coalition. Some individuals have much greater power within a group than others. Their power will depend upon how they can affect other group members’ incentive structure. Their power is based upon the four elements mentioned above. These categories are deliberately broad and they include authority, persuasion, and so on, that Barry wants to exclude from the analysis of power. I can see no particular analytic gain in treating these forms of power as separate categories. The information I choose to give to you (or withhold from you) may persuade you to go one way rather than another. This seems to me an important form of power. In the army I may follow the orders of my superior officer because he is in authority, or I may follow the suggestions of my doctor because she is an authority, but it seems to add nothing to the analysis to deny that in giving orders the officer is not powerful, nor in making suggestions that the doctor is not powerful. That is not to deny that the categories of authority, persuasion, and so on are never useful, but it is to suggest that they are more usefully employed as subcategories of power rather than different concepts altogether. If we can measure all of the resources that everyone has in these four categories, and everyone knows what everyone else’s resources are, and knows that they know what everyone’s resources are (in other words, everyone has complete information), then we can predict what everyone will do. Indeed, we no longer need the term ‘power’, for it will no longer do any work. All we need for analysis are the resources. Indeed, it is for this reason, surely, that ‘power’ was so rarely used in economics until recently. The only reference I could find in an old (pre-game theory) microeconomics textbook was ‘monopoly power’, and here the use is normative. It points to the fact that monopolists are able to gain ‘excess’ monopoly



Resources, power and systematic luck115

profits. Economists do use the term ‘power’ more often now, and they do so because of the greater use of game theory. Barry says ‘Harsanyi’s analysis, like that of Dowding, actually makes power a function of the pay-offs of the parties’ (2002: 183, n. 15). That is not true. It is true that what we expect to happen is a function of the pay-offs of the parties, but then that is analytic, given von Neumann– Morgernstern utility. My analysis of power focuses on the resources that people have which lead them to behave as they do, given their expected payoff. What Barry says is true of Harsanyi’s analysis in that with complete information we no longer need the concept of power. It has been analysed to its component parts, the resources of the players, given their utility functions. So we can know what everyone will do, given their utility (understood as von Neumann–Morgernstern ‘choice-based’ and not ‘experiential’ utility) and their resources. However, when we do not have complete information, then outcomes are not determined by examining the four sets of resources alone. A key element here is informational asymmetries. It is why the fifth resource I add to Harsanyi’s account – reputation – is so important. It is this fifth element that means we cannot predict what will happen from examining the actors’ four Harsanyi resources. Bizarrely, Barry follows this reasoning entirely, but rather than seeing it is the natural development of the resource account, he uses it as a critique. Barry quotes me as saying twice that A’s power is not the same as the resources that the actor has at his disposal. He takes this to be self-criticism of my resource account. But on both occasions I am explicitly discussing Harsanyi’s account. It is because Harsanyi’s account assumes complete information that we need to add reputation as a fifth resource.10 Barry simply assumes that asymmetric information about the other four power resources provides the ‘logical’ gap between the means to some end and individual’s abilities. Though even he admits that this ‘gap’ is filled by an agent’s ‘means’ when he says ‘I shall, therefore, follow the line that equates A’s power over B with the ability to change B’s behavior and then specifies the means: the belief on the part of B that A can make him worse off’ (Barry 2002: 163). Precisely. Actors may use their reputation to get more than they otherwise would, given the other four sets of resources they have. But actors may also be treated differently because of the resources they have, even when they have no intention of using them, perhaps because they do not realize they have those resources. One may gain a reputation without realizing one has it. This is one, though not the only, source of luck.11 We might compare the resource account of power with Sen’s account of capabilities (1991a, 1992, 1999). Capabilities for Sen are what a person can get; and at times he describes a person’s capability as her ability to get what she values (1992: 30). As Morriss points out, the term ‘power’ better

116 Luck

describes what Sen means than the term ‘capability’, since the latter has usually been taken to mean an undeveloped faculty rather than a simple counterfactual description (Morriss 2002: xxiii). Indeed if one replaces the term ‘capability’ throughout Sen’s writings with the more ordinary terms ‘power’ or ability, they become much easier to read, and his arguments more straightforward. One of Sen’s aims in his work is to press for the alleviating of poverty in the world and for greater equality. Sen sets the terms of some of this debate and has pressed against the ‘welfarist’ view that what should be equalized is resources (Sen 1982c). Whilst the term ‘resources’ in the context of power here is not the same as the use of the term ‘resources’ in that debate, the underlying issues are similar. Sen would not, I think, be happy with replacing his term ‘capability’ with ‘ability’ or ‘power’, and then analysing them in terms of resources. Empirically measuring power and luck The rule of anticipated reactions suggests that B may not do something because she is fearful of how A may respond. This works whether or not A would actually respond in the way B fears, and even when A is unaware that he has the reputation that he might respond in that way. Is A powerful due to that reputation? If he has the resources to respond in the manner B fears, then he has that outcome power. If he deliberately engendered the reputation to make B fearful in the way she is and has the resources to respond in the way she fears, he is socially powerful to the degree of that reputation and those resources. If he deliberately engendered the reputation to make B fearful in the way she is but does not have the resources actually to respond in that manner, then he is socially powerful to the degree of the reputation. If he has not deliberately engendered that reputation, then he is not socially powerful. Social power, as I define it, relies upon the intention to alter the incentive structure of others. I repeat, he is outcome-powerful, however, if he could respond in that way. Social power bears more responsibility than mere outcome power, since it has an extra intentional element. If A finds everyone doing as he wishes, even without A asking, then he may be powerful. Or he may be lucky. If everyone does as he wishes because they think he is a local Mafioso, and he deliberately engendered that reputation, he has that power up to his reputation. (And as Gambetta (1993) demonstrates, sometimes the reputation is virtually all the power the Mafia have.) If he has that reputation perchance (he just happens to like wearing sunglasses and sharp suits), he is merely lucky. The causal story is different and, I take it, so is the moral responsibility. One of the problems facing the empirical researcher in examining the social structure is why some groups seem to do so well. In the UK in the



Resources, power and systematic luck117

post-war period, the farming community has seemingly got what they wanted from the government more often than would appear rational, given the power resources it commands. The farmers are a tiny proportion of the electorate. They mostly reside in safe Conservative seats and virtually never vote Labour, no matter what the Labour Party says and does. But Labour and Conservative governments have consistently favoured them. This is not the place to retell that story, but one reason the farmers have done so well is that their interests, for one reason or another, have coincided with those of the government (Smith 1990; Dowding 1991). They have been lucky. That is not to say they have no power. They are well organized and at times can lobby hard and well. Even so, they seem to get more of what they want than other groups with similar resources. To put the point Barry’s way, the farmers have got what they wanted because they did not face much resistance from governments. We can see this form of luck in the simple voting models that Barry was studying when he suggested the concept of luck. Each member of a winning coalition has, as a member of that coalition, the power of the coalition. The coalition is powerful. But each member shares that power through luck: the luck of sharing preferences with other members of the winning coalition. Together consumers may have social power (in Barry’s sense) if they can coalesce to bring about regulations in their interests but not those of firms. Production specifications change because of the outcome power of each consumer that together leads (at least in theory) to optimal outcomes for consumers and firms. Each consumer uses his outcome power (to buy a product) and as a by-product firms change their behaviour. And we can measure this power by looking at what consumers spend their money on. The case of voting is a little different. But again each voter spends her vote, and if she is lucky to share preferences she may get the government of her choice. Each vote counts as one vote, but the votes all together decide the election. The votes together are the (outcome) power of the voters, and to which the parties listen. The case of capitalists is similar. On my account of the luck of capitalists, if capitalists just go about their business – buying and selling stocks and shares – and ignore governments, then governments will still (by and large) do what capitalists want them to. Governments do so because they know that the capitalists’ business of buying and selling stock, especially in these days of globalization, will lead to disinvestments in countries with less favourable conditions than others. That is the capital market. Capitalists will behave like consumers, buying stocks and shares, whilst governments behave like firms trying to influence them through their economic policy. But that is not social power in Barry’s terms. Or it would only be so, if governments wanted to pursue policies at variance with that

118 Luck

of capital. (Perhaps capitalists have social power against left-wing but not right-wing governments?) Nor is it social power in my terms, unless capitalists buy and sell shares in order to affect economic policy. I do not think the real world works precisely like that. I do think that capitalists intervene. I do think they make threats and offers to governments. I do think they buy and sell at times in order to affect economic policy. Nevertheless, I also think that in a global capitalist economy the power of capital is even more ubiquitous. Governments will do what capitalists want simply because it is a capitalist world, and capitalists do not need to intervene as much as they do in order to get what they want. This is their luck, and it is systematic, because it attaches to the social location. It is systematic because in a global capitalist world, democratic governments’ interests are tied to those of capitalists. My account here is supposed to give greater analytic clout to the phrase ‘the power of the system’ – to explain how interests and incentives operate to systematically advantage some over others. This might be called ‘systemic power’ (Stone 1980), but I prefer ‘systematic luck’, because it shows how it can operate without the intentions of actors. It is true that capitalists’ systematic luck is based on their outcome powers as they ‘go about their business’, but I believe that the concept helps us to understand how the power structure operates. It is also true that the social power of capital gives it a ‘brooding presence’ or reputation that leads government to anticipate reactions. Perhaps, especially with regard to capital, these different effects are hard to disentangle empirically, and it is probably not worthwhile trying to do so. But then the concept of systematic luck was not designed for that purpose. It was designed rather to explain why some seem to get more than their power resources, in comparison to others, suggest they should. Capitalists are not the best example of that, but I mentioned them, so I have tried to defend the concept in those terms. What systematic luck might teach us It might be hard to distinguish whether the reputation of some political actor was gained deliberately or by chance, and so to disentangle whether anticipated reactions are due to power or luck. It might be hard to disentangle whether a government acts due to fear of deliberate response from international financiers or fear that its actions will merely elicit a response that will be bad for the economy. Actually I doubt in most cases we will ever care. In that sense the world is messy, and clearing up some messes are simply not worth the candle. Nevertheless, I do believe that doing analytic philosophy is worthwhile over all. My defence of the resource account of power is essentially practical. It allows us to empirically examine the power



Resources, power and systematic luck119

structure of urban communities (the natural home of power studies), and internationally. We gain an idea of the powers of different groups by looking at the resources relative to other groups, and to see the likely coalitions based on sets of convergent interests. I think luck, and systematic luck help us to understand why some groups seem to do better than they otherwise would, given their resources and the fact they are not part of (explicitly acting) coalitions. Normatively I think it suggests how we may help equalize power – by equalizing across the resources of different people and groups. I also think it may help us to suggest that other tactics that crude power games may be more worthwhile in the long run. Since September 11th the USA is pursuing a policy (so it says) of wiping out international terrorism. It is targeting both people and countries. By and large, force is the means. It is the most powerful nation on the planet so that might seem to be the most obvious way of fighting those who have different preferences. However, persuasion and preference shaping, might, in the long run, be a better policy. If actors share preferences they are more likely to work together. This increases their power as a coalition, but does not increase their individual power; rather the increased coalitional power of the actors is a result of their luck in sharing preferences. What can this teach us about power more widely understood? One way in which actors can try to make themselves systematically lucky is to change the preferences of other actors. Persuasion, or winning rational arguments is one way, but the use of unconditional incentives is the most powerful. For the West, reducing world poverty, brokering to help secure peaceful solutions to long-term local conflicts, trying to limit the influence of anti-liberal forces, encouraging the development of liberal democracy, in other words trying to change the preferences of peoples and their governments, may be of much greater long-term benefit, than the use of force. In this sense, using one’s power to create future luck may be the best form of power to have. Notes  1 My use of the term ‘luck’ is slightly different from Barry’s, as pointed out by Peter Morriss (2002: xxxvii).  2 Here I have removed from the original essay two paragraphs identical to paragraphs in Chapter 4.  3 This Weberian definition follows his earlier work.  4 See Barry (2002: 161), for the claim that consumers, voters and capitalists have social power in Barry’s sense.  5 The ease of capital organizing in relation to workers, voters and consumers and what this means for the structure of society was pointed out by Claus Offe (1975).  6 Barry originally held a resource-based account (1991a, 1991b), then changed

120 Luck

 7  8  9

10

11

his mind when reviewing Morriss’s Power (Barry 1988), changed it again in a review written for the publishers of Dowding (1991), and has now (Barry 2002) changed his mind a third time. Given that Brian Barry is not renowned for changing his mind easily, this perhaps shows the inherent difficulty of coherently conceptualizing power caused by linguistic imprecision. I hope to change his mind a fourth time. Some things we do ‘accidentally’ we might not be able to do deliberately. I might accidentally burn my hand in a fire, but be psychologically incapable of putting my hand in deliberately. I described these definitions as ‘ostensive categorical distinctions’ in order to make the point they are not natural kinds, merely definitions helpful for empirical analysis. Preference shaping should be distinguished from persuasion. I may get you to eat more boiled than fried potatoes by giving you information about their respective effects on health or by putting a tax on cooking oil. The former changes preferences by persuasion, the latter by shaping. The former does not change your material conditions, just your beliefs; the latter changes the material conditions under which choice is made (Dowding 1991: chs 3, 7; Dunleavy 1991: 117–44). In fairness to Barry, that reputation is a fifth resource is not as explicit as it should be in Dowding (1991), but it is explicit in Dowding (1996); see also Dowding (1995: 146). Barry in a private communication suggests that one cannot use ‘reputation’ as a resource ‘since it refers to beliefs about what use people will make of their resources (in the other senses)’. But it is strange to say that trustworthiness, honesty, and so on are not resources of people in a community. Social capital is surely a property of individuals as well as being a relationship between them. Barry’s favourite philosopher also lists reputation as a power resource: ‘Reputation of power, is power; because it draweth with it the adherence of those that need protection’ (Hobbes 1651/1982: Part 1, ch. 10, 150). I suspect therefore that power must not only be analysed but also conceptualized within a game-theoretic and not a game-form framework.

6 Shaping future luck

Introduction Power is a disposition, analysable ‘counteractually’ by taking into account possible preference changes. Being powerful is getting what you want, but possible resistance is also built into the definition. This analysis leads us to a trivial truth: that no single person nor group can really carry out its own political agenda. In other words, no one has real or absolute political power. This trivial truth leads us to an important one. Every powerful actor is powerful because of the resources they bring to a bargain with other actors. Social power always depends upon a coalition of mutual or allied interests. Dictators rely upon many other people: their army, police, secret police, their cabinet, and so on. All of these people, or some subset of them, could conspire and overthrow the dictator. Dictators survive by forming coalitions with others and stopping rival coalitions forming by sowing doubts in others’ minds and turning potential partners against each other. The dictator offers positive and negative incentives to the others in order to gain their support and stop their challenge. In order to understand even the most obvious examples of social power, we need to understand the nature of coalition formation and the nature of bargaining. It can be deduced that all political power is a form of reciprocal or bargaining power. Studying individual or group power involves understanding the resources that these actors possess. These resources include both ‘external’ resources – money, legal rights, institutional authority– and ‘internal’ resources – physical strength, determination, persuasiveness. My account of power (Dowding 1991, 1996), based on the bargaining models of Harsanyi (1976a, 1976b) and later non-cooperative game theory, suggests that we need to study the resources of different groups in society, understand their preferences and model their relationship with one another. In this, five sets of resources are important: (1) knowledge or information, (2) legitimate authority, (3) unconditional incentives to affect

122 Luck

the interests of others, (4) conditional incentives to affect the interests of others, (5) reputation. Resources of power Information and knowledge Having information and knowledge is an important resource. Collecting information can be a costly business and supplying information often gives groups power over others. Governments rely upon their bureaucracies to supply them with the information necessary to produce and implement policy, and public bureaucracies rely upon professional groups, lobby groups and other organized groups and industries to supply much of that information. An important part of empowering citizens is not only a free press, but an investigative one. One might think that collecting information is not a cost for journalists, since it is part and parcel of their job, but that job becomes easier if information is supplied. This is why industries, lobbies, political parties and governments employ public relations consultants to try to ensure that the press gets the information they want them to have. Sometimes the group’s machine can be so effective as to rebound. When Shell wanted to dump the Brent Spar oil rig in the Atlantic Ocean, Greenpeace rallied public opinion against the company and the British government, forcing Shell, much to the Conservative government’s dismay, to back down. Later, the press felt that it had been manipulated by Greenpeace, through being invited onto their ship to share the excitement of the activists as they took on the oil company in mid-ocean. Journalists came to believe that the oil company had some good environmental arguments for Atlantic dumping as the least damaging option, which were not effectively conveyed. The BBC brought in new guidelines for its journalists because of this case. Legitimate authority Legitimate authority is obviously a source of power. To have the force of law on one’s side leads others to be more likely to comply. Authority also comes with knowledge. Professional groups often speak with authority on issues where they claim specialized knowledge. Thus doctors claim authority on health issues, teachers on educational affairs. When the Conservative government wished to introduce radical reforms in Britain’s schools against the dominant wishes of the teaching lobby, it felt the need to break down the authority of the main teachers’ lobbies. It attacked the power resources of the educational establishment, suggesting that ­teachers’ unions were led by Marxists, that teaching policy in schools was



Shaping future luck123

driven by ‘trendy’ ideas created by teacher training in universities set up during the heyday of left-wing ideological hegemony, and that in the end parents know best. These tactics were typified by the Secretary of State for Education, who launched a slanderous assault upon a leading educationalist (subsequently appointed by Blair’s government to an important educational policy role), describing him as a ‘nutter’ and claiming to ‘fear for Birmingham with this madman let loose, wandering around the streets, frightening the children’. Clearly, this strategy was designed to undermine the legitimacy of teachers, their unions, teaching establishments and educationalists to speak with a professional voice on the care of children, their major power resource. Politicians continue to utilize the undermining of professional legitimacy in disputes with other professional public services, such as the health service and universities. Unconditional and conditional incentives The third and fourth sets of resources are less obvious. They are the unconditional incentives to affect the interests of others and the conditional incentives to affect the interests of others. One actor may provide another with a number of unconditional incentives that affect the second’s calculations about his actions. They are unconditional in the sense that the second person bears the costs or receives the advantages no matter what he does. For example, government top-sliced resources from the central government grant to local authorities for urban regeneration schemes chosen through competition. This led councils to create many public–private partnerships that go against the preferences (or ideology) of many councillors. Unconditionally changing the incentive structures of actors alters their behaviour. Conditional incentives are more straightforward. They are threats or offers, or throffers (a combination of threats and offers), which take the form ‘if you do x, then I will do y’. There is an important asymmetry between threats and offers: offers cost more when they succeed and threats cost more when they fail. Increasing an offer is likely to increase the probability of incurring the costs of success; whilst increasing a threat, as long as it remains credible, is likely to decrease the probability of incurring the costs of carrying it out. Keeping threats credible provides an important limitation on the degree of the threat. Schelling (1966, 35–6) argues that some threats are inherently credible, other inherently incredible. If there were no limitations on the credibility of threats, then it would always be worthwhile to make the threat as great as possible. One way of making a threat credible is to pre-commit oneself to carrying it out if the subject does not comply. Governments may pre-commit the

124 Luck

state by passing legislation entailing heavy punishment, though even here, if the punishment seems out of proportion to the crime, juries may refuse to convict and judges try to find mitigating circumstances to reduce the penalty. Attempts to pre-commit are another form of threat; for example, a trade union balloting its members for strike action if management does not meet certain conditions. Reputation Making credible threats and carrying them out is very important to reputation. If one gets the reputation for making idle threats, one loses all credibility. The business of the Mafia is providing protection, and Diego Gambetta has shown how important reputation is for the Mafia and how cheap protection can be if one’s reputation is good: By far the most striking feature of a mafioso’s reputation is that it saves directly on production costs. Car manufacturers benefit from a good reputation, but they still have to produce cars. By contrast, a reputation for credible protection and protection itself tend to be one and the same thing. The more robust the reputation of a protection firm, the less the need to have recourse to the resources which support the reputation. (1993: 44)

The Mafia gains its reputation through deliberate acts it carries out, even if it does not in fact have the resources that those who fear it think it has. It is also possible to gain a fearsome reputation even if one does not intend it. Someone may be thought of as a Mafioso simply because he wears sharp suits and sunglasses, never smiles and has been overheard denying the existence of the Mafia. The ‘rule of anticipated reactions’ (Friedrich 1941) suggests that people will not press demands if they feel they will not get anywhere. This rule may work in favour of some groups, either because they have acted to engender this reaction in others or, though less often, through luck. Others believe that they will react unfavourably to some demand, and so do not make the demand, when in fact the response may be rather different. Reputation is not only important to those who make threats, it is also important to those who may suffer them. We cannot simply measure the value to B of A’s threat to her. If B acts to maximize her expected utility, then the action she should choose in the absence of A’s threat will have a higher utility than the action to which she is coerced. If t1 is the measure of the disutility to B of the sanction A threatens, then the threat will be successful if t1 > u1 − u2, where u1 is the utility to B of carrying out her preferred course of action without the threat and u2 the utility of carrying out the coerced action. The difference u1 − u2 that A can make to B’s welfare is the measure of A’s power. In fact, A does not necessarily have that much power. In the



Shaping future luck125

same way that one may wish to develop a reputation for toughness in a repeated Chicken game, one may find it advantageous to refuse to comply with threats in order to develop a reputation for stubbornness. There will be costs to A of carrying out his threat, and thus he will want to threaten those who are most likely to comply. Hence there are incentives to B to develop a reputation as someone who does not comply with threats. Reputation is a key element in bargaining and game theory. It depends on players having incomplete and imperfect information. If players had complete and perfect information, there would be no room for reputation, for there would be no room for players to pretend they were anything other than what they are. Thus the importance of information arises from asymmetry in information. It shows that we cannot simply read off actors’ power from their resources – the so-called ‘vehicle fallacy’ (Morriss 1987: 18) which equates power with its vehicle – and shows that determinate game-theoretic accounts are impossible. It also demonstrates that individuals’ preferences and their power can be formed by the way they play the game, rather than the game being played by them given their preferences and resources. We should also note that stubbornness can apply to offers as well as threats, and indeed to Harsanyi’s first three resource categories (Dowding 1991: 77–9). Power and luck In a famous article Brian Barry (1991a) asked the question: is it better to be powerful or lucky? The obvious answer is both; but Barry suggested that in many ways, being lucky is better than being powerful. Of course, in an academic context, the concepts of both ‘power’ and ‘luck’ take on specific meanings which may not correspond completely with ordinary usage. Barry defines power as ‘getting what you want despite resistance’ and luck as ‘getting what you want without trying’. This distinction is illuminating when trying to model the power of actors in different social situations. Barry’s argument was developed in a critique of applying cooperative theory to measuring the power of actors in committees (committees in this context means any body which makes decisions by some voting rules), which has been extensively applied by political scientists to parliaments around the world – for example, to the relative power of EU nations through decisions in the Council of Ministers. There are various different indices designed to measure power, but we do not need to go into their differences (see Felsenthal and Machover 1998). It is sufficient to note that power indices are constructed by assigning a score to each voter v for all winning coalitions C in which v is pivotal (or decisive) where C − v does not constitute a winning coalition. Thus your power in

126 Luck

a voting game is defined by the number of winning coalitions you could create, or destroy, by which way you vote. The scores are then normalized so the sum of the power of all voters equals 1. The justification for this is that power in voting games is zero-sum. The winners win what the losers lose. The power of each nation in the Council of Ministers for example, or the power of coalitions of nations such as ‘large nations’ or ‘poor nations’, can then be calculated given the weighted votes of each and the precise rules governing what constitutes a winning coalition. This is of obvious interest since, for example, it has been argued that the largest member states (Germany, the UK, France and Italy) have more power when the blocking majority is 27 than when it is 23, making John Major’s attempt to stop EU enlargement in 1993 irrational (Johnston 1995). Other writers have examined the actual and potential voting rules to see the relative powers of nations in terms of the GDP, population and other factors, to question the wisdom of the actual voting rules which exist in the Council (Widgrén 1994; Lane and Mæland 1995, 1996; Hosli 1996). Similar applications of the indices have been made to nations and political groups within the European Parliament (Herne and Nurmi 1993; Lane et al. 1995; Hosli 1997). As far as this analysis goes, there is nothing wrong with using power indices in this manner. They can be used to demonstrate the relative voting power (or voting resources) of voters or groups of voters within voting assemblies. However, it is a mistake to confuse voting power, which is the difference an actor can make in getting what she wants, with the probability that some actor will actually achieve what she wants. Actors may also achieve what they want in voting games without needing to vote at all, thereby getting what they want without trying, or may achieve what they want when their vote is irrelevant since their preferences can be secured by a majority without their help. This then is luck (Barry 1991a; Dowding 1991, 1996). In other words, it is a mistake to confuse getting what you want with your power. Let me give an example. Take the simplest power index (Shapley and Shubik 1969) and a simple case. The Shapley–Shubik index calculates the voting power of each player by imagining the vote for or against any proposal is taken in sequence. The power of each actor is then the summation of the number of times their vote is the vote which secures a majority, divided by the number of possible ordered sequences. In other words, the power of any given voter is the probability that that individual is the final member of a minimum winning coalition. Consider the decisiveness and luck when there are three voters and three options. Each voter may vote for one of the options. For the purposes of calculating the decisiveness of each voter, there are 27 separate votes possible: nine when voter I votes a and II and III vote a or b or c; nine where I votes b



Shaping future luck127

and voter II and III vote a, b or c; and nine where voter I votes c, and voters II and III vote a, b or c. In this case, the pivot is the voter, always either II or III who is the second voter, in the Shapley–Shubik sequence, to vote a when I does, b when I does, or c when I does, assuming voter I votes first. When I votes a, II is pivotal three times, and voter III four times, twice when voting a with voter I and twice when voting either b or c with voter II. Similarly, when voter I votes b and when he votes c. Thus, when voter I votes first, he is never the pivot, but II is pivotal nine times and III pivotal twelve times. There is no pivot six times. Figure 6.1 sets out the possible voting combinations. However, the three voters win more often than just when they are pivotal. Voter I wins as often as voters II and III, despite never being the pivot. Each finds himself in the majority 15/27ths of the time. Voter II is a winner without being the pivot six times, and voter III three times. Hence in the full sequence each wins 48 times out of 162. Again, there are 38 out All combinations of votes when I votes first, II second and III third Voters (order) Policies

I

II

III

I

II

III

I

II

III

a a a a a a a a a

a a a b c b c b c

a b c c b b c a a

b b b b b b b b b

b b b a c a c a c

b a c c a a c b b

c c c c c c c c c

c c c a b a b a b

c a b b a a b c c

Notes Bold indicates successful, underline the pivotal vote. I never pivots, II pivots 9 times, III pivots 12 times. Each voter is successful 15 times. 6 votes, no result. So I is lucky 15/27; II decisive 9/27, lucky 6/27; III decisive 12/27, lucky 3/27. If 6 votes ‘arbitrarily’ assigned to voter I as ‘dictator’ (‘pivot’), then: I pivots 6 times; II pivots 9 times; III pivots 12 times: I is successful 21 times; II and III successful 15 times. So, I decisive 6/27, lucky 15/27; II decisive 9/27, lucky 6/27; III decisive 12/27, lucky 3/27. But order of voting is arbitrary, so the ‘decisiveness’ of I is luck, so reassign: I is lucky 21/27; II decisive 9/27, lucky 6/27; III decisive 12/27, lucky 3/27. There are six orderings of this nature: (I, II, III), (I, III, II), (II, I, III), (II, III, I), (III, I, II), (III, II, I). So each is decisive = 42/162 = 0.26 lucky = 60/162 = 0.37 successful = 102/162 = 0.63

6.1  Three voters, three options

128 Luck

of 162 occurrences with no winner. Thus each voter is successful 90 times, and can claim to be decisive 42 times, and thus lucky 48 out of 162 times. If we assume (in order to keep the game zero-sum) that each possible vote must have a winner, then we must arbitrarily assign the 36 occasions when there is no majority. Let us assign it to the person who votes first (voter I in the sequence portrayed in Figure 6.1), thus making each decisive on 54 out of 162 occasions. However, claiming that a voter is decisive because we have arbitrarily imposed a winner which happens to coincide with his choice is an odd use of the term ‘decisiveness’ in Barry’s formulation. The ‘order’ of voting is just a device used by Shapley and Shubik in explaining the mechanics of their index. It makes much more sense to say that the voter is lucky rather than decisive when their choice is arbitrarily imposed. Hence, each is successful 102/162 times (0.63), decisive 42/162 times (0.26) and lucky 69/162 times (0.37). One gets what one wants through luck more often than through decisiveness. This is true of the simple three-person case. Luck shoots up in probability as we add more people, since it is far less likely that one’s vote will be decisive. For example, in the USA, if there are two presidential candidates, and the probability of each voter choosing one or the other is equal, then each voter’s probability of being decisive has been calculated as 1/16,666 (Mueller 1989). However, even this small probability of being decisive is misleading, since the maximum probability of being decisive occurs when the probability of voting for one or other candidate is equal across all voters, but the gradient around the function is very steep. If the expected probability of the vote for one of the candidates is 0.499, then the probability of a voter being decisive falls to 1/1090 (Carling 1995). This makes clear the obvious point that the citizens who get the president they want are lucky not powerful, whether or not they bothered to vote for him. What do we learn from this? The decisiveness and luck of an actor varies according to the preferences of other actors, but an actor’s power remains the same. In the context of voting games, we can see that the probability of one’s preferences being in line with those of the majority (assuming majority vote) is more crucial than any actual voting power. If one is always in the majority, then one always gets what one wants – but through luck not power. How likely is it that one’s preferences will agree with the majority? If luck is evenly distributed, then the probability – for simple majority votes with two alternatives may be 1/2. Yet why should luck be evenly distributed? Only if preferences are random. In fact, most of our preferences are in some way structurally determined (or what I have called ‘structurally suggested’: Dowding 1991, 1996). We have them because of the sorts of people we are. Thus some people may be ‘systematically lucky’. They get what they want because of the way society is structured. Their interests are secured because their



Shaping future luck129

preferences are non-randomly related to what society chooses. Their luck is attached to their social location. Let us return to the EU example. One of the major criticisms of the powerindex approach, as it has been applied to power in the Council of Ministers, is that it ignores ‘natural coalitions’ by assuming all possible coalitions are equiprobable. If actors share preferences they are more likely to vote together. This increases their power as a coalition, but not their individual power; rather, the increased coalitional power of the actors is a result of their luck in sharing preferences. Where those coincident preferences are not mere happenstance, but depend on structurally induced preferences, such as ‘being relatively poor’, ‘having large agricultural interests’, ‘northern’ versus ‘southern’ states, ‘large’ versus ‘small’ states, and so on, then we may see this luck as ‘systematic’ or structured by exogenous factors. When constitution building, such systematic preference equivalences may need to be taken into account, in order to defend what may otherwise be persistent minorities (Grofman et al. 1992). Attempts to incorporate connectedness into power-index calculations (Lane and Mæland 1995; Widgrén 1994) confuse the probability of coalitions forming with the normalized voting power of the constituent members of the coalitions. The probability of one’s preferences lining up with others in natural coalitions is separate from one’s voting power. Being central in policy space is some form of luck. It may lead to having greater power, for being pivotal in this sense may then enable a country to trade votes in one issue area for votes in another. But that is a separate issue. Preferences and the creation of luck Rational choice theory has traditionally operated with the assumption of fixed preferences (Stigler and Becker 1977), though what this means has also traditionally been misunderstood. The fixed preferences of Stigler and Becker are simply preferences defined as what people want. What people want is determined by their relationships to one another and where they are in their life cycle. Their preferences, in other words, are completely structurally determined. Thus as people get older their desire for sports equipment may decline and their need for health-care products may increase. Small firms prefer few regulations, larger companies may prefer strict regulations simply because their size gives them a comparative advantage in the market. Preference, in other words, is a relationship between variables defined in terms of a particular use of the rationality assumption: the instrumental rationality of choosing the best means for your ends, no matter what those ends are. Your self-interest leads you to prefer whatever is functionally best

130 Luck

for you, given your position in society. In this sense, your preferences do not change, just your position. To critics it is obvious that people change their minds, and new desires and wants are created. Thus critics find it strange when rational choice writers suggest preferences are fixed. But they mean fixed in this manner. Examining preferences in this way allows us to model changes in preferences which are structurally determined by the position occupied in society. We can then begin to understand how actors can manipulate the preferences of others in order to bring them into line with their own preferences, creating their own luck and systematic luck. When I write of ‘changing preferences’, then, I mean changing the social position of individuals rather than a person changing their mind about an issue due to new information – for example, deciding that GM foods pose potential future health risks and therefore shifting from a position of support to one of opposition. This is not to deny that controlling information is not a power resource, nor that such shifts in beliefs are not important, but rather to indicate that they will be examined not on an individual but a socialstructural basis (see below). There are various means by which governments, particularly, manipulate the structure of preferences in society. Dunleavy (1991) has examined these strategies most extensively, calling them ‘preference-shaping’ strategies. Governments use social expediency, exploiting events to bolster their own support, most obviously when they are involved in external conflict and use nationalism and patriotism to rally the country around the flag. They are also able to adjust social relativities through the pork barrel, diverting resources to secure their key supporters and trying to buy support at the margins. But such actions are contingent and achieve only short-term gains. Longer-term action is required to secure future luck, especially future systematic luck. Here governments need to make longlasting and favourable changes in the social structure, using three main means: manipulation of institutions, manipulation of strategic behaviour and changing individual preferences. Manipulation of institutions One way of making such changes to the social structure involves changing not the preferences of any one individual, but instead changing institutions so that the same social geography leads to different outcomes. Changing the way in which we aggregate preferences by altering electoral systems, gerrymandering constituencies, devolution or shifting responsibilities to and from different tiers of government are all well-known means by which institutions can be manipulated to change the structure of future coalitions, ensuring that the relative levels of luck of different social groups



Shaping future luck131

alters. One may also alter outcomes not by changing the institutions themselves but by changing their composition. Encouraging some groups into different local authority jurisdictions and encouraging others to leave can shift the balance of power within different local governments. British housing policy since the end of the First World War can be seen in this light, as Labour councils saw the advantages of public housing, whilst Conservative governments wanted to disperse council stock and encourage greater numbers of private homeowners, their more natural supporters. Local welfare policy can also have the effect of creating large-scale mass movement for economic benefit, shifting political fortunes (Cebula 1979; Cebula and Koch 1989). British housing policy also illustrates the ways in which the size of different social groups can be shifted. By cutting subsidies to public housing rents, reducing the ability of councils to build new housing stock and repair old stock and, above all, by its ‘right to buy’ policy, the Thatcher government shifted a mass of council tenants into the private housing market. The creation of Housing Action Trusts made it more difficult for elected local authorities to manipulate housing policy locally. By the end of the 1980s housing tenure was more correlated with partisan choice than social class (Dunleavy 1991). Manipulation of strategic behaviour Social choice makes a distinction between ‘naive preferences’ and ‘strategic’ or ‘sophisticated preferences’. Voting naively is to vote for one’s most preferred candidate no matter what her chances of winning are. However, if one understands the workings of a given electoral system, and knows the preferences of the other voters, one can try to ensure oneself a better outcome by voting in a more sophisticated or strategic manner. In a plurality vote election, for example, voting for the Liberal Democrat candidate rather than the Conservative in a safe Labour ward may increase the chances of getting one’s second most preferred winner. More complex voting systems are more open to such strategic manipulation. One may be able to manipulate such strategic behaviour by persuading voters that it is optimal. During the 1990s the Labour Party accused the Liberal Democrats of creating false polls showing them running second in by-elections to the Tories for precisely this reason. But support for non-democratic regimes may also be bolstered by such manipulation. Timur Kuran (1995) deals with preference change through the manipulation of preference falsification in closed and authoritarian societies. Preference falsification is the expression of false preferences, motivated by the fear of social disapproval which may be incurred by expressing privately held views. This type of preference falsification may lead everyone

132 Luck Individual utility

Total utility Reputational utility Intrinsic utility Public preference

6.2  Public and private preference

to falsify their preferences, though in fact they share private beliefs. Figure 6.2 shows how one may falsify one’s true private preferences in order to maximize one’s overall utility through expressive and reputational utility, which is dependent upon how others sees one, rather than through one’s own intrinsic utility. But if everyone is misrepresenting their own intrinsic utility in order to gain reputational utility, then what everyone believes to be the true utility of society is in fact driven by everyone’s desire for reputational utility. Democratic conformism may lead people not to express views because of fear of the censure of others, even where those others privately share the nonconformist beliefs. Anticipating the censure, each embraces the majority view, censuring those who do not. Whilst some may gain greater expressive utility than others and fear censure less, they too may go along with the majority view when the majority is large enough. Kuran’s formal model in fact assumes that everyone will go along with the majority view if they are alone in their beliefs. Manipulating people’s views about what others believe may thus have a strong effect upon the way they behave in society. Such processes may lead people to change their private views too. Changing individual preferences As mentioned above, individuals’ preferences for different outcomes may be affected by new information about the world. Knowledge is power, and control over information is an important power resource. But structurally altering individuals’ preferences is a means by which actors may try to ensure their future coalition partners and systematic luck. Tying households’ preferences over the economy to the mortgage rate by encouraging homeownership means that the interests of capital become the interest of



Shaping future luck133

everyone. Households with large mortgages have a much greater i­nterest in the state of the financial markets than they do even in the rate of pay rises, since a 1 per cent drop in the mortgage rate will bring greater financial rewards for many people than a 5 per cent pay rise. By tying a set of interests to those of your own, you can gain coalitional advantages by creating, in a systematic manner, your own future luck. Conclusion We can see from this type of analysis that getting what you want may rely upon ensuring others’ preferences fit into a set of interests aligned with your own. We learn that to be lucky, or systematically lucky, to nonrandomly share preferences with the majority, or with the most powerful is as important, if not more so, than using one’s resources. One of Britain’s problems in dealing with its European partners has been that it is out on a limb in many respects, as an island nation on the periphery of Europe, with traditional trading links with the Commonwealth and the ‘special relationship’ with the USA which has non-randomly given it different interests, especially from the other three largest nations, with which it could otherwise form a very powerful coalition. One way in which actors can try to make themselves systematically lucky is to change the preferences of other actors. Persuasion, or winning rational arguments, is one way, but the use of unconditional incentives is the most powerful. For the West, encouraging the development of capitalism and liberal democracy and trying to limit the influence of anti-liberal forces in order to change the preferences of peoples and their governments will be of much greater long-term benefit than any other form of influence or use of power. In this sense, using one’s power to create future systematic luck is the best form of power to have.

7 Luck, equality and responsibility

Talk about ‘differences that are a matter of luck’ and ‘differences for which people are not responsible’ and people being ‘worse off through no fault of their own’ has been engaged in uncritically, without adequately examining what it could mean. (Hurley 2003: 204)

Introduction Responsibility has become a central issue in egalitarian debates. Many egalitarians argue that the only justifications for inequalities in society are those which result from the choices responsibly and freely made by individuals. If one person is less well off than another, not due to the choices each makes, then the first has a legitimate claim of injustice. The label ‘luck’ has been attached to the generation of inequalities that do not derive from choice responsibly and freely made. So luck is the other side of responsibility. For luck egalitarians, all inequalities deriving from luck should be expunged from the just society. Or, rather more carefully, they feel that inequalities deriving from brute luck should be expunged from the just society.1 Taking responsibility for the outcomes one receives has long been a clarion call for those on the right of the political spectrum. For egalitarians to give a central role to responsibility allows them to avoid the charge that equalizing outcomes will deny responsibility by removing the expression of individual agency. Distinguishing the vagaries of life over which we have no control (brute luck) from those which we do (option luck) is the route some egalitarians have taken to overcome such a critique. After all, it is easy to demonstrate that in today’s society inequality does not map on to responsibility (Barry 2005). I will argue, however, that the neat trick of taking responsibility for outcomes out of the right-wing camp does not work as well as it might for egalitarians. The distinction between brute and



Luck, equality and responsibility135

option luck is not easy to maintain outside of simple examples, and redu­ cing brute luck can reduce the scope of individual responsibility. Rather, we need to decide what levels of inequalities are acceptable and let agents act within those bounds. Considering how unequal society should be, given responsibility for outcomes, requires a decision over the object of equality. I do not discuss that difficult subject here, though I note that some potential objectives sit less well with the discussion of responsibility and luck than others. In what follows, I tend to assume that the objective of equality is material benefits (money). Despite sophisticated discussion over the ‘equality of what?’ question, most of the examples given in the literature concerning luck and responsibility seem to rely upon material benefits as the object of equalization. I point out what broader accounts might entail and show that moving beyond material benefits only exacerbates the problems I identify. Responsibility In fact, I am not intending to discuss in any depth the thorny issue of responsibility. I shall discuss what we might mean by ‘luck’ in some detail, as I believe that concept has been taken somewhat for granted. One of my aims is to argue that we cannot distinguish luck and responsibility in either theory or practice to the degree theorists have assumed. I take a minimal approach to responsibility. A great deal of the literature on responsibility deals with the issue of free will and determinism. I am not concerned at all with this literature. The approach to responsibility I adopt is compatible, I believe, with either a free will or a deterministic approach to responsibility (that is, a deterministic approach which still allows the assignation of responsibility to people: see, for example, Dennett 2003). I assume we assign responsibility to people dependent upon how much control we believe they have over their course of action. Bracketing out the freewill–determinism issue allows us to suggest that people have more control over the outcomes dependent on their actions based upon (a) the number of viable elements in their opportunity set and (b) the relationship between the actions they perform and the outcomes those actions lead to. On the first aspect, it is obvious that if an opportunity set has only one element in it, we cannot assign responsibility for ‘choosing’ it. However, it is not clear that we assign more responsibility for any chosen element the more elements there are. In the formal freedom of choice literature, the elements in an opportunity set are simply given. However, in real life there may be alternatives available that people feel ought to be chosen,

136 Luck

but the costs are such that they choose another option. When assigning responsibility for actions we take account of the potential costs of actions as well as their likely outcomes. We assign responsibility given the number of viable options in an opportunity set. However, individuals must be able to distinguish the alternatives. If there is no diversity among them, then we cannot expect a person to accept responsibility for any particular choice. If the alternatives all appear ­identical – matches from a matchbox – then people will pick rather than choose (Ullmann-Margalit and Morgenbesser 1977). To the extent that there are no reasons for picking one rather than another, we cannot assign responsibility for one alternative (such as a match from a box) being picked rather than another. Diversity here needs to be an objective characteristic of the alternatives. When challenged over why one has chosen one course of action rather than another, it is not sufficient to say that one is indifferent to each, since to oneself they are indistinguishable. If one course of action affects others greatly, then responsibility can be assigned to the picker notwithstanding her indifference to the outcomes. Not being able to distinguish the alternatives can be offered as an excuse against responsibility for outcomes if one did not have sufficient information. Not being aware that one course of action would deleteriously affect others can be an excuse against being held morally responsible for the outcome especially where, given the circumstances, one cannot be expected to have had that information. The importance of information shows the limits to the number of options in an opportunity set. Assuming that there are costs to collecting information, then an optimal opportunity set will be finite. If there are too many alternatives or there is too much information to process about the alternatives in an opportunity set, rational choice becomes more rather than less difficult (Dowding and van Hees 2008b). Again, the amount of information that we can expect someone to have processed in a decision is important for the assignation of responsibility. Finally, there is the relationship between the actions we perform and the outcomes that result. Again, in the formal literature on freedom of choice, the elements of an opportunity set are thought to be well defined, in the sense that choosing them leads directly to that element. Action and outcome become the same thing. In real life, however, the actions we choose in order to bring about an outcome and the outcome itself are not so directly correlated. How far we assign responsibility to someone for the outcome itself depends upon how we view the relationship between the action and the outcome. We may not blame someone for a bad outcome given what she (and we) might have expected to happen given the actions taken. This leads us back into the idea of luck.



Luck, equality and responsibility137

Three types of luck Dworkin (1981, 2000) first made the distinction between brute and option luck in relation to his insurance market justification for equalizing resources. One way of making the distinction is expressed by John Roemer (1996: 248): Option luck is the outcome of a gamble explicitly taken, while brute luck is an outcome in which no gamble was entered into. Being struck by lightning when no insurance was available against that calamity is brute luck: being struck by lightning if insurance was available, whether or not you were insured, is a matter of option luck … Thus the presence of insurance markets transforms events of brute luck into events of option luck. Dworkin’s view is that it is fair for persons to suffer the consequences of option luck, for persons decide how much to insure against those kinds of event. Brute luck, however, is a morally arbitrary (and hence unfair) way of distributing resources.

The basis of Roemer’s distinction is that someone cannot complain about the bad consequences that may follow from a risk that one has chosen to take. After all, failing to insure against some bad outcome allows one to keep the money that would otherwise have been invested in the insurance. An attempt to gain advantage over those who do insure has been thwarted by events, but events that one can predict with some probability.2 Brute luck is not normally used in quite this manner, however. Rather than describing chance events for which there was no prediction or insurance, brute luck is often used to describe the conditions in which one finds oneself from birth. This is a necessary condition of life from each person’s point of view, though the distribution of life chances in any society can be known, given other known facts about that society. I will make a distinction between two forms of brute luck: ‘contingent brute luck’ describes events as in the quotation from Roemer and ‘personal identity brute luck’ (or more simply ‘personal identity [PI] luck’ (Dowding 1991: 64–5) describes the conditions one has from birth. Of the second, Cohen (1995: 229) suggests ‘no luck is bruter than that of how one is born, raised and circumstanced’. Some are born with different qualities (such as beauty or the capability to develop certain types of skill or talent), or with disabilities; and people are born into different positions in the social and economic structure. To the extent that the position one is born into naturally endows (dis)advantages – be they institutional, economic, social, educational, and so on – these constitute PI luck. Some (dis)advantages might be inherent in our very genetic make-up – physical or mental impairments, beauty, and so on.3 They might be attached to someone of any social class. Others might come with the class – but in either case they constitute PI

138 Luck

luck. Contingent brute luck is constituted of the events which occur during one’s life not correlated with social circumstances that (dis)advantage one. These are completely outside one’s control – the kind of simple bad luck that might hit anyone, such as being struck by lightning or contracting some disabling disease.4 We should note in passing that whilst contingent brute luck can be turned into option luck with the right insurance scheme, it requires the insurance scheme to be operable and in operation, otherwise option luck is to contingent brute luck as Kaldor–Hicks compensation is to no compensation. A bad day at the office matters when the work that day is an examination for promotion. And, even then, identical outcomes do not follow from identical efforts when the normal bounds of marking entail that the same scripts might have reversed marks on different occasions. I leave aside these distinctions between types of luck to be returned to below. First I want to interrogate the notion of luck itself. What precisely do we mean by ‘luck’? Interrogating luck In contrast to ‘responsibility’ and ‘choice’, there is relatively little writing devoted to the precise meaning of the concept of ‘luck’. Perhaps this is because the idea of ‘luck’ seems straightforward. In the egalitarian literature luck is usually thought of as those events which are outside of one’s control. However, luck cannot straightforwardly be defined in terms of control for three reasons. First, ‘control’ is not a binary relation. There are no events which are completely within our control, nor any completely out of our control. Even events we ordinarily think of as being completely in someone’s control occur with some probability between 0 and 1, even if that probability often approaches 1. Other events considered to be in our control occur with probabilities much less than 1. Events often do not turn out entirely as we intended, even when we consider ourselves to be successfully controlling them. Second, most, if not all, of our efforts require joint action – if not others actively cooperating with us, then at least not actively opposing us. Third, not all events outside of our control are normally considered to be part of our luck. The sun rising each morning is outside of our control, but we do not normally consider ourselves to be lucky each daybreak (Pritchard 2005). It might seem obvious that luck has something to do with probabilities. And in the case of option luck, one could have insured against the (low) probability of calamity, but one preferred to keep one’s money. If the calamity had not befallen one, then one would gain advantage over those



Luck, equality and responsibility139

similarly placed who had insured. However, once calamity strikes, then one is worse off than those who had insured. One entered the game, played the odds, and what results results. We might point out that risk takers sometimes win big against the odds; and sometimes they lose. That is life. Everyone can choose to be a risk taker or to behave conservatively. Taking responsibility for one’s actions is to accept the consequences that follow. It is not merely option luck that concerns us here, but whether the acceptance of responsibilities includes the playing out of contingent brute bad luck following choices freely and responsibly made. We can question the relationship between luck and probabilities. We might suggest that luck is not so much to do with probabilities as expectations. One may not expect one’s house to be struck by lightning. When it is we call this bad luck. Not insuring one’s house (at the correct odds for the likelihood) is not bad luck. One has not insured against the expectations properly constituted. That is why there is no justified claim for failures to insure. But again, luck is not simply the difference between our expectations about what might occur and what actually occurs. A premiership club losing to a conference team may be against my expectations, but has nothing to do with my luck if I am completely indifferent to the football results. Luck has to be measured in losses or gains, not simply in failures in expectations. Consider two lotteries: one with a ten million to one chance of winning £1,000 (for, say, a 1p stake) and one with a million to one chance of winning £1m (for, say, a £1 stake). Consider two people each with one ticket for each of the lotteries. Is the winner of the first lottery luckier than the winner of the second? The winner of the first lottery must have had lower expectations of winning (one in ten million) than the winner of the second (one in a million). But surely we think of the second winner as the luckier, since his gain is the greater. What matters then is the difference in rewards for two individuals, each of whom behaved similarly with similar insurance; but things worked out well for one, badly for the other. How lucky or unlucky we think of either of the individuals depends upon the difference in the value of the respective outcomes. The contingent working out of the probabilities has determined who was lucky and who not, but how lucky one is considered to be is determined by the difference in the value of the outcomes. I will refer to the economic and social system of rewards as the ‘reward structure’. I will say more about what might be meant by the reward structure below. It could be considered to be equivalent to what Rawls (1971) calls the ‘basic structure’ – that is, the major social institutions of government and regulation of trade and personal relations. However, it could also be given an expansive interpretation to cover those other elements that

140 Luck

Cohen (1997, 2000) believes need to be covered by egalitarian concerns, namely: the conventions that non-coercively govern our actions; the social ethos that shapes our attitudes to what is right and wrong; and the personal choices that people make. How might one measure luck? We could view the amount of luck one has as a relationship between the expected outcome and the actual outcome. A measure of luck could be thought as something like AV − EV, where EV is the expected value of the outcome and AV is the actual one. Over a series of trials, we would expect AV to approach EV. However, luck is often associated with events that only happen once in a lifetime. My expected monetary payoff (EV) in a £1m lottery with a million to one chance of winning and where my ticket costs £1 is 0. We should note that the seemingly odd consequence of examining ‘luck’ in this sense is that my actual payoff will never equal my expected payoff. For I either win – getting £999,999 (£1m minus my stake) – or I do not win – with a payoff of −£1. Thus the monetary payoff equals my luck in the lottery. It seems naturally odd to consider someone who does not win a million in one lottery as ‘unlucky’ even to the extent of −£1, since their not winning is, with high probability, precisely what we expect. Perhaps it is better to consider luck as being related to the distribution of outcomes. We do not expect to win the lottery, so the expected outcome is −£1; only if we win do we want to apply the term ‘luck’ to it. Bad luck is then only associated with outcomes that are unexpected. There are no unlucky losers in the lottery, only lucky winners. Such an interpretation requires an analysis of rational expectations of the odds of an event happening, but I will take this analysis no further here.5 In a similar vein, Pritchard (2005: ch. 5) places two conditions on an event being lucky. 1) If an event is lucky, then it is an event that occurs in the actual world but which does not occur in a wide class of the nearest possible worlds where the relevant initial conditions for that event are the same as in the actual world. 2) If an event is lucky, then it is an event that is significant to the agent concerned (or would be significant, were the agent to be availed of the relevant facts). We might see option luck or contingent brute luck this way. We cannot view personal identity luck in quite these terms. Personal identity luck is not subject to the first condition. It is only subject to the second. No person has any expectations about how we are to be placed in society prior to their birth, though given our knowledge of any society we can have expectations



Luck, equality and responsibility141

about the distribution of how people will be placed in that society.6 The only way in which we could apply the type of formula suggested above (p. 140) is if expectations were to be taken as some average on the distribution of rewards. However, I will argue that even for choices made, whilst we ordinarily think of luck as something not expected or of low probability, it is the second condition that is most important in our calculation of how lucky someone is.7 One consequence of equating luck with the reward structure is to suggest that luck egalitarians are wrong to say that inequalities are unjustified to the extent they derive from brute bad luck (see, for example, Arneson 1989, 1999; Cohen 1989, 2000; Roemer 1996, 1998), though by this statement I do not mean any criticism of luck egalitarianism. Since luck is defined in terms of the reward structure, it is the reward structure that creates unjustified inequality and it is by examining the reward structure that we must justify inequality, rather than by the luck it creates. Luck is simply the label we paste on to unjustified inequalities as determined by the reward structure. To the extent that there is a clear distinction between option and brute luck, then the conclusion that brute bad luck needs to be abolished and inequalities justified by decisions responsibly made might follow. If the distinction is hard to maintain beyond some clear examples, such an analysis is more problematic. If attention needs to be directed towards the correct reward structure, then different views about just rewards, the ends of the good life, and how far we want to encourage responsibility as opposed to equality become significant. This is because we cannot hold the concept of ‘responsibility’ static any more than we can the concept of ‘luck’. Changing the reward structure will not only affect the differences in luck, it will also affect incentives to behave in one or another manner, which will reflect upon how responsible people are. That is, there is no fixed amount of luck or responsibility in a society. Rather, the reward structure will determine how much luck and responsibility there is for each person in that society. We need to examine the relationship between luck and probabilities a little more closely to see how difficult it can be to distinguish brute and option luck; and to examine the relationship between the reward structure and responsibility. Luck and probabilities Imagine a simple Bournelli trial of coin tosses. In each trial we have a pattern of outcomes with probability p (success) and 1 − p (failure). What patterns do we describe as good/bad luck and what do we describe as ‘to be expected’? For each trial we know the probability of heads is 0.5 and

142 Luck

for tails is 0.5. Let us concentrate upon the probability of getting heads as success. Each time the coin comes down heads, we see this as success. If the coin comes down heads on a trial, we can see this as good luck. Sometimes winning the toss can be very important to an outcome. In cricket, winning the toss can, at times, virtually decide the outcome of the match. But in any coin toss, with a fair coin, coming down heads has a probability of 0.5. Once more, we can see that how much luck we assign to any given coin toss is determined by the reward. Tossing a coin in a laboratory to, say, generate a random set of outcomes involves no luck. Tossing the coin to see who gets the £1 million does. Now consider a sequence of tosses. If we toss the coin 100 times, we should expect there to be sequences where it comes down heads more often than 50 per cent of the time. We could look at the sequence 6–15 and see eight heads, and at the sequence 53–62 and see only two heads. Both sequences are expected in the sense that we can assign the same probability to each sequence occurring. Again, however, if we are betting on the coin coming down heads, the first sequence might be called lucky, the second unlucky. But how much luck do we assign to each sequence? The question is how can one empirically apply the concept of luck to a sequence of coin tosses? We know that the 50–50 split will not necessarily be seen in a short sequence. We can work out the probability of any given split that we do see. However, if we see someone tossing a coin and the heads-to-tails split seemed to favour heads when he bets on heads, and tails when he bets on tails, then we might begin to think that the coin is not fair. If it is established that the coin is fair, then we might think it is his tossing that is not fair. We might begin to think that the sequence is determined by skill. The question is at what point would we think the sequence was determined by skill? We have a hypothetical answer about expectations about sequences, and we have an actual set of sequences. The actual set does not fit our expectations, yet we also have an expectation about how often such a set of sequences would not fit our expectations. We have an underdetermination problem. We can extrapolate from the evidence either that the agent is skilfully tossing the coin to get heads (or tails) more often than we should expect or, for the set of sequences we have witnessed, the agent has been lucky. On one account he is lucky, on another he is skilful. How can we determine which our case is? In the following, I will suggest that this simple example of judging luck or skill in coin tossing shows the difficulty of judging whether someone has ended up where they are through contingent brute luck or through option luck. That is, whether they can be judged to be responsible for their position in life or whether it is down to luck. I ignore personal identity luck here, to be returned to later.



Luck, equality and responsibility143

Path dependency We have seen the difficulties of understanding luck in simple sequences of coin tosses and then considered how we might view luck with regard to level of skill at coin tossing. One way of judging skill would be to examine independent evidence. We might glean this by looking at the total population and examining the correspondence between how people look when tossing the coin, and the types of sequences they reach. For example, for any group of coin-tossers who seem to have better luck than others, we could examine the manner in which they tossed coins to see if we can tell how they are able to generate more heads than the unlucky group. Correlating success with some evidence independent of success – such as the manner of the coin-tosser – might allow us to judge skill. If we are correct in our inferences, we could then teach that skill to enable others to correctly toss heads more often than we would expect by chance. The analogy is to how players look when playing sports. We develop theories about good batting at cricket, hitting at baseball, golf swings, using a tennis racket. Our theory about how to play, together with how well people do, can teach us some of the differences between skill and luck. But even here, we know that ‘looking right’ and getting the right results is not perfectly correlated. Some good players manage to break the rules, and many people ‘look right’, yet do not gain good results. Nevertheless, the coincidence of results and the way people seem to play allows us to make some judgements about how skilful people are at sports and what constitutes a lucky break. Does the relationship of skill at sports teach us anything about contingent brute luck and responsibility in other areas of life? It can perhaps teach us something about path dependency and how luck affects responsibility. One important aspect of doing well at sport is temperament. Temperament both enables us to work hard at training and affects the outcome of matches. In some sports, temperament is the key to success. There are many good snooker players playing in clubs, but to keep one’s form in high-tension matches requires a nerve that some people do not seem to have. This is so in other sports too. Some players in tennis are thought to have the ‘big match temperament’, whilst others are seen as ‘chokers’ who play well, but do not win as much and all too often lose the big games. Sometimes the player who appears less skilful wins because she seems to have this big match temperament. She may be beaten more often than not by her opponent, until it comes to the big tournaments and then she seems to always prevail. We might think the big match temperament is a type of skill. We might hope to find some material manifestation of this skill independently of the results one achieves due to it. For example, we might think

144 Luck

it appears in some DNA sequence. However, given what we know about the manner in which genes are switched on and off by environmental conditions this is likely to be a forlorn hope, since the big match temperament is likely to be highly path dependent. One player, Martha, may be thought to have the big match temperament, whilst another, Cherie, may be a ‘choker’. These are genuine facets about their play, but each may have developed those temperaments fortuitously. Perhaps the first time they met in a final Cherie was off-colour and lost. The second time she was on form, but Martha had a couple of lucky calls and just edged out Cherie. Now Martha has a winning sequence of two, and Cherie a losing sequence of two. This fact about their history of matches – that occurred because of luck rather than skill (by ­supposition) – now affects future matches. Perhaps in their third match Cherie starts to play badly as she closes in on victory, in part caused by her losing the two previous finals. Then she loses a fourth time through the run of the ball (luck). By now she can never beat Martha in a big final because she is, indeed, a choker. The early sequence of Martha and Cherie’s matches is an outcome space that is typically patterned into different subsets, and that patterning gains its significance from how it is perceived. It was chance, really, that Martha beat Cherie twice, but the fact that the pattern matters to them affects the way they play in future finals, which then alters future patterns of the outcome space. Future patterning would have its own contingent brute luck without the past patterning affecting today’s play. But the players’ behaviour has been altered by their perception of the past patterning. The patterning of a subspace of sequences of coin tosses gains significance depending upon what was riding on each sequence. Similar sequencing has occurred between Martha and Cherie, but with the added difference that previous sequencing (by supposition caused through brute luck) has affected subsequent sequencing by the choices made by the players. Cherie may not be as aggressive as she normally is in her play because of the ‘choking effect’ and this affects the subspace she currently occupies. Brute luck is affecting her choice – choice otherwise made with full responsibility. In other words, contingent brute bad luck affects the nature of choices one makes, not simply now, but in the future. But that means that it affects our attitudes towards risk and so how we make choices when faced with options. Past contingent brute luck also affects one’s reactions to future option luck: but by how much? How unlucky has Cherie been in continually losing to Martha in the final? We have considered the probabilities in terms of the patterning of outcome subspace and how that affects future patterns. But that does not determine how unlucky she has been. Luck depends on the reward



Luck, equality and responsibility145

structure and the nature of what we feel should be equalized. If money is the object of equality, how unlucky Cherie is depends upon the relative difference between first and second prize. If respect and status is important, then it depends upon how much respect and status ride for each on the results of their finals. The degree that chance played in creating the sequences of losses for Cherie against Martha plays a role in our judgement of their respective luck. But how unlucky Cherie is, and how lucky Martha, depends upon the rewards that were consequent upon the sequence of outcomes. If the rewards for winning are only marginally superior to those of losing, then Martha has only had a little bit of good luck and Cherie a little bit of bad luck. If the rewards are very different, then the amount of luck each had – dependent upon the playing out of chance and its subsequent effect upon future behaviour – will be consequently higher. An egalitarian should want the rewards consequent upon winning not to be too great. But how much is too great? A radical egalitarian would want the rewards for winning and losing to be the same. Precisely what we include in rewards, though, is important. We might imagine the monetary rewards of winning and losing to be identical. But can we imagine the entirety of rewards being equal? Some lists of the relevant thing to be equalized include ‘respect’ or the ‘social bases of respect’. Can we really imagine no difference in how the players or spectators would feel about winning or losing? If two players compete, then their very nature will determine that there are rewards for winning – or there would be no competition. There has to be some motivation for them to compete: tennis is not simply patting a ball over a net. So there will be some inequalities. One argument for Rawls’s difference principle is that inequalities are justified to the extent that they exist for the benefit of the worst-off group in society. One incentive argument suggests that in order to motivate people to do well, greater differences in ‘prizes’ for succeeding are required than, say, merely greater respect. Material rewards generate greater output of goods and services, and with the correct tax and transfer schemes this greater output can be to the advantage of the worst off. All would agree that current reward structures do not exist for the worst-off groups, and the degree of inequality that exists cannot be justified by the difference principle. There is disagreement, however, over whether this is because markets are inherently bad at generating the correct incentive schemes or whether we simply do not have working labour markets. Certainly, few could seriously maintain that we have a labour market rather than a privileged self-serving club for remuneration of company executives. Some maintain that with a finely tuned basic structure, labour

146 Luck

markets could operate for the worst off by creating the right incentives for rewarding talents and hard work.8 Quite how egalitarian such a society would be – whether the economic rent demanded by high producers would mean great inequality, even if their productivity made the worst off better off, or whether improving the lot of the worst off would entail that such economic rents must be relatively low – is not something we can judge without a clear understanding of both the measurement device and empirics about the incentive effects of the economy in terms of that device. I do not wish to get into that discussion here. Rather, I want to examine how the nature of reward systems affects responsibility in egalitarian schema. Patterns and responsibility We assign responsibility according to the amount of control that individuals have over the outcomes that result from their actions. The number of viable alternatives in their opportunity set limits their choice of actions. The assignation of responsibility given any opportunity set depends on how diverse is the choice and the relationship between the action and the outcomes. Here I concentrate upon the latter problem. The sporting analogy was designed to show how difficult it might be to assign responsibility in some cases. We can assign expectations over the likelihood of sequences of heads and tails (what I refer to as a ‘pattern’) in finite series of coin tosses. If an individual appears to get heads more often than we should expect over a long enough sequence, we might suspect that the pattern is not random. However, often the data will underdetermine whether a pattern is random or forced in some manner. Furthermore, given that confidence (or temperament) is important for successfully instigating particular types of patterning of outcomes, past outcomes affect one’s present skill levels. As long as luck is assigned to chance events, and responsibility to those more under one’s control (skill), then our assignation of responsibility will depend upon our judgements about the conditioning of confidence and temperament. This demonstrates, however, that present responsibility for outcomes might well be based upon past contingent brute luck. If we recognize that inequalities are justified by the degree of responsibility that people have for outcomes (based on their behaviour now), we will not be removing brute luck entirely from the reward structure. And we should not. Contingent brute bad luck is something that enables character to develop, just as contingent brute good luck might.9 How do we judge individual responsibility in relation to patterns in behaviour dependent upon choices made? To return to the sporting analogy: it was suggested that temperament is something that is important



Luck, equality and responsibility147

in sporting events. Ordinarily we would place temperament within the ambit of personal responsibility. Developing a big match temperament is something that coaches try to train in their protégés. Certainly we might see it not only as an aspect of but as an integral part of personality. I argued, however, that it is also something that might be developed through contingent brute luck. If we try to reduce the element of luck in our egalitarian society by reducing the differences in rewards for winning and losing, we are also likely to affect the development of temperament. The greater the differences between winning and losing, the more likely that features of temperament will emerge. If there is little at stake in a game, then ‘choking’ is less likely. If we make the results too egalitarian, do we lose elements of responsibility? If the sporting analogy has any utility for life more generally, the problem it illustrates for the luck egalitarian is that we need incentives that will not exacerbate the effects of contingent brute luck without unduly affecting the incentives for people to make choices and take on responsibilities. To return to the characterization of responsibility above (pp. 135–6), we assign responsibility to the extent that people have a viable set of distinguishable alternatives. To make responsible choices, those alternatives must be distinguishable. If it makes no difference whether one wins or loses a tennis match, or becomes a lawyer, doctor or gardener, because any potential difference is compensated for, then there is no distinction and therefore no responsibility. Of course, there has to be a difference between winning and losing in tennis as long as we keep score. There are, of course, distinguishing features of lawyers, doctors and gardeners outside of any material rewards. But to the extent there are these distinguishing features, there are incentives to choose one job or profession over the others. Those incentives include the nature of the job, one’s fitness to take it on, reputational effects and material benefits. We might want to equalize many of those different incentives. However, it might seem that if we equalize over all of them, then we are left with no choices – only ‘picks’ or ­‘assignments’ – and no responsibility for choices. In other words, the attempt to distinguish responsibility and luck for differential outcomes does not allow us to bring responsibility into the conception of equality rather than as a rival desideratum. Perhaps the situation is not so bleak for egalitarians. In order for individuals to be motivated to choose alternatives from opportunity sets, they need to be able to find them distinctive. This does not imply that everything must be the same for everyone. That would only be so if all humans were stirred by precisely the same motivations. But people have different desires, talents, and so on. If there were the perfect niche for each person to fulfil their lives, and each niche were equally socially valued, then we

148 Luck

could have responsible choosing of niches with equal social valuation. Individuals would maximize for themselves and equalize across society. We could not be assured, however, that such a fitting of people into niches would be socially optimal. No one might want to clean the streets. Incentives Given there is no reason to think that people naturally want to fill all available niches, incentives are required for people to choose some occupations rather than others. Some egalitarians have argued that Rawls’s difference principle must be applied to moral psychology as well the basic structure (Narveson 1976, 1978; Cohen 2000). Cohen has argued that we need to understand that it is not enough for the basic structure or constitution of society to be geared towards equality if other interdependent factors are not also egalitarian in character. The conventions of society, the social ethos that shapes people’s attitudes and the personal choices that people make must all be geared towards egalitarian aims. It is not my aim to criticize Cohen’s account in any detail. Estlund (1998), Williams (1998), Pogge (2000) and others have done so. It is sufficient to say that I believe that the best way of changing personal choices towards greater fraternity and equality is to turn the social ethos and the conventions of society towards those aims. And those are most easily manipulated in that direction by the institutions of the basic structure (the argument of Pogge 2000). We need to understand what such an argument means for responsibility. We could simply define the just society as one where everyone does what is best. There are indeed social niches and everyone responsibly works out where their niche is and fills it. Or we could think more seriously about the need for incentives. I began by suggesting that the object to be equalized was very important for the discussion of how we are to view the issues addressed here. I have assumed throughout – usually in the background, sometimes in the foreground – that the object of equality is material benefits or, more simply, money. I do not believe, as it happens, that human psychology will ever allow us to equalize income across possible jobs, let alone wealth. But for the sake of argument, we can imagine a society where the reward structure materially equalizes across all positions. Unless the state directs what people do, we can still imagine a labour market of sorts. There will be numerous positions to be filled. There can be objective criteria for filling these positions – qualifications gained, human capital, and so on. If someone is qualified for several jobs, what would be their reasons for choosing one over another? If Martha had the capability to be a doctor, an architect or a gardener and there are no material incentives for one over



Luck, equality and responsibility149

the other, she will choose in terms of the satisfaction she expects to gain. So she chooses to be a gardener. This means one fewer gardening jobs going, so Michael, who could have been a gardener, gets a job working in a restaurant. He gains less satisfaction from that position than as a gardener. Now, it might be said that Martha should have taken a job as a doctor, leaving the gardening job for Michael. That might maximize the position of the least well-off person – Michael. But if material rewards have been equalized, Martha, unhappy as a doctor, becomes the least well off and Michael, happy as a gardener, should perhaps take the job in the restaurant. In other words, once material rewards are equalized, incentives for responsible choice must come from somewhere else. Do these need to be equalized too? If so, we may have problems, since we have reduced what we can trade for those rewards. And we cannot be assured that everyone ideally fits into a particular niche for social optimality. If we are to leave room for responsibility, we must leave some room for inequalities somewhere. We do not get rid of contingent brute luck in such a situation either. Someone may choose a career because of the interest it might bring, but, through no fault of their own, discover that the nature of the job changes over time. An academic may choose a career in what she thinks will be an exciting research programme, but regrets the choice as someone else soon makes all the exciting discoveries and now there is only tedious follow-up work. Fully functioning perfect labour markets are supposed to materially reward talent and effort and oil some of the differences in job satisfaction. Of course, supposing that this occurs is almost as fantastical as expecting each person morally to find their perfect niche. ‘Talent’ is a term taken for granted in much of the egalitarian literature. As we have seen, judging degrees of talent is difficult. Talent may include some genetic predisposition or capabilities, but is largely a result of development involving luck as well as training. We may judge relative talents within disciplines reasonably well. We may be able to make judgements about who constitutes the best sets of lawyers, doctors, academics, road-sweepers, sportspeople, and so on. That is, talents (together with associated luck) may emerge within labour categories; but we can only judge across categories when we have some currency by which to measure the judgement. Salary and respect are two obvious candidates. However, we cannot be assured that either works well. Perhaps perfect labour markets might provide relative salary levels on which to place such judgements. We can be assured that current labour markets signally fail. Brute luck comes into effect with salary levels as shortages, fashion or technological change may destroy carefully reasoned grounds for job choice. Trading the better-paid though boring

150 Luck

accountancy job for the interesting career as an academic might end up a poor decision as academia becomes a minefield of administration rather than research. There may be some relationship between talent and effort and rewards through well-functioning labour markets. But there is a lot of luck involved too.10 All this is by way of repeating that whilst responsible choosing is involved, so too will be contingent brute luck. They go together. That is not to say that the effects of luck cannot be reduced. Formal equal opportunities and ensuring the respect of rights will do so. Ensuring that job satisfaction can remain as high as possible through labour laws will do so. And setting the terms of material rewards to prevent such great disparity between the highest and lowest paid will do so. I have been arguing that to place emphasis on responsibility and luck in justifying levels of inequality is to examine the epiphenomena of problems. The reward structure largely determines the degree of both. Egalitarians need to answer crucial empirical questions about what levels of inequality are acceptable. How great should the differentials be between the highest and lowest earners? What is the highest sustainable level of basic income? What level of disability might require extra help? Historically there have been suggestions as to the relevant differentials that might be justified. The manifesto group of the Labour Party in the early 1980s suggested differentials of top-to-bottom earners of something like 10:1 whilst Proudhon suggested a 4:1 ratio. Picking ratios out of the air may not be the best procedure. We can be assured, however, that current levels of inequality are unjustified. And a start towards working out justified inequalities would be to pick off the unjustified ones first. Personal identity luck I set aside personal identity luck as a form of brute luck. If luck is measured as some form of relationship between expected outcomes and actual outcomes, then lowly returns from a lowly position in life can hardly be called bad luck. After all, if one is in a lower class, with uneducated parents, poor upbringing, and so on, then the poor outcomes one receives can hardly be unexpected. If we measure luck along the lines suggested on pages 140–1, then two people on their deathbeds may have had identical luck, but one lived a life of relative poverty in relation to the other. Of course, the point at which we measure brute luck at birth is not from one’s circumstances at birth, but from some imagined original position behind a veil of ignorance. And what we are measuring with personal identity luck at birth is the social structure as measured by expected rewards of birth into different social classes.



Luck, equality and responsibility151

Seen in this manner, personal identity luck is a measurement of social class and caste; and a measure of expected lifestyle for people with different capabilities (such as ‘natural talents’ or disabilities).11 To that extent, we can best hope to measure the differences in personal identity luck by looking at the outcomes of difference classes or what John Roemer (1998) calls ‘types’ of people using his Equality of Opportunity (EoP) measure. Roemer believes his measure captures the effects of brute luck. I think that is too broad a claim. Rather, it captures the average effects of personal identity luck across type. We can see why Roemer’s approach does not capture responsibility from his response to one of the early criticisms of his EoP measure (Roemer 1998: 21–2). Roemer’s measure is designed to capture autonomous effort – or responsibility. He does this by selecting his types by a vector of characteristics outside of the control of people (personal identity luck), with each member of a type sharing similar characteristics. The assumption is that each type faces similar opportunity sets, and how well they do through their choice of elements within the sets is a measure of their responsibility. So we can look at the average salary of each centile in two types (men and women) and the differences across the types will show the difference in opportunity to earn money between the two types. Now an underlying assumption of such statistical analysis is that the residual we are trying to capture (and other error terms) are equally normally distributed across types. That is, any factor which affects the selection of variables within opportunity sets not contained in the data, affects each type in precisely the same manner. For that reason choice of types is important.12 Brian Barry constructed an example of two types: an academic and an Asian type, who, as it happens, have identical educational results.13 However, the Asian group’s success derives from sheer hard work, impelled by their cultural milieu, whereas the academic type breezed through with little effort. If we want to reward effort, Barry suggests, the Asian type should receive higher rewards than the academic type; but Roemer’s formula treats both types equally. Roemer (1998: 21–2) responds that if indeed the pain of hard work is what is to be rewarded, Barry is correct; but he is trying to capture autonomously taken effort. The comparison in autonomously taken effort is between Asian children within type, not in comparison with another (academic) type. In itself, this is a reasonable response. But what it reveals, of course, is the assumption that autonomously chosen effort is equally normally distributed across types. What reason do we have for making such an assumption? To continue the education example, it may be that the top centile in each group has been successful, due not to autonomous choice, but because their parents are pushy relative to other parents of that type. The difference between

152 Luck

the Asian type and a non-Asian type might be tracking relative parenting pushiness across types (culture), but the differences within both the Asian and non-Asian groups is capturing relative parenting pushiness within each type. Or again, we can capture systematic factors across types affecting health (bad personal identity luck), but relative health levels within types may be capturing the effects of contingent brute luck rather than autonomously chosen lifestyles. Moreover, however we define types, and whatever it is we measure, whilst there are inequalities (captured by dividing the types into centiles) these problems will emerge. This criticism of Roemer’s EoP measure is directed at his interpretation of what it measures. It is not directed at the measure itself. EoP measures the systematic effects of personal identity luck as captured by relevant social and ethnic classes as well as by gender. Using such a measure to reduce relative inequalities across such types would be a massive step forward for equality. I do not believe, however, that the inequalities that would be left – those captured by placing individuals into centiles within types – would simply be those left by personal responsibility, for the reasons argued above. Conclusion I have examined the claim that inequality can only be justified in terms of choices freely and responsibly made and that egalitarianism requires the effects of brute bad luck to be removed. On the face of it, this seems a very reasonable egalitarian claim. I have argued, however, that luck is largely defined in terms of the reward structure. That in itself should not be a problem for egalitarians. Changes in the reward structure will reduce brute bad luck. However, changing the reward structure will, as those on the right of the ideological spectrum claim, affect the degree of responsibility that people can take for their choices. Luck and responsibility are epiphenomena of the incentives that people have to choose from the opportunity sets available. One conclusion is that egalitarians need to take a closer look at what reasonable rewards should be. They need to make empirical claims about how great the differential rewards for different types of behaviour should be. I have also shown that luck and responsibility are more closely bound than often imagined. Our personalities and characters are created by our history and that will include past contingent brute luck. Luck will reverberate throughout our lives and we cannot hope to remove bad luck entirely from the rewards we receive. Some may now make better (more responsible) decisions than others simply because of luckier decisions in the past. It is hard to disentangle luck and responsibility.



Luck, equality and responsibility153

We can measure, as John Roemer has importantly shown, average differences in personal identity luck based on types of individual. We can work to reduce the inequities that exist across those types. I have argued, however, that we cannot be assured his measure captures responsibility as opposed to luck within types. Contingent brute luck may be equally normally distributed within types, as are autonomously derived effort levels. We do want to assign responsibility to people. We do want to reward people who have, through their own efforts, done well. At the individual level, in both our personal and professional lives, we can seek to ensure that that happens. Governments can seek to reduce inequality and should utilize the results of social research that demonstrates systematic effects creating inequities. It should seek to reduce the inequalities across types and classes of people. Governments should not seek to reward responsibility, but should leave that to others. Instead, it should seek to reduce inequalities and welfare losses dependent upon failures in responsibility. We know there are many failures in human rational action. We know people systematically under-insure against calamity. We know people underinsure against infirmity. We know evolution has fitted people to like fatty, salty and sugary foods that are now in super-abundance in the developed world. We know children are not the best judges of their educational needs; that parents will fight to give advantage to their own children; and that the distribution of good family backgrounds is not equitable. Government can regulate to try to overcome all of these problems and more. It should not be concerned to reward the distribution of responsibility over decisions within these areas, but simply to raise the threshold of poor decisions where we know they are systematically deficient. Notes  1 Dworkin’s distinction between ‘option’ and ‘brute’ luck is considered below.  2 The literature assumes that the insurance market is efficient and correctly judges the probabilities of chance events. However, if one is subject to the vagaries of life against which one cannot insure, then no blame can be attached to one.  3 Though even here, of course, what counts as beauty or mental impairment depends upon the nature of society (its structure and culture).  4 In fact, elements of the genetic side of personal identity luck might also be correlated with class – if ‘beauty’ is inherited and enables upward social mobility, then it too will be correlated with class. Similarly, inherited susceptibility to disease may be more likely to emerge where social conditions are more stressful, again hitting the lower social classes to a greater extent. Even lightning might strike outdoor workers more often, so contingent brute luck may not be completely contingent. I ignore that possibility here and assume contingent

154 Luck

 5

 6  7

 8

 9

10 11

12

13

brute luck is equally normally distributed among all social classes. That is, it is the normal distribution of such brute luck that makes it contingent brute luck rather than personal identity luck. We cannot assume that luck is the difference between the actual outcome and the most probable one, as Richard Bradley pointed out to me. Whilst this seems to be a reasonable account of luck when the most probable outcome is highly likely, it is not when the differences between the outcomes are small. Consider the lottery where you have a one in six chance of winning each of £10, £20, £30 and £40 and a two in six chance of winning nothing. The most probable outcome is nothing, but we might describe someone who won nothing as unlucky. Perhaps it depends on how the outcomes are described – winning nothing or winning something. For that reason, such personal identity brute bad luck might be better called ‘fortune’, a term closer to natural English. I make nothing of the linguistic point; what matters is the conceptual analysis and what it entails. Pritchard’s analysis of luck is written with regard to problems in the theory of knowledge and how to analyse epistemic luck: that is, the luck associated with ‘knowing something’ by chance as in the Gettier (1963) examples. How I analyse luck here is not meant as any form of criticism of his analysis in the context in which he is writing. Some argue that the operation of markets leads to inequalities that luck egalitarianism allows, but they are surely wrong (Anderson 1999; Scheffler 2003). However, I am querying the neat option luck–brute luck distinction that such a claim relies upon. In the comedy science fiction TV series Red Dwarf, Rimmer (one of life’s ‘losers’) thinks avoiding being held back a year at school was good luck; he discovers it was bad luck when he meets his superhero self from another possible world where being held back a year led him to transform his attitude to life. This was well understood by the management guru W. Edwards Deming, who argued that performance pay outside of strict confines was illusory, as there is so much random variation in performance. I use the term ‘capability’ in its dictionary definition as a ‘potential ability’, rather than Sen and Nussbaum’s preferred use of it to mean ‘ability’ (Sen 1984, 1987a, 1987b, 1993, 1999; Nussbaum 2000; see Morriss 2002; Dowding 2006c). As far as I can see, most of the criticisms of Roemer’s EoP approach (for example, Phillips 2006) merely give examples where this assumption clearly cannot be satisfied. This means that his formula cannot be used in those cases, but does not undermine the formula itself. Barry constructed the example at the American Political Science Association conference in 1996. It is discussed by Roemer (1998: 21–4).

8 Luck and leadership

Some prime ministers, such as Margaret Thatcher and Tony Blair, are considered strong leaders. Others, such as John Major and Gordon Brown, are perceived as weak. I will consider why this is. I do not challenge those perceptions, though part of my argument is that prime ministers, or indeed any leader, might be weak, at least in part, because they are perceived to be weak. One view of strong leadership is that it is a personal characteristic. It is as though some are born strong and others weak. Whilst genetic inheritance of character traits cannot be denied, seeing the genetic lottery as the only form of luck is blinkered. I suggest that luck plays a significant role in both perception and the reality of strong and weak leadership. The circumstances in which leaders take power, the events they face and the context in which they rule are of vital importance in cementing their leadership. It is perhaps not coincidental that my two examples of weak leadership are prime ministers who immediately succeeded my two examples of strong leadership. Whilst the generational difference between Thatcher and Major makes it problematic to imagine her replacing him as leader of the Conservative Party, it is an eminently reasonable counterfactual that Blair could have followed Brown. Had that happened, would our respective judgements about their leadership qualities be reversed? Perhaps not: decisions taken and personality might have led Brown to be less successful than Blair; and subsequently Blair more successful than Brown. But, almost certainly, had Brown become prime minister in 1997, his premiership would not have been the disaster his actual one proved to be. My argument about luck, however, is not the simple one that in one context a Blair succeeds, but in another context (following a successful prime minister, but with a tired party and sceptical public) he would fail; thus the luck of the draw determines one’s fate. Everyone can see that that can be true. Rather, the relationship of events and one’s reactions to them

156 Luck

creates an impression not only on the observer, but also on the person themself. That impression can then determine’ their future reactions to events, which then give further impressions of their qualities. The interaction of events and one’s reactions to them create both the impression and the reality of how strong or weak a leader is. Second, how others perceive a leader’s actions affect their responses to them. So perceptions affect leadership qualities in a second way. I first say a little about the literature of leadership; then I explain the nature of luck; I next consider in a bit more depth the nature of perceptions of political leadership. Throughout, I illustrate my argument with the cases of the four prime ministers I mention in my opening sentences. Leadership studies Margaret Levi’s 2006 American Political Science Association presidential address suggests that a great gap in our understanding of politics concerns the quality of political leadership (Levi 2006). There are models of what leaders must do to thrive and survive (for example, Bueno de Mesquita et al. 2003) and there is a growing literature that examines the coordinative, informational and strategic roles that leaders play. Fiorina and Shepsle (1989) describe leaders by type: agents, entrepreneurs or agenda setters. Under the first type, agents, leaders are followers helping their principals to coordinate their actions by helping aggregate preferences and information. Dewan and Myatt (2007) model this as public information provided by a leader and examine conditions under which people follow the leader, suggesting a ‘Michels’ Ratio’ where a parameter of the ‘need for direction’ controls how much it matters what direction everyone coordinates on, affecting how much the leader can dictate. They develop this further, endogenizing how leaders emerge based on the clarity of their messages and sense of direction. As the followers’ concern for cohesion increases, the clarity of the leader’s message grows in importance relative to the realities of the world. Sometimes coordinating activity is more important than being right (Dewan and Myatt 2008, 2012). The greater the opportunity a leader has to direct the group rather than simply coordinate activities within it, the more the leader resembles an entrepreneur rather than an agent. Most models of leaders as entrepreneurs see them as putting together coalitions (rather than leading pre-formed groups), often manipulating information, framing issues and building new lasting political coalitions (Riker 1982, 1986). As agenda setters, leaders deploy the institutional resources their role gives them to ensure that their preferences are promoted (Cox and McCubbins 2005; Cox 2006). The three types are not mutually exclusive: any leader might fulfil



Luck and leadership157

each role at different times. All three roles are essentially informational ones. Leaders collate information and then use it to secure their own and their followers’ interests (Ahlquist and Levi 2011). Some studies examine ‘everyday’ or ‘transactional’ leadership – managing processes within frameworks accepted by people. Others concentrate upon the ‘great leaders’ who are ‘transformative’, bringing about radical departures (Hargrove 1989). Such transformative leaders are those who articulate a new way of looking at the world and advocate a specific direction, taking their followers with them (Schofield 2006). All of these models of leadership indicate the structural conditions of leadership and can help us retrospectively analyse successful leadership. They do not, however, provide information on the specific qualities that people bring to these informational and strategic roles, in the sense Levi (2006) seems to demand: finding how we can ex ante acknowledge the attributes that qualify specific individuals for leadership. This is because, I will argue, context and luck explain more about the success or failure of any given leader than any set of personal characteristics or traits. To be sure, in order to reach the top of the political ladder, perhaps to attain ministerial posts in a parliamentary democracy, individuals will demonstrate qualities that can be systematically analysed (Dalvean 2012a, 2012b). But the specific characteristics of who emerges from that group as chief executives and, of those, who are later recognized as strong or great might not be susceptible to further evaluation. There is a large literature on the nature of leadership qualities (for reviews, see Grint 2000; Messick and Kramer 2004; Porter and McLoughlin 2006; Hunter et al. 2007; Mumford et al. 2008; Avolio et al. 2009). Much of this psychological literature is about leaders in private sector organizations, but some also discusses political leaders (Bell et al. 1999; Peele 2005: Morrell and Hartley 2006).1 Some combine analytics with personal experience (Keohane 2010). Much of the literature falls at the first hurdle since it commits the cardinal sin of social scientific research: selection on the dependent variable. But there is an impressive set of studies with better research design that does consider personal characteristics. That literature shows there can be little doubt that personal psychological characteristics help some people to leadership positions, and that such characteristics also help determine which leaders are viewed as good or bad, strong or weak, progressive or regressive, and so on (Hogan and Kaiser 2005). However, most studies of leadership in organizations concentrate upon personal traits or specific decisions; good leadership then becomes a checklist of good qualities. Such checklists are not always consistent because, of course, qualities that are useful in some circumstances (‘a good leader will show willingness to take a risk’) might prove disastrous

158 Luck

in other circumstances (‘a good leader will be cautious where conditions dictate’) (Boin et al. 2005: ch. 1). Moreover, case studies of leadership are written in retrospect, so taking a given risk looks justified. Where a policy proves disastrous, those involved might report on the chaotic decisionmaking process, even though it might not qualitatively differ from processes producing policies deemed successful. This ‘narrative fallacy’ (Taleb 2007) or specification problem (Dawes 1996) is a general predicament with retrospective case studies (Dawes 2001). Of course, personal qualities enhance good leadership, but structural or institutional features will also help select which among those psychological factors will determine who gets to the top and succeeds in different contexts. An organization might need a conciliatory leader following a period with a strong-willed leader who perhaps led well out of crises but later failed the organization. The qualities that help select a US president may not be those enabling a person to lead a British political party or succeed in the parliamentary arena. To succeed in the British parliament, politicians must be able to deal with parliamentary questions, come over well on the media and be able to manage complex departments. British prime ministers have had many different traits and personalities, but certainly, over the last fifty years, each, in their own way, was sufficiently quick-witted or able to deflect criticisms to succeed in a parliamentary setting. The requirements for success as US president are somewhat different. Presidents receive much more protection from such hostile environments as the British parliament. The closest equivalent is the presidential debates that candidates face, but these rule-bound occasions are still very different. Some US presidents might have been found wanting in the British context; as might some British prime ministers in the rather different cultural context of a US presidency. From the outset we can see that whilst personal attributes must play a role in leadership success, the same personal characteristics might not be successful in all institutional contexts. This fact demonstrates, at the very least, how important the individual and structural relationship is in any account of leadership. The idea that both structure and agency are important in social outcomes is widely accepted in social science (see Chapter 2). In leadership studies (Grint 2000) it is labelled the ‘contingency approach’, departing from what has been the dominant mode of analysis, concentration on personal characteristics or ‘traits’, thus privileging individual over structure. More recently, the literature has moved towards more contextualized accounts (Hunter et al. 2007; Yammarino and Dansereau 2008), examining the types of issues leaders face and the institutional forms in which they operate. ‘Situational approaches’ suggest that different sorts of leaders emerge in different sorts of circumstances: leadership is a niche to be filled



Luck and leadership159

(Grint 2000). There is a conceptual debate here over whether ‘leaders’ are seen primarily as people who fulfil an organizational or group role or whether ‘leadership’ is given an independent status, meaning that not all who hold such roles are considered ‘leaders’. In other words, is ‘leadership’ defined by trait or by structural position? Here I assume that ‘leader’ means the role, but that we can talk of ‘good’ or ‘bad’, ‘strong’ or ‘weak’ leaders in terms of their traits or at least the decisions they make in the context in which they make them. I offer no objective definition of what constitutes good, bad, strong or weak leadership (see, for example, Hogan and Kaiser 2005, for some definitions). Rather, I offer some account of how leadership is perceived by people: that is, why some leaders might be thought to be ‘strong’ or ‘weak’. My argument is that how we view individual leadership traits is, in part, dependent upon their emergent context. Moreover, that perception is a self-perception as well as a perception by other observers. The self-perception then interacts with future situations, often perpetuating those characteristics from ‘the inside’ so to speak, and forms part of the expectations of others, affecting their future interactions with the leader. Thus perceptions alter the leader’s behaviour when faced with similar situations, and also others’ behaviour. Thus the perception of good or bad, weak or strong leadership becomes the reality. Even without self-reflection, the perception that a given leader is weak can be perpetuated once formed, for two reasons. First, actions that might appear strong if carried out by a person with a reputation for strength might be seen as the action of some weak person who is trying to be strong. Second, and more importantly, the reactions of others in response will depend upon how they view the leader. Someone seen as strong who shouts at a subordinate might cause that subordinate to quail and obey. However, a leader who is perceived to be weak and shouts will be laughed at or shouted at in return, further weakening him. One’s reactions in these cases might be an unconscious result of the prior perception of others and of one’s ­self-perception. Various physical responses are associated with dominance in personal relations. Keeping one’s eyes down is submissive, whereas staring at or just above the other’s eyes is a sign of dominance (Strongman and Champness 1968; Kleinke 1986). Self-confidence can unconsciously affect such eye movements or make it difficult to maintain eye contact, even when one is aware of these facts and trying not to show submissiveness. Of course, these actions and reactions lead to conscious and unconscious strategies in people. Game-theoretic analysis of strategic interaction demonstrates just how important reputation can be; the mere fact of a reputation for weakness can create (further) weakness (see, for example, Binmore 2007: 282–3). The literature on leadership has tended to ignore the interactive and strategic nature of inter-agent relationships, thus missing

160 Luck

an important aspect of the contextual character of leadership traits. In other words, as has been noted in other contexts in the social sciences (see Chapter 2), traits and context (individual and structure) are not simply different sides of leadership, both of which need to be taken into account, but are deeply implicated with each other. I explain first how leadership qualities might emerge through luck; and then suggest how patterns of relationships can lead observers to perceive some leaders as weak and others strong. It is through those external perceptions that the subject and those with whom they interact come to believe that those weaknesses and strengths are realities. Luck and dominance I first argue that the relative strength or dominance that one person might develop over another can occur through luck. I am not making the obvious point that sometimes someone might make a decision which happens to come off through sheer luck and because of that we consider them a good decision maker. So, for example, two people in identical situations each make the same risky decision; for one it succeeds spectacularly, for the other it fails. In retrospect we might make the judgement that the one who got it right (through sheer luck) is the good decision maker. The one who got it wrong we see as having poor judgement. When such decisions are seen as cementing authority, then the first person, in retrospect, is regarded as a strong leader; the latter as a weak one. My argument follows from this simple account, but extends into real characteristics of the individual. I am not simply making Napoleon’s point that he preferred his generals to be lucky. Rather, I am suggesting that the actual strength or determination (the trait) of someone occurs through luck.2 The luck I have just been describing is the beginning of my account, not its entirety. Let us be clear. Imagine for a moment the simplistic story that there are individual traits that emerge in behaviour, which are genetically determined. There might be a gene for ‘determination’, or ‘stubbornness’ or ‘collegiality’. We pause for a moment to consider what this means. In a simple account we identify a given gene as an allele in a specific locus by the effects it has, in this case, on behaviour. An allele can be regarded as having a certain selection coefficient relative to another at the same locus at a given moment. The coefficient is a number that can be treated algebraically and inferences drawn from one locus to other loci. In that sense, it is the statistical effects by which we identify the genes, and it is by those effects that we denote them. So across subjects we note differences in a specific locus on a chromosome and see statistical differences in behaviour. Not all who have the gene will display those behavioural



Luck and leadership161

characteristics, and some of those without the gene might well display some of them.3 In other words, in context those with the gene are more likely to display those behavioural qualities. My argument does not depend upon such an account of ‘genes for’ some behaviour being discoverable, though it might explain, in a process-­ tracking fashion, why some with such genes (if independently identified) emerge with those behavioural characteristics. My argument can assume that two people with regard to any ‘gene for’, say, determination, are identical. But for one, through sheer luck, the behavioural trait emerges, and so she is at first perceived to be more ‘determined’ (though by other measures is no more so), whilst the second is perceived to be hesitant and indecisive. Then, because of that perception, the first actually does become more determined (by those other measures) and the second (by those other measures) more hesitant. This is the luck through which leadership qualities might emerge. I will explain that possible emergence through a sporting analogy. What is viewed as the ‘luck’ of a person depends upon the identification of a pattern associated with a reward (see Chapter 7). Both the pattern and the reward are necessary for the assignation of luck. Given a set of rewards (in a sporting contest, in a lottery, in the ‘leadership stakes’), the amount of luck is determined. For example, if the winner of a lottery is thought lucky to have won, then we note that where there are lotteries there always will be someone who is lucky. Or, with a nod to Napoleon, whilst prior to the fighting the outcomes of battles are in doubt, there will always be lucky generals. To the extent that life is a series of lotteries – there are always winners and losers – then there will always be some people who are luckier than others. The response to the claim that winners and losers are determined by their traits is that the emergence of traits is only statistically correlated with genes; why some people with those genes have the observed effects and others do not is due to context. Where the patterns we see in these statistical correlations are associated with rewards, that context is what we call luck. The genetic lottery is one kind of luck and the context is another kind of luck.4 When considering leadership, certain modes of behaviour – broadly, those associated with competency – are thought to correlate with good leadership (McClelland 1973). Whatever those modes of behaviour are, I will call them ‘temperament’. In sport, temperament enables us both to get on with the job and to work hard at training, and also affects the outcome of matches. Some players are thought to have the ‘big match temperament’, whilst others are seen as ‘chokers’ who often play well but do not win as much as they should and who are particularly prone to losing in the big matches. Sometimes the less skilful player wins because she seems to display the ‘big match temperament’. She might be beaten more often than

162 Luck

not, until the big tournaments when she always seems to prevail. The ‘big match temperament’ might be thought of as a type of skill. We might hope to discover some material manifestation of this skill independently of the results thought to be achieved by it. That is, to find the ‘gene for’ big match temperament. That is likely to be a forlorn hope. Not only is it unlikely that complex behaviours can be discovered in simple gene sequences but, even if they could, given what we know about the manner in which genes are switched on and off by environmental conditions, such things as a big match temperament are likely to be highly path dependent and any potential statistical relationships are unlikely to emerge.5 The Australian golfer Greg Norman was a ‘choker’, but when did he become so? Was it after Jack Nicklaus’s charge at the 1986 Masters? The year following Larry Mize’s lucky chip? Or was it during one of the other three playoffs he lost at grand slams? He undoubtedly choked in the final round at the Masters in 1996; and in the 2008 Open Championship blew a two-shot lead. What we might see in Norman’s grand slam history is an outcome space that is typically patterned into different subsets in which Nicklaus and Mize beat Norman. However, the fact that the pattern matters to Norman has an effect on his play in future final rounds in major tournaments. Because the previous pattern matters to Norman, it becomes the context that affects his future actions. In other words, the previous pattern, which by assumption was (bad) luck, then alters future patterns of the outcome space. Not only does this pattern affect Norman’s behaviour, it also affects that of his rivals; they know that, despite leading in previous major championships, Norman is typically caught and beaten. This pattern of behaviour, by assumption initially due to luck, becomes a genuine (predictable and real) pattern simply because it is seen as a pattern by Norman and other observers. In that sense, patterns of behaviour take on the opposite characteristics of ‘patterns’ in fair coin tosses. The cause of the gambler’s fallacy is seeing a pattern in random data that is not there, and then making predictions about future data based upon a false pattern. With any self-reflexive animal, a pattern (which by assumption is luck and thus not a real pattern) becomes a real pattern because the self-reflexive animal thinks it is one.6 Even if Greg Norman claims the pattern is not real, he might subconsciously believe it is and this then affects his current play, ensuring that the pattern is real. We note that what makes the pattern real is its predictability under some probability distribution, just as a ‘gene for’ is real by its identification in a specific locus on a chromosome and inferences drawn from its statistical effects on behaviour. And we note that this pattern remains real whether or not any ‘gene for’ is identified. Real patterns are identified by their predictions.



Luck and leadership163

With some patterns, such as addiction, recognizing the pattern and then fighting to change it might be the only way to overcome it. In sport, players might indeed try to fight the choking pattern, but the fact that sport is competitive and their opponents might think the pattern is real will make breaking it all the harder. Given the strategic nature of leadership, if others have beliefs about a leader’s behaviour, changing that behaviour from a bad pattern (‘weak’ leadership) to a good one (‘strong’ leadership) might prove exceedingly difficult. Future patterning would have its own luck without the past patterning affecting today’s play. However, players’ behaviour is altered by their perception of past patterning. As we saw, the patterning of the space of sequences of coin tosses gains significance depending upon what was riding on each sequence. Similar sequencing has occurred for Norman, but with the added difference that previous sequencing (by supposition caused through luck) has affected subsequent sequencing. Norman’s choking traits were, at least in part, caused by luck. We can conclude from this analysis of the deep role that luck plays in the development of character traits that there are no recipes for great success as such. Or, rather, the recipe for great success is just as likely to be a recipe for disaster. Play safe or take risks. The latter leads to success or failure, the former to a more moderate outcome; or – in Taleb’s (2007) terms – black swans can destroy even safe bets; betting on the arrival of black swans can bring great rewards, but is even more likely to lead to failure. More can be said about these possibilities for leadership and how the public perceives them. Perceptions of leadership quality The sporting analogy has demonstrated how genuine traits might emerge through luck. What I have been calling ‘temperament’ might be the result of beliefs about one’s abilities that have been generated by past successes, some part of which was due to the luck of the patterning of outcome space in which one is successful. That temperament is then itself causally important in future patterning. Similarly, certain good leadership traits are genuine in the sense that some people have them and others do not. But the reason those good leadership traits emerge is through luck. They emerge because of the outcomes brought about by decisions the person made and, importantly, how others viewed those outcomes. Now, of course leaders can manipulate the portrayal of outcomes to their own advantage – John Major portrayed a rather poor showing in the leadership challenge he forced as though it was a great success that cemented his command of the Conservative Party (Cowley 1996). However, other aspects of the strength

164 Luck P

A

B

(a) Leader dominant

B

P

A

(b) Leader concessionary

8.1  Spatial representation of leader reputation

of leadership can be illustrated in terms of simple models of their context. In a very simple spatial analogy we can see how perceptions of leadership quality can be determined by context. Imagine the lines in Figure 8.1 represent some issue dimension, such as a left–right ideological dimension or a hawk–dove dimension on foreign policy. Whatever the issue is, it crosses the legitimate boundary of policy concern of two different departments headed by different ministers. We have three players: a minister from Department A, one from Department B, plus the prime minister (P). The non-strategic (or ‘naive’) preferred policy points (their ‘bliss points’) are marked on the dimension. Cabinet government requires discussion across departments, either on a bilateral basis or through the cabinet committee system. The details are not important. Whatever happens requires some bargain between the two ministers and also the agreement of the prime minister. Imagine, due to her greater institutional power, the leader can force her own bliss point; and in fact does so. Our question is, given the leader has forced her view in each situation, how does this outcome create our perception of the leader? In 1(a) the leader is an outlier to the left of the two cabinet ministers. She has forced her own bliss point even though it clearly departs from those of her colleagues. Despite opposition, she was determined and in the end forced the compliance of her cabinet ministers. In forcing her bliss point despite resistance she has demonstrated her power in the classical sense of ‘getting what you want despite resistance’ (Dahl 1957; Weber 1978). In 1(b) she also achieved her bliss point and, given her resources, could have forced that bliss point no matter what her ministers desired. However, here her bliss point precisely bisects the bliss points of her two cabinet ministers. Even without the prime minister’s involvement, attaining her bliss point would be a reasonable expectation of any negotiation between two ministers with approximately equal power. In the first case, the prime minister seems dominant; in the second concessionary. Even if she forced through the policy at her bliss point against strong opposition from both



Luck and leadership165

ministers, since they were opposing from opposite sides her strength in forcing this situation might well be viewed as the enforcement of a compromise between two powerful ministers. We can note that, as drawn, the prime minister’s bliss point has not changed, only the position of minister B. Imagine that the qualities (personal characteristics of the prime ministers) are identical in each situation; the only difference is how they are viewed, given the situation. It is the situation of each that characterizes our views of the prime ministers in question. If the ‘leadership prize’ is ‘dominance’, then the prime minister in 1(a) is lucky. If ‘getting what you want with least effort’ is the prize, then the prime minister in 1(b) is lucky. What matters for my argument is how the prime minister is regarded in each case. In the first, her strength is displayed for all to see. In the second, observers might consider that she has negotiated a tricky problem between two powerful ministers whom she has to somehow satisfy, and view her as relatively weak. Of course, real life is more complex than this simple example. We might know the bliss points of the actors prior to negotiations taking place and that knowledge will affect our judgement about what is going on in the situation. And players have both strategic as well as naive preferences. Player A might pretend to be at the prime minister’s bliss point in 1(a), hoping that she would not want to be seen to favour A too much and hence shift policy slightly towards B (and A’s bliss point). The prime minister might be very careful not to reveal her true preferences until very late in the game, and so on. But I want to make a very simple point. In the UK Margaret Thatcher and Tony Blair are both regarded as having been strong prime ministers; and they were. However, they were also preference outliers within their own cabinets, both being further to the right than the median figure. John Major was perceived to be a weak leader, and he was; but he was also a median voter within his cabinet. He was seen to be concessionary, sometimes giving way to one side, to the other side on others, and sometimes making compromises. But, I would suggest, he forced his true bliss point just as often as Blair or Thatcher.7 If it is true that Thatcher, Major and Blair got what they wanted approximately equally as often, then why do I agree with common opinion that Thatcher and Blair were both strong leaders and Major a weak one? Why do I not claim that this is only our perception and not the actual fact? I do so precisely because of the interactiveness (and so ‘path dependency’) of perception and reality. It is because Thatcher and Blair were seen to be strong leaders that they were able to be so. Because Major was perceived to be weak, even when he did force issues – and he did, repeatedly – he was seen to be doing so from a weak position. And in each case their cabinet colleagues responded accordingly.

166 Luck

Success and strong leadership Thatcher was not generally perceived to be a strong leader in the first two years of her premiership (strident from a misogynistic viewpoint perhaps, but not strong).8 Not concessionary certainly; but she was perceived as a weak leader presiding over a split cabinet. She did not get her way on industrial relations, on Ireland, nor on other matters.9 Whilst she held firm on the economy, the recession and mass unemployment that her government seemed to do little to resolve did not make her appear a great leader. In her decision to send troops to win back the Falkland Islands she took a risk (against elite opinion), and it came off. In that and subsequently she was seen to be a leader with authority. Furthermore, those controversial economic decisions taken early in her premiership have subsequently been viewed in a more favourable light by economists and public opinion. Thatcher was seen to be strong partly because she went against elite and public opinion and has been viewed as being right. (Whether or not she was ‘really’ correct is irrelevant to my argument.) It is not enough to be considered to be doing the right thing if you wish to be viewed as a great leader. Rather, you need to be perceived as doing the right thing against the weight of public opinion (Canes-Wrone et al. 2001). I hypothesize therefore that how ‘great’ or ‘strong’ or ‘authoritative’ we judge public leaders to be is based in part upon the relationship between their views and those of ‘society’ (say the majority or median voter) and how often the leader was perceived to be correct. (And since perceptions of what is ‘correct’ can vary over time, how strong a leader is viewed to be at time t1 might change when t1 is viewed at time t2 – as indeed in retrospect have Thatcher’s leadership qualities in her first two years.) I illustrate this point in Figure 8.2. The idea illustrated in Figure 8.2 is that a leader thought of as choosing correctly, but in line with the median voter’s views, is only given credit for following the median voter. They can be thought of as pandering to the electorate (Canes-Wrone 2006). The leader considered to have chosen against the wishes of the majority is given extra credit, not only for being correct, but being correct against the odds.10 Similarly, the leader who is considered to have made the wrong decisions and who was in line with the median voter on the issue is seen as weak. The leader who chooses wrongly and against prevailing opinion is perceived as disastrous. Such a person can be thought of as having gambled against the odds and lost. Because she was radical and controversial, Thatcher was not a popular leader during her time in office, having the second lowest public net satisfaction score measured by polls whilst she was prime minister. (Gordon Brown had the lowest.) Retrospectively, however, she was rated much higher by the public and second highest (to Attlee) by academics (Theakston and Gill 2011). Certainly this is because (like Attlee) she was a radical prime



Luck and leadership167 Amount of authority A

A-A When leader goes against majority opinion F-F When leader follows majority opinion

F

F 100:0

A* 50:50 Perception that leader is correct

A 0:100

8.2  Authority of leader

minister who made long-term changes to the fabric of social and economic structure, often against prevailing public and elite opinion.11 Figure 8.2 presents a stylized picture of this analysis. When the perception that the leader is correct is about even, then both the leader following public opinion (represented by F-F) and the leader who opposes majority opinion (A-A) are awarded about the same authority. I have no empirical evidence of that claim, but if the shapes of the ‘authority curves’ A-A and F-F are roughly correct, the figure illustrates the idea that with only a slight increase in the perception of the probability that the leader is correct, the leader who chooses against public opinion is attributed much greater authority. Going wrong ‘All political lives, unless they are cut off in midstream at a happy juncture, end in failure, because that is the nature of politics and of human affairs’ (Powell 1977: 151). It is the nature of political life that issues change, problems arise, and eventually the public will want to move on from positions currently adopted. It often seems that the stronger the leader, barring Powell’s happy juncture, the more spectacular the fall. One obvious example is Margaret Thatcher, who seemed completely mystified by the way in which she was forced to stand down, told one by one by her cabinet ministers that they felt she ought to resign, although she had received more votes in the leadership election than any other contender. Their reasons are not straightforward, but one important factor was the poll tax. This

168 Luck

was perhaps the single most disastrous policy of her government in terms of her administration’s popularity. It was an inefficient tax that led more people to lose than to gain, and was perceived to be inequitable and pernicious by most people. The story of how it came to be government policy has been well told and involved a measure of mischance as well as miscalculation (Butler et al. 1994). Nevertheless, it was Thatcher’s determination to see it legislated that was the key factor in its becoming government policy. There are no easy answers as to why powerful leaders fall, but the poll tax failure is a paradigmatic example of what often occurs. The analysis I am offering here suggests one cause. In the same way as leaders are seen as powerful and right against prevailing opinion, leaders see themselves in that light. Leaders can too readily come to believe the perceptions, the myths, of themselves. Thatcher’s early strength came about through coalitions in her cabinet, getting allies into important positions. She did not, as is sometimes thought, always get her way. She triumphed over the miners in 1984, for example, but that victory was set up by careful planning, having three years earlier conceded to them, persuaded by colleagues that it was not the right time to take them on. She made tough decisions against majority views and prevailed. It is easy for successful leaders to think that they can win the same trick over and again. But it is the nature of risky choices that the odds are against your always winning. Conclusion Levi (2006) suggests the lack of good models of strong leadership is a serious lacuna in political science. Part of my aim here is to suggest that there are good reasons why we cannot fill completely that lacuna. I do not argue that we cannot have models of leadership, and considered several in the opening section. These are general structural accounts of leadership roles. They specify what leadership is and categorize leaders in several, often overlapping types. Other models, such as those of Dewan and Myatt (2007, 2008, 2012) are more explanatory, predicting relationships between, for example, the need for coordinated action per se and the direction people desire. My argument is orthogonal to those models. We can describe types of leaders and the roles they display. We can explain why one prime minister was more successful than another in terms of the decisions they made and the successes they enjoyed. And we can describe some good qualities that leaders display when making those decisions and dealing with colleagues. What we cannot do, I argue, is specify a set of rules of behaviour that explain success. Telling an entrepreneur to behave like Richard Branson, or a prime minister to model themself on Margaret Thatcher, is not necessarily going to help them. It might have precisely the opposite effect.



Luck and leadership169

There are almost certainly qualities that lead some to be more successful than others in politics. And we might expect to find over a large enough study of cases some qualities that seem to lead some to be more successful prime ministers than others. However, the qualities that emerge as important in any given case might well have done so through luck. At a different time and place those same qualities might have led to failure. Furthermore, the qualities that we perceive – and because we perceive them, they do really exist – in successful leaders also emerge through luck; it is the pattern discerned in behaviour that leads us to treat people in specific ways that feed into the perpetuation of that behaviour and thus the pattern. I have argued that we cannot understand leadership by looking at the qualities of any given leader, nor indeed of all leaders in isolation from (i) the issues they faced, (ii) whether the decisions they made were in line with what (a) elite opinion and (b) public opinion thought correct, and (iii) how far, given the stochastic nature of the outcomes of public policymaking, those decisions were seen to be correct. Underlying all these judgements is luck. Luck, first, that a leader is perceived in a particular way given (i)–(iii) and second how those perceptions caused the leader to view their own leadership credentials, and thus make decisions in the future given (i)–(iii), and so on. Leadership qualities and environment continually interact, and who becomes r­ ecognized as a great public leader depends as much upon the elements of luck I have identified as on any potential leadership qualities they enjoy. Nothing in this analysis, however, denies that those who are ­perceived as great leaders actually were great, or that those who are ­perceived to be weak, actually were weak. It does deny, however, that we can have a recipe for advising how to become a great leader, or for picking future great leaders, without that recipe being equally one for disastrous leaders or predicting failures. It is far easier to advise on how to be safe than how to be spectacular – ­assuming one wants spectacular success rather than spectacular failure. Notes  1 There is also, of course, a literature on the rhetoric and strategy of leadership; but good rhetoric is dependent upon time and culture, and strategic ­considerations – such as those adopted by Riker’s heresthetic politicians – partly rely upon spotting openings (Riker 1986, 1996; and for British examples, see McLean 2001). These are interesting ways of exploring leadership that I do not consider here.  2 See Chapter 7 for details of the analysis of luck which I then direct at the issue of responsibility and just rewards in the context of ‘luck egalitarianism’.  3 A Danish study suggests that the adoption of children from one honest family into another gives the child a 13.5 per cent probability of getting into trouble with the law. This probability increases to 14.7 per cent if the adopting family

170 Luck includes criminals. Being adopted into an honest family from a criminal one increases the probability to 20 per cent; whilst if both adopting and biological parents are criminals the probability rises to 24.5 per cent. From this evidence Mednick et al. (1984) conclude that genetic factors predispose behaviour in criminal contexts.  4 I call the first ‘personal identity luck’ (Dowding 1991). I suggest (see Chapter 7) that personal identity luck is one aspect of Dworkin’s ‘brute luck’ (1981); the second (which we might call ‘context luck’) is the other aspect.  5 And indeed the simplistic idea of looking for markers to correlate with behaviour is not the best way to think about the relationship between genotype and phenotype. However, the simplistic account enables us to consider just how important chance environmental factors are in the development of genuine personal characteristics.  6 The idea of a ‘real pattern’ I use here is from Dennett (1998).  7 Most people might imagine Thatcher and Blair got what they wanted more often than Major; however, before coming to that conclusion one must look at the full records (see Young 1989; Major 1999; Seldon 2005).  8 The Daily Mail even suggested in April 1981 that her leadership showed that women were not up to such leadership roles.  9 Though, in retrospect, tactically she got it right by getting her men (with the exception of John Biffen, first Chief Secretary and then Trade) into key economic roles. This can be seen as an important leadership trait. Again, the Falklands War helped her, as she was able to shift Biffen to be Lord President (April 1982), moving the powerful Pym first to Foreign Secretary following Carrington’s resignation, and then out of cabinet following the subsequent general election. From April 1982 she had her people in all the key economic positions. 10 Brennan (1998) similarly argues that politicians who follow their stated principles are given credit in the form of ‘trust’ over those who simply follow majority opinion. What is surprising, perhaps, is how often politicians follow majority opinion when the risks seem high. Attacking the government over a risky policy like the war in Iraq might be thought to bring longer-term rewards than slavishly following the opinion of the day, if one’s assessment of the risks is given credence. The respective positions of Barack Obama and Hillary Clinton over Iraq might not have been unimportant in the former’s victory in the Democratic primaries in 2008. 11 The stochastic nature of most policy decisions means that ‘getting it right’ is often hard to judge. Therefore we often make judgements about people and their leadership qualities based on issues where it is easy to judge. Thatcher won the Falklands War. Did she win education policy? Or civil service reform? Or even economic policy? These are much harder to judge. Lesson for leaders: win some easy-to-judge decisions against prevailing opinion. (Opinion in the UK was split over whether to fight or negotiate over the Falklands, but certainly prevailing elite opinion was to negotiate.)

Part III

Freedom

9 Choice: its increase and its value

One of the great claims of Margaret Thatcher’s administration, continued by John Major’s, is that the choice of the individual has been increased. They have allowed individuals the freedom to buy their own council houses and they are introducing choice into the health service and into education. They have increased share ownership by selling state assets and have augmented choice through more widespread ownership of material goods and resources, as revealed by official statistics on ownership of televisions, videos and other consumer durables. This claim of increasing choice is often put on a par with increasing individual liberty – the two are mentioned in the same breath. Some philosophers, notably Hillel Steiner and Gerry Cohen, conflate choice and liberty; they define the scope of an individual’s negative liberty by the number of options which others do not physically stop her from choosing (Steiner 1974; Cohen 1988). This equates liberty with whatever one could choose to do. Steiner defends this conception as the only consistent version of negative liberty, even if it is one which most negative libertarians would not ordinarily wish to defend – which is precisely why Cohen develops it. However, liberty and choice are generally thought of as two separate concepts by philosophers who are often cited as part of the Thatcherite nexus. Hayek writes: The range of physical possibilities from which a person can choose at a given moment has no direct relevance to freedom. The rock climber on a difficult pitch who sees only one way out to save his life is unquestionably free, though we would hardly say he has any choice. (1960: 12)

and again: Whether [a person] is free or not does not depend upon his range of choice but on whether he can expect to shape his course of action in accordance with his

174 Freedom present intention, or whether somebody else has power so to manipulate the conditions as to make him act according to that person’s will rather than his own. (1960: 13)

Similarly, Milton and Rose Friedman do not equate freedom and choice in Free to Choose. Rather, government must guarantee liberty so that the choice provided by capitalism may be enjoyed: Unless there is … protection [from coercion], we are not really free to choose. The armed robber’s ‘Your money or your life’ offers me a choice, but no one would describe it as a free choice or the subsequent exchange as voluntary. (Friedman and Friedman 1990: 29)

Thus when Des King (1987: 38) describes the New Right as believing that ‘the free market system of social organization generates innovation and wealth in the most desirable fashion by maximising choice and liberty’, choice and liberty are two separate values. Nevertheless, having a choice is generally considered desirable, either intrinsically – as a value in itself – or instrumentally – as something which leads to other valuable things. Andrew Reeve (1990: 98) has suggested that individual choice ‘“straddles” the values of welfare and liberty’. It does so since we tend to think of the free person as someone making choices; the choice that one makes reveals one’s preferences, thereby giving some indication of the amount of individual utility or welfare provided by the options available. Reeve reviews the concept of choice in different strands of liberalism and concludes that the policy implications of the different views depend to some extent upon their conception of ‘choice’. He argues that those versions which model individual choice upon ‘market choice’ do not ‘capture the reasons for which individual choice has been valued’ (Reeve 1990: 99). I shall examine this claim, and also develop some of Reeve’s arguments about the different conceptions of choice and what this entails for public policy. I shall suggest, first, that the very notion of increasing, let alone maximizing, choice is inherently problematic Second, I shall suggest that even where choice has unproblematically been increased, it is not always certain that this is desirable. Finally, I find a disjunction between what we value about choice and what we value about the market. This suggests that when weighing the value of introducing a market into areas where it has not before operated, this value is only tenuously linked with increasing consumer choice. First, we must dispose of what we do not mean by ‘choice’. There is a sense in which everyone always has a choice. The rock climber on the difficult pitch can choose to let go, or try an impossible route. No matter what



Choice: its increase and its value175

incentives there are to act one way or another, the stubborn can always refuse to do so. Thus, if there is only one type of car available, we do not claim that consumers have a choice (with respect to cars) to buy a car or not to buy a car. Increasing consumer choice means giving them another positive option. Although difficult to represent formally, choice requires at least two positive options, a and b, rather than the negative choice between a and not-a. There are three main ways in which the word ‘choice’ is used in relation to liberty. The first means something like ‘opportunity set’. When we write phrases like ‘offering a wider choice’, we mean that one or more alternatives have been added to the opportunity set. Thus a choice between a and b becomes a choice between a, b and c. Or, when we suggest that consumers prefer having a choice to not having a choice, we mean that they prefer to have at least two alternatives in an opportunity set. This is probably the way the term ‘choice’ is most widely used. It is closely connected to its second form of usage, where ‘choice’ means ‘the act or faculty of choosing’. In relation to liberty, ‘choice’ is used in connection with individuals’ abilities to choose. A free person is one who is capable of choice; such capability may include certain conditions of rationality, skills promoted through education, as well as the relevant information on which to base a decision.1 A brief perusal of the literature, however, suggests that the term ‘choice’ is most often used in neither of these senses, but rather to designate that option from the opportunity set which the person actually selects. Thus, if I prefer alternative a in the opportunity set {a, b, c}, a is said to be my choice. Or the word ‘choice’ is used in this third sense to designate each of the options in the opportunity set: ‘which of the three choices do you prefer?’. The three uses of the term can be seen in the torturous sentence: ‘As someone capable of choice[2], my choice [3] from the choice[1] offered is x’.2 There is some confusion between the first and third uses of the term, but that confusion is not intrinsic and can be easily avoided by calling the first ‘opportunity set’ and the third ‘alternative’ or ‘option’.3 However, the conflation of the two uses does, I think, underlie certain problems for the interpretation of ‘increasing choice’, as I shall argue below. One may be capable of choosing and be a free person in that sense, but if one has a very restricted set of options, that capability will not be realized. Similarly, ‘choice’ as alternative chosen only makes sense when there are a number of alternatives available. When we talk about ‘increasing’ or ‘maximizing’ choice, which of the three uses are we referring to? It can be interpreted under all three, but with very different implications.

176 Freedom

Increasing choice The idea of increasing choice can be elucidated under all three conceptions of choice, although under some it seems to relate more to increasing the value of the options contained within an opportunity set. In the second sense, choice can be increased by increasing the abilities of people to choose or by increasing the number of capable choosers. Educating people, both by helping them to learn how to rationally consider and make decisions and by broadening their knowledge, will help in this respect. Less directly, but just as importantly, providing greater information about the same number of products in a market may help to increase choice in this sense. This is not what is usually meant by increasing choice. However, it seems to me that if choice is intrinsically valued (see below), then this is the sense in which it is intrinsically valued. We value choice under this meaning because moral beings are choosing creatures and increasing choice is thus more about creating conditions under which individuals are able to make decision than about actually offering a wider range of alternatives. But I shall say no more here about this conception of choice, since it is not in the mainstream use of the term and is anyway parasitic upon the ‘opportunity set’ interpretation: one cannot be a choosing creature unless there are a number of alternatives from which to choose. The only way we can make sense of ‘increasing choice’ in the ‘alternative’ sense of choice is by increasing the value of the chosen alternative. If, for example, my most-preferred alternative from those on offer is x, but x is replaced within the opportunity set by y, which I prefer to x, then my choice might be said to have increased because the value of the alternative I choose has risen for me. This version of increasing choice, however, seems a very opaque way of saying that I am getting something better than I otherwise would. ‘Increasing choice’ is often used in this sense, though usually conflated with the first, ‘opportunity set’, sense (as demonstrated below). If we have an opportunity set of {a, b, c}, to which is added d, to form {a,  b, c, d}, then choice has seemingly been increased. This is surely the least problematic version of ‘increasing choice’. But why should we merely want a larger number of alternatives to choose from, if the most-preferred alternative remains the same? Unless we gain utility through the act of choosing itself and the act is enhanced by adding alternatives, then the value of adding less-valued alternatives must be nil. If, for example, there are three options, of which a is preferred to b and b to c, but to such a slight degree that one would prefer a choice between b and c to getting a outright, then choice per se must have an intrinsic value.4 Gerald Dworkin (1982: 60) suggests this is irrational, although without saying exactly why. It is not irrational if choice does indeed have an



Choice: its increase and its value177

intrinsic value, although we might try to demonstrate formally that giving choice such an intrinsic value is irrational because it causes intransitivity by the following: {a > b > c} and ({b > c} > {a}) By removing the brackets it seems that a > b > c >a, which is intransitive and therefore irrational. But we have not justified the removal of the brackets. Neither b nor c is preferred to a individually; rather, the choice between the two is preferred to just having a on its own. The brackets were added to try to represent this and if choice has intrinsic value, then there is nothing irrational in this and no justification for removing the brackets, for they perform a role in demonstrating this. I think readers will have to decide for themselves whether choice has an intrinsic value for them in this sense. Would you swap something which you wanted more for something you wanted less, just to have a choice between that less-valued object and an even lower-valued one? It seems to me that choice in this sense has per se no intrinsic value, but without empirical evidence I would not care to generalize to the effect that my views are shared by most people, if not everyone. To me the only value choice holds is derivative. Indeed, adding an option to an opportunity set might be disadvantageous if it were not the alternative that one would choose. There are costs associated with choice. In order to make a rational choice one must consider the alternatives available and weigh up the pros and cons of each possibility. Often the costs of considering a further alternative might be negligible. Adding a new brand of toothpaste to those available would not make one’s choice much more costly. Indeed, if one were happy with the brand customarily used, one may not even notice the addition on the shelves. If one were to try the new brand and dislike it, the costs of the error are low. But, in certain areas of greater importance to us, a wider range of options, or even being given a choice at all, may not be welcome. The costs are just too high. For example, moving from a simple compulsory pension scheme (like SERPS) to opening up a compulsory choice between competing pension plans may offer new alternatives for individuals, which may be more highly valued for some of them, but the choice itself may not be valued at all. The trouble and stress involved in having to calculate which of the competing schemes is best for one, especially with long time horizons involved and their very great importance, may render the fact of choice unwelcome.5 Although adding alternatives to an opportunity set does not necessarily bring added value to it, we might still value the added options even

178 Freedom

though we do not actually choose them. Surely we can recognize that one opportunity set is better than another, even if the option we would choose is contained in both? Ordinarily, we would take it as a joke if a car salesman claimed to be offering greater choice than last week if, in the meantime, he had increased his stock by adding a dozen wrecks to the quality cars in his showroom – although, of course, a mechanic who fixed cars for a hobby might be pleased by this addition.6 But we would not take it amiss if he added a new range of models, even if it is not a range that appeals to us. The point here is that increasing choice in this sense assumes a range of preference scales in which the options are not all placed in the same order. The greater the range of products available, the greater the chances that all individuals will have their preferences satisfied. It seems, therefore, that adding alternatives, even irrelevant ones, must be said to increase choice, although it also follows that increased choice is not necessarily something to be valued for its own sake (for further arguments, see Barry 1989; Dworkin 1982). What are valued are the options contained within an opportunity set, and therefore the most-valued opportunity set should be one that brings the most value to society as a whole. This, however, has paradoxical consequences, for valuing an opportunity set in these terms means that it is possible to increase the value of an opportunity set to society as a whole by decreasing the number of alternatives and their total value to each individual chooser. Collective action problems take this form; that old favourite, the Prisoners’ Dilemma, is just one example. Consider a Prisoners’ Dilemma with the following payoffs: Prisoner 2 C       D C    7, 7     1, 8 Prisoner 1 D    8, 1     2, 2 If both try to get their most-preferred payoff of 8 by choosing the line D, they both end up with payoff 2, so the overall payoff to society is 4. If one chooses C (to get 7) and the other D to get 8, then one gets 8 and the other 1, so the overall payoff to society is 9. We know that rational players will play D, as it is the dominant strategy, so society will end up with the lower total payoff of 4. If, however, the option of choosing the line to get payoff 8 were to be removed, then both will choose line C in order to get payoff 7, with total payoff to society of 14. In order to achieve the greatest benefit to each person in society, we remove everyone’s most-preferred option from their opportunity set.



Choice: its increase and its value179

Changing opportunity sets Simply adding options to or subtracting them from opportunity sets is one way of increasing or decreasing choice. However, rarely in public policy do we simply add or subtract options; much more often we replace one alternative by another. This creates a far greater range of problems for evaluating the worth of choice. Consider the following opportunity sets: OS1 {a, b} OS2 {b, c} OS3 {a, b, c} OS1 and OS2 appear to be incommensurable, each offering two alternatives, but not the same two alternatives. OS3 is an obvious increase in choice over both. However, it seems just false to say that OS1 and OS2 are incommensurable if I am able to compare a to b to c and am able to say I prefer a to both b and c. What we might say is that OS1 and OS2 are incommensurable with regard to the choice they offer without further information about individual preference rankings. If we are able to compare the two opportunity sets and OS1 is preferred, then it may be said to provide an increased choice over OS2. This is the way in which the ‘alternative’ version of ‘choice’ is often thought to lead to increased choice – although, as I noted above, it is rather an opaque use of the term. But this analysis also introduces other problems. Consider the opportunity sets: OS3 {a, b, c} OS4 {a, b, c, d, e} OS5 {f} If I prefer f to a–e, then, on the above method, OS5 provides an increased choice over OS4. Again, this seems simply wrong when OS4 has five alternatives and OS5 only one. Surely there is greater choice in OS3 and OS4 than OS5, even if I were to prefer to have no choice and just receive f. It seems that to assess whether choice has been increased, we should only take into account the number of alternatives in opportunity sets and not their relative value. We might question, however, whether or not we value increased choice on this criterion. Why should I value an opportunity set with five alternatives over one with three, when the only ­alternative I want is in both? Why should I value the increased choice if I prefer one alternative in the smaller opportunity set to all five in the greater?

180 Freedom

Unless we can demonstrate that choice does indeed have an intrinsic value, we have no reason for valuing the number of alternatives at all. What I want is having those things I value the most, not having a choice among a set of things I do not value at all. Therefore, if we are to value increasing choice, perhaps we have to take account of the evaluation of the alternatives. But further, and more importantly, there are logical reasons why we may have to consider the value of the alternatives as well as their actual number. That is because their evaluation is a large part of the way in which we individuate the number of alternatives available in an opportunity set. This requires some explanation. The problem of individuation The formal analysis above has simply assumed that we have a set of alternatives of any given number, and the alternatives can be simply denoted by letter: a, b, and so on. Whether {a, b, c} is a set containing three options or a set of ten depends upon some principle for individuating them from one another; and that, it turns out, is less straightforward than such set notation (or indeed ordinary English) makes it appear. Consider the following twelve phrases:   1   2   3   4   5   6   7   8   9 10 11 12

Somewhere to live A two-bedroomed house A semi-detached house with a good aspect A house by the sea A dwelling needing extensive renovation A house with scope for modification A rat-infested hovel A house with spacious rooms A house offered for sale at half its market value A home with a large well-kept garden A home with easy access to a fast rail link A building 100 yards from a main railway line

These are twelve descriptions which could as easily be true of one building as of twelve. If they were descriptions of the same building, then we could not be said to be offering someone twelve alternatives if we showed them the list and said we would sell them the house they chose from that list. However, someone might want to buy the house on the strength of one of the descriptions, even if that entailed buying the other eleven, although they would not want to buy a building under any of the other descriptions. For example, it would be perfectly rational to want to buy something



Choice: its increase and its value181

which was (truly) being offered for sale at half its market value, even if it is a rat-infested hovel within 100 yards of a mainline railway. One might be asked, ‘Do you want to buy a rat-infested hovel?’, to which one replies, ‘Not really’. But then we are told, ‘It’s for sale at half its true market value’. To which someone with money to spare for a while may rationally respond, ‘Okay then, it’s a deal’. No matter how many ways a given object could be described, it is only one object. It is true that we may value an object differently under one description than we do under another, but this does not make it two objects. In other words, our evaluation of objects does not necessarily respect Leibniz’s law, which states that all objects are identical with themselves (Schick 1982). We cannot actually increase choice by redescribing some object, but we may well alter the way we evaluate the choice that is available. It seems to me that in order to decide how great a choice an individual has, we have to individuate options within an opportunity set by the manner in which the chooser individuates them, with the proviso that Leibniz’s law of self-identity is not broken. The number of options which constitute a choice to an individual is not simply the number she can differentiate one from another. It must include the condition that they can sensibly be evaluated differently, even if she does not evaluate them differently. I may be indifferent between several options and yet still recognize them as different options. This is not a trivial point. If I ask for a light for my cigar and am offered a box of matches, we would not want to say that I have a choice because there are 35 matches in the box. I am aware that the 35 matches are not the same match, but I am indifferent as to which match I use. However, if I am offered a box of matches by one person and a cigarette lighter by another, I have been offered a choice, even though, as it happens, I am indifferent as to whether my cigar is lit by a match or a cigarette lighter. The proof of this is that some people do prefer their cigars to be lit by matches rather than cigarette lighters, on the grounds that lighter fuel affects the taste of the cigar. Thus, although I am indifferent to the choice, I can recognize that a choice has been offered.7 One way of representing this is suggested by the distinction introduced by Edna Ullmann-Margalit and Sidney Morganbesser (1977: 757) between picking and choosing: We speak of choosing among alternatives when the act of taking (doing) one of them is determined by the differences in one’s preferences over them. When preferences are completely symmetrical, where one is completely indifferent with regard to the alternatives, we shall refer to the act of taking (doing) one of them as an act of picking.

182 Freedom

Thus, one picks a match from the box containing them, but chooses to use the matches rather than the cigarette lighter. (Of course, one may choose one match rather than another if one can see a difference: one match, say, is cracked and may break on striking.) Having weighed up the merits of each alternative, one can see no reason to choose one over the other, so one picks one. The weighing-up process can be considered a part of the choice or decision process, but once completed, if one is truly indifferent, then one just picks which alternative to have or do). Where once is truly indifferent between two alternatives, one has no reason for preferring one over the other and need not give reasons for picking one rather than the other. But if reasons can be given for choosing one option over another, then a choice has truly been offered – even if I merely pick one of them. Thus the ability to give a reason for choosing one alternative rather than another is what distinguishes a and b in an opportunity set. If ‘a’ and ‘b’ are just different descriptions of the same natural object, then I may have believed I had a choice, but Leibniz’s law of self-identity proves that in fact I did not. Note that one still has a reason for picking one or the other even where one is indifferent between them, as long as one is not indifferent between having one or the other and having neither. Buridan’s hungry ass, who could not tell the difference between two bundles of hay, is often wrongly thought to have been unable to choose either of them because he was indifferent between them. But if, in fact, the ass had been merely indifferent, he would have picked one of the bundles of hay. His not choosing either bundle demonstrates unconnected preferences (Sen 1982d). He did not connect his preference for food with his indifference between the two bundles of hay. The point of the story is that he had a very strong preference for eating over not eating, but did not connect this preference with the two bundles of hay before him. Despite being hungry, he was too stupid to eat because he became confused by his indifference. Having unconnected preferences over alternatives makes choice between them impossible. Often we have unconnected preferences because we have no idea what the supposed choice between different alternatives actually entails. Lack of knowledge of the implications of different types of food for our health means that we do not connect our preference for eating, say greasy fish and chips with our preference for staying healthy. In fact, the health market generally offers good examples of unconnected preferences. Individuals do not generally have well-defined indifference curves over the primary characteristics of health care, for they are often ignorant of their own illnesses. For many minor ailments this is not true. We are aware when we have a bad knee and are able to discover fairly easily whether a given doctor can help our knee problems. But for many more serious ailments we are ignorant of the progress of recovery or deterioration,



Choice: its increase and its value183

ignorant even of whether or not we are suffering. For example, the elderly do not always understand their own needs for health care, nor do individuals suffering from chronic depression. These more general types of consumer ignorance can hardly be overcome by the state, except through care orders put through the courts or under the law, but they illustrate a less dramatic ignorance of the nature of illnesses, which restricts rational choice of health options. Consumers are just not in a position to judge whether or not an operation is necessary for the problems they face, or which treatment would be best. This information can be provided by their doctors, but such information processed by patients has first been selected and edited by their doctors, who may choose to communicate it in a particular manner.8 Under either private or public provision, doctors generally advise patients on treatments and are averse to forcing treatment upon their patients. How much doctors tell their patients varies with what information they feel patients can successfully process, which is one reason why doctors tend to spend longer discussing diagnoses with their better-educated patients. The way to increase choice for patients here would be to enable them to have second or third opinions from different doctors, perhaps operating from different practices. Choosing one alternative over another, as opposed to merely picking one rather than another, is distinguished by the chooser being able to give reasons why one was preferred to the other. The distinction cannot therefore be revealed merely by watching another’s behaviour, although that is not to say that it might not be revealed by close study (Dowding 1991: ch. 3). Obviously the more one knows about the alternatives, the more reasoned the decision in favour of one rather than another. Indeed, if one knows literally nothing about two alternatives, then choice, as opposed to picking, is impossible. One important piece of information needed by an individual to choose between two supposed alternatives is that they are not in fact identical. But the individual also needs more information about an option than merely being able to distinguish it from the others in order to make a reasoned choice. She requires the information that would be most likely to lead her to choose, or not choose, that particular option. The more information one receives about the alternatives in an opportunity set, the more likely it is that one will satisfy this condition. The less information one receives about the alternatives, the more likely that our preferences for those alternatives will be underdetermined. An example of the underdetermination of preferences with regard to alternatives is that of the choice between schools for one’s children. All three major political parties in Britain have announced that they will aid parental choice by insisting that schools publish their examination results.

184 Freedom

Will this aid choice? It will do so only if the information made available is of sufficient quality, and it is doubtful that this can be achieved. I take it that parents want their children to go to the school where (among other things) those children will have the best chance of getting as good a set of examination results as they are able to achieve. Publishing the percentage of children who achieve high grades from the total of those entered for examinations will not demonstrate this if schools only enter those children who they think stand a chance of getting high grades. Your child will not get any grade for an examination if she is not entered for it. We know that some schools achieve a very high percentage of passes and high grades by not allowing weak pupils to enter for examinations. Thus, ceteris paribus, each child who goes to that school actually has a lower chance of passing any examination than she would at a school where everyone is entered for all examinations. Therefore, ceteris paribus, your child actually has a better chance at schools where the pass rate as a whole is lower. We could try to counter this tactic by requiring schools also to publish the percentage of students entered for examinations, as well as their pass rates. This would provide more information by which to judge schools, although again it may not actually reveal what parents really want to know. Results are one thing, but the quality of entrants is another. In fact, ‘parental choice’ over schools is a myth. All that can be offered is the right to apply to any school. Once catchment area barriers are abolished, schools will be able to select from those who apply and a pecking order of schools will develop, thereby ensuring that most parents will not send their children to their most-preferred school (Barry 1989). The problem of choice in education or health care or anything else is not really what we are interested in. The real question is whether the market provides the best educational or health system or whatever. We need to consider both the merits of the market outside of the choice issue and to answer our initial question about the value of choice itself. The value of markets Advocating markets on the basis that they increase consumers’ choice is doubly wrong: first, because choice is not necessarily something to be valued and, second, because increased choice itself is not what is valuable about markets. At best, increased choice is a by-product of what is valuable about markets. We value the market process because it provides us with what we want and, where it fails to provide us with what we want, we replace it with something else. The market often provides us with what we want because it provides us with a range of alternatives. From these



Choice: its increase and its value185

alternatives we make our choices, and successful products continue in production, whilst the unsuccessful disappear from the shelves. The market is a preference-revealing device. Having a wide range of alternatives in some market is instrumentally useful, for it enables a diversity of preferences to be revealed; but we value the range of alternatives because there is a wide range of tastes. Market diversity is good because preferences are diverse. If we all want exactly the same thing, then there is no point in having a wide range of alternatives. Dworkin (1982: 58) writes: One of the ways in which increased choice contributes to the welfare of individuals is by increasing the probability that they will satisfy their desires. People want various things – goods and services, status, affection, power, health, security – and their chances of getting these things are often enhanced if they have more options to choose among. My chance of finding a shirt I like is greater if I have ten different shirts to choose among than if I have only two.

This is an important reason for wanting more rather than less choice. I suggested that merely adding alternatives to an opportunity set does not add value to it if the additions are not ones the chooser wants. However, if we make no assumptions about the desirability of the options, then we would want to have a wider rather than a lesser range, since it increases our chances of getting what we want. This happens, I think, not only because the wider the range the greater the probability that our pre-formed preferences will be satisfied, but also because we can discover our preferences through the very act of choice itself. There is an ancient debate concerning individuals’ perceptions of colour. If we produced a gradual shading of yellow into orange into red, and then showed the sequence with one ‘stage’ missed out, would the perceiver notice it and know what that shade would look like? I do not think the debate was ever satisfactorily settled; but, either way, individuals do see shades for the first time and discover they like them. One hears people say, ‘I bought this shirt because I thought it was an interesting shade of green’, and so on. You discover you like something when you see it or try it. Beyond the trivial colour example, one can also discover preferences through the process of making a choice or decision. In a complex choice, where various reasons for and against the different alternatives compete for attention, one may actually discover which of the reasons weigh most heavily by the very act of making the choice. Making the choice reveals preferences not only to the observer, but to the decision maker himself. Choice can be an act of discovery. Indeed, for Joseph Raz the whole point of society providing a wide range of choices – both trivial and important – is

186 Freedom

that it allows the individual to discover herself. Choice is what brings autonomy (Raz 1986). There is another reason why markets are valued. The fact that valued products succeed, whilst ones which by comparison are less valued fail, gives a measure of control to the consumer. Their very act of making choices between alternatives leads consumers to control the type of products which are made available. Surely one of the reasons for valuing increased choice in areas of state provision is that it is supposed to increase the control of the citizen over that provision.9 Such control is greatest where it is easy to shift from using one variety of a product to another. It is least where shifting from one to another is costly. Such control is greatest where product comparison is easy, non-existent where preferences are unconnected. The question to be examined in any area of supposed increased consumer choice is whether alternatives genuinely lead to greater control. Where we are not sure about the value of the alternatives to us, then we may not want a choice. As Scanlon (1988: 179) writes: The same interest which sometimes makes choice valuable – the desire that outcomes should coincide with one’s preferences – can at other times provide reasons for wanting outcomes to be determined in some other way. When I go to an exotic restaurant with my sophisticated friends, the chances of getting a meal that accords with my preferences may be increased if someone else does the ordering.

Where markets fail to bring control to the consumer, then control must be sought by other means. This is precisely the justification for democratically elected bodies having oversight of institutions which make public provision. Hospitals and schools that have ‘opted out’ are being taken away from such control as once existed. The ‘control’ arguments require as careful an analysis as I have tried to give to ‘choice’ arguments. These comments suggest that, whatever the merits of Conservative policies, claims to have ‘increased choice’ need careful examination. Increased choice, as opposed to better products or efficient markets, is not necessarily something to be valued at all. First, the whole notion of ‘increased choice’ is problematic and, second, it is not obvious that we should always value it anyway. Rather, what we value is getting what we want. Markets are often good at that. They do it by offering us a choice of products; but it is the goods we value, not the choice itself. In any particular area of public policy, the usefulness of the market must be examined in relation to the ease of shifting from one alternative to another, the costs of making decisions and the ability of individuals to have clearly defined preference schedules. Whether or not it brings greater choice is not something to be



Choice: its increase and its value187

valued at all. The value of choice in the market is merely instrumental in that it enables preferences to be revealed or discovered. Unless one revels in making choices and prefers weighing up alternatives to actually having them, ‘increasing choice’ itself is valueless. Notes 1 The significance of choice for Scanlon, who recognizes the different ways in which the term can be used, arises from the fact that individuals have this faculty (Scanlon 1988). 2 Reeve (1990: 116) identifies four means of ‘choice’, but I do not see any difference between his first use – there is a set of options – and his third – ‘the opportunity to choose’ – for we would hardly describe a set of options as a choice for a person if that person were not going to have the opportunity to choose between them. 3 We should note in passing that the term ‘option’ is more wide-ranging than ‘alternative’. Saying that one has a choice among several alternatives entails that choosing one excludes choosing the others. Alternatives are mutually exclusive; no such implication is conveyed by ‘option’. Generally speaking, I shall use ‘alternatives’; occasionally I use ‘option’. Everything I say about alternatives can be extended, with a little revision, to options. 4 I owe this formulation of the possible intrinsic value of this notion of choice to Bob Goodin. 5 Hence the almost universal condemnation of plans in 1985 to abolish SERPS (Weale 1990: 205–7). 6 This caveat is important and will be returned to in the final section. 7 I might consider the reasons that sway others to be idiotic or almost unintelligible, but they count as reasons if we can recognize them as such. For example, when I was a student at Keele University, the Chancellor, Princess Margaret, came to the students’ ball, and soon reached for her cigarettes. The President of the Students’ Union rushed up with a lit match, but the Princess refused to suck. An aide then came forward with a lighter, informing the hapless student that ‘the Princess does not have her cigarettes lit with matches’. I can understand that someone may think themselves too grand to have their cigarettes lit with a match, even if I think their reasons are daft. 8 This point was forcefully made by Klim McPherson on the BBC programme Brass Tacks, February 1989. 9 Bob Goodin brought this aspect to my attention.

10 The value of choice in public policy Keith Dowding and Peter John

Introduction Since a major speech Tony Blair gave on 16 October 2001, Labour governments in the UK have pushed a choice agenda in public service provision (Perri 6 2003). They have argued that choice can improve service quality in many areas of public service as well as being an intrinsic good in itself. We define choice as being instrumentally valuable in the sense that increasing choice in public services brings welfare gains through efficiency by the signals that choice gives to providers (generally, though not exclusively, through market or quasi-market processes). We define choice as being intrinsically valuable if it is desired for itself, even though why it is desired might be further explicated (for example, choice enhances individual autonomy). (In that sense, any intrinsic value can be further justified instrumentally.) The choice agenda, both when first introduced by the Conservatives under Margaret Thatcher and in its new guise under Blair (where the intrinsic benefits have been highlighted), is controversial. How far market and quasi-market mechanisms bring efficiency gains has been questioned, though the mainstream view is that they have brought efficiency gains in some areas but not others. More importantly, it has been argued that increasing citizen choice empowers the middle classes more than the lower classes, and so is inequitable. We shall initially set aside the instrumental value of choice when discussing the value of choice; it re-emerges in later sections. We do not comment on the equity issue, not because it is unimportant, but because our focus is upon the claims of choice having an independent intrinsic value. We query the intrinsic value of choice as we define that term here. Even if choice does not increase efficiency, and even if it increases inequity, some might argue that increasing choice in public service could be justified on the grounds of its intrinsic value. Surveys have demonstrated that people say they value



The value of choice in public policy189

choice in service provision, and people in the lower classes seem to value increasing choice even more than people in the middle classes (Le Grand 2007). We interrogate the concept of ‘increasing choice’. We examine what increasing choice might mean, and show that it is surprisingly difficult to measure whether choice has been increased in any area of public provision in a non-trivial and uncontroversial manner. The dilemma of measurement is whether to simply count the number of alternatives (a so-called ‘cardinality’ measure) or whether to take into account the welfare gains of those alternatives. The former measurement can make increasing choice seem trivial and valueless; the latter means it is possible to increase choice whilst reducing the number of alternatives: that seems counter-intuitive (as well-being against government policy). The latter type of measurement also closes the gap between any intrinsic value choice might have and the instrumental value of choice as we have defined that term. If it is welfare gains by which we measure choice, then it seems choice has no intrinsic value beyond the efficiency gains by which we defined its instrumental value. We begin by explicating the measurement problem by informally describing some social choice results that demonstrate it. We then explicate what these results mean for government’s policy of attempting to increase citizen choice. We suggest that increasing choice by increasing alternatives does not necessarily empower people (Newman and Vidler 2006: 203–4), and perhaps the value of choice comes via empowerment. We define increasing alternatives (the cardinality measure) as ‘hard’ choice – hard in the sense it is easy to compile and provide targets. We define providing information, advice and helping people to make choices as ‘soft’ choice – because targetsetting and oversight is more difficult. Through the lens of the dilemma we have set out, and in terms of the hard and soft choice definitions, we look at some of the empirical attempts to measure whether the British government has succeeded in increasing choice. We find that the empirical attempts to measure gains in increasing choice largely rest upon welfare gains and not any supposed intrinsic value. Our analytic argument and review of the empirical literature together suggest that we should concentrate attention upon the welfare gains of any public policy (which should also include equity considerations), rather than any supposed intrinsic benefits of the choice agenda. What is choice and how might it be measured? In the social policy literature the very idea of increasing choice is not normally considered problematic. An examination of the formal literature from the social choice tradition demonstrates that in fact this is not so. The

190 Freedom

very meaning of ‘choice’ and how increasing it might be measured are deeply problematic (Kreps 1979; Sen 1988, 1990a, 1991; Pattanaik and Xu 1990, 1998; Dowding 1992; Arrow 1995; Sugden 1998, 2003; van Hees 2000a). We explain these problems here and suggest why this is a problem for the choice agenda in social policy. The central question posed in the social choice literature was set out in one of the early contributions by Pattanaik and Xu (1990). Their ‘flavour of an impossibility theorem’ (that is, a flavour of a logical contradiction) is fashioned by three axioms which intuitively seem plausible and produce a simple counting rule for measuring choice. However, the simple counting rule has counter-intuitive implications. In their presentation, Pattanaik and Xu (1990) present three axioms: 1 Indifference between no-choice situations: opportunity sets consisting of only one item yield the same amount of choice. An ‘opportunity set’ consists of those possibilities that a person might be able to choose in some domain. The justification of this axiom is that an opportunity set of only one item yields no choice at all, hence we should be indifferent between two sets each offering no choice. 2 Strict monotonicity: for all distinct alternatives x, y, the opportunity set consisting of both x and y yields strictly more choice than one consisting of only x or only y. The justification seems obvious. A set where I can choose from two items surely yields more choice than a set with only one of those items. 3 Independence: for all opportunity sets A and B and alternatives x which belong to neither A nor B, A gives at least at as much choice as B if, and only if, the union of A and {x} gives at least as much choice as the union of B and {x}. The idea here is that adding or subtracting the same element from any two opportunity sets should not affect their evaluation with respect to each other. That is, if we prefer set A to set B, then we should still prefer set A to set B when element x is added to each of them. Pattanaik and Xu (1990) show that these three axioms yield a unique cardinality rule where choice is measured simply by the number of items in an opportunity set. In other words, to measure choice we simply need to count how many items are open to the chooser. In a public policy context, the cardinality rule is applied when we say a choice of four hospitals is greater than having just one; or a choice of three schools is greater than having just two. However, some very straightforward examples demonstrate that this simple cardinality rule does not correspond to our intuitions about what constitutes an increase in choice. In Pattanaik and Xu’s



The value of choice in public policy191

example, adding the alternative ‘blue car’ to the set {red car, train} does not seem to add as much choice as adding ‘blue car’ to the set {bicycle, train}. In a public policy context, adding ‘Roman Catholic school’ to the set {Christian school, Muslim school} does not add as much as to the set {Muslim school, non-faith school}. Now note here that this is not a question of one set having more valued choice, but having more choice itself. To be sure, in both examples, there are two opportunity sets with two items in them. By the simple counting rule each set has the same amount of choice. But the question is whether they each offer equal choice. Do we consider we have a choice between identical cans of beer (Sugden 1998)? The problem is a deep one about how we designate objects so they constitute different objects of choice. The social choice literature offers two responses to this problem. One is to include ‘diversity’ within the counting rule; the other is to include the value of the alternatives within the measure. There have been several attempts to measure diversity of alternatives within opportunity sets (Bossert et al. 1994, 2003; Sugden 1998), which display various degrees of technical ingenuity. However, Martin van Hees (2004) demonstrates that they are doomed to failure. It is not possible to combine the cardinality rule with diversity without one rule dominating. The impossibility result can be illustrated with a simple example. Imagine measuring diversity along a single dimension. In Figure 10.1(a), the two items w and x are closer together than the two items v and z in Figure 10.1(b). It appears therefore that 10.1(b) provides greater diversity and so we might argue that 10.1(b) provides greater choice. Of course, it might be countered, even in this figure a voter might prefer the choice between a moderate socialist and a moderate conservative to one between a radical communist and a radical libertarian; so it is not obvious that 11.1(b) provides more choice than 11.1(a), let alone greater utility. The point is made even more manifest when 11.1(b) is compared to 11.1(c). Here we have three items w, x and w

x

(a) v

z

(b) w

x

y

(c)

10.1  Cardinality and diversity

192 Freedom

y. They are still contained within the bounds of the broader set 11.1(b), but there are now three items. Does 11.1(b) or 11.1(c) contain the most choice? Do we value three alternatives closer together or two further apart? Our answer might depend on where we locate our ideal or bliss point on that dimension. Do we want more options or fewer more diverse ones? In a public policy context, do we prefer the choice of three hospitals which all offer much the same surgical techniques or two hospitals that offer radically different treatments for the condition? Do we prefer a choice of three faith schools or one faith and one non-faith school? Diversity in this literature does not have to be measured in terms of utility. Indeed Bossert et al. (1994) base their measure on an objective function from the physical sciences. But the most obvious and important way would be to include valuation of the alternatives. If we believe that what matters is how much we value the options, then we are suggesting that the degree of choice is defined by the amount of utility we would gain from the choice we make. Thus freedom of choice is defined by what economists call ‘indirect utility’ (Kreps 1990b: 45–8). This is the utility the agent would receive through his most-preferred alternative at the time of making that choice, and that itself might depend on the choices others make. Some writers have suggested that freedom of choice should be defined by the indirect or potential utility, since it either provides ‘flexibility’ for choosers (Kreps 1979; Arrow 1995) or enables greater individual autonomy and community diversity (Sugden 2003; Bavetta and Gualla 2003). Van Hees’s (2004) result suggests, however, that any measure of freedom of choice must allow either number or diversity to dominate the calculation. In a public policy context, either the number of hospitals or schools, or the character of the alternatives, will determine how we measure the increase in choice. Either way is open to criticism, because of counter-intuitive examples, and they cannot be combined in any way that avoids such examples. We should note, however, that these ‘impossibility results’ do not demonstrate that we can never be sure we have increased choice in any policy domain. Such impossibility results demonstrate only such measures cannot be applied to all cases. They do not show that they cannot, non-controversially, be applied to some cases. Sometimes it is obvious that the change from situation A to situation B has, non-controversially, increased choice. Adding a non-faith school to a menu of faith schools noncontroversially increases choice, for example. It is only when the opportunity sets are not subsets of each other that problems might arise. Scarcity, however, ensures that the choice agenda cannot satisfy that requirement. The conceptual problems we have considered concern unconstrained choice. Once feasibility constraints are added to the mix, further problems emerge. One is individuation (see Chapter, 9, pp 180–4). People value



The value of choice in public policy193

objects under different descriptions; thus how a choice is described also evaluates that choice. Consider choice of schools. Imagine a situation where parents could order the set of schools for their child. If there were too many subscribers for any particular school, final entry would be by lottery, with those not attaining their first choice entering the lottery for their second choice, and so on. For parents whose main concern is entry to the school with the best GCSE results, such a system might seem fair. If they fail to get their first choice, they stand a chance of getting their child into their second choice, and so on. However, parents who want their child to go to the nearest school would lose out. If the nearest school was also the school that had the best GCSE results, their ability to choose the local school for their child would be reduced by comparison with any system where ‘catchment area’ plays a major role. So the implementation structure increases choice for some parents (by the cardinality rule), but reduces it for others (by reducing the probability of their successfully choosing their most-preferred alternative). The increase for some at the expense of others is due to the different, though reasonable, criteria we have for measuring increased choice. Enhancing choice under one criterion (number) reduces choice under another criterion (evaluation). Thus the spaces under which choices emerge and under which they might be evaluated become important. Another example of where reducing choice whilst increasing it happened to one of the authors. When he had a minor operation, prior to the introduction of the government’s choice agenda, he and the surgeon agreed a mutually convenient date. After the implementation of choice, when undergoing the same operation, the author had to be asked which hospital he wanted; once that choice was made, he was simply assigned a date. It might be possible to introduce a choice of surgeon/hospital without losing the choice of dates, but since the latter was not an explicit part of the government’s choice target implementation in the author’s health area, this resulted in one choice being substituted for another. As it happens, the author was indifferent over choice of hospital (or, rather, wanted his local hospital – which he would have got anyway), but much more interested in having some control over the precise day of the operation. Of course, such an anecdote only illustrates the problem; it does not show how general it is. The value of choice Choice might be valued instrumentally – for what it brings – or intrinsically – for what it is. The instrumental value of choice is that it gives signals for providers that increase both allocative and productive efficiency. Market efficiency is driven by choice. With a large number of consumers and multiple providers, consumers can choose products in terms of their type,

194 Freedom

quality and price. Given heterogeneous tastes, a large number of different types of product can be produced and some firms can specialize. Product variety occurs due to these market signals. Furthermore, such allocative signals also drive productive efficiency, since price, as well as type and quality, matters to consumers. Thus consumer choice drives both allocative and productive efficiency. Quasi-market processes are supposed to mirror the competitive pressures and produce these efficiencies. The main difference between quasiand standard markets is that in the former the organizations are not developed to provide profits. How this affects their incentives has been much discussed and will not be repeated here (Le Grand and Bartlett 1993; Hardy and Wistow 1995; Flynn 1999). We can see that competition in the form of contracting-out has undoubtedly produced efficiencies in some areas of the public sector, such as refuse collection. However, generally speaking, as we learn from transaction costs economics (Williamson 1975, 1996), such efficiencies are most readily achieved in areas where contracts are easiest to specify. Their success in the health market has been difficult to demonstrate (Propper et al. 2006). Similarly, whilst Tiebout competition has been shown to operate in the sense that households do move to try to attain better tax–service packages, no one has definitively demonstrated efficiency gains through that process (Dowding et al. 1994; Dowding and Mergoupis 2003; Donahue 1997). Thus, whilst hard choice might provide efficiency gains, demonstrating this outside of public services for which simple contracts can be written has proved problematic. However, if people value choice in itself, then providing it even without efficiency gains would still be justifiable. Why might choice be intrinsically valuable? If one prefers alternative x to all other alternatives, why would one prefer the set {x, y, z} to {x, y} to {x} (Chapter 9)? Add any costs choosing might bring, and then surely the preference revelation is likely to be contrariwise. Indeed a Which? survey suggests people would prefer ‘access and quality’ to more choice in the areas of pensions, health and education (recounted in Le Grand 2007). The Labour government also discovered this in focus group research when planning their 2005 election campaign (see Guardian, 27 March 2004). And, as Le Grand points out, this finding can only show that people do not intrinsically prefer choice to quality, since choice may be the process by which quality is enhanced. The only reason people might prefer more choice to less if they end up with the same product anyway is that they enjoy choosing. Schwartz (2004: 18), for example, reports that 93 per cent of American female teenagers state shopping to be their favourite activity. Or perhaps the security and control that choosing provides is preferable without actually being pleasurable. It has also been shown that people



The value of choice in public policy195

have stated preferences over the process by which health care is provided, as well as the success of the care itself (Salkeld 1998; Ratcliffe and Buxton 1999; Brouwer et al. 2005). Such preferences are sometimes called ‘process utility’. People undoubtedly have preferences for the processes of product provision as well as for the product itself and they gain utility from choice. One reason that choice might have intrinsic value is that we discover our preferences for outcomes through the act of choosing. It is only by having a menu of alternatives that we discover that we prefer one to the others and what features of those products we value the most. Choice might be valued intrinsically in another sense: choice is constitutive of autonomy. It allows autonomy, or at least leads people to believe they are autonomous (Bavetta and Gualla 2003; Sugden 2003). It gives people the feeling that they are at least partly in control of the public services they consume, rather than having them forced upon them. There might be value in the autonomy that choosing brings. People would prefer to be consulted by their GP on the type of treatments they might receive and what specialists they might see. They would prefer to be given the set of alternatives by their specialist or surgeon, once it is established that surgery is a possible solution to their problem, rather than simply be assigned a treatment by their GP or be told by their surgeon that they need an operation. Treating people as autonomous beings able to make decisions for themselves – once the alternatives, the risks and possibilities are laid open to them – might be preferable for most people, even if they recognize that by choosing for themselves they might end up with a treatment bringing lower welfare – say in the form of pain relief – than if the doctors simply assigned a treatment to them. That is one way that choice, or autonomy, might have intrinsic value. And if choice is to have any intrinsic value above and beyond instrumental welfare gains, then there must, at least in theory, be examples where we prefer choice even if it brings lower value than the outcome produced by another procedure. Furthermore, people are habituated to having more choice in the private sector, from television stations, to holidays, to types of camera, phones, bread, sweets, and so on; and where the internet allows people to work through complex options through clicking boxes on a website rather than reading vast manuals. In virtually every area, the private sector offers more options than in the past, and so people might come to expect choice from the public sector too. It is not so much a question of what they value as what they expect. On consideration, someone might agree that television, in some ways, was more fun when there were fewer options; at the same time they might still be surprised (and complain) when they find only five channels on their hotel TV set. Dissatisfaction has as much to do with expectations as with the quality of the product. Thus choice might be indirectly

196 Freedom

valued. A person might give no intrinsic value to choice – she would not swap one opportunity set A for another larger set B if her top alternative in A was only fractionally more valuable to her than her top alternative in B – but would still be disappointed (her expectations for choice unfulfilled) if A did not have the range of choice she was expecting. In terms of choosing opportunity sets, prior to knowing the value of the alternatives, she would always prefer the larger set (with the caveat on size given below). The costs of choice So much for the potential benefits of choice: what are the costs? We might identify four types: (1) welfare costs, (2) information costs, (3) transition costs and (4) psychological costs. Welfare costs Welfare costs have already been encountered: whilst patients feel more autonomous by choosing their treatment, it is possible that leaving everything up to the professional might bring greater welfare overall in terms of, say, quality-adjusted life years (QALYs). In fact, stated preference evidence suggests people are willing to swap welfare in QALY terms for that autonomy (Olsen and Smith 2001). Similarly, having choice over pensions might make some better off, but bad choices by others mean that a single state or company pension brings overall welfare gains. In order to see if the efficiency and autonomy gains outweigh potential welfare losses, we might simply look at the costs and benefits of increasing choice. Information costs Information costs can be so high that increasing choice by adding alternatives actually makes choice impossible. Imagine an opportunity set of infinite size. How could one choose one’s most-preferred alternative? A set that big is too large even to comprehend, let alone find out information about each alternative to distinguish them, far less decide which is best. One could only subdivide the set in some manner and then rationally choose between the alternatives in the subset: a form of ‘picking’ (Ullmann-Margalit and Morgenbesser 1977). Thus adding options to large sets need not add any instrumental value and indeed might subtract such value, since subdivision through picking may remove the best options. Furthermore, sets large enough to confuse people would neither have intrinsic value nor increase autonomy. Transition costs We can think of practical examples of rational ‘non-choosing’. Sometimes it is claimed that many consumers are ‘not rational’, since they insist



The value of choice in public policy197

on sticking long-term with their current utility providers: ‘By shopping around the average consumer could save themselves £40 a year on their gas and electricity bills’. Now, assuming ‘shopping around’ needs to be done annually, is not itself enjoyable, and takes around four hours to collect the ­relevant information, work out the best deal and administer the switch, then it is not rational to shop around if the marginal value of a consumer’s time is more than £10 per hour. It should be noted that the example shows an individual is not necessarily irrational if she does not shop around for the best utility provider. It shows that there may be no individual advantage for a consumer to spend time switching between utility providers, even if she could reduce her annual bills. It does not show that opening up utility provision to market forces does not bring consumer advantages. Using a similar example, Schwartz (2004: 25) claims, ‘the problem is that state regulators aren’t around there anymore to make sure that consumers don’t get ripped off. In an era of deregulation, even if you keep what you’ve always had, you may end up paying substantially more for the same service.’ Well, you might. But equally, deregulation might bring competitive advantages which drive down prices overall as companies become more efficient in a competitive market. But hang on, the critic says, ‘how can the gas market be efficient if consumers cannot “rationally” choose between companies?’ Well, the example only shows that it is not rational for consumers who value their time at more than £10 per hour; there might be plenty who value it at less. Furthermore, markets do not need to be driven by ‘rational’ consumers. Companies often have marketing drives to persuade people to switch to them, and some consumers do – perhaps because there are initial price advantages, even if these disappear as their original supplier fights back and reduces its prices, so the consumer would have been better off in the long run not switching. And it would not be rational to switch back, since they would have to spend too long finding out the precise costs and calculating the respective benefit. What matters for the market overall is that there are some switchers – rational or not – and that companies believe they are fighting for their market share and do so by becoming more efficient. Thus all consumers, even those who never shift providers, might gain through market processes. In fact, it has never been established, even theoretically, what number of informed consumers is required to drive a market. It has been found that only 5 per cent of furniture buyers and 8 per cent of appliance buyers gather much information prior to purchase (Claxton et al. 1974). Teske et al. (1993) argue that a small number of informed parents can create a market for state schools. The task of demonstrating the value of choice in an empirical study is horrendous.

198 Freedom

Under this argument, it is the market that has provided the efficiency gains; and the market requires there to be choice. But no individual may value choice in itself, and no individual need gain instrumental value through their own choice (that is, all ‘switchers’ might have been better off sticking with their initial provider had the market operated as it did with all those who switched), even though everyone has gained through the operation of the market. More simply, choice can be instrumentally rational for a community, even if it is not instrumentally rational for any one member of it. And it may be more rational for some communities than others, when choice encourages some groups to cluster around the provision of a good, accessing the improved efficiency from voice or random spatial variation in service quality – meaning that the middle classes ‘club together’ to get the better schools and hospitals. There is a large literature on the equity impacts of choice and, depending on the nature of the study (see the review of the evidence by the House of Commons Public Administration Committee 2005), the evidence seems to point in both directions. Some proponents of choice accept the idea of user selection via cream-skimming, but argue that overall welfare benefits outweigh this. Others, such as Julian Le Grand (2007), suggest ways of trying to end cream-skimming. Psychological costs Finally, we have ‘psychological costs’. There are many ways in which greater choice might bring psychological costs. Some might be result from information costs. Schwartz (2004: 19–20) recounts an experiment in a US gourmet food store. The experimenters displayed a set of exotic jams: customers were invited to taste the jams and received a coupon for a dollar off if they bought a jar. Under one condition, the consumers were offered a choice of six jams; in another a choice of 24. In both conditions customers could spend their coupon on any of the 24 jams. The larger display attracted more people, and about the same number tasted about the same number of jams. However, around 30 per cent of people who had a choice of six jams bought some; only 3 per cent of those who faced the array of 24 did so. A second experiment in the laboratory, involving chocolate, produced similar results (Iyengar and Lepper 2000). We might speculate that there are greater costs involved in working out which jam one prefers from the larger array. Furthermore, choosing from the larger set might diminish enjoyment of the one actually chosen. Consumers might feel regret or have doubts about their choice, knowing there were other jams they had not tried: in other words, being forced to pick leads to regret that one cannot rationally choose. Schwartz (2004: 122) suggests that if we assume that opportunity costs take away from the



The value of choice in public policy199

overall desirability of the most-preferred option and that we will feel the opportunity costs associated with many of the options we reject, then the more alternatives there are from which to choose, the greater our experience of the opportunity costs will be. And the greater our experience of the opportunity costs, the less satisfaction we will derive from our chosen alternative. In addition, some households have cost-free or cheap leisure time where choosing can be part of the pleasure (akin to long trips round shopping centres), whereas other households may have more pressures from day-to-day life – numbers of children, meeting bills, dealing with crime, facing poor health, giving support to relatives – making choice less enjoyable. Hence for them it is more rational not to make choices – i­ronically the very households that are likely to consume most public services. Lower education may also decrease the value people place upon choice, because scrutinizing the documentation takes longer. With greater mobility and access to wider networks, more affluent groups may have access to greater ‘soft’ information to help them make more informed and less costly choices. What does this type of psychological cost mean for the introduction of choice into public services in the UK? Not much, perhaps: after all, choice over hospitals has increased to three or four. This is more like the array of six than twenty-four jams. But, again, choice means less for the poor who are more likely to want their closest producer – something they would receive anyway. It might, though, have more relevance for everyone for other choices. The Conservative government in the 1980s encouraged the growth of choice in pension plans by inducing employers to close down simple company schemes, usually based on a share of final salary, and allowing the financial sector to open up private schemes offering a wider variety. Here people faced a much broader set of alternatives. Moreover, pension schemes, owing to their base assumptions, mathematical derivations and complexity, are notoriously difficult to understand. Here the psychological costs of choosing might be great, not only because of the variety, but also the complexity and the importance of the decisions. The shift away from state- or company-provided schemes to a mix of company, state and private ones also places much greater responsibility upon individuals, who might experience stress, both whilst making the decision and subsequently. As evidence for this type of stress, Schwartz (2004: 116) reports a survey where, of those who did not have cancer, 65 per cent said they would prefer to choose their treatment; but of those who actually had cancer, 88 per cent said they would prefer not to choose their treatment. When it comes to the crunch, people would sooner not take responsibility

200 Freedom

for a decision that brings pain and might make an important difference to their chances of survival. Of course, people can feel regret even when they do not make choices. One may regret choosing one cancer treatment rather than another. But one may also regret not having a cancer treatment that one only learns about later: which regret is likely to be greater is moot. ‘Introducing choice’ into public services does not always introduce new possibilities; rather, it enables people to have some input into selecting among the available options. In that sense, bringing choice into public services is not necessarily introducing regret: it might simply change its character. Having said that, there is a great deal more stress involved in choosing a medical treatment, or school, than in a choice of jam. Introducing choice into public services We have seen that increasing the number of alternatives (‘hard choice’) does not necessarily entail increasing choice in any valued sense; indeed, it may induce costs. How hard choice is implemented may also decrease hard choice over alternatives not covered by regulatory accounting procedures. We need to examine the space of choice. Where hard choice is induced through market or quasi-market procedures, efficiency might be improved. There have been few quantified attempts to measure the effects of choice in public services and those that have been conducted have generally not sought to control for other factors affecting the quality of output. (For some reviews of the evidence, see Williams and Rossiter 2004; FarringtonDouglas and Allen 2005; Institute for Political and Economic Governance 2006; Le Grand 2007.) In this section we review some increases in choice in British public services from the conceptual perspective we have adopted. Education In the UK parents have had greater choice over schools for their children since 1988. This has generally been popular and it would almost certainly be controversial if government were to attempt to reverse this facility. A survey of parents found that 90 per cent were satisfied with the outcome and 70 per cent were satisfied with the process of school choice (Flatney et al. 2001: 15). In 2007 an attempt in Brighton to bring greater equity into the process, by introducing a lottery system for the most popular schools, caused some local outcry. Satisfaction with the outcome of the choice process is certainly due to the fact that 90 per cent of parents get their first or second choice of school for their children. Variance in satisfaction depends upon the complexity of the process as implemented by different local education authorities. And lower rates of satisfaction were found



The value of choice in public policy201

among parents in London areas where there are shortages of secondary school places (Flatney et al. 2001: 15). Health Whilst there has been some examination of choice in the health market in England, it is not at all clear how generalizable these results are. There is little evidence as to whether the introduction of choice over hospitals has been successful. The London Patient Choice Project (LPCP) shows hard choice was welcomed, but this was a special initiative for patients who had been waiting for more than six months for treatment; it involved help with the costs of getting to and from the hospital chosen (Dawson et al. 2005). Patients were also provided with Patient Care Providers (PCA) to advise them. If the costs of going to hospitals farther afield are defrayed, and that as a result the waiting time can be drastically reduced, choice is welcomed; but what if those conditions do not obtain? Indeed, the LPCP may only show that when more resources are put into a service, satisfaction increases. Furthermore, evidence from other European countries shows that the referring doctor’s advice is more important than the patient’s views in hospital choice (Thomson and Dixon 2004), though this may simply be a result of patients heeding expert advice in areas where they feel they are out of their depth. Patients in Manchester were more likely to make their own choice if advised by a PCA (Barber et al. 2004), and those in London not taking up the offer of choice were on the whole less satisfied (Coulter et al. 2005) – though the causal path is not obvious here. It is also not obvious if the doctor’s ‘choice’ is actually being replaced by the PCA’s. Providing information is a key element of patients taking up choice opportunities (Barber et al. 2004; Le Maistre et al. 2004; Coulter et al. 2005). On the other hand, in the US, fear of litigation has meant some doctors will not advise patients and simply list the options (Le Grand 2007). Furthermore, crude measures of ‘success rates’ for surgeons might be misleading and lead to ‘gaming’ (Hood 2002), where doctors choose not to operate on high-risk patients in order to preserve their success rate. Le Grand (2007) examines various strategies for trying to encourage what he calls ‘knightly behaviour’ rather than gaming to poor targets. So the evidence for the success of hard choice in the hospital service is weak at best. Choice in health care should be more about soft choice – ­discussion with doctors and care professionals about the types and appropriateness of treatments available to that specific patient. An international study found that, of the six countries studied, the UK was poorest at involving patients in soft choice (Schoen et al. 2004). Some doctors may underestimate patients’ ability to make choices and manage choice procedures

202 Freedom

themselves (Farrell 2004). The LPCP found that under a third of eligible patients were actually offered a choice by their consultant (Coulter et al. 2005). Another study found that out of 24 users across four localities, 60 per cent were given no information about the options presented to them; of those who received domiciliary care, none was offered a choice of provider or sector (Knapp et al. 2001). Medical training now teaches patient involvement (Modernising Medical Careers 2005), something that implies that introducing soft choice is a long-term strategy. The need for information and the complexity of information processing suggests some patients need more guidance than others. In order to maintain equity, it is necessary to ensure that professionals are rewarded for helping those most in need rather than punished by crude productivity measures. Choice and quality Has choice also helped to improve quality? Here the evidence is even sparser. There is some evidence of progressive outcomes in some services from enhanced choice schemes, especially in relation to maintaining a mixed clientele, and some gains in efficiency and responsiveness. Parental choice for secondary school selection may have led to more middle-class parents remaining within state provision rather than using private schools (Flatney et al. 2001). The 2001–3 Choice Based Letting experiment in social housing increased social housing take-up, especially among working households and ethnic minorities, and was valued as transparent and open, even though the scheme was relatively complicated (ODPM 2004). One large-scale study concluded that parental choice had increased competition between schools locally, with schools responding to consumer demands and failing schools closing (Bradley et al. 1999). According to Perri 6 (2003), between 1993 and 1997, efficiency gains occurred, with many schools improving exam results and attendance – both of which are difficult to link directly to parental choice and are more likely to be the result of a package of reforms providing incentives for schools to improve (Bradley et al. 1999). The regulation of secondary school places limits on responsiveness and competition. Only large and well-resourced schools are able to bring major innovations to the curriculum, and parents have had limited scope and resources to set up new schools. Here, centrally directed standards stand in the way of greater flexibility and choice (DfES 2005). In terms of younger children, the Nursery Voucher Scheme shows that (quasi-) markets might in fact reduce choice. The range of alternatives decreased through parents making choices to send children to schools rather than to nurseries and childminders. With their larger budgets, schools were able to expand in response to higher demand, resulting in private and voluntary-sector provision declining in some areas. Again,



The value of choice in public policy203

however, the effect of regulation of the pre-school curriculum may have been a major contribution to the narrowing of choice through this process (Education and Employment Committee 1997). Conclusion If efficiency is really the aim, then introducing competition might be justified where contracts are relatively easy to specify. Where they are difficult, the benefits of competition are less obvious. Where quality gains have been achieved, it is not always possible to judge whether it is top-down performance management and target-setting or any new managerial quasi-market processes that are responsible, since the government has been pressing several buttons simultaneously. Even then, bureaucratic gaming might ensure that the targets are reached rather than that quality has improved overall. If increasing choice is the aim, then markets can reduce choice if some providers go out of business or some services are dropped through low demand. Whilst surveys suggest people want better services rather than choice per se, choice might drive better services through competition. In some areas, hard choice seems popular: for example, in schools, though even here obtaining one’s first choice is important to its popularity. We have argued that the conceptual problems in measuring choice mean that it is extremely problematic to claim that increasing hard choice is valued. We have queried whether and to what extent choice is always desirable. We have also examined some empirical evidence about the successes and failures of the choice agenda in the UK. And we have noted just how difficult it is to really examine those successes and failures. We will conclude by suggesting some criteria to help in these judgements. First, we need to consider the nature of choice offered. Often, adding extra alternatives may also exclude some and those excluded might be the most valued. Second, market and quasi-market choice might reduce options (as some providers go out of business) as well as increase them. What is the real objective: efficiency or wider choice? Third, the most important aspect of the choice agenda is implementation. Patient choice is about offering real alternatives that bring advantages to people, not simply a menu of alternatives. For that reason ‘soft’ expansion in the form of medical practitioners talking through potential procedures and letting patients have their say in their treatment may well bring much greater benefits than ‘hard’ expansion of targeted numbers of alternatives. Soft expansion requires training and providing incentives for ‘knightly’ behaviour (Le Grand 2003), but is much more important than the hard expansion which might benefit only a few people. Encouraging ‘knightly’ behaviour necessitates ensuring, at the

204 Freedom

very least, that it is not discouraged by simplistic targets. GPs and hospitals need to be encouraged to spend time with their most difficult patients, so data on the social and medical backgrounds of patients are required when judging ‘throughput’. Similarly, judging schools simply on the basis of the examination results without factoring in students’ socio-economic backgrounds encourages cream-skimming. Advantage Premium is a scheme to give extra funding for children from a failed school (O’Shaughnessy and Leslie 2005), whilst Bowles and Gintis (1998) suggest a voucher scheme where the voucher has variable value depending on the socio-economic background of the pupil and the socio-economic composition of the school (see also Le Grand 2007). Fourth, costs and benefits need to be addressed. Would the money spent on implementing the choice agenda be better spent elsewhere? Even more importantly, what are the costs and benefits to consumers? Choice can be stressful. In some areas – such as pension policy or major surgery – the benefits of choice to some might be outweighed by the stress it causes others. Of course, patients always need to give consent for treatment, but doctors might judge that directing some is preferable to leaving decisions in their hands. Again, the choice agenda might best be left soft, with street-level bureaucrats given discretion to decide how best it can be implemented. The idea that we should not have a public service where ‘one size fits all’ can be applied to the process of offering choice, as well as to the idea of offering more than one alternative. Most of our arguments on the criteria for judging choice rely upon the idea that choice is not represented by Pattanaik and Xu’s (1990, 1998) cardinality rule. Policies that simply lead to a menu that enables the government to state that consumers now have a choice of possible alternatives whereas once they had none may add little to public welfare. Indeed, given the costs of implementation, there may be welfare loss. Rather, choice must be viewed more broadly in terms of the welfare benefits it might bring, efficiency gains through competition and information, and the feelings of autonomy it enhances. These are all to be encouraged; on the other hand, one must be aware of the costs that implementing choice brings about. Doctors spending more time with patients, having PCAs to help with ­decisions – these cost money and mean lower throughput. And choice can be costly for the chooser too. Choice does not come free and those costs need to be examined carefully. However, with a fully rounded analysis and the right preconditions in place, the choice agenda might well be welfareenhancing for the British state in the years to come.

11 Republican freedom, rights and the coalition problem

Introduction In recent years the concept of republican freedom has been put forward most notably by Quentin Skinner (1990, 1992, 1998, 2008) and Philip Pettit (1993b, 1996, 1997, 2001, 2002).1 Initially, Skinner saw republican freedom as a type of negative liberty that emphasized certain aspects or normatively important features (Skinner 1990). Now he and republican freedom’s other great defender, Philip Pettit, proclaim republican freedom to be a different and superior conception of freedom. Negative freedom is freedom from interference. Republican freedom is freedom from domination. Freedom from domination is a broader notion than mere interference, according to republican writers, because whilst an actor i might not be interfered with in her pursuit of some activity x, she is able to do x merely through the grace of some other actor j. The paradigm example is the slave who is allowed to pursue various activities beyond his duties to his master, but only through the master’s indulgence. For example, Sarah might be able to fly her kite, go to the shops or walk down to the riverbank when she is not required by her mistress to perform her duties. Similarly, Fred might also be able to fly his kite, go to the shops or walk along the riverbank when he is not required to carry out his duties as required by his employer. The difference is that Fred’s employer has no power or control over Fred in his free time. He cannot stop Fred from flying his kite if that is what Fred chooses to do. But Sarah, a slave, is only able to fly her kite due to the kind disposition of her mistress. If her mistress chose, Sarah would have to stay in her room when not performing her duties. Despite neither of them actually being interfered with when flying their kites, Fred is free; Sarah is not. Sarah’s ability to fly her kite is ‘arbitrary’, subject to the capricious will of her mistress. Fred’s ability is not arbitrary and is subject only to his own will. Pettit suggests that this makes Fred’s ability to fly his kite more resilient or robust than Sarah’s, on the grounds

206 Freedom

that it is not subject to the arbitrary will of others. Since the republican concept of freedom tracks his abilities and not Sarah’s, this means that republican freedom is more resilient than negative freedom. Contrary to this republican account, negative libertarians, represented by the ‘pure negative libertarians’ who have recently become republican freedom’s most trenchant critics (Carter 1999, 2008; Kramer 2003, 2008), say that Sarah is less free than Fred overall, even though she is free to fly her kite. Fred is free to fly his kite whether his employer approves or not. Sarah is free to fly her kite only when her mistress approves. Sarah has less freedom because she does not have the freedom to fly her kite when her mistress disapproves. She does have the freedom to fly her kite, but her overall freedom is lower. Carter (2008) suggests there is an equivalent judgement over most freedoms and losses of freedom between republican and negative freedom. In most cases they will agree when freedoms are gained or lost by acts of people or through the nature of constitutions, laws or even conventions. There are cases where they disagree. For example, Pettit (though not, it appears, Skinner) holds that constraints that promote the interests or welfare of a person do not constitute unfreedoms, nor need those constraints result from a contract freely undertaken. Pettit (1997: 56, n. 3) even suggests that the prisoner duly jailed under the rule of law has not been made unfree to do what he would have done if not constrained by his prison bars. Rather he has been made non-free to carry out some activities. Republican freedom, at least as discussed by Pettit, is thus trivalent, contrary to most accounts of negative freedom (and notably the pure negative account upon we concentrate), which is bivalent. I propose to ignore that (positive) element of republican freedom and concentrate upon the negative aspect of republican freedom – those elements about which Carter believes negative and republican freedom reach equivalent judgements on gains or losses in freedom. I ask whether republican freedom really is different from and superior to negative liberty, concentrating upon two charges of the negative libertarians. The first is that republican liberty is moralized. Republicans deny this: they deny that republican liberty requires a specific theory of justice or morality or a Kantian notion of liberty as rational behaviour. However, the charge of moralization made by negative libertarians is a weaker charge than that. Republican liberty can be considered moralized in part because of its trivalent nature, which depends upon normative judgements about the nature of external constraints.2 If some constraints are not to count as liberty-reducing, then we have to have an account of what constitutes justified and unjustified constraint. That is a form of moralization. A second form is introduced when some aspects of freedom are considered to be more



Republican freedom, rights, the coalition problem207

valuable than others. I suggest below that some comments of Pettit introduce welfarist considerations into the measurement of liberty. I focus, however, on the concept of domination. Without a very clear analytic account of domination, distinguishing negative and republican liberty is problematic, especially in relation to the claim that republican freedom is more resilient than negative freedom. I consider the ‘coalition problem’ in terms of republican liberty, arguing that what at first appears to be an easily answerable problem becomes a major obstacle for the attempt to distinguish republican and negative liberty with regard to their resilience. I end by suggesting that the real distinction marked in the paradigm master–slave example is not a distinct conception of freedom, but the familiar liberal distinction between freedoms and rights. Sarah is free to fly her kite when her mistress does not object, though unfree when her mistress objects; Fred is free to fly his kite no matter what his employer thinks about the matter. There is a sense, to be discussed in the penultimate section, that Fred has a right to fly his kite, whereas Sarah does not. To the extent that individual rights entail greater duties on others than do mere liberties, then, under the right institutional conditions, they might be more resilient. Interference Republican freedom is freedom from domination and not merely freedom from interference; but surely someone who dominates must interfere? How can I dominate you unless I interfere with you somehow? Certainly some of the discussion of domination in Pettit and Skinner suggests such interference in the sense that those dominated often have to fawn on those who dominate them. Sarah only gets to fly her kite because of the way she pleads with and pleases her mistress. However, Pettit also argues that this is not a necessary component of domination. One loses republican freedom if one is dominated, even if one is not aware that one is dominated.3 For the pure negative libertarian, if someone does x, then they must have been free to do x. Doing x proves you are free to do x, albeit under the circumstances in which you were able to do it. So, for a pure negative libertarian, if you achieve x under trying circumstances – say, despite the best efforts of your dominant master – you were ‘free to achieve x despite the best efforts of your dominant master’. For the republican, merely achieving x does not demonstrate you were free to do x. This is so because if your action in doing x has the wrong character, then even if you succeed in doing x, you cannot be held to be free to do x. Non-republicans likewise have criticized the pure negative account in this regard, suggesting that succeeding in doing x is not sufficient for one to

208 Freedom

be said to be free do x, arguing that doing x has to have the right character (Dowding and van Hees 2008b). However, the republican criticism of pure negative libertarian freedom is not identical to that account. In Dowding and van Hees (2008b), doing x does not count as being free to do x if one’s doing of x was due to the will of another, and could not have been accomplished due to one’s own will. (If one did x due to the will of another, that does not count as an exercise of freedom and does not prove one is free to do x, though if one could have willed doing x, then one would have been free to have done x. At best, doing x under the will of another is only evidence that one might have been free to do x.) The republican account is very different. Here the doing of x can still be an act of choice of one’s own will, and might even be described as an act freely chosen. However, it does not count as freedom if the act one carried out could have been stopped by another person or group of people. If it could have been stopped by the choice of others, then the act was not a free one. It was done only with the grace and agreement of others. At first sight, explained in this way, the republican account of freedom appears hopeless. After all, virtually any action that one does could be stopped by another person or group of people. Depending on the precise situation in which a person i is at any given moment – whether walking on a lonely hilltop or crowded street, say – it might be physically impossible for some person j or group of people {j, k, l} to stop a person from taking the next step or thousand steps. However, pretty much anything we do or plan to do could be stopped by some coalition of others. In that sense, any action x a person i takes is the success of a coalition of people including i, since for i to be successful in her action x requires the non-performance of certain actions w, y z or x by a coalition of others (Braham 2006). In a negative libertarian sense, therefore, the republican account of freedom seems to suggest that no one is free (beyond trivial actions of the next few moments), since they could always be stopped by a coalition of other actors. I call this ‘the coalition problem’ for republican liberty. Pettit (1997: 54–5) recognizes the coalition problem and answers it in his account of domination, discussed below. However, I shall argue that whilst republicans have recognized the coalition problem, it cannot easily be dealt with in a nonmoralized fashion. Domination and freedom measurement under coercion Republicans will proclaim that of course they do not characterize any and every potential interference with the performance of some action x as a restriction on freedom. The fact that a coalition of other ordinary folks could indeed stop a person i from performing an action x does not in



Republican freedom, rights, the coalition problem209

itself demonstrate that i is not free to do x. Pettit (1997: 23) says, ‘interference need not involve the exercise of a capacity for arbitrary interference, only the exercise of a much more constrained ability’. The precise details of this ‘more constrained ability’ are not spelled out, however. Pettit goes on to argue that there are three aspects of domination, to the extent that some agent has dominating power if (a) they have the capacity to interfere, (b) on an arbitrary basis, and (c) in certain choices that the other is in a position to make (Pettit 1997: 52). The third condition provides a scope aspect to domination; thus an employer may dominate his employee in the workplace but not outside it. I do not discuss that condition since it seems straightforward. The first two are of the utmost importance in Pettit’s account, however, and I address both in several different contexts. In a dictionary sense, domination means the exercise of ruling power, sway or ascendancy. This suggests that a dominator need not always get his way, but that his will tends to prevail. In a continuing relationship, for example, where sometimes one person prevails, sometimes the other, in roughly equal proportions, neither would dominate. Where the relationship between two people is a one-off and one of them must prevail (say there is conflict with no compromise), the nature of the decision over who prevails would determine whether or not there was domination. Obviously, tossing a coin (or indeed any agreed process to reach a decision) would seem to preserve equality. Where there is a battle of wills, but one prevails, there need be no domination. Only where we see a power relationship measured by some means outside of the outcome itself – their relative resources, say – where we think the more powerful is almost bound to prevail, might we think there is domination. The specific conditions are important to our judgement. We note here that the probability of the dominator prevailing is a key element in the identification of domination. Now, a person j might tend to prevail in a relationship without our suggesting that j dominates i. For example, i might simply recognize that j has superior judgement in various matters, and hence tend to go along with j. Many authority relationships are like that: the follower does what the authority suggests for content-independent reasons. That is, the follower  i does action x, not because i has specific reasons for doing x but rather because the authority j has told i to do it. The content of the reasons for doing x belong to j, not i, whose reasons for doing x are contained in j’s orders or suggestions. However, if i follows j’s orders willingly, because i has recognized j’s superior judgement in these matters or believes that in these circumstances a leader is necessary and j is the legitimate leader, then i’s action is still willing and j need not be considered a dominator. In that case, i can be considered free in doing x.

210 Freedom

We perceive j to be a dominator if i does the actions unwillingly or, if willingly, j’s authority is illegitimate in some regard. Either way, we are moralizing the role of the dominator. In the first sense, because we are not merely noting the physical actions of i in doing some act (it has to be a willing act) and in the second because we are labelling j’s ascendancy as illegitimate – a moral claim. Pettit might suggest that he only needs to refer to ‘avowed interests’, and whilst these are values, they are not moral values. In other words, there is some third way between purely physical descriptions and moral values. However, any account in terms of avowed interests still requires a moral theory, such as the welfarism I discuss below. The second way is more strongly moralized than the first, and I return to that aspect below. Let us consider coercion in the calculation of freedom. So the first way in which the pure negative libertarian and the republican writers differ is over whether i is free to do x when i does x unwillingly because of the coercion of j. The republicans state that i is not free to do x. The pure negative libertarians say that i may have done x unfreely, but doing x proves that i is free to do x. The loss in freedom that might accompany i’s doing x consists in the freedom to do any non-x, say action y, that i could have done had she not been forced to do x. In other words, both the pure negative libertarians and the republicans agree that i has lost freedom when doing x; they differ in their characterization of what constitutes that loss of freedom. In Carter’s terms, they have equivalent judgements about the case qua loss of liberty, but differ in the nature of that loss of liberty. It is true that in any specific measure of freedom, the pure negative libertarians and the republicans might differ in details, but agree in the relative freedom judgements across different societies. The pure negative libertarians would allow that the action x performed under coercion would enter into the count of the list of actions which i is free to do. The republicans would not. If x were the only action i could perform (say otherwise she would be killed and unable to perform any action), then in the pure negative libertarians’ calculation, the overall freedom of i would include the entry of one item, namely x. For the republicans, there would be no entry into i’s overall freedom, since x would not count. However, in relative terms, comparing i’s situation when forced to do x with i’s situation when not so forced (because the dominator does not exist), the calculations might not differ importantly. Both recognize a loss of freedom for i in this situation relative to other situations. Ian Carter (2008: 63–4), for example, sees no important difference between the two accounts in this regard.4 Philip Pettit perceives an important difference between republican and pure negative accounts with regard to coercion, believing that the republican account fits better with how people ordinarily view coercion and better with standard decision-theoretic analysis. The most obvious problem with



Republican freedom, rights, the coalition problem211

the pure negative theory, according to Pettit, is that it denies the most straightforward explanation of what has happened when there is a threat: that is, that one option in an opportunity set is replaced by another. He characterizes the pure libertarians’ ‘main thesis’ by the negative claim that freedom of choice is not affected by anything other than the removal of an option: I interfere with you and impact on your freedom only when I block you from doing something; I do not have any impact on your freedom to choose between various options just by coercively threatening, for example, to punish [you for] your choice of one or another alternative. (Pettit 2008: 119)

For instance, for pure negative libertarians, in the highwayman example, with the threat ‘Your money or your life’ what you lose is a conjunctive alternative – that of keeping-your-money-and-keeping-your-life – and thus the threat reduces your ‘overall’ liberty. It does not affect the alternatives in your opportunity set of ‘keeping your money’ and ‘not keeping your money’. These remain possibilities. Pettit (2008: 119) says: I follow the standard common decision-theoretic view that the choice is a set of mutually exclusive, jointly exhaustive options and that in the contingent context of the choice … each option is one that the agent can just choose: it is within his or her power of choice.

Since the conjunctive option is not in the original set, the threat does not reduce that original opportunity set, even if affects ‘overall liberty’ outside of the original set considered by the example. Pettit’s claims here are somewhat problematic. He believes his way of considering the matter is closer to standard decision theoretic practice. But that is not the case. When the highwayman says, ‘Your money or your life’, he is, according to Pettit, altering one of the options in your opportunity set. Now, standardly in decision theory, the ‘mutually exclusive, jointly exhaustive options’ are called ‘alternatives’, signalling that in the menu of choice, within the opportunity set, only one can be chosen. Pettit seems to think that the opportunity set once the highwayman has intervened is x, y. That is, you have two alternatives. In one (we will fondly imagine, following the literature) you keep your money, but lose your life; in the other you lose your money, but keep your life.5 So what are the two opportunity sets we are considering? Assuming one can alienate one’s money and commit suicide, the person has, prior to the highwayman’s intervention, the alternatives of keeping-your-money-andyour-life; not-keeping-your-money-and-keeping-your-life; keeping-yourmoney-and-not-keeping-your-life; and not-keeping-your-money-and-notkeeping-your-life. These are the four ‘mutually exclusive, jointly exhaustive

212 Freedom

options’ pertinent to the issue in question. After the highwayman’s intervention, according to the usual analysis, the first alternative disappears, the others remain. The standard decision-theoretic analysis favours the pure negative versions, not Pettit’s analysis of the situation.6 Pettit’s way of dealing with the highwayman example is to suggest that the identity of the alternative changes after his intervention. He defines ‘changing an alternative’ thus: and if that option is changed in a way that engages your values – whether or not these are the right values, by some independent metric – then it is thereby made into a different option … I will replace an option before you if I take steps that, by your own lights, makes a different evaluative profile. (Pettit 2008: 121; see also Pettit 1997: 53)

But what does this mean in the highwayman example? It is not immediately clear from Pettit’s discussion, since he does not work the example through with his preferred way of considering alternatives. The most obvious way in which a highwayman affects one’s options is to replace the alternative ‘keep one’s money and live’ with ‘keep one’s money and be killed’, with the latter perhaps evaluatively different from ‘keep one’s money and kill oneself’. Now, it is possible that if we were to take a poll of people on how to analyse the highwayman example, suitably described, as many or more might choose Pettit’s preferred description over the pure negative libertarian’s preferred one – though I have not conducted such a poll, and the pure negative libertarians might be sceptical that people would plump for Pettit’s description once the beauty of the negative libertarian’s description had been explained to everyone. However, even if we were to give Pettit his point – that his way of describing the event is closer to what ordinary folk might think – his way of describing the issue brings a bigger problem for the republican account in terms of measuring liberty. A problem that further moralization is required to avoid. In standard decision-theoretic analyses of the measurement of freedom, we cannot say under Pettit’s way of describing the highwayman’s intervention that this intervention has reduced freedom. Or, at least, we cannot do so without introducing some form of moralization. For example, if we say that the highwayman replaces the option ‘blue car’ in our opportunity set {blue car, red car} with the item ‘black car’, to give us the alternative opportunity set {black car, red car} we have the same number of items as if he had not intervened.7 To say that we prefer the alternative ‘blue car’ to ‘black car’ is to enter an evaluative element into the measurement of freedom. It also leads to the well-known measurement problem (Pattanaik and Xu 1990, 1998). What if the highwayman removes ‘blue car’ from the opportunity set to replace it with ‘black car’ and ‘green car’? We now



Republican freedom, rights, the coalition problem213

have an opportunity set with three items {black car, green car, red car} in it rather than two {blue car, red car}. By Pettit’s account, if blue car is the preferred alternative of the person we are considering, then anyone removing that option and replacing with a wider range has reduced our freedom. To say the least, it is not obvious that such actions are libertyreducing; this has spawned a large literature (see Dowding and van Hees 2009). Republicans might wish to accept this way of characterizing the measurement of freedom that suggests individual welfare needs to be taken into account when measuring freedom. And, indeed, their positive account of freedom where non-arbitrary intervention in the interests of a person does not restrict liberty is suggestive of such a welfaristic measure of freedom. It might be claimed that the threat of killing you if you do not hand over your money and, by Pettit’s account, thereby changing one’s opportunity set, is of a rather different order than changing the colour of a car in an opportunity set. However, Pettit is claiming that his analysis is closer to standard decision-theoretic analysis of options than the pure negative one. To claim that a different order of evaluative importance changes the standard decision-theoretic analysis requires some work that he has not yet engaged upon. Furthermore, it would require that the greater evaluative element can be explained in a manner that does not moralize freedom (if, that is, the republicans wish to continue with the claim that their notion of freedom is non-moralized), and it is not obvious what form such an argument could take. It is not clear why someone who replaces one opportunity set with another has necessarily affected one’s freedom, except where one is a subset of the other, when under any simple counting rule freedom would thus have been reduced. However, where one is not a subset of the other, simple counting rules cannot distinguish across opportunity sets without bringing in some evaluation of the alternatives. Any such analysis is moralized in the sense that it relies upon the argument that the items (overall) in one set are worth more than the items in another, no matter what their number is. I return once more to the problems of a welfarist justification of republican freedom below. Republicans want to distinguish between different types of cases based on the intentions of the person altering the opportunity sets (Pettit 1997: 52; also citing Miller 1990). For example, if the change from blue to black cars was due to the interference of a company pursuing higher profits by offering only black cars, as black paint is cheaper than blue, then we might judge a person’s opportunity set has altered, but this has no important effect on their overall freedom. However, if the intervention was due to the arbitrary will of the intervener – he did not like blue and was determined that blue cars would not come into his field of vision – then the intervention

214 Freedom

might be one of dominance. I consider these sorts of issues below, but note here that this kind of argument is not represented in standard decisiontheoretic analyses, and it is that claim that is discounted here. By standard accounts, changing options rather than removing them does not reduce freedom by cardinal measures, unless it reduces it in terms of diversity of alternatives or of individual welfare.8 Domination and reasonable expectations Another attempt by republicans to distinguish their account of liberty from the pure libertarian involves the claim that a person might be dominated without being aware of it. Someone might be unfree to fly his kite at time t, even though he is able to fly his kite at that time t, but unaware that there is a dominant agent who could stop him if that agent chose to do so. If the dominant agent had different desires, the person could not fly his kite at time t. The negative libertarian might respond that in that case the individual still loses the conjunctive freedoms as in the highwayman or master–slave examples above. However, this claim by republican writers, taken at face value, produces the ‘coalition problem’. The coalition problem arises because there is always a set of people J who could stop any agent i from doing x at any time, if they chose to do so. Pettit discounts the coalition problem by distinguishing actual coalitions from virtual ones. The distinction between actual coalitions and virtual ones is a key element of Pettit’s argument that requires close scrutiny. There are two ways an actual coalition might be distinguished from a virtual one: through organization or by structured preferences. The first overcomes costs problems, given potential desires to thwart an agent from performing x at time t; the second leads to dominating behaviour without organization. However, the republicans, if they are to be true to some of their claims, must discount the second set of dominating behaviour as domination leading to lack of freedom. This gives the republican account a counterintuitive air. We need to look at the analytic account of domination given by Pettit more carefully than we have thus far. He says that ‘One agent dominates another if and only if they have a certain power over that other, in particular a power of interference on an arbitrary basis’ (Pettit 1997: 52). The arbitrary basis is that which is based on the whim, or we might say the preferences, of the dominator. Arbitrary does not imply non-predictable; rather that their power can be used if they so will it and that it is unconstrained by other forces. Pettit goes on to say that the dominating party has to be an agent, not a system or network, but can be a corporate entity



Republican freedom, rights, the coalition problem215

(Pettit 1997: 53). As we saw above, he formally defines domination by saying three conditions must hold: a person has dominating power if (a) they have the capacity to interfere, (b) on an arbitrary basis, and (c) in certain choices that the other is in a position to make (Pettit 1997: 52). These three conditions are sufficient for domination, according to Pettit (1997: 58), though he goes on to argue that domination generally requires common knowledge: that is, the dominator and the dominated must be aware that the dominator has the power to interfere on whim (Pettit 1997: 63–4). The exception to this claim is backroom manipulation (Pettit 1997: 60), though Pettit does not develop this account. Skinner also argues that, whilst common knowledge is not required for the lack of republican freedom, it is generally present. The coalition problem for the republican account is that for any individual i at any given time, there is always a coalition of others {j, k, l} who could stop i from doing x. This is not a problem for the pure negative account, since this fact only discounts a set of conjunctive freedoms, and such a set is discounted for all i, thus reducing everyone’s relative overall freedom by the same amount. It can thus be ignored within the measure of freedom. However, it is a problem for the republicans’ account, since they want to discount someone as being free to do x even when they do x: that is, a coalition could stop them but does not in fact do so. One potential solution to the problem is that such coalitions restrict freedom by trivial degrees, given their low probability of formation. As the probability that a virtual coalition might become an actual one rises, the degree of freedom would decrease. Pettit and Skinner cannot make this move since they believe that the possibility of arbitrary interference, even when there are low probabilities, is still domination. This coalition problem is such a serious objection to the republican account that republicans have to entirely discount possible coalitions as being dominating agents. The objection may not appear to be a serious one, since it seems obvious that what Pettit calls a virtual coalition is not a dominating one. The fact that my neighbours could gang up together to stop me crossing the road does not mean they are my dominators. I do not worry about such a virtual coalition becoming an actual one, and I do not change my behaviour because of that possibility. However, given the republicans have discounted considering probabilities of actual interference in their account of domination, analytically distinguishing the virtual from the actual is not a trivial exercise. Consider crossing the road. Often when we cross the road, we can see cars that could run us down. Car drivers could speed up, they could swerve to hit us, and so on. They do not, and we do not expect them to and behave accordingly. The fact that car drivers do not dominate pedestrians

216 Freedom

(even though pedestrians must be wary of them) is based on expectations. Indeed, aggressive pedestrians can cross roads, forcing car drivers to take their foot off the accelerator or brake in order to avoid them. The pedestrians know that the car drivers do not want to knock them down. Why is that? For one thing, most car drivers have some care for others and simply do not want to hurt pedestrians, even if they annoy them when crossing the road. For another, it is not in the self-interest of the driver – their car might be damaged, they may have to stop and their journey be delayed, the police might be called and if some fault is found with the driver, then further costs must be paid. This analysis all looks satisfactory from the republican viewpoint. There is a duty of care backed up by state sanctions that ensure the more powerful actor – the car driver – does not act arbitrarily in deciding whether or not to impede pedestrians. Indeed, Pettit (2001: 38–9) considers that actual domination occurs where ‘virtual’ control exists. This is a case of anticipated reactions upon the part of the dominated, even where the dominator is unaware of the fact of these anticipated reactions. Consider the paranoid pedestrian. He is convinced that car drivers are attempting to run him down. He is convinced whenever he sees more than two people together that they will try to knock him to the ground and steal his goods and chattels. He is convinced that these virtual coalitions are actual ones, and he behaves accordingly. What are we to make of the paranoid pedestrian? Is he dominated by others? On the surface, Pettit’s conditions are satisfied. The drivers (a) have the capacity to interfere with (c) the pedestrian crossing the road, and could do so (b) on an arbitrary basis. And given that the drivers realize this, even though they have no intention of doing so, it is common knowledge. Moreover, this knowledge makes the paranoid pedestrian change his behaviour. True, the drivers might face state sanctions, but they could certainly get away with it sometimes. We could simply argue that the paranoid pedestrian is indeed paranoid and has unreasonable expectations about the behaviour of others. In other words, we can place a ‘reasonableness’ criterion on the expected behaviour of more powerful agents. It is not always reasonable to expect someone with the power to make arbitrary decisions in some aspect of your lives so to do; nor is it reasonable to expect some potential coalition so to do. However, that is the upshot of Kramer’s (2003: 135–41) example of the giant who has the resources to dominate a village, but because of his gentle nature does not do so. The point of the story is that, given his disposition (and in one variant he even takes to the hills to get out of the villagers’ way), the villagers have no reasonable expectation that he will interfere with them. Kramer argues the giant does not interfere with the villagers’ freedom, even though he could do so. His dispositions towards using his power and the villagers’



Republican freedom, rights, the coalition problem217

reasonable expectations about his likely behaviour ensure that there is no constraint on their freedom. Kramer (2003: 141–2) expands the example into one concerning Lennie and George, characters not dissimilar from those in Steinbeck’s Of Mice and Men. Lennie is big and strong and could dominate George, but it so happens that George’s wit and irascible disposition lead us to think the direction of domination is the other way round. It would seem, as Kramer suggests, that character or personality could itself be part of the nature of dominance. The republicans claim that people might only appear free due to the whim of the dominator, but if that whim is part of their character and it is their character rather than other physical resources that makes them dominant, then surely changing their nature must change them into non-dominators. The challenge to republican freedom here is twofold. First, dispositions cannot be removed entirely from an account of domination; and second, expectations about behaviour are important in determining whether or not someone loses freedom because of some unequal power relationship with another. I take each in turn. How important are dispositions not to interfere with others in the freedom we enjoy in our society? It might be argued that it is not because they do not have the disposition to do so that car drivers do not run down pedestrians or shoppers mug each other. Rather, it is the system of enforceable laws that ensures the reasonable running of society. No doubt our system of law provides strong incentives for law-abiding behaviour and the law would tend to break down without enforcement. However, this is not the full story, since people do not break laws on occasions when they could and almost certainly would get away with it. A large part of our law-abiding behaviour is dispositional or habitual. And surely that is not unimportant in an account of domination. Do we want to say that a big person is a dominator whenever he could get away with forcing someone to do something? Or a car driver whenever there is an open road with no witnesses? We certainly do not behave as if that were the case. Pettit does not agree with this characterization of domination. He says: I do not see if someone powerful is endogenously restrained in some measure from actually interfering with others – restrained, say, by a shift of attitude or habits – then any potential victim of interference is liberated in corresponding measure: the measure, presumably, in which he or she would be liberated if the restraint came from increased protection. The potential victims suffer a loss of freedom, so it is suggested, only in direct proportion to the controller’s probability of actually interfering. (Pettit 2008: 123)

In other words, domination would be reduced more by having greater protection on country roads (say closed-circuit cameras operating on all

218 Freedom

stretches of road) than by encouraging the moral disposition not to attempt to run down pedestrians whenever one could reasonably get away with it. Surely, however, creating dispositions and reasonable expectations about the behaviour of others is an important cement of society that ensures freedom as robustly as any system of state-enforced sanctions. Many of us would consider a society with fewer closed-circuit cameras a freer society than one with many, even though the latter might control behaviour more tightly. Pettit discounts the idea that habits or dispositions can themselves be part of a system of dominance that reduces freedom outside of any organized coalitional behaviour. However, consider the place of women in society. Many societies exist where women are considerably under the thumb of men, both legally and by custom and habit. In many societies, and indeed in western societies in the past, women were constrained in social and political activity by law. They could not vote, had limited property rights and restricted rights in their own person and over their children, and so on. These legal restrictions have increasingly been removed, especially in the West, but to some degree the world over. However, even as legal restraints have been removed, the domination of men over women has not commensurately declined. Women might be protected by law in many degrees from discrimination in property rights, in the workplace and in the home, but habitual domination from dispositions instilled by past practices is not so easily removed. Whilst the increased protection of the law should not be underestimated in consideration of female emancipation and liberty, neither should lesser domination through habitual discriminatory behaviour. One of the most important aspects of liberation comes from habitual and dispositional change, often coming a generation or two later than constitutional and legal changes. But this latter aspect of liberty seemingly cannot be countenanced by the republican notion of liberty as defended by Pettit. To put the point in its crudest form, according to republican liberty as espoused by Pettit, it would seem that two societies that had the same constitutions, regulatory institutions and enforcement practices would have the same amount of liberty, even if, in the first, one class of people (women, say) had to continually resort to the law if they wanted their liberties respected, whereas in the other they rarely had to do so. After all, in both societies the men have the same capacity to act, they have the same physical powers and the legal constraints are identical. The only difference in the two societies is in the dispositional nature and habitual behaviour of men. I believe most people would agree that women had more liberty in the second society; and furthermore this could be captured by measuring the conjunctive actions they are able to take.



Republican freedom, rights, the coalition problem219

Dispositions cannot be removed from an account of domination, as they are an important aspect of the agential account, and domination is not simply a measure of the agent’s power as given by her resources. Domination is set not only by the terms of the relative resources, but also by reasonable expectations about how those resources might be used in different contexts. Women in abusive households come to hold expectations about the behaviour of men in those households; as do women in non-abusive households. In some households, men are dominators, in others they are not; and that fact is not entirely due to the relative physical strengths of the two parties, nor to state-enforced sanctions. It is due to the behaviour and expectations of both parties; their dispositions to act in one way rather than another. Moreover, learning those dispositions is surely a vital aspect of a moral and a free non-dominating society. Expectations are relevant to Pettit’s distinction between virtual and actual coalitions. When does a virtual coalition become an actual one? Consider coalitions in legislatures. Coalitions of legislators are usually considered to be sets of legislators who vote together. Whenever a set J of legislators vote together, they form an actual coalition. For some such coalitions, the fact of voting together is all that constitutes this actual coalition. They might be members of parliament who have never voted together as a bloc before and will never do so again, but came together since they shared preferences on this issue and this issue only. We would not consider that this coalition dominates the parliament, even though it got its way on this particular vote. However, some coalitions are long-lasting and stable. Sometimes groups of legislators vote together on all issues, and if they are in a majority they might be considered to dominate the parliament. (Groups that are smaller than a majority can dominate too, if they form a pivotal position between two other coalitions.) What distinguishes a stable and long-lasting dominant coalition from the transitory is the stability of preferences and the structure of organized interests in a parliament? Distinguishing these behaviourally is not unproblematic (Krehbiel 1993), since the organizational structure as an explanation implies that legislators vote together despite a heterogeneity of preferences, but people form political parties because of shared interests and some homogeneity of preferences. Nevertheless, analytically there are two separate bases of dominant coalitions: party organization and homogeneity of preferences. This would make it appear that the one reason a virtual coalition is not an actual one is the fact that its members do not share preferences. Pettit’s account of the distinction between potential and actual coalitions might seem robust with regard to organization, but potential coalitions might soon become actual where individuals share interests and quickly

220 Freedom

see that the actions of an agent i might be incompatible or threatening to those interests. Here the expectation that some person might act in a way contrary to a potential dominator would determine that the coalition would indeed take on the dominating role. However, the expectation that the individual i might take some action x is surely determined in part by whether i or other people of i’s type (call them the I group) are likely to take that action x. Think of relationships between different ethnic groups. They might live in harmony whilst one group does not believe another will have many members who will carry out action x. However, if there is an expectation that some threshold number of I group will do x, this leads a coalition within the J group to behave in a manner that dominates members of I. We cannot discount expectations and subsequent dispositions to act as separate from the characterization of dominance. Pettit recognizes such possibilities, saying that in such cases: There is virtual domination, we might say, but not actual domination. Virtual domination may be something for republicans to guard against, of course, because of the future danger it represents. But it does not yet constitute the central evil to which they are opposed. (Pettit 1997: 54–5)

However, the problem is not that we need to guard against actual domination rather than virtual domination; it lies in distinguishing the two. Is actual group interference domination? The coalition problem teaches us that potential group interference cannot be considered domination. But does actual group interference always count as domination? Say i wishes to buy the latest Dan Brown thriller, but is unable to because the stocks have all been purchased by others. Republicans would not be forced to say that i is unfree to buy The Lost Symbol just because others had got there first. The group of people who bought all the copies does not count as a dominator, nor does the person who bought the last copy in the shops. Negative libertarians are committed to this unfreedom. The person buying the last copy has made i unfree to purchase that copy. The ‘interference’ here is that others have legitimately used their own freedom to buy, which unfortunately has meant the shops have run out, leaving i disappointed.9 For republicans, this does not count as domination. If, however, a coalition of people conspired deliberately to stop i from buying The Lost Symbol, the republicans might agree that i has been made unfree. But this second example requires some further analysis for that judgement to be made. Let us imagine that a group J decided that they wanted to stop i from reading The Lost Symbol the first weekend it is on sale. Let us imagine that



Republican freedom, rights, the coalition problem221

they commit no crime in their conspiracy. Say some of them visit him to slow him down on the way to the shops; others advise that shop A has sold out; another takes him to shop B knowing it has sold out; another promises to read his copy quickly and lend it to i, but then reneges on the deal. Has the group made i unfree to read The Lost Symbol or merely non-free? Certainly, if they did this in i’s interests then, according to Pettit, they have not made him unfree. Let us say i has an important exam on the Tuesday. The group wants him to revise, and knows that he will not be able to tear himself away from a new Dan Brown. If he thinks he is going to get to read it on the Monday, he might revise over the weekend; then the broken promise of a loan means he can revise on the Monday too. But what if group J’s intentions have nothing to do with i’s interests? What if it is just a joke, or the basis of a bet (person k will read the latest Dan Brown before i does), or simply malicious? Here the republican is surely tempted to suggest i has been made unfree. If they do not resist that temptation, they will be committed to the view that domination occurs whenever a group has the power to stop some person i from doing x and is capriciously minded to do so. However, this makes the will to stop a person an important aspect of domination. A person or group with the power to thwart i and who does so because of the right sort of intention becomes a dominator. It might seem that some person or group j that could easily intervene because of some power they have (legal or physical) making them ‘obviously’ a dominator and others who could less easily intervene only become dominators when they actually intervene. We might try to mark the distinction more formally than by some vague gesturing towards expectations that some person or group might intervene by judging this in terms of i’s reaction to the (potential) dominator. If i treats some j as though she were a dominator (even when she does not intervene), she can be considered one. Some remarks by Skinner and Pettit suggest that this might be one route to take. They discuss at some length the manner in which people might be free to do x in the pure negative sense by being cunning or obsequious enough to get x despite the presence of dominating j.10 But this route means that some people might be dominant without doing anything or realizing their effect. Tony might be treated as though he were a Mafioso simply because he wears sunglasses and a sharp suit. Or, in Kramer’s example, the villagers might never do anything that they think the gentle giant might disapprove of, even if he never had the intention of stopping them. The ‘rule of anticipated reactions’ (Schattschneider 1960) is well known in power studies: individuals react to powerful individuals by doing their wishes without the powerful having to do anything. However, can we really denote a class of people as dominators simply because that is how

222 Freedom

they are treated? Certainly we have to use some criterion of ‘reasonableness’ in the anticipation or we allow everyone to be dominators when the paranoid pedestrian is around.11 There might well be systemic features of society that lead to dominant relationships, as we saw above in the relationship between men and women in most, if not all, societies. But, as we saw, such reasonable expectations are often based upon prior actions due to disposition and habit, and hence the will of the dominator is not left out of the domination relationship. And the nature of the reaction will also add a moralized element to the definition of republican freedom. If someone’s response to me is deferential, then we have one domination relationship; if it is dictatorial we have another. The measurement question is also implicated here. For the pure negative libertarians, if there is no one actually stopping a woman from going out unveiled, then she is negatively free to do so. However, she may feel that she is unable to do so because she would be shunned by her society. Those who hold pure negative accounts of liberty measure the loss of freedom by the conjunctive freedoms she loses if she were to walk about unveiled. It is not clear in this example whether Carter’s equivalence thesis holds or the relative freedom as measured by republicans and pure negative libertarians differs.12 Thus, in a comparative context, the pure negative libertarian will see less freedom in a society characterized by such general domination than one without. Similarly, if there are aspects of domination by a single person affecting others’ behaviour that too will cause freedom to be reduced in the societies where these occur. Thus, in the comparative context, whilst the precise characterization of specific situations may differ, republicans and pure negative libertarians will agree on measures of freedom across societies with or without these features of domination. Without a specific measure of republican freedom, which has not been provided by the republicans, we cannot tell whether there is any difference between republican measures of freedom and those of the pure negative libertarians. Some measurement issues I have skirted around the measurement issue so far, largely because it is dealt with extensively elsewhere (see Chapters 9 and 10; Dowding and van Hees 2007, 2009; van Hees 1998, 1999a, 2000a, 2000b, 2004). However, I note that the pure negative libertarians, and others, have explicit measures of freedom because they wish to produce a formula for measuring freedom across societies or groups in societies. It is this aim that motivates some of the more arcane aspects of the discussion. The republicans do not produce any such formula and indeed might see it as irrelevant. At times, for republicans, the measurement issue seems purely



Republican freedom, rights, the coalition problem223

nominal. Either a person is free or not. The ‘free citizen’ is free; the slave is not free – no matter what either of them is able to accomplish in their lives. At other times, dominance is seen as coming with scope modifiers, so a person is dominated in one sphere of life (the employment situation, say), but not in another (their home life); and within a given sphere, dominance is discussed in terms of more or less. That is, one employer might be more dominant than another, suggesting that freedom from dominance can come in degrees. From a physical standpoint, examining which actions should be counted in a republican calculus is problematic; from a legal standpoint less so. After all, we can compare the constitutional rights of citizens and the legal rights of employees, wives, pedestrians, and so on across societies, in order to see how they are protected. Pettit makes clear that it is not only the laws as codified that matter, but also how they are enforced. This again can be measured, though here we get closer to the pure negative libertarians’ measures. For example, in two societies with identical employment laws, we could measure custom by looking at dismissal rates of employees relative to unfair dismissal proceedings or we could estimate the number of battered wives and the conviction rate for domestic violence, and so on. Crude measures perhaps, but at least we can see how we might empirically calculate ordinal or interval measures for comparative analysis. That is, we can at least give quasi-orderings of freedom. The measurement issues are relevant to Pettit’s resilience claim. He suggests that republican freedom is more resilient or robust than negative freedom in the sense it is less precarious. It is less likely to wither under attack. This is surely an empirical claim, but we can examine its theoretical justification. Resilience Pettit suggests that republican freedom is more resilient than pure negative freedom. What precisely does he mean? He might mean that where two people have the same abilities to do some x – such as fly a kite – where that ability is correctly described as a republican freedom, the probability that they will retain that freedom if the world changes is greater than if that ability could only be described as a negative freedom. Pettit suggests that the difference can be seen in modal terms: someone might enjoy noninterference in the actual world: because of a quite precarious contingency: say, because it happens that certain powerful individuals have a liking for you or that you are able to keep out of the way of such individuals or ingratiate yourself with them. In this sense you enjoy non-interference in the actual world … not enjoy it robustly

224 Freedom or resiliently … you do not enjoy it in the range of readily accessible worlds – a range of nearby possible worlds – where this or that contingent condition is varied; you do not enjoy it resiliently. (Pettit 1997: 24)

However, without such dominating powerful people, the freedom from non-interference is more resilient – it is enjoyed in a set of nearby possible worlds which share this feature of not having dominant powerful people. Resilience is related to the arbitrariness condition of domination in the following way. If we assume that the disposition to interfere is normally distributed across all societies, then the society where the least actual interference and least domination will occur is one where people have the least capacity to so interfere. Thus it might be thought that equalizing power will make the possibility of dominance lower. A second way that one society might have more resilient freedom than another is where constitutional and legal provisions, duly backed by implementation and regulatory processes, can protect the freedoms of others, even where other forms of power inequality hold. I consider each in turn. Equalizing power might reduce the possibility of dominance, for the simple reason that arranging collective action against a person i is more costly than acting individually. Whilst it is always possible for a group J to dominate an individual i, the members of J have to agree on the target and coordinate their activities. We might think this would happen less often where resources are equalized. Of course, society might require there to be some level of unequal powers. Hierarchy is often required in complex organizations in order to pursue collective goals, and powers are sequenced accordingly. This does enable the possibility of dominance, but if democratic procedures are in place throughout society, such that organizations have agreed aims and the powers of those higher up are constrained by their organizational duties, then the possibility of dominance might be minimized. In Pettit’s terms, they are constrained in making arbitrary decisions. This constraint leads us into the second possibility for reducing the possibility of dominance: legal provisions. I suggest that republicanism is better specified by concentrating upon the legal aspect, for two reasons. First (following Kramer and Carter), equalizing power will increase pure negative freedom as well as republican liberty, since it will increase the set of conjunctively possible actions (the equivalent judgement thesis). Second, it is difficult to specify a coherent account of domination that does not involve dispositions and habitual behaviour in the characterization of dominant actors. Even if individuals are given equal power, one group might be dominant and another subordinate, simply because of the dispositions and habitual behaviour of each. Equalizing resources outside of dispositions is not sufficient, as



Republican freedom, rights, the coalition problem225

demonstrated by the examples of Lennie and George and the subordination of women even after legal equalization.13 Thus we will conclude that if there is something different to be found in republican accounts that makes republican liberty more resilient, it might be in what republican accounts specify in legal provisions. Rights and republican freedom The resilience of any person i’s freedom is not going to be reliant purely upon any philosophical concept, though there might be some correlation between the resilience of some types of freedoms compared to others overall. To return to the kite-flying example, we might think that Fred’s freedom to fly his kite is more resilient than Sarah’s. First, he has a Hohfeldian liberty to fly the kite; he is under no duty not to. Sarah, however, has a duty not to fly the kite when her mistress forbids her. Second, Fred has a claim right with regard to his kite in the sense that others have a duty to respect his property in the kite; Sarah lacks this, since, as a slave, she can have no property. Let us consider Sarah’s condition more carefully. She is under a duty not to undertake activities unless her mistress approves of them. Therefore her situation could be described in two ways: first, she is at liberty to fly her kite, as kite flying is an activity her mistress approves of, and therefore Sarah is free from any duty not to partake in it. Or, second, Sarah’s mistress disapproves of kite flying, but allows Sarah to fly her kite anyway; Sarah’s duty to not fly the kite has therefore been temporarily waived by her mistress, the claim holder. Under the first description, Sarah is not party to a right/duty relationship; she is at liberty, but her liberty to fly the kite is not protected by other claim-rights (unlike Fred, whose liberty is protected by his claim-right to his kite). Under the second description, Sarah is party to a right/duty relationship which has temporarily been waived, and again she does not have the protection of other claim-rights. Sarah may be less free under the second description, as her mistress may choose to enforce (stop waiving) her right at any time. Though, of course, even under the first description where there exists no right/duty relationship, Sarah’s mistress could still choose to change her mind and deem kite flying inappropriate, in which case Sarah would then have a duty to desist from it and a right/duty would come into being. Sarah’s mistress, then, has the power to set the rules of the game and decide when a right/duty exists and when it does not. This might appear a different and even greater power than deciding to enforce/waive a right. Furthermore, we can note that Sarah, when it comes to freedom and when she is under a duty as in the second description, does not have the

226 Freedom

same sort of freedom as Fred to oppose her mistress. If Fred knows his employer disapproves of kite flying, but decides to continue nonetheless, he has not committed a moral or legal wrong. He was unconstrained in his actions to disobey his employer. Indeed, if he was fired by his employer for flying his kite, Fred may have the protection of unfair dismissal laws. If Sarah chooses to fly her kite anyway, she is not only disobeying her mistress, but breaching a duty with all the legal and moral weight that that entails. In the instance of such a breach, Sarah’s mistress may seek to enforce her claim, which may involve punishment for Sarah. It is possible that Fred might be as constrained from kite flying as Sarah, if his employer disapproves and Fred fears this will affect his promotion prospects. In such a case, we can, of course, see the employer as a dominator, and employment laws might provide some protection. My point is that, whilst we expect free citizens to have freedoms more resilient that slaves in the sense Pettit means, it does not mean in practice that each slave’s freedoms (in the sense of what they are able to do) are more likely to be lost. If, for some reason, masters tended to be liberal about their slaves flying kites, but employers tended to disapprove, then being a free citizen rather than a slave would not make one’s freedom to fly a kite more resilient.14 The fact that Fred owns his kite gives others duties not to stop him kite flying that Sarah who cannot own a kite does not enjoy. These claim-rights of Fred might also give greater resilience to his kite-flying activities. In that sense, claim-rights seem to give greater resilience to freedoms; free citizens have such rights, whereas slaves do not (or at least free citizens have a much larger set of such rights – there might be some laws governing masters’ duties to slaves). Claim-rights, however, do not track republican freedom. And we must be clear that any such claim of resilience is only a tendency that relies upon a ceteris paribus clause about dispositions – as demonstrated in the example of the freedom of women in two societies which differ only in the disposition of men. There might well be societies whose laws (and their interpretation) are identical, but differ simply due to the dispositions of members of those societies. What we might say is that certain types of freedoms, notably claim-rights, might be more resilient than others given the nature of their protection, because the nature of the protection varies with the type of freedom. Such distinctions will rely upon legal provisions and interpretations of rights (broadly understood) and freedoms. I take this to be the intuition of republicans. However, that relation, as seen from the complex specification I have given (and not fully worked through here), is not straightforward, and will always be subject to the dispositional objection. There is no logical relation between legal provision and material freedom, nor any empirical one-to-one relation (and hence no logical or empirical one-to-one relation)



Republican freedom, rights, the coalition problem227

between man-made laws and resilience. There might well be an empirical correlation between resilience and freedoms instantiated in rights or Hohfeldian claims, rather than in mere Hohfeldian liberties or privileges. But these thoughts about the resilience of some freedoms in relation to others do not track well against republican versus pure negative accounts of liberty. Conclusion First, the distinction between the resilience of a right X for a class of people and the actual ability of a given member of that class to do a specific instance of x is an important one. Too often in political philosophy considerations of individual rights begin with the consideration of i’s ability to do x given the potential constraints of some person j. Perhaps it would be better if more attention was directed towards the rights and freedoms of classes of people to do x-type actions. The lesser freedom of women compared to men to do some action x might be rather more important than the actual ability of some woman or some man to carry out that action. Considering freedoms at this collective level does not imply we are not discussing individual freedoms. Second, my conclusion that, roughly speaking, resilience might track better against rights than (republican) freedom might horrify Quentin Skinner, who promotes republican values against liberal ones at least partly in response to liberal concentration upon ‘the rights of’ rather than ‘the duties of’ citizens. However, I do not feel that his arguments in that regard depend upon the republican account of freedom. Indeed, I am as appalled as he by the abuse of rights in liberal thought. Some of that horror is provoked by the undefended move of those who try to justify Hohfeldian claim-rights when, at best, they have arguments only to justify Hohfeldian liberties. To the extent that liberalism is about increasing freedom, it might seem that shackling people with duties is inimical to liberalism. Yet maximizing the amount of freedom in a society does not imply that some should not be restrained for the benefit of others. Duties towards others might well increase freedom overall, and it is such duties that might lead to greater resilience. Thinking seriously about measures of freedom, as the pure negative libertarians have done, even if their specific measures are problematic (van Hees 2000a; Dowding and van Hees 2009), is a step in the direction of demonstrating that fact. Furthermore, the pluralism of values inherent in many accounts of liberalism might allow for the trading-off of liberty for welfare, or fairness, if that is indeed required. Making duties the basis of morality is no sillier than making rights fundamental.

228 Freedom

Notes  1 Skinner would sooner call this neo-Roman freedom but has reluctantly accepted the ‘Republican’ moniker.  2 Kramer’s account is also trivalent, but in his case the trivalence depends upon a distinction between self-imposed or naturally imposed inabilities, on the one hand, and other-imposed inabilities, on the other hand. Kramer’s trivalence therefore purports to be causal and physical. Thus moralization does not follow from trivalence itself, but in Pettit’s case from the nature of the trivalence.  3 Whilst Skinner undoubtedly should take this position, he appears to adopt a contrary position in Liberty before Liberalism (1998: 84).  4 We cannot be assured that the linearity in relative calculation of coercive situations such as these relative to alternative non-coercive situations would remain between any erstwhile republican measure of overall liberty and that of the rival pure negative libertarian measures. However, since the republicans have not elucidated any formula for measuring overall freedom and as this is not the major point of contention, I leave that issue aside. I merely note here that the characterization of wherein the lack of freedom comes with coercion differs, but that both sides recognize that coercion does indeed reduce liberty.  5 The reality of the highwayman, of course, is that you have ‘no choice’, in the sense that you have a dominant strategy of handing over your money, since the alternatives are for the highwayman to take your money and not kill you or to kill you and take your money.  6 In the real highwayman case (see note 5), all the alternatives where you keep your money disappear after his intervention. Hence you have only two alternatives: ‘hand over your money’ or ‘not hand over your money’ – which are equivalent to ‘not die’, ‘die’: that is, they are conjunctives.  7 He says, ‘Hand over your blue car, you can take my old Black Bess’.  8 Whether diversity always reduces to welfare issues is a moot point.  9 I note here that it seems natural to place the term ‘interference’ in scare quotes. We would not normally consider this to be a form of interference with someone’s freedom. 10 Of course, the pure negative freedom they have is the freedom to do x through cunning or fawning. 11 One way of analysing such reasonableness would be in terms of the probabilities of those actions or examining the counterfactuals to see that the paranoid pedestrian is not unfree. 12 One pure negative libertarian disagrees. According to Hillel Steiner (1974, 1994: 52–4, 2003), the law of conservation of freedom entails that no society has more freedom than another. However, he is alone in holding that conception. It is not implied by either Carter’s or Kramer’s measure. 13 A man might dominate in the home through greater physical presence, but the habitual dispositional behaviour of men and women – for the latter to



Republican freedom, rights, the coalition problem229

avoid conflict and the former to thrive on it – might lead to continued domination even if all other resources were equalized. 14 In what sense could this be true? Only if there was some necessary relation between liberality with regard to kite flying and being a master, which seems too bizarre to contemplate. A more plausible example might be that in a capitalist democracy local politicians might sell the open spaces (used for kite flying) to developers to ensure a buoyant local economy and enhance their re-election prospects; a local monopolistic dictator might appreciate the open fields and retain them.

12 The construction of rights Keith Dowding and Martin van Hees

In what sense, if any, do rights exist? If rights are instantiated in law but difficult to exercise in practice, do people really have those rights? Can we compare across countries to see what rights people have both materially and formally? We map out below the senses in which rights can be said to have existence. We suggest a framework for analysing how rights might be measured and compared. The framework is supposed to be relatively neutral between competing conceptions of rights, though we do argue that rights cannot form the basis of morality or a system of justice. They are not foundational, since no system of rights worth the name can be ‘co-possible’ (Nozick 1974) or more precisely ‘compossible’ (Steiner 1994). A system of rights is compossible to the extent that all persons can exercise their right and there will be no conflicting duties. In other words, we suggest that there are no principles of justice that can deliver a set of rights that do not contain contradictory judgements about the permissibility of actions. There will always be occasions when people cannot exercise their rights simultaneously. Contrary to what some have argued, it does not follow that at least one person does not have a right to do x simply because two people cannot exercise a right to do x simultaneously. Having a right does not entail being able to exercise it. Nevertheless, if rights cannot be exercised, they hardly seem worth the name. Did Zimbabweans really have the right to vote in the presidential elections of 9–10 March 2002? That election demonstrated how easy it is to disenfranchise voters, simply by taking excessive time to check names against the list of registered voters; whilst voters are just as effectively disenfranchised by intimidation, whether organized by the state, political parties or wandering thugs. US citizens have the right to vote, but first they must register and then there must be polling stations which recognize that registration and remain open long enough for registered voters to cast their ballot. Some voters claimed they were disenfranchised at the US presidential



The construction of rights231

election in 2000, as polling officers would not accept their registration details. The ease of registration differs across US states, still more across the world, whilst the costs of voting also differ. At what level of difficulty and cost can we say that citizens effectively lose the right to vote? At what level of health care or educational opportunity do people have rights to education or health care? Can we expect poor nations to provide the same level of rights as rich nations? How might we weight and compare rights across countries? These are not easy questions to answer, but we try to provide a framework through which such questions may be addressed. We are concerned here with judging in what sense people have rights. We provide a framework by which to judge to what extent rights exist and to what extent they can be exercised. Whilst no framework could be normatively neutral, our framework can encompass different substantive rights theories. We begin by examining the non-compossibility of rights, demonstrating formally that rights are not compossible, or, to the extent that they are, they are not the rights about which we ordinarily write and speak. According to Steiner (1994: 2–3), the compossibility of a set of rights is a necessary condition for the plausibility of any theory of justice that yields that set. Nevertheless, rights and liberty are fundamental to liberalism, and clashes of rights are (part of) what politics is all about (see Primus 1999 for a historical account of such clashes and changing rights talk in US history; see also Wellman 1998). The fact that rights clash, and that at times my right to free speech may need to be curtailed by your right to privacy, or my right of free speech is curtailed by your right of free speech, does not mean we do not have such rights. But in what sense do we have them? Rawls (1971, 1982) argues that as long as our rights in these regards are approximately equal, then we have a just liberal system. He says that as long as the ‘central range of application’ of basic liberties is provided for, the principles of justice are fulfilled. For Rawls, basic liberties are specified by institutional rules and duties that form a framework of legally protected paths and opportunities. As he recognizes, of course, differential material or educational means entail that not everyone may exercise their rights equally. Rawls introduces the idea of the ‘worth’ of liberty to suggest how useful to people these rights are given the probability they can exercise them. If someone has no chance of exercising a right, then the right is barely worth the name. We do not try to adjudicate this particular issue here. Rather, we provide a framework for defining rights of different sorts – what we call the existence conditions of rights – and then suggest a way of handling whether or not we should judge those rights can be said to exist in the forms we define for a given individual, a group or society as whole. Such judgements are

232 Freedom

made relative to the expectations we may have for exercising rights, given the social and economic circumstances of different societies. We specify the sense in which someone can be said to have a right (the senses in which a right can exist) and then the probability that they can exercise it. In doing so, we specify, in a fashion broadly compatible with Rawls’s (1982) discussion, how to identify worthwhile rights and how to do so relative to different circumstances. Although rights can, and should, play an important role in the theory of justice, they cannot form the basis of a theory of justice. Rights are essentially constructed. The extent to which rights exist is not all-or-nothing when it comes to respecting or exercising those rights. People have rights even when they cannot exercise them. First, they may have them formally, when they do not have them materially. Second, people may have rights materially, even though they cannot always exercise them. Having a right entails some expectation that it might be realized, but does not ensure it. We believe that this way of representing rights both makes sense formally and is more in tune with everyday moral language than other formal accounts of rights discussed here. First, however, we argue that the allure of rights compossibility should be avoided. The non-compossibility of Sen rights Rights may seem obviously non-compossible. The most striking examples are entitlements, though rights in any form are not manifestly compossible. For example, if I claim a right to free speech and you claim a right not to be subject to abuse, something must give way if I want to verbally abuse you. Of course, we may need to sort out whether my right to free speech includes the right to say the precise words I intend, and whether your right not to be abused is covered by those words. Typically rights conflict in these sorts of ways. We can trivially make rights compossible by creating a dictator who can assign rights as she sees fit to ensure that no rights ever conflict. (Steiner (1994) calls such a dictator the Adjudicator.) She could be careful and assign type rights so that no conflict occurs – the easiest way would be assign all rights to herself and none to others. Or she could decide the assignation of a right after a conflict between two sets of putative rights has been claimed.1 Juridical decisions decide which rights are to be restricted or given greatest weight in any conflict, and they might be said to be part of a construction towards a formally compossible rights set. However, any formally specified rights set worth the name is unable to deliver a compossible set of rights without such an ‘external’ adjudication. External, that is, to the rights set as understood to be in conflict (Steiner 1998: 265ff.).



The construction of rights233

Furthermore, in practice any external adjudication is likely to involve considerations that are not strictly rights-based (as the discussion of cases in Wellman 1995: ch. 7 shows). In order to overcome the triviality of such a set of compossible rights, formal theorists normally assign a set of reasonable conditions and then see whether conflicts emerge. For example, we can see a simple proof of the non-compossibility of rights in terms of Amartya Sen’s account of minimal liberalism.2 One way of representing Sen is to view society as a social decision function: that is, as a function F that defines possible combinations of individual preference orderings R1,….,Rn to a ‘social preference’ R = F(R1,….,Rn).3 Whatever the feasibility constraints on the alternatives open to society, F provides the answer to the question of what is best for society by generating the social preference relation R. For any situation, the best social state relative to R is chosen from the set of feasible social states. Obviously, it is thereby assumed that the social preference relation always contains a best element.4 Sen defines a condition of minimal liberalism (ML), suggesting it is a necessary but not a sufficient condition for rights holding. ML states that there are at least two individuals, each of whom has at least one pair of alternatives over which he is decisive. An individual is decisive over a pair of social states (x, y) if it is always the case that if the person strictly prefers x to y, then society should strictly prefer x to y, and, conversely, if that person strictly prefers y to x, then society should strictly prefer y to x. The condition of minimal liberalism, ML, now states that there are at least two individuals who are decisive for some pair of distinct alternatives: that is, there are individuals i and j and distinct pairs of alternatives (x, y) and (z, v), such that [yPix → yPx and xPiy → xPy] and [zPjv → zPv and vPjz → vPz]. Sen’s liberal paradox states that condition ML cannot be satisfied simultaneously with the (weak) Pareto-condition (PAR) and the condition of Universal Domain (UD). According to the Pareto-condition, a strict preference that is shared by all individuals should also be represented in the social preference relation: if all individuals strictly prefer some alternative x to some other alternative y, then society should also strictly prefer x to y. The condition of Universal Domain demands that a social preference relation is generated for any logically possible configuration of individual orderings: no individual preference orderings are excluded a priori. The incompatibility of the three conditions is easy to demonstrate. Take the simple case in which there are only three alternatives, x, y, z, and two individuals. Assume the two individuals, i and j, have rights over x and y, and over y and z, respectively. Given UD, suppose that individual i strictly prefers y to x and x to z, whilst j strictly prefers x to z and z to y. We then see that i’s rights over x and y implies that yPx and that j’s rights over y and

234 Freedom

z implies zPy. The Pareto principle yields xPz, which means that we have xPzPyPx: the social preference relation is cyclic and hence does not contain a best alternative. Sen’s results show that, under the conditions UD and PAR, a person’s decisiveness over one pair of alternatives is incompatible with other persons having any rights at all. To see rights in terms of decisiveness over pairs of alternatives means that, in a very strong sense, individual rights are incompatible. However, the incompatibility of Sen rights only follows if the social decision function satisfies UD and PAR. It can easily be shown that if one of these conditions is dropped, compossibility remains possible. Indeed, much of the early literature directly following Sen’s theorem consisted of the derivation of such possibility results arising from relaxing the conditions (see Wriglesworth 1985 for an overview). To what extent can such relaxations be seen as showing the compossibility of rights? A relaxation of UD – that is, a restriction of the domain of the social decision function – can be interpreted in two ways. First, it could be used normatively. It has, for instance, been shown that the liberal paradox arises in cases in which individuals are ‘meddlesome’, and it has been argued that rights should not protect ‘meddlesomeness’ (Blau 1975). We could ban ‘deviant preference profiles’ from the social decision mechanism (Goodin 1986). Alternatively, rather than abandoning UD, a weaker notion of decisiveness could be used – a person’s strict preference regarding a pair of alternatives over which she has a right should only be reflected in the social preference relation if the person is non-meddlesome. It is not the place here to assess the normative appeal of abandoning UD or of using domain information to weaken the definition of rights. However, such approaches are rather ad hoc. Defining meddlesome preferences is not straightforward. To call them ‘meddlesome’ already labels them as ‘wrong’, but we might call some of them ‘caring preferences’ or, more neutrally, ‘other-regarding’. Most of us want our loved ones to care about us, even if this means they have preferences regarding our well-being that differ from our own. Abandoning UD may be one route out of Sen’s paradox, but it is a route with almost as many problems as the theorem itself. Second, relaxing UD can be justified ontologically. On this view, the domain can be restricted because the situations in which conflicts between rights occur do not ‘really’ belong to the realm of the possible. The restricted domain is thought to describe more correctly the set of possible worlds: some configurations of individual preference orderings may, for instance, be psychologically impossible and may therefore never be realized; conflicts in those psychologically impossible worlds need not bother us. Such an approach ensures rights compossibility by adopting an alternative interpretation of the notion of possible worlds; it does not deny that under a



The construction of rights235

logical interpretation of that notion, Sen’s liberal paradox establishes the non-compossibility of rights. More importantly, however, the claim that the preference profiles that lead to problems may be logically possible but impossible on some other criterion simply cannot be sustained. There is, for instance, nothing odd about Sen’s original illustration of the paradox – the Prude and the Lewd have to decide whether or not to read Lady Chatterley’s Lover. Whatever one’s metaphysical opinions regarding the notion of ‘possibility’, that situation certainly seems possible. What about relaxing the Pareto condition? Can that secure the compossibility of Sen rights? Such a relaxation is Sen’s (1982e) suggestion. On his view, ML is more important than PAR and the latter therefore has to yield. However the Pareto condition itself describes certain rights, to wit the rights of society. After all, the Pareto condition states that the group of all individuals is decisive for any pair of alternatives. Hence, we cannot really say, as we did above, that Sen’s result only established the noncompossibility of rights given a social decision mechanism satisfying UD and PAR. Instead, we should say that, given UD, the result establishes the non-compossibility of individual and certain group rights (see in this respect Stevens and Foster’s (1978) extension to conflicts between group rights; see also Gekker 1985). It could be argued that a theory of justice need not attach much importance to the group rights exemplified by PAR, and so the result merely shows that some allocations of rights are non-compossible. And, indeed, Sen’s arguments for a relaxation of the Pareto condition could be interpreted as a powerful defence of such a position. However, even without group rights, there is a serious compossibility problem with Sen rights. Alan Gibbard (1974) posed the problem, in what is known as Gibbard’s liberal paradox. He showed that with a slight formal strengthening of ML, compatible with the spirit of Sen’s defence of ML, impossibility results arise even when the Pareto condition is dropped altogether. The condition ML does not specify over which pairs of alternatives an individual is decisive – it states only that at least two individuals are decisive over some pairs. In his various examples, however, Sen assumes that individuals are decisive over those pairs of alternatives that differ only with respect to some aspect pertaining to that individual’s life. Gibbard’s strengthening of ML essentially comes down to this same assumption. Assume that each alternative can be described as a vector of features or characteristics and let, for instance, some feature – say component a of the vector – stand for person i reading or not reading a particular book (Lady’s Chatterley’s Lover, say), whereas another feature – component b – describes whether j reads a copy of that same book. Gibbard’s condition would be satisfied if i is decisive for any pair of alternatives that differ only with respect

236 Freedom

to i reading the book or not (that is, which differ only with respect to their ath component) and if j is decisive for all pairs of alternatives that differ only with respect to j reading the book or not (that is, which differ only with respect to their bth component). Gibbard subsequently shows that there is no social decision function that satisfies such a strengthened version of ML and that also satisfies UD. In other words, given UD, any allocation of individual rights turns out to be non-compossible. Many authors have suggested that the Sen approach is basically flawed. As Gibbard (1982: 597–8) himself later formulated it: These liberal paradoxes carry with them, then, an air of sophistry: they must in some way be creating problems that do not really exist … To talk about the paradoxes, then, is to explore the role of one kind of mathematics in thought about social norms and organization. What is it about the mathematical apparatus of social choice theory that apparently so misapplies to questions of liberty?

In our opinion, there are at least two problems with the social choice theoretic account of rights. First, a Sen right is defined as a form of power: having a right means being able to determine a part of the social preference relation.5 However, we often think of people having rights without having power. Consider a right to buy tickets for some concert. Obviously, one can have such a right even though one cannot ensure that one will indeed get a ticket, say because there are more people who want to attend the concert than there are available tickets. Indeed, to say that people lack a certain power does not necessarily imply that people lack certain rights or liberties or that the rights of individuals are non-compossible. The latter conclusion only follows if the notion of non-compossible rights is considerably stretched: it now refers to the impossibility of mutually compatible powers. And this is precisely what happens in the Sen approach – if each person’s preference for attending the concert leads to a veto of the outcomes in which she does not have a ticket, a cycle of the social preference relation emerges. Second, even if all rights were powers, defining rights in terms of decisiveness makes it difficult to distinguish between having rights and exercising them. If a person strictly prefers x to y and is decisive over (x, y), then, on Sen’s account of rights, x is socially preferred to y. The fact that a person has a strict preference and is decisive entails that the social preference relation is partly fixed, that is, that having rights is equivalent to exercising them. But we normally make a distinction between having a right and exercising it. If I have a right to read a book, I have that right whether or not I ever read the book. One way of avoiding the conclusion that, in the social choice approach, one cannot distinguish between having and exercising a right is to say that



The construction of rights237

the preferences of an individual can be interpreted as choices made by that individual (see, for instance, Sen 1992: 148). The rights that individuals have are described by the sets for which they are decisive, and the preferences of the individuals describe the ways in which they decide to exercise them. Suppose I strictly prefer x to y – that is, suppose I choose x when faced with a choice between x and y. This choice is hypothetical, since I am confronted not only with x and y, but also with many other alternatives. In this approach, a person exercises her rights by making certain hypothetical choices. Two observations are in order. First, the non-compossibility of Sen rights does not disappear. The impossibility result still remains valid, but can now be understood as the impossibility of translating all relevant hypothetical choices into one coherent social preference relation: individuals cannot always exercise their rights simultaneously. Second, both the exercise and possible violation of rights are about the choices we actually make, not about hypothetical ones. Thus, the social choice definition of rights leads to counter-intuitive conclusions, since all rights are assumed to be powers, and the distinction between having rights and exercising them evaporates. Dissatisfaction with these counter-intuitive consequences has led to a different way of formalizing rights: viz., through the use of game-theoretic tools. In the next section we shall briefly sketch the outlines of the game-theoretic approach. Game forms and effectivity functions Recently, social choice writers have developed game-theoretic approaches to the study of rights and freedoms (Gärdenfors 1981; Gaertner et al. 1992; Hammond 1996; Pattanaik 1996; Fleurbaey and van Hees 2000). The basic idea underlying game-theoretic approaches is that rights do not depend on individuals’ preferences, as is the case with a definition in terms of decisiveness, but on the things individuals are and are not allowed to do: that is, on their admissible strategies. The approaches were designed to formally characterize rights in a manner preferable to the ML condition, but were not designed to overcome the ‘impossibility’ result (though see van Hees 1999, 2000a). Two types of game-theoretic approaches can be distinguished. The first, based on work originally by Gärdenfors (1981), suggests that rights may be represented by effectivity functions. In this approach rights are described in terms of sets of outcomes of which a person can secure elements. The right to wear a black shirt means a person is effective for a set of outcomes that are all characterized by that person wearing a black shirt. If the right is exercised, one of those outcomes is realized. The second approach represents rights directly in terms of a normal game form. A game form is a game

238 Freedom

where the utility function of each player remains unassigned. So it includes a list of the players; for each player a list of alternative strategies; and an outcome from every combination of strategies that may be chosen. Taking the notion of ‘admissibility’ as a primitive term – that is, a term that is not further explicated – some of the strategies of the individuals are labelled ‘admissible’ whilst others are not. The rights of individuals are subsequently specified by the freedom each has to choose any of their admissible strategies and/or by the obligation not to choose a non-admissible strategy (Gaertner et al. 1992: 173; Suzumura 1991: 229). The two approaches are not independent of each other. One way of viewing the effectivity approach is to say that it specifies how to distil the exact specification of rights from a given normal form game. Of course, we then need to make clear under what circumstances a person can be said to be effective for a set of outcomes: that is, what it means to say that a person can ‘secure’ a certain outcome. Two types of effectivity have been used: a-effectivity and b-effectivity. The first is defined as an individual’s power to force an outcome from any set: given a normal game form, a person is a-effective for a set of outcomes under any strategy that, regardless of whatever admissible strategies the others adopt, leads to an outcome belonging to that set. The notion of b-effectivity is much weaker since it does not guarantee that a person can secure some outcome no matter what anyone else does – rather, it is defined by the absence of a blocking power by the others: a person is b-effective for a set A if the group consisting of all others is not a-effective for a set disjoint from A. There is a potential problem, however, with inducing ‘rights as effectivity’ from a normal game form if it is assumed that the normal game form contains only feasible strategies (some of which are admissible, others non-admissible). In that case, effectivity functions may be too strong for an account of rights. Take, for example, my right to wear a blue shirt. Presumably, I have such a right even though I am unable to choose one – say because I messed up the colours of my clothes whilst doing my laundry. That is, even though everyone else has adopted admissible strategies I may not be effective (whether in the a- or b-sense) for a set of outcomes in which I wear a blue shirt. This sort of problem can be accommodated if it is assumed that the game form from which effectivity functions are derived can differ across individuals (van Hees 1995). We then assign to each individual a normal game form that describes that individual’s ‘deontic realm’. To see whether a person has a certain right, we examine whether they are effective for the relevant set of outcomes in the game form that has been specifically assigned to that person: that is, whether someone is effective for that set in their own deontic realm. Since individuals may have been assigned different game forms – that is, may have different deontic



The construction of rights239

realms – the problems described need not arise. It is, for instance, perfectly possible that I have the power to see to it that a certain outcome is realized in my deontic realm, whereas you have in your deontic realm the power to see to the realization of the contrary state of affairs.6 Although the notion of a deontic realm can be interpreted in different ways, it is clear that insofar as the deontic realms of individuals differ, they denote different ‘worlds’ (whether possible or not). But that means that the problems are not solved by showing that rights are compossible after all – that they can exist in one and the same possible world – but by implicitly abandoning the notion of non-compossibility of rights. Furthermore, there is a problem that underlies all the game-theoretic approaches to rights that have been developed so far: their neglect of the intentions of individuals. For instance, contrary to what the effectivity approach suggests, rights can exist even where I am not effective (however defined) in bringing about some outcome. What matters also is whether someone has deliberately stopped me from being effective in an illegal manner. Imagine I am on my way to the polls, but am held up in heavy traffic and they are closed by the time I arrive. My right to vote has not been infringed, even though my strategy of voting – driving to the polls – was not effective – certainly not in the a-sense, but perhaps also not in the b-sense. My right to vote has not even been infringed if the heavy traffic was caused by an accident due to someone doing something illegal – driving faster than the legal limit, say. My right to vote would only be infringed if someone did something illegal in order to stop me from voting, such as a refusal to recognize my registration card as legitimate. To accommodate these problems, a game-theoretic account of rights should also refer to the intentions that individuals have and thus far this has not been done. Steiner’s rights compossibility In both Sen’s and the effectivity approach, rights are conceived in terms of ways of realizing certain outcomes. Sen defends his account on the grounds that actions and their consequences are closely related. Steiner too, though he sees rights as pertaining to actions, understands rights compossibility as concerning the object-temporal or spatio-temporal coincidence of the extensional specification of actions. In that sense, actions and their outcomes are closely linked. Furthermore, both the social choice and the effectivity approach define rights in terms of powers. Steiner’s account of rights, though not formalized in the same way, also sees rights much in terms of powers, and so similar compossibility problems arise.7 In terms of the Hohfeldian classification, Steiner views rights as claims or immunities such that the right-holder possesses the powers to enforce or

240 Freedom

waive the constraints that are logically implied by it. For instance, if A has a claim that B performs x, then B has the duty to perform x unless A waives this duty. In our discussion of Steiner’s theory, we shall not focus on these claims or immunities as such, but on the liberties that are implied by them. Steiner makes clear that the exercise of rights that form part of a compossible set of rights should issue in the performance of actions that are compossible. We argue that this requirement cannot be satisfied in a meaningful way, which entails that the compossibility of rights as conceived by Steiner forms a problem.8 An act-type is a class of acts for which each act-token is an instantiation of that act-type. ‘Freedom of expression’ is an act-type, for which my saying ‘Down with the government’ on the pavement at Whitehall on Thursday 8 March 2001 at 2.15pm precisely is an act-token. But my statement at that time is also an act-token of the class of acts of saying ‘Down with the government’. We normally think of rights and liberties as act-types or, rather more carefully, we think of them as claims for act-types. The right to free speech is a liberty, immunity against the removal of the liberty, and rights against certain types of interference with the exercise of that liberty. Thus the right to be able to express myself freely on controversial political, social and scientific matters without hindrance from the state, with indeed protection from the state against others who may illegitimately try to stop me, is a complex set of interrelated rights, liberties and immunities.9 Whether or not I want to express myself, or ever do express myself, on social and political matters does not affect whether or not my right to do so is protected. (This means both the protection of my liberty-to-speak by my immunity and the protection of my ability-to-exercise-my-liberty by my rights.) Similarly, a right to education or decent health care is the right that the state will facilitate provision of education and health care. This is so whether or not I play truant from school or never feel the need to go to the doctor. In that sense we ordinarily think of rights as being about act-types. But if rights are about act-types, then rights cannot be compossible. Or, at least, to the extent that rights are about act-types, we cannot guarantee that the exercise of those rights on any occasion is compossible. We cannot all give our conflicting views about the government at the same place and time, since my occupying this place at this time entails that you cannot occupy the same place at the same time. To the extent that resources are limited, one person’s health care may need to be sacrificed for another’s, whilst the amount devoted to the special educational needs of some may be more limited than many would wish. For these reasons, Steiner develops an account of rights compossibility specifically generated in terms of act-tokens. Immediately, we can see that Steiner’s account of rights is not the language of ordinary rights talk. The



The construction of rights241

justification of such a move has to be that it is the only way to make rights talk coherent and the basis of a system of justice. However, if the cost of that coherence is to divorce it entirely from our moral world, the cost is too high. Our argument against Steiner is that his account of rights is so divorced that the cost is indeed too high. Steiner identifies an act-token by its extensional description, indicating its physical components. There can only be one act-token (or a particular acttype) answering to the same extensional description that has the same set of physical components. So he defines action compossibility: ‘Two actions, A and B, are incompossible if there is partial (either object-temporal or spatio-temporal) coincidence between the extensional description of A and either (i) B’s extensional description, or (ii) C’s extensional description if C is a prerequisite of B’ (Steiner 1994: 37, italics removed). From this, Steiner argues that a person, i, is unfree to do an action if i’s control of at least one of its physical components is actually or subjunctively denied by another person j. This implies, according to Steiner, that j must actually or subjunctively possess at least one of the physical components denied to i. The subjunctive element is important to Steiner but its nature may not be immediately clear. The subjunctive exclusion refers to an actual exclusion inoperative only because the individual did not attempt the action. Thus if i cannot leave a room because the door is locked, i is unfree to leave the room even if i does not attempt to do so (and has no desire to do so). It presumably includes the case where the door is not locked, but had i attempted to leave the room, it would have been locked – the bolt would be drawn quickly as i’s hand touches the handle. Does it include the case where the door could have been locked had i chosen to leave the room, but would not have been, since, j, the person who could have locked the door, would not have done so? We may assume j had no desire to lock i in the room. Steiner is not absolutely clear on this. One would think from his discussion (1994: 37–8) that in such a case, i would be free. We would think so, because i has the subjunctive freedom to leave the room since, had the attempt been made, i would have succeeded. Person i would have left the room, even though j could have prevented it had j so chosen. Control over actions, or the actual or subjunctive possession of the physical components of i’s freedom, is thus existent only given that others with preventative powers happen to not to use them. The freedom i has in the actual world does not extend to other possible worlds where others’ intentions and actions are different. A person’s control therefore is a control only of sorts: it is a control given that the person faces no (effective) resistance; it is not the control the person would have if there were resistance from others that is not there. This might be termed ‘outcome power’ (Dowding 1991: 48).

242 Freedom

However, things are not so simple. The problem arises not when we consider i’s liberty, but when we consider i’s liberty together with j’s. Steiner insists that any subjunctive freedoms a person has, only that person can have. No two people can both have the same subjunctive freedoms. Thus i may be subjunctively free to post a letter in the letterbox down the road in two minutes’ time, but only if there is no one else who is subjunctively free to post a letter at the same box in exactly two minutes’ time. Steiner (1994: 41) says, ‘Like actual possession, subjunctive possession cannot be ascribed to more than one person for any one time’. The fact that I am able to go and post a letter two minutes from now might suggest that I have the freedom to perform that particular act-token. After all, just as in the example of leaving the room, the freedom would be absent only if there were resistance from others. However, on Steiner’s view, a freedom is always a possession, and in this case – even in the absence of any potential resistance – it cannot be the case that I (subjunctively or actually) possess the same physical components of that act: there may be a host of others who could arrive at the letterbox just ahead of me. The fact that they do not, since none of them wants to post a letter at that precise moment, is thereby irrelevant. Steiner, in a discussion of an objection raised by Taylor (1982: 142–4) to an earlier account of his theory of freedom (Steiner 1974), acknowledges that it need not be the case that everything is always in the possession of one person or another (Steiner 1994: 40). However, what his account of subjunctive freedoms implies is that almost nothing can be in the possession of one person or another: it would seem that no one has any subjunctive freedoms beyond what others could physically stop them from doing. This means that a person’s subjunctive freedoms do not extend very far since, beyond the next few seconds, a lot of people could do actions that would stop one doing much (some could start a nuclear war, killing us within in the next hour or two). To put the point more bluntly, the account of subjunctive freedom must be wrong. Subjunctives rule every possible act we do and do not do but could do. It means we are never free, except in the very limited sense that what we actually do we are free to do, plus a few acts in the very near future that no one could possibly stop us doing in the actual world. This is simply not what we mean when we talk about human freedom. If i’s attempt to post a letter is thwarted by a queue of people at the letterbox, that does not mean i is not free to post a letter. It means i cannot always post a letter precisely when and where i wants. Or, stated differently, i does not have the freedom to perform this particular act-token. Steiner’s account of rights builds on this account of freedom and thus encounters similar problems. His choice account of rights is based upon each person’s freedom being constrained only by the legitimate (or admissible) actions of others. Where j illegitimately locks the door on i, then i’s



The construction of rights243

rights have not been recognized; where j legitimately locks the door on i, then i’s rights have not been breached. In neither case, however, is i free to leave the room. However, i’s rights only cover the cases where freedom to do some action x is not overwhelmed by some other person j’s legitimate actions. Thus i does not have the right to post a letter in two minutes’ time, given that others subjunctively may also be queuing to post a letter at that exact time. (To the extent that a set of people may legitimately initiate a nuclear war at any given moment, nobody has any rights beyond the next couple of hours.)10 Steiner (1994: 86–101) tries to accommodate these problems by drawing on Bentham’s familiar distinction between vested and naked liberties. A liberty (or permission) refers to the absence of a duty, and should not be confused with Steiner’s descriptive notion of a freedom as discussed thus far. A vested liberty is a liberty that is protected by the duties that others have, whereas a naked liberty lacks such protection. The problem discussed so far can be formulated in terms of vested liberties: we have argued that what we commonly perceive as freedoms can never be seen as ‘completely vested’. Since, by definition, a naked liberty lacks ‘protection’, it cannot constitute possession of part of the world, and hence cannot form a freedom in Steiner’s sense. Steiner (1994: 87) does not disagree with this: ‘the salience of nakedness, in the creation of incompossibilities is clear enough’. It is for this reason that he argues that rights should consist of vested liberties (Steiner 1994: 89): that is, of liberties surrounded by ‘an impenetrable perimeter’. However, the resulting rights catalogue becomes rather peculiar if all rights are completely vested: that is, if there is indeed an impenetrable perimeter. If we want to say that I have the right to post a letter, we would have to stipulate that my vested liberty only exists between, say, 3.00 and 3.05am – at other time-slots others have the right to post a letter. Indeed, most activities (to demonstrate, to practise a religion, to go to a concert, to surf the web, and so on) that would normally be thought of as the exercise of rights (a possible instantiation of an act-type right) should be made permissible to an individual or small group of individuals only during some particular slots in time or space. In fact, not even all individuals can be guaranteed such rights. Should we, for instance, say that at least some individuals lack the right to practise a particular religion if it is physically impossible that all individuals attend a Sunday service? The existence of rights Much of the discussion thus far has concerned itself with the sense in which a right exists, though this question has not been addressed directly. Before



The construction of rights245

Less trivially, however, we may wish to consider how we can represent different kinds of existence. We shall present four different ways in which rights can exist (cf. Rescher 1978). When we speak about the existence of individual human beings, we speak of the existence of particulars as contingent members of the world.11 This existence can, of course, be described in terms of a first-order predicate logic that contains the existence predicate E. We then write Ek, where k is an individual constant.12 The first way in which rights can be seen to exist is in terms of the existence of properties of individuals. To say that an individual right exists is, then, to say that certain first-order predicates apply to certain individuals. Existence is seen as a second-order predicate that is applied to first-order predicates in combination with the particulars to which those first-order predicates applies. At this point, if a first-order predicate R denotes a certain right, we discuss not the existence of such a right, but the existence of this right for a particular individual k. In a way, the right of k can be seen as a particular, where the right ‘as such’ is a universal. To indicate that we refer here to rights that are being proclaimed by some moral or legal theory without necessarily being respected, we call this type of existence formal. (i) The formal existence of a right as a particular: Σ(Rk) What can we say about the truth-conditions of such a formula in a world w? One might claim that Σ(Rk) only holds if (Ek & Rk). Thus an individual’s right to breathe fresh air exists if, and only if, the individual exists and if she indeed has this right. This implies that future generations cannot now have the right to breathe fresh air in the future. Whilst this may be acceptable to some (it does not imply, for example, that we do not have obligations to future generations, so that they can, in the future, have such rights), we may not want to exclude the existence of such rights on logical grounds only. We might therefore be inclined to infer that Σ(Rk) only presupposes Rk rather than (Ek & Rk). But then we are still making a particular ontological presupposition, to wit, the existence of the universal R. If we say some individual possesses a right R, we assume that R is a right that exists which individuals may or may not possess. Now let Σ be the second-order predicate describing the existence of first-order predicates. Hence, we now also distinguish (ii) The formal existence of a right as a universal: ΣR and we claim that Σ(Rk) should be defined partly in terms of ΣR. That is, we believe that to say an individual formally possesses a particular right

246 Freedom

presupposes that the right formally exists as a universal. For instance, a bill of rights stipulates which rights exist formally as universals, and this in turn determines which rights formally exist as particulars. With respect to formal existence, the universals thus have ontological priority. The question of the formal existence of a right as a particular shifts to the formal existence of the corresponding universal: that is, the formal existence of the freedom or right as such. It is here that a prior theory T is needed that recognizes this existence.13 Such recognition makes clear which rights there are formally, and also which individuals possess those rights. It does not, however, establish the more ‘definite’ or ‘concrete’ existence of rights when they are respected or exercised. To capture the ‘material’ existence of rights, we introduce a different existence predicate, e. (iii) The material existence of a right as a particular: e(Rk). What can we say about the truth-conditions of formulas of type (iii)? In this case, we again have recourse to the prior theory T. That is, the theory also needs to stipulate the conditions that should be fulfilled before we can say that the rights a person formally possesses are respected or exercised. Assume that the fulfilment of these conditions can be described in terms of a formula δ of the formal language at hand, and that the formal existence of a right is a necessary requirement for its material existence. We might then hold that e(Rk) ⇔ Σ(Rk) & δ. Thus an individual right exists materially if it exists formally and if it is respected or exercised. The distinction between formal and material existence can in turn be applied to rights sui generis. We can say, for instance, that although some society formally acknowledges some right, it is being violated – without our thereby referring to the violation of rights of specific individuals. (iv) The material existence of a right as a universal: eR. Whereas the formal existence of rights as universals can be established without reference to the rights as particulars that individuals possess formally, the ontological priority is reversed with respect to the material existence of rights as universals: a society can formally acknowledge a right sui generis, but to say that those rights also exist materially, we first have to examine which rights materially exist as particulars. To summarize, we have argued that to begin to analyse rights by their exercise is mistaken, but it is surely correct that rights ‘exist’ in the most ‘definite’ or ‘concrete’ sense when they are exercised or respected by others’ actions. In some important sense, rights essentially ‘exist’ in terms



The construction of rights247

of (i) and (ii): otherwise we would not be able to claim that rights are not being respected by the laws of some state or through actions of members of society. However, rights also have a more concrete existence when they are respected or exercised. This is captured by (iii) and (iv). Thus the analysis of rights should begin at the level of the formal existence of rights sui generis (level ii), proceed to the formal existence of rights as particulars (level i), and then proceed to the level of material existence of rights as particulars (level iii) and rights sui generis (level iv), respectively. Of course, the nature of the existence of rights under (i) and (ii) is a form of ‘bootstrapping’. Rights exist in the sense that they are thought to exist (whether through moral theory, beliefs underlying behaviour or by law), but thinking they exist leads to a more concrete existence through behaviour itself or state action. Rights can thus be seen in terms of expectations. Through our recognition of others’ rights we behave differently towards them. In other words, rights are part of our social institutions that may be defined in terms of equilibrium strategies. There is a large and growing literature that examines social institutions in terms of equilibrium arrangements (see, for example, Schotter 1981; Ostrom 1991; Calvert 1995). Specific applications of institutional theory to the analysis of rights can be found in Sened (1997) and van Hees (2000a). Our rights theories both track such equilibrium strategies and also criticize or try to modify such behaviour. Often groups may try to construct new rights by extending previous claims to new domains. If such rights can form new social institutions, then these groups may be successful. In other words, moral systems and ways of representing them grow both through argument and behaviour; the latter (as Rawls importantly noted) putting constraints upon the successes of the former. This section merely forms a sketch of the nature of the existence conditions of rights. As said, the crucial parts of those conditions – the conditions δ and the theory of which they form a part – have not yet been specified. Leaving aside the issue of the material existence of rights as universals, we shall analyse in the next section how the condition δ may be cashed in: that is, what it means to say that an individual right exists materially. Below we examine what that analysis entails for the other parts of the theory T. Measuring rights and freedoms These considerations lead us to examining the expectations that one may have about exercising one’s rights. We briefly outline the possibilities of an approach to rights that circumvents some of the problems discussed pertaining to non-compossibility. In two important aspects we deviate from the three approaches to rights discussed above. First, we focus on the

248 Freedom

freedom to perform act-types rather than act-tokens, thus coming closer to ordinary rights vocabulary. Second, we reverse the relation between judgements about which specific freedoms an individual has, and judgements about the extent (or degree) of a person’s (overall) freedom. Rather than first examining which freedoms an individual has and on the basis of that information examining the extent of a person’s freedom, we derive a judgement about what freedoms a person (materially) has from judgements about the extent of that person’s freedom. Although we shall limit the analysis to showing how a measurement of (overall) freedom can help to establish the existence of (particular) freedoms, we shall also claim that the approach can be used to establish the existence of other types of rights. We want to ascertain whether a person (materially) possesses a certain act-type right R and we assume that the individual possesses the right in the formal sense. Let r1,..,rk denote act-tokens instantiating the right. R may stand, for instance, for freedom of speech; r1 may stand for shouting ‘down with the government’ at Whitehall at a given time and date, r2 may stand for shouting the same thing at Piccadilly Circus, and so on. Now we want to derive a ‘freedom-function’ G that assigns a value between 0 and 1 to each R – the value G(R) describing the extent to which a person is free to do R. Suppose that for each act-token ri, we can define the probability p(ri) that the act-token will, in the relevant manner, not be prevented. Assuming that these probabilities determine to what extent a person is free to perform R, we have G(R) = G(p(r1),…,p(rk)). That is, the extent to which a person is free to do R depends on the probabilities with which each of the relevant act-tokens will not be prevented. It might be objected by those taking an all-or-nothing view to the existence of freedom that what we are specifying here is ‘the probability of a person’s being free to perform R’ rather than the extent to which they are free to do R.14 To the extent that our analysis is consistent with the ‘all-ornothing’ view of liberty or freedom, there is perhaps nothing wrong with interpreting freedom in this manner. We take it that our analysis teases out what most people mean when they talk about freedom. However, the fact that everything can be translated from everyday language to a clumsier formulation is not a reason for preferring the latter. Especially so, when one considers that the implication of the clumsy formulation is that what people value in a free society is not the extent of their freedom, but ‘the probability of being free to do R, S, …’. We claim that the extent of someone’s freedom is specified by their probability of being free to do R, S, …, but see no point in declaring that people are wrong about what they claim they value. In other words, analysing these claims in terms of a formal language (a translation) is preferable to attempting to change the claims into a language most people would not recognize.



The construction of rights249

For Steiner, rights pertain to act-tokens, not to act-types. If one nevertheless wants to make claims about the existence of act-type rights, it seems there is only one way of carrying out such an analysis that is compatible with Steiner’s approach. As we saw, having a right, for Steiner, means possessing a part of the world. If a person has the act-type ‘freedom of speech’, then there should be at least one act-token instantiating it that the person can be sure to realize. The main non-compossibility problem described thus far – the impossibility of ensuring that a person can perform some acttoken – can now be taken to mean that the probability that some particular act-token will not be prevented will usually not be 1. If δ indeed stipulates that the value of at least one of the relevant probabilities should be 1, then rights are non-compossible. But this is too strong. We cannot be said not to possess the right of free speech, the right to attend a church service, the right to post a letter, and so on, simply because in some circumstances we are not to be able to exercise them. We cannot say that the material existence of a right depends on at least one of the relevant act-token probabilities having a value of 1. It seems much more natural to suppose that the existence will depend on the value that the function G assigns, and this value may be positive even if all of the relevant act-token probabilities are smaller than 1. Indeed, as the extensive recent literature on measuring freedom suggests, there are many ways in which the function G could be defined (see, for example, Chapter 9; Pattanaik and Xu 1990, 1998; Sen 1991b; Arrow 1995; Carter 1999; Sugden 1998; Rosenbaum 2000; van Hees 2000a). One possibility is for the value to be determined by the act-token with the highest probability of not being prevented. Thus if at Piccadilly Circus, I have the highest probability of not being prevented from protesting, then that probability describes the extent of my freedom of speech. Another possibility might be to take the average values of the various probabilities. For instance, one could argue that the extent of freedom is severely limited if one is only permitted to protest at a particular place – the other probabilities should therefore also be taken into account. Deciding between these and the many other possibilities may be a very difficult task, and this is not the place to explore this issue further.15 Let us simply suppose that we have a particular G with which we can establish the extent to which a person is free to perform various act-types. This information is used subsequently to define which act-types a person is free to perform. Again, an all-or-nothing approach is counter-intuitive: we do not say that a person lacks the freedom to perform a particular act-type if, and only if, the extent of that person’s freedom to perform that act-type is 1. It may be more plausible to claim that for each act-type right a specific threshold-value exists: if the relevant G-value exceeds that value, the

250 Freedom

person has the right; otherwise she does not. Although there is no particular point in time or public place at which I can be absolutely sure of having the opportunity to protest against the government, the value of G will be sufficiently high to say that my right is unaffected.16 The proposition δ, which is used to determine whether an individual right exists materially, is thus true. Such a method departs from the usual approaches to the measurement of freedom. Rather than starting with the freedoms an individual has and subsequently building a judgement about the extent of the person’s freedom, the order is reversed. First, we examine the extent to which a person is free to perform certain act-types. That extent, which is specified by G, and a specification of the various threshold values, subsequently determines which (material) freedoms and rights an individual is said to have. In other words, first we measure how free individuals are and then we determine which freedoms and rights they have.17 Although we illustrated the approach only with respect to the establishment of freedoms, the framework is general enough to derive judgements about the existence of other types of rights. This is just a brief outline of an alternative approach. Many questions still need to be addressed. Particularly relevant is the question how the various Hohfeld types of rights can be distinguished within our framework. Although the probabilities pertain to the freedom to perform an act and thus to liberty-rights only, the other Hohfeld types can be described in terms of these liberty-rights. For instance, to say that A has a claim-right that B performs x means that B has the liberty to perform x and does not have the liberty not to perform x, and it also entails that A has the liberty to perform an action (i.e. to waive the duty) to bring about a state of affairs in which B does have the liberty to abstain from performing x. Moreover, many questions need to be addressed with respect to the nature of G and the various threshold values. We shall not go into these questions here, but we mention various advantages of the approach that may help to motivate further exploration. First of all, the use of threshold values enables us to make clear that for some rights act-token incompossibilities matter more heavily than for others. At the same time, the costs of raising the probabilities of rights being exercised also matter. For example, access to medical care may be thought to be more important than being able to post a letter at a nearby letterbox. But ensuring that one may have one’s health problems sorted out, no matter what they are, is far more costly than ensuring good access to letterboxes. Both importance and cost need to be considered when considering threshold values. Second, the approach may enable us to incorporate the relation between intentions and rights we mentioned above. After all, others’ intentions



The construction of rights251

affect the probabilities. If others intend to prevent me from going to the voting booth, and they have the means to carry out – or at least try to carry out – these intentions, then the probability that I will be able to vote will affect the relevant G-value. My right is violated only when the relevant G-value falls below the threshold value. The threshold value may be much lower when non-intentional acts stop me voting. For example, if the journey to the voting booth is very long and hazardous with a high probability of failure, even though no one is trying to stop me, we may feel that I do not have the right to vote. (Or imagine there are so few voting booths open, for such a short period of time, that it is not possible for all those registered to vote to do so. As recent events have shown, this example is not fantastic.) In this case, the low probability of my being able to exercise my right should be obvious, and one may legitimately expect the state to increase that probability – by increasing the number of voting booths, allowing postal votes, and so on. This approach allows us to recognize, however, that one’s right to vote is not always violated when someone does something to stop me, as in the traffic accident example discussed above. What matters is the probability of being able to exercise the right and the reasons why that probability holds. Third, the approach enables us to make clearer the distinction between rights and freedoms. Freedom is a generic term that refers to all possible acts, whereas rights refer to acts the probabilities of which are affected by legal or moral considerations. My right to vote implies that I am protected against others acting against me in certain ways and for which provisions are made to entail that I have a given minimum probability (the threshold value) of being able to exercise the right. Without that legal protection, the relevant probability would be lower. Fourth, we may note that this approach enables us to handle positive rights and negative rights in much the same manner. In both cases we are looking at the probabilities of our rights being, in the first case, respected and, in the second, exercised. It thereby also enables us to give due consideration to the moral value of rights and to the costs of protecting them. One may expect that any society may value the same negative and positive rights equally, but not expect a poor society to able to protect negative rights so strongly or provide so many positive rights as a rich one. Conclusion Rights are non-compossible in an important sense. Rights as defined in both social-choice-theoretic and the game-theoretic approaches are noncompossible. Steiner’s sophisticated attempt to specify a form of rights as act-tokens also fails to deliver rights compossibility in any interesting

252 Freedom

sense. If one is only free to do some action if one is subjunctively free to do it and others are not so subjunctively free, then the rights one has over such actions are non-existent or vanishingly small. Whilst his account is not strictly logically contradictory, when examined it looks more like a reductio ad absurdum of rights theory than a defence. Certainly it is far from the nature of rights as recognized in traditional political theory and in everyday moral talk. The non-compossibility of rights is an embarrassment especially to those who want to argue that rights form the foundation or basis of justice. Those who see rights as an important component of our moral and political thought can view their non-compossibility with equanimity. We can still attempt to measure and quantify individual rights by the extent of individual freedoms to carry out actions and not be subject to the domination of others within the scope of the rights presupposed by moral theory. The extent of these freedom judgements does not require that individuals necessarily are able to carry out everything they wish under the terms of the rights. That is what is meant by rights non-compossibility. We cannot all exercise all of our rights simultaneously, but it does not follow from that that we do not have those rights. Where rights clash, we examine the grounds of the rights and other moral considerations to see which rightsclaim should prevail, under each circumstance. Moral theories are not complete – they develop over time – but our moral theory does guide such decisions. If rights are non-compossible, but form an important part of moral and political theory, it is natural to question the nature of their existence. Quantifying their existence conditions is not a trivial matter. Rights exist in many senses. They exist as presupposed by a moral or political theory. They exist as claims under such a theory. They exist in the sense that individuals recognize their existence and act accordingly. They exist as set out in legal documents subject to defence by the state and private litigation. We have begun to tease out some of these existences and place them in the context of their exercise and recognition. Whilst rights do not guarantee their own exercise, they have only a weak existence (a theoretical) existence if there is no chance of their being respected by others or of their exercise. This suggests that we need to examine the probability that individuals’ rights will be respected and that they may be able to exercise their rights. These judgements will depend upon prior moral considerations – which includes what types of actions and constraints need to be taken into account in the probability judgements, what kind of freedom function we understand, and so on. There is no reason to think that every right should have the same probability of being successfully exercised. Many considerations, both pragmatic and



The construction of rights253

moral, enter into consideration of what threshold value – or probability – each exercise should take. Different threshold values may exist for different sorts of rights, and such threshold values may have to be traded one for another in different societies. ‘Overall’ comparison of rights or freedoms across societies may be extremely difficult or even impossible. However, certain categories of rights and freedoms may be compared. Can the notion of compossibility of rights also be applied to our alternative conception of rights? Theoretically, it can. However, it is doubtful whether it is possible to give a proof of compossibility or non-compossibility to a reasonable set of rights within our framework. Given the logical relations between various right types, and the interdependence of the various probability distributions, constructing a (non-trivial) set of rights that can be shown to be compossible (or non-compossible) would be complex.18 Yet it can be safely inferred that since our probabilistic account of rights is weaker than the other accounts of rights we have described, compossibility results are more likely to occur. More importantly, however, the issue of compossibility has lost its urgency in our framework. Compossibility may be one consideration relevant in the construction of rights, and hence in establishing, for instance, the various threshold values, but there is in our framework no reason for giving it a priori overriding importance. Our formal approach to the nature and measurement of rights and freedom is more in tune with everyday intuitions, and standard accounts of rights in political philosophy than extant formal accounts. It provides a framework to start an assessment of the extent of rights in different nations understood in different senses, and allows us to measure rights and freedoms both through the statutory and institutional forms in which they may be identified in different countries and through their exercise. The number of people who exercise rights in one country relative to the number in another, taking into account other relevant features, gives us an indication of the probability that such rights can be exercised. The distribution of such exercise across different social groups also gives an indication of the bias in any system of rights. Notes  1 We say more about this issue below.  2 We use the term ‘right’ in the broadest sense. We examine the compossibility and existence of rights as they are discussed in both philosophical and everyday moral discourse. Thus the ‘right to free speech’ here may be seen as a liberty to speak on various topics, an immunity against being divested of that liberty, and rights against certain types of interference with that liberty. Similarly, the Sen rights specified in his example of the right to read a book

254 Freedom seemingly require, at least, a liberty and a property right. The co-possible rights Nozick (1974: 166) refers to must be both. We do not attempt to sort out these issues conceptually; rather, we deliberately use the term ‘right’ somewhat loosely, but claim that our analysis can be applied easily to a more technical Hohfeldian analysis of rights, entitlements, privileges, liberties, and so on. In places we suggest how that application to other Hohfeldian categories may be accomplished. Especially clear statements of a Hohfeldian position are given by Kramer (1998, 2001).  3 R describes the weak preference relation (‘at least as good as’). From R the strict preference relation P (‘better than’) and the indifference relation I (‘equally as good as’) are derived in the usual way.  4 This implies that the social preference relation needs to be acyclic and complete.  5 Note that we do not use the term ‘power’ in the Hohfeldian sense, in which it forms a type of right: that is, the one which describes the permissibility of bringing about changes in normative (legal or moral) relations. For a game-theoretic description of the various Hohfeld types of right, see van Hees (1995).  6 Note, incidentally, that an approach in which each individual has been assigned her own game form generalizes the one in which there is only one game form from which all effectivity functions are derived. The latter then simply forms a special case, to wit, the one in which all individuals have been assigned the same game form.  7 We have chosen to discuss Steiner in some detail, despite his taking a choicebased approach to rights as opposed to the more popular interest-based approach, since his work most carefully and thoroughly examines the issue of compossibility.  8 Some people maintain that ‘When we say rights conflict, what we really mean is that the duties they imply are not compossible’ (Waldron 1989: 506). But this is too restrictive. Wellman (1995: 201) points out that ‘Feinberg’s [1980] example reminds us that rights involve a variety of Hohfeldian positions. Any of these could conflict logically with its opposite. Just as the liberty of doing some act is the logical contradictory of a duty not to this act, so one’s power of effecting some legal or moral consequence is logically incompatible with one’s disability to effect this consequence.’  9 The nature of what the right of free speech defends is important. The fact that people can go to the shops and verbally enquire about the price of goods, or talk about personal matters with friends, relatives or their doctor, is not enough for there to be free speech. The content of what one is allowed to speak is what counts. Of course, a defence of the right of freedom of speech need not be based on the importance for the speaker, but may also be in terms of what listeners gain from it (Mill 1848). 10 As explained in the previous section, this conclusion only holds if the deontic realms of the individuals coincide. However, it is clear that the notion of possession as used by Steiner refers to possession of a part of the world that is the

The grammar of rights and freedoms

13 Social choice and the grammar of rights and freedoms

The tools of social choice and game theory are being marshalled in order to get a new handle on old concepts in political philosophy. There is a large literature on the nature of equality as a value and what we should expect to be equalized (see, for example, Roemer 1996, 1998). There is a growing literature on the nature of liberty, its potential measurability (for example, Jones and Sugden 1982; Sen 1988, 1989, 1990a, 1990b, 1990c, 1991b, 1992, 1997, 1999; Bossert et al. 1994; Gravel 1994; Laslier 1998; Sugden 1998; Carter 1999; van Hees and Wissenburg 1999), and on the concept and form of rights (for example, Sugden 1985; Gaertner et al. 1992; Gaertner 1993; Deb 1994; Fleurbaey and Gaertner 1996; Hammond 1996; Pattanaik 1996; Sen 1997; Fleurbaey and van Hees 2000; van Hees 2000b). Much of this literature on rights was inspired by Sen’s ‘paradox’ of the impossibility of the Paretian liberal; Sen has been a key writer in the literature on the concept and measurement of freedom; whilst his ‘Equality of What?’ article (1982c) inspired much of the work on the nature of equality and its maximand. The justification for using the tools of social choice and game theory is that it leads us to re-examine our intuitions and understand the true or valid implications to be drawn from approaching these concepts in one way rather than another (Carter 1999: 17–18; van Hees and Wissenburg 1999: 80–1). Where intuitions are found to be contradictory or wanting in other regards, they need to be altered or given up; though, equally, if the results of analytic research prove to be too far adrift from our intuitions, we may wish to consider the appropriateness of the particular analytic framework adopted. This is not quite the same as the ‘reflective equilibrium’ of Rawls (1972), which is concerned with the way in which we develop grand theories of ethics and politics from our intuitions, reflecting back on the intuitions from the implications drawn from the theory. The interrogation of theory by intuitions, and intuitions by theory leads to reflective equilibrium (Rawls 1972: 48–51).1 Here, rather, we are discussing



The grammar of rights and freedoms257

the language in which we can represent both our intuitions and our grand theories. The analytic language adopted must be rich enough and utilized well enough to translate the intuitions and grand theorizing of ordinary language and traditional political philosophy if it is to fulfil the role of challenging the implications that may be drawn from them. Do social choice and game theory handle rights adequately? I argue that the grammar of rights and freedom developed by traditional legal and political philosophy has all too often been ignored by writers in the social choice tradition, much to the latter’s detriment. I do not suggest that formal approaches cannot appropriate the correct grammar; my aim in that regard is much more modest – merely to point out their failure to do so.2 Minimal Liberalism (ML) Recent work on rights in social choice literature has grown from a critique of Sen’s (1970) characterization of rights in terms of his Condition L* or Minimal Liberalism (ML) (see also Sen 1982a). The ‘impossibility’ result has produced an enormous literature discussing whether or not it is possible to produce Pareto-efficient outcomes in a liberal society.3 We need not be distracted by that debate here. We are only concerned with the characterization of rights contained in ML to see if they are adequate and consider the nature of the inadequacy if they are not. Condition ML states that there are at least two individuals for each of whom there is at least one pair of alternatives over which she is decisive. A person is decisive if there is a pair x, y such that it is always the case that if she prefers x (respectively to y) to y (respectively to x), then society should prefer x (respectively to y) to y (respectively to x). The intuitive problem with Sen’s characterization is that the domain of the rights-holding – or that which a person is said to be decisive over – is broader than that over which we normally think of people as having rights. This can be illustrated with his own example of the rights of i and j to read Lady Chatterley’s Lover. In Sen’s example, i (the Lewd) would like to read the book. Person j (the Prude) does not want to read the book. However, the Lewd would prefer the situation where the Prude reads the book to the situation where he reads it himself. Similarly, the Prude would rather read the book himself than have Lewd reading it. Sen’s impossibility theorem was designed to show that each player choosing in terms of his rights to read or not read the book would mean that the outcome both preferred (the Paretian or welfarist outcome) would not be attained.4 This is the paradox of the Paretian liberal. It can be represented, below, as a standard normal-form game, which looks like a Prisoners’ Dilemma (PD) where R means read, and ~R means not read the book.5

258 Freedom

Prude (  j )   R  

~R

   ~R  y   w Lewd (i )     R   x   z 13.1  Sen’s Liberal Paradox as a toy game

Thus outcome y (which from now on I will usually call (~R, R)) is what the Pareto principle suggests should be chosen and is both Lewd and Prude’s second choice. Since R is a dominant strategy for Lewd and ~R for Prude, (hereafter (R, ~R)) game theory says will z be the outcome. The ‘paradox’ of the single play PD has no ‘solution’ either. Of course, this outcome does not fit our intuitions, and most people when thinking about the PD for the first time feel that surely people can come to some arrangement to get out of the fix. But one of the functions of game theory and toy games like singleplay PD is to allow us to inspect our intuitions and see if they provide the correct answer. In this case our intuitions are wrong, probably because we rarely, if ever, face a single-play PD. As is well known, there are two general ways in which the optimal (~R, R) outcome can be attained in a PD. Either we have some outside enforcer who makes Lewd and Prude play the relevant strategies or the game is iterated, in which case the optimal solution becomes attainable. Sen’s Paretian liberal paradox assumes that both Lewd and Prude are able to exercise their rights by reading or not reading the book. Exercising their rights as predicted by their preferences will lead Lewd to read and Prude not to read. As it happens, Prude would prefer to forgo his exercise of the right not to read if it would cause the Lewd not to read the book; and Lewd would prefer to forgo his exercise of his right to read, if this would cause Prude to read it. Thus the exercising of their rights to read or not read as predicted by their preferences will lead to a suboptimal outcome. We are not concerned with the results of this analysis here, only the analysis of rights under condition ML. Sen makes it clear that ML does not sufficiently characterize rights. It does not provide a definition of rights. However, he claims that any such definition must include within it the characterization contained in ML. In other words, ML is a necessary but not a sufficient condition for rights-holding. I do not challenge condition ML here (though see below). But the formal characterization of rights has moved beyond Sen’s condition as the game-form approach has developed. The doubts about ML may proceed from two sources. The first is that Arrow’s theorem may be proved via Sen’s theorem (Binmore 1994: 131–2). This means that, under Sen’s analysis, no more than one person



The grammar of rights and freedoms259

can have a right. Though mentioned in some introductory texts (Craven 1992: 111–15), this fact has received little attention in the literature (though see Chapter 12). The second, related, reason is that it is doubtful that individuals have rights over outcomes (or social states); rather, they have rights over strategies. Thus in the Lady Chatterley example above, Prude and Lewd have rights over the strategies R and ~R. The game-form approach develops this insight.6 The game-form approach Recently, social choice writers have developed the game-form approach to the study of rights and freedoms (Sugden 1985; Riley 1989, 1990; Gaertner 1993; Gaertner et al. 1992; Hammond 1995, 1996; Pattanaik 1996; Pattanaik and Suzumura 1996; van Hees 1999, 2000b, 2004; Fleurbaey and van Hees 2000). A game form is a game where the utility function of each player remains unassigned. So it includes a list of the players; for each player, a list of alternative strategies; and an outcome from every combination of strategies that might be chosen. Rights and freedoms in the game-form approach are then analysed as sets of permissible strategies. The game-form approach was designed to characterize rights and freedoms in response to Sen’s Paretian liberal paradox. It was not intended to overcome the ‘impossibility’ result, but rather to formally characterize rights in a manner preferable to condition ML (though see van Hees 2000b, 2004). There are two types of game-form approach. The first, based on Gardenfors (1981), suggests rights may be represented by effectivity functions. The problem with effectivity functions is that they define not rights but powers. The a-effectivity function defines an individual’s power to force an outcome from any set over which he is a-effective. This is a very strong notion of a right, entailing as it does that a person has admissible strategies that secure outcomes regardless of what anyone else does. β-effectivity does not guarantee that an individual can secure some outcome no matter what anyone else does; rather, it is defined by others’ powers to interfere. This seems to be a conceptual mistake. Hart (1973) long ago argued that one’s right to x is not compromised by your inability to x. Even if one wants to think about some types of rights in terms of ‘effective freedom’, one need only specify a threshold probability of success well below 1 for some rights (see Chapter 12). The second game-form approach provides a much more promising avenue of exploration. Normal game forms The social-choice literature has been much concerned with examples where meddlesome preferences cause problems. In the game-form context,

260 Freedom

Gaertner et al. (1992), using an example from Gibbard (1974), consider a case of the right to choose the colour of one’s shirt. In the example, two individuals, i and j, have two shirts each, one white (w), one blue (b). The simple game form characterizes each person’s right as a strategy, either W   j W B

W

w, w

w, b

  i

B

b, w

b, b

13.2  Shirt game: rights as strategies

or B, giving four potential outcomes (w,w), (b, w), (w, b) and (b, b) represented in the following matrix. Gaertner et al. (1992) examine three games under this game form, but the one considered here is the conformist–nonconformist game, where the players order the outcomes: i: bb > ww > bw > wb, j: wb > bw > bb > ww. Here i is a conformist who prefers to wear white when j wears white, and blue when j wears blue; whilst j is a nonconformist who prefers to wear   B   W j

B

1, 3

3, 2



W

4, 1

2, 4

i

13.3  Conformist non-conformist shirt game

blue when i wears white, and white when i wears blue. The normal form game appears below (with the orderings for each player: 1 > 2 > 3 > 4): There is no equilibrium in pure strategies for this game. Gaertner et al. (1992) analyse the game with each player playing maximin, so i will play B and j W. The justification for choosing maximin is that each chooses in ignorance of the other and chooses only in terms of his preference for shirt colour, not in terms of the conformity condition. If the game is played repeatedly, the option of mixed strategies must suggest itself, since the best response for i to j’s pure strategy W is W, and j’s best response to i’s pure strategy W is B. However, the choice of strategies makes no difference to their analysis of rights. In no strategy choice can a player be assured of his Sen right, because there is no guarantee that either will be decisive over the outcome pairings. However, as each has only a right over a strategy choice, a game-form right has not been violated. This seems more in keeping with standard understandings of rights in political discourse and philosophy.



The grammar of rights and freedoms261

Because the definitions of rights that Sen and Gibbard provide are not satisfied by the rights given in the game-form approach and are thus found wanting, Gaertner et al. (1992: 163) suggest the following weaker formulation: ‘For every individual i, there exists distinct social alternatives x and y such that i is locally decisive over {x, y}’. Here i is locally decisive when x and y are the only two alternatives and i strictly prefers one to the other. They suggest this is preferable to Sen’s formulation, provided one makes the ‘very plausible assumption’ that ‘For all individuals i, and for all social alternatives z and w, if z and w are the only available social alternatives and if i strictly prefers z to w, i will choose z and reject w whenever he is empowered to act on behalf of society’ (Gaertner et al. 1992: 168). The social-choice language here will be anathema to traditional political philosophers. The idea that an individual is empowered to act on behalf of society whenever he exercises his rights is completely alien. Barry (1991c: 93–107) says as much in his contribution to the Sen paradox, to which Sen replied that the language does not matter, only the analysis (Sen 1986: 224–5). However, I argue below that because these examples do not begin from the right place they go astray almost immediately. In order to be able to analyse even simple examples, like the right to read a book or choose what colour of shirt to wear today, we must begin by considering the grounds of the rights-claims. Only after we have understood the moral grounds upon which such rights-claims are based, can we consider how to formally analyse such cases. Only then can we understand their limits and existential scope. The grammar of rights talk is conceptually important, and because it has not been taken seriously, the formal social-choice and game-form approaches go astray. Fleurbaey and van Hees (2000: 296) analyse the Gibbard conformist– nonconformist game in two ways. First, given a legal game form, we can derive rights of individuals through an associated a-effectivity function. If i has an admissible strategy that leads, regardless of what others do, to an outcome in which i wears a blue shirt, then i can be said to have the right to wear a blue shirt. Second, given a specification of the rights of individuals, we can seek an ordinary game form that implements those rights. We then look for game forms whose associated a-effectivity functions coincide with the specified rights. This is the approach Fleurbaey and van Hees (2000) adopt. Effectivity functions in the game-form approach to rights are defined in terms of which someone is effective in bringing about some outcome. Fleurbaey and van Hees (2000) do this through a characterization of positive and negative, active and passive rights. The problems with their account (specified below) are due to this characterization rather than to the formalism itself. What we might say about their formalism is that it

262 Freedom

demonstrates why rights cannot be characterized as positive in their terms, and why their versions of active and passive rights do not constitute a good characterization. The grammar of rights and freedoms The social-choice literature does not carefully distinguish between rights, freedoms and powers. Both grammatically and conceptually they are different, and these differences are not trivial. Grammatically, it is easy to see they differ. Verbal oddities abound if one simply substitutes the word ‘freedom’ or ‘liberty’ for right or vice versa (White 1984). For example, we can be more or less free, but we cannot have more or less of a right – we either have it or not. Even though what we have a right to do and what we are free to do may be the same, being free to do x does not imply having a right to do x. We may be free to ask a favour of another, but we do not have the right to do so. We may be free to enter Australia on a tourist visa, but we do not have a right to do so. Conversely, we may have a right not to be impeded in some action, but not be at liberty (in Hohfeld’s sense, see below) to carry it out (Kramer 1998: 16).7 Traditional political and legal philosophy has identified at least ten different ways of analysing the simple statement ‘i has a right to x’ (see Waldron 1984, for example). We do not need to go into each of these, merely to acknowledge that social-choice and game-form approaches should take note of them. I will utilize Hohfeld’s (1919) categories to distinguish different sets of rights from freedom, then compare with how the formal approaches use rights and freedom, in order to show that the formal approaches are not, as yet, really characterizing rights as they should. Hohfeld produced eight categories, four of which produce legal positions entailing one another and four of which produce ‘jural contradictories’ (see Figure 13.1).8 The vertical arrows indicate mutual entailment. The diagonal arrows indicate mutual contradiction. So someone’s right implies a duty upon someone and someone’s duty implies someone’s right. If someone has a duty to do something for someone, he is not at liberty to do it – liberty implies a choice over whether to do it or not. If I am at liberty to do something, that implies a ‘no-right’ for everyone else (there is no one who has a right overriding my liberty, though many other people may be at liberty to do the same thing). Hohfeld also refers to a right in Figure 13.4 as a ‘claim’, and I will refer to such rights as ‘claim-rights’ hereafter (with the term ‘right’ standing for rights more broadly, including liberties and claimrights). I shall not refer to all his categories. One important distinction, that will be put to some work below, is the contrast between Hohfeld’s



The grammar of rights and freedoms263 Right

Duty

Liberty

Power

No-right

Liability

Immunity

Disability

13.4  Hohfeld’s categories

claim-rights and liberties. A claim-right ‘must be specified by reference to the actions of the people who bear the correlative duties – rather than the actions of the people who hold the rights – whereas liberties must be specified by reference to the action of the people who hold the liberties’ (Kramer 1998: 13). What are the grounds of rights? Traditionally, there are two main theories: the ‘choice theory’ associated with Hart (1973) and, more recently, Steiner (1994, 1998, 2001) and the ‘interest theory’ associated originally with Bentham (1843) and subsequently with others (Lyons 1969; MacCormick 1977; Raz 1984, 1986; Waldron 1988; Kramer 1998, 2001; see also Tuck (1979) for a historical account of the conflict between the two approaches and Steiner (1994) for a different take on them). The choice theory singles out the rights-bearer by virtue of the power he has over the duty in question. That is, person i may discharge person j with regard to j’s duties toward i. It is this control over the duty that makes i the rights-holder. But this entails that those unable to waive such duties – dead people, babies, animals, and so on – cannot have rights. It is also sometimes argued that certain fundamental rights, such as the right not to be maimed, tortured or killed, cannot be waived. The benefit theory is based on the idea that j’s duties towards i are based on some aspect of i’s interest, and that interest is so closely related to the duty that we can know a priori that if the duty is not carried out, the interest will not have been advanced. (So i’s right to be repaid the loan he made to j is contained in j’s duty to pay i back. But person k, who is awaiting payment from i, does not have any rights against j, even though i may not pay k until the loan is paid back.) Either way, when putative rights conflict, we must examine the foundations or bases upon which these rights are thought to rest. If one right is to be overridden in favour of another, this will be because the grounds of one are thought to be more important or urgent than those of the other. We must search for answers to rights conflict by the justification for their assertion not in the assertion itself.

264 Freedom

Back to ML Sen’s condition ML is supposed to be a necessary but not sufficient condition for a right. It states that there are at least two individuals for each of whom there is at least one pair of alternatives over which she is decisive. Does it really give a necessary condition for rights-holding? Consider person i. For two pairs of alternatives x, y and z, w, person i is decisive if society should prefer whatever alternative person i prefers. ‘Prefers’ here is a term of art in social choice (Sen 1986: 224), not to be taken in its literal meaning. According to Sen, we may interpret the phrase in a number of different ways. We can interpret decisive to mean i has a right to choose x or y. In that sense, i’s decisiveness can be seen as a power. This should entail power in Hohfeld’s sense, but is properly interpreted as power in the causal sense as the infinitive form of able. This would be under Sen’s ‘choice’ interpretation. It could also mean, under Sen’s ‘desire’ interpretation, that if i prefers x to y, then the constitution of society should allow x (see, for example, Sen (1992: 142–3) for an account of the choice and desire interpretations of preference functions). Either way, in Sen’s interpretation of ML, if xPi y, then x rather than y should obtain. Here only the easier-to-understand choice interpretation is considered, though below I will turn briefly to the desire interpretation. The choice interpretation is easier to understand since, given x and y are alternatives, and the only alternatives in the set, then either x or z must obtain, and choosing x seems the most straightforward way of understanding this concept. What is the x or y that might obtain? They are, in the parlance of social choice, ‘social states’. Social-choice writers usually describe social states as ‘complete descriptions’ of a given situation. However, when examples are discussed, the complete description gives way to a partial description that may be seen as an outcome from a set of actions, such as two people independently choosing to read a book. The first odd thing to note about someone having a right over a social state is that, ordinarily, people do not have rights over social states or outcomes, only over actions (their own or others). Even in Sen’s example, the right is ‘to (not) read Lady Chatterley’s Lover’, not the right ‘to be (not) reading Lady Chatterley’s Lover’. The difference between the two sentences may seem small, but is significant. The right to read Lady Chatterley’s Lover does not entail the freedom to read it now. The right here seems to be a Hohfeldian liberty, and such a liberty entails that, given other conditions giving access to a copy (conditions specified by other (claim) rights – see below), the individual can be in a state of reading at some time. It does not give her the liberty to be reading it at all times (when driving a car, for example). But the liberty to read the book does not entail that if an



The grammar of rights and freedoms265

individual is not reading the book she is exercising that liberty. Choosing to read the book is an exercise of such a liberty. Choosing not to read it is an exercise of the liberty. The former entails reading the book, the latter entails not reading it. But whilst having read the book entails that I have exercised my liberty to read it, not having read the book does not entail having exercised the liberty not to read it. I may simply not have read it.9 (I may not be aware of its existence, for example.) Only if one has chosen not to read it, has ‘the right (liberty) to not read it’ been exercised. The choice to exercise such liberty in one way or another may depend upon other conditions. For example, we may promise not to read the book – that is the upshot of the ‘contractual solution’ to Sen’s theorem. However, our reasons for choosing to exercise this liberty are irrelevant to the grounds of the right. Rights are not justified in terms of people’s reasons for exercising them (and certainly not grounded in reasons which depend upon others’ choice of whether to exercise them). We might have various grounds for the liberty of reading Lady Chatterley’s Lover, the most obvious being freedom of speech or thought. Given these grounds, any one of the four possible outcomes of the choices to read or not read may ensue, but the grounds do not justify any of the outcomes. That is, the social state ‘i reading and j not reading Lady Chatterley’s Lover’ is not justified by the grounds of the rights of i and j. The social state is caused by the actions of i and j, and i and j’s actions may be justified (though not fully: see below) by their having the rights to act as they do. But justifications do not follow all the way down. Public rights of way do not justify crashing one’s car into someone else’s. Fault in accidents may be determined by priority, with the basic rule that whoever occupies a part of road has priority over others, unless other rules of the road are broken. So the outcome is caused, but not justified, by the choice given the grounds of the rights. In that sense, it is false to say that the liberty to read Lady Chatterley’s Lover is a right over a pair of social states. In one sense, the right over pairs of states seems correct. Given j is at liberty to read Lady Chatterley’s Lover, then what happens over one aspect of the social state – the second entry in the outcome pairs (R, R), (~R, R) and (R, ~R), (~R, ~R) is determined by j. Person i has no rights over that entry. Person i only has liberties of the sort specified above over the first entry in the outcome pairs (R, R) (R, ~R) and (~R, R), (~R, ~R) – where ~R entails not reading the book at any time and R reading it at some time. In other words, i is at liberty to read the book. The outcome is a result of two sets of liberties being exercised in one way or another. In this way we can see a disjunction between rights over actions and rights over social states or outcomes. This is hidden in Sen’s example by distinguishing the social states in terms of variables (w, x, y, z) and giving

266 Freedom

preferences over the social states and not the actions. (There is another disjunction, of course. One may act on a right because it will lead to some outcome, but the right does not guarantee the outcome. I may read Lady Chatterley’s Lover because I think it is going to erotically stimulate me, but in fact it bores me to tears. Again, the grounds of the right are not contained in my preferences over exercising the right, so the satisfaction of those preferences is irrelevant to the question of whether or not the right has been satisfied.) This is made even clearer in Gibbard’s game of choosing which colour of shirt to wear, specified above. The liberty of choosing our shirt colour may be grounded in the general right of freedom of expression. We may have strategic reasons for our choice, as the grounds do not constrain the reasons for freedom of expression. For a Sen right given by condition ML, i must be decisive over either {(w, w), (b, w)} or over {(b, b), (w, b)} and j must be decisive over either {(b, w), (b, b)} or over {(w, b), (w, w)}. Thus the conformist (i) gets to choose a shirt the same colour as j’s and the nonconformist (j) gets to choose a shirt a different colour from i’s But the right to choose our shirt colour does not entail this. Person i may choose to wear blue because j does, and j may choose white because i chose blue. The grounds for these liberties do not give rights over any of the four possible outcomes. No one has ‘a right to dress in conformity with one’s neighbour’; no one has ‘a right to dress differently from one’s neighbour’. Whether or not we may copy another’s clothes may be open to litigation based on rights contained in fashion patents; we may even perhaps obtain a restraining order on someone for harassment if they always dress as we do and we can show this causes genuine harm. But the grounds of the rights leading to these outcomes (forcing nonconformity, for example) are not based on any rights to dress similarly or differently to others. The grounds for such rights are far more general than this. The distinction between negative and positive freedom is well known, though its precise character is contested. Without entering into dispute, we can note that negative freedom – freedom from constraint – and positive or effective freedom – capability of attaining one’s ends – are both negative concepts in another sense. In law, liberty – whether negative or positive in Berlin’s terms10 – implies an absence of any duty (in morals, an absence of any obligation) to do otherwise. If we want to do something that we are effectively free to do, then we may do it. There is no duty not to. But the mere absence of an obligation or duty not to do something does not entitle one to do it. In contrast, a claim-right is a positive concept. If we have a claim-right to do something, we are entitled to do it. So the liberty of reading Lady Chatterley’s Lover does not entitle me to read it (Peacock and Rowley 1972).



The grammar of rights and freedoms267

I may be negatively free to read it because of that right (no one is stopping me), yet not effectively free (there is no copy available). I may also be entitled to read it, if, say, I own a copy to which I have title. But entitlement comes under a different right – a claim-right and is not contained in the right (liberty) to read it given to me under freedom of expression. I may also be entitled to read it if I am a member of a library and the library has a copy. That entitlement will be constrained by the rules governing lending and borrowing from that particular library, such that I may not be able to read it on the day I want to. Again, my entitlement to read the book is not given to me by my (general) right to read the book grounded in freedom of speech, but by the (special) lending claim-rights I have as a member of the library. The positive nature of a claim-right of this sort, in contrast to the negative nature of freedom consists in an antecedent characteristic, namely the presence of a title in the holder of the right. It is therefore not at all clear that ML provides a necessary condition for rights-holding or for the exercise of rights, as ML uses the notion of a social state, and the concept of a social state includes within it a description which is not contained within the grounds of a right. It is, of course, true that, given the assumption that the four outcomes are the only ones possible, i’s choice, as it happens, is one between one of the two pairs of alternatives. In that sense, therefore, i is choosing one social state over another. However, since i is not choosing that social state alone, in another sense i is not choosing the social state. He is choosing only one element of its description. This is not puzzling, because it is not puzzling to see how two people who crash their cars on a public right of way did not choose to crash them, despite their exercising their liberty.11 The authorities may have to decide, in any particular car accident, who had priority and who was most at fault. The authorities would take a completely different view of the crash if ‘choosing to crash’ entered into the description of either’s reasons for action. Certainly, the liberty exercised by each when driving is not going to be grounded in their reasons for choosing thus, if the desire to crash was part of that reason.12 It can be seen, therefore, that the description of the right is important to its grounds, which are not directly related to reasons for exercising the right in one way or another, and may have nothing to do with correct (and certainly not complete) descriptions of the social state obtained through (joint) action. To confuse, as ML does, the description of the social state obtained as a necessary feature of (the exercise of) the right is a mistake. The game-form approach, I take it, is an attempt to get away from social states entering into the analysis of rights. This is an important move; however, the game-form approach still uses examples that are analysed before the general grounds of the rights under consideration have been investigated.

268 Freedom

Back to game forms Fleurbaey and van Hees (2000: 296) suggest that Gibbard’s example can be elucidated in two ways. First, given a legal game form, one can derive rights of individuals through an associated a-effectivity function. If i has an admissible strategy that leads, regardless of what others do, to an outcome in which i wears a blue shirt, then i can be said to have the right to wear a blue shirt. This may be a sufficient condition for a right. My wearing a blue shirt, by using an admissible strategy, regardless of what others do, may show that I have a right to wear a blue shirt. However, it is not necessary, since I can have the right to wear a blue shirt even where the outcome of my wearing a blue shirt does not obtain. The entire work of the sufficient condition of the ‘right to wear a blue shirt’ is captured in the phrase ‘admissible strategy’. As long as we have carried out an admissible strategy, we have the right. But ‘admissible strategy’ is not sufficient for a claim-right. We may be able to always reach outcomes through admissible strategies, but this does not demonstrate that we have a claim-right to that outcome. Just because we are free to do some x, or obtain some x, does not mean we have a claim-right to x. Indeed, it is not clear that we can always obtain some desired outcome, such as wearing a blue shirt, even if we are at liberty to choose what colour of shirt to wear. We can always think of conditions where the desired outcome is unobtainable without the rights-holder’s rights being violated. Of course, we may have strong moral views on lack of freedom to exercise rights, but that is not the same as asserting that the liberty or claim-right does not exist.13 The second approach to the issue is the one Fleurbaey and van Hees (2000) actually adopt: given a specification of the rights of individuals, we can look for an ordinary game form that implements those rights and then for game forms whose associated a-effectivity functions coincides with the specified rights. Effectivity functions in the game-form approach to rights are defined in terms of which someone is effective in bringing about some outcome. But again, effectivity functions are too strong for any account of rights. For example, i’s right to wear a blue shirt is not contained by the set of outcomes in which he wears one. Person i also has the right to wear a blue shirt in outcomes where he does not wear one under at least two sets of conditions: (1) where he chooses to wear a white shirt; (2) where he wants to wear a blue shirt, but is unable to choose one (but where his right to wear one has not been violated: that is, everyone else has only followed admissible strategies).14



The grammar of rights and freedoms269

Fleubaey and van Hees’s game-form approach can accommodate these problems but not the following. Rights still exist, even where we are not effective in bringing about some outcome and someone has not followed an admissible strategy. What matters is whether someone has deliberately stopped us from being effective in an inadmissible manner. Imagine I am on my way to the polls, but am held up in heavy traffic, and the polls are closed when I get there. My right to vote has not been infringed, even though my strategy of voting – driving to the polls – was not effective. My right to vote has not even been infringed if the heavy traffic was caused by an accident due to someone doing something illegal – driving faster than the legal limit. My right to vote would only be infringed if someone did something illegal in order to stop me voting, such as a refusal to recognize my registration card as legitimate (see Chapter 12). Fleurbaey and van Hees (2000) introduce a distinction between negative and positive rights. In Sen’s example, some person i’s negative right implies that some person j should refrain from any action that prevents me from reading Lady Chatterley’s Lover. This is too strong. Person j may buy the only available copy and thereby prevent i from reading it. Positive rights guarantee against the inaction of others, so that my right to read a book entails that I actually have the opportunity of doing so – for example, the state must make it available in a library (including both liberties and claims). But this is too strong, unless we understand opportunity not to entail that I am always able to read it when I wish. For example, it does not imply that no one else can borrow it on the exact dates and times I want to borrow it. Fleurbaey and van Hees (2000) define a ranking of systems of legal rules on the basis of the amount of freedom they grant to agents and then define a partial ordering over the set of systems of rules, saying that a system of rules maximizes freedom in a given subset if it is maximal element in this subset for the ordering. They define the state of nature where no legal system exists as the system where all strategies are admissible. Maximal freedom may mean minimal power. It also means no claim-rights. The approach may be able to characterize freedom in some sense, but does not capture Hohfeldian liberty, effective freedom or other intuitive notions, nor does it capture claim-rights. Game-form writers have liked to take up in Feinberg’s account of active and passive rights (Gaertner et al. 1992; Fleurbaey and van Hees 2000). For Feinberg (1973: 60), active rights are ‘rights to act or not act as one chooses’ and passive rights are ‘rights not to be done to by others in certain ways’. Active rights include the right to vote, to follow our chosen religion, to smoke, and so on, whereas passive rights include the integrity of our bodies, property, a clean environment, safe working conditions, education for our

270 Freedom

children. But Fleurbaey and van Hees (2000) choose not to use the distinction in Feinberg’s manner of imposing obligations on others; it should, they say (2000: 305), ‘rather be seen as a distinction between rights that secure access to (or the opportunity to choose) certain actions, and rights that secure a certain state of affairs’. They are correct that active/passive distinction cannot be made in terms of obligations. Active rights oblige others not to interfere in our actions in some ways. But passive rights also involve the obligations of others to forbear from some actions, and whilst their approach may help to overcome debates over positive and negative duties, and over acts and omissions (see Bennett 1983), it also has problems of its own. A different distinction that captures Feinburg’s point is the one noted above between positive rights – the right of receiving something – and negative rights – the rights of forbearance. These differ in the obligations owed to us by others. White (1984) sees this as a distinction between our rights to receive and our rights to act. Our right to receive may need the assistance of others, but not necessarily oblige them or give them a duty; but our right to act does oblige others not to interfere (White 1984: 17). Receiving something – an education, food, health care – is a positive right. It does not entail that anyone is generally obliged to do anything, but it imposes special obligations on people – teachers, doctors, nurses or other agents of the state – to help us (cf. Steiner 1994: ch. 3). But even these special obligations do not entail the outcomes of the positive rights. My right not to starve (or, better, my right to be nourished) can only oblige the state to feed me if the state has the food available. My right to be educated is only a right for others to try to teach me, not to actually educate me, which, at the very least, requires my cooperation. My right to health care does not entail a right to be healthy and, if the health services are strapped for cash, may not even include all the help that may be possible. It is worth recalling what some of these rights are. Active rights include the right to vote, to follow one’s chosen religion, to smoke, and so on. Fleurbaey and van Hees’s (2000) formalization does not capture the logical form of these examples. As we saw above, rights still exist even where we are not effective in bringing about some outcome. What matters is whether someone has deliberately stopped us from being effective in an illegal manner. My ‘right to smoke’ (a liberty) is not infringed even if all retailers render the strategy of my smoking infeasible by not buying cigarettes from wholesalers in order to sell on to me. My active right to pursue the religion of my choice does not make it incumbent on anyone to provide the strategies that make this possible – churches, preachers, and so on. We can see that the Fleurbaey and van Hees version of the passive– active distinction has problems and that their formalism has not captured the elements of rights they are attempting. They say (2000: 308):



The grammar of rights and freedoms271 Active rights guarantee access to an action. Passive rights are, in a sense, stronger because they guarantee more than the access to a state of affairs. They also guarantee the state of affairs itself, although this guarantee is still limited insofar as it bears only on J’s obstruction.

But the nature of the obstruction is important to whether or not it counts as a rights violation: notably the intentions of others using admissible or inadmissible strategies. The problem is that it is hard, if not impossible, to define all obstructions as rights violations or not, in advance of considering the case. Rights must be about admissible strategies, but admissibility cannot be described simply in terms of the outcomes that it allows. Conceptual work on the nature of admissible strategies given the grounds of claim-rights and liberties is required. Similar problems emerge in Fleurbaey and van Hees’s account of β-effectivity functions. A β-effectivity function does not entail that an individual’s freedom to realize a state of affairs means she has a strategy that always leads to it. The b-effectivity function only entails that no other individual may see to it that the state of affairs will not be realized. Again, however, this function is still too strong. What matters is the nature of the obstruction, not the obstruction itself. Conclusion The problems I have addressed with the formal approaches to rights and freedom may be seen in terms of the type–token distinction. Analysts have begun by examining particular examples – act-tokens – and then trying to characterize rights by their intuitions over the simple examples. However, this is not the correct way to analyse rights. One must begin by considering the grounds of rights, which typically concern act-types. Steiner (1994) argues that rights need to be analysed in terms of act-tokens in order to ensure compossibility. Chapter 2 proves that rights characterized as act-types cannot be compossible, and as act-tokens the class of compossible ones are ‘vanishingly small’ and do not look much like traditional rights. This means that rights cannot be the grounds of any theory of justice. Indeed, most of the criticisms I have advanced here have been based on the view that we need to look at the moral grounds of rights in order to be able to tell whether, when individuals cannot exercise them, they are not being respected. We cannot expect to be able to exercise our rights at all times, but it does not follow that we do not have them. Chapter 2 shows how we may formally examine rights in this way. Game-form approaches may seem to provide the possibility of avoiding this problem, by specifying

272 Freedom

rights as permissible strategies or effectivity functions. However, we have seen that these approaches define rights far too strictly to conform to everyday discourse (our intuitions) and general philosophical analysis of these concepts (grand theorizing). It is true that one answer to the confrontation of formal analysis with less formal and intuitive understandings is that intuitions may be wrong or contradictory. However, I hope to have shown that this is not the case. By ignoring the grammar of rights and freedom, and ignoring the analysis of rights developed over many years of philosophical analysis, and plunging straight into the analysis of simple examples without considering the justification of the rights they are supposedly analysing, formal writers have gone astray. In even a simple example such as the right to read a book, the entailments for the state of affairs that may legitimately be expected require analysis of the right to read (as a liberty) and the entitlement to read (given by relevant property rights). Neither, without other conditions, entails ML, which therefore cannot be a necessary condition of rightsholding. This is so, whether or not we want to define rights and liberties as negative or positive concepts. Only when formal writers examine the grounds of rights and understand the complexity of rights and liberties talk will they begin to analyse their logical form. This will involve, I believe, examining the existence conditions of rights prior to analysis of their exercise through strategies in game forms. Notes  1 See Goodman (1955: 65–8) for the first use of this idea.  2 We may remain sceptical, however, given the paucity of the informational base of social choice and game theory (van Hees 2000: 344) – a paucity necessary for its strengths of formal manipulability, allowing insight into specific issues.  3 The ‘impossibility’ result is only an impossibility result in the language of social choice. It does not demonstrate that there is no possible world where all outcomes are ‘minimally liberal’ and Pareto-efficient. There are possible worlds where minimal liberalism and Pareto-efficiency (happen to) reign. Rather, it shows that there are possible worlds that are minimally liberal and Pareto-inefficient. The full set of possible worlds relevant to the questions is generated through condition U that allows each person to take on any preference ordering. The ‘impossibility of the Paretian liberal’ is actually a possibility, not an impossibility, theorem (as indeed is Arrow’s theorem).  4 More formally, Sen’s proof shows that there is no social choice function F that meets the following conditions: (1) Unrestricted domain (U): there are no restrictions on the domain W. No (combinations of) consistent individual preferences should be ruled out. (2) Pareto-efficiency (P): if all individuals



The grammar of rights and freedoms273

strictly prefer x to y, then the social preference ordering should rank x above y: xPy. (∀i: xPiy) → xPy. (3) Minimal Liberalism (ML): there are at least two individuals for each of whom there is at least one pair of alternatives over which the individual is decisive. A person is decisive if there is a pair of social states (x, y), such that it is always the case that if the person prefers x (respectively to y) to y (respectively to x), then society should prefer x (respectively to y) to y (respectively to x). (This is the weakened version of the condition of liberalism, L, which allots such powers of decisiveness to all individuals in society: see Sen 1970: 87.) More precisely: $i,j: i≠j $w, x, y, z: (w, x) ≠ (y, z) such that {(wPix→wPx)∧(xPiw→xPw)} & {(yPiz→yPz)∧(zPiy→zPy)}. Note that if it were the case that the pair (w, x) is identical to the pair (y, z), ML could never be satisfied. These pairs, therefore, have at most one state in common. For Sen, if i has a right with respect to the choice of w or x, then she is decisive with regards to w and x. Being decisive is necessary but not sufficient for having a right with regards to w and x (see Sen 1970: ch. 6). ML holds when at least two distinct individuals i and j have such a right.  5 See, for example, Fine (1975); Aldrich (1977); Sugden (1985); Sen (1986: 227); Hardin (1988: 111); Riley (1989, 1990); Gaertner (1993); van Hees (2000b). The game version is not identical to Sen’s social-choice theorem. Strictly speaking, the theorem demonstrates that his three axioms lead to the empty set. They cannot simultaneously be satisfied. However, another way of viewing this is to say that any outcome must break one of his conditions. In the game representation, P is broken.  6 Sen challenges the game-form approach. He suggests it may have some utility, but his own axiomatic approach is preferable overall, for it allows us to see rights and freedom problems more easily, since justification of the procedures (strategies) must be given in terms of consequences (social states) (Sen 1995: 14–15, 1996: 31, 1997: 24–9). In fact, Sen must challenge the game-form approach, as it obviously suggests the ‘contractual solution’ to his liberal paradox through the players trading their rights over the strategies.  7 Kramer’s example is a factory polluting a lake suing a group for vandalizing the outlet pipes. The factory’s claim against the group may be secure even if its pollution is illegal.  8 I follow Kramer (1998) in calling these ‘contradictories’ not ‘opposites’, and I follow him (and others who use the term ‘bare liberty) in calling Hohfeld’s privilege ‘liberty’. Both the terms ‘privilege’ and ‘liberty’ are potentially misleading, but I feel liberty is more closely akin to natural language, which is important when we are representing theories which should capture most aspects of our moral intuitions.  9 It does entail not having exercised the right to read it. Indeed, we may doubt whether the right to read a book entails a right ‘not to read it’ in the way formalized by Sen. Rather, it simply entails that we may not exercise the right to read. We have the right to read the Highway Code, but that right does not disappear when we are required to read it when learning to drive, nor our right

274 Freedom

10 11 12

13 14

to read Lady Chatterley’s Lover when required to read it for an English literature exam. Note that some aspects of positive freedom, however, may entail claim-rights. Again, we must keep in mind the distinction between right of way and priority in British road law. The only difference is that in the Sen example, reformulated as a Prisoners’ Dilemma, each has a dominant strategy. This means that neither need know what the other is choosing when making their choice – their preferences are independent of the other. But the dominant strategy also entails that they will choose as they do, even if they are aware that a suboptimal outcome will result. However, even in this case it is not true in one sense that they ‘chose’ the suboptimal outcome. The suboptimal outcome was the result of their choices, but the reasons for their choices were their independent preferences – their dominant strategy. Sen (1982d) understands this where, for example, he discusses the Prisoners’ Dilemma as a problem for revealed preference analysis. Some may claim that the nature of the right is given by the terms under which it can be exercised. See Chapter 12 for why I do not think this is the best way of approaching this issue. Furthermore, of course i may not have the right to wear a blue shirt even under conditions where i does wear a blue shirt. Individual i’s right to wear a blue shirt is derived from his right to choose what colour of shirt he wears. He does not have the right to wear a blue shirt where he is not able to choose his shirt colour – say in prison when he has to wear prison garb (this is true even when the prison garb includes blue shirts).

References

References

Acemoglu, D. and Robinson, J. A. (2006) Economic Origins of Dictatorship and Democracy (Cambridge: Cambridge University Press). Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. J. and Sanford, R. N. (1950) The Authoritarian Personality (New York: Harper and Row). Ahlquist, J. S. and Levi, M. (2011) ‘Leadership: what it means, what it does, and what we want to know about it’, Annual Review of Political Science 14: 1–24. Aldrich, J. (1977) ‘The dilemma of a Paretian liberal: some consequences of Sen’s theorem’, Public Choice 30: 1–21. Altemeyer, R. (1996) The Authoritarian Specter (Cambridge, MA: Harvard University Press). Anderson, E. (1999) ‘What is the point of equality?’, Ethics 109: 287–337. Archer, M. (1984) ‘Morphogenesis versus structuration: on combining action and structure’, British Journal of Sociology 33: 455–83. Archer, M. (1995) Social Theory: The Morphogenic Approach (Cambridge: Cambridge University Press). Arendt, H. (1972) ‘On violence’, in her Crises of the Republic (New York: Harcourt Brace Jovanovich). Arendt, H. (2004) The Origins of Totalitarianism (New York: Schocken). Arneson, R. J. (1989) ‘Equality of opportunity for welfare’, Philosophical Studies 56: 77–93. Arneson, R. J. (1999) ‘Equality of opportunity for welfare defended and recanted’, Journal of Political Philosophy 7 (4): 488–97. Arrow, K. J. (1951/63) Social Choice and Individual Values (New Haven: Yale University Press). Arrow, K. J. (1995) ‘A note on freedom and flexibility’, in K. Basu, P. Pattanaik and K. Suzumura (eds), Choice, Welfare and Development: A Festschrift in Honour of Amartya K. Sen (Oxford: Clarendon Press). Avolio, B., Walumbwa, F. O. and Weber, T. J. (2009) ‘Leadership: current theories, research, and future directions’, Annual Review of Psychology 66: 421–49. Axelrod, R. (1984) The Evolution of Cooperation (New York: Basic Books).

276 References Bachrach, P. and Baratz, M. (1970) Power and Poverty: Theory and Practice (New York: Oxford University Press). Banfield, E. (1961) Political Influence (New York: Free Press). Barber, S., Gordon-Dseagu, V. and Rafferty, J. (2004) Survey of Patients Offered Choice of Hospital in the Greater Manchester Area (London: College of Health). Barnes, B. (1988) The Nature of Power (Cambridge: Polity). Barry, B. (1974) ‘The economic approach to the analysis of power and conflict’, Government and Opposition 9: 189–223. Barry, B. (1988) ‘The uses of “power”’, Government and Opposition 23 (3): 340–53. Barry, B. (1989) ‘Does society exist? The case for socialism’, Fabian Tract 536 (London: Fabian Society). Barry, B. (1991a) ‘Is it better to be powerful or lucky?’, in his Democracy and Power: Essays in Political Theory, Vol. 1 (Oxford: Clarendon Press). Barry, B. (1991b) ‘Power: an economic analysis’, in his Democracy and Power: Essays in Political Theory, Vol. 1 (Oxford: Clarendon Press). Barry, B. (1991c) ‘Lady Chatterley’s Lover and Doctor Fischer’s Bomb Party: liberalism, Pareto optimality, and the problem of objectional preferences’, in his Liberty and Justice: Essays in Political Theory, Vol. 2 (Oxford: Clarendon Press). Barry, B. (2002) ‘Capitalists rule, OK? Some puzzles about power’, Politics, Philosophy and Economics 1: 155–84. Barry, B. (2003) ‘Capitalists rule, OK? A commentary on Keith Dowding’, Politics, Philosophy and Economics 2: 323–41. Barry, B. (2005) Why Social Justice Matters (Cambridge: Polity). Baumgartner, F. and Jones, B. (1993) Agendas and Instability in American Politics (Chicago: Chicago University Press). Bavetta, S. and Gualla, F. (2003) ‘Autonomy – freedom and deliberation’, Journal of Theoretical Politics 15: 424–43. Bell, D. S., Hargrove, E. C. and Theakston, K. (1999) ‘Skill in context: a comparison of politicians’, Presidential Studies Quarterly 29: 528–48. Bell, S. and Hindmoor, A. (2015) Masters of the Universe But Slave of the Market: Bankers and the Great Financial Meltdown … and How Some Banks Avoided the Carnage (Cambridge, MA: Harvard University Press). Bennett, J. (1983) ‘Positive and negative relevance’, American Philosophical Quarterly 20: 185–94. Bentham, J. (1843) ‘Anarchical fallacies’, in The Works of Jeremy Bentham, ed. J. Bowring (Edinburgh: William Tait). Benton, T. (1988) ‘Review of Peter Morriss, Power’, Sociology 22: 491–3. Binmore, K. (1994) Game Theory and the Social Contract, Vol. 1: Playing Fair (Cambridge, MA: MIT Press). Binmore, K. (2007) Playing for Real: A Text on Game Theory (Oxford: Oxford University Press). Blau, P. (1975) ‘Liberal values and independence’, Review of Economic Studies 42: 395–401. Boin, A., ‘t Hart, P., Stern, E. and Sundelius, B. (2005) The Politics of Crisis Manage­ ment: Public Leadership under Pressure (Cambridge: Cambridge University Press).



References277

Bossert, W., Pattanaik, P. K. and Xu, Y. (1994) ‘Ranking opportunity sets: an axiomatic approach’, Journal of Economic Theory 63: 326–45. Bossert, W., Pattanaik, P. K. and Xu Y. (2003) ‘Similarity of options and the measurement of diversity’, Journal of Theoretical Politics 15: 405–22. Bourdieu, P. (1977) Outline of a Theory of Practice (Cambridge: Cambridge University Press). Bowles, S. and Gintis, H. (1998) ‘Efficient redistribution: new rules for markets, states and communities’, in E. O. Wright (ed.), Recasting Egalitarianism: New Rules for Communities, States and Markets (London: Verso). Bradley, S., Johnes, G. and Millington, J. (1999) ‘School choice, competition, and the efficiency of secondary schools in England’, Working Paper EC 3/99, Centre for Research into the Economics of Education, Lancaster University. Braham, M. (2006) ‘Measuring specific freedom’, Economics and Philosophy 22: 317–33. Braham, M. and Holler, M. (2005) ‘The impossibility of a preference-based power index’, Journal of Theoretical Politics 17 (1): 137–57. Brams, S. J. (1975) Game Theory and Politics (New York: Free Press). Brennan, G. (1998) ‘Democratic trust: a rational-choice perspective’, in V. Braithewaite and M. Levi (eds), Trust and Governance (New York: Russell Sage Foundation). Brouwer, W. B., van Exel, N. J., van der Berg, B., van den Bos, G. A. and Koopmanschap, M. A. (2005) ‘Process utility from providing informal care: the benefits of caring’, Health Policy 74 (1): 85–99. Bryant, C. G. A. and Jary, D. (eds) (1996) Anthony Giddens: Critical Assessments (London: Routledge). Bueno de Mesquita, B., Smith, A., Siverson, R. M. and Morrow, J. D. (2003) The Logic of Political Survival (Cambridge, MA: MIT Press). Butler, D., Adonis, A. and Travers, T. (1994) Failure in British Government: The Politics of the Poll Tax (Oxford: Oxford University Press). Calmy-Rey, M., Nye, J., Vasella, D., Kellenberger, J, Wasescha, L., de Weck, R. and Brunschwig Graf, M. (2007) Soft Gouvernance (Geneva: Fondation pour Genève). Calvert, R. L. (1995) ‘The rational choice theory of institutions: cooperation, coordination and communication’, in J. S. Banks and E. A. Hanushek (eds), Modern Political Economy: Old Topics, New Directions (Cambridge: Cambridge University Press). Canes-Wrone, B. (2006) Who Leads Whom? Presidents, Policy, and the Public (Cambridge, MA: MIT Press). Canes-Wrone, B., Herron, M. C. and Shotts K. W. (2001) ‘Leadership and pandering: a theory of executive policymaking’, American Journal of Political Science 45 (3): 532–50. Carling, A. (1995) ‘The paradox of voting and theory of social evolution’, in K. Dowding and D. King (eds), Preferences, Institutions and Rational Choice (Oxford: Clarendon Press). Carter, I. (1995) ‘The independent value of freedom’, Ethics 105: 819–45. Carter, I. (1999) A Measure of Freedom (Oxford: Oxford University Press).

278 References Carter, I. (2004) ‘Choice, freedom and freedom of choice’, Social Choice and Welfare 22: 61–81. Carter, I. (2008) ‘How are power and unfreedom related?’, in C. Laborde and J. Maynor (eds), Republicanism and Political Theory (Oxford: Blackwell). Carter, I. (2011) ‘Value-freeness and value-neutrality in the analysis of political concepts’, MS. Castells, M. (1980) City, Class and Power (London: Macmillan). Castells, M., Goh, L. and Kwok, R. Y.-W. (1990) The Shek Kip Mei Syndrome: Economic Development and Public Housing in Hong Kong and Singapore (London: Pion). Cebula, R. J. (1979) ‘A survey on the literature on the migration impact of state and local government policies’, Public Finance 34: 69–84. Cebula, R. J. and Koch, J. V. (1989) ‘Welfare policies and migration of the poor in the United States: an empirical note’, Public Choice 61: 171–6. Chalmers, D. J. (2011) ‘Verbal disputes’, Philosophical Review 120 (4): 515–66. Claxton, J., Fry, J. and Portis, B. (1974) ‘A taxonomy of prepurchase information gathering patterns’, Journal of Consumer Research 1: 35–42. Clegg, S. (1989) Frameworks of Power (London: Sage). Cohen, G. A. (1976) Karl Marx’s Theory of History: A Defence (Oxford: Clarendon Press). Cohen, G. A. (1988) History, Labour and Freedom: Themes from Marx (Oxford: Clarendon Press). Cohen, G. A. (1989) ‘On the currency of egalitarian justice’, Ethics 99: 906–44. Cohen, G. A. (1995) Self-Ownership, Freedom and Equality (Cambridge: Cambridge University Press). Cohen, G. A. (1997) ‘Where the action is: on the site of distributive justice’, Philosophy and Public Affairs 26: 3–30. Cohen, G. A. (2000) If You’re an Egalitarian, How Come You’re So Rich? (Cambridge, MA: Harvard University Press). Cohen, G. A. (2008) Rescuing Justice and Equality (Cambridge, MA: Harvard University Press). Connolly, W. E. (1983) The Terms of Political Discourse, 2nd edn (Oxford: Martin Robertson). Coulter, A., Le Maistre, N. and Henderson, L. (2005) Evaluation of London Patient Choice: Patients’ Experience of Choosing Where to Undergo Surgical Treatment (Oxford: Picker Institute). Cowley, P. (1996) ‘How did he do that? The second round of the 1990 Conservative leadership election’, in D. M. Farrell, D. Broughton, D. Denver and J. Fisher (eds), British Election and Parties Yearbook (London: Frank Cass). Cox, G. W. (2006) ‘The organization of democratic legislatures’, in B. R. Weingast and D. A. Wittman (eds), The Oxford Handbook of Political Economy (Oxford: Oxford University Press). Cox, G. W. and McCubbins, M. D. (2005) Setting the Agenda: Responsible Party Government in the House of Representatives (Cambridge: Cambridge University Press).



References279

Craven, J. (1992) Social Choice: A Framework for Collective Decisions and Individual Judgements (Cambridge: Cambridge University Press). Dahl, R. A. (1957) ‘The concept of power’, Behavioral Science 2: 201–15. Dahl, R. A. (1961) Who Governs? Democracy and Power in an American City (New Haven: Yale University Press). Dahl, R. A. (1968) ‘Power’, in D. L. Sils (ed.), International Encyclopaedia of the Social Sciences, Vol. 12 (New York: Free Press). Dalvean, M. (2012a) ‘Predicting cabinet ministers: a psychological approach’, in K. Dowding and C. Lewis (eds), Ministerial Careers and Accountability in the Australian Commonwealth (Canberra: ANU E Press). Dalvean, M. (2012b) ‘The selection of cabinet ministers in the Australian federal parliament’, PhD thesis, Australian National University. Dawes, R. M. (1996) ‘Comment: counterfactual inferences as instances of statistical inferences’, in P. E. Tetlock and A. Belkin (eds), Counterfactual Thought Experiments in World Politics: Logical, Methodological and Psychological Issues (Princeton: Princeton University Press). Dawes, R. M. (2001) Everyday Irrationality: How Pseudo-Scientists, Lunatics, and the Rest of Us Systematically Fail to Think Rationally (Colorado: Westview). Dawson, D., Jacobs, R., Martin, S. and Smith, P. (2005) Evaluation of the London Patient Choice Project: System Wide Impacts (York: Centre for Health Economics). Deb, R. (1994) ‘Waiver, effectivity and rights as game forms’, Economica 61: 167–78. Dennett, D. C. (1984) Elbow Room: The Varieties of Free Will Worth Wanting (Cambridge, MA: Bradford Books). Dennett, D. C. (1987) The Intentional Stance (Cambridge, MA: MIT Press). Dennett, D. C. (1998) ‘Real patterns’, in his Brainchildren: Essays on Designing Minds (Harmondsworth: Penguin). Dennett, D. C. (2003) Freedom Evolves (London: Allen Lane). Dewan, T. and Myatt, D. P. (2007) ‘Leading the party: coordination, direction, and communication’, American Political Science Review 101 (4): 827–45. Dewan, T. and Myatt, D. P. (2008) ‘The qualities of leadership: coordination, direction, and communication’, American Political Science Review 102 (3): 351–68. Dewan, T. and Myatt, D. P. (2012) ‘On the rhetorical strategies of leaders: speaking clearly, standing back, and stepping down’, Journal of Theoretical Politics 24 (4): 431–60. DfES (2005) The White Paper: Delivering School Choice (London: Department for Education and Skills). Diamond, J. (1998) Guns, Germs and Steel: A Short History of Everybody for the Last 13,000 Years (London: Vintage). Domhoff, G. W. (1983) Who Rules America Now? (Englewood Cliffs: Prentice-Hall). Domhoff, G. W. (1986) ‘The growth machine and the power elite: a challenge to pluralists and Marxists alike’, in R. Waste (ed.), Community Power: Directions for Future Research (Beverly Hills: Sage). Donahue, J. D. (1997) ‘Tiebout? Or not Tiebout? The market metaphor and America’s devolution debate’, Journal of Economic Perspectives 11 (4): 73–82.

280 References Dowding, K. (1990) ‘Ableness and ability: Morriss on power and counteractuals’, Department of Government Working Papers 10 (Uxbridge: Brunel University). Dowding, K. (1991) Rational Choice and Political Power (Aldershot: Edward Elgar). Dowding, K. (1992) ‘Choice: its increase and its value’, British Journal of Political Science 22: 301–14. Dowding, K. (1994) ‘The compatibility of behaviouralism, rational choice and “new institutionalism”’, Journal of Theoretical Politics 6: 105–17. Dowding, K. (1995) ‘Model or metaphor? A critical review of the policy network approach’, Political Studies 45: 136–58. Dowding, K. (1996) Power (Buckingham: Open University Press). Dowding, K. (1999) ‘Shaping future luck’, Journal of Conflict Processes and Change 4: 1–11. Dowding, K. (2000) ‘Institutionalist research on the European Union’, European Union Politics 1: 147–68. Dowding, K. (2001a) ‘There must be end to confusion: policy networks, intellectual fatigue, and the need for political science methods courses in British universities’, Political Studies 49 (1): 89–105. Dowding, K. (2001b) ‘Explaining urban regimes’, International Journal of Urban and Regional Research 25: 7–19. Dowding, K. (2002) ‘Revealed preference and external reference’, Rationality and Society 14: 259–84. Dowding, K. (2006a) ‘Can populism be defended? William Riker, Gerry Mackie and the interpretation of democracy’, Government and Opposition 41 (3): 327–46. Dowding, K. (2006b) ‘Three-dimensional power: a discussion of Steven Lukes, Power: A Radical View, 2nd Edition’, Political Studies Review 4: 136–45. Dowding, K. (2006c) ‘Can “capabilities” reconcile freedom and equality?’, Journal of Political Philosophy 14 (3): 323–36. Dowding, K. (2008a) [trans. F. Aguiar] ‘Defensa de la preferencia revelada’, Revista Internacional de Sociología 49: 9–31. Dowding, K. (2008b) ‘Rational choice perspectives’, in S. Clegg, M. Haugaard and C. Lester (eds), Handbook of Power (London: Sage). Dowding, Keith (2008c) ‘What is welfare and how might it be measured?’, in D. Ross and H. Kincaid (eds), Oxford Handbook of the Philosophy of Economics (Oxford: Oxford University Press). Dowding, K. (2010) ‘Luck, equality and responsibility’, Critical Review of International Social and Political Philosophy 13 (1): 71–92. Dowding, K. (ed.) (2011) Encyclopedia of Power (Beverly Hills: Sage). Dowding, K. (2013) ‘Prime-ministerial power: institutional and personal factors’, in P. Strangio, P. ‘t Hart and J. Walter (eds), Understanding Prime-Ministerial Performance: Comparative Perspectives (Oxford: Oxford University Press). Dowding, K. (2016) The Philosophy and Methods of Political Science (London: Palgrave). Dowding, K., Dunleavy, P., King, D. and Margetts, M. (1995) ‘Rational choice and community power structures’, Political Studies 42 (3): 265–77.



References281

Dowding, K. and Hindmoor, A. (1997) ‘The usual suspects: rational choice, socialism and political theory’, New Political Economy 2 (3): 451–63. Dowding, K., John, P. and Biggs, S. (1994) ‘Tiebout: a survey of the empirical literature’, Urban Studies 31: 767–97. Dowding, K. and Mergoupis, T. (2003) ‘Fragmentation, fiscal mobility and efficiency’, Journal of Politics 61 (4): 1190–207. Dowding, K. and van Hees, M. (2003) ‘The construction of rights’, American Political Science Review 97 (2): 281–93. Dowding, K. and van Hees, M. (2007) ‘Counterfactual success and negative freedom’, Economics and Philosophy 23: 141–62. Dowding, K. and van Hees, M. (2008a) ‘In praise of manipulation’, British Journal of Political Science 38 (1): 1–16. Dowding, K. and van Hees, M. (2008b) ‘Freedom, coercion and ability’, in M. Braham and F. Steffen (eds), Power, Freedom and Voting: Essays in Honour of Manfred J. Holler (Berlin: Springer). Dowding, K. and van Hees, M. (2009) ‘Freedom of choice’, in P. Anand, P. K. Pattanaik and C. Puppe (eds), Oxford Handbook of Rational and Social Choice (Oxford: Oxford University Press). Dowding, K. and Verbeek, B. (2000) ‘The impossibility of Paretian welfarism’, paper presented at the European Public Choice Conference, Paris, 18–21 April. Drèze, J. and Sen, A. (1999) The Amartya Sen and Jean Drèze Omnibus (Oxford: Oxford University Press). Dunleavy, P. (1991) Bureaucracy, Democracy and Public Choice (Hemel Hempstead: Harvester Wheatsheaf). Dunleavy, P. and Margetts, H. (1995) ‘The rational basis for belief in the democratic myth’, in K. Dowding and D. King (eds), Preferences, Institutions and Rational Choice (Oxford: Clarendon Press). Dworkin, R. (1981) ‘What is equality? Part 2: equality of resources’, Philosophy and Public Affairs 10: 283–345. Dworkin, G. (1982) ‘Is more choice better than less?’, in P. French, T. Uehling and H. Wettstein (eds), Midwest Studies in Philosophy, Vol. 7 (Minneapolis: University of Minnesota Press). Dworkin, R. (2000) Sovereign Virtue: The Theory and Practice of Equality (Cambridge, MA: Harvard University Press). Education and Employment Committee (1997) ‘The operation of the nursery education voucher scheme’, House of Commons Papers, 1996–7 session, 12 Mar. Elkin, Stephen L. (1987) City and Regime in the American Republic (Chicago: Chicago University Press). Elster, J. (1978) Logic and Society (London: Wiley). Estlund, D. (1998) ‘Liberalism, equality and fraternity in Cohen’s critique of Rawls’, Journal of Political Philosophy 6: 99–112. Farrell, C. (2004) Patient and Public Involvement in Health: The Evidence for Policy Implementation (London: Department of Health). Farrington-Douglas. J. and Allen, J. (2005) Equitable Choice for Health (London: Institute of Public Policy Research).

282 References Feinberg, J. (1973) Social Philosophy (Englewood Cliffs: Prentice-Hall). Feinberg, J. (1980) Rights, Justice, and the Bounds of Liberty (Princeton: Princeton University Press). Felsenthal, D. S. and Machover, M. (1998) The Measurement of Voting Power: Theory and Practice, Problems and Paradoxes (Aldershot: Edward Elgar). Fine, B. J. (1975) ‘Individual liberalism in a Paretian society’, Journal of Political Economy 83: 1277–82. Fiorina, M. P. and Shepsle, K. A. (1989) ‘Formal theories of leadership: agents, agenda setters and entrepreneurs’, in B. D. Jones (ed.), Leadership and Politics (Lawrence: Kansas University Press). Flatney, J., Connolly, H., Higgins, V., Williams, J., Coldron, J., Stephenson, K., Logie, A. and Smith, N. (2001) Parents’ Experiences of the Process of Choosing a Secondary School (London: Department for Education and Skills). Fleurbaey, M. and Gaertner, W. (1996) ‘Admissibility and feasibility in game forms’, Analyse und Kritik 16: 54–66. Fleurbaey, M. and van Hees, M. (2000) ‘On rights in game forms’, Synthese 123: 295–326. Flynn, R. (1999) New Managerialism in the Public Sector (Buckingham: Open University Press). Fogel, R. W. (1964) Railroads and American Economic Growth (Baltimore: Johns Hopkins University Press). Foucault, M. (1980) Power/Knowledge: Selected Interviews and Other Writings 1972–1977 (Brighton: Harvester). Foucault, M. (1982) ‘The subject and power’, Critical Inquiry 8 (4): 777–95. Foucault, M. (2008) The Birth of Biopolitics: Lectures at the Collège de France, 1978– 1979 (Basingstoke: Palgrave Macmillan). Friedman, M. and Friedman, R. (1990) Free to Choose (London: Pan). Friedrich, C. (1941) Constitutional Government and Democracy (New York: W. W. Norton). Gaertner, W. (1993) ‘Rights and game forms: types of preference orderings and Pareto inefficiency’, in W. E. Diewart, K. Spremann and F. Stehling (eds), Mathematical Modeling in Economics: Essays in Honor of Wolfgang Eichhorn (Berlin: Springer). Gaertner, W., Pattanaik, P. and Suzumura, K. (1992) ‘Individual rights revisited’, Economica 59: 161–77. Gallie, W. B. (1956) ‘Essentially contested concepts’, Proceedings of the Aristotelian Society 56 (2): 167–98. Gambetta, D. (1993) The Sicilian Mafia: The Business of Private Protection (Cambridge, MA: Harvard University Press). Gärdenfors, P. (1981) ‘Rights, games and social choice’, Noûs 15: 341–56. Garrett, G. and Tsebelis, G. (1999) ‘Why resist the temptation to apply power indices to the EU?’, Journal of Theoretical Politics 11 (3): 291–308. Gaventa, J. (1980) Power and Powerlessness: Quiescence and Rebellion in an Appalachian Valley (Oxford: Clarendon Press).



References283

Gekker, R. (1985) ‘On the inconsistency of a weak version of the libertarian p ­ rinciple’, Economic Letters 18: 21–5. Gettier, E. (1963) ‘Is justified true belief knowledge?’, Analysis 23: 121–3. Gibbard, A. (1973) ‘Manipulation of voting schemes: a general result’, Econometrica 41: 587–601. Gibbard, A. (1974) ‘A Pareto-consistent libertarian claim’, Journal of Economic Theory 7: 388–410. Gibbard, A. (1982) ‘Rights and the theory of social choice’, in L. J. Cohen, J. Los, H. Pfeiffer and K.-P. Podewski (eds), Logic, Methodology and Philosophy of Science, Vol. 6 (Amsterdam: North Holland). Giddens, A. (1979) The Central Problems of Social Theory (London: Macmillan). Giddens, A. (1984) The Constitution of Society (Cambridge: Polity). Goodin, R. E. (1980) Manipulatory Politics (New Haven: Yale University Press). Goodin, R. E. (1986) ‘Laundering preferences’, in J. Elster and A. Hylland (eds),  Foundations of Social Choice Theory (Cambridge: Cambridge University Press). Goodman, N. (1955) Fact, Fiction and Forecast (Cambridge, MA: Harvard University Press). Gramsci, A. (1971) Selections from the Prison Notebooks, ed. and trans. Q. Hoare and G. Nowell Smith (London: Lawrence and Wishart). Gravel, N. (1994) ‘Can a ranking of opportunity sets attach an intrinsic value to freedom of choice?’, American Economic Review 84 (2): 454–8. Grint, K. (2000) The Arts of Leadership (Oxford: Oxford University Press). Grofman, B., Handley, L. and Niemi, R. G. (1992) Minority Representation and the Quest for Voting Equality (Cambridge: Cambridge University Press). Gurr, T. R. and King, D. S. (1987) The State and the City (Chicago: University of Chicago Press). Hammond, P. (1995) ‘Social choice of individual and group rights’, in W. A. Barnett, H. Moulin, M. Salles and N. Schofield (eds), Social Choice, Welfare and Ethics (Cambridge: Cambridge University Press). Hammond, P. (1996) ‘Game forms versus social choice rules as models of rights’, in K. J. Arrow, A. Sen and K. Suzumura (eds), Social Choice Re-examined (Basingstoke: Macmillan). Hardin, R. (1982) Collective Action (Baltimore: Resources for the Future). Hardin, R. (1988) Morality within the Limits of Reason (Chicago: University of Chicago Press). Hardy, B. and Wistow, G. (1995) ‘Securing quality through contracts? The development of quasi-markets for social care in Britain’, Australian Journal of Public Administration 98 (2): 25–35. Hargrove, E. C. (1989) ‘Two conceptions of institutional leadership’, in B. D. Jones (ed.), Leadership and Politics (Lawrence: University of Kansas Press). Harsanyi, John (1976a) ‘Measurement of Social Power, Opportunity Costs, and the Theory of Two-Person Bargaining Games’, in his Essays on Ethics, Social Behaviour and Scientific Explanation (Dordrecht, D. Reidel). Harsanyi, John (1976b) ‘Measurement of Social Power in n-Person Reciprocal

284 References Power Situations’, in his Essays on Ethics, Social Behaviour and Scientific Explanation (Dordrecht, D. Reidel). Hart, H. L. A. (1973) ‘Bentham on legal rights’, in his Essays on Bentham (London: Macmillan). Haugaard, M. (1997) The Constitution of Power (Manchester: Manchester University Press). Haugaard, M. (2010) ‘Power: a “family resemblance” concept’, European Journal of Cultural Studies 13: 419–38. Hayek, F. A. (1960) The Constitution of Liberty (London: Routledge & Kegan Paul). Herne, K. and Nurmi, H. (1993) ‘The distribution of a priori voting power in the EC Council of Ministers and the European Parliament’, Scandinavian Political Studies 16: 269–84. Hindess, B. (1989) Political Choice and Social Structure (Aldershot: Edward Elgar). Hindmoor, A. and McGeechan, J. (2013) ‘Luck, systematic luck and business power: lucky all the way down or trying hard to get what it wants without trying?’, Political Studies 61 (4): 834–49. Hintikka, J. (1969) ‘Existential presuppositions and their elimination’, in his Models for Modalities: Selected Essays (Dordrecht: D. Reidel). Hirschman, A. O. (1970) Exit, Voice and Loyalty (Cambridge, MA: Harvard University Press). Hobbes, T. (1651/1982) Leviathan, ed. C. B. MacPherson (Harmondsworth: Penguin). Hogan, R. and Kaiser, R. B. (2005) ‘What we know about leadership’, Review of General Psychology 9 (2): 169–80. Hohfeld, W. N. (1919) Fundamental Legal Conceptions as Applied in Judicial Reasoning (New Haven: Yale University Press). Holtz, A. (1997) ‘Study: young brains especially susceptible to alcohol’, CNN Interactive Health: www.cnn.com/HEALTH/9701/09/nfm/youth.alcohol/ index.html, accessed 4 Mar. 2003. Hood, C. (2002) ‘The risk game and the blame game’, Government and Opposition 31: 15–37. Hosli, M. O. (1996) ‘Coalitions and power: effects of qualified majority voting in the Council of the European Union’, Journal of Common Market Studies 34: 255–73. Hosli, M. O. (1997) ‘Voting strength in the European Parliament: the influence of national and partisan actors’, European Journal of Political Research 31: 351–66. House of Commons Public Administration Committee (2005) Choice, Voice and Public Services, HC 49 (1) (London: HMSO). Hunter, F. (1953) Community Power Structure: A Study of Decision Makers (Chapel Hill: University of North Carolina Press). Hunter, S. T., Bedell-Avers, K. E. and Mumford, M. D. (2007), ‘The typical leadership study: assumptions, implications and potential remedies’, Leadership Quarterly 18 (3): 435–46. Hurley, S. L. (2003) Justice, Luck and Knowledge (Cambridge, MA: Harvard University Press).



References285

Institute for Political and Economic Governance (2006) Choice in Public Services (Manchester: University of Manchester Press). Isaac, J. (1987) Power and Marxist Theory (Ithaca: Cornell University Press). Isaac, J. (2006) ‘Thoughts on the concept of power’, paper presented to American Political Science Association conference, Philadelphia. Iyengar, S. and Lepper, M. (2000) ‘When is choice demotivating? Can one desire too much of a good thing?’, Journal of Personality and Social Psychology 79: 995–1006. Johnston, R. J. (1995) ‘The conflict over qualified majority voting in the European Council of Ministers: an analysis of the UK negotiating stance using power indices’, British Journal of Political Science 25: 245–54. Jones, B., Baumgartner, F. and True, J. (1998) ‘Policy punctuations: US budget authority, 1947–1995’, Journal of Politics 60: 1–33. Jones, B., Sulkin, T. and Larsen, H. (2003) ‘Policy punctuations in American political institutions’, American Political Science Review 97: 151–69. Jones, P., and Sugden, R. (1982) ‘Evaluating choice’, International Review of Law and Economics 2: 47–65. Jordan, G. and Richardson, J. (1987) British Politics and the Policy Process (London: Unwin Hyman). Kenny, A. (1975) Will, Freedom and Power (Oxford: Blackwell). Keohane, N. O. (2010) Thinking about Leadership (Princeton: Princeton University Press). King, D. S. (1987) The New Right: Politics, Markets and Citizenship (Basingstoke: Macmillan). Kleinke, C. L. (1986) ‘Gaze and eye contact: a research review’, Psychological Bulletin 100 (1): 78–100. Knapp, M., Hardy, B. and Forder, J. (2001) ‘Commissioning for quality: ten years of social care markets in England’, Journal of Social Policy 30: 283–306. Kramer, M. (1998) ‘Rights without trimmings’, in M. Kramer, N. E. Simmonds and H. Steiner, A Debate over Rights (Oxford: Oxford University Press). Kramer, M. (2001) ‘Getting rights right’, in M. Kramer (ed.), Rights, Wrongs and Responsibilities (London: Palgrave). Kramer, M. (2002) ‘Why freedoms do not exist in degrees’, Political Studies 50 (2): 230–43. Kramer, M. (2003) The Quality of Freedom (Oxford: Oxford University Press). Kramer, M. (2008) ‘Liberty and domination’, in C. Laborde and J. Maynor (eds), Republicanism and Political Theory (Oxford: Blackwell). Krehbiel, K. (1993) ‘Where’s the party?’, Legislative Studies Quarterly 23: 235–66. Kreps, D. M. (1979) ‘A representation theorem for “preference for flexibility”’, Econometrica 47: 565–77. Kreps, D. M. (1990a) Game Theory and Economic Modelling (Oxford: Clarendon Press). Kreps, D. M. (1990b) A Course in Microeconomic Theory (New York: Harvester Wheatsheaf). Kripke, S. (1980) Naming and Necessity (Oxford: Blackwell). Kukathas, C. (2016) ‘Jusiticitus’, in M. Knoll (ed.), Pluralism and Conflict: Distributive Justice beyond Rawls and Consensus (Berlin: De Gruyter).

286 References Kuran, T. (1995) Private Truths, Public Lies: The Social Consequences of Preference Falsification (Cambridge, MA: Harvard University Press). Lane, J.-E. and Mæland, R. (1995) ‘Voting power under the EU constitution’, Journal of Theoretical Politics 7: 223–30. Lane, J.-E. and Mæland, R. (1996) ‘Constitutional power in the EU’, NEMEU Working Papers 96,2 Enschede: Twente University. Lane, J.-E., Mæland, R. and Berg, S. (1995) ‘The EU parliament, seats, states and political parties’, Journal of Theoretical Politics 7: 395–400. Laslier, J.-F. (ed.) (1998) Freedom in Economics (London: Routledge). Le Grand, J. (2003) Motivation, Agency and Public Policy (Oxford: Oxford University Press). Le Grand, J. (2007) The Other Invisible Hand: Delivering Public Services through Choice and Competition (Princeton: Princeton University Press). Le Grand, J. and Bartlett, W. (eds) (1993) Quasi-Markets and Social Policy (Basingstoke: Macmillan). Le Maistre, N., Reeves, R. and Coulter, A. (2004) Patients’ Experience of CHD Choice (Oxford: Picker Institute). Levi, M. (2006) ‘Why we need a new theory of government’, Perspectives on Politics 4 (1): 5–20. Lindblom, C. E. (1977) Politics and Markets (New York: Basic Books). Little, R. (2007) The Balance of Power in International Relations: Metaphors, Myths and Models (New York: Cambridge University Press). Logan, J. and Molotch, H. (1984) ‘Tensions in the growth machine: overcoming resistance to value-free development’, Social Problems 31: 483–99. Logan, J. and Molotch, H. (1987) Urban Fortunes: The Political Economy of Place (Berkeley: University of California Press). Luhmann, N. (1979) Trust and Power (New York: Free Press). Lukes, S. (2005) Power: A Radical View, 2nd edn (Basingstoke: Palgrave Macmillan). Lukes, S. and Haglund, L. (2005) ‘Power and luck’, Archives Européennes de Sociologie 46: 45–66. Lyons, D. (1969) ‘Rights, claimants and beneficiaries’, American Philosophical Quarterly 6: 173–85. MacCallum, G. C. (1972) ‘Negative and positive freedom’, in W. G. Runciman and Q. Skinner (eds), Philosophy, Politics and Society, Fourth Series (Oxford: Blackwell). McClelland, D. C. (1973) ‘Testing for competence rather than intelligence’, American Psychologist 28 (1): 1–14. MacCormick, D. N. (1977) ‘Rights in legislation’, in P. M. S. Hacker and J. Raz (eds), Law, Morality and Society (Oxford: Oxford University Press). Machover, M. and Felsenthal, D. S. (2011a) ‘I-power’, in K. Dowding (ed.), Encyclopedia of Power (Beverly Hills: Sage). Machover, M. and Felsenthal, D. S. (2011b) ‘P-power’, in K. Dowding (ed.), Encyclopedia of Power (Beverly Hills: Sage). McLean, I. (2001) Rational Choice and British Politics: An Analysis of Rhetoric and Manipulation from Peel to Blair (Oxford: Oxford University Press). Major, J. (1999) The Autobiography (London: HarperCollins).



References287

Malkin, J. and Wildavsky, A. (1991) ‘Why the traditional distinction between public and private goods should be abandoned’, Journal of Theoretical Politics 3: 355–78. Mann, M. (1986) The Sources of Social Power, Vol. 1: A History of Power from the Beginning to AD 1760 (Cambridge: Cambridge University Press). Mann, M. (1993) The Sources of Social Power, Vol. 2: The Rise of Classes and Nation States 1760–1914 (Cambridge: Cambridge University Press). Mansbridge, J., with J. Bohman, S. Chambers, D. Estlund, A. Follesdal, A. Fugn, C. Lafont, B. Manin and J. L. Marti (2010) ‘The place of self-interest and the role of power in deliberative democracy’, Journal of Political Philosophy 18: 64–100. Martin, C. B. (1994) ‘Dispositions and conditionals’, Philosophical Quarterly 44: 1–8. Martin, J. L. (2001) ‘The authoritarian personality, fifty years on: what questions are there for political psychology?’, Political Psychology 22: 1–26. Mednick, S. A., Gabrielli, W. F. and Hutchings, B. (1984) ‘Genetic influences in criminal convictions: evidence from an adoption cohort’, Science 224: 891–4. Messick, D. M. and Kramer, R. M. (eds), (2004) The Psychology of Leadership: New Perspectives and Research (Mahwah: Lawrence Erlbaum). Milgrom, P. and Roberts, J. (1992) Economics, Organization and Management (Englewood Cliffs: Prentice-Hall). Mill, J. S. (1848) On Liberty (London: John W. Parker and Sons). Miller, D. (1990) Market, State and Community (Oxford: Oxford University Press). Modernising Medical Careers (2005) Curriculum for the Foundation Years in Postgraduate Education and Training (London: Modernising Medical Careers). Moe, T. (2005) ‘Power and political institutions’, Perspectives on Politics 3: 215–34. Molotch, H. (1976) ‘The city as a growth machine’, American Journal of Sociology 82: 309–30. Molotch, H. (1979) ‘Capital and neighborhood in the United States’, Urban Affairs Quarterly 14: 289–312. Morgenthau, H. (1949) Politics among Nations: The Struggle for Power and Peace (Ithaca: Cornell University Press). Morrell, K. and Hartley, J. (2006) ‘A model of political leadership’, Human Relations, 59 (4), 483–504. Morriss, P. (1987) Power: A Philosophical Analysis (Manchester: Manchester University Press). Morriss, P. (2002) Power: A Philosophical Analysis, 2nd edn (Manchester: Manchester University Press). Mueller, D. C. (1989) Public Choice II: A Revised Edition of Public Choice (New York: Cambridge University Press). Mumford, M. D., Antes, A. L., Caughron, J. J. and Friedrich, T. L. (2008) ‘Charismatic, ideological and pragmatic leadership: multi-level influences on emergence and performance’, Leadership Quarterly 19 (1): 144–60. Narveson, J. F. (1976) ‘A puzzle about economic justice in Rawls’ theory’, Social Theory and Practice 4: 1–27. Narveson, J. F. (1978) ‘Rawls on equal distribution of wealth’, Philosophia 7: 281–92. Newman, M. and Vidler, E. (2006) ‘Discriminating customers, responsible patients,

288 References empowered users: consumerism and the modernisation of health care’, Journal of Social Policy 35: 193–209. NIAAA (1992) ‘The genetics of alcoholism’, National Institute on Alcohol Abuse and Alcoholism, no. 18, PH 357: www.niaaa.nih.gov/publications/aa09-text. htm, accessed 4 Mar. 2003. NIAAA (2000a) ‘Why do some people drink too much? Role of genetic and psychological influences’, National Institute on Alcohol Abuse and Alcoholism, 10th Special Report to the United States Congress on Alcohol and Health, 24(1): www.niaaa.nih.gov, accessed 4 Mar. 2003. NIAAA (2000b) ‘From genes to geography: the cutting edge of alcohol research’, National Institute on Alcohol Abuse and Alcoholism, no. 48: www.niaaa.nih. gov/publications/aa09-text.htm, accessed 4 Mar. 2003. Nozick, R. (1974) Anarchy, State and Utopia (Oxford: Basil Blackwell). Nussbaum, M. (2000) Women and Human Development: The Capability Approach (Cambridge: Cambridge University Press). Nye, J. (1990) Bound to Lead: The Changing Nature of American Power (New York: Basic Books). Nye, J. (2004) Soft Power: The Means to Success in World Politics (Washington, DC: Public Affairs Institute). ODPM (2004) Applicants’ Perspectives on Choice-Based Lettings (London: Office of the Deputy Prime Minister). Offe, C. (1975) ‘Two logics of collective action’, in his Disorganized Capitalism (Cambridge: Polity). Olsen, J. A. and Smith, R. D. (2001) ‘Theory versus practice: a review of “willingness-to-pay” in health and health care’, Health Economics 10 (1): 39–52. Olson, M. (1971) The Logic of Collective Action: Public Goods and the Theory of Groups, 2nd edn (Cambridge, MA: Harvard University Press). Oppenheim, F. (1981) Political Concepts: A Reconstruction (Oxford: Basil Blackwell). Ordeshook, P. (1992) A Politcal Theory Primer (London: Routledge). O’Shaughnessy, J. and Leslie, C. (2005) More Good School Places (London: Policy Exchange). Ostrom, E. (1991) ‘Rational choice theory and institutional analysis: toward complementarity’, American Political Science Review 85: 237–43. Ostrom, V., Tiebout, C. and Warren, R. (1961) ‘The organization of government in metropolitan areas: a theoretical inquiry’, American Political Science Review 55: 831–42. Otnes, P. (1988) The Sociology of Consumption (Atlantic Highlands: Humanities Press). Pansardi, P. (2011) ‘Power to and power over’, in K. Dowding (ed.), Encyclopedia of Power (Beverly Hills: Sage). Parikh, R. (2002) ‘Social software’, Synthese 132: 187–211. Parsons, T. (1952) The Social System (London, Tavistock). Parsons, T. (1963) ‘On the concept of power’, Proceedings of the American Philosophical Society 107: 232–62. Pattanaik, P. K. (1996) ‘On modelling individual rights: some conceptual issues’,



References289

in K. J. Arrow, A. Sen and K. Suzumura (eds), Social Choice Re-examined (Basingstoke: Macmillan). Pattanaik, P. K., and Suzumura, K. (1996) ‘Individual rights and social evaluation: a conceptual framework’, Oxford Economic Papers 48 (2): 194–212. Pattanaik, P. K. and Xu, Y. (1990) ‘On ranking opportunity sets in terms of freedom of choice’, Recherches Economiques de Louvain 56: 383–90. Pattanaik, P. K. and Xu, Y. (1998) ‘On preference and freedom’, Theory and Decision 44: 173–98. Peacock, A. T. and Rowley, C. (1972) ‘Pareto optimality and the political economy of liberalism’, Journal of Political Economy 80 (3): 476–90. Peele, G. (2005) ‘Leadership and politics: a case for a closer relationship’, Leadership 1 (2): 187–204. Perri 6 (1998) ‘Ownership and new politics of the public interest services’, Political Quarterly 69: 404–14. Perri 6 (2003) ‘Giving consumers of British public services more choice: what can be learned from recent history?’, Journal of Social Policy 32: 239–70. Peterson, P. E. (1981) City Limits (Chicago: University of Chicago Press). Pettit, P. (1993a) The Common Mind (New York: Oxford University Press). Pettit, P. (1993b) ‘Negative liberty, liberal and republican’. European Journal of Philosophy 1: 15–38. Pettit, P. (1996) ‘Freedom as anti-power’, Ethics 106: 576–604. Pettit, P. (1997) Republicanism: A Theory of Freedom and Government (Oxford: Oxford University Press). Pettit, P. (2001) A Theory of Freedom: From the Psychology to the Politics of Agency (Cambridge: Polity). Pettit, P. (2002) ‘Keeping republicanism simple: on a difference with Quentin Skinner’, Political Theory 30: 339–56. Pettit, P. (2008) ‘Republican freedom: three axioms, four theorems’, in C. Laborde and J. Maynor (eds), Republicanism and Political Theory (Oxford: Blackwell). Phillips, A. (2006) ‘“Really equal”: opportunities and autonomy’, Journal of Political Philosophy 14: 18–32. Philp, M. E. (1992) ‘Michel Foucault’, in W. E. Connolly et al. (eds), Blackwell Encyclopedia of Political Thought (Oxford: Blackwell). Pogge, T. (2000) ‘On the site of distributive justice: reflections on Cohen and Murphy’, Philosophy and Public Affairs 29: 137–69. Polsby, N. (1980) Community Power and Political Theory, 2nd edn (New Haven: Yale University Press). Porter, L. W. and McLoughlin, G. B. (2006) ‘Leadership and the organizational context: like the weather?’, Leadership Quarterly 17 (4): 559–76. Powell, E. (1977) Joseph Chamberlain (London: Thames and Hudson). Primus, R. A. (1999). The American Language of Rights (Cambridge: Cambridge University Press). Pritchard, D. (2005) Epistemic Luck (Oxford: Oxford University Press). Propper, C., Wilson, D. and Burgess, S. (2006) ‘Extending choice in English health

290 References care: the implications of the economic evidence’, Journal of Social Policy 35: 537–57. Przeworski, A. (1985) Capitalism and Social Democracy (Cambridge: Cambridge University Press). Quine, W. V. O. (1939) ‘Designation and existence’, Journal of Philosophy 36: 701–9. Quine, W. V. O. (1953) From a Logical Point of View (New York: Harper and Row). Quine, W. V. O. (1970) Philosophy of Logic Englewood Cliffs: Prentice-Hall). Rasmusen, E. (1989) Games and Information: An Introduction (Oxford: Blackwell). Ratcliffe, J. and Buxton, M.. (1999) ‘Patients’ preferences regarding the process and outcomes of life-saving technology’, International Journal of Technology Assessment in Health Care 15: 340–51. Raub, W. and Weesie, J. (1990) ‘Reputation and efficiency in social interactions: an example of network effects’, American Journal of Sociology 96: 626–54. Rawls, J. (1971) A Theory of Justice (Oxford: Oxford University Press). Rawls, J. (1982) ‘The basic liberties and their priority’, in S. M. McMurrin (ed.), The Tanner Lectures on Human Values III (Salt Lake City: University of Utah Press). Rawls, J. (1999) A Theory of Justice, revised edn (Cambridge, MA: Belknap Press). Raz, J. (1984) ‘On the nature of rights’, Mind 93: 194–214. Raz, J. (1986) The Morality of Freedom (Oxford: Clarendon Press). Reeve, A. (1990) ‘Individual choice and the retreat from Utilitarianism’, in L. Allison (ed.), The Utilitarian Response: The Contemporary Viability of Utilitarian Political Philosophy (London: Sage). Rescher, N. (1978) ‘The equivocality of existence’, in N. Rescher (ed.), Studies in Ontology: American Philosophical Quarterly Monograph Series (Oxford: Basil Blackwell). Riker, R. (1969) ‘Some ambiguities in the notion of power’, in R. Bell, D. V. Edwards and R. J. Wagner (eds), Political Power: A Reader (London: Collier-Macmillan). Riker, W. H. (1982) Liberalism against Populism: A Confrontation between the Theory of Democracy and the Theory of Social Choice (San Francisco: W. H. Freeman and Co.). Riker, W. H. (1986) The Art of Political Manipulation (New Haven: Yale University Press). Riker, W. H (1996) The Strategy of Rhetoric (New Haven: Yale University Press). Riley, J. (1989) ‘Rights to liberty in purely private matters: part I’, Economics and Philosophy 5: 121–66. Riley, J. (1990) ‘Rights to liberty in purely private matters: part II’, Economics and Philosophy 6: 27–64. Roberts, R. (1985) ‘Reputation in games and markets’, in A. Roth (ed.), GameTheoretic Models of Bargaining (Cambridge: Cambridge University Press). Roemer, J. E. (1996) Theories of Distributive Justice (Cambridge, MA: Harvard University Press). Roemer, J. E. (1998) Equality of Opportunity (Cambridge, MA: Harvard University Press). Rose, N. (2007) The Politics of Life Itself: Biomedicine, Power and Subjectivity in the Twenty-First Century (Princeton: Princeton University Press).



References291

Rosenbaum, E. F. (2000) ‘On measuring freedom’, Journal of Theoretical Politics 12 (2): 205–27. Ross, D. (2005) Economic Theory and Cognitive Science: Microexplanation (Cambridge, MA: MIT Press). Ross, D. (2014) Philosophy of Economics (Palgrave: Macmillan). Salkeld, G. (1998) ‘What are the benefits of preventive health care?’, Health Care Analysis 6 (2): 106–12. Sanders, D. (1993a) ‘Economic influences on the vote: modeling electoral decisions’, in I. Budge and D. McKay (eds), Developing Democracy: Research in Honour of J. P. F. Blondel (London: Sage). Sanders, D. (1993b) ‘Forecasting the 1992 British general election outcome: the performance of an economic model’, in D. Denver, P. Norris, D. Broughton and C. Rallings (eds), British Elections and Parties Yearbook 1993 (Hemel Hempstead: Harvester Wheatsheaf). Sanders, D. (1996) ‘Economic performance, management competence and the outcome of the next general election’, Political Studies 44: 203–31. Satterthwaite, M. (1975) ‘Strategy proofness and Arrow’s conditions: existence and correspondence theorems for voting procedures and social welfare functions’, Journal of Economic Theory 10: 187–217. Savas, E. S. (1987) Privatization: The Key to Better Government (Chatham, NJ: Chatham House Publishers). Scanlon, T. M. (1988) ‘The significance of choice’, in S. M. McMurrin (ed.), The Tanner Lectures on Human Values VIII (Salt Lake City: University of Utah Press). Schattschneider, E. E. (1960) The Semi-Sovereign People: A Realist’s View of Democracy in America (New York: Holt, Rinehart and Winston). Scheffler, S. (2003) ‘What is egalitarianism?’, Philosophy and Public Affairs 31: 5–39. Schelling, T. (1966) Arms and Influence (New Haven: Yale University Press). Schick, F. (1982) ‘Under which descriptions’?, in A. Sen and B. Williams (eds), Utilitarianism and Beyond (Cambridge: Cambridge University Press). Schneider, M. (1989) The Competitive City (Pittsburgh: Pittsburgh University Press). Schoen, C., Osborn, R., Huynh, P., Doty, M., Davis, K., Zapert, K. and Peugh, J. (2004) ‘Primary care and health system performance: adults’ experiences in five countries’, Health Affairs 10: 487–503. Schofield, N. (2006) Architects of Political Change: Constitutional Quandaries and Social Choice Theory (Cambridge: Cambridge University Press). Schotter, A. (1981) The Economic Theory of Social Institutions (Cambridge: Cambridge University Press). Schwartz, B. (2004) The Paradox of Choice: Why More Is Less (New York: HarperCollins). Seldon, A. (2005) Blair (London: Free Press). Selton, R. (1976) ‘The chain store paradox’, Theory and Decision 9: 127–59. Sen, A. (1970) Collective Choice and Social Welfare (Edinburgh: Oliver and Boyd). Sen, A. (1982a) ‘The impossibility of a Paretian liberal’, in his Choice, Welfare and Measurement (Oxford: Basil Blackwell).

292 References Sen, A. (1982b) ‘Liberty as control: an appraisal’, Midwest Studies in Philosophy 7: 207–21. Sen, A. (1982c) ‘Equality of what?’, in his Choice, Welfare and Measurement (Oxford: Basil Blackwell). Sen, A. (1982d) ‘Behaviour and the concept of preference’, in his Choice, Welfare and Measurement (Oxford: Basil Blackwell). Sen, A. (1982e) ‘Liberty, unanimity and rights’, in his Choice, Welfare and Measurement (Oxford: Basil Blackwell). Sen, A. (1984) Resources, Values and Development (Oxford: Blackwell). Sen, A. (1986) ‘Foundations of social choice theory: an epilogue,’ in J. Elster and A. Hylland (eds), Foundations of Social Choice Theory (Cambridge: Cambridge University Press). Sen, A. (1987a) Commodities and Capabilities (New Delhi: Oxford University Press). Sen, A. (1987b) ‘The standard of living: lecture II: lives and capabilities’, in his The Standard of Living (Cambridge: Cambridge University Press). Sen, A. (1988) ‘Freedom of choice: concept and content’, European Economic Review 32: 269–94. Sen, A. (1989) ‘Food and freedom’, World Development 16: 769–81. Sen, A. (1990a) ‘Welfare, freedom and social choice’, Recherches Economiques de Louvain 56, 451–85. Sen, A. (1990b) ‘Individual freedom as social commitment’, New York Review of Books, 14 June: 49–54. Sen, A. (1990c) ‘Justice: means versus freedoms’, Philosophy and Public Affairs 19 (2): 111–21. Sen, A. (1991a) Inequality Re-examined (Oxford: Oxford University Press). Sen, A. (1991b) ‘Welfare, preference and freedom’, Journal of Econometrics 50: 15–29. Sen, A. (1992) ‘Minimal liberty’, Economica 59: 139–59. Sen, A. (1993) ‘Capability and well-being’, in M. C. Nussbaum and A. Sen, The Quality of Life (Oxford: Oxford University Press). Sen, A. (1995) ‘Rationality and social choice’, American Economic Review 85 (1) 1–24. Sen, A. (1996) ‘Welfare economics and two approaches to rights’, in J. C. Pardo and F. Schneider (eds), Current Issues in Public Choice (Cheltenham: Edward Elgar). Sen, A. (1997) ‘Individual preference as the basis of social choice’, in K. J. Arrow, A. Sen and K. Suzumura (eds), Social Choice Re-examined (Houndmills: Macmillan). Sen, A. (1999) Development as Freedom (Oxford: Oxford University Press). Sened, I. (1997) The Political Institution of Private Property (Cambridge: Cambridge University Press). Shapley, L. S. and Shubik, M. (1969) ‘Measurement of power in political systems’, in R. Bell, D. V. Edwards and R. J. Wagner (eds), Political Power: A Reader (London: Collier-Macmillan). Shepsle, K. A. (1979) ‘Institutional arrangements and equilibrium in multidimensional voting models’, American Journal of Political Science 23: 27–59.



References293

Shepsle, K. A. and B. R. Weingast (1981) ‘Structure-induced equilibrium and legislative choice’, Public Choice 37: 503–19. Simon, H. A. (1953) ‘Notes on the observation and measurement of political power’, Journal of Politics 15: 500–16. Skinner, Q. (1990) ‘The republican ideal of political liberty’, in G. Bock, Q. Skinner and M. Viroli (eds), Machiavelli and Republicanism (Cambridge: Cambridge University Press). Skinner, Q. (1992) ‘On justice, the common good and the priority of liberty’, in C. Mouffe (ed.), Dimensions of Radical Democracy (London: Verso). Skinner, Q. (1998) Liberty before Liberalism (Cambridge: Cambridge University Press). Skinner, Q. (2008) ‘Freedom as the absence of arbitrary power’, in C. Laborde and J. Maynor (eds), Republicanism and Political Theory (Oxford: Blackwell). Smith, M. (1990) The Politics of Agricultural Support in Britain: The Development of the Agricultural Policy Community (Aldershot: Dartmouth). Steiner, H. (1974) ‘Individual liberty’, Proceedings of the Aristotelian Society 75: 33–50. Steiner, H. (1994) An Essay on Rights (Oxford: Blackwell). Steiner, H. (1998) ‘Working rights’, in M. H. Kramer, N. E. Simmonds and H. Steiner, A Debate over Rights (Oxford: Oxford University Press). Steiner, H. (2001) ‘Choice and circumstance’, in M. Kramer (ed.), Rights, Wrongs and Responsibilities (Basingstoke: Palgrave). Steiner, H. (2003) ‘Equality, incommensurability and rights’, in L. Meyer, S. Paulson and T. Pogge (eds), Rights, Culture and the Law: Essays after Joseph Raz (Oxford: Oxford University Press). Stevens, D. N. and Foster, J. E. (1978) ‘The possibility of democratic pluralism’, Economica 45: 391–400. Stigler, G. and Becker, G. (1977) ‘De gustibus non est disputandum’, American Economic Review 67: 76–90. Stone, C. (1980) ‘Systemic power in community decision making: a restatement of stratification theory’, American Political Science Review 74: 978–90. Stone, C. (1987) ‘The study of the politics of urban developments’, in C. Stone and H. T. Sanders (eds), The Politics of Urban Development (Lawrence: University of Kansas Press). Stone, C. (1989a) Regime Politics: Governing Atlanta 1946–1988 (Lawrence: University of Kansas Press). Stone, C. (1989b) ‘Paradigms, power and urban leadership’, in B. D. Jones (ed.), Leadership and Politics (Lawrence: University of Kansas Press). Strongman, K. T. and Champness, B. G. (1968) ‘Dominance hierarchies and conflict in eye contact’, Acta Psychologica 28 (3): 376–86. Sugden, R. (1985) ‘Liberty, preference and choice’, Economics and Philosophy 1: 213–19. Sugden, R. (1998) ‘The metric of opportunity’, Economics and Philosophy 14: 307–37. Sugden, R. (2003) ‘Opportunity as a space for individuality: its value and impossibility of measuring it’, Ethics 113: 783–809.

294 References Sunstein, C. R. and Thaler, R. R. (2008) Nudge: Improving Decisions about Health, Wealth and Happiness (New Haven: Yale University Press). Sutton, J. (1986) ‘Non-cooperative bargaining theory: an introduction’, Review of Economic Studies 176: 709–24. Suzumura, K. (1991) ‘Alternative approaches to libertarian rights in the theory of social choice’, in K. J. Arrow (ed.), Issues in Contemporary Economics, Vol. 1: Markets and Welfare (Basingstoke: Macmillan). Taleb, N. N. (2007) The Black Swan: The Impact of the Highly Improbable (New York: Random House). Taylor, M. (1982) Community, Anarchy and Liberty (Cambridge: Cambridge University Press). Teske, P., Schneider, M., Mintrom, M. and Best, S. (1993) ‘Establishing the micro foundations of a macro theory: information, movers, and the competitive local market for public goods’, American Political Science Review 87: 702–13. Theakston, K. and Gill, M. (2011) ‘The postwar premiership league’, Political Quarterly 82 (1): 67–80. Thomson, S. and Dixon, A. (2004) ‘Choices in healthcare: the European experience’, Euro Observer 6 (4): 1–4. Tsebelis, G. (2002) Veto Players: How Political Institutions Work (Princeton: Princeton University Press). Tuck, R. (1979) Natural Rights Theories (Cambridge: Cambridge University Press). Ullmann-Margalit, E. and Morgenbesser, S. (1977) ‘Picking and choosing’, Social Research 44: 757–84. Van Hees, M. (1995) Rights and Decisions: Formal Models of Law and Liberalism (Dordrecht: Kluwer). Van Hees, M. (1998) ‘On the analysis of negative freedom’, Theory and Decision 45: 175–97. Van Hees, M. (1999) ‘Liberalism, efficiency, and stability: some possibility results’, Journal of Economic Theory 88: 294–309. Van Hees, M. (2000a) Legal Reductionism and Freedom (Dordrecht: Kluwer). Van Hees, M. (2000b) ‘Negative freedom and the liberal paradoxes’, Rationality and Society 12: 335–52. Van Hees, M. (2004) ‘Freedom of choice and diversity of options: some difficulties’, Social Choice and Welfare 22: 253–66. Van Hees, M. and Wissenburg, M. (1999) ‘Freedom and opportunity’, Political Studies 47 (1): 67–82. Waldron, J. (1984) ‘Introduction’, in J. Waldron (ed.), Theories of Rights (Oxford: Oxford University Press). Waldron, J. (1988) The Right to Private Property (Oxford: Oxford University Press). Waldron¸ J. (1989) ‘Rights in conflict’, Ethics 99: 503–16. Walker, J. (1983) ‘The origins and maintentance of interest groups in America’, American Political Science Review 77: 390–406. Ward, H. (1987) ‘Structural power: a contradiction in terms?’, Political Studies 35: 593–610.



References295

Weale, A. (1990) ‘Social policy’, in P. Dunleavy, A. Gamble and G. Peele (eds), Developments in British Politics 3 (London: Macmillan). Weber, M. (1978) Economy and Society, Vols 1 and 2, ed. G. Roth and C. Wittich (Berkeley: University of California Press). Wellman, C. (1995) Real Rights (New York: Oxford University Press). Wellman, C. (1998) The Proliferation of Rights: Moral Progress or Empty Rhetoric? (Boulder: Westview Press). White, A. R. (1984) Rights (Oxford: Clarendon Press). Widgrén, M. (1994) ‘Voting power in the EC and the consequences of two different enlargements’, European Economic Review 38: 1153–70. Williams, A. (1998) ‘Incentives, inequality and publicity’, Philosophy and Public Affairs 27: 225–47. Williams, J. and Rossiter, A. (2004) Choice: The Evidence (London: Social Market Foundation). Williamson, O. E. (1975) Markets and Hierarchies (London: Collier-Macmillan). Williamson, O. E. (1996) The Mechanisms of Government (New York: Oxford University Press). Wriglesworth, J. L. (1985) Libertarian Conflicts in Social Choice (Cambridge: Cambridge University Press). Yammarino, F. J. and Dansereau, F. (2008) ‘Multi-level nature of and multi-level approaches to leadership’, Leadership Quarterly 19 (1): 135–41. Young, H. (1989) One of Us: A Biography of Margaret Thatcher (London: Macmillan).

Index

ability 30, 42, 50, 60, 70, 73–6, 82–4, 88, 90–6, 143–6, 151 ableness 83–4, 88, 91–6 agents 47–50, 52–3 as roles 9–10, 50, 56–7, 61, 95–6, 98, 130 agent–structure distinction 47–67, 98 arbitrary will 214–22 Arrow Theorem 3, 66 n.2, 258 Attlee, Clement 166 autonomy see free will Blair, Tony 155, 165, 170 n.7, 188 blame 7, 15 blame fallacy 69–70, 107 Brown, Gordon 155, 166 capabilities 82–3, 99, 115–16, 138, 148–9, 154 n.11, 175 capitalists 7, 41, 43, 105, 117–18 cat 33, 35–6 centre of gravity 34 choice 14, 20–2, 135, 147–8, 173–87, 188–204 and individuation 180–4, 191, 192–3 and markets 184–6, 199–200 and power 189 and public service 200–3 health 201–2 housing 202 school choice 183–4, 200–1, 202–3 and quality 202–3 and unconnected preferences 182–3 cardinality 189 costs of 196–200 information 196 psychological 198–200 transition 196–8 welfare 196

diversity 191, 192 flexibility 192 increasing 176–8, 203–4 measurement of 176–8, 189–93, 212–14, 222–3 Pattanaik and Xu axioms 190 picking 136, 147, 181–2, 196 see also freedom coalitions 11–12, 14, 22, 112, 119, 125–9, 208 coalition problem 208, 214–20 collective action problem 8, 16, 23, 68–9, 75, 107, 111, 138 community power debate 68–81 conceptual analysis 1–8, 33–9, 106 as non-normative as possible 7, 31–40, 44–5 different conceptual partitions 5–6, 32, 45–6, 108–9 normative dispute 6–7 subscript strategy 3–4, 5–7 verbal disagreement 4–5, 6, 94 control 138 Council of Ministers 125–6 counterfactuals 91–5, 116 Decisional Studies 68–9 decisiveness 16, 17, 91–5, 108, 111, 125–9, 264 democracy 3, 44, 54–5, 73, 119, 133, 157, 229 n.14 dictators 121 difference principle 145 dispositional properties 83–9 dispositions 209–10, 214–20 diversity 136 DNA see genes domination see freedom



Index297

Equality of Opportunity (EoP) measure 151–2, 154 n.12 essential contestability 1–6, 29, 30, 31–2, 45–6, 62 historical element in 1–2 explanation 33, 59–60, 62–4 farmers 41, 116–17 freedom 19 and happiness 35 bivalency 206 distinct from right 251 domination 207–25 highwayman example 211–12 independent or intrinsic value 20, 24, 35, 177, 193, 194–6 instrumental value 20, 35, 193, 197–8 interference 207–8 master–slave 205–6, 225–7 measurement see choice, measurement; rights, measurement negative and positive 266 pure negative 206, 207–8, 210–14, 224, 228 n.12 republican freedom 12, 22, 205–29 resilience or robustness of 22, 223–7 subjunctive 242–3, 252 see also choice; rights free will 19, 20, 55–6, 57, 59, 65, 66–7 nn. 7, 8, 135, 186, 192 gambler’s fallacy 34, 162 genes 34, 90, 95, 137, 144, 153–4 n.4, 155, 160–1, 169–70 n.3 gentle giant 216–17 Global Financial Crisis 41 growth coalitions 73 habitus 56–7 incentives 123–4, 148–50 individuation 180–4 information or knowledge 122, 136 interference see freedom intoxication 85–6, 99–100 nn. 2, 3 judges 91–5, 100 n. 8 Kaldar-Hicks compensation 138 Lady Chatterley’s Lover 235, 257, 259, 264–7, 269, 273–4 n.9 leadership 14, 155–70 bliss point 164–5

qualities 157–9 reputation 164–9 types 156–7 legitimate authority 122–3 luck 11, 13, 14–19, 52, 70, 72–3, 74, 91, 94, 105, 115, 117, 125–30, 133–5, 137–54, 160–3 Barry luck 17–19 brute luck 18, 134–5, 137–44, 149–51, 153 n.1, 154 n.6 contingent brute luck 18–19, 137–44, 149 coin tosses 141–3 measuring luck 138–42 option luck 18, 134–5, 137–44 path dependency143–6 personal identity (PI) luck 14, 17–19, 137–44, 150–2, 154 n.6 see also systematic luck luck egalitarianism 18–19, 70, 134–54 Major, John 155, 165, 170 n.7, 173 manipulation 131–2 methodological individualism 47–50 Napoleon 160, 161 narrative fallacy 158 natural kinds 34–5, 36 opportunity set 135–6, 176–80, 194, 212–14, 264–7 paranoid pedestrian 216, 222 Paretian Liberal 23 patterns 33, 143–8, 162–3 picking see under choice pluralism 68 power agenda power 30 agential accounts 9, 14, 47–51, 62–4 and game theory 69, 84, 93, 100 n.7, 105, 120 and interests 38–9 appearance becomes reality 14, 155–6, 160–70 as clash of wills 110–11, 112 as a disposition 83–4, 112, 115, 121 dynamic 8, 10, 29 exercise fallacy 9, 13, 83–4 extensional account 84–9, 94–5, 109 hegemonic 39–40, 41, 42, 43, 44–6 I-power 30 outcome power 17, 22, 70, 110, 113, 241

298 Index power (cont.) P-power 30 perceptual argument 14 power indices 12, 16, 17, 30, 73, 91–5, 125–9 Shapley–Shubik 73, 94, 126–8 power over 6, 8, 29, 32–3 power to 6, 8, 29 preference-free 12 resource account 10, 72–3, 89–91, 112–16, 121–5 social power 70, 113, 116 soft power 13, 30, 39, 43–6 static 8, 10, 59 structural power 8–9, 14, 63 see also structure conceptual argument 9, 52, 106 redundancy argument 9, 51–2, 106 symbolic power (force) 58 vehicle fallacy 83–9, 98–9, 109 veto power 38 zero sum 37 prediction 33, 46 preference-shaping 113, 129–33 Prisoners’ Dilemma 69, 178, 258, 274 n.12 privileged group 41 public goods 77–8 QALY 196 quasi market 194–5 real patterns 5, 34, 45–6, 143–8, 162–3 Red Dwarf 154 n.9 regime theory 75, 79–80, 106 republican freedom see under freedom reputation 10–11, 76–7, 90, 115, 120 n.10, 124–5, 159, 164–8 responsibility 7, 8, 14–17, 18–22, 66 n.4, 133–6, 146–8 reward structure 19, 139–41, 152 rights 22–4, 230–5 active and passive 269–70 act-type and act-token 240–1, 247–51 adjudicator 232 and voting 230–1, 253, 269, 270 as Hohfeldian claims 22, 24, 225–7, 239–40, 253–4 n.2, 262–4 as Hohfeldian privilege or liberty 23, 24, 225–7, 253–4 n.2, 262–3 compossibility 230–1, 232–7, 239–43, 247, 251–2, 267 distinct from freedom 251

exercising 232, 240, 246–7, 268–70, 274 n.13 existence of 243–7 formal 232, 244–6, 247 as a particular 245 as a universal 245–6 game-form representation 237–9, 259–62, 264–7, 268–71 admissible or permissible strategy 269, 272 grammar 262–3 legal 232, 244 material 232, 244, 246–7 as a particular 246 as a universal 246 measuring 247–51 moral 242 positive and negative 251 Sen rights 232–7, 257–8, 264–7 threshold value 249–53 see also freedom rule of anticipated reactions 116, 118, 216, 221 Sen’s Impossibility Theorem 233–7, 257–9, 264–7 shirt games 238, 259–62, 274 n.14 skill 143–6 social software 61 special 31, 46 specification problem 158 structural suggestion 49, 130 structure 9, 51–6, 65 deep structure 52, 55, 56–7, 64–5 structure-induced equilibrium 41, 54 systematic luck 6–7, 14, 22, 40–6, 105–7, 117–19, 129, 133 talent 18–19, 143–6, 149 151 see also skill; temperament; traits temperament 143–4, 146–7, 161–2, 163–4 see also skill; talent; traits Thatcher, Margaret 155, 165–8, 170 nn.7–9, 173, 188 housing policy 113, 131 traits 158, 160–3 see also skill; talent; temperament type–token 271 utility indirect 192 process 20–4, 195 whisky 85–6, 99–100 n.2