Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Political freedom : a defence of natural rights republicanism Hellewell, Jamie Scott 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_february_hellewell_jamie.pdf [ 1.3MB ]
Metadata
JSON: 24-1.0135606.json
JSON-LD: 24-1.0135606-ld.json
RDF/XML (Pretty): 24-1.0135606-rdf.xml
RDF/JSON: 24-1.0135606-rdf.json
Turtle: 24-1.0135606-turtle.txt
N-Triples: 24-1.0135606-rdf-ntriples.txt
Original Record: 24-1.0135606-source.json
Full Text
24-1.0135606-fulltext.txt
Citation
24-1.0135606.ris

Full Text

Political Freedom:  A Defence of Natural Rights Republicanism ?by ?Jamie Scott Hellewell ???A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in  The Faculty of Graduate and Postdoctoral Studies (Philosophy) The University of British Columbia (Vancouver) November, 2014 ???© Jamie Scott Hellewell, 2014
?Abstract ?What is political freedom? Many, especially liberal, philosophers have followed Isaiah Berlin in insisting we understand liberty as essentially about the absence of interference in the life choices of individuals. However “neo-Roman republicans,” prominently Quentin Skinner and Philip Pettit, have challenged Berlin’s view. They argue that political freedom is better understood as the absence of domination, where “domination” consists in being under another person’s power in the manner of a slave or a despot’s subject. Such forms of power undermine freedom in virtue of being arbitrary -- i.e. unaccountable to the interests of those subject to them. In consequence, neo-republicans hold that freedom is realized through the democratic control of social life.  In this dissertation, I argue that neither of these alternatives is entirely satisfactory. Berlin’s non-interference view captures a basic and compelling liberal insight, namely that there ought to be limits on how far any community, democratic or otherwise, can intrude upon the lives of individuals. Yet it neglects the way in which arbitrary power can undermine freedom even in the absence of actual interferences, leading Berlin to embrace the implausible conclusion that freedom is in principle compatible with despotic rule. By contrast, the neo-republican view supports a more plausible account of the relation between freedom and power, and thus between freedom and democratic government. Yet, I argue, it is unable to justify or provide the proper criteria for setting limits on the state’s authority over individuals. It thus risks licensing a ‘tyranny of the majority.’  ii Given these challenges, I defend an alternative account of political freedom: While I accept the republican claim that freedom consists in non-domination rather than non-interference, I argue that the neo-Roman account of the nature of domination is problematic, resting on a flawed conception of arbitrary power. I look to Locke for a more plausible account of arbitrary power and thus the ideal of non-domination, one which defines political freedom against a background of natural rights. The resulting view -- “natural rights republicanism” -- plausibly accommodates both liberal concerns for individual sovereignty and republican sensitivity to relations of power. ?iiiPreface ?Jamie Hellewell is the sole author of this thesis. He was responsible for the design of the research program, the research itself and the composition of the thesis. This thesis is original, unpublished work. ?ivTable of Contents ?Abstract………………………………………………………………………………………….ii Preface………………………………………………………………………………………….iv Table of Contents………………………………………………………………………………v  List of Tables…………………………………………………………………………………..xi Acknowledgements………………………………………………………………………….xii Chapter 1: Introduction……………………………………………………………………….1  1.1  Topics in the philosophy of freedom………………………………………………….2  1.1.1  Conceiving political freedom…….….…………………………………………..2  1.1.2  The value of political freedom……..……………………………………………5  1.1.3  The distribution of freedom….………………………………………………….8  1.1.4  A ‘theory of freedom’…..……………………………………………………….11  1.2  “Two Concepts of Liberty”..………………………………………………………….14   1.2.1  Negative and positive liberty.………………………………………………….14  1.2.2  ‘Freedom’ versus ‘ability’…….……………………………….………………..15  1.2.3  The argument(s) of “Two Concepts”…….……………………………………17  1.3. Liberty after “Two Concepts”..……………………………………………………….23  1.3.1 Freedom as non-interference: the empirical approach…………….……….24  1.3.2. Freedom as non-interference: the rights-based approach……….………..29  1.3.3. General objections to the negative concept of liberty………………………34  1.3.4. Freedom as effective capacity…………………………..…………………….36  1.3.5. Freedom as personal autonomy………………………….…………………..43 v 1.3.6. Freedom as non-domination…………………………………………………..50  1.4. The problem of this dissertation…………..………………………………………..56  1.4.1. Freedom as an external social relation…..………………………………….57  1.4.2. Non-interference versus non-domination……………………………………58  1.4.3. Contending insights……….……………………………………………………59  1.4.4. The conjunctive approach……………………………………….…………….61  1.4.5. The Lockean alternative……………………………………………………….64  1.4.6. Chapter breakdown…………….………………………………………………65 Chapter 2: Republican “Freedom” and its Liberal Critics……………………………67  2.1. Freedom as self-mastery..…………………………………………………………..68  2.1.1. ‘Neo-Athenian’ and ‘neo-Roman’ republicanism……………………………69  2.1.2. ‘Freedom’ in Rousseau’s republic…………………………………………….71  2.2. Varieties of ‘self-mastery’: the individualistic interpretations……………………..72  2.2.1. Individual self-mastery…………………………………………………………73  2.2.2. Primitive self-mastery………………………………………………………….73  2.2.3. Rational autonomy……………………………………………………………..74  2.2.4. Procedural versus substantive autonomy……………………………………76  2.2.5. Individual autonomy and the coordination problem…………….…………..79  2.3. Varieties of self-mastery: the collective interpretations………….………………..82  2.3.1. The ‘metaphysical’ interpretation………………………………….………….82  2.3.2. The ‘participationist’ interpretation: issues……………………….………….84  2.3.3. The ‘populist’ interpretation………………………………………….………..88  2.3.4. The ‘deliberative’ interpretation……………………………………….………89 vi 2.3.5. The ‘rationalist’ interpretation………………………………………………….91  2.3.6. Summary………………………………………………………………………..92  2.4. The classic liberal critique of republican “freedom”…………………………..…..94  2.4.1. Benjamin Constant…………………………………………………….……….94  2.4.2. John Stuart Mill…………………………………………………………………97  2.4.3. Isaiah Berlin……………………………………………………………………100  2.4.4. Individual sovereignty and negative liberty………………………….……..101  2.4.5. Berlin’s critique of the ‘rationalist interpretation’…………………….……..103  2.4.6. Divided self objection…………………………………………………………105  2.4.7. Value pluralism objection…………….………………………………………108  2.5. Defining the liberal insight……………………………………………….………….111  2.5.1. Constant versus Berlin………………………………………………………..112  2.6. Conclusion……………………………………………………………………………115 Chapter 3: Freedom as non-Domination………………………………………………..117  3.1. Neo-Roman republicanism…………………………………………………………117  3.1.1. Neo-Roman versus neo-Athenian freedom……………….………………..118  3.1.2. Neo-Roman republicanism: three fundamental claims….………………..119  3.2. Non-domination and non-interference………………………………………….…120  3.2.1. Freedom as non-interference………………………………………………..120  3.2.2. Freedom as non-domination…………………………………………………121  3.2.3. Non-interference and non-domination contrasted…………………………123  3.3. The case for the non-domination conception…….………………………………124  3.3.1. Self-editing effects………………………….…………………………………124 vii 3.3.2. Slavery and despotism……………………………………………………….127  3.4. The liberal rejoinder…………………………………………………………………136  3.4.1. Actual and subjunctive interferences…………………………………….…137  3.4.2. Subjunctive interferences and probabilities………………………………..139  3.4.3. Republican objections answered…………………….…………….………..140  3.5. A hollow triumph?………..………………………………………………………….144  3.5.1. Threats versus subjunctive interferences………………………….……….144  3.5.2. Two further reasons to endorse the amended formulation……………….146  3.5.3. Summary………………………………………………………………………154  3.6. Freedom in chains?…………………………………………………………………155  3.6.1. Three objections………………………………………………………………155  3.6.2. Replies…………………………………………………………………………157  3.7. Summary and conclusion……………………………………………………….…163 Chapter 4: The Nature of Arbitrary Power……………………………………………..167  4.1. Introduction……………………………………………………………………….…167  4.2. What is arbitrary power?…………………………………………………………..168  4.2.1. The procedural condition……………………………………………………..169  4.2.2. The interest-tracking condition………………………………………………171  4.3. The free republic……………………………………………………………..……..173  4.4. Tyranny of the majority?………..…………………………………………………..178  4.4.1. Initial statement of the problem……………………………………………..178  4.4.2. Pettit’s counter-majoritarian condition………………………………………180  4.4.3. Are the counter-majoritarian measures sufficient?………………………..181 viii 4.4.4. Can the counter-majoritarian constraints be justified?……………………182  4.5. Rethinking arbitrary power……………………………………………………..…..186  4.5.1. The ‘objective interest’ interpretation……………………………………….188  4.5.2. The ‘subjective preference’ interpretation………………………………….190  4.5.3. The ‘deliberative’ interpretation(s)…………………………………………..193  4.5.4. The ‘deliberative-democratic’ interpretation: problems…….……………..195  4.5.5. The ‘deliberative-communitarian’ interpretation: problems….……………198  4.6. Beyond the interest-tracking condition…………………………………….…..…202 Chapter 5: Locke’s Natural Rights Republicanism: A Defence…………………….206  5.1. Interpreting Locke on political freedom……………………………….………….206  5.1.1. The negative liberty interpretation…………………………………………..207  5.1.2. The positive liberty interpretation……………………………………………208  5.1.3. The republican interpretation………………………………………………..211  5.2. Freedom in the state of nature……………………….…………….……………..213  5.2.1. Locke on ‘natural liberty’……………………………………………………..214  5.2.2. Why natural liberty is unsustainable………………………………………..215  5.3. Freedom under government……………………………….………….…………..217  5.3.1. Locke on arbitrary power: the procedural condition…………….…………217  5.3.2. Locke on arbitrary power: the substantive condition…………….………..219  5.3.3. Locke on freedom and democracy……………………………………….…224  5.4. Lockean freedom defended……………………………..…………………………227  5.4.1. Lockean freedom and democracy: a refinement…………………………..228  5.4.2. Lockean freedom and non-interference: a clarification…………….……..229 ix 5.4.3. Lockean freedom and property rights: a qualification…………………….230  5.4.4. Lockean freedom and the ‘liberal insight’…………………………………..231  5.4.5. Lockean freedom and the ‘republican insight’……………………………..232  5.4.6. Lockean freedom and arbitrary power………………………….…………..235  5.5. Natural rights republicanism…………………………………..……….…………..238  5.5.1. Berlin’s narrative of positive and negative liberty……………….…………238  5.5.2. The republican revisionist narrative……………………………….………..240  5.5.3. Rethinking the republican heritage………………………………….………242  5.6. Conclusion……………………………………………………..………………….…248 Chapter 6: Objections and Replies………………………………………………………249  6.1. Introduction…………………………………..………………………………………249  6.2. Objections to Locke’s catalogue of natural rights…..……………………………252  6.3. Objections to the idea of ‘natural’ rights…………………………..………………256  6.4. Objections to “moralizing” liberty……………………..……………………………261  6.4.1.Two ways of moralizing ‘freedom’……………………………………………261  6.4.2. What’s wrong with moralizing freedom?……………………………………263  6.4.3. Reply to the objection from pluralism……………………………………….267  6.4.4. Reply to the objection from the independent value of freedom………….273  6.4.5. Reply to the circularity objection…………………………………………….281  6.5. Conclusion……………………………………………………..…………………….286 Bibliography…………………………………………………..……………………………..287 ?xList of Tables Table 1.1: Lockean Freedom in Logical Space………………………………………….65 Table 6.1: Locke and Nozick Contrasted………………………..…….………………..251 ?xiAcknowledgements ?I wish to thank my advisor, Paul Russell, for his insight into and passionate interest in the topic of political freedom, and for his patient and conversational approach to my research process. I wish to thank Scott Anderson and Matt Bedke, the other members of my dissertation committee, for their detailed and helpful comments on my dissertation draft. I have also benefited greatly from seminar courses and conversations with Scott Anderson, especially one on the topic of coercion which had an important influence on my thinking about the nature of political freedom.  I am grateful to the UBC Philosophy Department as a whole for its support and commitment to providing a rigorous philosophical learning and research environment. ?I am very thankful to my wife April Griffin and to other friends and family who have been a consistent and encouraging support through the whole Ph.D. process.  ?I was supported at UBC by a Graduate Entrance Scholarship, a Sunlife Financial Graduate Scholarship, and an Institute for Humane Studies Fellowship. From 2007 - 2010, I was supported by a Doctoral Fellowship awarded by the Social Sciences and Humanities Research Council of Canada. ?xiiChapter 1 Introduction ?“The world has never had a good definition of the word ‘liberty,’” remarked Abraham Lincoln, speaking to a Baltimore audience in 1864, “We all declare for liberty; but in using the same word we do not all mean the same thing.” Lincoln’s observation might just as easily be applied to discussions of the ideal of political freedom in contemporary political philosophy. While most participants to such discussions claim freedom as one of the central political values of a just or well-ordered society, there is significant disagreement about how to conceive of this ideal. Indeed, a variety of conceptions of freedom now prevail in the literature, and with them a variety of competing views of the “free society.”  In this introductory chapter, I review the relevant philosophical literature with a view to: ?(a) Mapping the key philosophical issues involved in a theory of political freedom; (b) Producing a catalogue of what I take to be the most influential and important contemporary conceptions of the ideal political freedom; (c) Indicating what I take to be the chief difficulties with or objections to each of these conceptions; (d) Framing the problem of this dissertation. ?11.1. Topics in the philosophy of freedom  Philosophical discussions of political freedom are largely focused around three basic questions: (1) What is political freedom? (2) What is the value of political freedom? (3) How ought we to distribute political freedom?  1.1.1. Conceiving political freedom   The first of these three questions concerns the problem of how best to define or conceive of the ideal of political freedom. Proposed accounts of the notion of freedom tend to take the form “Freedom is...” or “Freedom consists in...”, and involve offering a set of necessary and sufficient conditions or defining criteria for some person or group to count as free, or free with respect to a particular action or set of actions. Following current practice, I shall call any such account a conception of freedom.   Connected with this question, philosophers have disagreed about whether there is in fact one or more than one “concept” of freedom.  Does our inherited political 1discourse contain multiple, distinct ideals each going under the name “freedom,” or is there instead a single unified concept of freedom concerning which only one proposed conceptual analysis can be correct? Benjamin Constant (1819) was perhaps the first to suggest that our conceptual heritage includes distinct ideals of freedom -- the “liberty of the ancients” and the “liberty of the moderns” -- each of which ought to be seen as having semantically legitimate claim to the title of “freedom.”  In this past century, Isaiah 2Berlin -- with his distinction between the “positive” and “negative” concepts of liberty -- is 2 See, e.g., MacCallum, “Negative and Positive Freedom”; Feinberg, Social Philosophy; Oppenheim, 1Dimensions of Freedom: An Analysis; Taylor, “What Wrong with Negative Liberty” Constant, “The Liberty of the Ancients compared with that of the Moderns”2certainly the most well-known advocate of this view.  Let’s call this the conceptual 3pluralism thesis.  In an influential early response to Berlin’s essay, Gerald C. MacCallum argued that the distinction between positive and negative liberty is based upon a confusion, that -- contra the conceptual pluralism thesis -- there is a single concept of freedom underlying the apparent multiplicity of uses.  He suggests that we regard the concept of 4freedom as always implying a triadic relation: ‘an agent X is free from obstacle Y to do some action Z.’ This is the only concept of freedom, MacCallum insists. Questions appropriately arise in determining the range of objects that legitimately fall within the domain of the three variables. Should only human individuals count under X, or might animals, groups, etc. also be included? Should only external barriers count under Y, or should internal psychological barriers be included? Should only barriers intentionally imposed by other humans count, or should even natural barriers be included? Or, perhaps, all humanly-alterable barriers? The differences Berlin attributes to distinct concepts of liberty, are actually, on MacCallum’s account, differences about how to set the range of the variables involved in the triadic relation.  Other authors, however, have rejected MacCallum’s single-concept approach. Charles Taylor, for instance, has argued that the positive concept of freedom as self-mastery is what he calls an “exercise” rather than an “opportunity” concept, and so does 3 Berlin, “Two Concepts of Liberty”3 MacCallum, “Negative and Positive Freedom” 4not fit MacCallum’s schema after all.  Quentin Skinner has similarly claimed the 5republican conception of freedom as non-dependency cannot be cashed out in terms of obstacles to potential actions.  Berlin himself rejected MacCallum’s overtures, insisting 6that it is an error to think that anytime we aspire to be free from some Y we have any defined conception of a Z we would then be free for: A “man struggling against his chains or a people against enslavement need not consciously aim at any definite further state.”  7 In any case, it seems to me that this debate is ultimately beside the point. Even if MacCallum is right about the shared underlying structure of competing ideals of freedom, it does not follow from this that we are really dealing with a “single” concept of freedom,  or that the differences between the competing accounts are not significant. In 8fact, all the issues of substance remain. Consequently, I think it is best to leave the one-versus-multiple concepts debate to the side. An ecumenical way to move forward, would be to characterize the contending views, in Rawlsian language, as offering different “conceptions” of the ideal of freedom. The philosophical question concerning which conception of freedom we ought to adopt, then, can plausibly be understood as a 4 Taylor, “What’s Wrong with Negative Liberty”; In “Liberty: One Concept Too Many?”, however, Eric 5Nelson has argued against re-establishing the negative-positive concept-dichotomy along lines suggested by Taylor. He attempts to show that views which think about freedom in terms of “self-mastery” can be fully captured by reference to the absence of internal barriers to opportunities one would otherwise have. Introducing talk about exercise versus opportunity concepts only further muddies the waters of substantive debates about freedom. He, thus, argues that we follow MacCallum in insisting that there is a single concept of freedom captured by the triadic relation above, and that the differences among the contending views are really differences about how to best interpret the three variables. “The Paradoxes of Political Liberty,” 206 “Two Concepts,” xliii n.17 Consider an analogy. Suppose that all conceptions of “love” exhibit the follow structure: There is a lover 8L, who stands in a relation of attraction and/or admiration A, to some beloved B. It does not follow from this structural commonality that there are no substantively different conceptions of love or forms of love -- e.g. consider the important differences between the love of friendship, the love of parents for children, the love of romantic partners, etc. Sometimes the devil is in the variables. question about which conception of freedom best answers to our more particular judgements about political freedom in “reflective equilibrium.”  This approach will enable 9us to move on to focusing on more substantive disagreements. 1.1.2. The value of political freedom  The second fundamental question around which philosophical discussions of freedom revolve concerns the value of freedom. What is the value of political freedom? Not everyone has thought that political freedom is a valuable thing for human society. Joseph de Maistre, writing in the aftermath of the Jacobin Terror, considered liberty to be a positively dangerous thing. In contrast to the optimism of the philosophes, de Maistre viewed human nature as dark, vicious and bloody, and thus insisted that the only hope for civilized existence rested in its containment and restraint, not its liberation.    10 However, de Maistre’s views notwithstanding, virtually everyone now writing on the topic believes that freedom is a valuable thing -- in many cases the most valuable thing. However philosophers differ concerning (a) whether its value is merely instrumental or at least partly intrinsic; they disagree about (b) what sort of value freedom realizes; and they disagree about (c) whether freedom as such is valuable (i.e. regardless what it is freedom for) or whether only specific freedoms are valuable.  (a) Is being free good for its own sake or is it good simply in virtue of other positive consequences that tend to result from people’s being free? That is, is freedom 5 Rawls, A Theory of Justice, 189 de Maistre writes: “Religion and slavery” are the “two anchors of society”; “All greatness, all power, all 10social order depends upon the executioner; he is the terror of human society and the tie that holds it together. Take away this incomprehensible force from the world, and at that very moment order is superseded by chaos, thrones fall, and society disappears” (as quoted in Berlin, Freedom and its Betrayal, 144, 149)intrinsically valuable or merely instrumentally valuable? John Stuart Mill (1859) appears to take the latter view. Though On Liberty is among the great philosophical defenses of personal freedom, Mill is clear that he regards “utility as the ultimate appeal on all ethical questions.”  And so he undertakes to defend the ‘harm principle’  on grounds 11 12that the wide scope of individual liberty it implies will lead in the long-run to the greatest happiness for the greatest number. The whole argument of On Liberty is thus structured around providing a purely instrumental argument in favour of individual liberty.  On the other hand, without denying the instrumental value of freedom, most philosophers take freedom to be something that is also of great intrinsic value. Thus Berlin, in a rare disagreement with Mill, insists that “coercion is, insofar as it frustrates human desires, bad as such... while non-interference, which is the opposite of coercion, is good as such.”  Or, in Rousseau’s more emphatic words: “To renounce our freedom 13is to renounce our character as men, the rights, and even the duties, of humanity. No compensation is possible for anyone who renounces everything.”   14 (b) A second question regarding the value of freedom concerns what sort of value freedom is supposed to realize. Philosophers offer different accounts of both the instrumental and the intrinsic value of freedom. Mill locates the value of freedom in the fact that it provides fertile ground for spontaneity, originality and genius, for the testing of new ideas and ways of living, the net result of which is a progress towards truth, 6 On Liberty, 1011 “...the sole end for which mankind are warranted, individually or collectively, in interfering with the 12liberty of action of any of their number is self-protection... the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others” (68). “Two Concepts,” 12813 The Social Contract, 5014towards more fulfilling ways of living, and in general the advancement of the permanent interests of humanity as a progressive species.  John Gray sees freedom as valuable 15in that it enables people of diverse religious and philosophical points of view to coexist and even cooperate in peaceful toleration.  Others view freedom, especially economic 16freedom, as valuable because free markets are better instruments for promoting economic growth and satisfying human needs and preferences.   17 Those who emphasize the intrinsic value of freedom also differ on what value they take the condition of freedom to consist in. For those who conceive of freedom as non-interference, the value of freedom consists in not being coerced, of not being subject to humanly-imposed obstacles to your choices. For those who conceive of freedom in terms of rational autonomy, the value of freedom might reside in rising above one’s empirical nature, or in realizing one’s true or authentic self. For those who conceive of freedom as the non-domination, the value of freedom might be thought to be connected to the recognition of one’s status as an equal among one’s fellow citizens.   (c) A third question regarding the value of freedom concerns whether freedom should be seen as “non-specifically valuable” or whether it is only specific freedoms or “liberties” that should be seen as valuable. Ronald Dworkin, who conceives of freedom in terms of non-interference , thinks that non-interference as such is not of any 18particular value; rather it is only specific sorts of liberties that matter -- i.e. freedom of 7 On Liberty, 7015 Isaiah Berlin16 cf. Hayek, The Constitution of Liberty17 More precisely, Dworkin conceives of freedom as the absence of interference in those actions one has 18a right to undertake. Dworkin’s view is thus an instance of what I refer to below as the “rights-based approach” to the non-interference conception of freedom.conscience and belief, freedom of expression, freedom of religion, freedom of association, freedom to own and exchange property, etc. Likewise, Rawls defines the first principle of justice as follows: “Each person has the same indefeasible claim to a fully adequate scheme of equal basic liberties, which scheme is compatible with the same scheme of liberties for all.”  The “equal basic liberties” includes the liberty of 19conscience and freedom of association, freedom of speech and liberty of the person, and the rights to vote, to hold public office, and to be treated in accordance with the rule of law. However, others, like Ian Carter, have insisted that the value of liberty resides not merely in the value of specific liberties (i.e. liberties for certain kinds of actions only) but in the value of liberty as such -- i.e. “non-specific” overall liberty.   201.1.3. The distribution of freedom  The third fundamental question around which philosophical discussions of freedom revolve concerns the distribution of freedom. The most widely accepted conceptions of freedom consist in definitions of what is required for an individual to be free, or for an individual to be free with respect to a certain action or range of actions. However, given individuals live in society with other individuals with at points conflicting aims, the question of naturally arises how the freedom of each ought to be coordinated or balanced against the freedom of all the rest. We can think of this as a question concerning the nature of the free society.   In On Liberty, Mill describes freedom as the absence of interference (either in the form of coercion or social pressure) in an individual’s ability to choose for himself how to 8 A Theory of Justice, 5319 Carter, A Measure of Freedom, 3220act and live his life.  Nonetheless, Mill recognizes that the freedom of any given 21individual is bound to impact others in various potentially harmful ways, including by imposing on their freedom. He thus undertakes to define and defend a principle according to which the boundaries of individual freedom ought to be drawn -- namely, the Harm Principle. According to the Harm Principle, “the sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number is self-protection... the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others.”  This principle embodies one possible answer to the distribution problem, 22namely that each individual in a free society shall be accorded the maximum possible range of non-interference in their personal activities (including thoughts, words and deeds), consistent with not harming others impacted by those activities.   But, unsurprisingly, there are other possible distributive principles: One possibility is that freedom in society ought to be maximized overall (the maximization principle); another is that each individual ought to accorded an equal degree of freedom (the equality principle) ; a third is that each individual ought to be accorded the maximum 23degree of freedom possible that is consistent with a like degree of freedom for everyone else (the maximum equal distribution principle); and, of course, we might also opt for a maximin or a leximin  distribution principle.  249 On Liberty, 6821 Ibid.22 cf. Steiner, “The Natural Right to Equal Freedom” and An Essay on Rights23 cf. van Parijs, Real Freedom for All: What (if anything) can Justify Capitalism?, 2524 One of the complications in determining an appropriate principle for distributing freedom concerns the question of whether freedom is the only good to be distributed or whether there are other social goods that a theory of distributive justice must take into account. If we take the latter view, then many of the principles just mentioned will likely be too strong, since satisfying them will leave little room for political values other than freedom. Berlin himself never offered anything like an explicit principle for the distribution of freedom, but he clearly held that freedom was just one among a number of important but competing political values against which it had to be balanced. To the extent that Berlin suggests a distributive principle at all, it is that there ought to be a certain minimum area of non-interference according to each individual that ought on no account to be violated  (call it the ‘guaranteed minimum principle’). As to the 25boundaries of this guaranteed minimum area of non-interference, Berlin does not offer a clear principle in the manner of Mill, but instead is content to defer to the evolving list of standard liberal “liberties.”    Rawls’ distributive principle also assumes that liberty is only one of the goods that a theory of justice must take into account -- though, he gives liberty a priority place.  Moreover, because he accepts a version of the specific freedoms thesis, he 26frames the issue of distribution as a question of how to distribute certain civil and political “liberties” -- what he calls the “basic liberties” -- in accordance with justice :  2710 “Two Concepts,” 12425 Rawls declines to address the question of how best to conceive of freedom directly, though he claims to 26accept MacCallum’s view that there is really just one concept of freedom as captured by the triadic formula (A Theory of Justice, 177).  Rawls’ list of civil liberties includes liberty of conscience, freedom of association, freedom of speech 27and liberty of the person; his list of political liberties includes the right to vote, to hold public office, and to be treated in accordance with the rule of law.Each person in society is to have as much of the “basic liberties” as possible, consistent with every other person having an equal amount.   Before moving on, a brief clarification is called for. I’ve framed the question of the nature of the free society as a question concerning the distribution of individual freedom. In doing so, I wish to assume neither that freedom is a commodity-like good, nor that it need be defined in such a way that the freedom of individuals can or must come into conflict. For instance, freedom might be best conceived of as self-realization -- the recognition and realization of one’s authentic or rational purposes. Talk of “distributing” self-realization is, admittedly, misplaced. In Chapter Two, I consider a conception of freedom according to which being free consists in actively participating in the collective decision-making process by which your community is governed. Understood in this way, there is nothing preventing the freedom of every person from being fully realized, since there is no necessary conflict between the freedom of each. In using the language of “distribution,” I only want to stress the point that an account of the free society must answer the question of how the freedom of an individual is to be coordinated, balanced, or reconciled with the freedom of others in society. Recognizing this point is consistent with holding that freedom cannot be simply distributed like apples; it is also consistent with holding that the complete freedom of the individual can be reconciled without loss with the complete freedom of everyone else in the community. 1.1.4. A ‘theory of freedom’  Philosophical discussions of political freedom, thus, encompass a range of issues revolving around three fundamental questions: How should freedom be conceived? What is the value of freedom? How should freedom be distributed? The 11majority of recent scholarship has focused most explicitly on the first of these questions, no doubt due to the influence of Isaiah Berlin’s influential writing on the topic. The aim of this dissertation is likewise to articulate and defend a particular conception of freedom. However, I would now like to propose the following methodological principle: Conceptions of freedom ought to be evaluated in light of the overall theory of freedom in which they are embedded. What is a ‘theory of freedom?’ I shall call a ‘theory of freedom,’ any view which provides a philosophical account of: (1) how ‘freedom’ is to be defined or conceived, (2) the value of freedom so defined, and (3) the principle or principles of just distribution of freedom so defined.   While a ‘theory of freedom’ must contain each of these components, it need not neatly distinguish them in the manner that, e.g., Mill does. For instance, someone might define freedom as the ability to engage in activities that constitute human flourishing. Here the account of the value of freedom (i.e. it is instrumental for achieving human flourishing) is implicit in the conception of freedom itself (i.e. the ability to engage in activities that constitute flourishing). Someone else may define freedom as the absence of interference in those activities an individual has a right to undertake.  Here a 28principle of just distribution is built into the very concept of freedom (i.e. the concept is “moralized”) -- though, of course, some elaboration of the content of an individual’s rights will be necessary to yield determinate judgements about what interferences constitute infringements of liberty and which do not. Adjusting my methodological principle slightly to accommodate these complexities, I propose that any conception of freedom ought be assessed in light of the general theory of freedom in which it is embedded or which it contains (in the manner just described). 12 cf. Nozick, Anarchy, State and Utopia; Rothbard, The Ethics of Liberty28 The rationale for this methodological principle is as follows. In proposing or criticizing various conceptions of freedom, what we are ultimately concerned with, I submit, is the question of how we ought to arrange our basic social and political institutions in order to properly protect, preserve, or realize freedom. Answering the conceptual question is necessary to define the value that we are seeking to realize, but it may not be sufficient to provide an informative political ideal. For instance, suppose you think that freedom is best conceived of as an absence of interference by other human beings in the choices an individual could otherwise make. That in itself will not tell us how far the non-interfered choices of individuals in society ought to extend. After all, one could, consistent with this conceptual claim, think that liberty so-defined is a dangerous or unattractive thing. Some account of the value of freedom is needed. Further, suppose you think that liberty so-defined is valuable, maybe even the supreme political value. That in itself still tells us little about who ought to be free to do what, when and where.  It would be unreasonable to expect that interference could be 29avoided altogether, if for no other reason that among the various actions individuals might undertake are those that would interfere in the action possibilities of others. Some principle is required for determining how freedom ought to be distributed. In short, except in the context of an overall theory of freedom, it may be difficult to assess the merits of a particular conception of freedom.  Having laid out the key philosophical issues involved in the topic of political freedom, and having proposed a methodological principle according to which proposed conceptions of freedom ought to be evaluated, I now want to turn to a selective review 13 Bernard Williams makes a similar point in “From Freedom to Liberty: The Construction of a Political 29Value.”of the recent and relevant scholarly literature. I’ll begin by sketching the argument of Isaiah Berlin’s seminal paper, “Two Concepts of Liberty,” then I will highlight and discuss some of the leading philosophical approaches to the ideal of political freedom in the post-“Two Concepts” literature, and finally narrow in on the recent debates between “Republican” and “Liberal” conceptions of freedom.   1.2. “Two Concepts of Liberty”  In 1958, liberal philosopher Isaiah Berlin delivered his inaugural address to an Oxford University audience under the title “Two Concepts of Liberty,” an address that was subsequently published in a collection of Berlin's papers entitled Four Essays on Liberty (1969). “Two Concepts” is largely responsible for instigating and setting the terms for the debate concerning the ideal of freedom that has unfolded in the past 50 years. For that reason, among others, it is the proper place to begin in conducting a review of that debate. 1.2.1. Negative and positive liberty  In “Two Concepts,” Berlin undertakes a critical examination of what he takes to be the two most influential conceptions of political freedom in modern societies -- what he calls the ‘negative’ and the ‘positive’ concepts of liberty. According to the ‘negative concept’ of liberty, freedom consists in the absence of interference by other human beings in what a person could otherwise do.  By contrast, according to the ‘positive 30concept’ of liberty, freedom consists in the positive exercise of self-mastery or self-government.  Despite their surface similarity, these two concepts represent rival and, at 31points, incompatible ideals, according to Berlin, and in fact speak to distinct fundamental 14 “Two Concepts,” 12230 Ibid., 13131concerns. The first ‘negative’ sense of freedom “is involved in the answer to the question ‘What is the area within which the subject - a person or group of persons - is or should be left to do or be what he is able to do or be, without interference by other persons?’; Whereas the second, 'positive' sense of freedom, “is involved in the answer to the question ‘What, or who, is the source of control or interference that can determine someone to do, or be, this rather than that?’” The first concerns itself with how far individuals are governed, while the second concerns itself with by whom they are governed.  321.2.2. ‘Freedom’ versus ‘ability’  It is important to emphasize that Berlin’s negative-positive distinction is not equivalent to the distinction that is commonly made between “formal” and “effective” freedom. An individual is formally free insofar as there are no legal impediments to their undertaking certain courses of action -- i.e. they are permitted to so act. Yet a person can be formally free to, e.g., to attend university and yet lack the ability to do so because they can’t pay the tuition fees. An individual has effective or “real” freedom insofar as they are actually able to undertake an action (in this case attend the university). Berlin’s positive-negative distinction is sometimes characterized along similar lines -- namely, that negative freedom is “freedom from” whereas positive 15 Ibid., 13032 Berlin does, at one point in “Two Concepts,” employ the language of “freedom from” and “freedom to” to 33characterize the distinction between negative and positive liberty (131). However, context makes clear that he does not have anything like the formal-effective distinction in mind: “The desire to be governed by myself, or at any rate to participate in the process by which my life is to be controlled, may be as deep a wish as that for a free area of action, and perhaps historically older. But it is not a desire for the same thing.... For it is this -- the ‘positive’ conception of liberty: not freedom from, but freedom to -- to lead one prescribed form of life -- which the adherents of the ‘negative’ notion represent as being, at times, no better than a specious disguise for brutal tyranny.”  freedom is “freedom to.”  However, this is a mistake.  Berlin’s conception of positive 33freedom is, as we’ve noted, an ideal of rational self-mastery.  Nonetheless, Berlin does think that the idea of freedom ought to be distinguished from the idea of ability. Citing Helvétius, Berlin notes, “The free man is the man who is not in irons, nor imprisoned in a gaol, nor terrorized like a slave by the fear of punishment... it is not lack of freedom not to fly like an eagle or swim like a whale.”  34“Mere incapacity to attain a goal,” Berlin insists, “is not lack of political freedom”; you lack political freedom only if you are prevented from attaining a goal by the interferences of other human beings.”   35 It follows from this distinction between freedom and ability that the mere fact that other human beings fail to provide me with opportunities to do things I might like to do, to “enable” me, does not mean that they thereby diminish my liberty. However, it does not follow from this distinction that having a wide set of opportunities or abilities is not a 16 Ibid., 122 n.234 Ibid., 122. It’s unfortunate that Berlin is inconsistent concerning whether “interference” is to include only 35those actions undertaken deliberately to prevent a person from attaining their goal, or whether the unintentional but nonetheless humanly-caused obstacles are also to be included. For example, on the one hand he writes: “Coercion implies the deliberate interference of other human beings within the area in which I could otherwise act” (122; italics added); on the other hand, he writes: “The criterion of oppression is the part that I believe to be played by other human beings, directly or indirectly, with or without the intention of doing so, in frustrating my wishes. By being free in this sense I mean not being interfered with by others” (123; italics added). I believe that this latter passage represents Berlin’s considered view, since he writes in the 1969 “Introduction” to “Two Concepts” that freedom is diminished by the “closing of doors” to possible actions that are the “intended or unintended” results of “alterable human practices” (xl). In any case, Berlin’s inconsistency doesn’t take away from his essential point: namely, that those incapacities that find their causal source in something other than arrangements of human beings do not constitute limitations on political liberty.good thing, or that providing them ought not to be one of the goals of social policy. It’s just that this is a goal distinct from the goal of preserving a person’s liberty. Berlin happily acknowledges that liberty, in the negative sense, might be of little value to a person if the opportunities it leaves them are sparse. Thus he admits that a person’s lack of ability to do the things he would like may condition the value of his liberty, even if it does not diminish his liberty itself.   361.2.3. The argument(s) of “Two Concepts”  The overarching aim of “Two Concepts” is to make the case against the adoption of the positive concept of liberty and in favour of adopting the negative concept as a guiding political ideal. However, the argument Berlin offers is complex and multi-faceted. Part of the reason for this is that, while the negative concept of liberty is presented as remaining relatively clear and stable, the positive concept is presented as taking on a variety of distinct and evolving forms.  In what follows I want to highlight what I take to 37be the three most significant strands of argument in Berlin’s essay: (i) the “inner citadel” objection; (ii) the “divided self” objection; and (iii) the objection from “value pluralism.” The first objection is best understood as an argument against what I will call the “primitive” form of the ideal of self-mastery, whereas the second and third objections ought to be read as applying to more developed versions. (i) The “inner citadel” objection  According to the positive concept of liberty, freedom consists in self-mastery. But what does it mean to be one’s own master? In its most primitive form, the desire to be one’s own master is the desire to direct one’s life and actions according to one’s own 17 Ibid., liii36 Ibid., 131-237“consciously” conceived “purposes” ; To stand in relation to one’s self as a master does 38to a slave, able to command and direct one’s actions according to ends that find their source in one’s own “reason and will.”  In short, to be free is to be able to do or be 39whatever it is you have set your will to do or be.   Berlin’s “inner citadel” objection is directed against conceiving of political freedom in terms of this primitive form of the ideal of self-mastery.  Because freedom consists in 40being able to do what you will, on this view, it follows that if you are forbidden or otherwise prevented from doing something you have not set your will upon to do, your freedom is not thereby diminished. However, and herein lies the problem, this implies that a person can increase her freedom with equal effectiveness either by removing the external obstacles to her will or, perversely, by simply repudiating the frustrated aims. I can achieve freedom not only by removing an obstacle to something I desire to do, but also by merely overcoming or “mastering” the desire to do it, by bringing my will in line with the constraints I face rather than the other way around. Even a slave can be accounted free on this view, so long as he, like the stoic Epictetus, can bring himself to be content with his chains. Freedom is thereby achieved by retreating into an “inner citadel” of desires that are invulnerable to external forces. Conceiving of political freedom in these terms is thus consistent with a disturbing form of political quietism.  (ii) The “divided self” objection  While the inner citadel argument is directed against the ideal of self-mastery in its primitive form, Berlin’s primary preoccupation is with more developed versions of this 18 Ibid., 13138 Ibid., 13539 Actually, the argument presents an objection to any view according to which freedom consists in being 40able to do what you desire to do, or not being prevented or interfered with in doing what you desire to do. ideal. He writes: “on the face of it, [“negative” and “positive” liberty] seem concepts at no great logical distance from each other -- no more than negative and positive ways of saying much the same thing. Yet the ‘positive’ and ‘negative’ notions of freedom historically developed in divergent directions not always by logically reputable steps, until, in the end, they came into direct conflict with each other.”   41 Though there is significant variation in the precise content of these developments, there are really two basic steps in the evolution of the positive concept of liberty. The first step involves a shift away from viewing impediments to freedom as exclusively external towards the view that there might also be “internal,” psychological obstacles to freedom.  Berlin writes: ? ‘I am my own master’; ‘I am a slave to no man’; but may I not (as Platonists or   Hegelians tend to say) be a slave to nature? Or to my own ‘unbridled’ passions?   Are these not species of the identical genus ‘slave’ -- some political or legal,   others moral and spiritual? Have not men had the experience of liberating    themselves from spiritual slavery, or slavery to nature...  42?The thought that one might be “mastered” by one’s own desires and inclinations, and the corresponding shift towards thinking of freedom as requiring a kind of control not only over external realities but also over one’s own motivations, according to Berlin, leads to a divided view of the self. He writes: ?19 Ibid., 13141 Ibid., 13242 ...and do they not in the course of it become aware, on the one hand, of a self   which dominates, and, on the other, of something in them which is brought to   heel? This dominant self is then variously identified with reason, with my ‘higher   nature’, with the self which calculates and aims at what will satisfy it in the long   run, with my ‘real’, or ‘ideal’, or ‘autonomous’ self, or with my ‘self’ at its best’;   which is then contrasted with irrational impulse, uncontrolled desires, my ‘lower’   nature, the pursuit of immediate pleasures, my ‘empirical’ or ‘heteronomous’ self,   swept by every gust of desire and passion, needing to be rigidly disciplined if it is   ever to rise to the full height of its ‘real’ nature.  43?So, on the one hand there is the “higher” or “true” self, and on the other there is the “lower” or “empirical” self. Self-mastery is a kind of psychological achievement of having realized one’s “true” self by bringing one’s desires and aims under the governance of reason.  Berlin identifies a second step in the development of the ideal of self-mastery, one which expands the conception of the self. He writes:  ? Presently the two selves may be represented as divided by an even larger gap:   the real self may be conceived as something wider than the individual (as the   term is normally understood), as a social ‘whole’ of which the individual is an   element or aspect: a tribe, a race, a church, a state, the great society of the living  and the dead and the yet unborn. This entity is then identified as being the ‘true’  20 Ibid.43 self which, by imposing its collective, or ‘organic’ will upon its recalcitrant  ‘members’, achieves its own, and therefore their ‘higher’ freedom.  44?Berlin no doubt has in mind especially the metaphysics of Hegel and idealist followers like T.H. Green here, though perhaps also the views of both the Nazis and the nationalist and marxist movements for self-determination among colonial peoples. On such views the self that is doing the mastering appears to be some reified social whole, rather than the individual persons that constitute it. In any case, while this second step may represent a historically significant development of the positive concept of liberty in Berlin’s time, the force of the divided self objection does not ultimately depend on it. The crucial, and from Berlin’s perspective dangerous step is the first one.   That is, once freedom is conceived of as a psychological achievement of having brought one’s aims and desires under the governance of one’s reason, it follows that individuals may fail to be free even though not coerced by external agents. Individuals who seek to act on desires and aims that are contrary to reason or to their higher nature do not act freely. This, in turn, invites the possibility of others, those better able to perceive their “true” ends, “forcing them to be free,” to use Rousseau’s paradoxical phrase. Actually, there are two related problems here: First, it follows from this conception of self-mastery that coercing or otherwise interfering with the actions of individuals in the grip of their “empirical selves” does not diminish their freedom. Second, it follows that coercing such individuals into doing what they would do if they were “rational,” or forcing them to become “rational” (perhaps in re-education camps!), amounts to increasing their freedom. These implications, Berlin insists, convert the ideal 21 Ibid.44of freedom into a license for coercion, one which at its extreme constitutes a justification for totalitarianism in the very name of “true” freedom.  (iii) The objection from “value pluralism”  The third of Berlin’s objections to the positive concept of liberty as self-mastery depends on what has been called his “value pluralism.” Berlin sees an implicit commitment to a form of rationalism -- “rationalist monism” -- about human values in the (developed) positive concept liberty. Rationalist monism is the view that there is some rationally discoverable single fundamental value or ordinal ranking of values according to which any ethical question can, at least in principle, be given a single correct answer.  Freedom, according to the (developed) concept of positive liberty, consists in the direction of one’s actions in conformity with this rational standard, rather than being dragged along by one’s empirical desires and inclinations.   But, Berlin insists, rational monism is false, and so the conception of freedom as rational self-mastery rests on a mistake. Rational monism is false because values are in fact irreducibly and incommensurably plural, and “in perpetual rivalry with one another.”  Further, it is an illusion to think that “all values can be graded on a scale, so 45that it is a mere matter of inspection to determine the highest... to represent moral decision as an operation which a slide-rule could, in principle, perform.”  This is not to 46claim that there are not some values that are more important than others; nor is it to claim that there is no such thing as an irrational choice. Instead, it is just to claim that many decisions involve choices among values between which reason cannot adjudicate, and for which, consequently, there is not a single correct or “rational” choice 22 Ibid., 17145 Ibid.46about what to do. But the positive concept of liberty, Berlin suggests, depends for its coherence on the assumption that reason speaks with a clear and determinate voice.  The truth of the doctrine of value pluralism also, on Berlin’s view, provides positive support for the negative concept of liberty. If many of the most significant decisions in life involve choices which are not susceptible to simple rationalistic resolution, then it is “more humane” to leave such decisions in the hands of the individuals directly faced with them  -- to preserve a domain of non-interference in 47which individuals may enjoy the dignity of making for themselves the choices that they will have to live with. 1.3. Liberty after “Two Concepts”  The response to the arguments of Berlin’s essay has been voluminous, and so I won’t conduct a comprehensive review here. Instead, I will focus on highlighting the leading proposals for conceiving of political freedom within the post-“Two Concepts” literature. Some of these are attempts to develop and defend alternatives to negative liberty, alternatives which go beyond the stripped down non-interference ideal, but in ways that attempt to avoid the dangers Berlin warned about in his discussion of positive liberty. Others accept Berlin’s basic point that freedom is essentially about non-interference, but develop the negative concept in distinctive ways. In what follows, I will attempt to represent a range of proposed conceptions of freedom of both sorts, ones I take to be most promising or influential. That is, I discuss the conception of freedom as (a) “empirical” non-interference; (b) non-interference with rights; (c) effective capacity; (d) individual autonomy; and (e) non-domination. In line with the methodological 23 Ibid.47principle defended above, I will consider the prospects of each of these proposed conceptions in light of the theory of freedom they are embedded in.   1.3.1 Freedom as non-interference: the empirical approach.  While Berlin’s negative concept of liberty has been much criticized, a significant number of philosophers have accepted his basic point that freedom ought to be conceived of as the absence of interference in an individual’s choices. Yet, this basic point has been developed in different ways. Perhaps the most important division is between those who define the non-interference ideal in a strictly “empirical” way and those who adopt a rights-based -- or “moralized” -- definition. The strictly empirical version holds that any interference with an action a person could otherwise perform constitutes a diminishment of her freedom; whereas, according to the moralized definition, only interferences with actions a person has a right to perform constitute a diminishment of her freedom.  With respect to the definition of liberty, Berlin clearly falls on the empirical side of the divide, since he insists that any interference, however well justified, constitutes a diminishment of an individual’s freedom. Normative judgments concerning the justification of interference ought not, on this view, to play any role in the identification of instances of freedom and unfreedom. However, when it comes to the question of how free a person is -- that is, the question of individual’s overall degree of freedom -- Berlin takes a different perspective. We might expect that assessing the extent of an individual’s freedom should be a simple matter of counting up the number of options interfered with by others and comparing this with the number of options that remain. However, Berlin does not think that a person’s overall freedom can be adequately 24judged in purely quantitative terms. In a footnote to the “Introduction,” Berlin explains that the extent of a person’s freedom is a function of (a) how many opportunities are open to him (which can only be given an impressionistic accounting), (b) how easy or difficult each of these is to actualize, (c) how important each of these are in light of the person’s plan of life, circumstances, and character, (d) how far the opportunities are closed or opened by deliberate human acts, and (e) what value not merely the agent but also the general sentiment of the society in which he lives puts on these possibilities.  48Berlin insists that, because of these multiple and interacting factors that go into determining how free an individual is, there is no hope of designing any kind of quantitative measure of freedom or even of arriving at uncontroversial and determinate judgements about the comparative freedom of different individuals (say in different political communities).  At best, we can make contestable, somewhat impressionistic 49comparisons of individuals’ extent of freedom on the whole.  However, this conclusion has been challenged in the recent work of Ian Carter. Carter maintains that the degree of a person’s overall freedom should be understood in a purely quantitative manner. Roughly: the extent of a person’s freedom is determined by the proportion of their total possible actions that are blocked by the actions of other 25 Ibid., 130 n. 148 It is easy to see why Berlin is skeptical about being able to arrive at anything more precise or 49demonstrable: He thinks there is no possibility of counting opportunities -- in part, no doubt, because of the difficulty of defining what counts as an opportunity, the sheer number of them, and the difficulty of determining when two actions are really two ways of doing the same thing and when they are genuinely distinct options. Further, not only would we need to count options, we would also have to scale each option on some metric of difficulty (where the difficulty is the result of human interference). We would also need to find a way to determine how important various options are in light of an individual’s character, plan of life and circumstances, and then design a schema for assigning different weights to options based on how important they are. And, we would also need to adjust the weightings of different options based on how strongly both the agent and his society in general values them. Such a task would not only be highly complex, it would involve too many judgements about weight-assignments which border on arbitrary.people.  Carter calls this the “purely empirical” approach to political freedom, since it 50ignores issues of the relative importance of the different action possibilities in the assessment of a person’s overall degree of freedom, as well as questions about the moral justification of the constraints. A person’s overall freedom is simply about what fraction of their possible actions are interfered with.    The “purely empirical” approach faces an obvious objection, however. Charles Taylor illustrates the problem clearly in the following passage:   ? Consider the following diabolical defense of Albania as a free country. We    recognize that religion has been abolished in Albania, whereas it hasn't been in   Britain. But on the other hand there are probably far fewer traffic lights per head   in Tirana than in London. (I haven't checked for myself, but this is a very plausible  assumption.) Suppose an apologist for Albanian socialism were nevertheless to   claim that this country was freer than Britain, because the number of acts    restricted was far smaller. After all, only a minority of Londoners practice some   religion in public places, but all have to negotiate their way through traffic. Those   who do practice a religion generally do so on one day of the week, while they are  held up at traffic lights every day. In sheer quantitative terms, the number of acts   restricted by traffic lights must be greater than that restricted by a ban on public   religious practice. So if Britain is considered a free society, why not Albania?  51?26 Carter, A Measure of Freedom, 171-250 Taylor, “What’s Wrong with Negative Liberty?”, 21951Taylor’s point seems well taken: Few people would seriously dispute the claim that the freedom of Albanians is more significantly diminished by the abolition of religion than is the freedom on Londoners by traffic lights. Moreover, it also seems plausible to think that the reason we make this judgement is that we regard restrictions on matters of fundamental human significance to be greater diminishments of liberty than restrictions on more trivial ones. But that claim contradicts the basic assumption of the pure empirical conception of freedom -- that a person’s freedom is a simple function of the number of actions interfered with.   Carter, however, resists this conclusion. He argues that, while a person may value certain “specific” freedoms more than others on grounds that the activities they leave her free to engage in are of special importance, this ought not to affect our judgement about her overall degree of freedom. The value of a specific sort of freedom is one thing; its weight in the calculation of a person’s overall degree of a freedom is another. (I discuss Carter’s empirical approach in further detail in Chapter 6).  From the perspective of a theory of freedom, the empirical non-interference conception is incomplete in the absence of a distributive principle or principles. A simple maximization principle would seem to be unsatisfactory since, as with maximizing forms of utilitarianism, it is in principle consistent with vastly unequal distributions of negative liberty. Steiner favours an equal freedom principle, which requires that each individual be accorded the same degree of non-interference.  On the face of it, this principle is 52also unattractive, since requiring merely equal shares of negative liberty is consistent with each member have little or even no freedom -- so long as their degree of freedom 27 Steiner, “The Natural Right to Equal Freedom”52is equal.  A more promising principle of distribution advocated by, among others, 53Narveson , is the principle of maximum equal freedom. Here each individual is to be 54accorded the maximum possible degree of non-interference possible, consistent with every other individual having the same degree. Yet, we might wonder whether it would be better to permit some inequalities in degrees of freedom if doing so would make even the least free more free than in a condition of equal freedom (maximin freedom) or to make each person as free as possible starting with least free and working our way up (leximin freedom).    Berlin appears to favour something like a guaranteed minimum principle of freedom, where each person ought to be according a certain basic minimum degree of non-interference. However he nowhere offers an account of what that basic minimum ought to include, instead preferring to simply to defer to the evolving list of “liberties” advocated by historical liberalism. Part of the reason for this is, no doubt, his belief that degrees of freedom can in principle only be assessed in an impressionistic way. Given his rejection of the possibility of a quantitative measure of a freedom, talk of maximum or maximum equal non-interference lacks any clear and distinct meaning. A further, and more significant reason for Berlin’s guaranteed minimum principle is that while freedom is a fundamental political value, it is not the only political value -- he even allows that it 28 Steiner’s justification for the principle depends in part on his belief that the total amount of freedom in 53any given society is fixed, and that the only point on which variation is possible is on the question of distribution (1994: 52-3). On such an assumption, a principle of simple equality is not implausible, since equal freedom is necessarily equivalent to maximum equal freedom. However, this assumption is surely mistaken. A simple illustration is sufficient to demonstrate this fact: Suppose there are three individuals, each with loaded guns and hold up in trenches. Each individual is able to shoot either of the others if they step out of their trench, and each has issued a threat to do so if they do step out. The degree of interference in the action possibilities of each individual is very high -- the vast majority of actions they could otherwise perform are blocked. Now compare this scenario to one in which the three individuals are not armed and have not issued any threats. It seems obvious that the aggregate degree of non-interference will be less.  Narveson, “Equality versus Liberty: Advantage, Liberty”54may not be the most important one. In consequence, adopting a principle of maximum or maximum equal freedom (or maximim or leximin freedom for that matter) would come at too great a cost to other goods we have reason to value.  On this point, Carter 55agrees: he claims that freedom is one -- but only one -- on the goods that a theory of justice must provide principles of distribution for.  For this reason, the optimizing 56principles must be rejected, and some other principle which makes room for restrictions on freedom for the sake of distributing other important goods must be adopted instead.  1.3.2. Freedom as non-interference: the rights-based approach  In contrast to the pure empirical approach, some authors have advocated a conception of freedom according to which a person is free to the extent that they are not interfered with by other human beings in actions they have a right to perform. This conception is, thus, “moralized” in the sense previously explained.  The rights-based approach is most widely associated with Robert Nozick  and with libertarianism more 57generally.  However, it is also advocated by non-libertarian liberals, including, e.g., 58Ronald Dworkin.   59 Dworkin argues that Berlin’s view that any interference, however well justified, counts as a diminishment of a person’s liberty is implausible. “Suppose I want to murder my critics,” Dworkin notes. “The law will stop me from doing that, and the law will 29 Berlin does not offer anything like an overarching principle for how these various values, including 55freedom, ought to combined in a just or well-ordered society. In fact, he appears to view the attempt to offer such a principle as stemming from a dangerous and misconceived urge for a “final solution” based on the conviction that all that is good can be combined in a “harmonious” whole -- i.e. as based on the mistaken assumption of rational monism. A Measure of Freedom, 68-956 Nozick, Anarchy, State, and Utopia57 cf. Rothbard, The Ethics of Liberty58 Dworkin, “Do Liberal Values Conflict?” 59therefore, on Berlin’s account, compromise my liberty.”  That is a highly counter-60intuitive implication, on Dworkin’s view. Thus he proposes the following:  ? ...liberty isn’t freedom to do whatever you might want to do; it’s freedom to do   whatever you like so long as you respect the moral rights, properly understood,   of others. It’s freedom to spend your own rightful resources or deal with your own  rightful property in whatever way seems best to you. But so understood your   liberty doesn’t include freedom to take over the resources of someone else, or   injure him in ways you have no right to do.  61?Though this rights-based conception of freedom has been subject to criticism, including from liberal quarters, it is arguably consonant with the main current of the liberal tradition, which emphasizes the centrality of individual rights, and closely associates liberty with enjoyment of them. Liberals regularly use the terms “liberties” and “rights” interchangeably, and historically have tended to defer to “rights” in marking off the appropriate zones of non-interference individuals ought to enjoy.    62 Now, from the point of view of a theory of freedom, the rights-based conception is incomplete in the absence of a specification of the content of those rights. Dworkin holds that there is not a general right to non-interference as such, and in fact that non-30 Ibid., 8860 Ibid., 9261 Italian liberal political theorist Norberto Bobbio goes so far as to claim that “[t]he philosophical 62presupposition of the liberal state... is found to be in the doctrine of natural rights” (2005: 5). And from a more hostile point of view, Jurgen Habermas concurs: “According to the liberal view, the citizen’s status is determined primarily by the individual rights he or she has vis-a-vis the state and other citizens”; moreover, these “individual rights are negative rights that guarantee a domain of freedom of choice within which legal persons are freed from external compulsion” (2005: 240-241).interference as such is not valuable in itself.  Instead, he holds that individuals have 63rights to equal specific liberties -- i.e. freedom of thought and conscience, speech, peaceful association, freedom to own and exchange private property, etc. These specific liberties, in turn, are grounded in and given definition by reference to his core principle of justice that every person is entitled to equal concern and respect in the design of the structure of society.  In Anarchy, State and Utopia, Robert Nozick also implicitly endorses a rights-based definition of negative liberty. Nozick defines the content of an individual’s rights as follows: First, each individual has, initially at least, a right to self-ownership, i.e. to full “control” ownership of their self, including their body and powers. Second, each individual can come to have ownership rights over things in the external world in accordance with the “entitlement theory of justice.” On the entitlement theory of justice, a person can come to have an ownership right over a certain “holding” either through transfer or by original acquisition. In the former case, a person has a right to the holding just in case, roughly, it was voluntarily transferred from another person who had the ownership right to it. In the latter case, a person can acquire an original right to a holding by mixing their labour with a part of the external world, provided that it is not already owned by someone else and provided that no one who could previously have used that now-owned part of the world is made worse off by the acquisition. In light of Nozick’s account of the content of individual rights, we can say that a person’s freedom is violated iff they are interfered with by other human beings in doing (or not doing) what they have a right to do (or not do) in accordance with the principle of self-ownership and 31 cf. Kymlicka, Contemporary Political Philosophy, 14363the entitlement theory of justice in holdings. Interferences in actions that an individual does not have a right to perform do not count as violations of freedom.    Nozick’s rights-based conception of freedom has been challenged on various grounds. Some authors endorse the rights-based approach in general, but object to Nozick’s account of the original acquisition of rights to ownership in external objects. For instance, some so called “left-libertarians” hold that an individual can acquire an original private ownership right to a particular resource only so long as she leaves an equally valuable share of that resource for others  or so long as she leaves enough for others 64to have an opportunity for well-being that is at least as good as the opportunity for well-being she obtained in using or appropriating the resource.    65 Other authors reject the definition of freedom in terms of rights altogether, objecting to the “essential moralization” of the concept of liberty. As David Zimmerman explains, a concept is essentially moralized if its “conditions of application contain an ineliminable reference to moral rightness or wrongness.”  The rights-based version of 66the ideal of freedom as non-interference is certainly moralized in this sense, and as such has been subject to a variety of objections.   Some critics have alleged that, by moralizing the concept of freedom, the rights-based ideal denies the independent value of “freedom” (i.e. non-interference) itself. It turns out, the objection goes, that what one values is not the ability of individuals to act 32 cf. Steiner, An Essay on Rights64 f. Otsuka, Libertarianism without Inequality65 Zimmerman, “Taking Liberties: The Perils of ‘Moralizing’ Freedom and  Coercion in Social Theory and 66Practice,” 577free from the interference, but rather only the protection of property rights. The value of non-interference as such drops out of the picture altogether.  67 Other critics have charged proponents of the moralized conception of freedom with a kind of circularity. G.A. Cohen has leveled this charge against libertarians, and Nozick in particular: ? Libertarians want to say that interferences with people’s use of their private   property are unacceptable because they are, quite obviously, abridgements of   freedom, and that the reason why protection of private property does not similarly  abridge the freedom of non-owners is that owners have a right to exclude others   from their property and non-owners consequently have no right to use it....   Thereby Nozick locks himself in a circle. For Nozick, there is justice, which is to   say no violation of anyone’s rights, when there is lack of coercion, which means   that there is justice when there is no restriction on freedom. But freedom is then   itself defined in terms of non-violation of rights, and the result is a tight    definitional circle and no purchase either on the concept of freedom of on the   concept of justice.  68?Cohen’s objection, then, is that the libertarian position is circular because on the one hand it claims to offer a conception of justice whose chief virtue is that it places priority on the preservation and respect for individual freedom, and yet on the other hand 33 cf. Christman, “Review of Philip Pettit’s A Theory of Freedom and Government,” 203; Lovett, “What 67Count as Arbitrary Power?”, 142; Carter, “How are Power and unfreedom Related?”, 71-72, 142 Cohen, Self-ownership, Freedom and Equality, 60-168defines freedom in terms of a conceptually prior conception of justice (from which rights proceed).  69 A further objection to the moralized conception of freedom is that it leads to implausible conclusions. As Cohen says, “it entails that a properly convicted murderer is not rendered unfree when he is justifiably imprisoned.”  Such a conclusion, he insists, 70flies in the face of ordinary use of the word freedom, according to which there can be both just and unjust restrictions on freedom. I discuss each these objections in more detail in Chapter 6.  1.3.3. General objections to the negative concept of liberty  On a more general note, many philosophers find the whole approach of conceiving of freedom in terms of non-interference -- whether empirical or rights-based -- problematic. Though I’ll discuss the basis of these objections in more detail in subsequent sections, it is worth mentioning them at this point.   Some object to the negative concept of liberty along the following lines: What good is it if no one interferes with you in doing some thing if you aren’t actually able to do that thing? Suppose you haven’t been afforded the opportunity to acquire basic literacy skills. The mere fact that no one prevents you from picking up a book isn’t sufficient, surely, to render you free to read whatever you like, or to do those things for which reading is a prerequisite. Being free, it seems, is better understood in terms of ability, as having a certain set of capacities or opportunities, rather than merely as the absence of interference.  34 For a similar conclusion about rights-based conceptions of freedom, see Christman, “Review,” 20369 Cohen, “Illusions about Private Property and Freedom,” 22870 Others object that conceiving of freedom in terms of non-interference falls short of capturing what ultimately makes individual freedom valuable -- namely, the individual’s capacity for authentic or autonomous self-direction. The mere fact that no one coerces you is no guarantee that your actual choices will reflect aims and purposes that are genuinely your own. You may be in grip of some irrational fixation, or some maladjusted emotional disposition, or some belief-system that is the result of one’s unreflective or oppressive socialization. In other words, there may be “internal obstacles” that render your actions unfree, even though they are not subject to the interference of other human beings. True freedom must consist in the realization of your authentic aims, the direction of your life according to your own autonomous will. The absence of interference, then, is at least not a sufficient condition for freedom. And it may not even be a necessary condition: When Odysseus had his crew tie him to the mast in order that he would not, upon hearing the Sirens’ song, be lured by the intoxicating music to his death, he was surely thereby subject to interference; yet that interference served to preserve rather than diminish his freedom.   Still others object that the non-interference conception is blind to the way the mere vulnerability to the power of another agent to interfere arbitrarily in one’s activities diminishes one’s freedom. Even if a particular powerful agent happens not to interfere in certain choices an individual makes, the mere fact that that agent could do so at his pleasure places the vulnerable party in a condition of being dominated. Individuals subject to domination, even when not actually interfered with, will tend to self-edit their actions in anticipation of how the dominant party might react. And even if they don’t self-35edit, the mere fact that the dominant party possesses the capacity to invigilate and intervene at will places the vulnerable party in a condition of being under his control.   1.3.4. Freedom as effective capacity   Those impressed by the force of the first of these objections insist that the idea of freedom must be more closely associated with the notion of ability or capacity -- sometimes called “effective freedom.”  As noted, Berlin maintained that the idea of 71political liberty should be kept entirely distinct from that of ability, since liberty is essentially about relations between human beings, not about their natural endowments. However, in the post-“Two Concepts” literature, both Phillip van Parijs and Amartya Sen have advocated versions of the conception of freedom as capacity.  In his book Real Freedom for All, van Parijs argues that the negative concept of liberty is an impoverished ideal and that “real” freedom consists not merely in the absence of coercion but in actually be able to do the things you might want to do. At points van Parijs seems to equate negative liberty with “formal freedom” -- i.e. with being permitted (by law) to do a particular thing; as I’ve already noted, this is a mistake. One can be permitted by law to something and yet still be interfered with in doing it (and vice versa). Nonetheless, van Parijs’ guiding conviction does not depend on this conflation, but consists in the thought that being free must mean actually being able to do the thing you are free to do.   van Parijs is unflinching in his commitment to this capacity-based conception of freedom. He insists that any obstacle at all to a person’s doing something counts against their freedom -- it doesn’t matter whether or not the obstacle was deliberately imposed by other humans or not, it doesn’t matter whether it was even caused by other 36 Sen, Development as Freedom, 1971humans, in fact it doesn’t even matter whether other humans are capable of doing anything to remove the obstacle.  Thus van Parijs holds that a person’s freedom is 72diminished if he is unable to swim across a lake because his lungs would give out or even if he is unable to travel at the speed of light because the laws of physics preclude it.    73 van Parijs defines a “free society” as one which meets the following three conditions:  (1) There is some well enforced structure of rights (security).  (2) This structure is such that each person owns herself (self-ownership). (3) This structure is such that each person has the greatest possible opportunity to do whatever she might want to do (leximin opportunity).  74?The first condition requires that there exist a system of law that accords to each individual a set of rights that ensures the security of their freedom. The second condition requires that this set of rights reflect the principle that each individual owns themselves, in contrast to the claim of collectivism that each individual is owned by the society as a whole. Self-ownership, on van Parijs’ view, comes in degrees depending on the extent of things a person is able to do with herself.  The third condition requires that the 75system of rights be designed so that each person is afforded the greatest possible 37 Real Freedom for All, 2372 Ibid.73 Ibid., 2574 Ibid., 675extent of opportunity to do the things she might want to do with herself.  This condition 76is to be understood more precisely as leximin opportunity:  ? ...in a free society, the person with least opportunities has opportunities that are   no smaller than those enjoyed by the person with the least opportunities under   any other feasible arrangement; in case there exists another feasible    arrangement that does just as good for the person with least opportunities, then   the next person up the scale in a free society must have opportunities no smaller   than the second person up the scale of opportunities under this arrangement,   and so on.  77?38 van Parijs claims that there can be clashes between the three conditions, and that a full 76characterization of the free society would require specification of standards by which these conflicts should be settled (25). However, he insists that a defeasible or “soft” priority should be given to the first over the second and the second over the third conditions (26).  However, these latter claims -- that the three conditions can conflict and that trade-offs might be necessary -- are puzzling. Talk of conflict, priority and trade-offs would make sense if the three conditions embodied three independent values; however, this does not appear to be the case here.  According to van Parijs, the extent of a person’s “self-ownership” is determined by the range of things she is able to do with herself (6) -- i.e. by the size of her opportunity set. Thus, the third condition (leximin opportunity) is really a specification of the distributive principle according to which the extent of each individual’s self-ownership (the value enshrined in the second condition) is to be determined, not a distinct value consideration. Further, self-ownership (the second condition) is supposed to be the value which the system of legally-enshrined rights required by the first condition is designed to secure and protect. And so the second and first conditions are also not intelligibly viewed as competing values, requiring prioritizing or trade-offs. It is perhaps possible to see a kind of conflict between the first two conditions -- between the degree of an individual’s self-ownership and its security. On this reading, van Parijs might be claiming that we should opt for a somewhat more restricted set of opportunities that are protected by a stable regime of legal rights over a more extensive set of opportunities that lack this security. However, even if this is his intended meaning, it is difficult to see this as giving priority to the first over the second condition, since that would have the absurd consequence that the security of liberty ought to have a higher priority than liberty itself. It is even more difficult to see what sort of conflict he envisions between the second and third conditions. Obviously, for any given individual it will be the case that their opportunity set (which constitutes their degree of self-ownership) is less extensive on a leximin distribution than it would be on some other distribution (e.g. one in which all the resources available are devoted to enhancing their personal opportunities). In this sense, there will clearly be conflicts between self-ownership and leximin opportunity. But if this is the kind of conflict van Parijs has in mind, it is unclear why he would assign priority to self-ownership over leximin opportunity -- since this would effectively negate his commitment to leximin distribution. (In fact, it would render his conception of a free society incoherent, since one cannot consistently maximize the opportunity set of every individual)     Ibid., 2577 van Parijs’ view, then, is that “real” freedom consists in the ability to do what one might want to do, that the extent of an individual’s freedom is determined by the size of their opportunity set, and that a free society is one in which opportunities are distributed in accordance with the leximin principle.   One objection to van Parijs’ view is that his distributive principle severs all connection between the opportunities people have and the contribution they make to producing those opportunities. This has seemed to many to be problematic both from the point of view of justice (why should “lazy” unemployed surfers receive the same benefits as hard working teachers?)  and, in light of its incentive structure, from that of 78productivity (under such a scheme, mightn’t some of those hard-working teachers choose to hit the beaches instead?). With respect to the latter point, van Parijs’ distribution principle may be at odds with the basic impulse of his theory of freedom -- namely to provide as much opportunity for everyone as possible -- insofar as the incentives it provides will tend to reduce the total opportunities available for distribution.    However, there are other worries that stem more directly from his conception of individual freedom, rather than his distributive principles for a free society. An individual’s freedom is diminished, according to van Parijs’s definition, to the extent that he is unable to do the things he might want to do -- regardless what the source of that inability is. As noted, this approach carries with it the counter-intuitive implication that inabilities which have their source in the laws of nature count as diminishments of a person’s freedom. If I am unable to travel at the speed of light or leap tall buildings in a single bound, I am to that extent less free. van Parijs insists that while using the word “free” in this way involves some “stretching,” it is still “fully intelligible”; and, indeed, that 39 Gintis, “Review of Real Freedom for All: What (if Anything) Can Justify Capitalism?”, 182 78“stating that I am not free to travel faster than the speed of light is only slightly odd, if at all.” However, it seems to me that the plausibility of this reply depends on an equivocation between two distinct senses of the word “free” in ordinary usage, between its meaning in a specifically political context (i.e. political freedom) and its meaning in a broader sort of context. It is perfectly intelligible to say of a person whose leg has just been released after being trapped under the weight of a fallen tree that he is now “free.” But it would indeed be a stretch to say that his political freedom had been diminished by the tree. After all, we would use the word “free” in the same sense if the lifting of the tree had “freed” some inanimate object that had been caught underneath it, though we would not describe that object as thereby being made politically free. Underlying this difference, I believe, is the fairly entrenched presumption that political freedom concerns the character of our relations with other intentionally-acting human beings.   However, even if van Parijs were to alter his conception of individual freedom so that only those inabilities that other human beings could conceivably remedy count as constraints on freedom, it still carries what seem to be counter-intuitive implications. If I have a kidney disorder that prevents me from doing certain things that I might want to do, it would be a stretch to claim that you have diminished my freedom by not donating one of your healthy kidneys to me, or to claim that my political freedom has been infringed because there are people out there that could remedy my situation. It would be even more counter-intuitive to claim, as van Parijs’ view appears to commit him to doing, that there would be no difference with respect to a person’s freedom whether he was unable to do certain things because his kidneys were failing (and no one donated a healthy one) or was unable to do certain things because the health authority had “re-40assigned” his functioning kidneys to someone else. Underlying these intuitions is the conviction that mere inability -- even humanly-remediable inability -- does not constitute a diminishment of freedom in the sense relevant to politics.   Like van Parijs, Amartya Sen has argued for a conception of freedom that ties it more closely with the notion of ability or capacity. However, unlike van Parijs, Sen conceives of freedom as consisting in both opportunity and process aspects. The opportunity aspect of freedom consists in the set of capacities a person has to perform certain actions or “functionings” (her abilities or “real” freedoms, in van Parijs’ sense). The process aspect of freedom consists in avenues a person has to participate in the political process by which the community is governed and, in particular, by which opportunity-distributing policies are determined, as well as “negative liberties” for individuals not to be interfered with in matters of exclusively personal concern or right. Sen’s view, commonly referred to as the “capabilities approach,” is thus best understood as a hybrid of the non-domination (see 1.4.6 below), non-interference and the effective capacity conceptions of freedom. “Unfreedom” can arise either through inadequate processes (as when there are violations of political rights or civil liberties) or through the inadequate opportunities people have to realize their basic capacities (as when they lack access to health care, security, education, food, work, etc.).   As attractive as Sen’s view may be as a normative account of development, where it has been most influential, it faces a serious problem as a theory of political freedom. How we are to understand the relationship between the opportunity and process aspects of freedom -- and, indeed, between the non-interference and participatory elements of the process aspect? Would an expansion of opportunity by 41non-democratic means still count as any increase in freedom? Would an increase of democratic control at the expense of individual “negative” liberty? How would we judge the relative freedom of citizens in a democratic, but opportunity-poor country compared to those in an opportunity-rich but non-democratic one? Would trading-off process aspects of freedom to achieve gains in opportunity aspects count as a violation of freedom or an enhancement of it? What if it turns out that democratic decision-making leads to policies that are sub-optimal for the provision of opportunities or fair distributions of opportunities?   Underlying these various questions are really two difficulties: First, it is unclear that Sen really has offered a single conception of freedom, rather than just adjoining three distinct conceptions. After all, what is the underlying ideal that the self-government, negative liberty and opportunity conditions can plausibly be seen as different aspects of? If someone defined justice as a social arrangement in which social utility is maximized and in which benefits are distributed according to merit and in which people enjoy equality of opportunity, we could rightly complain that no single underlying principle or ideal could give coherence to such a conception. Likewise, by defining freedom as a condition in which individuals are accorded a broad set of opportunities for action and in which they have been able to participate in the decision-making process by which opportunities are distributed and in which individual negative liberties are protected, is Sen not merely conjoining fundamentally distinct ideals? Second, even if we do not think this conjunctive approach is merely ad hoc, it is unclear what the relation of priority between the different aspects of freedom is supposed to be or what principles might be employed in cases of conflict.     421.3.5. Freedom as personal autonomy  A second class of alternatives developed in opposition to negative liberty are those conceptions that define freedom by reference to ideals of individual “self-realization,” “authenticity,” or “autonomy.” Common to such conceptions is a conviction that, for an agent to be free, it is not enough that there be an absence of external obstacles (human interference, lack of opportunity) with their action possibilities. Their choices must be the product of their “real will” -- their rational, true, or authentic self. Freedom is in part a psychological achievement of “self-government” or “self-creation”. I shall call conceptions of freedom that have this general form “autonomy-based conceptions of freedom.”  Charles Taylor argues along these lines in an influential essay. Negative liberty, in restricting the sort of conditions that can undermine freedom to external obstacles, fails to take account of the way internal factors can undermine a person’s realization of their own deepest purposes and aims. In the simplest cases, an irrational fear, spiteful temperament, or attachment to creature comforts may prevent us from pursuing the career we want, maintaining relationships we value, or following through on a life project. These cases involve the thwarting of consciously held aims by our own wayward desires, emotions, and habits -- but, crucially, in the absence of any external preventing conditions.   One way to make sense of this is by reference to what Harry Frankfurt has called first and second order desires. First order desires are those inclinations and preferences one has for things in the world other than one’s desires themselves -- e.g. the desire to own a new car, find a soul mate, stop that annoying dripping sound, eat chocolate, 43become fabulously rich, or master French.  Second order desires are the desires one has about one’s first order desires, i.e. desires about which first order desires one wishes to influence one’s actions. Thus I may have a first order desire to watch M.A.S.H. re-runs all day, but a second order desire not to be moved by this first order desire but instead to be moved by my competing first order desire to clean the house or be a productive person. A person is autonomous, on this account, to the extent that the desires they act on are the ones they desire to act on. Incorporating this conception of autonomy into a conception of freedom would yield something like: A person is free to the extent that (a) he is effectively motivated to act only on the first order desires that his second order desires endorse AND (b) there are no external obstacles to his acting on the basis of these (endorsed) desires.  Taylor however appears to want to go further and suggest that some internal barriers may interfere with an agent’s self-realization without his even recognizing them as barriers -- e.g. if he fails to recognize his own most important purposes. Translating this claim into the language of first and second order desires, Taylor seems to claim that acting on a particular first order desire may be non-autonomous even though the agent has no second-order desire not to act on it and may have a second order desire to act on it. Thus, it is not enough that one act on the desires one wants to act on, one must also want to act on the desires one ought (objectively) to want to act on. Hence Taylor claims that Charles Manson is plausibly understood as unfree even though he clearly identifies wholeheartedly with his homicidal desires. Taylor thus proposes the follow alternative to the negative conception of freedom: A person is free to the extent that she (a) is able to recognize adequately her most important purposes, (b) is able to 44overcome or neutralize any motivational fetters, and (c) is free of any external obstacles to realizing such purposes.  Perhaps the most clearly articulated and influential version of the autonomy-based conception of freedom is John Christman’s. The sorts of internal barriers Christman focuses on are those where the desires, preferences, and character by which a person governs her actions have been conditioned by oppressive processes. Christman gives an example of a woman who is raised in a culture which fiercely inculcates in her the idea that women should never aspire to be anything but subservient and humble domestic companions to their husbands, no matter how unhappy this makes them or how abusive their husbands are. Even after this women moves to a new cultural setting in which opportunities abound for women, she remains with her abusive husband and, indeed, does not wish to act in any other way. Her socialization has been effective enough to shape her very desires and sense of self. Such a woman, Christman insists, is not free -- even after any external barriers to her possible actions have been removed.  Christman’s view of autonomy, however, differs from both the Frankfurt-style account and the more externalist-leaning account of Taylor. On Christman’s view, the Frankfurt-style view is problematic because second order desires can be every bit as inauthentic as first order ones. The women in the example just given is a case in point, since she doesn’t remain loyal to her husband out of a weakness of will, but because she identifies with the values of domestic subordination. The problem with the Frankfurt-style account is that it is not sensitive to the way that even higher order desires can be the product of manipulation or social oppression. On the other hand, Christman rejects 45the externalist solution to this problem proposed by Taylor in the Charles Manson case; an account of personal autonomy, on his view, should remain neutral between different conceptions of the good life.   Thus Christman proposes a conception of autonomy that is procedural, internalist and historical. Roughly: a person is autonomous (relative to a particular aim) if he would not repudiate that aim were he to be in a position to rationally reflect on the historical processes by which that aim was formed. This account is procedural because whether an aim is autonomous does not depend on the content of the aim (contra Taylor), but only on the procedure by which the aim was formed. The account is also internalist: The rationality required is simply the capacity to judge the processes and aims on the basis of the rest of one’s own values or “motivational set” -- not on the basis of some objective standard of “rational ends” or “important purposes.” And finally, Chistman’s account is historical: Whether or not an aim is autonomous depends partly on its causal history, not merely on the structure of the will as on the Frankfurt account.   A qualification is in order, however -- one that turns out to be quite consequential. Christman does not require an agent to have actually reflected on her aims in the prescribed manner, but only that, were she to do so, she would not repudiate them. Only hypothetical non-repudiation is required. Requiring actual reflection on each desire and choice for it to count as autonomous would be impossibly demanding and, moreover, would betray an overly-cognitive bias. Life choices can be wholehearted and authentic without going through the mill of ratiocination (Christman, 2009: 213).  Drawing these points together, Christman’s autonomy-based conception of freedom can be captured as follows: A person is free to the extent that (a) her aims are 46such that she would not repudiate them were she to rationally reflect upon them in light of the causal process by which they were formed and in light of her other values, beliefs and aims AND (b) there are no external obstacles to her acting on the basis of these aims.     Autonomy-based conceptions of freedom are clearly versions of the positive concept of liberty criticized by Berlin, and so any such account must address Berlin’s three fundamental objections -- (1) the “inner citadel” objection; (2) the “divided self” objection; and (3) the objection from “value pluralism.” In my view, none of them can adequately defuse all three.   A conception of freedom that incorporates the Frankfurt-style account of autonomy can avoid the value pluralism objection and the most troubling aspects of the divided self objection, but is particularly vulnerable to the inner citadel objection. A Frankfurt-style account doesn’t define autonomy as the governing of one’s actions by some objective (externalist) standard of rational ends, but only as the governing of one’s actions according to one’s own second-order desires -- whatever they happen to be. This view is clearly consistent with the value pluralism thesis. A Frankfurt style account also avoids the form of the divided self objection that most worried Berlin, since it does not license coercion of individuals on the basis of some standard of rational ends external to those individuals’ own subjective motivations. It would license coercion of individuals if doing so would accord with their own second-order desires -- e.g. forcing an unhappy drug addict into recovery; but this, at worst, would support mild forms of paternalism. However, the inner citadel objection is crippling for the Frankfurt style account, since on this account an individual’s freedom can be increased with equal 47effectiveness by altering her desires to accord with the obstacles she faces rather than removing the obstacles.    Taylor’s externalist version of the autonomy-based conception of freedom, by contrast, avoids the inner citadel objection, but is vulnerable to the divided self objection. It avoids the inner citadel objection because a person’s freedom consists in her being unobstructed in acting on her most important purposes, regardless whether those purposes receive recognition in her second-order desires. Simply altering a person’s desires will do nothing to alter her true purposes. Yet, because it ties the notion of freedom to the realizing of one’s “true” (i.e. objectively determined) purposes, it does not count as diminishments of freedom the coercion of individuals who fail to recognize or even reject such purposes for themselves. Taylor attempts to avoid the latter implication by claiming that the true purposes for each person will be unique to that person, and so epistemically inaccessible to, e.g., state agencies unfamiliar with the particularities of that person qua individual.    However, this is a puzzling and seemingly self-defeating reply. Any serviceable conception of political freedom must at least serve as a basis for a norm by which the legitimate use of the coercive power of the state can be delimited. If we define freedom as the absence of obstacles to a person’s autonomous aims, but then insist that what those autonomous aims are is epistemically inaccessible, how can we arrive at any clear criteria for determining which actions the state may interfere with without diminishing individual freedom and which they may not? Taylor’s externalist version of the autonomy-based conception of freedom thus faces the following troubling dilemma: Either the true purposes of individuals are publicly accessible (in which case the logic of 48coercing individuals in the name of their subjectively unrecognized true purposes, and thus their freedom, remains intact) or they are not publicly accessible (in which case we are left without a socially-determinable criterion for identifying infringements of freedom).   Unfortunately, Christman’s version of the autonomy-based conception of freedom, while it has certain advantages, is ultimately vulnerable to the same worrisome dilemma. One advantage is that, given his procedural and internalist account of autonomy, it is perfectly compatible with the doctrine of value pluralism. Moreover, his view is also well-positioned to defuse the force inner citadel objection. For Christman, it is not sufficient to count as free that a person’s actual desires -- even the second order ones she happens to identify with -- are not frustrated by external constraints. Instead it must be the case that the aims she would have if she had not been subject to oppressive or manipulative socializing forces are not frustrated by external obstacles. Thus a “contented slave” is not free if the source of his contentment is a resignation to his fate in the face the constraints he faces or the product of oppressive socialization.   However, Christman’s view, like Taylor’s, appears unable to provide a satisfactory response to Berlin’s divided self objection. That is, on Christman’s view only those external obstacles that constrain the autonomous aims of individuals diminish their freedom. Constraints on the actions of non-autonomous agents do not diminish their freedom, even if these constraints frustrate their actual self-identified desires. Moreover, given the rather demanding conditions Christman places on autonomous aims -- they must survive deep reflection on the causal source of one’s aims -- many actions of sane adult persons will fall short of being autonomous, and thus interfering with them will by 49definition not diminish their freedom. This appears to open the door wide to the licensing of coercive state action in the name of true freedom.   Christman, like Taylor, anticipates this objection and, again like Taylor, rejects this conclusion on grounds that state agencies lack the epistemic perspicuity to discern what the autonomous aims of particular individuals might be. In Christman’s case this reply is at least less ad hoc; Given the internalist character of his account, it is not unreasonable to think that what will count as an autonomous aim will be unique to individuals, and relativized to their perspectives and constellation of other values. Hence, it is also not unreasonable to conclude that states lack the ability to make such highly discriminating judgements, and so cannot justifiably employ coercive policies on the basis of such judgements. However, the difficulty with this response is that it lands Christman squarely in the second horn of dilemma raised against Taylor’s view -- namely that if the autonomous aims of individuals are not publicly accessible, then we are left without a socially-determinable criterion for identifying infringements of freedom.  Even if an autonomy-based conception of political freedom could manage to overcome Berlin’s three objections, from the perspective of a theory of freedom, there would still be a need to provide an account of a free society. Because even the autonomous aims of individuals are bound to conflict, some principle or principles will be required to determine who ought to be free to do what. I return to this point in Chapter 2.  1.3.6. Freedom as non-domination  In my view the most compelling alternative to the negative concept of liberty has been advanced by advocates of what is commonly called neo-Roman republicanism. According to this view, freedom ought to be conceived of as the absence of subjection 50to the arbitrary power of another agent. Unfreedom, then, is fundamentally about being under another person’s power, rather than about the number of interferences one faces.  But not just any kind of power -- power that is arbitrary, that is exercised at the discretionary will or pleasure of the power-holder. When a person is subject to the arbitrary power of another they are, to use Philip Pettit’s phrase, dominated.   The paradigmatically dominated person is the slave, and a consideration of the nature of the slave’s condition helps to illustrate the contrast between the conception of freedom as non-domination and negative liberty. A slave is the property of his master, who may direct the slave to do or not to do what he wishes in virtue of the power he holds over him. His power over the slave is arbitrary in that he may exercise it in whatever manner he pleases. Now masters can be more or less benevolent towards their slaves and can take a more or less permissive approach to the management of the slave’s life. A “house slave” may be granted considerable latitude for acting and choosing according to his own judgement, and may be graced with special benefits not accorded to other slaves, including time for leisure and personal pursuits. But these benefits are privileges given at the discretion of the master, and liable to be withdrawn at any time should it so please that master. For this reason, the “house slave” remains in every bit as much a condition of domination as the less fortunate and more heavily coerced “field slaves,” though he enjoys a lesser degree of interference in his daily activities.  The ideal of non-domination thus takes a contrasting view of the essence of freedom, and in fact renders different judgements concerning the freedom or lack thereof of individuals in different conditions. As Pettit argues, on the conception of 51freedom as non-domination, an absence of interference is neither a necessary nor a sufficient condition for freedom. It is not necessary since, as with the relation between the master and his “house slave,” one person may possess a power to interfere arbitrarily in the choices of another, but decline to actually interfere. This is also the case, Republicans have historically insisted, in relations between absolute monarchs and their subjects. While such rulers may at times leave their subjects more or less un-harassed, these subjects enjoy this absence of interference only at the discretion and by the leave of the ruler, and so are not what republican writers historically called ‘freemen.’  The absence of interference is also not a sufficient condition for freedom. Only interferences issuing from an arbitrary power are instances of domination; interferences made on a non-arbitrary basis do not diminish liberty. Thus, when a police officer, acting within the law, exercises his powers of arrest and detainment to interfere with the plans of a would-be murderer, he does not thereby diminish that person’s political freedom.  Part of the appeal of this conception of freedom as non-domination is that individuals subject to an arbitrary power will tend to self-edit their actions, even in the absence of actual interference. Because it is in the nature of arbitrary power that its possessor exercises her power at her pleasure or will, it tends to be unpredictable. This is especially true in cases where the power-holder is a single individual -- as with the absolute monarch or the slave-holder. In such cases, a change of mood, an alteration of perceived interests, a passing fancy, a paranoid resentment, or any number of changing facts about the power-holder’s motivations might alter her disposition towards her subjects. This places subjects in a perpetual state of uncertainty, and may well cause 52them to self-edit their actions in various ways: They may seek to fawn and toady or otherwise ingratiate themselves to the powerful in order to stay in their good graces, or they may avoid actions which might draw attention or cause offence, seeking to stay under the radar.   However, much turns on what is meant by “arbitrary power.” Without an account of which forms of power are arbitrary and which are not, the neo-Roman conception of freedom as non-domination is incomplete. According to Pettit’s now standard account, arbitrary power is power that is not forced to track the avowed interests of those subject to it. Two features of this account bear emphasizing. First, in order for power to count as non-arbitrary, it is not enough that the power-holder happens to use that power in way that aligns with the interests of persons subject to it, he must be forced or constrained to do so. There must be some mechanism by which to ensure that such power is exercised in way that is accountable and responsive to the interests of those subject to it. Second, it is the avowed -- i.e. self-affirmed -- interests of those subject that non-arbitrary power must to forced to track, not the power holder’s ideas about what their real interests are.  Given this understanding of arbitrary power, neo-Republicans maintain that there exists a necessary and non-instrumental relation between freedom and democratic government. Power exists in various forms and is present in many sorts of relationships within society -- e.g. between employers and employees, landlords and tenants, husbands and wives, parents and children, merchants and customers, and more generally between any individual or group that has access to the means to injure, rob or kill and those within their reach. If freedom (as non-domination) is to be preserved, 53these various powers must be constrained in such a way that they are not exercised in ways that are detrimental to the interests of those subject to them. Pettit insists, plausibly, that the only feasible way to do this is the establishment of a political authority that has the power to regulate such power relations in the interests of those subject to them. However, while that may solve the problem of domination between private agents in society -- the problem of dominium -- it risks doing so by creating a new problem of domination between the political authority and its subjects or citizens -- a problem of imperium. For this reason, the power wielded by the political authority must itself be subject to constraint -- to ensure it operates in the interests of those subject to it. That is, there must be mechanisms in place that effectively hold it accountable and responsive to those interests. For this reason, freedom necessarily requires some form of democratic control of the political authority. A free society, then, is one in which all social power is brought under the regulatory power of a political authority, which in turn is subject to effective democratic contestation and control.   A number of objections have been raised concerning the neo-Republican theory of freedom, three of which seem especially significant: (i) The “gentle giant” objection, (ii) the “freedom in chains” objection, and (iii) the moralization objection. (i) The “gentle giant” objection   According to the neo-Republican account, an individual is dominated just in case some other agent has the power to interfere in her choices merely at his will or pleasure -- i.e. arbitrarily. It follows from this account, as we’ve seen, that freedom can be undermined even in the absence of actual interference. The mere possession of the 54power to interfere arbitrarily is sufficient. However, according to Matthew Kramer, such a view leads to implausible conclusions, as the following story illustrates:  ? Suppose that, in a community not far from the hills, a gigantic person G is born.   From adolescence onward, G is far larger and stronger and swifter and more   intelligent than any of his compatriots. If he wished, he could arrogate to himself   an autocratic sway over his community by threatening to engage in rampages   and by coercing some of the residents into serving as his henchmen. Were G so   inclined, no one would dare to resist his bidding... In fact, however, he loathes the  idea of becoming a tyrant; his principal desire is to seclude himself altogether   from his community. He does indeed depart therefrom, in order to reside in a   cave in the nearby hills where he contentedly feeds off natural fruits and wildlife   and where he spends his time in solitary reflection and reading and exercise. In   these circumstances, G is a dominator (according to Pettit’s criteria for that   status), but he is not significantly reducing the overall liberty of anyone else. ?It seems, then, that the mere possession of the power to interfere (even arbitrarily) cannot be sufficient to undermine the freedom of someone vulnerable to that power.   (ii) The “freedom in chains” objection  Because on the neo-republican account it is only the subjection to arbitrary forms of power that constitute domination, it follows that interferences undertaken on a non-arbitrary basis do not diminish a person’s freedom. That is, non-arbitrary interferences do not reduce a person’s freedom. But such a view, critics claim, implies highly counter-55intuitive judgements concerning particular cases. Suppose someone builds a fence across a path others could previously have walked along. If his building the fence is not arbitrary (say he is granted the power to do so by the relevant well-constituted and legitimate political authority), then the fence does not diminish others’ freedom, though it clearly prevents them from doing something they could previously do. Even more dramatically, suppose a particular person is locked in a prison cell, serving a lengthy sentence. If it turns out that that person has not been arbitrarily detained -- suppose he is guilty of a serious crime and has been convicted in a court of law -- then it follows, on the neo-Republican account, that his current state of confinement has not in fact diminished his freedom.  (iii) The “moralization” objection   Related to the second objection, some critics have claimed that the neo-republican account is ultimately guilty of moralizing the concept of political freedom. It certainly discriminates between forms of power that undermine the freedom of those subject to it and forms of power that do not (and in consequence between kinds of interferences that undermine freedom and kinds that do not). The basis for these discriminations is, to be sure, the standard of arbitrariness. But underlying the language of “arbitrariness,” the critics insist, must be some conception of what is normatively justified. If that is right, then the neo-republican conception is, contrary to the intentions of its proponents, a moralized one.  1.4. The problem of this dissertation  Having identified and discussed what I take to be the leading contemporary approaches to the question of how to conceive of political freedom, I now want to 56??narrow in and frame the problem of this dissertation.  1.4.1. Freedom as an external social relation  In my view, the non-interference and non-domination ideals represent the most promising approaches to the theory of political freedom. While I will not attempt a further defence of this claim here, the conception of freedom as ability seems to me to gloss over what is distinctive about of our interest in political freedom -- namely, that we desire not to be unduly controlled or intruded upon or thwarted by other human beings. On some elaborations, the conception of freedom as individual autonomy is similarly inattentive to the social character of the ideal of political freedom. Taylor includes maladjusted emotional dispositions among the “inner obstacles” to freedom, but these hardly seem the sorts of concerns about which people might reasonably object to others that their liberty had been violated. Christman’s approach at least focuses on inner obstacles that are the result of oppressive socialization pressures. However, even Christman’s view seems to miss the mark concerning our most basic interest in political freedom -- since it ends up caught in a dilemma between either vastly over-permitting coercion or conceding that the ideal of freedom as autonomy is unsuited as a norm for setting the bounds on legitimate social coercion or control.  On both the non-interference and non-domination conceptions, by contrast, the ideal of political freedom as essentially concerned with the relations between human beings -- in particular, with those relations insofar as they involve the exercise of (or power to exercise) force by some people to direct the actions and lives of others. 57Moreover, on both views the conditions of freedom are entirely external to the individual whose freedom is in question. Neither view makes room for the idea of “inner obstacles” to political freedom; thus both are able to steer clear of the dilemmas and dangers of conceiving of freedom as involving a kind of psychological achievement of aligning one’s aims with one’s real or rational will. For these reasons each of these views seem like promising starting points for a theory of political freedom.  1.4.2. Non-interference versus non-domination  The basic difference between the non-interference and non-domination views is that, according to the first, freedom is only diminished when an action a person could otherwise have taken is actually interfered with by other human agents (or would likely be interfered with if he attempted it), whereas, according to the second, freedom is diminished by the mere possibility of or power for interference (on an arbitrary basis).     This difference has large implications for understanding of the nature of a free society. If one conceives of freedom as non-interference, then the degree of an individual’s freedom will depend on the extent to which he is interfered with by other human beings in doing what he could otherwise do. Liberals who have been attracted to this view have, consequently, conceived of the free society as one in which each of its members are accorded an appropriately wide domain in which they may act free from the interference of others. It’s true that, even within the liberal camp, there are differences concerning what principle ought to determine the bounds of the domain of individual non-interference. Moreover, according to some (those who favour the “moralized” approach) the principle of distributing non-interference is built into the very concept of liberty; whereas, according to others, the principle of distribution is external 58to the definition of liberty. Nonetheless, the essential point remains -- namely, that a free society is defined in terms of its having protected certain areas of human activity from intrusion by others. Moreover, liberals have tended to produce a roughly overlapping set of “liberties” that are of fundamental significance -- freedom of belief and conscience, freedom of association, freedom of speech, freedom of religion, freedom to own and exchange property, etc.      If, on the other hand, one conceives of freedom as non-domination, then the degree of an individual’s freedom will depend on the extent to which he is subject to arbitrary power. As we’ve seen, neo-Republicans who advocate this view have tended to define “arbitrary power” as power that is not forced to track the avowed interests of those subject to it. In consequence of defining freedom in this way, as we’ve noted, neo-republicans conceive of a free society as one in which all social power is brought under the effective control of those subject to it by means of an appropriately constituted and democratic political authority.  1.4.3. Contending insights  I believe that, underlying each these conceptions of the free society, there is an important insight that is lacking in its rival. Liberals are right, it seems to me, that any acceptable account of the free society must include a limit on the power and authority of the state -- democratic or otherwise -- to interfere in the lives of individual members. The neo-Republican view, by associating the ideal of freedom so closely with democratic government, risks obscuring the danger of the tyranny of the majority. Yet Republicans are right, in my view, to insist that freedom can be undermined by the mere existence of arbitrary power. When individuals live in conditions in which others have a power to 59interfere in their actions at their pleasure, that in itself diminishes their freedom -- even in those instances where that power is not manifested in actual interferences. Thus, in demanding her freedom, the slave is generally not demanding merely that she be granted more time for leisure and personal pursuits, or be allowed to marry and have a family, or observe the practices of her religion; instead, she is demanding that she no longer be under the power of the master. If the master were to object to her demand on grounds that he is a kindly and liberal slave-holder, he would be missing the essential point. Likewise, on the non-interference view, as Berlin himself acknowledges, it is theoretically possible to have a “liberal despot” -- an unaccountable ruler that nonetheless happens to give his subjects a wide berth of un-interfered with choice.  However, in line with republican claims, this seems mistaken to me. To put it in terms I think most apt: The mere fact that someone arrogates to himself the authority to decide what others may or may not do is an affront to liberty. For this reason it seems to me that a free society is necessarily one in which the political power is constrained in a way that makes it accountable and responsive to the claims of those subject to it. However, this intuition is one the liberal non-interference view is ultimately unable to account for.  The central question motivating this dissertation, then, is whether there is a way to conceive of political freedom that captures the basic insights of the liberal and republican views: The liberal insight that freedom requires that there be a limit to how far others, including a democratic state, can interfere in the lives of individuals, and the republican insight that not just interferences but also relations of power as such can undermine a person’s freedom. I believe the answer to this question is yes. My aim in 60this dissertation is to articulate and defend a conception of freedom which I think successfully reconciles these insights in a plausible way.  1.4.4. The conjunctive approach  However, the view I am proposing must be distinguished from another approach that similarly seeks to reconcile republican and liberal claims -- namely, what we might call the “conjunctive approach.” According to it, freedom consists in the absence of both interference and domination. Here the thought is that all actual interferences diminish a person’s freedom, but that, in addition, freedom is also diminished by the mere vulnerability to arbitrary power. This appears to be the view advocated by Quentin Skinner in his earlier writings. He writes: “[F]reedom is restricted not only by actual interference or the threat of it, but also by the mere knowledge that we are living in dependence on the goodwill of others.”   Skinner, however, appears to have moved away from this conjunctive approach in later writings, opting for the simple non-domination view advocated by Pettit -- no doubt under the influence of Pettit’s arguments in this direction. As Pettit notes, the only difference between the conjunctive account and the simple non-domination account is that the former counts non-arbitrary interferences as diminishments of freedom while the latter does not. But, Pettit argues, the conjunctive approach is inferior in this respect for two reasons. In the first place, it is out of step with the historical republican tradition -- since according to that tradition laws that are not made on an arbitrary basis do not diminish liberty, but in fact constitute liberty.   In the second place, and more importantly, the conjunctive approach depends on a confused account of the nature of the wrong involved in infringements of freedom. 61One might think that the wrong involved in both domination and interference is the fact that both restrict choice. However, if that were the case, Pettit argues, there could be no principled reason to exclude restrictions on choice that are the result of natural inabilities or the unintended consequences of human actions from the list of items that diminish freedom. This would reduce the ideal of liberty to that of ability, a result Pettit thinks Skinner will find unattractive. Instead, what makes dominating interference uniquely evil is the fact that it restricts choice by subjecting an individual to the arbitrary will of another human being.  It thereby leaves the dominated in a particularly insecure position, being subject to the whim of the other; moreover, it involves an asymmetry of status, where the dominated is not accorded the respect of one that is equally worth being attended to or heard. However, these features of dominating interference, Pettit insists, are not present in non-dominating interference. Thus, the conjunctive approach represents an unstable middle ground position between the view that the wrong involved in constraining liberty is the restriction of choice and the view that the wrong involved is a particularly social one that leaves the dominated person in a vulnerable and degraded condition vis-a-vis other persons.   Charles Larmore accepts that the ideals of non-domination and non-interference indeed represent distinct ends embodying distinct values (roughly the value of having the status of an equal among others in the community and the value not being coerced into doing what one might otherwise not do). However, rather than seeking to conjoin these values into a single ideal of freedom, as on the conjunctive approach advocated by the early Skinner, he suggests viewing the two ideals as simply that -- two distinct goals going under the name of freedom. Each of them, he is happy to acknowledge, 62represents a valid and attractive political goal; and both of them ought to be accorded their due in our social and political institutions.   Larmore’s approach has a certain attraction since it recognizes the insights underlying both the proponents of both negative liberty and republican freedom. And, interestingly, it bears a strong resemblance to Berlin’s own value pluralist approach:   ? Positive liberty... is a valid universal goal. I do not know why I should have been   held to doubt this, or, for that matter, the further proposition, that democratic   self-government is a fundamental human need, something valuable in itself,   whether or not it clashes with the claims of negative liberty or of any other goal...   What I am mainly concerned to establish is that, whatever may be the common   ground between them, and whatever is liable to graver distortion, negative and   positive liberty are not the same thing. ?However, there are certain disadvantages to this facile pluralism. Most significantly, from the point of view of a theory of freedom, it leaves one unclear as to what a free society ought to look like. What sort of balance between non-domination and non-interference ought we to strike? What principles could we employ to determine when one ought to be given priority over the other? Without some further specification as to the relation between these values, it is unclear how the pluralist approach -- or Skinner’s conjunctive approach for that matter -- could provide normative guidance for the evaluation and construction of the political institutions of a free society. In light of this problem, it seems to me that, if it were possible to have a unified conception of freedom 63that managed to reconcile the insights of the republican and negative liberal views, that would be preferable.  1.4.5. The Lockean alternative  In this dissertation I propose just such a reconciling view. While absent from contemporary discussions of the concept of political freedom, this view is not entirely new. Its clearest historical expression can, I think, be found in John Locke’s Second Treatise. According to Locke, political freedom does not consist in “a liberty for everyone to do what he lists, to live as he pleases, and not to be tied to any laws”; rather “freedom of men under government is, to have a standing rule to live by... and not to be subject to the inconstant, uncertain, unknown, arbitrary will of another man...” In other words, Locke rejects the non-interference view and instead conceives of political freedom as non-domination. However, Locke thinks of non-domination differently than contemporary neo-Roman republicans, because he defines “arbitrary power” differently. As I shall argue, according to the Lockean view, arbitrary power is power that is forced to respect and uphold the natural or pre-political rights of those subject to it. “Freedom” consists in the absence of subjection to arbitrary power so defined.   Locke’s conception of freedom is thus a species of the non-domination view, but is distinctive in that it defines arbitrary power against the backdrop of natural rights. The following table (table 1.1) helps to clarify the Lockean conception by putting it in logical space with other prominent views. ???64Table 1.1. Lockean Freedom in Logical Space ??My fundamental aim in the dissertation is to offer a clear articulation of this Lockean alternative, and to argue that it is superior to both the neo-Roman conception of freedom as non-domination and the liberal conception of freedom as non-interference.  1.4.6. Chapter breakdown  In the next chapter (Chapter 2), my primary objective is to bring to light what I take to be the fundamental underlying insight of the historic liberal critique of republicanism. Contrary to Berlin’s view that the fundamental claim of the classical liberal tradition is that freedom consists in not being interfered with, I suggest that the commitment to the non-interference ideal is at best a consequence of a more basic commitment. That more basic commitment is a belief in an inviolable sovereignty of individual persons.  In Chapter 3, I explain what I take to be the fundamental insight of the republican conception of freedom -- one that leads them to reject the non-interference ideal. I Freedom as the... non-Moralized MoralizedNon-interference  ...absence of interference in actions one could otherwise perform. (Bentham, Mill, Berlin)...absence of interference in actions one has a right to perform.  (Nozick, Dworkin) Non-domination ...absence of subjection to arbitrary power, where non-arbitrary power is power that is forced to track the avowed interests of those subject to it.  (Pettit, Skinner)...absence of subjection of arbitrary power, where arbitrary power is power that is not forced to track the rightful interests of those subject to it. (Locke)65defend the republican position that freedom is best understood in terms of non-domination against recent objections from proponents on the non-interference ideal.  However, in Chapter 4, I argue that the neo-Roman account of the ideal of non-domination is vulnerable to the worry that led many classical liberals to prefer the non-interference ideal -- namely, concern about a tyranny of the majority. That this objection holds is not obvious, and has been denied by proponents of the neo-Roman position. These denials notwithstanding, I argue that their version of the ideal of non-domination is indeed unable to provide a satisfactory solution to the problem of majoritarianism. I locate the source of the problem, not in the conception of freedom as non-domination as such, but in the neo-Roman account of the nature of arbitrary power.   In Chapter 5, I articulate a different version of the republican conception of freedom as non-domination, one given its clearest historical formulation in John Locke’s Second Treatise. The crucial distinguishing feature of the Lockean view is that it defines arbitrary power as power that is forced to respect and uphold the natural rights of those subject to it. The resulting conception of freedom -- the “natural rights republican conception of freedom” -- I argue is able to accommodate the basic liberal concern identified in Chapter 2 without abandoning the fundamental republican insights defended in Chapter 3.  In Chapter 6, I defend the Lockean view against a number of objections against to the idea of defining liberty by reference to natural rights. In particular, I address concerns about Locke’s particular account of the content of natural rights, concerns about the very idea of natural rights, and concerns about adopting a “moralized” conception of freedom.
66Chapter 2 Republican “Freedom” and its Liberal Critics ?The aim of this chapter is to identify what I take to be the fundamental concern of the classic liberal critique of the republican ideal of freedom, and thus in part to explain the attraction of the “negative” or non-interference conception of freedom. The classic liberal critique however is most clearly directed against a particular version of the republican ideal of freedom -- namely, what is sometimes called the “neo-Athenian” version. For this reason, I begin by providing an analysis of this “neo-Athenian” understanding of freedom. The neo-Athenian view in general is that political freedom consists in collective self-mastery; however, there are differences even within this general approach. I highlight three significant strands: (a) the populist interpretation; (b) the deliberative interpretation; and (c) the rationalist interpretation. The classic liberal critique of republicanism takes aim at the definition of freedom as collective self-mastery under each of these three interpretations. I draw on the political writings of Benjamin Constant, John Stuart Mill and Isaiah Berlin to outline the content of that critique. While I agree with the negative point of these liberal objections, I take issue with Berlin’s diagnosis of the basic conviction underlying them. On Berlin’s view, the basic liberal conviction is that freedom consists in not being interfered with by others. However, I suggest that the acceptance of the non-interference ideal of freedom is, at best, a secondary consequence of a more basic commitment -- namely, a commitment to the idea of an inviolable individual sovereignty.   672.1. Freedom as self-mastery  At the heart of the republican tradition is a shared conviction that the fundamental antimony of freedom is slavery. A slave is the property of his master, who may direct the slave to do or not to do what he wishes in virtue of the power he holds over him. The essence of slavery is, thus, not to be found in the extent to which a slave’s actions are coerced, but rather in the fact that he is subject to another, under the power and authority of another, that he must take his direction from the will of another rather than from his own will.   The view that slavery is freedom’s opposite has deep historical roots. In Justinian’s Digest of Roman law, perhaps the most influential classical discussion of the concept of liberty, we read that “the fundamental division within the law of peoples is that all men and women are either free or slaves.” And further that “slavery,” formally defined, is “an institution of the ius gentium by which someone is, contrary to nature, subjected to the dominion of someone else.”  A slave, in contrast to a free citizen, is 79sub potestate, under the power or subject to the will of another.   80 But given this definition of slavery as subjection to the dominion or will of another person, it has seemed clear to historical republicans that it is not only those individuals that are literally the property of a master that must be accounted unfree, but also, for instance, those that are the ‘subjects’ of an absolute ruler -- like a monarch or an 68 Cited in Skinner, “Classical Liberty and the Coming of the English Civil War,” 979 Ibid.80emperor.  An absolute ruler, just like the literal slave-holder, holds an arbitrary power 81and dominion over his subjects, able to direct those subjects to act or not act at his pleasure and prerogative. For this reason, according to historical republicans, no people can be accounted free so long as they remain the subjects of such a king or emperor -- they are no more than slaves by another name.  822.1.1. ‘Neo-Athenian’ and ‘neo-Roman’ republicanism  But while republicans are unified in seeing unfreedom as essentially a condition of being dominated, of standing in relation to another as a slave does to her master, important differences emerge concerning what it means to be free. It is now common to distinguish between ‘neo-Athenian’ and ‘neo-Roman’ forms of republicanism. Neo-Athenian republicans tend to identify freedom with the positive condition of self-mastery, understood as the activity of self-rule or self-legislation, i.e. with being subject only to laws or constraints that one has given oneself. Neo-Roman republicans identify freedom merely with the negative condition of not being mastered or ruled by others.  Thus 83Laborde and Maynor write of the former that it endorses “the Aristotelian view that real freedom consists in self-mastery or self-realization in a community with others,” and further that “followers of the neo-Athenian tradition” believe that ‘liberty’ is “definitionally 69 It is perhaps not surprising then that, at the level of popular discourse, the term ‘republican’ has at 81many times been used as a simple synonym for ‘anti-monarchist.’ Indeed the birth of the Roman republic is dated from the overthrow of the Roman monarchy, its end with the rise of the imperium.  In truth though, the republican tradition exhibits a more ambivalent attitude towards monarchy, in many cases favouring a “mixed” form of government over its complete abolishment. This ambivalence is partly explained by divisions within republican thinking, to be explained below. However, these complexities aside, the essential point here remains: that absolute rulers (and so at least absolute monarchies) are inherently incompatible with the freedom of subjects. See Pettit, Republicanism, 32-382 Ibid., 21-283linked to popular participation” in “self-governing communities.”  By contrast, according 84to the neo-Roman strand of republicanism, freedom consists not in self-mastery or self-rule, but rather in the absence of mastery or rule by others. “The people, neo-Roman writers from Machiavelli through Harrington assured us, did not want to rule: instead, they wanted not to be ruled, or at least not to be ruled in a particular way” -- i.e. as subjects to the arbitrary power of an emperor or king.    85 The distinction between ‘neo-Athenian’ and ’neo-Roman’ conceptions of freedom is associated with the distinction between what is called ‘developmental’ or ‘perfectionist’ republicanism on the one hand and ‘protective’ or ‘political’ republicanism on the other. ‘Developmental’ republicanism emphasizes “the intrinsic values of political participation for the enhancement of decision-making and development of the citizenry,”  and so fits neatly with the idea that freedom consists in self-mastery through 86self-legislation. ‘Protective’ republicanism places “primary stress on the instrumental value of political participation for the protection of citizens’ objectives and interests,”  87and so is a more natural fit with the neo-Roman conception of freedom. While both developmental and protective republicanism emphasize the necessity of citizen virtue and political participation, the former emphasizes it for its own sake, while the latter emphasizes it as a necessary means for protecting the pursuit of private ends. The former view is thus a kind of “perfectionist” republicanism, while the latter is a kind of “political” republicanism.   8870 Laborde and Maynor, “The Republican Contribution to Contemporary Political Theory,” 384 Ibid.85 Held, Models of Democracy, 4586 Ibid.87 Weithman, “Political Republicanism and Perfectionist Republicanism” 88 In this chapter, I will focus my attention on the ‘neo-Athenian’ conception of freedom as self-mastery, since it is against the backdrop of this strand of thinking that the classical liberal critique of “republican liberty” is brought into sharpest relief; indeed, the critics I am concerned with are most plausibly interpreted as directing their arguments against this more ambitious republican ideal. In Chapter 4, I will turn to consider what force these liberal objections might have against the other, neo-Roman, strand of thinking about freedom. 2.1.2. ‘Freedom’ in Rousseau’s republic  As its name suggests, the neo-Athenian conception of freedom as self-mastery finds its classical inspiration in the model of direct democracy that prevailed in the ancient Athenian city-state. The most influential modern expression of this ideal -- i.e. of neo-Athenian republicanism -- is to be found in Rousseau’s Social Contract (1762).   In Book One of The Social Contract, Rousseau famously writes that, “Man is born free, and everywhere he is in chains,”  and further that “[t]o renounce one’s liberty 89is to renounce one’s quality as a man,” an act for which “there is no possible compensation.”  Nonetheless, he acknowledges that human beings have reached a 90point in their history “at which the obstacles to [their] self-preservation in the state of nature are too great to be overcome,” and that this pre-political, primitive condition can no longer subsist.  So the fundamental problem he sets himself to solve is this: To 91“[f]ind a form of association which will defend and protect, with the whole of its joint strength, the person and property of each associate, and under which each of them, 71 Rousseau, The Social Contract, Bk. 1, Ch. 189 Ibid., Bk. 1, Ch. 4 90 Ibid., Bk. 1, Ch. 691uniting himself to all, will obey himself alone, and remain as free as before.”  92Rousseau’s solution involves each individual transferring all claims to natural liberty and right to the community as a whole, but in return gaining an equal voice and consideration in the formation of the General Will, which, as sovereign, legislates for the community.   But it is crucial to notice in what the sense of the word ‘freedom’ this solution can be said to have met the requirement of fully preserving the freedom of each individual. It does not preserve the natural liberty to do whatever is in one’s power to do; many things which a natural individual could previously do may now be forbidden by law and subject to sanction. Nor does it consist in the protection of certain inviolable domains of non-interference for private choice. Instead, the freedom Rousseau’s solution preserves is understood as a matter of being subject only to laws which one has given oneself, through participation as an equal in the deliberations that form the basis for the general will. “Moral liberty,” writes Rousseau, is “the only thing that makes man truly the master of himself; for to be driven by our appetites alone is slavery, while to obey a law that we have imposed on ourselves is freedom.”  932.2. Varieties of ‘self-mastery’: the individualistic interpretations  So Rousseau appears to conceive of freedom as self-mastery, and to interpret ‘self-mastery’ as self-legislation -- i.e. as ‘being subject only to laws which you have given your self.’ However, even so specified, the ideal of self-mastery is in fact open to a variety of interpretations. In what follows, I aim to provide an analytical catalogue of the most salient interpretations of the ideal of freedom as self-mastery.  72 Ibid.92 Ibid., Bk. 1, Ch. 8932.2.1. Individual self-mastery  How might one interpret the notion of ‘self-mastery?’ The first distinction to be drawn is between individualistic and collective interpretations of this ideal. Rousseau appears to take the condition of self-mastery to be satisfied so long as the individual in question is subject only to constraints or laws that find their source in the collective decision-making of a whole (i.e. the ‘General Will’) of which that individual is an equal and contributing part. Rousseau, thus, appears to conceive of freedom as collective self-mastery. I will return to consider the collective interpretation in Section 2.3; and, in fact, it is this interpretation that will be the primary concern of the present chapter. However, for now I would like to focus on the individualistic reading. According to the individualist interpretation of the ideal of self-mastery, a person is free to the extent that she has the power to direct her life according to her own individual will. 2.2.2. Primitive self-mastery  The individualistic interpretation of the ‘self-mastery’ itself admits of further sub-divisions, depending on how one understands the nature of an individual’s ‘will.’ According to the most primitive version, to be self-mastered is simply to stand in the same relation to oneself as a master does to his slave. As noted, to be a master is to possess a power over one’s slave such that one is able to direct him to act or refrain from acting in various ways according to one’s own will or pleasure -- i.e. one’s arbitrium. To be one’s own master, would be to have this same power of direction over one’s own life and choices: to be able to do what one wants, rather than be to be compelled to do what someone else wants. Let’s call the notion of self-mastery under this interpretation the ‘primitive self-mastery’ interpretation. 73? Primitive Self-Mastery: The possession of a power to direct one’s actions    according to one’s own will (i.e. arbitrium), which is not subject to the like power   of another.  ?2.2.3. Rational autonomy  The ideal of self-mastery, however, is normally thought to imply more than mere primitive self-mastery. After all, even a wanton -- someone that unreflectively follows his strongest present desire -- might count as free in this sense. Thus, self-mastery is often thought to require that one direct one’s actions according to one’s ‘real’ or ‘rational’ will. This interpretation of ‘self-mastery’ is distinguished from primitive self-mastery in two ways: In the first place, it departs from primitive self-mastery in holding that to be free it is not enough to possess a power to direct one’s own actions according to one’s arbitrium; but rather one must direct one’s actions according to some rational set of principles or higher purposes. Rousseau expresses a thought along these lines when he claims that “to be driven by our appetites alone is slavery, while to obey a law that we have imposed on ourselves is freedom.”  To be one’s own master is to govern 94oneself according to laws rather than arbitrary preferences -- not necessarily laws in the sense of formal legal statutes, but in the sense of some consistent and intelligible set of reasons. It is thus an ideal of “autonomy” as philosophers commonly use that term, an ideal of auto (“self”) nomos (“law”). This interpretation of self-mastery also departs from primitive self-mastery in making it an “exercise” rather than an “opportunity” concept.  9574 Ibid.94 Taylor, “What’s Wrong with Negative Liberty?”, 21395That is, it requires not merely the possession of the capacity for directing one’s actions and life according to rational principles, but the actual (successful) exercising of this capacity.   Isaiah Berlin describes the step from the notion of primitive self-mastery to the notion of self-mastery as involving a kind of achievement of psychological self-regulation as follows: ? One way of making this clear is in terms of the independent momentum which   the, initially perhaps quite harmless, metaphor of self-mastery acquired. ‘I am my   own master’; ‘I am a slave to no man’; but may I not (as Platonists or Hegelians   tend to say) be a slave to nature? Or to my own ‘unbridled’ passions? Are these   not species of the identical genus ‘slave’ -- some political or legal, others moral   and spiritual? Have not men had the experience of liberating themselves from   spiritual slavery, or slavery to nature, and do they not in the course of it become   aware, on the one hand, of a self which dominates, and, on the other, of    something in them which is brought to heel? This dominant self is then variously   identified with reason, with my ‘higher nature’, with the self which calculates and   aims at what will satisfy it in the long run, with my ‘real’, or ‘ideal’, or  ‘autonomous’ self, or with my ‘self’ at its best’; which is then contrasted with   irrational impulse, uncontrolled desires, my ‘lower’ nature, the pursuit of    immediate pleasures, my ‘empirical’ or ‘heteronomous’ self, swept by every gust   of desire and passion, needing to be rigidly disciplined if it is ever to rise to the   full height of its ‘real’ nature.  9675 “Two Concepts,” 13296?From this passage we see that, on Berlin’s view, the notion of self-mastery evolves in such a way as to ‘split’ the self between ‘higher’ and ‘lower’ psychological elements, between “reason” on the one hand and the “desires and passions” on the other. Freedom (self-mastery) is then identified with an achievement of directing one’s life and actions by the determinations of one’s reason -- thereby overcoming ‘inner obstacles’ such as ‘irrational impulses,’ ‘uncontrolled desires,’ ‘immediate pleasures,’ inordinate ‘passions’ (e.g. irrational fears), and, broadening the category beyond what Berlin includes in this passage, unexamined dogmas, irrational beliefs, ‘false consciousness’ and the like.  Let’s call the notion of self-mastery under this interpretation ‘rational 97autonomy.’     ? Rational Autonomy: The possession and exercise of a power to direct one’s   actions according to one’s reason.   ?2.2.4. Procedural versus substantive autonomy  This interpretation of the ideal of self-mastery as rational autonomy, however, remains somewhat indeterminate. What might it mean to direct one’s actions or govern one’s life ‘according to one’s reason’? In particular, what understanding of ‘reason’ is to be pre-supposed? Berlin, at times, seems to gloss over important differences in views about the nature of practical ‘reasons’ and, thus, about what constitutes rational autonomy. While we need not (and could not) discuss all the intricacies of those differences here, it will be helpful to at least distinguish between two general 76 Taylor, “What’s Wrong with Negative Liberty?”, 21597approaches to the idea of rational autonomy, corresponding to two general understandings of the nature of practical rationality.   98 In the first place, we might interpret the idea of rational autonomy along proceduralist lines. Suppose we hold a broadly “Humean” or “internalist” conception of practical reasons, such that a person, ‘P,’ only has a reason to do X if doing X would satisfy some desire that P has -- or, in Bernard Williams more nuanced formulation, if there is a sound deliberative route from P’s “subjective motivational set” to her desiring to do X.  In that case, we will be inclined to adopt a proceduralist conception of rational 99autonomy. According to a proceduralist view, an agent’s ends are deemed autonomous so long as they would be self-endorsed under certain ‘internalist’ conditions of rational reflection. That is, rational autonomy consists in the exercise of the capacity to reflect one’s preferences in the light of one’s total set of values and interests, deliberate competently about their consistency with and acceptability in light of these other values and interests, and either endorse or repudiate those preferences on the basis of such deliberations. Being rationally autonomous, then, implies only that one’s ends satisfy certain procedural conditions; it does not imply anything about the ‘content’ of those ends. Let’s call this interpretation of the ideal of self-mastery ‘procedural autonomy.’ ?77 Some readers will note the absence of a discussion of the Kantian conception of autonomy here. The 98Kantian view does not fit neatly in either the procedural or the substantive category, since it holds both that all practical reasons are internalist in character and that there are some ends which all rational creatures have regardless the particular contents of their subjective motivation sets. Moreover, because a fully “Kantian-autonomous” agent always respects the rational aims of other agents, it is not clear that such a view generates a coordination problem (2.2.5). In section 2.3.5 I discuss what I call the “rationalist interpretation” of the collective self-mastery ideal — a view that is perhaps closest to the Kantian view.    Williams, “Internal and External Reasons”99 Procedural Autonomy: The possession and exercise of a capacity to direct one’s   actions for the sake of ends which one endorses (or would endorse) on the basis   of personal reflection upon and competent deliberation from one’s other values   and interests. ?By contrast, substantive accounts interpret the “reason” of rational autonomy along ‘externalist’ lines.  That is, this view presupposes that there are some objective or 100universally valid truths -- perhaps derived from human nature -- about which ends individuals have a reason to adopt irrespective of their ‘subjective motivational set.’ Thus, directing one’s actions in accordance with reason requires identifying and acting upon ends that meet some objective standard of human flourishing. Let us call this interpretation of the ideal of self-mastery ‘substantive autonomy.’ ? Substantive Autonomy: The possession and exercise of one’s power for directing  one’s actions according to ends that conform with an objective rational standard   of human flourishing. ?78 In contemporary discussions of autonomy, the term ‘substantive autonomy,’ it seems, is used to mean 100at least two distinct things, though often without clearly distinguishing these. That is, the term ‘substantive autonomy’ is sometimes used to refer to the view according to which a person is autonomous only if their set of ends has a certain content. A person that happily endorsed her own enslavement or domestic subordination on the basis of competent procedural reasoning, would not thereby count as autonomous on this view because of the content of her aims. This version of ‘substantive autonomy’ is perfectionist in orientation. Alternatively, the term ‘substantive autonomy’ is sometimes used to refer to the view that in order to be autonomous certain external, independence-securing conditions must be met (cf. Christman, “Relational Autonomy, Liberal Individualism, and the Social Constitution of Selves”). My use of the term here is of the former, perfectionist, sort.  It’s clear that Berlin was especially focused on the latter, substantive ideal of autonomy  -- though he evidently includes within its scope a range views concerning 101the basis for determining what count of rational ends. 2.2.5. Individual autonomy and the coordination problem   The individualistic interpretations of the ideal of self-mastery, however, face a coordination problem. If freedom consists in an individual’s directing of her actions according to her own rational will, how is the freedom of various members of a given society to be reconciled? For a variety of familiar reasons, it seems inevitable that such individual powers of rational self-determination will come into conflict with each other in society. If the conception of freedom as individual self-mastery is to serve as the basis for a theory of freedom (as defined in Chapter One), there must be some solution to the coordination problem.   It is uncontroversial, I assume, that the ordinary preferences of individuals in any political community are bound to come into conflict. Some resources, goods or opportunities are scarce such that individuals’ preferences for those things outstrip the supply. In such cases it is impossible for everyone’s preferences to be satisfied. Additionally, some individuals may have preferences concerning other individuals or which would impact other individuals in ways they would prefer not to be. For instance, David may strongly prefer to marry Alice, but Alice may have an equally strong preference not to be married to David. In this case it would be impossible for David’s preference to be satisfied without thwarting Alice’s, and vice versa. Beneficiaries of a 79 Not exclusively though, as the italicized portion of this excerpt from the passage quoted above 101indicates: “This dominant self is then variously identified with reason, with my ‘higher nature’, with the self which calculates and aims at what will satisfy it in the long run, with my ‘real’, or ‘ideal’, or ‘autonomous’ self, or with my ‘self’ at its best’” (“Two Concepts,” 132; emphasis added).certain mining operation may prefer that the operation be allowed to continue in spite of negative impacts on the drinking water of an adjacent community. Here the satisfaction of the preferences of one group will produce side effects that are dis-preferred by another.   Conflicts among the ordinary preferences of individuals in a community are ubiquitous. Moreover there is no reason to think that such conflicts will disappear when it comes to their autonomously-formed ends. David might carefully reflect on his desire to marry Alice in light of his ‘subjective motivational set’ and still conclude, by sound reasoning, that marrying Alice is the best thing to do, while Alice might, by a like process of critical reflection, arrive at the conclusion that marrying David is not the best thing to do. So long as the ‘subjective motivation sets’ of individuals differ, as they inevitably do, there is every reason to think that conflicts in individuals’ (procedurally) autonomous ends will persist.   The same conclusion also holds, I believe, on a substantive account of autonomy. I defined ‘substantive autonomy’ as ‘the possession and exercise of one’s power for directing one’s actions according to ends that conform with an objective rational standard of human flourishing. Complete conceptions of the good life or human flourishing, no doubt, vary according to what Rawls calls the distinct ‘comprehensive doctrines’ individuals hold to -- at points rendering conflicting judgements as to which ends an individual ought to pursue, or which virtues they ought to have. However, leaving this difficulty aside for the moment, it seems reasonable to assume that virtually any conception of the good life or human flourishing is going to include the satisfaction 80of basic human needs (e.g. physical security, food, health, shelter, family/community) and the development of certain basic human capacities.    102 However, even on this minimal account of human flourishing, the pursuit of ends which serve the flourishing of some individuals (and so the pursuit of which would count as ‘rational’ for those individuals and, thus, autonomous in the substantive sense) will conflict with and hinder the flourishing of other individuals (and so count against their autonomy). For example, suppose that among the beneficiaries of the mining operation in the previous example there are families of employees who depend on the income they earn from working at the mine to pay for their housing, their food, their children’s education, health care bills, etc. On the other hand, members of the local community depend on their access to clean drinking water for their continued health, and for the long-term viability of their community. So even on a minimal conception of human flourishing, the substantive autonomy of some individuals is liable to conflict with the substantive autonomy of others. Moreover, as we enrich our conception of human flourishing beyond the minimum, the extent of potential conflict between the rational aims of autonomous individuals is only going to be amplified.    In sum, given that what we are after is a theory of freedom, it seems clear that the individualistic version of the ideal of freedom as self-mastery -- under each of the interpretations noted -- faces a coordination problem. Individual wills, even rational ones, are bound to collide. Some principle or procedure to adjudicate such conflicts is required if the conception of freedom as self-mastery is to provide normative guidance as a political ideal.  81 cf. Nussbaum, Women and Human Development: The Capabilities Approach102 The coordination problem is at least one reason one might be inclined to prefer a collective interpretation of the ideal of self-mastery, rather than an individualistic one. In this line, Berlin writes: “Those who believed in freedom as rational self-direction were bound, sooner or later, to consider how this was to be applied... to [an individual’s] relations with other members of society. Even the most individualistic among them -- and Rousseau, Kant, and Fichte certainly began as individualists -- came at some point to ask themselves whether a rational life not only for the individual, but also for society, was possible, and if so, how it was to be achieved. I wish to be free to live as my rational will (my ‘real self’) commands, but so must others be. How am I to avoid collisions with their wills?”  One thought, to be considered below, is that the wills of 103each member of society can be reconciled and united into a ‘General Will’ through some collective reasoning process.  2.3. Varieties of self-mastery: the collective interpretations  Whereas the individualistic interpretations of the ideal of self-mastery require that individuals possess (and exercise) a personal power to direct their own lives (according to reason), collective interpretations of the ideal associate freedom instead with the exercising of a collective power to direct the social life of the community. As with the individualistic version, the collective version of the ideal of the self-mastery comes in a variety of forms.  2.3.1. The ‘metaphysical’ interpretation  The first distinction to be drawn concerns who the bearer of freedom is. Is the bearer of freedom some collective or supra-personal entity itself, or is the bearer of 82 “Two Concepts,” 145103freedom the individual in virtue of the role they play or the position they occupy in the collective body that governs the social whole? Berlin writes of the former view: “...the real self may be conceived as something wider than the individual (as the term is normally understood), as a social ‘whole’ of which the individual is an element or aspect.”   And further that “[t]his wish to assert the ‘personality’ of my class, or group 104or nation,” includes both the desire that “the group... not be interfered with by outside masters”  and, even more importantly, the desire that the group impose “its collective, 105or ‘organic’, single will” in the government of its “members.”  Indeed, according to 106Berlin, this desire for collective self-mastery or self-determination is “what makes nationalism or Marxism attractive to nations which have been ruled by alien masters.”   107 The first version of the collective interpretation of the ideal of self-mastery, then, depends on a metaphysical reification of the social ‘self’ with its own interests and will that is not a simple function of the individual preferences of its members (even the majority preferences of the members) or of the joint deliberations of those individual members. Instead, it views individual members as merely parts of an organic whole, which is the fundamental bearer of freedom. Let us call this the ‘metaphysical’ interpretation of the ideal of collective self-mastery. ?83 Ibid., 132104 Ibid., 160105 Ibid., 132106 Ibid., 159107 Metaphysical collective self-mastery: The collective possession and exercise of a  power to direct the actions of individual members according to the will or interests  of the social body or collective ‘self.’  ?2.3.2. The ‘participationist’ interpretation: issues  In contrast to the metaphysical interpretation, however, we might hold that the bearer of freedom is the individual in virtue of the role they play or the position they occupy in the collective body that governs the community. Thus it is sometimes said that an individual’s liberty consists in their “participation” in collective will-formation or decision-making process. Here the general thought is that an individual is free insofar as she shares in the power to which she is subject and, consequently, has a voice in the making of the laws by which she is governed.  However, this view can take a variety of forms depending on how the collective decision-making process or “general will” by which the community is governed is conceived. In the first place, there is the question of whether it is necessary that each member participate directly in the making of the laws by which the community is to be governed or whether it is enough that each member participate indirectly through the election of representatives who make the laws on their behalf.   There are at least three distinct rationales for insisting on the direct participation of “free” citizens in the making of laws. The first is a skepticism about the very possibility of the will of the people being faithfully “represented.” Rousseau takes this view when he writes that the sovereignty of the people “cannot be represented” since individual representatives don’t cease to have “particular wills” and interests and thus, by their 84“very nature” tend to “partiality” rather than “equality.”  Moreover, even if the particular 108will of the representative happens to agree with that of the general will at the time of appointment, it is “impossible for the agreement to be lasting and constant” since new circumstances call for new decisions which may not have any basis in the wills of those represented.  Thus, Rousseau provocatively claims that “[t]he people of England 109regards itself as free; but it is grossly mistaken; it is free only during the election of members of parliament. As soon as they are elected, slavery overtakes it, and it is nothing.”  A second rationale, also given by Rousseau, is that direct participation in 110collective political decision-making is necessary for the preservation and stability of the republic: “As soon as public service ceases to be the chief business of the citizens, and they would rather serve with their money than with their persons, the State is not far from its fall.”  A third rationale, one consonant with the “perfectionist republicanism” 111alluded to earlier, is that direct participation in political decision-making is essential as a means for the development and exercise of those public virtues that are necessary for the full flourishing of individual citizens.  A second question concerns the nature of the decision procedure by which the laws are made, rather than the constituency of those directly engaged in making them. 85 Rousseau, The Social Contract, Bk. 2, Ch. 1108 Rousseau writes: “In reality, if it is not impossible for a particular will to agree on some point with the 109general will, it is at least impossible for the agreement to be lasting and constant; for the particular will tends, by its very nature, to partiality, while the general will tends to equality. It is even more impossible to have any guarantee of this agreement; for even if it should always exist, it would be the effect not of art, but of chance. The Sovereign may indeed say: "I now will actually what this man wills, or at least what he says he wills"; but it cannot say: "What he wills tomorrow, I too shall will" because it is absurd for the will to bind itself for the future, nor is it incumbent on any will to consent to anything that is not for the good of the being who wills” (ibid). Ibid., Bk. 3, Ch. 15110 Ibid.111One possible answer is to require unanimity on each decision made or law passed. This view has the virtue of retaining a transparent connection between the will of each member and the collective or general will according to which those members are to be governed. On the other hand, a consensus requirement is usually considered impossibly demanding, particularly for large political communities. A second possibility is to require a simple majority-rule for any and all collective decisions. This view reverses the virtue and vice of the consensus view: It has the virtue of being practically realistic as a decision procedure for large and diverse communities; it has the vice of clouding the connection between the wills of individual members who find themselves in the minority and the general will. A third possibility is to require consensus (or at least something more than a simple majority) for the most basic or important decisions, and to require a simple majority for less basic or important decisions. Rousseau appears to endorse a version of this third option: requiring a consensus among members to initially lay down their natural liberties and rights in order to make a social pact to constitute a political body , a necessary clause of which pact is that a majority vote on all other 112matters is always binding,  though how substantial a majority will depend on how 113“important and serious the issue” and how “urgently” a decision is needed.  114 A third question regarding the character of the collective decision-making process concerns the form of the deliberations that are to serve as the basis for decisions that constitute the general will. One possibility is that each individual participates in the 86 “There is one sole law that by its nature demands unanimous consent: it is the social pact. For civil 112association is the most completely voluntary of acts; each man having been born free and master of himself, no one, under any pretext at all, may enslave him without his consent” (Ibid., Bk. 4, Ch. 2). “Except for this original contract, a majority vote is always binding on all the others; that is a direct 113consequence of the contract” (ibid.). Ibid.114political process with the aim of securing an agreement (whether a majority agreement or a consensus agreement) that best serves their private interests. Another possibility is that each individual aims to secure an agreement that best serves the common good. This second option appears to be Rousseau’s view; he writes: “There is often a difference between the will of everyone and the general will; the latter is concerned only with the common interest, while the former is concerned with private interests, and is the sum total of individual wants.”  On the first view, deliberations will take what Pettit 115calls a ‘bargain-based’ form; that is, “different interest groups try to secure a mutually beneficial agreement, with each looking for an agreement that will demand the smallest possible concession from them.”  On the second view, deliberations will take a 116‘debate-based’ form; that is, “different parties try to agree on what arrangement answers best to considerations that they can all recognize as relevant.”   117 The non-metaphysical -- or ‘participationist’ -- version of the ideal of collective self-mastery is thus susceptible to a variety of distinct interpretations depending on how one answers each of these three questions -- i.e. the question of direct versus indirect participation, the question of consensus, majority, or mixed decision procedures, and the question of bargain-based or debate-based forms of deliberation. Different versions of the ideal of collective self-mastery result as answers to these questions are “packaged” in different ways. While it is not necessary to run through all the possible variations, I would like to pay special attention to three distinct and influential 87 Ibid., Bk. 2, Ch. 3115 Republicanism, 187116 Ibid.117configurations -- what I will call the “populist,” “deliberative,” and “rationalist” interpretations.  2.3.3. The ‘populist’ interpretation  Underlying the ‘participationist’ interpretations of the ideal of collective self-mastery, is a conviction that individuals are free only insofar as they share equally in the power to which they are subject and, consequently, have an equal voice in the making of the laws by which they governed. But, as we’ve also noted, in political communities not all members agree about what laws ought to govern them. Interests clash and values conflict. One way of resolving this problem is to interpret the ideal along “majoritarian” or “populist” lines.   “Populism” is understood in different ways. Margret Canovan describes populism as the view that democracy should reflect the “pure and undiluted will of the people.”  118Other accounts emphasize the “anti-elitist” character of populism , favouring rule by 119the “common sense of the common person.” Some accounts associate populism with a commitment to direct democratic decision-making.  I will employ the term “populism” 120here to include (a) a commitment to a form of democracy in which all political decisions are made in accordance with the will of the majority and in which (b) individuals vote on the basis of whatever their existing preferences and ideas are rather than exclusively on those opinions that are “rational” or that emerge from a collective deliberation about the common good. For various reasons, I exclude from the definition a commitment to direct 88 Canovan, Populism118 Albertazzi and McDonnell, Twenty-First Century Populism: The Spectre of Western European 119Democracy, 3 Pettit, Republicanism, 8120participation in law making, instead keeping the view neutral between direct and indirect participation.  ? Populist collective self-mastery: An individual is free insofar as she is subject only  to laws agreed to, directly or indirectly, by the majority of community members on  the basis of their pre-existing preferences and ideas.    ?2.3.4. The ‘deliberative’ interpretation  A second historically prominent interpretation of the ideal of collective self-mastery takes its inspiration from the democracy of classical Athens -- and so has perhaps the best claim to the title “neo-Athenian” republicanism. I shall call this the “deliberative interpretation.” The deliberative interpretation differs from the populist view in two important respects: In the first place, it requires that each individual subject to the collective authority be directly involved in the deliberations and decisions that form the basis of the laws in order to count as “free.”  In the second place, it requires that the 121collective decision-making process must take a ‘debate-based’ form: participants deliberative together about the best way to achieve the common good. It is not a contest of individual interests and values, but a reasoned discussion about what is best for the community as a whole.  ? Deliberative collective self-mastery: An individual is free insofar as she is subject   only to laws agreed to by the majority on the basis of a collective deliberative  89 Classical Athenian democracy did not operate under a system of representation; instead, all citizens 121had the right and duty to participate directly in the assembly or ekklêsia, which body was responsible for, among other things, the making of laws by which the community was to be governed. process aimed at the common good in which she is an equal and direct    participant. ?The ideal of collective self-mastery under the deliberative interpretation comes closest, in my view, to the conception of freedom expressed by Rousseau in the Social Contract.  Moreover, there is a natural affinity between this way of conceiving of 122political freedom and the “developmental” or “perfectionist” form of republicanism I referred to earlier. Clearly the deliberative ideal requires a fairly demanding level of public virtue for its realization. Individual citizens must be willing to devote substantial time and energy to participation in political life, they must be ready to attempt to understand the situations and points of view of other citizens, they must be capable of offering public-regarding justifications for the policies they favour, and must be prepared 90 While I cannot provide a full defense of this claim here, I would like to head off two obvious objections: 122The first pertains to the role of majority decision-making. Earlier I suggested that Rousseau’s view is best seen as combining consensus and majority decision procedures. In the case of the original compacting to form a social body, it is true that he requires unanimity. However, he goes on to claim that those unwilling to so compact be considered “foreigners among citizens” and further that so long as they “live within [the state’s] boundaries” they must “submit to its sovereignty” (Social Contract, Bk. 4, Ch. 2). It seems then, that the choice is between submission to the majority will and exile. Moreover, though the creation of the social compact requires unanimous agreement among compactors, all subsequent decisions are subject to the principle of majority rule -- in fact, Rousseau appears to hold that the commitment to such a principle is a necessary implication of the act of forming a social contract. He writes: “Except for this original contract, a majority vote is always binding on all others; that is a direct consequence of the contract” (ibid.; italics added).   Another reason one might object to characterizing Rousseau as advocating an essentially Athenian view of freedom is his explicit rejection of the Athenian model of democracy (cf. Social Contract, Bk. 2, Ch. 4; Political Economy, 8). However, the reason Rousseau objects to the Athenian model of democracy is neither the participatory nor the deliberative character of its assembly, but rather what he perceived to be the lack of separation between legislative and executive/judicial functions (between the “sovereign” and the “government”). When the legislating assembly extends its reach to making judgements and decisions on particular cases, its deliberations and decisions no longer reflect the general will, but is overtaken by factions with particular or partial interests in the case at hand (cf. Social Contract, Bk. 2, Ch. 4; Political Economy, 8). Rousseau’s objection to the Athenian model is thus not an objection to the ideal of rule by a collective deliberative body aiming at the common good, but rather to the failure of the Athenian model to realize this very ideal -- i.e. its failure to adequately separate the task of making laws from the task of implementing those laws in particular cases corrupted the assembly itself so that it no longer aimed at the ‘common good’ and thus no longer constituted the ‘general will.’ For this very reason Rousseau writes that, “Athens was in reality not a democracy, but an extremely tyrannical aristocracy, controlled by philosophers and orators” (Political Economy, 8). For an alternative interpretation of Rousseau, see Neuhouser (“Freedom, Dependence, and the General Will”).to reconsider their own goals and exercise self-restraint for the sake of the common good.  While it is consistent with the definition of the deliberative ideal to view these 123civic virtues as merely instrumentally valuable -- i.e. as necessary means or “costs” for realizing and sustaining freedom -- advocates of the deliberative ideal often view them as intrinsically valuable elements of human flourishing in their own right, and thus see the deliberative ideal as compelling in part because it provides a forum in which such virtues can be exercised and developed.  Hence that deliberative ideal of self-mastery 124finds a natural home in a more generally perfectionist-leaning form of republicanism. 2.3.5. The ‘rationalist’ interpretation  While the populist and deliberative interpretations of the ideal of collective self-mastery differ in important respects, they share certain features as well. Both understand the ‘general will’ as the outcome of an actual decision-making process, in which individual members participate through the expression of their avowed preferences and opinions. They are also both committed to a decision-making process that falls short of consensus, deferring, at least in the end, to the principle of majority-rule. The latter point follows naturally from the former: Because actual collective deliberations on a political scale rarely if ever end in unanimity (even when robustly ‘debate-based’), some such principle is necessary if the community is to come to determinate decisions.   A third interpretation of the ideal of collective self-mastery differs from the previous two in requiring that the laws by which the community is governed be those that would be agreed to by all members were they to engage in an ideally rationally 91 Weithman, “Political Republicanism and Perfectionist Republicanism,” 294123 Ibid.124process of deliberation about how best to coordinate their aims. This interpretation is both less and more demanding than the populist and deliberative views. It is less demanding in that it requires of the laws by which the community is governed only that they would be agreed to under ideal deliberative conditions; only hypothetical agreement is required, not the “empirical” agreement of actually participating individuals operating under sub-optimal rational conditions. It is more demanding in that it requires that the laws be ones which would be agreed to by all members; consensus, not a mere majority, is required. Let’s call this the “rationalist interpretation” of the ideal of collective self-mastery. ? Rationalist collective self-mastery: An individual is free insofar as she is subject   only to laws that would be agreed to by all members of the collective were they to  engage in an ideally rational deliberation about the best way to coordinate their   aims.   125?2.3.6. Summary  Earlier I noted that, at the heart of the republican tradition, is a shared conviction that the fundamental antimony of freedom is slavery, the subjection to the arbitrary will 92 Though perhaps controversially, one historical figure that might be viewed as advocating such a 125‘rationalist republican’ view is Immanuel Kant. This ideal of the “kingdom of ends” -- perhaps better called a “republic of ends” -- serves as the fundamental normative principle guiding Kant’s political philosophy. Thus, he writes of the “original contract” that it is an “idea of reason” that constrains the sovereign to “give his laws in such a way that they could have arisen from the united will of a whole people and to regard each subject, insofar as he wants to be a citizen, as if he has joined in voting for such a will” (“On the Common Saying: That May Be Correct in Theory, But it is of No Use in Practice,” 8:297). The “original contract,” then, is for Kant not a historical event, but an “idea of reason” defined by the set of laws that would be agreed by each citizen were they to constitute a “republic of ends.” The republic of ends thus serves as the rational normative standard to which all legitimate legislation must conform. No law may be promulgated, Kant insists, that “a whole people could not possibly give its consent to” (8:297). The consent that is required for legitimate legislation is thus merely hypothetical consent -- i.e. what each person would consent to were they to take the perspective of the republic of ends.and power of another person. And, yet, while “neo-Roman” republicans define freedom negatively as the absence of mastery by others, “neo-Athenian” republicans define freedom positively as self-mastery. This ideal of self-mastery, however, as it should now be clear, is subject to a variety of interpretations.   According to one line of thinking, the ideal ought to be understood individualistically -- i.e. to be free is to direct your life solely according to your own individual will. The individualistic reading can, in turn, be interpreted in a variety of ways, depending on how one interprets the “individual will” that does the directing -- i.e. as ‘primitive self-mastery,’ as ‘procedural autonomy,’ or as ‘substantive autonomy.’ The individualistic interpretations of self-mastery, however, face a coordination problem: Namely, individual wills, however conceived, will tend to come into conflict. But a theory of political freedom must provide some principle by which the various wills of individuals existing together in a political community can be reconciled or coordinated.   The collective reading of the ideal of self-mastery solves this problem by defining freedom in terms of collective self-government. However, this approach takes a variety of forms. According to the ‘metaphysical interpretation,’ the collective is understood as an ‘organic body’ or ‘social whole’ that is itself the bearer of freedom. By contrast, non-metaphysical interpretations insist that individual human beings are the bearers of freedom, but that they do so in virtue of their contribution to or position in the ‘general will’ by which the community is ruled. I identified what I take to be the three most prominent configurations of this view -- the ‘populist interpretation,’ the ‘deliberative interpretation’ and the ‘rationalist interpretation.’  93 I now want to turn attention to criticisms of the conception of freedom as collective self-mastery -- in particular, to criticisms advanced from within the classical liberal tradition.  2.4. The classic liberal critique of republican “freedom” 2.4.1. Benjamin Constant   Among the earliest and most perceptive critics of Rousseau’s attempt to re-assert the ideal of freedom as collective self-mastery in the modern context is Benjamin Constant.   In a famous essay, Constant (1819) distinguishes between two kinds of liberty -- the “liberty of the ancients” and the “liberty of the moderns.”  The liberty cherished by 126the moderns is concerned fundamentally with the protection of individual rights against incursion -- rights to free expression of opinion, due process, mobility, property ownership, voluntary association, religion, etc. The liberty of the ancients, by contrast, “consisted in exercising collectively, but directly, several parts of the complete sovereignty.” There are two crucial elements to the “ancient” ideal of liberty: First, that the freedom of each citizen consists in his “direct,” “active and constant” participation in the deliberations and decision-making by which the polis is governed. Second, the governing political body possesses a “complete” or “absolute” sovereignty over all the affairs of those under its authority. In consequence, Constant observes, “among the ancients the individual, almost always sovereign in public affairs, was a slave in all his private relations.”  12794 Constant, “The Liberty of the Ancients compared with that of the Moderns”126 Ibid.127 Constant sees in the writings of Rousseau and other republicans an attempt to re-assert this ancient ideal of freedom in the modern context -- to dangerous effect. “Rousseau and many others, had mistaken, just as the ancients did, the authority of the social body for liberty”; further, they “believed that everything should give way before the collective will, and that all restrictions on individual rights would be amply compensated by participation in social power.” By thus “transposing into our modern age an extent of social power, of collective sovereignty, which belonged to other centuries,” they have “furnished deadly pretexts for more than one kind of tyranny.”  128 Constant views the “transposing” of the ideal of collective self-mastery onto modern states as problematic for three reasons. In the first place, in modern states the requirement that each citizen directly participate in the legislative decision-making process is too impractical and costly. In ancient republics this ideal was feasible only because of the vastly smaller size of the states involved and because citizenship was accorded only to a fraction of the total population -- a fraction that enjoyed the leisure-time required for full participation due to the institution of slavery. In large modern states, however, it would be impractical for all citizens to be directly involved in the legislative process. Moreover, even it were not impossible, the burden on individual citizens would be too great to be an attractive option for them. Most citizens in modern states do not enjoy the leisure afforded by having slaves; moreover, the great advantages available through commercial enterprise has made most citizens unwilling to forgo them in order to devote the time and energy required for “direct” and “constant” participation in social power.  95 Ibid.128 Constant’s second and third objections are directed at the other element of the ancient ideal of freedom: the doctrine of (absolute) collective sovereignty. According to the ancient ideal, freedom requires that the community as a whole possess a power and authority over every aspect of the lives of its members.  One reason Constant views 129this requirement as problematic is that he thinks that collective political bodies tend to make poor decisions concerning many aspects of social life -- particularly in the economy. Constant claims that state intervention in the economy is “always a trouble and an embarrassment. Every time collective power wishes to meddle with private speculations, it harasses the speculators. Every time governments pretend to do our own business, they do it more incompetently and expensively than we would.”   130 Constant’s most fundamental objection to the “liberty” of the ancients, though, is not a consequentialist one. It is rather that the doctrine of collective sovereignty is inconsistent with the sovereignty of individuals. Individuals, he insists, have certain inviolable rights that cannot be extinguished by governments of any form. In his biography he writes: ? For forty years I have defended the same principle: freedom in everything, in   religion, in philosophy, in literature, in industry, in politics — and by freedom I   mean the triumph of the individual both over an authority that would wish to   govern by despotic means and over the masses who claim the right to make a  96 Rousseau affirms this conclusion in his description of the social contract as involving a complete 129transfer by every member of all his natural liberties and rights. Following Hobbes, he views this condition as necessary, since a merely partial transfer would result in a divided sovereignty -- which, he insists, is impossible since there could be none with the authority to judge between them in the case of disputes. The result would be a failure to establish sovereignty at all. Sovereignty must be absolute and undivided or non-existent.  Constant, “The Liberty of the Ancients compared with that of the Moderns”130 minority subservient to a majority…The majority has the right to oblige the    minority to respect public order, but everything which does not disturb public   order, everything which is purely personal such as our opinions, everything   which, in giving expression to opinions, does no harm to others either by    provoking physical violence or opposing contrary opinions, everything which, in   industry, allows a rival industry to flourish freely — all this is something individual   that cannot legitimately be surrendered to the power of the state.   131?The “liberty” of the ancients is thus a “pretext” for a “kind of tyranny”  by the majority 132over those in the minority, one that threatens to violate the rights or “individual sovereignty” of those subject to it.  2.4.2. John Stuart Mill  Writing four decades later, John Stuart Mill picks up and develops Constant’s basic criticisms of republican freedom in his masterful On Liberty (1859). Mill introduces his argument with a quasi-historical sketch of the phases in the understanding of the ideal liberty. In the earliest phase, “liberty...  meant protection against the tyranny of the political rulers.”  That is, the demand for “liberty” was conceived of as a demand that 133the power of rulers over the community be subject to certain limitations -- i.e. that individual subjects enjoy certain “immunities” from the scope of that power. However, “[a] time... came in progress of human affairs, when men ceased to think it a necessity of nature that their governors should be an independent power opposed in interest to 97 Constant, Ma Vie, 67 131 Constant, “The liberty of the ancients”132 Mill, On Liberty, 59133their own.”  Consequently, “[w]hat was now wanted was that the rulers should be 134identified with the people, that their interest and will should he the interest and will of the nation.”  Freedom, in this second phase, came to be identified with popular self-135government or self-rule. Since “[t]he nation did not need to be protected against its own will,” it was no longer thought important to think of freedom as a “limitation of the power itself,” since “[t]here was no fear of its tyrannizing over itself.”  However, in the third 136phase, Mill insists, it has become clear that this conclusion contained a fatal error: ? It was now perceived that such phrases as ‘self-government’, and ‘the power of   the people over themselves,’ do not express the true state of the case. The  ‘people’ who exercise the power are not always the same people with those over   whom it is exercised; and the ‘self-government’ spoken of is not the government   of each by himself, but of each by all the rest. The will of the people, moreover,   practically means the will of the most numerous or the most active part of the   people...  137?Thus, like Constant, Mill sees the ideal of freedom as collective self-mastery as licensing “the tyranny of the majority.”   But, while Mill shares Constant’s fundamental worry about republican freedom, his views differ from Constant’s in several important respects. In the first place, Mill is 98 Ibid., 60134 Ibid., 61135 Ibid.136 Ibid., 62137most clearly addressing his objections to what I have called the “populist” interpretation of the ideal of collective self-mastery , whereas Constant focuses his attention of what 138I have called the “deliberative” interpretation . Secondly, while neither of them offers a 139very explicit or precise definition of their preferred alternative to the conception of freedom as collective self-mastery, Constant clearly thinks “modern liberty” must be defined in terms of individual rights, whereas Mill appears to think of liberty as the absence of any “interference” by other human beings in what an individual desires to do -- whether by “physical force” or the “moral coercion of public opinion.”  Thirdly, related 140to the last point, while Constant rests his defense of individual sovereignty, in part, on the grounds that individuals have certain inviolable or natural rights, Mill insists on building his case for the same conclusion on purely utilitarian grounds . 141 Nonetheless, Constant and Mill are at least united in seeing the conception of freedom as self-mastery as flawed on grounds that it fails to place limits on collective sovereignty and so licenses a tyranny of the majority. And while they may not endorse the same definition of freedom, they are at least both explicit in claiming to be defending a wide scope of “individual sovereignty.”  Further, Mill’s basic argument is that a 14299 This is clear from the fact that Mill identifies the demand for self-rule with a demand that “the various 138magistrates of the state should be [the people’s] tenants or delegates, revocable at their pleasure.” A demand which “became the prominent object of the exertions of the popular party wherever any such party existed...” (ibid., 60). As can be seen from the fact that Constant’s first objection to republican freedom is that it is both 139practically impossible and undesirable for the citizens of modern states to be directly and constantly involved in the deliberations and decision-making of the political authority. Though, it should be noted that Mill, like Constant, does describe the aim of the “harm principle” as 140providing a principle according to which we ought to determine the domain in which “the individual is sovereign” (On Liberty, 69; emphasis added). “It is proper to state that I forgo any advantage which could be derived to my argument from the idea 141of abstract right as a thing independent of utility. I regard utility as the ultimate appeal on all ethical questions... (Mill, On Liberty, 69-70). See footnote 61.142society that affords its members the wide scope of individual liberty implied by the harm principle will be one that produces a greater amount of “utility” understood in the “largest sense, grounded on the permanent interests of man as a progressive being.”  This 143claim bears at least a passing resemblance to Constant’s view that a society governed by a centralized and absolute political authority tends to produce less desirable results than one in which private individuals are left to conduct their own affairs according to their own judgement.   2.4.3. Isaiah Berlin  Isaiah Berlin’s “Two Concepts of Liberty” contains arguably the most comprehensive liberal critique of the ideal of freedom as collective self-mastery. Berlin echoes the basic objection to republican “freedom” raised by Mill and Constant, namely that this ideal effectively licenses a “tyranny of the majority.”  He writes that, as “Mill 144explained,” “government by the people was not, in his sense, necessarily freedom at all. For those who govern are not necessarily the same ‘people’ as those who are governed, and democratic self-government is not the government ‘of each by himself’ but, at best, of ‘each by the rest’.”  There is “no great difference” between this “and 145any other kind of tyranny which encroaches upon men’s activities beyond the sacred frontiers of private life.”  Likewise, referring to Constant, he writes that, “He reasonably 146asked why a man should deeply care whether he was crushed by a popular government or by a monarch...”; and, further, “He pointed out that the transference... of unlimited 100 Ibid., 70143 “Two Concepts,” 63144 Ibid.145 Ibid.146authority... from one set of hands to another does not increase liberty, but merely shifts the burden of slavery.”    147 Berlin also agrees, at points, with the basic diagnosis of the fatal error of the ideal of freedom as popular self-government: the doctrine of (absolute) collective sovereignty. He writes: “The liberals of the first half of the nineteenth century correctly foresaw... that the sovereignty of the people could easily destroy that of individuals” (ibid). “Unlimited authority” or “absolute sovereignty” in “anybody’s grasp was bound... sooner or later to destroy somebody”; the “mere fact of the accumulation of power itself, wherever it might happen to be” was the “real cause of oppression” since liberty is “endangered by the mere existence of absolute authority as such.”  “[T]he doctrine of absolute sovereignty 148was a tyrannical doctrine in itself. ” 149 What is most original in “Two Concepts,” at least as concerns the ideal of collective self-mastery, is that (a) Berlin identifies the liberal concern for individual sovereignty with a commitment to the “negative” concept of liberty and (b) he extends the liberal critique to encompass what I have called the ‘rationalist’ interpretation of the ideal.  2.4.4. Individual sovereignty and negative liberty  As noted, Constant sought to assert the inviolability of a certain domain of individual sovereignty against the claim to absolute popular sovereignty implicit in the ideal of freedom as collective self-mastery. Throughout his essay, Berlin presents himself as being of one mind with Constant, suggesting that “[n]o one saw the conflict 101 Ibid.147 Ibid.148 Ibid., 164149between the two types of liberty more clearly.”  At points he gives voice to his basic 150objection against positive liberty in Constant’s terms, warning that “the sovereignty of the people could easily destroy that of individuals,”  and suggesting that “no society is 151free unless it is governed by… two interrelated principles: first, that no power, but only rights, can be regarded as absolute... and, second, that there should be frontiers, not artificially drawn, within which individuals should be inviolable.”  However, elsewhere, 152Berlin equates the acceptance of the idea of individual sovereignty with the acceptance of the “negative” or non-interference conception of freedom. Thus, he writes that “[Constant] saw that the main problem for those who desire ‘negative’, individual freedom is not who wields this authority, but how much authority should be placed in any set of hands.”  Berlin’s identification of individual sovereignty with the enjoyment 153of non-interference underlies his framing of the basic contrast between ‘negative’ and ‘positive’ liberty: “But there is no necessary connection between individual liberty and democratic rule. The answer to the question ‘Who governs me?’ is logically distinct from the question ‘How far does government interfere with me?’ It is in this difference that the great contrast between the two concepts of negative and positive liberty, in the end, consists. For the ‘positive’ sense of liberty comes to light if we try to answer the question, not ‘What am I free to do or be?’, but ‘By whom am I ruled?’.”  154102 “Two Concepts,” 163150 Ibid.151 Ibid., 165; emphasis added.152 Ibid., 162153 Ibid., 130154 However, it is worth flagging at this point that, contra Berlin, the acceptance of an inviolable individual sovereignty does not logically commit one to the non-interference conception of freedom. In fact, the idea of “sovereignty,” as it is normally understood, implies something richer than merely ‘a domain in which one is not-interfered with.’ It has to do, rather, with the possession of a supreme authority or power within a certain domain. It thus implies a right or fact of control or the power to rule. This power must be supreme in the sense that it is superior to all other powers or authorities within its domain. The ideas of individual sovereignty and non-interference are thus not equivalent. One can enjoy non-interference within a certain domain without enjoying a sovereignty within that domain -- as when another agent possesses a power or authority to interfere with you in that domain, but opts to leave you unhindered. Understood in this sense, individual sovereignty requires not merely that one not be interfered with in one’s private affairs, but that no one else possesses the power or authority to do so. I will return to this point in section 2.5.1 below. 2.4.5. Berlin’s critique of the ‘rationalist interpretation’  The problem of the tyranny of the majority raised by Constant and Mill applies most clearly to what I have called the “deliberative” and “populist” interpretations of the ideal of collective self-mastery. Part of the originality of “Two Concepts” is that it extends the liberal critique of republican freedom to encompass this rationalist version of the ideal. In fact, I would suggest that the critique of ‘rationalist republicanism’ is the most persistent and central aim of the essay.   As I’ve already observed, Berlin describes the rationalist version of the ideal of freedom as self-mastery as emerging by a series of conceptual steps. In the first step, 103freedom is identified with what I earlier called “primitive” self-mastery: To be free is to have the power to direct one’s actions according to one’s own will, to be one’s own master, rather than being subject to the will of someone else. In the second step, the definition of “slavery” is broadened to include non-rational psychological influences on the will.  In consequence, the ideal of self-mastery now requires that an individual 155direct her actions according to her “rational” or “true” will -- a psychological achievement of rational self-regulation rather than merely a condition of not being subjected to another person’s power. In the third step, the reason according to which individual wills are directed is identified with the collective rationality of the community as a whole. This development is a consequence of the need to coordinate the individually rational but potentially conflicting aims of individuals into a coherent system or free society.     Now, as I have shown, the collective interpretation of the ideal of self-mastery actually comes in different varieties: the populist, the deliberative and the rationalist. The populist and deliberative versions both require the actual participation of individual members in an actual collective decision-making process. The populist version takes as inputs to the decision-making process whatever preferences individuals have -- i.e. what Berlin would call their “empirical” preferences. The deliberative version takes as inputs to the decision-making process the opinions individual members arrive at after deliberating together about what social arrangement would best serve their common interests. Because both of these versions define self-mastery in terms of actual decision-making processes, and because actual decision-making processes usually do not result in consensus, both of them ultimately defer to the preferences or opinions of 104 “[M]ay I not... be a slave to nature? Or to my own ‘unbridled’ passions? Are these not species of the 155identical genus ‘slave’ -- some political or legal, others moral and spiritual?” (Ibid., 132).the majority. As we’ve seen, Berlin shares the worry raised by Constant and Mill that they, thereby, license a tyranny of the majority.   The rationalist version, however, does not defer to the majority.  According to it, an individual is free insofar as she is subject only to laws that would be agreed to by all members of the collective were they to engage in an ideally rational deliberation about the best way to coordinate their aims. It thus requires a kind of consensus -- though only a hypothetical one. Berlin raises two main objections to the rationalist ideal of collective self-mastery: (1) the “divided self” objection, and (2) the “value pluralism” objection.  2.4.6. Divided self objection  The “divided self” objection seeks to undermine this conception of freedom by appeal to the consequences such a view has for the relation between coercion and freedom. The source of the problem is to be found already in the second step Berlin identifies in the evolution of the ideal of self-mastery. At this stage, to be “free” requires both the absence of external slavery, but also the absence of internal slavery, in the form of irrational aims and desires. It follows that an action can be unfree if it proceeds from an irrational desire, even if the agent in question is not subject to any external coercive powers or interferences. To be free is to have one’s actions directed by one’s “rational” (i.e. “true”) self, not merely by one’s “empirical” self.  But such a view carries counter-intuitive and dangerous implications, according to Berlin. It licenses the coercion of “empirical” individuals in the name of their “rational” selves, i.e their true but misperceived wills. Actually, there is a stronger and a weaker formulation of this objection.  105 On the stronger formulation, other agents who perceive the individual’s “rational” or “true” will better than she does herself can, in effect, “force her to be free” by coercing her into doing what she would do if she were fully rational. Berlin sees this result as problematic partly because he thinks this sort of logic serves as a convenient rationalization for those with power to coerce others in the name of their “true freedom.” But, even if it is employed genuinely, rather than as a mere pretext, Berlin clearly believes that the idea that individuals may be “forced to be free” serves as a kind of reductio ad absurdum for the underlying definition of freedom.   The soundness of this stronger formulation of the divided self objection is questionable though. If the proponent of the ideal of rational self-mastery holds, as I think is most plausible, that for a person to be free with respect to a particular action it must be the case both that the action be rationally motivated AND that it not be coerced by others , then “forcing others to be free” would in fact be impossible -- since such an 156action would no longer be motivated by the agent’s own reasoning. One might be able to force someone to be free by employing coercion to get them to be more rational (e.g. by forced education), but even then their participation in the process would not be free, even if the outcome of the process is that the person is capable of greater freedom.  In any case, the force of Berlin’s objection does not rest on the stronger “forced to be free” formulation. According to the weaker formulation of the objection, the ideal of rational self-mastery is problematic because it implies that coercive interference with the actions of sane adult individuals whose aims are not sufficiently rational at least does not diminish their freedom -- since these were not free actions to start with. This weaker formulation is enough to generate the problems Berlin is concerned about. It licenses 106 cf. Christman, “Liberalism and Positive Individual Freedom,” 347156those who are more enlightened to coerce sane, adult human beings into doing what they would do if they were sufficiently rational without thereby diminishing their freedom. Taken to its logical extreme, Berlin insists that this way of thinking about freedom can end up serving the cause of justifying totalitarianism -- an absolute and despotic rule of “empirical” individuals -- by defining away the plea for liberty.   As noted, the divided self problem arises even on the individual version of the ideal of rational self-mastery (Berlin’s “step two”); however, it applies with at least equal force to the collective version (“step three”) we are presently focused on. According to it, an individual is free insofar as she is subject only to laws that would be agreed to by all members of the collective were they to engage in an ideally rational deliberation about the best way to coordinate their aims. As with the individual version, the collective version implies a distinction between the empirical will of an individual and his rational will -- only in this case his rational will is understood as what he would agree to were he to deliberate rationally with other members of the community under a consensus constraint. Moreover, as with the individual version, coercion of the empirical will of a person does not diminish his freedom so long as it does not constrain his rational will. Consequently, the collective version of the ideal of rational self-mastery is subject to the same troubling implications as its individualist cousin. In fact, it might seem that the problem is amplified in this case, since there is likely to be an even greater distance between what an individual actually wills and what he would will subject to the constraint of consensus decision-making than there would be between what he actually wills and what he would will if he were to reason effectively about his personal ends.   ?1072.4.7. Value pluralism objection  Whereas the force of the divided self objection depends on the counter-intuitive and unattractive consequences of conceiving of freedom as collective rational self-mastery, Berlin’s second main objection questions the truth of one of its basic assumptions -- namely, the assumption that there must exist some agreement that all citizens would converge on in an ideal deliberation.   Why assume that individuals in a given society would all agree on how to order their collective life together if they were only to deliberate rationally? According to Berlin, the basis for this assumption must be the belief in value “monism”  -- i.e. the belief that 157there exists a single underlying value or ordinal ranking of values in terms of which all practical questions can in principle be given a single correct answer.  If a “single 158criterion” of this sort exists, then it would be reasonable to suppose that “in an ideal world of wholly good and rational men” it would be possible “to reach clear-cut and certain answers” to all ethical problems.   159 However, Berlin insists that value monism is false. Values are irreducibly and incommensurably plural, and so there cannot a single criterion according to which all ethical decisions can be given a single rational answer. Instead, moral decision often involves a simple choice wherein something of value must be sacrificed. This explains 108 “Two Concepts”, 170157 Berlin initially suggests that an even stronger claim underlies the assumption of a rational consensus -- 158namely, the “conviction that all the positive values in which men have believed must, in the end, be compatible, and perhaps even entail one another” (ibid., 167) such that no sacrifices would be necessary in an ideal “final solution” (ibid., 168). If by this he means that all the things people value can be jointly maximized in a coherent and feasible social order, then it is doubtful that anyone has every seriously held such a view. Even a value monist (see below) can deny this claim: the value monist need only claim that there is a single underlying value according to which all good things can be ranked and moral decisions made, not that no good thing will need to be sacrificed for the sake of maximizing good overall. In any case, the assumption that an ideal rational consensus must exist need not depend on the stronger claim. Ibid., 170159why many of the most difficult moral decisions carry with them an unresolved element of regret and loss. But given that many ethical decisions will be of this sort, there is no reason to suppose that all rational individuals will make the same choices. And if they would not, or at least if there is no reasonable grounds to think they would, then the rationalist republican is left without a determinate standard according to which they can judge whether individuals in a given society are indeed free.     I must confess that I find this argument quite compelling; however, not everyone is convinced that value monism is mistaken.  The question of whether value monism 160or value pluralism is true remains the subject of ongoing debate in contemporary moral philosophy.  However, Berlin’s objection to rationalist republicanism can be made on 161the basis of a weaker, less-controversial claim -- namely: even if there does exist some agreement that all citizens would converge on in an ideal deliberation, what that agreement would consist in is rationally inaccessible to any actual person. One reason for thinking this is that, particularly in modern societies, there exists a reasonable pluralism of world views, or what Rawls calls “comprehensive doctrines.” Citizens have different views about God or transcendence, about the meaning or purpose of life, about right and wrong, good and evil. These background commitments will result in different, and at points conflicting views about how society ought to be organized. Moreover, the differences between these background commitments may not be subject to resolution by any shared conception of rationality. In consequence, the rationalist republican is left without a rationally accessible and determinate standard according to which they can judge whether individuals in a given society are indeed free.   109 cf. Hurka, “Monism, Pluralism and Rational Regret”160 For an excellent survey of the contemporary pluralism-monism debate, see Mason, “Value Pluralism”161 Each of Berlin’s two main objections offers an independent reason to reject the rationalist republican ideal of freedom; however, in concert, their force is amplified. Reflecting on the first objection, it might seem that there isn’t anything terribly mysterious or troubling about distinguishing between what a person actually wills and what they would will if they were to consider their options from a more rational point of view. Furthermore, in certain cases, it seems quite unobjectionable to interfere in what a person is actually intending to do on grounds that, from a better vantage point, we know he doesn’t really want to do it. After all, as Mill (and before him Locke) pointed out: If a man is headed for a crossing where, unbeknownst to him, the bridge has been washed out, we may reasonably suppose that he doesn’t really want to fall in the river, and may interfere to prevent him from continuing on his course, without thereby impugning his will.   However, this case is atypical in that it involves reconstructing what an individual would will concerning a very specific course of action from a hypothetical position that involves adding one simple, uncontroversial bit of empirical information of obvious significance. When it comes to issues more typical of political decision-making, we are generally dealing with matters far more complex and subject to uncertainties and indeterminacies. They typically involve larger questions concerning what a good life consists in, the nature of human existence, broader principles of right and wrong, and much more complicated and disputed empirical facts. Judgements concerning what a person would will from a more rational perspective are thus highly dubious. Moreover, interference with a person’s actions on grounds of such judgements seems inconsistent with her liberty. The force of this point is only amplified when the “rational perspective” 110we are making judgements from is the ideal collective deliberations of the community as a whole. Here the distance between the actual will of the individual being interfered with and her “rational” will seems potentially very great. Moreover, in light of Berlin’s second objection, we have little reason to suppose there even is such a unified collective rational perspective, or at least that if there is that we could have rational access to it. The ideal of freedom as collective rational self-mastery thus seems at best to license interferences with individuals lives on grounds that are unreasonable even if sincere, and at worst to constitute a mere pretext for those with power to impose their own will on the less powerful.   2.5. Defining the liberal insight  Let’s sum up the results of the last sections: ‘Neo-Athenian’ republicans hold that freedom consists in collective self-mastery or self-government. This ideal has been interpreted in different ways, the most prominent among them being the populist, the deliberative and the rationalist interpretations. Liberals like Constant, Mill and Berlin have objected to the first two versions on grounds that they license a tyranny of the majority. Berlin extended this liberal critique to the third, rationalist, interpretation on similar grounds -- namely, it provides a too easy, and in fact misconceived, justification for exercising coercive power over the lives of recalcitrant individuals in the name of freedom, or at least without its objection.      How then should we understand the basic liberal insight that underlies these objections? In my view, Berlin is right when he claims that the central unifying insight of republicanism’s liberal critics is that the authority of government, whatever its form, must 111be kept at bay if individual freedom is to be preserved.  However, Berlin goes on to 162claim that underlying this insight must be the view that “individual freedom” consists in non-interference.  In what follows, I aim to challenge this second claim. In particular, I 163will argue that the basic liberal insight is equally consistent with Constant’s view that “individual freedom” consists in “individual sovereignty.” Moreover, because the definition of freedom as individual sovereignty is not equivalent to the definition of freedom as non-interference, Constant and Berlin offer different views of what is it means for “the authority of government to be kept at bay.” 2.5.1. Constant versus Berlin  According to Constant, the fundamental error implicit in the conception of freedom as collective self-mastery is the doctrine of (absolute) popular sovereignty. That is, it recognizes no limit on the power and authority of the collective to rule over the lives of its members. Thus, on Constant’s view, the fundamental liberal -- or “modern” -- insight is that freedom requires the recognition and protection of an inviolable sovereignty of individuals. While this individual sovereignty is inviolable, it is of course not absolute in the sense of having unlimited scope (otherwise the sovereignty of each would be in conflict with that of every other). Constant sees the boundaries of the domain in which each individual is sovereign as defined by the now familiar list of “liberal” rights.  112 “Two Concepts,” 126162 Berlin wrote as though the commitment to “negative liberty” is what unified not just the classical 163English political philosophers (including Hobbes, Locke, Bentham and Mill) but also other “liberal” thinkers like Constant and Tocqueville in France, and Jefferson and Paine in America (ibid., 123-4; 126). However, while there is certainly a line of thinking running from Hobbes through to Bentham and Mill that is explicitly committed to the conception of freedom as non-interference, others in this list are not plausibly interpreted in this way (see Chapter 5, section 5 of this dissertation). As I argue below, Constant does not define freedom as non-interference; nor, as I shall argue in Chapter 5, does Locke or the American Founders. There is not, in my view, a shared conception of political freedom underlying the views of these various figures, but rather merely a shared basic concern as defined by the “liberal insight.” Berlin claims to agree with Constant’s diagnosis of the error contained in the republican conception of freedom, namely the doctrine of absolute collective sovereignty. This error is common to each of the populist, deliberative and rationalist interpretations of self-mastery. Each of them views rule by the decision-making of the collective body -- whether bargain-based or debate-based, majoritarian or consensual, actual or hypothetical -- as the sole defining criterion of freedom. Berlin also claims to endorse both Constant’s contrary assertion of the doctrine of individual sovereignty. However, he equates the idea of “individual sovereignty” with the enjoyment of a domain of negative liberty. Thus, on Berlin’s view, the underlying the basic liberal insight is the view that freedom ought to be conceived of as non-interference and that a free society is one in which each individual is accorded a certain domain in which they are not interfered with by others.    Yet, these analyses are not equivalent: one can enjoy a certain domain of non-interference without being sovereign within that domain. To be sovereign in a domain is not merely not to be interfered with by others within that domain, but rather to possess a power or authority over that domain that is not, in turn, subject to the authority or power of others (i.e. it is supreme authority).  One can enjoy not being interfered with in an 164action or set of actions without being sovereign in this sense, as exemplified by the case of the slave with a permissive master. Non-interference does not entail individual sovereignty.  113 It is generally accepted that the notion of “sovereignty” denotes a supreme authority within a certain 164domain, and that authority is “supreme” just in case it is superior to all other authorities under its purview (Philpott, “Sovereignty”). But it is also important to note that authority can be supreme, without being absolute. Absoluteness refers not to the extent or character of sovereignty, which must always be supreme, but rather to the scope of matters over which a holder of authority is sovereign. Constant’s notion of “individual sovereignty” is of a supreme but non-absolute authority or power -- i.e. a supreme authority over a certain limited domain of action types (namely, those one has a natural right to).   In reality then, Berlin does not share Constant’s view that freedom consists in individual sovereignty. Moreover, Berlin does not even share Constant’s view about the underlying source of the error of the ideal of collective self-mastery. Both think that defining freedom in terms of popular sovereignty is problematic, but they offer different accounts of what the error is. On Constant’s view, the error is one of mis-locating sovereignty, i.e. of ascribing an absolute sovereignty to the people as a collective -- since this entails the denial of individual sovereignty. On Berlin’s view the error is to think that the issue of sovereignty is relevant to individual freedom at all. Thus, he claims that “positive” liberty is concerned with by whom I am ruled, whereas “negative” liberty is concerned with how far I am interfered with.  It’s true that Berlin sees the 165acceptance of the doctrine of absolute popular sovereignty as tending to lead to the violation of the negative liberty of individuals. However, in principle at least, it need not do so. If the collective were to sovereignly decide to allow an appropriately wide degree of non-interference to individual citizens, then it’s sovereignty would be fully consistent with the freedom of individual members. In short, for Berlin, the issue of sovereignty is irrelevant to the definition of freedom. It is for this reason that he can claim that “it is perfectly conceivable that a liberal-minded despot would allow his subjects a large measure of freedom.”  166 Of course, Berlin and Constant’s views will yield many equivalent judgements about the freedom or lack thereof of individuals in different conditions. For instance, both will see freedom as fatally compromised when a democratic government enforces laws that severely constrain the ability of individuals to form voluntary associations, 114 “Two Concepts,” 130165 Ibid., 129166express their opinions freely in public, practice their religion in way that does not prevent others from doing the same, or possess and exchange property. These common judgements, no doubt, form the basis for their shared insight that the republican conception of freedom licenses an unacceptable majoritarianism. However, this does not negate the fact that underlying these common judgements lie distinct bases. On Berlin’s view, such acts of majoritarian “tyranny” compromise freedom because they involve interference with what individual’s could otherwise do; On Constant’s view, they compromise freedom because they involve an arrogation of authority by the majority over domains of life in which individuals are rightly sovereign.        Thus, when Berlin claims that the authority of government must be kept at bay if individual freedom is to be preserved, he in fact means that the authority of government ought not to be exercised in such a way that it interferes with the choices of individuals beyond an acceptable limit. Whereas, when Constant makes this claim, he means that the mere assumption of such an authority -- whether or not it is manifests in actual interference -- violates the freedom of individuals.  2.6. Conclusion  In this last section, I claimed that the basic insight underlying the liberal objections to the neo-Athenian conception of freedom is that it recognizes no limit on the power and authority of the collective to rule over the lives of individual members. That insight, I argued, can be supported either by accepting Berlin’s conception of freedom as non-interference or by accepting Constant’s conception of freedom as individual sovereignty. However, while both these views are inconsistent with the definition of freedom in terms of popular sovereignty, they are not equivalent. In 115particular, one can accept the basic liberal insight without accepting the definition of freedom as non-interference.   This last conclusion is important for the argument of the chapters that follow. In the next chapter (Chapter 3), I will argue that there are in fact good reasons to reject the non-interference conception of liberty in favour of the conception of freedom as non-domination. Doing so, it should now be clear, needn’t involve denying the basic liberal insight. Indeed, in Chapter 4 I will argue that the neo-Roman interpretation of the ideal of non-domination is flawed precisely because it entails such a denial. Like the neo-Athenian ideal considered in this chapter, it is ultimately committed to the doctrine of popular sovereignty. In Chapter 5, I propose and defend a distinct version of the ideal of non-domination -- one that carries an implicit commitment to the doctrine of individual sovereignty, and so avoids the problems of both neo-Athenian and neo-Roman republicanism.
116Chapter 3 Freedom as non-Domination  ?In Chapter 2, I distinguished two streams of thinking about freedom within the republican tradition -- the “neo-Athenian” and the “neo-Roman.” I focused on the latter, according to which freedom is thought to consist in collective self-mastery or self-rule. That approach, as liberal critics have argued, ultimately falters because it fails to place any limits on the authority of “the people” to control the lives of individuals. The lesson Isaiah Berlin drew from this is that freedom ought to be understood as the absence of interference -- whether by a democratic body or anyone else -- in the choices of individuals. In this chapter, I consider the counter-critique advanced by “neo-Roman” republicans against the ideal of freedom as non-interference, particularly as articulated by Quentin Skinner and Philip Pettit. I argue that their objections are ultimately compelling, and that there is good reason to favour their conception of freedom as non-domination. 3.1. Neo-Roman republicanism  In the last chapter I suggested that the basic unifying conviction of republicans is that the fundamental antinomy to freedom is slavery. To have a master that can direct your actions according to his will and pleasure is the essence of unfreedom -- whether that master is a literal slave-holder, an absolute monarch, or some other possessor of arbitrary power. Slavery is not merely about being constrained or interfered with in this or that action, it is about being subordinated to another, under their power and will.  1173.1.1. Neo-Roman versus neo-Athenian freedom  However, while republicans are unified in their understanding of unfreedom, they do not all agree about what freedom consists in. Neo-Athenian republicans define freedom as self-mastery -- as being one’s own master, having the power to direct one’s own life according to one’s own will. For this reason they have tended to look to the ancient Athenian democracy as a model for the free republic, a model in which each citizen participates in the collective deliberations and legislation by which the community is governed.  In contrast, neo-Roman republicans define freedom in a merely negative fashion: Freedom consists in the absence of mastery -- or “domination” -- by others. The basic aim here is not to rule, but rather not to be ruled by others. For this reason they have looked to the Roman Republic rather to the Athenian democracy as their model for the free republic. In the Roman Republic, the people did not necessarily have a hand in the making of laws so much as a power to contest and perhaps veto laws that did not accord with their interests. The key features in this model are the dispersion of power across different bodies with avenues for contestation and mutual invigilation.      167 Neo-Roman republicans thus define freedom as non-domination, and have sought to repudiate a couple tendencies of the ideal of self-mastery. In the first place, while they share the view that the freedom of citizens requires that there be mechanisms of democratic accountability, they have tended to favour a contestatory rather than a participatory model of democracy. The contestatory model requires only 118 In “Are Freedom and Liberty Twins?”, Hanna Pitkin writes: “The Roman plebs struggled not for 167democracy but for protection, not for public power but for private security. Of course they sought public, institutionalised guarantees of that security. But libertas... was ‘passive’, ‘defensive’, ‘predominantly negative’” (534-5).that citizens have avenues to effectively challenge political decisions that conflict with their interests, rather than that they actually be involved in the making of those decisions. In the second place, neo-Roman republicans have distanced themselves from the rationalist tendencies associated with the ideal of self-mastery. Being ‘free’ consists merely in not being under the dominion of other people, and so it does not require any psychological achievements of rational self-regulation or self-realization. The conditions of freedom are entirely external. For this reason, the ideal of freedom as non-domination is not vulnerable to the “divided self objection” (2.4.6) that plagues at least certain forms of the ideal of self-mastery.  3.1.2. Neo-Roman republicanism: three fundamental claims  The neo-Roman theory of political freedom can be summarized in three basic claims: (1) Freedom consists in the absence of domination by others (rather than the absence interference), where to be dominated is to be subject to the arbitrary power of another agent; (2) Arbitrary power is power that is not forced to track the interests of those subject to it; (3) In order for all social powers to be forced to track the interests of those subject to them, they must be subject to democratic mechanisms of contestation and control.  ?The aim of this chapter is to consider the case for the first of these three basic claims -- that freedom ought to be conceived of as non-domination rather than non-interference. In the next chapter (Chapter 4) I will focus on the second and third claims in more detail, and subject them to critical examination.  1193.2. Non-domination and non-interference  Claim ‘1’ states that freedom ought to be conceived of as non-domination rather than non-interference. To begin, let’s try to make the contrast between these two conceptions of freedom more precise. 3.2.1. Freedom as non-interference  According to Berlin’s “negative” concept of liberty, an individual’s freedom is diminished if and only if another human agent interferes with her ability to do something she could otherwise have done. Thus, a particular person ‘P’ is free with respect to a particular action ‘A’ so long as no other person or persons interferes with P’s ability to do A.   While Berlin does not specify with precision what is to count as an “interference,” he clearly thinks it requires an intentional action by a human being that prevents another person from doing something.  He also cites Helvetius approvingly, where he writes: 168“The free man is the man who is not in irons, not imprisoned in a gaol, nor terrorized like a slave by the fear of punishment…” , suggesting Berlin understands interference 169primarily in terms of the use or threat of force. If that is right, then we can interpret Berlin’s view as follows: An action counts as an “interference” if it involves: (a) Using force to physically compel another person to do or not do something -- e.g. by manhandling them, blocking them with a wall or a fence, prodding them with a sharp or electrified object, evacuating them by releasing a noxious gas into the area, etc.; (b) Rendering an action another person could perform more difficult, costly or painful beyond a certain threshold -- e.g. by placing land mines along their path, by confiscating 120 “Two Concepts,” 122 168 Ibid.169their property for engaging in certain religious practices, or installing a device that emits an extremely loud or high frequency sound in the area in which they frequent; (c) An explicit or implicit threat to either physically compel or cause harm to a person or someone dear to them if that person does not comply with a demand for them to do or not do something. I shall refer to interferences of any of these sorts simply as “the use or threat of force.”  On the non-interference conception of freedom, it follows that laws, insofar as they forbid or require certain actions and attach sanctions to non-compliance, always restrict freedom. Such laws are essentially conditional threats. Of course, it’s true that the overall impact of a set of laws may be to increase an individual’s freedom if the interferences they prevent are greater than the interferences they impose. However, even in this case, they do so only at some cost to liberty.   3.2.2. Freedom as non-domination   Philip Pettit has offered the clearest analysis of the conception of freedom as non-domination. On Pettit’s account, an individual’s freedom is diminished if and only if another human agent possesses the power to interfere on an arbitrary basis with the actions that individual could otherwise perform.  Freedom is thus threatened not 170simply by actual interferences, but by being subject to the mere possibility of interference. Yet not just any power to interfere diminishes freedom; rather only the power to interfere arbitrarily -- i.e. according to the power holder’s mere will or pleasure. When a person is subject to the arbitrary power of another they are, to use Pettit’s phrase, dominated. Thus, according to the conception of freedom as non-domination, a 121 Pettit, Republicanism, 52170particular person ‘P’ is free with respect to a particular action ‘A’ so long as no other person or persons has the power to interfere arbitrarily with P’s ability to do A.   But what constitutes arbitrary power? What distinguishes it from non-arbitrary and thus non-freedom-diminishing power? In the next chapter I will consider this question in more detail, but for now it will be necessary to at least adopt a provisional account of non-arbitrary power. Most uncontroversially, arbitrary power is, as Pettit claims, power that is exercised merely at the will or pleasure of the power holder.  It is 171thus a kind of discretionary power. It is not constrained by anything outside the whims or desires of the one holding the power; in particular, it is not constrained by principle, procedure or law. For this reason it widely held that non-arbitrary power is power that is subject to something like the rule of law. It thus has a more stable and predictable character to the way it is exercised. One paradigm case of arbitrary power is that held by an absolute ruler, since such a ruler stands above any law, and may override or bend the law to suit his will, rather than the other way around. Another paradigm case is the power a master holds over his slaves, since his slaves are possessions and may be used or abused according to his pleasure. Arbitrary power need not be exercised in a nefarious or cruel manner. A dictator may fancy himself “father of the nation,” and seek to use his power in what he deems benevolent ways. But the point remains that whether or not he uses his power for noble or ignoble purposes is entirely up to him.   It is widely believed that while something like a rule of law -- or “procedural” -- condition is necessary for power to count as non-arbitrary, it is not sufficient. After all, the law, even if vigorously enforced, may be designed to serve the interests of those in power, or at least contrary to the interests of those subject to it -- without any means for 122 Ibid., 55171them to dispute or remedy the situation. For this reason, Pettit places a second condition on power that is to count as non-arbitrary -- namely, that it must be “forced to track the interests and ideas” of those subject to it.  This condition requires not only 172that non-arbitrary power be exercised in ways that conform with the interests of its subjects, but that it must be “forced” or constrained to do so. For the purposes of this chapter I wish to weaken this condition somewhat, for reasons that will become clear in the next. According to this weakened condition, non-arbitrary power is power that is accountable and responsive to the relevant interests or claims of those subject to it. Let’s call this the “accountability condition.” According to our provisional definition, then, arbitrary power is power that fails to meet either (a) the “procedural condition” or (b) the “accountability condition”; non-arbitrary power is power that meets both. 3.2.3. Non-interference and non-domination contrasted  The ideals of freedom as non-domination and non-interference take contrasting views of the essence of freedom, and in fact render different judgements about the freedom of individuals in different conditions. On the non-domination account, freedom is essentially about the relations of power between different social agents; on the non-interference account, freedom is essentially about the degree to which individuals actually interfere with each other.   In consequence, on the non-domination account, interference is neither a necessary nor a sufficient condition for someone’s freedom to be diminished. It is not necessary since an individual may be subject to the arbitrary power of another agent, and yet that other agent may decline to actually interfere with her ability to undertake a particular action or set of actions. On the non-domination account, it remains the case 123 Ibid.172that such an individual is unfree with respect to those action possibilities, even though not they are not actually interfered with. However, on the non-interference account the individual remains free. Interference is also not sufficient to diminish an individual’s freedom on the non-domination account. If another agent interferes with ability of an individual to perform a particular action, but does so on a non-arbitrary basis, that agent does not diminish that individual’s freedom. Thus, for example, when a police office, operating within the law, exercises her powers of arrest and detainment to prevent a would-be murderer from carrying out his homicidal plans, she does not thereby diminish his liberty. On the non-interference account, however, she does.   3.3. The case for the non-domination conception   It should be clear then that the non-domination and non-interference conceptions of freedoms constitute distinct alternatives. Which is the superior account of our ideal of political freedom? I think there are at least two reasons to favour the non-domination account: (1) The non-domination account is able to register certain “self-editing effects” as symptomatic of a lack of freedom, while the non-interference account is not; (2) The non-domination account captures what we find most inimical to freedom in both slavery and despotism -- namely the element of “alien control” -- while the non-interference account does not. Let’s consider these in turn. 3.3.1. Self-editing effects  According to our provisional definition, power can be arbitrary either in the sense that it is not constrained to operate according to law or in the sense that it is not accountable and responsive to the relevant interests of those subject to it. When power is arbitrary in the first sense it will tend to be exercised in unpredictable ways. Without a 124stable, public set of rules or procedures constraining their actions, those holding power will have license to use that power in whatever way they see fit in the moment. Any number of factors may influence how they exercise their power -- their perception of their own interests, their judgement about what would be best for others, a passing fancy, a paranoid fixation, an unconscious resentment, a naively embraced ideology, etc. In the absence of the constraint of law, arbitrary power will therefore tend to be exercised in unpredictable ways. This unpredictability is likely to be amplified if such power is also arbitrary in the second sense, and especially so if a single individual is the power holder -- as in the case of an absolute dictator or slave-holder. From the point of view of those subject to such power, it will be difficult to anticipate how such a powerful person is likely to respond to various actions they may wish to undertake. Such subjects will have to make their choices in conditions of increased uncertainty and vulnerability.  Of course, power that is arbitrary in this way need not be entirely unpredictable. It may be that the power holder has a fairly fixed character, set of habits, or interests such that one is able to make a good guess as to how he is likely to respond in certain situations. For instance, a slave may come to know his master’s tendencies well enough to know when he is likely to lash out. Moreover, even if a subject of arbitrary power doesn’t have intimate knowledge of the particularities of the power holder, as e.g. in the case of an ordinary subject of a king, she can at least rely on general knowledge about human nature, about the tendencies of those in positions of power, and whatever other facts about the king and his circumstances are commonly known. Nonetheless, while these considerations may reduce the unpredictability of arbitrary power, they do not reduce it to the point that having a law-governed form of power would. Moreover, such 125predictions cannot be counted on in the same way that a reliable legal regime can (after all, a slave cannot object to a beating by his master on grounds that the master is normally in a pleasant mood on Sunday mornings!), and it seems likely that subjects of such power will be inclined to take a more cautious or risk-averse approach to how they depend on such predictions. Arbitrary power thus exhibits only a weak predictability, in comparison to the strong predictability of legal power.   If it is true that arbitrary power (in the first sense) tends to be unpredictable in these ways, then it is also likely that individuals subject to such power will tend to self-edit their actions in various ways in anticipation of how the power-holder might otherwise respond.  There are at least three types of self-editing effects we might 173expect to see: First, individuals will do things they might not otherwise have done to curry favour with the powerful. They might engage in public flattery of the power-holder, they may display posters of the “beloved father” in their homes and businesses, they may bestow gifts and honours on the powerful, etc. Second, individuals will avoid doing things that they might otherwise have done in order to avoid attracting notice or detection. They will avoid public actions or speech that is likely to cause offense, they will take potentially displeasing activities underground where they can be done in secret. Third, individuals will comply with requests by the powerful to do things they might not otherwise have been inclined to do. They will “volunteer” their sons for the war effort, they will invite the king’s soldiers to stay in their homes or eat their food, they will join the party and “enthusiastically” attend its rallies, etc.    Two points should be noted about these self-editing effects. In the first place, they depend on what I called the weak predictability of arbitrary power. If the powerful 126 See Pettit, Republicanism, 87; Skinner, “A Third Concept of Liberty”, 406-8173were effectively and reliably constrained by law, there would be no need on the part of its subjects to self-edit their actions. They could simply conduct their lives in the assurance that, so long as they obey the law, they need not fear any (additional) adverse responses from those in power. Strong predictability would render the self-editing measures unnecessary. On the other hand, if the powerful were entirely unpredictable, there would be no point in self-editing one’s actions, since one would have no basis for determining what sort of actions would be more likely to minimize unwanted interferences. In the second place, and most importantly, these self-editing effects intuitively constitute diminishments of the freedom of those displaying them, even though they need not involve any actual interferences or threats of interference on the part of the powerful. The mere possibility of such interference is enough to generate these anticipatory measures of self-editing. If that is right, then we have one reason to favour the non-domination account of freedom over the non-interference account. 3.3.2. Slavery and despotism  A second reason to favour the non-domination account is that it does a better job capturing what we find objectionable about both slavery and despotism. (i) Slavery  If anyone lacks freedom, surely a slave does. Let’s define slavery as a condition in which (a) one person (the slave) is under the power and authority of another person (the master), and (b) in which the master’ ability to exercise his power over the slave is not, in turn, subject to any one else’s power or authority.   Slavery comes in different forms depending on the scope of the power and authority the master holds over the life of the slave. At the extreme end, a master’s 127power and authority over the slave consists in absolute ownership -- as is the case with “chattel” slavery -- and so is unlimited in scope. However, some slaves are subject to a more limited form of subjection: In more circumscribed forms of slavery, there are legal or customary constraints on the master’s authority over the slave -- e.g. where the slave retains some rights to own property, marry freely, purchase his freedom, not be killed without cause, etc. However, within these limits, it remains the case that the master may do with the slave what he wills. Enslavement thus comes in degrees, but the essence of the condition remains that the slave is under an arbitrary power and authority, at least with respect to a broad range of their potential actions.   Though it hasn’t always been seen in this light, in our time it is difficult to see slavery as anything other than a degrading and morally abhorrent social institution. This assessment is surely based in large part on the fact that we value freedom so highly and see in slavery the very negation of the idea of ourselves as naturally free and equal.  Advocates of the non-domination and non-interference conceptions of freedom, however, offer alternative accounts of what it is about the slave’s condition that renders them unfree.   According to the non-domination account, a slave is unfree by definition -- since slavery is a condition in which one person is subject to the arbitrary power of another. It is true that we can talk about degrees of freedom in various slaves’ lives, but that degree of freedom depends on the scope of the master’s power and authority over him not in the manner in which that power is exercised.  According to the non-interference account, there is nothing in the definition of slavery that entails that slaves are unfree. To be unfree is to be actually interfered with, 128not merely to be under someone’s power in virtue of which one is vulnerable to interference. Nonetheless, the whole point of the institution of slavery is to get people to do things that they wouldn’t be inclined to do voluntarily. Masters generally want slaves so that they can get them to do the work! Slaves thus tend to be highly coerced individuals in consequence of their subordinate position, and so endure high degrees of interference-unfreedom. As with the non-domination account, the unfreedom of slaves comes in degrees; however, in this case their degree of unfreedom depends on the extent to which the master actually interferes with her ability to do the things she might otherwise have wanted to do, rather than in the degree to which the master has the power and authority to do so.   Both conceptions of freedom, then, can make sense of the fact that we consider slavery objectionable from the point of view of freedom. However, it seems to me the non-domination account captures something important that the non-interference account misses. The mere fact that it is up to the master’s discretion whether his slave may do this or that thing, rather than up to the slave herself, constitutes an offense to the slave’s freedom. Even if the master is so gracious or permissive as to allow the slave a measure of leisure and personal pursuit, it remains the case that the slave enjoys such non-interference only by the master’s leave, a grace that is liable to be withdrawn at the master’s whim. There is something about the master’s position of control in itself that seems at odds with freedom.    This point can perhaps be brought out more clearly by considering the positions of three hypothetical individuals, Georgia, Marcus, and Dave:  ?129Georgia is the slave of a particularly exacting master who exercises control over nearly every aspect of her life. She spends long days working in the field under the whip; what she eats, where she may go, where she sleeps and for how long, who she may associate with, everything is closely prescribed, monitored and enforced.  ?Marcus is also a slave, but of a less heavy-handed and controlling master. His activities are mostly controlled during the ten working hours of the day, but he is allowed to spend his non-work hours in broad range of leisure activities, he is allowed to marry and form relationships as he chooses, and is generally afforded a wide degree of non-interference outside work.  ?Dave is an employee rather than a slave, one who is under contract to work ten hours a day, during which time his activities are controlled by the employer to the same extent that Marcus’s are by his master. But outside of work his time is his own. He is able to do all the things Marcus is able to do in his off time, only, in his case, his ability to do these things is not dependent on his employer’s permission, but rather is simply his legal right, one which is protected by the authorities.  ?By hypothesis then, Georgia and Marcus are equivalent in the extent to which they are under the power of their respective masters, and thus in their degree of domination; and Marcus and Dave enjoy an equivalent degree of non-interference in their lives. Thus, according to the non-domination conception of freedom, Georgia and Marcus are 130equally unfree, and Dave is more free than both them. According to the non-interference account, Marcus and Dave are equally free, and both are freer than Georgia.   I am happy to acknowledge that our initial intuitions about these cases point in opposite directions. It seems right to say that, in a certain sense, Marcus has more freedom than Georgia. He enjoys more liberty to do the things he wants to do. (I will return to this point indue course). The contrast I want to focus on, however, is the one between Marcus and Dave. On that score, it seems right to say that Dave is freer than Marcus. After all, even in his leisure time Marcus is subject to the control and authority of his master -- since, unlike Dave, he enjoys the non-interference he does only by the leave of his master.   How should we make sense of this intuition about the comparative freedom of Marcus and Dave? It seems to me that we can articulate the intuition in two different “tones.” The first expresses itself by reference to what Pettit has called alien control.  174Though Dave and Marcus enjoy an equivalent degree of non-interference, Marcus’s “liberty” is entirely under the control of his master. Moreover, this control is not accountable or responsive to Marcus’s desires or aims; it is simply at the discretion of his master, and so is an entirely alien control. On the other hand, Dave enjoys at least part of his liberty independent of his employer’s control. And for this reason Dave seems free in a way Marcus is not.  We can also give expression to this intuition in a more morally-charged voice. Suppose Marcus reaches a point where he is ready to demand his emancipation. We might imagine the following exchange between him and his master: ?131 “Republican Freedom: Three Axioms, Four Theorems,” 106174Marcus: I no longer wish to be your slave, I want to be a free man. I want to be able to determine for myself what I will do with my life, and not be dependent on you to tell me what I may do, with whom I may do it, and when I may come and go.  ?Master: (Somewhat offended) Have I not been a gracious and accommodating master? Have I not granted you a generous amount of leisure, and have I not allowed you a great degree of freedom in what you can do with your self outside your duties? Isn’t your condition so much better and freer than the slaves of other masters? ?Marcus: It’s true, you have treated me better than Georgia’s master treats her. But even so I remain ever at your mercy, dependent on your good will. What liberty I have is not really my own, since even as I enjoy it, it remains yours to give or to withhold.    Master: (Indignant) What impudence! You repay my generosity with complaining! Other slaves would gladly trade places you. They would rejoice at the good fortune of serving under such a light-handed and indulgent master! ?Marcus: (Defiant) By what right do you claim the authority to rule over me? I am not like one of the horses in your stables. I am a human being every bit as much as you! Who are you to arrogate to yourself the authority to lord it over me? Did I appoint you ruler over me?  ?132It seems to me that the morally-charged language of “arrogated authority” and “right” is more fitting than the merely empirical language of “alien control” to express the impulse behind our intuition that Marcus lacks freedom in an important sense that Dave does not. However, I won’t press this point here. For the purpose of this chapter, we can suppose that this difference amounts to no more than a difference in expressive tone. The crucial point is that the non-interference conception of freedom is blind to the way in which the mere fact of control or authority is central to why we think of slavery as fundamentally inimical to freedom. (ii) Despotism  It seems to me that the same point holds for the case of despotism. Despotism is a form of government in which a single agent rules with absolute power. The despot -- whether a king, an emperor, a general, or a group of party elites -- stands to his subjects in the same relation as the master does to his slaves. He possesses a power in virtue of which he controls the lives of his subjects, whether or not he always exercises that control by interfering in their activities. Moreover, that power is not constrained by law nor accountable or responsive to those subject to it.   Now it is unlikely that any actual despot will be able to achieve the degree of control over his subjects that a master might over his slaves. For one thing, exercising control requires that one be able to monitor compliance. A master could conceivably achieve something like complete monitoring of his slaves’ activities, but a despotic ruler is unlikely to be able to do so. However, this difference is a matter of degree rather than kind. 133 Following our approach to the case of slavery, let’s consider the positions of the subjects in three hypothetical countries: ?Country 1 is ruled by an absolute tyrant. The tyrant has consolidated to himself a near absolute power over his subjects. He closely monitors the activities of his subjects through a network of informants, narrowly prescribes what they can and cannot do in nearly every aspect of their lives, and applies overwhelming force to punish infractions.  ?Country 2 is ruled by a liberal despot. Like the absolute tyrant, he also possesses absolute power. However, he rules with a lighter hand, and indeed fancies himself a benevolent dictator -- or as he prefers, “father of the nation.” He does require of his subjects that they obey certain commands that he deems necessary for good order and prosperity, but beyond this he is largely content not to interfere with economic, religious, or social activities of individual subjects.  ?Country 3 is a liberal democracy. The citizens of country 3 are subject to a set of laws that impose more or less the same constraints on their activities as those faced by subjects in country 2, and guarantee them more or less the same liberties. However, in their case the government is constrained to operate within the laws by a constitution, a complex set of checks and balances, an independent judiciary, and regular elections. ?By hypothesis, the subjects of countries 1 and 2 are equivalent in the extent to which they are under the power of their respective rulers, and thus in their degree of 134domination; and the subjects of country 2 and the citizens of country 3 enjoy an equivalent degree of non-interference in their lives. Thus, according to the non-domination conception of freedom, the subjects of 1 and 2 are equally unfree, and the citizens of 3 are more free than both them. According to the non-interference account, the subjects and citizens of countries 2 and 3 are equally free, and both groups are freer than those in 1.  As in the case of slavery, it seems to me that our initial intuitions about these scenarios point in opposite directions. It seems right to say that, in a certain sense, the subjects of the liberal despot have more freedom than the subjects of the absolute tyrant. They enjoy more liberty to do the things they might want to do. However, again, the contrast I want to focus on here is between scenarios 2 and 3.   And once again, it seems right to say that the citizens of the liberal democracy are free in a sense that the subjects of the liberal despot are not. The subjects of the liberal despot remain entirely under the power of the despot. Though they enjoy a wide degree of non-interference, their enjoyment of this “liberty” is under the control of the despot. Moreover, this control is not accountable or responsive to them; it is wielded entirely at the discretion of the despot, and so is an entirely alien control. By contrast, the power of government in the liberal democracy is itself constrained to operate within the bounds of the law, and is designed to be accountable and responsive to the claims of its citizens. It is thus the citizens themselves that ultimately possess the power of control over their enjoyment of their “liberties.”    We can also imagine the subjects of the liberal despot giving voice to this claim in the morally-charged language of the slave: “By what right do you claim the authority 135to rule over us? We are not like the horses in your stables. We are human beings every bit as much as you! Are all people not born free and equal? Who are you to arrogate to yourself the authority to lord it over us? Did we appoint you ruler over us?”   The conception of freedom as non-interference, however, is insensitive to the sense of unfreedom involved in both the slave’s dependence on the authority of his master and the subject’s dependence on the power of the despot. Berlin himself appears content to accept this result, at least in the latter instance. He writes: “[I]t is perfectly conceivable that a liberal-minded despot would allow his subjects a large measure of personal freedom. The despot who leaves his subjects a wide area of liberty may be unjust... but provided he does not curb their liberty, or at least curbs it less than many other regimes, he meets with Mill’s specification [of negative liberty].”  However, 175in my view, this seems at best a kind of biting of the bullet. A conception of freedom that is able to mark the difference between the conditions of Marcus and Dave and between the subjects of a liberal despot and the citizens of a liberal democracy is at least superior in that respect. If that is right, then we have a second reason to favour the non-domination account of freedom over the non-interference account. 3.4. The liberal rejoinder  The effort to supplant the non-interference account with the non-domination account has not, however, gone unchallenged. Ian Carter and Matthew Kramer in particular have sought to undermine the case for the neo-republican conception of freedom. Their overall strategy has been to argue, first, that the non-interference conception can be given a more sophisticated formulation than is generally supposed 136 “Two Concepts,” 129175by its republican critics. This more sophisticated version of the non-interference account, they insist, can capture the most compelling insights of the non-domination account. Moreover, they have argued, the remaining differences between the two accounts tell in favour on the non-interference conception. 3.4.1. Actual and subjunctive interferences  According to neo-republicans, the non-interference ideal is inadequate in part because freedom can be undermined even in the absence of actual interference -- i.e. when another agent holds the power to interfere on an arbitrary basis, but happens not to do so. This claim is supported by the reality of self-editing effects but also by our intuitions about cases involving alien control in itself. However, according to Ian Carter, these criticisms of the non-interference view depend on an unnecessarily simple account of what constitutes an “interference.” In particular, he claims that “interferences” ought to be understood to include both “actual” and “subjunctive” interferences.  176 Carter distinguishes between actual and subjunctive interference as follows: An agent ‘X’ actually interferes with the action possibilities of an individual ‘Y’ just in case X actually does something that prevents Y from doing something she otherwise could do. Actual interference includes, paradigmatically, X manhandling Y into doing or not doing something, X constructing a wall or a fence that prevents Y from doing something, or X incapacitating Y by locking her in a cell, handcuffing her, drugging her, or appropriating her property . An agent X subjunctively interferes with the action possibilities of an 177individual Y just in case, were Y to attempt to do some action Z, X would actually 137 “How are Power and unfreedom Related?”, 67176 One need not hold that just any taking of a person’s property diminishes their non-interference 177freedom (though Carter’s view commits him to this view). However, certain instances obviously do. If I steal your car, you are now unable to travel as far or as easily as you previously could; If I take your computer, you may now no longer be able to access information you previously could. interfere with her doing so. For instance, if it is the case that the guards standing outside the Parliament buildings would arrest and detain a protestor were he to attempt to enter the restricted area, then those guards are subjunctively interfering with that protestor’s action possibilities.  Carter maintains that the non-interference account of freedom ought to be understood as stipulating that both actual and subjunctive interferences diminish an individual’s freedom. Thus the mere fact that the Parliamentary guards are standing ready to prevent the protestor from entering the restricted zone in itself diminishes the protestor’s freedom. Not only is this a plausible account of the way people often speak about interferences in this context, Carter insists that the thought that “interference” ought to include subjunctive interference is already implicit in traditional formulations of the non-interference conception of freedom. As I’ve noted, on Berlin’s account of the negative concept of liberty, “interferences” are understood to include both the use and threat of force to secure complying action. Threats are often particular kinds of subjunctive interferences since they often work not by actually interfering with a particular action, but by promising to do so should the threatened person not comply. Our example of the guards and the protestor clearly fits this pattern.   However, other cases of threats don’t, or at least not so obviously. Some threats work not by promising to prevent an action should it be attempted, but instead by promising to inflict other kinds of harms on an individual should he fail to comply with the threateners demand. For instance, a mob boss may threaten to break the kneecaps of anyone who turns informant. Alternatively, a threat may work by promising to harm some third party the threatened individual cares about -- e.g. the informant’s family. 138Carter accommodates these sort of threats by reference to what he calls “compossible actions.”  Two actions, J and K, are “compossible” for a particular agent X, just in case 178X can perform both J and K and further, if X were to perform either one of them, then he would still be able to perform the other. When the mobster issues the threat, “If you dare talk to the police, I’ll break your kneecaps,” he is not thereby interfering with your freedom to talk to the police, but he is interfering with your freedom to perform the previously compossible actions of talking to the police and walking about.    1793.4.2. Subjunctive interferences and probabilities   Carter further nuances the non-interference account by introducing a probabilistic element to its definition of freedom.  The motivation for this is clear once we have 180included subjunctive interferences in the category of freedom-diminishing interference. To say that an agent X has subjunctively interfered with a particular action of someone else Y involves making a claim about what X would do were Y to attempt that action. However, such claims are always probabilistic in character: we are really making a claim about what X is likely to do. Even if the mob boss is very powerful and is genuine in his intention to carry out his threatened action, it is still possible that you might ignore that threat and yet not have your knees broken. The mob boss could suffer a sudden heart attack, you might successfully flee the country, a rival gang might offer you protection to spite their competitor, etc. 139 Carter, A Measure of Freedom, 80-1178 As an analysis of what we ordinarily take ourselves to mean, this is perhaps a bit odd. It is even more 179of a stretch if we try to extend it to the case where the threatener promises harm to a third party: “If you dare talk to the police, you’re grammy might take an unfortunate fall down the nursing home steps.” Here we would have to say something like, while the threatener hasn’t interfered with your freedom to talk to the police, he has interfered with your freedom to perform the previously compossible actions of talking to the police and enjoying your grammy’s company. However, for our purposes here, there is no need to worry about the somewhat awkwardly sounding character of these analyses. “How are Power and unfreedom Related?”, 70; A Measure of Freedom, 190180 In light of this, Carter refines the definition of “subjunctive interference” as follows:  X subjunctively interferes with Y’s possible action Z just in case, were Y to attempt Z, X would with a suitably high degree of probability actually interfere with Y’s doing Z. Putting these various points together, on Carter’s sophisticated formulation of the ideal of freedom as non-interference, a person is free with respect to a particular action (or compossible set of actions) just in case no other agent actually interferes with her performing that action (or compossible set of actions) or likely would interfere with her doing so if she were to attempt it. 3.4.3. Republican objections answered  Armed with this sophisticated formulation of the ideal of non-interference, Carter insists that the reasons offered in favour of the superiority of the non-domination account fall short, and in fact point in the opposite direction. Those reasons were (a) that the non-interference account is implausibly blind to certain self-editing effects in dominated individuals, and (b) that the non-interference account fails to recognize a central feature of what we consider freedom-compromising about both slavery and despotism -- namely the element of “alien control.”  To start, it is important to note that, even on the non-domination account, the fact that an individual self-edits her actions in anticipation of possible responses by those with power does not in itself constitute her unfreedom. Rather self-editing is a symptom of her unfreedom. Her unfreedom consists in being subject to arbitrary power, whether or not she chooses to self-edit her actions in response to her awareness of her dominated condition. However, the prevalence of self-editing effects among dominated individuals is supposed to give us one reason to think that freedom must consist in non-140domination. Self-editing effects highlight the poverty of the non-interference account of freedom because they involve individuals not doing what they otherwise might do under the pressure of the superior force of others, and yet apparently without involving any diminishment of their freedom. Still, the second, more basic reason to favour the non-domination account, is that the non-interference account fails to recognize the sense in which freedom is undermined by the mere fact of “alien control,” as illustrated by the cases of slavery and despotism.   However, Carter thinks the force of both these points depends on assuming the unsophisticated version of the conception of non-interference. Suppose one agent X has the power to interfere (arbitrarily) in a particular action possibility of another agent Y. Now consider two possible scenarios, the first in which X is likely to interfere in the action possibility of Y should she attempt it, and the second in which X is very unlikely to do so. According to the sophisticated non-interference account, in the first scenario X subjunctively interferes with Y’s performing the action, and so diminishes her freedom even if he doesn’t actually interfere. Thus, the sophisticated non-interference account renders an “equivalent judgement” to that of the non-domination account concerning the state of Y’s freedom with respect to this action.  If Y, aware of her dominated 181condition, were to self-edit her actions, the proponent of the non-interference conception of freedom would be every bit as able to view this as a symptom of her lack of freedom as the proponent of the non-domination account.   Where the two views differ is with respect to the second scenario, the one in which powerful X is very unlikely to actually interfere with the action of Y were she to attempt it. In that scenario, Y is judged unfree on the non-domination account, but free 141 Carter, “How are Power and unfreedom Related?”, 66-7181on the non-interference account. Moreover, were Y to self-edit her actions in light of her awareness of X’s power to interfere, even though X is very unlikely to actually interfere, this would count as a symptom of her unfreedom on the non-domination account, but not on the non-interference account.   However, in Carter’s view, these remaining differences concerning the second scenario tell in favour of the non-interference account and against the non-domination account. To see why, consider the following illustration offered by Matthew Kramer:  ? Suppose that, in a community not far from the hills, a gigantic person G is born.   From adolescence onward, G is far larger and stronger and swifter and more   intelligent than any of his compatriots. If he wished, he could arrogate to himself   an autocratic sway over his community by threatening to engage in rampages   and by coercing some of the residents into serving as his henchmen. Were G so   inclined, no one would dare to resist his bidding... In fact, however, he loathes the  idea of becoming a tyrant; his principle desire is to seclude himself altogether   from his community. He does indeed depart therefrom, in order to reside in a   cave in the nearby hills where he contentedly feeds off natural fruits and wildlife   and where he spends his time in solitary reflection and reading and exercise. In   these circumstances, G is a dominator (according to Pettit’s criteria for that   status), but he is not significantly reducing the overall liberty of anyone else.   182?Though admittedly far-fetched, the gentle giant story nicely illustrates Kramer and Carter’s basic point -- namely, the mere power to interfere (i.e. the mere possibility of 142 Kramer, “Liberty and Domination,” 47182arbitrary interference) doesn’t seem enough to undermine the freedom of others. It is rather the probability of interference that seems crucial, which is precisely what the sophisticated non-interference account claims.  Moreover, a similar point holds with respect to self-editing effects. Consider the following scenario: Suppose Robinson is shipwrecked on a deserted Island, and has sustained serious injuries in the process, including a massive wound to his leg. By shear luck, a medical doctor, out alone on a sailing adventure, arrives at the Island and sees the injured Robinson. Dr. Friday grabs his first aid kit and approaches him with an offer to help. However, Robinson, in addition to being physically injured, is delirious from lack of hydration. In his altered mental state, he believes Dr. Friday is a cannibal aiming to torture and eat him. Now given the compromised physical condition of Robinson and the doctor’s possession of a scalpel, it happens to be the case that Friday has the power to physically constrain, torture and eat Robinson. If he were to exercise this power, there would be no one else around to witness it, never mind interfere with his doing so. However, Friday only wants to help. Unfortunately, Robinson, in his state of delirium, limps away as fast as he can to hide. Once hidden, he remains as quiet as he can in an effort to avoid detection by his would-be torturer.      As with the case of the gentle giant, Dr. Friday in this story meets the criteria of a dominator with respect to Robinson. Further, Robinson clearly self-edits his actions in response to the doctor’s very real power to interfere with Robinson’s future action possibilities -- he runs off on what is a very painful leg and he hides quietly to avoid detection. Nonetheless, it seems quite implausible to say that Dr. Friday has diminished Robinson’s freedom in anyway, or to say that Robinson’s self-editing reactions are 143symptoms of a condition of domination. But that, again, appears to tell in favour of the sophisticated non-interference account of freedom, since Dr. Friday possesses the power to interfere arbitrarily but is only very unlikely to do so.   3.5. A hollow triumph?  Should we view these counter-arguments by Carter and Kramer as decisive? In my view, while they are right as far as they go, they involve a much bigger departure from the traditional non-interference conception than their authors acknowledge; in fact, as I will argue below, the resulting view is more plausibly understood as a variant of the non-domination account than a variant of the non-interference account. To see why, we need to examine more closely the notion of “subjunctive interference,” and consider whether all that is included within the scope of its definition is plausibly termed an “interference.” 3.5.1. Threats versus subjunctive interferences  According to Berlin’s original non-interference conception, an individual’s freedom is diminished either by direct applications of force or by threats of force. Threats most certainly need not be explicit to count as threats. When a guard raises his gun as you approach and says “Keep back,” there is a clear sense in which he is issuing a threat, though not a fully articulated one. If the mob boss makes “an offer you can’t refuse” we all know we are supposed to read an implicit threat into his superficially non-threatening “offer.” When a police officer informs us that it is illegal to be in the park after dark, he is probably doing more than striking up conversation. Both explicit and implicit threats are reasonably viewed as interferences in the sense intended by the traditional non-interference conception of freedom. 144 “Subjunctive interference,” however, names a much broader category of phenomena. Recall: X subjunctively interferes with Y’s possible action Z just in case, were Y to attempt Z, X would with a suitably high degree of probability actually interfere with Y’s doing Z. Credible threats, both implicit and explicit, are clearly instances of subjunctive interference. But they are very particular kinds of subjunctive interferences -- namely ones in which the interferer communicates an intention to actually interfere to another person with the aim of thereby causing that other person to do/not do something they might otherwise have done/not done. However, subjunctive interferences need not be of this sort.   They can also take one the following two forms: (a) The interferer intends to actually interfere in the event another individual attempts to do something they don’t like, but doesn’t communicate this intention. Suppose David and Susan are married with children, and suppose further that David intends to take the children away should Susan decide to leave him, but keeps this intention secret from her. That in itself constitutes an interference with Susan’s freedom to parent her children. (b) The interferer is likely to actually interfere in the event that another individual attempts to do something they could otherwise do, but doesn’t have any present intention to do so. Suppose David has the power to take the children away (he has the financial resources to relocate, and there are no effective legal protections against his doing so). Given David’s superior social power and given certain facts about his beliefs and motivations, there is a strong probability that he would take the children if Susan were to leave him. But at present he hasn’t formed any intention to do so, and perhaps hasn’t even entertained the thought. That in itself still constitutes an interference with Susan’s freedom to parent her children.   145 Now, it does seem plausible to view both of these types of situations as diminishing the freedom of the vulnerable parties involved. However, it does not seem plausible to view them as cases of “interference.” Instead, they look much more like cases of domination without interference. In other words, to invoke the broad notion of “subjunctive interference” is really to concede to the fundamental insight of the non-domination conception of freedom in everything but name. It’s true, there remains a difference concerning whether freedom ought to be understood as the absence of vulnerability to the mere power to interfere (i.e. the mere possibility of interference) or whether it ought to be understood as the absence of vulnerability to another agent’s power to interfere that is likely to result in actual interference. However, this strikes me as a small difference -- much smaller than the difference between Carter’s sophisticated version of the non-interference conception and the traditional non-interference conception.  For this reason, I think advocates of the non-domination account ought simply to amend their view to acknowledge that domination consists in the subjection to the arbitrary power of another agent, where that power is likely to manifest in actual interference with a suitably high degree of probability. Such an amendment is in keeping with the basic republican conviction that it is vulnerability to a certain kind of power that is the essence of unfreedom. Moreover, it is consistent with continuing to hold that only those interferences that stem from arbitrary exercises of power diminish a person’s freedom. 3.5.2. Two further reasons to endorse the amended formulation  There are two further reasons republicans have to accept this amended version 146of the ideal of non-domination. The first is that the amended formulation confirms the traditional republican stance towards slavery and despotism. It might seem to do the opposite, since it no longer implies that slavery and despotism diminish freedom by definition. It is logically possible on the amended formulation to have a slave holder that is very unlikely to actually interfere with the slave’s activities, and so would not be compromising the slave’s freedom in virtue of his position of power. Likewise, it is logically possible to have a despot who possesses an absolute power over his subjects, and yet is very unlikely to actually interfere (at least with respect to a suitably wide range of activities). However, while no longer a simple analytic truth, the republican claim that slavery and despotism are necessarily opposed to freedom remains secure. Let’s focus on the despotism case to see why (though the same reasoning applies to the slavery case).   What would be required for a despot’s subject to be accounted free with respect to a particular action or set of actions, all the while remaining a subject? Two conditions would have to be met: First, it would have to be the case that, while the despot possesses the power to interfere arbitrarily in the subject’s activities, he doesn’t actually do so. Second, it would have to be the case that, not only does the despot not actually interfere arbitrarily in the subject’s activities, there is a very low probability that he would do so.   The first condition is easy enough to meet, at least with respect to a restricted range of actions or action types, since no despot interferes with every possible action of his subjects. Moreover, some despots could plausibly be viewed as interfering relatively little overall, approaching the status of Berlin’s “liberal despot.” It might seem that the 147second condition could also be met, even if less frequently, for two sorts of reasons: First, it might be that it is in the interest of the ruler herself not to interfere in certain domains. A particular ruler may have no interest in interfering in the religious practices of some minority because doing so would be very costly and disruptive and would serve none of her purposes. Or she might have an interest in leaving individuals alone to engage in economic activities of production and exchange with the thought that doing so will increase her tax revenues over the long run.  A second reason a despot may 183be very unlikely to interfere in certain domains, is out of moral principle. Perhaps a particular ruler is firmly committed to some liberal version of the ideal of noblesse oblige and so refuses to unduly lord it over her subjects. If, for any of these reasons, a particular despot meets both the first and second condition, then on the amended non-domination conception her subjects would count as free (at least with respect to the relevant domain of actions).  However, in fact, the second condition is much more difficult to meet than it initially appears. As a first step to seeing why, we need to attend more closely to the nature of its probability requirement. When we say that a particular despot is unlikely to interfere do we mean that (a) as a matter of objective fact, he is unlikely to interfere or (b) on the basis of the available evidence, his subjects have good reason to believe that is unlikely to interfere? In other words, should the probability requirement be understood along the lines of an objective or an epistemic probability?   It turns out that neither view in itself is entirely satisfactory. The simple objective probability interpretation is unsatisfying since, even if it happens to be true that a given 148 The dictatorship of Augusto Pinochet bears some marks of this sort of strategic economic non-183interference. This has apparently led some to view his regime as a real life “liberal despotism,” but that conclusion is surely mistaken in light of its very illiberal policies towards political expression and dissent.  despot is unlikely to interfere, if the evidence available to his subjects suggests otherwise, they will tend to exhibit the previously discussed self-editing effects. Suppose, for instance, the despot has up to this point displayed a consistent propensity to interfere with his subjects lives, but has recently read Mill’s On Liberty and decided from here on to rule with a much lighter hand. Suppose further that, as a matter of fact, this change of heart will prove complete and permanent. His subjects will nonetheless have little reason to place any trust in this conversion; for all they know this may be a clever ruse or a passing fancy. Their best policy will remain one of caution and self-restraint.     184 On the other hand, the epistemic interpretation by itself is also unsatisfactory. Suppose that, as a matter of fact, the despot is very likely to interfere with his subjects’ actions, but the evidence available to his subjects suggests otherwise. Suppose, for instance, that the despot’s long-time advisor has recently passed away and been replaced with a counsellor who, while brilliant and persuasive, has a tyrannical frame of mind. Unbeknownst to them, the subjects are now very likely to be interfered with should they in anyway cross the despot’s purposes or cause him offence. In such a case, it seems right to say that the subjects are dominated, even though not subject to a power that, according to the evidence available to them, they have reason to believe is likely to interfere.    In light of these points, the best interpretation of the second condition is a hybrid one: That is, for a despot’s power to count as non-dominating (on the amended 149 The problem here is closely related to the problem of “bluffed” threats. If one person threatens another 184with harm should she refuse to comply with his demands, but in fact either doesn’t actually intend to interfere or lacks the power to interfere, so long as the threatened person is unaware of these facts, it remains the case that the threatener has diminished her freedom.   formulation) it can’t be the case either (a) that, as a matter of objective fact, he is likely to interfere or (b) that, based on the available evidence, those subject to his power have good reason to believe he is likely to interfere.     185 Given this understanding of what is required for a despot to count as non-dominating (and thus for his possession of absolute power to be consistent with the freedom of his subjects), there are a couple of reasons to think that no actual despots could meet these requirements.   (i) The corruption problem  The first depends on a straightforwardly empirical observation, namely that, as the time-worn adage attests, power corrupts, and absolute power corrupts absolutely. This conviction has long been a staple of republican polemics, though it by no means unique to them. The prophet Samuel is recorded as warning the ancient Israelites against the temptation to install a king over themselves on these very grounds:  ? These will be the ways of the king who will reign over you: he will take your sons   and appoint them to his chariots and to be his horsemen and to run before his   chariots. And he will appoint for himself commanders of thousands and    commanders of fifties, and some to plough his ground and to reap his harvest,  150 According to this hybrid account, an individual’s freedom is diminished if another person has the power 185to interfere in her actions, is very likely do so, and yet never does. That might seem like an odd result, particularly when the subjected party is totally unaware of their vulnerable position. How can someone’s freedom be diminished when none of their actions are blocked and, moreover, they are entirely unaware and thus unaffected by the threat of them being blocked? I’m not so sure this is an odd result, but in any case it is not a problem specific to this view. The same result holds for the non-interference conception of freedom, in cases where an individual is unaware that certain actions they could otherwise perform have been blocked (e.g. government officials have unbeknownst to me destroyed public documents I could otherwise have accessed, but had no intention of accessing). On either view, it is possible for a person’s freedom to be diminished without her being aware of it, psychologically affected by it, and without any of her actions being thwarted.  and to make his implements of war and the equipment of his chariots. He will   take your daughters to be perfumers and cooks and bakers. He will take the best   of your fields and vineyards and olive orchards and give them to his servants. He  will take the tenth of your grain and of your vineyards and give it to his officers   and to his servants. He will take your male servants and female servants and the   best of your young men and your donkeys, and put them to his work. He will take  the tenth of your flocks, and you shall be his slaves.   186?The validity of these observations does not depend on the belief that human nature is inherently corrupt, but only that it is prone to corruption when in positions of unconstrained power. When a person possess such unconstrained power, his natural desires for wealth, for honour, for pleasure, for power itself lack any of the ordinary checks -- including both formal and, to a large extent, informal checks (since people are more likely to flatter and defer to a prince rather than express disapprobation). Lacking these checks, a person with unconstrained power will tend to be overtaken by these desires.   But the problem of the corrupting influence of power can result not just from the “ignoble” but also from the “noble” passions. After all, as Hume rightly observes, human nature has also bequeathed to us a natural sympathy, and so it is not only our self-regarding desires that are left unchecked when in positions of unconstrained power. However, this fact also carries corrupting risks. As Mill argues in On Liberty, even the desire to interfere for the sake of other people’s own (perhaps unrecognized) good tends to result in unacceptable levels of coercion, based on invariably mistaken ideas 151 I Samuel 8: 11-17.186about what is in the best interest of those subjected to it. This problem is particularly acute in the case of despotism, since even when acting in good faith, the despot is unlikely to possess the information required to discern the interests of those subjects most remote from him.  The problem of corruption is amplified when we attend to the epistemic position from which subjects must assess the likelihood that the despot will interfere. As we’ve just seen, in order for a despot’s power to count as non-dominating with respect to a particular action or set of actions, it must not only be the case that the despot is unlikely to interfere in these activities. It must also be the case that, based on the available evidence, the subjects have reason to be confident that he is unlikely to interfere. But such subjects will usually have to make these judgements based mostly on general knowledge about the tendencies of people with power of this sort, since they are unlikely to have intimate access to the motivations and circumstances under which the despot is operating. Moreover, even if they have some such information, they must counterbalance this with the knowledge that human beings are often unstable in their character, particularly under changing circumstances. Thus, even if a particular despot is in fact unlikely to interfere, his subjects are unlikely to be have trustworthy epistemic access to this fact, and so will have to fall back on their knowledge of the general facts of absolute rulers, facts which give one little reason for optimism.  (ii)The infection problem  The second reason to doubt that a despot’s power could ever be non-dominating is the “infection problem.” No despot can avoid interfering with the actions of his subjects altogether if he is going to hold onto his power. Actually existing despots, then, 152will invariably interfere or be likely to interfere in at least some of their subjects’ activities. There is nothing particularly problematic about that, since even non-arbitrary political authorities will need to do so. However what’s different about the despot is that he is not constrained in any meaningful way by law or external checks. Thus, when and how he interferes need not conform to stable principle or at least any publicly available standard, and further there is nothing constraining him from deviating from principle in particular cases should it prove especially beneficial to him to do so.   Suppose then there is a particular despot that exhibits a relatively stable pattern of interfering quite frequently in certain types of actions and interfering very infrequently in others. It might seem that with respect to the latter category of actions the subjects ought to be accounted free on the amended non-domination view, since there is a low probability of the despot interfering and, moreover, this low probability is publicly known.  The difficulty is that, so long as there are even occasional deviations from the pattern where the deviations aren’t governed by any standard that can be known and confidently counted on, subjects will likely adopt a generalized approach of cautious self-editing even in the low-probability-interference-domains. A few exceptions can infect a whole domain of activity.  The infection problem is especially acute where the cost of attracting a response from the despot is high and the benefit of provocative action is low. In such cases, subjects will be inclined to self-edit their actions even though the likelihood of interference is small. Suppose, e.g., a particular despot has proven very unlikely to punish those that express public disagreement with his policies, but with a few 153prominent exceptions (perhaps only 1 in a 1000 such acts elicit a response).  But 187suppose that in those few cases the despot has imprisoned the perpetrators for extended lengths of time. For many people, even though the probability of interference is low, the high cost of such interference, should it occur, will be enough of a deterrent.  3.5.3. Summary  Upon closer examination, then, the traditional republican view that slavery and despotism are necessarily opposed to freedom holds up well on the amended (probabilistic) definition of non-domination. While the idea of a “liberal despotism,” or at least a quasi-liberal one, does not seem outside the range of possibility, the idea of a non-dominating despot does. For a despot not to dominate his subjects with respect to a particular range of actions it would have to be the case not only that he does not interfere, but that he is unlikely to interfere -- both in the sense there is actually a low probability that he will interfere and in the sense that his subjects, given their epistemic positions, have good reason to be confident he is unlikely to interfere. However there is good reason to think that such subjects could never really be sufficiently confident that the despot’s liberality will hold up. In the absence of the constraint of law, those wielding absolute power will tend towards both “ignoble” and “noble” corruption; moreover, even occasional arbitrary interferences that deviate from an otherwise liberal pattern can infect those areas of infrequent interference with enough uncertainty to generate self-editing effects -- particularly when the stakes are high.  154 A real-world illustration of this point is the case of Salmon Rushdie. While many people have written 187books that are at least as critical of Islam as Rushdie’s The Satanic Verses, only a tiny fraction of them have had their lives threatened as a result. Nonetheless, after the Rushdie affair, at least according to some accounts, writers have been much more cautious in expressing critical or unorthodox opinions about Islam. Rushdie himself claimed in an 2012 interview that The Satanic Verses would not be published today because of a climate of “fear and nervousness.” "Salman Rushdie: Satanic Verses 'would not be published today'."3.6. Freedom in chains?  Earlier I noted that, according to the republican view, the absence of interference is neither necessary nor sufficient for freedom. In Section 3.5, I offered reasons to accept the second of these claims -- namely, that there can be a diminishment of freedom even in the absence of interference. However, justifying the first claim might appear more difficult. How could it be that an individual is interfered with in performing a particular action, and yet is not thereby made less free? According to republicans, this is the case whenever the interferences are the result of an exercise of non-arbitrary power. However three closely related objections have been raised against this claim. 3.6.1. Three objections  First, it appears to lead to counter-intuitive implications in particular cases. It turns out, as a number of authors of noted,  that on the non-domination view a justly 188convicted criminal, though imprisoned in a cell, is not thereby made less free. Since the constraints he faces are not, by hypothesis, the result of an arbitrary exercise of power, they are not instances of domination. But surely such a person is unfree if anyone is.  Second, the republican position carries what appear to be implausible implications concerning the comparative extent of unfreedom of individuals in different situations. This problem arise in two sorts of cases. The first of which are cases in which two people are subject to equivalent interferences, but one of them is interfered with on an arbitrary basis and the other on a non-arbitrary basis. On the non-domination account, the latter suffers a diminishment of freedom while the former does not. For example, suppose that Eve is accustomed to walking home along a path that cuts 155 cf. Carter, “How are Power and unfreedom Related?”, 65188through a particular garden, but that one day a fence is erected blocking her from continuing to do so. Adam, her cousin in a neighbouring district, is confronted with an identical interference in walking along a path he normally takes on his way home. However, suppose that in Eve’s case, the fencing has been approved by the local (non-arbitrary) authorities on behalf of the farmer who owns the garden; whereas in Adam’s case the fencing contravenes a local bylaw protecting the right of way for pedestrians, and so constitutes an arbitrary interference. It would appear that both Eve and Adam have had their freedom diminished to an exactly equal extent. However, on the non-domination account, it turns out that while Adam’s freedom has been diminished, Eve’s has not been diminished at all.  A similar problem arises in cases where two people are equally subject to arbitrary power, but where one is interfered with more extensively than the other. On the non-domination account, both individuals have had their freedom diminished to an equal extent, even though one of them is prevented from doing fewer things she might want to do than the other. This problem appeared in our initial discussion of slavery and despotism (Section 3.3.2.), where we noted our intuitions in favour of the claim that Marcus (slave to a permissive master) is freer than Georgia (slave to a heavy-handed and exacting one) and that the subjects of a liberal despot are freer than the subjects of an intrusive tyrant.  According to a third objection, the claim that non-arbitrary interferences do not diminish freedom depends, ultimately, on a view about what interferences are morally justified.  However this implies that the concept of freedom is essentially moralized -- 189156 cf. Carter, “How are Power and unfreedom Related?”, 66; Christman, “Review of Philip Pettit’s A 189Theory of Freedom and Government,” 205that is, that view that whether or not a person is free with respect to a particular action depends, in part, on an underlying conception of justice. This result has seemed problematic to many philosophers who believe that our definition of freedom ought not to presuppose an answer to the question of how it should be distributed. 3.6.2. Replies  The validity of the third “moralization” objection has been denied by Pettit who, as we’ve seen, defines arbitrary power as power that is not forced to track the avowed interests of those subject to it.  Nothing in this definition, he insists, contains a moral 190judgement, never mind a claim about justice. Matters are not quite as simple as that, in my view; however I will defer further discussion of this issue to Chapters 4 and 6.   The first and second objections both suggest that, while an absence of interference may not be sufficient for freedom, it is at least necessary.  However, in my 191view, these objections are less compelling than they initially appear.  To begin, it must be acknowledged that the words “free” and “unfree” are often used in ways that have little or no connection to the ideal of political freedom. For instance, if a man’s leg has been pinned by a fallen tree, it would be perfectly intelligible to claim that, by lifting the tree, we have thereby “freed” him. For that matter, we might talk of allowing a stream to flow “freely” by breaking up an ice jam. These uses of the word “free,” while linguistically legitimate, obviously have little to do with the idea of political freedom. They involve something like a purely physicalist sense of the word, used to indicate the absence of physical obstruction to an object’s motion. The idea of 157 Pettit, “Reply to Carter, Christman, and Dagger,” 55190 This conclusion might incline us towards what I called in Chapter 1 the “conjunctive approach” -- the 191view that freedom requires the absence of both domination and interference. However, I’ve already given reasons against taking this approach. Moreover, as I will argue below, it is unnecessary.political freedom, however, ought to be understood as concerning the relations between human beings qua agents; it is an essentially social ideal. There is thus no contradiction in holding that, while the man pinned by the tree lacks the freedom to move, his political freedom has not thereby been diminished. It would be extremely odd if the man were to complain: “They call this a free country, but I’ve just spent the last six hours completely trapped under this tree!” We might wonder if shock or dehydration had affected his mental state; or, barring that, we might wonder whether he had somehow misunderstood what is meant by a “free” country. In any case, we would categorically reject any suggestion that being trapped by a fallen tree could affect one’s degree of political freedom.   Now, the case of the justly imprisoned criminal is admittedly quite different. After all, he has had his movements constrained by the intentional actions of human agents. Nonetheless, it seems to me that something analogous is going on here. It would of course be perfectly acceptable to say of the prisoner that he is not free to come and go as he might like, that he lacks most of the freedoms that others enjoy, and even that he is not a “free man.” However, suppose this prisoner were to make the following complaint: “They call this a free country, but I’ve spent the last seven years locked up in this prison!” It seems to me that we would reject this claim in much the same way as in the case of the man pinned by the tree. Perhaps not right off. We might initially suspend judgement and look more closely at the facts of the man’s case. Importantly though, the facts we would be looking into would not be those concerning whether he had indeed been in the prison for seven years or whether the gates had in fact been securely locked all that time. Instead, we would be interested in whether there was enough 158evidence to support the claim that the man had actually broken the law, whether the law was just, whether the man had received a fair trial, whether the punishment suited the crime, etc. If all these points checked out, we would confront the man with them and say something like: “Your liberty has not been violated in any way. Your imprisonment is totally justified.” In short, while it is true that he lacks “freedom” in a perfectly intelligible sense of the word, it is not true that his liberty has been violated or infringed.  Proponents of the non-interference account might object that our reaction to the prisoner’s complaint can be fully explained by reference to a distinction between justified versus unjustified diminishments of (one and the same) freedom: “We are not denying that he lacks freedom, just that he has any right to complain about it.” Yet, as a claim about how our ordinary discourse about political freedom actually works, I am not convinced. It seems to me that the notion of political freedom in “ordinary” use does contain a built-in discrimination between kinds of interferences, and that when someone claims that they lack political freedom we automatically take them to mean that they are subject to illegitimate constraints.  192 There is a second reason to insist on discriminating between kinds of interferences in our account of political freedom: Any view, if it is to fulfill the requirements of a “theory of freedom,” will have to do so. Even the proponents of a purely empirical non-interference conception of freedom, if they intend to offer a normative ideal for political life, will have to offer some principle by which to determine which interferences are consistent with a free society and which not. It’s true they will view this purely as a question of distributing, not defining freedom; however, 159 Bernard Williams makes a similar point in his “From Freedom to Liberty: The Construction of a Political 192Value.”discriminate they must. And, to bring the point home, on any plausible such theory, it will turn out that, while the imprisoned criminal lacks freedom, his liberty has not been violated; whereas, if he had been unjustly imprisoned, though his condition be the same, such a view will have to hold that he not only lacks freedom, but also that his liberty has been violated. The discrimination between interference-identical cases remains; only the language is different.    Let’s turn now to consider the second objection concerning comparative judgements about degrees of freedom. In the Adam and Eve example, the non-domination account implies that Adam has had his freedom diminished while Eve has not, even though they have had exactly the same action possibilities blocked by the very same means (i.e. a fence).  But the thought underlying this objection can’t be that 193Adam and Eve ought to come out as being equally unfree since, as the arguments of the previous sections show, there is a sense in which Adam’s freedom is diminished and Eve’s is not -- i.e. Adam is subject to “alien control” whereas Eve, by hypothesis, is not.  Instead, the thought must be that, while Adam and Eve’s unfreedom is not equal, 194it must be the case that Eve’s freedom is at least diminished to some extent. However, if that is indeed the claim, then this objection is really the same as the case of the justly 160 This objection, in its basic form, goes all the way back to Hobbes’ Leviathan, Chapter 21: “Of the 193Liberty of Subjects.” This point can be re-enforced by a consideration of the difference between the position of Charles and 194Alexandr in the following illustration: Charles is currently locked in a prison cell, serving time for murder. The homicide investigation unit of the local police uncovered evidence of his crime, arrested and detained him. He was subsequently found guilty in a court of law, and sentenced to a term of imprisonment. Alexandr is also currently locked in a prison cell. He had written a letter to a friend in which he made critical remarks about the president, which the authorities intercepted. The secret police picked him up one day and deposited him in the cell in which he now sits. Charles and Alexandr, we may presume, currently face roughly equivalent degrees of interference in their possibilities for action. According to the non-interference account of freedom, then, the imprisonment of Charles constitutes a diminishment of his freedom that is identical to that of Alexandr. However, this seems wrong. While Alexandr’s imprisonment is a paradigmatic case of the absence of political freedom, Charles’ case is not.   imprisoned criminal, and so subject to the same reply: There is a sense in which Eve is now unfree to walk through the garden, but it is not the case that her political freedom has thereby been violated.   The case of Marcus and Georgia, however, is different -- and more difficult. In this example, both individuals are subject to the same degree of domination, but one is interfered with much more extensively. On the non-domination account, Marcus and Georgia come out equally unfree, though Marcus’s position seems obviously less constrained than Georgia’s.  I am not confident that there is an entirely satisfactory way to accommodate this intuition within the ideal of non-domination. There are, however, a couple of points that can help blunt the force of the objection.  In the first place, we need not deny that, while Marcus and Georgia are equally dominated, Georgia’s domination is more intolerable. That is, we can distinguish between unfreedom and the “conditions of its disvalue.” Proponents of the non-interference account are themselves committed to a similar distinction: A common objection to the negative concept of liberty is that not being interfered with in doing X is of little use if one is not actually able to do X. In responding to this objection, Berlin distinguishes between liberty and the conditions of its value.  He argues that even 195though a lack of capacity may render one’s freedom less valuable, it doesn’t alter the extent of one’s freedom.   A similar point would presumably hold for the reverse position: Namely that two people experiencing the same extent of interference might nonetheless suffer different degrees of “disvalue” from this condition. Suppose Martin and James are both Ugandan 161 Berlin, “Two Concepts,” liii, 124; cf. Rawls, A Theory of Justice, 179195citizens and, consequently, are both subject to a law forbidding homosexual relationships under the threat of life imprisonment. On the non-interference account, both men are equally unfree to enter into such relationships. However, it turns out that Martin is gay and James is not. Thus Martin suffers a much more serious “disvalue” from his unfreedom than does David.     A similar distinction of behalf of the non-domination account may go some way to blunting the problem raised by the Marcus and Georgia case.  While both individuals 196are equally dominated and thus equally unfree, Georgia suffers more “disvalue” from her domination than does Marcus because she is subject to a more heavy-handed and exacting master. So there is at least nothing in the non-domination view that prevents us from viewing Georgia’s enslavement as more intolerable than Marcus’s.  A second point can also help blunt the force of the present objection. It may be that part of what underlies the objection is that thought that there must be a limit to how far any power -- regardless the form -- ought to be able to interfere in an individual’s activities if that person is be free. This is the thought that, in the last chapter, I argued underlies the classic liberal critique of republicanism. However, there are two ways this intuition might be accommodated: The first is to insist that freedom simply be defined as non-interference. The second is to define freedom as non-domination, but to insist that part of what makes arbitrary power arbitrary is that it is not subject to constraints that limit the extent of its interfering reach. In Chapter 5 I will argue for a definition of arbitrary power that does just that. On such a view, the ideal of freedom is sensitive to the extent of interference individuals face: It shares with the non-interference account 162 Pettit employs such a distinction on behalf of the non-domination account. See: Republicanism, 104; 196“Keeping Republican Freedom Simple: On a Difference with Quentin Skinner,” 347.the conviction that freedom requires that a person not actually be interfered with beyond certain limits (since that would by definition amount to an arbitrary interference, and thus a constitute an exercise of arbitrary power). Yet, it differs from the non-interference account in that it holds that not only must no one actually interfere beyond these limits, but also that no one have the power to do so.   It’s true, this approach does not directly solve the problem presented by the Marcus and Georgia cases, since both Marcus and Georgia are subject to a master whose power to interfere is unlimited. However, it helps explain the temptation to mark a difference between them: The interferences Marcus faces are limited, and so his position more closely resembles the position of a free person than does Georgia’s. Nonetheless, in the end this is an illusion, because Marcus enjoys the non-interference he does only by the leave of his master.   3.7. Summary and conclusion Let me sum up the conclusions of this chapter:   I argued that there are two compelling reasons to favour the non-domination conception of freedom over the non-interference conception: First, individuals subject to the arbitrary power of another agent are liable to “self-edit” their actions in various ways even in the absence of actual interference. Second, as the phenomena of slavery and despotism make clear, an individual can enjoy a certain degree of non-interference while still remaining unfree insofar as their enjoyment of this “liberty” is subject to the discretionary control or authority of another person.     However, recent proponents of the non-interference account have advanced several important objections to this conclusion. The first set of objections targets the 163republican claim that freedom can be diminished in the absence of interference. Ian Carter has argued that the plausibility of this claim depends of an overly-simplified interpretation of the notion of “interference.” On Carter’s more sophisticated interpretation, “interference” includes both actual and “subjunctive” interference, where X “subjunctively interferes” with Y just in case, if Y were to attempt to perform a particular action, X would be likely to interfere with Y’s doing so. This sophisticated version of the ideal of non-interference is able to accommodate the insight that a person’s freedom can be diminished in the absence of actual interference, without thereby jettisoning the claim that the absence of interference is necessary for freedom. Moreover, as both Carter and Kramer have argued, the remaining difference between the non-domination and non-interference tells in favour of the latter. In cases where one agent possesses the power to interfere arbitrarily in the actions of another but is very unlikely to do so -- as in the “gentle giant” example -- the non-domination view implausibly counts that as a diminishment of freedom, while the non-interference view does not.   I argued that Carter and Kramer are right to insist that it is only the probability that a powerful person will interfere that is relevant to freedom, not the mere possibility; however, wrong to think that this conclusion supports the non-interference ideal. “Subjunctive interferences,” as Carter defines them, are not plausibly viewed as instances of “interference” at all. Instead, the account that is best supported by these considerations is the “amended” non-domination account, according to which a person is free to the extent that they are not subject to the arbitrary power of another agent, where that power is likely, with a suitably high degree of probability, to result in actual 164interference. This amended version of the non-domination ideal is, I argued, fully consistent with the traditional republican conviction that slavery and despotism are necessarily antithetical to freedom.   The second set of objections target the republican claim that not all interferences diminish a person’s freedom (i.e. those that involve the exercise of non-arbitrary power). According to the first of these, the non-domination account is implausible because it implies that a justly convicted criminal is not unfree even though imprisoned. I argued that this objection is less compelling than it initially appears. In the first place, while it’s true that the imprisoned man lacks freedom in one (fully intelligible) sense, it is not obvious that his political freedom has been in anyway infringed. In the second place, I argued that even the proponents of the non-interference account, if they intend to offer a workable “theory of freedom,” will be forced to discriminate between interferences of the sort faced by the imprisoned criminal and others that involve less legitimate exercises of power.   According to the second objection, the non-domination account implies implausible judgements about the comparative extent of freedom of different individuals. It implies that Adam’s freedom is diminished while’s Eve’s is not, though they face exactly the same constraints on their actions by the very same means. Moreover, it implies that Marcus and Georgia are equally unfree, though Georgia’s actions are much more extensively constrained than Marcus’s. I argued that the problem in the Adam and Eve cases is really no different than in the imprisoned criminal case, and resolvable in the same fashion. However, the Marcus and Georgia cases are more difficult. I argued that the temptation to view their extent of unfreedom as different can be weakened in 165light of the distinction between unfreedom and the conditions of its “disvalue.” Moreover, I suggested that any remaining intuition favouring distinguishing these cases might be plausibly explained as an understandable but mistaken application of a genuine insight into the nature of arbitrary power.    If the arguments just summarized are sound, then we have good reason to favour the non-domination conception of freedom over the non-interference conception. However, in light of recent criticisms, we also have reason to amend the non-domination account along probabilistic lines. Moreover, as my discussion of the last objection indicates, it is clear that the adequacy of the ideal of non-domination will depend in part on how exactly we understand the nature of arbitrary power. In the next chapter I turn to examine several proposed accounts of arbitrary power.
166Chapter 4 The Nature of Arbitrary Power ?In the last chapter, I argued in favour of the view that freedom ought to be conceived of as non-domination rather than non-interference -- that is, that freedom consists in the absence of subjection to arbitrary power. However, I deliberately left open, as much as was possible, the question of how we ought to understand the nature of arbitrary power. According to the currently influential neo-Roman account, arbitrary power is power that is not forced to track the interests of those subject to it. In this chapter, I make a critical examination of this view under a number of proposed interpretations, and argue that none of them is ultimately able to provide a satisfactory answer to the problem of majoritarianism. I conclude that the “interest-tracking” account of arbitrary power, and thus the distinctively neo-Roman version of the ideal of non-domination, ought to be rejected. I suggest an alternative “rights-respecting” account, a view that I explain and defend in more detail in Chapter 5.    4.1. Introduction  In the last chapter, I suggested that the neo-Roman theory of political freedom can be summarized in three basic claims: (1) Freedom consists in the absence of domination by others (rather than the absence interference), where to be dominated is to be subject to the arbitrary power of another agent; (2) Arbitrary power is power that is not forced to track the interests of those subject to it; 167(3) In order for all social powers to be forced to track the interests of those subject to them, they must be subject to democratic mechanisms of contestation and control. ?I argued in favour of accepting the first of these claims. I now want to examine more closely the second and third claims. The second claim is, as we shall see, the pivotal one. Without a clear account of the nature of arbitrary power, the conception of freedom as non-domination asserted in the first claim will remain indeterminate. Moreover, the link between freedom and democracy asserted in the third claim is dependent on the definition of arbitrary power offered in the second -- both to justify the claim that there is such a connection, and to define the form of democratic government appropriate to a free republic.    The central question guiding this chapter is whether the neo-Roman conception of freedom as non-domination can serve as the basis for a theory of freedom that is not subject to the same basic problem that vitiates the neo-Athenian view. In Chapter 2, I diagnosed that problem as residing in the fact that the conception of freedom as collective self-mastery sets no limits on the authority of the collective to exercise control over the lives of individual members. So in this chapter we will pay special attention to how the sovereignty of individuals fares in the neo-Roman account of a free society.     4.2. What is arbitrary power?  We’ve already noted (3.2.2) that arbitrary power, most uncontroversially, is power that may be exercised “at the will or pleasure” -- i.e. the arbitrium -- of the power-holder. However, how exactly are we to understand this? Obviously it can’t mean that any power exercised willingly or with pleasure is thereby arbitrary. Suppose a local traffic 168enforcement officer issues a ticket to a motorist who has just run a red light. The officer exercises his authority to do so willingly, since no one has forced him to come to work or to pull over this particular motorist. Moreover, he takes a certain pleasure in knowing that enforcing the law against running red lights makes the roads safer for everyone. The idea must be rather that arbitrary power is power that may be exercised merely at the pleasure or will of the power-holder -- that is, unconstrained by anything or anyone external to the power-holder. If the officer had pulled over the motorist at random, or because he has a personal grudge against people driving red cars, that would be an arbitrary exercise of power. Yet, if arbitrary power is power that is not subject to constraints in how it may be exercised, what are the relevant constraints? 4.2.1. The procedural condition  One thought is that arbitrary power is power that is unconstrained by principle, procedure, or law. In the first scenario, the traffic officer’s ticketing of the motorist would count as an exercise of non-arbitrary power, then, because his power to detain the motorist and ticket him is subject to the local traffic laws. We are assuming, of course, that there is an effective set of checks in place to ensure that he only exercises such powers in accordance with the law -- that’s what it means to be constrained by law. In the second scenario, by contrast, the officer’s power is evidently not constrained by law. Perhaps the local enforcement agency is corrupt and allows its officers to issue tickets whenever they please since it generates revenue, or perhaps it lacks effective oversight measures to ensure its officers don’t abuse their power for the sake of personal vendettas.  169 The thought that power that is unconstrained by principle, procedure or law is arbitrary seems right. It certainly captures at least one essential aspect of what we ordinarily mean by “arbitrary” power -- namely, its discretionary character. Moreover, it also explains why individuals subjected to arbitrary power tend to self-edit their actions in the ways noted in the last chapter (3.3.1). When power is not constrained to operate according to law, principle, or procedure it tends to be unpredictable (at least in the strong sense of predictability — cf. 3.3.1), and so those subject to it are wise to act with caution and restraint if they wish to avoid drawing unwanted responses. For these reasons, I think we ought to accept that, in order for power to count as non-arbitrary, it must at least meet what Frank Lovett has called the “procedural condition”   197? Procedural condition: In order for power to count as non-arbitrary, its exercise   must be reliably constrained by effective rules, procedures or laws that are   common knowledge to all persons or groups concerned.  ?The requirement of the procedural condition for non-arbitrariness is roughly equivalent to the familiar ideal of the “rule of law.”  Of course, as the definition just noted makes 198clear, the mere existence of a codified body of law or set of formal procedures is not sufficient to render power non-arbitrary. It must be the case that the laws and procedures are effective -- i.e. “they must actually constrain how power is exercised,” and do so reliably.  199170 Lovett, “What Counts as Arbitrary Power?”, 139197 Ibid., 140198 Ibid., 139199 Moreover, they must also meet a publicity requirement: The laws and procedures by which power is constrained must be commonly known (and known to be effective and reliable). The reason for this should be clear: Even if one agent’s power to interfere in the activities of others is in fact effectively constrained by law, if those subject to this power are unaware of this fact or have no clear understanding of what those constraints consist in, they will have no basis for (a) believing that power is not likely to be exercised arbitrarily, (b) anticipating how it might be exercised, or (c) identifying and contesting perceived abuses of that power. They will thus be inclined to self-edit their activities in precisely the same way as if the power they are subject to was in fact unconstrained.  4.2.2. The interest-tracking condition  The procedural condition alone, however, seems inadequate to define the nature of arbitrary power -- or at least the notion of arbitrary power as it figures in the conception of freedom as non-domination. While something like a rule of law constraint seems necessary for power to count as non-arbitrary, it doesn’t seem sufficient. The reason for this is that procedural non-arbitrariness is “too easy to achieve.”  In 200principle at least, it seems fully possible to have a rigorous legal system that is, nonetheless, highly discriminatory and oppressive in the way it exercises power over those subject to it. In Saudi Arabia, for instance, women face a myriad of severe legal restrictions on their mobility, their choice of dress in public, their capacity to seek 171 Ibid., 141200employment, and their right to express themselves publicly and participate in politics.  201One of the particular legal restrictions they face is a rule against operating motor vehicles. Suppose, then, a Saudi traffic enforcement officer were to pull over and ticket a woman for driving a car. On a purely procedural conception of arbitrary power, the officer’s interference with the woman’s travel plans would not count as arbitrary -- since this interference is in accordance with the law and, we will assume, his powers of enforcement are effectively constrained to operate within that law. Nonetheless, it seems clear that his doing so infringes on this woman’s freedom and, moreover, that the law more generally puts women in a dominated position. The problem seems to be that the law itself is in some sense arbitrary -- after all, why should a person be prevented from driving a car simply because she is a woman?    For this reason, most republicans see the need to add a second, substantive condition to their account of non-arbitrary power. Thus the second basic claim of the neo-Roman republican theory of freedom states that non-arbitrary power must be constrained in a way that compels it to “track the interests” of those subject to that power. Let’s call this the “interest-tracking condition.” ? Interest-tracking condition: In order for power to count as non-arbitrary, its    exercise must be forced to track the interests of those subject to it. 172 The following is an excerpt from Freedom House’s annual report, Freedom in the World (2012), on the 201state of freedom in Saudi Arabia: “Women are not treated as equal members of society, and many laws discriminate against them. They were not permitted to vote in the 2005 or 2011 municipal elections, they may not legally drive cars, and their use of public facilities is restricted in some cases when men are present. By law and custom, Saudi women cannot travel within or outside of the country without a male relative. Unlike Saudi men, Saudi women cannot pass their citizenship to their children or foreign-born husbands. According to interpretations of Sharia in Saudi Arabia, daughters generally receive half the inheritance awarded to their brothers, and the testimony of one man is equal to that of two women. Moreover, Saudi women seeking access to the courts must be represented by a male. The religious police enforce a strict policy of gender segregation and often harass women, using physical punishment to ensure that they meet conservative standards of dress in public.”?In other words, it is not enough for power to be constrained by law, it must be constrained in a way that makes it accountable and responsive to the interests of those subject to it.  The interest-tracking condition appears to make sense of why a traffic officer’s pulling over and ticketing a motorist who has run a red light does not constitute an exercise of arbitrary power, while a traffic officer’s pulling over and ticketing a motorist for being a woman does. In the first case, the officer is acting within a law that serves the interests of those subject to it -- since all motorists presumably have an interest in the general observance of traffic light regulations. In the second case, the officer is acting within the law, but that law does not track the interests of many of those subject to it. It may track the interests of those who wish to maintain male social dominance, or at least of those employed as taxi drivers and chauffeurs; but it doesn’t seem to track the interests of women.     4.3. The free republic  According to the neo-Roman view, then, an individual is free (i.e. undominated) just in case she is not subject to the power of another to interfere with her actions in a way that is unconstrained by law and/or not forced to track her interests. Given this specification of the ideal of non-domination, how should we understand the free society? How should the non-domination of each member be reconciled or coordinated with the non-domination of every other?  173 Pettit identifies two possible general strategies for securing a free society: (a) the “strategy of reciprocal power” and (b) the “strategy of constitutional provision.”  202According to the strategy of reciprocal power each individual’s freedom may be secured by ensuring that he himself possesses a power sufficient to block or deter the exercise of interfering power by other agents. The aim then is to neutralize the capacity of any one individual to interfere arbitrarily in the actions of another by ensuring each person has enough power to directly hold in check the power of any other. The strategy of constitutional provision, by contrast, seeks to secure the non-domination of each individual, not by equalizing their individual power, but by establishing an overarching political authority with the power to regulate the power of private agents to interfere arbitrarily with each other. Here it is a third party -- the constitutional authority -- that ensures that the powers of private agents are constrained to operate according to laws that track the interests of those subject to such powers.    Pettit is rightly skeptical about reciprocal power as a stand-alone strategy for realizing a free society. After all, by what mechanism is the distribution of reciprocal powers to be achieved, if not by some central political authority? Perhaps the thought is that individuals in a state of nature will somehow naturally tend towards such a state. But that seems dubiously optimistic. Not unreasonably, Hobbes envisions the opposite outcome: a condition in which every individual has sufficient power to interfere arbitrarily in the lives of any other, and none has sufficient power to reliably resist such intrusions into their own. Pettit paints a somewhat different, but equally pessimistic picture: In the real world, he insists, individuals are not equal in their capacities to deter and defend; moreover, left to themselves, these inequalities are likely to accumulate over time, with 174 Pettit, Republicanism, 67202some more powerful individuals being able to recruit the services of henchmen to expand that power, leading ultimately to a society of “petty despotisms.”  Thus, while 203Pettit thinks the strategy of reciprocal power might have some value as a supplement to the strategy of constitutional provision, it is not feasible as a stand-alone approach to securing a free society.   A free society, then, requires the existence of a political authority with the power to constrain private agents from interfering arbitrarily in the affairs of others. However, this solution appears to solve one problem by creating another: That is, it appears to solve the problem of dominium (domination between private individuals) by creating a problem of imperium (domination of individuals by the state). After all, the political authority, to effectively prevent dominium, must possess a power to interfere in the actions of individuals far greater than any possessed by private agents.   However, the power of the political authority, though supreme, need not be dominating so long as it is non-arbitrary power -- that is, power that is constrained to operate in accordance with the law (procedural condition) and in the interests of those subject to it (interest-tracking condition). For this reason, according to Pettit, in a free republic, the political authority ought to take the form of a constitutional democracy.   204 It must be a constitutional democracy in order to satisfy the procedural condition. That means, in the first place, that the operations of the political authority itself must be subject to the rule of law: The laws “should be general and apply to everyone, including the legislators themselves; they should be promulgated and made known in advance to those to whom they apply; they should be intelligible, consistent, and not subject to 175 Pettit, Republicanism, 93203 ibid., Chapter 6204constant change...”; and government decisions and action ought wherever possible to be undertaken on a legal rather than a particularist basis (1997: 174). It also means that the powers which officials of the political authority require ought to be dispersed. This dispersion of power requirement includes both the need to ensure adequate separation between the bodies that carry out the legislative, judicial and executive functions, and the need to ensure that government agencies in general are subject to mechanisms of independent audit, appeal and accountability. The rule of law and dispersion of power constraints work in tandem to ensure the procedural condition for non-arbitrariness is met: The dispersion of power constraint ensures that political power is effectively and reliably constrained to operate according to the rule of law by adding checks and balances that protect against abuse of power.   205 The political authority in a free republic must also be a constitutional democracy. This second requirement is necessary in order to satisfy the interest-tracking condition for non-arbitrary power. Recall: the interest-tracking condition requires not merely that the political authority act in the interests of those subject to it, but that it be forced or constrained to do so. There must, therefore, be some effective and reliable mechanism or mechanisms in place that compel the political authority to act in ways that are accountable and responsive to the interests of those subject to it. Pettit is also quite clear that the relevant interests are the avowed  interests of the citizens -- i.e. the 206interests they take themselves to have, not necessarily their “real” interests according to some purportedly objective standard of a good or flourishing life. So specified, the 176 Pettit adds a third constitutional constraint -- “the counter-majoritarian condition.” I discuss this 205condition below.  Pettit, “The Common Good,” 151206interest-tracking condition requires that there be effective and reliable mechanisms by which the political authority is forced to track the avowed interests of those subject to it. In short, it requires that the political authority be subject to democratic control.   Pettit is careful, though, to specify that it is a contestatory rather than a consensualist model of democracy that best meets the requirement of the interest-tracking condition. A consensualist model of democracy would require that each of the decisions and actions of government actually be consented to by those subject to it -- i.e. that non-arbitrariness of public decisions comes from their “having originated or emerged according to some consensual process.”  If that is what was required for 207democratic control, then it would prove to be “an inaccessible ideal.”  A contestatory 208model of democracy, by contrast, requires of public decisions only that “if they conflict with the perceived interests and ideas of the citizens, then the citizens can effectively contest them.”  “What matters is not the historical origin of the decisions in some form 209of consent but their modal or counterfactual responsiveness to the possibility of contestation” (ibid). The primary role of the democratic mechanisms, then, is more closely analogous to that of an editor than that of a writer.   210 On the contestatory model, democracy is about more than the opportunity to elect representatives to the legislature. It is about the existence of a whole range of channels for contesting decisions and actions of officials and agencies in all three 177 Ibid., 185207 Ibid., 184208 Ibid., 185209 This point marks an important difference between the neo-Roman ideal of the free republic and that 210associated with the neo-Athenian ideal of collective self-mastery (Chapter 3). The latter approach tends to see democracy more idealistically as a vehicle for forming and giving expression to a collective self-legislating will.branches of government. Moreover, there must not only be channels for voicing contestations, there must be suitable forums in place for hearing and responding to those contestations. That is, the system must be designed in such a way that contestations are effective in amending decisions and policies.    4.4. Tyranny of the majority?  At this point, we might wonder whether the neo-Roman ideal of non-domination is vulnerable to the same basic worry that plagues the neo-Athenian ideal -- i.e. the problem of a tyranny of the majority. After all, like the neo-Athenian view, it draws a tight connection between the freedom of individuals and the existence of democratic forms of governance. In what follows, I will argue that it is vulnerable to this worry, though the route to seeing this is a complicated one. Let’s begin with the simplest presentation of the problem. 4.4.1. Initial statement of the problem  Thus far I’ve analyzed the neo-Roman account as being committed to the following sequence of claims: (1) Freedom consists in an absence of subjection to arbitrary power; (2) Arbitrary power is power that either (a) is not constrained to operate according to the rule of law (the procedural condition) or (b) is not forced to track the interests of those subject to it (the interest-tracking condition); (3) In order to ensure that individuals are not dominated by other private agents (dominium), there must exist a political authority with the power to prevent such agents from interfering arbitrarily; (4) In order to ensure that the political authority does not itself dominate those subject to its power (imperium), its power must itself be constrained to operate according to the rule of law and be forced to track the interests of those subject to it; (5) The requirement that 178the political authority’s power be forced to track the interests of those subject to it is fulfilled by existence of effective and reliable mechanisms of democratic contestation of its decisions and actions.   The basic problem with this picture is that the interests of citizens often conflict -- particularly if we are taking “interests” to mean avowed interests. So it will often be impossible for a political authority to track the interests of all those subject to it. But that presents us with a difficult problem: In cases of conflict, which interests should the political authority be designed to track? The fact that the interest-tracking condition is supposed to be realized through democratic mechanisms appears to suggest that the political authority ought to be designed to track the interests of the majority. However, that would land us right back in the problem of the tyranny of the majority. Or to put the point in republican language, it would seem to sanction the domination of the minority by the majority. After all, it would then be that case that individuals with minority interests are subject to a power that is not forced to track their interests.   In fact, the problem appears to be even more serious. Not only do those with minority interests meet the criteria of dominated individuals, even those whose interests are in line with the majority meet the criteria. The power of the political authority is not forced to track the interests of any individual qua individual. The fact that the political authority happens to act in accord with the interests of an individual in the majority is not sufficient for such an individual to count as non-dominated. Such a person is in a similar position to that of the slave with a permissive master or the subject of a liberal despot: The power he is subject to happens not to interfere with his aims, but his enjoyment of 179this good fortune is subject to the control of a collective body over which he, qua individual, lacks meaningful control.    4.4.2. Pettit’s counter-majoritarian condition  Pettit is not unaware of this tyranny of the majority problem; he acknowledges that if the political authority and law were subject to easy majority control it would constitute a form of arbitrary power over those in the minority.  To address this 211problem, he adds a third necessary element to the constitutional framework of a free republic -- i.e. the “counter-majoritarian condition.”  The counter-majoritarian condition 212requires that “at least where the basic and important laws are concerned... it should not be easy to change these laws... it should require more than the mere fact of majority support in the parliament or even in the population.”  Pettit allows that the specific 213counter-majoritarian measures adopted will vary from political community to political community; but, among the possibilities are requirements of a super-majority for changing more basic laws, including among those basic laws a bill of rights, requiring non-basic laws be passed by two differently constituted houses of parliament, etc.     The general effect of these measures is supposed to be to make changes to the law, especially the most basic and important laws, “pass along a particularly difficult route.”  They constrain the law-changing power of majorities by (a) raising the 214threshold for the size of majority needed to change more basic and important laws, (b) diversifying the constituency from which majorities are formed; (c) screening out fragile 180 Ibid., 181211 Ibid., 180212 Ibid., 181213 Ibid.214and short-lived majority interests (since temporary spikes in public opinion or weak preferences will not survive the demands of a complex legislative process); (d) carving out domains of individual rights for protection by basic laws.  There are two questions to ask concerning such counter-majoritarian measures: First, are they sufficient to secure against majoritarian domination? Second, is there an adequate basis in the neo-Roman conception of freedom to actually justify them?  4.4.3. Are the counter-majoritarian measures sufficient?  There are a couple of reasons to think that the answer to the first question is no. First, merely making it more difficult for majorities to change the (basic) laws will not be sufficient to rule out majoritarian domination, and may in fact have the opposite effect. Suppose the existing (basic) laws are themselves dominating -- that is, e.g., suppose they already favour the interests of the majority in ways that seriously compromise the interests of the minority. In that case, making them more difficult to change will actually entrench majoritarian domination. Making laws more difficult to change is a measure that cuts both ways. In fact, this is true even if the existing laws allow for minorities to dominate the majority. For example, suppose the law gives the right to private ownership of property to male citizens only. Constraints that make such laws more difficult to change will only serve to entrench domination. The proposed counter-majoritarian constraints on changes to laws could only serve the cause of non-domination if the existing laws are already non-dominating.  Second, even if we start from favourable constitutional conditions (i.e. the legal status quo is presently non-dominating), it might seem that the counter-majoritarian measures proposed still do nothing to change the dominating character of the 181democratic state vis-a-vis any given individual: After all, it remains the case that the probability that any given individual’s voice will end up being decisive on any measure is vanishingly small. From which it appears to follow that the laws by which citizens are governed are still not forced to track their avowed interests qua individuals.   However, I think it is more reasonable to conclude that the counter-majoritarian measures can go some way to reducing majoritarian domination (again, assuming we start from favourable legal conditions). For one, they screen out many majority views from becoming law simply by making the process slower and more difficult. Further, because the constraints have the effect of requiring (or making it more expedient to have) larger and more stable majority support for law-making, they make it less likely that any given individual will be subject to laws that they do not approve (since it lowers the probability that an individual will find themselves at odds with the majority decision). That is, the constraints make it such that (a) for any given individual ‘x,’ it is less likely that any particular law ‘y’ imposed by the democratic state, will be one that is contrary to x’s avowed interests AND (b) the fact that ‘a’ holds is a consequence in part of the fact that y is contrary to x’s avowed interests. Hence, while the constraints imposed do not force the laws to track the avowed interests of any given individual qua individual, it does raise the probability that an individual will find himself or herself a member of the majority view, which the system is forced to track. If that is right, then we can conclude that Pettit’s counter-majoritarian measures reduce the vulnerability of individuals with minority interests to alien control by majorities, but does not eliminate it. 4.4.4. Can the counter-majoritarian constraints be justified?  The second question we must consider is whether there is in fact an adequate 182basis in the neo-Roman conception of freedom to actually justify and define the content of Pettit’s counter-majoritarian constraints.    The answer to this second question, I think, is no. In cases of conflicts of interest, measures which constrain the majority from changing the law to conform to their interests will constitute a domination of the majority by the state. After all, those in the majority will now be subject to power that is not forced to track their interests, but instead tracks the interests of the minority. Suppose, for example, there is a constitutional provision which constrains the state from interfering in the religious practices of anyone in its jurisdiction and requires it to prevent such interferences by other private agents. Suppose further that in this jurisdiction the majority of citizens are adherents of a particular religion, and those adherents consider the religious views and practices of a certain minority group to be blasphemous. As a result the majority has an interest in interfering with any public expressions of the minority religious view. The constitutional provision protecting religious liberty would seem to be a clear case of a counter-majoritarian measure aimed at preventing majoritarian domination of the minority. It tracks the interests of the minority in being able to practice their faith. However, it does so, it seems, only by establishing a domination by the state over the majority -- since it constitutes a power to interfere with the actions of majority that is not forced to track, and indeed conflicts with, their interests in stamping out blasphemy.  It is difficult to see then how the neo-Roman ideal of non-domination could serve as a sufficient justificatory basis for counter-majoritarian constraints. The counter-majoritarian constraints protect against domination of minorities only by instituting domination of majorities. By itself, the neo-Roman conception of non-domination can tell 183us nothing about what to do in conditions of conflicting interests, never mind provide a criterion for determining when minority interests ought to be protected and when majority interests ought to hold sway.  Pettit at least appears to be aware that the ideal of non-domination cannot be fully satisfied for every individual in a given political community. In light of this, he considers two possible ways to understand how this ideal might function to regulate the state’s power in a free society: (a) as an absolute side-constraint or (b) as a good to be maximized.  215216 The side-constraint view clearly will not work. If non-domination were a side-constraint, then no law would be permitted to be passed nor any state power authorized which was not forced to track the avowed interests of every individual concerned. But a consensus requirement of this sort would be unacceptable for two reasons. In the first place, it would in practice mean that, in any world resembling our own, no law could ever be passed and no state power ever legitimized -- since every such law or power is bound to conflict with the interests of someone. The side-constraint view would, in effect, favour social anarchy -- not anything like the constitutional democratic state Pettit advocates in his theory of government. In the second place, given the side-constraint view would effectively favour social anarchy (to eliminate imperium), it would open the door wide to domination between private agents (i.e. dominium).  184 Ibid., 102215 These are not the only two possible options of course. For instance, someone might favour a 216maximum equal allocation of non-domination, or some other such principle. However, introducing a distributive principle of this sort would involve building a conception of justice into the very definition of a free society -- an approach Pettit has resisted. I discuss this issue in more detail in Section 5.1.   Pettit, however, endorses the view that non-domination be understood as a good to be maximized.  But this view, it seems to me, sends us right back to the problem of 217the tyranny of the majority. For what would be required if non-domination is a good to be maximized? Presumably that in any case where there are conflicts of interests, the law ought to be forced to track the avowed interests of the larger party of individuals. After all, if power is non-dominating to the extent that it tracks the interests of those subject to it, then maximizing non-domination would appear to require that power be forced to always track the interests of the many over the few. But if that is right, then it turns out that the counter-majoritarian constraints actually work against the maximization of non-domination, and so cannot in fact be justified by that ideal.   This same problem vitiates the very distinction between basic laws and non-basic laws that underpin the counter-majoritarian constraints. The counter-majoritarian measures Pettit highlights aim to constrain the ability of majorities to change law. However, if all the laws were very difficult or impossible to change, then the system would cease to be democratic in any meaningful sense, and certainly not in the robust sense the interest-tracking condition requires. Thus, an important feature of Pettit’s proposed counter-majoritarian constraints is that they distinguish between basic or fundamental laws which should be very difficult to change by majorities and other laws which should not be so difficult. The inclusion of a bill of rights among the basic laws most obviously carves out domains of individual choices that should not be subject to easy majoritarian control in the way that other matters normally would be.   However, how are we to understand the basis of this distinction between basic law that should not be forced to track the interests of the majority and non-basic law 185 Ibid.217which should? Pettit claims that the criteria for determining what the most basic and important laws are is simply the ideal of non-domination.  Yet, this does not answer 218our question: The ideal of non-domination requires that laws be forced to track the avowed interests of the governed. But our question is about the basis for distinguishing between cases where the laws ought to track the interests of the majority of the governed from cases where the laws ought instead to track the interests of the minority against the interests of the majority. Some other criterion is needed to tell us which domains of individual choice must be carved out from the scope of non-basic law which is subject to ordinary majority control. The ideal of non-domination as Pettit conceives it is simply insufficient to ground the distinction upon which his counter-majoritarian constraints (and, in particular, his advocacy a bill of rights) depend. 4.5. Rethinking arbitrary power  In light of this persistent problem with majoritarianism, we might be tempted to reject the conception of freedom as non-domination all together. Yet, as the arguments of Chapter 3 were intended to show, there are good reasons to think that there is something right about that conception. For this reason, I suggest an alternative approach. I’ve distinguished three basic claims of the neo-Roman theory of freedom: (1) Freedom consists in non-domination -- i.e. the absence of subjection to arbitrary power; (2) Arbitrary power is power that is (a) not subject to the rule of law and/or (b) not forced to track the interests of those subject to it; (3) In order for all social powers to be forced to track the interests of those subject to them, they must be subject to democratic control. I’d like to propose, then, that the source of the problem is not with ‘1,’ but 186 Ibid., 182218instead with ‘2’ or ‘3.’ Moreover, because the justification for ‘3’ is supposed to lie in ‘2’ (‘2b’ in particular), it seems to me ‘2’ is the right place to start. That is, perhaps the problem of majoritarianism is a product of this particular account of arbitrary power.   Ultimately I intend to argue that this is indeed the case -- and specifically that it is the interest-tracking condition that is the source of the problem. However, I want to consider first whether there might be another way to understand the interest-tracking condition than I have so far considered and whether this might provide a satisfactory solution to the problem.  Thankfully, Frank Lovett has recently provided a useful catalogue of alternative accounts of the nature of arbitrary power, at least within the broadly republican line of thinking.  Following Lovett, I will take as given that any satisfactory account of 219arbitrary power must at least hold that power is arbitrary if it fails to meet the procedural -- i.e. “rule of law” -- condition. Lovett surprisingly thinks the procedural condition is sufficient to define the nature of arbitrary power, defending what he calls the purely procedural account. I’ve already offered reasons to reject such a view however (4.2.2). For this reason, as Lovett himself acknowledges, most republicans see the need for a further “substantive” condition.  As we’ve seen, according to neo-Roman republicans, 220this further substantive condition consists in the claim that arbitrary power is power that is not constrained to “track the interests and ideas” of those subject to that power.  221 However, it turns out that the interest-tracking condition is subject to a variety of interpretations, depending on how we interpret the idea of “interests” invoked in the 187 Lovett, “What Counts as Arbitrary Power?”219 Ibid., 140220 Ibid.221condition. The notion of “interests” is a notoriously slippery one, subject to much philosophical controversy. Lovett identifies three prominent interpretations of the notion of “interests” advanced in this context: (i) The “objective interest” interpretation; (ii) the “welfarist” or subjective preference interpretation; and (iii) the “deliberative interpretation.”  2224.5.1. The ‘objective interest’ interpretation  On the objective interest interpretation, a person’s “interests” are understood as their “objectively-defined, normatively justifiable interests.”  In other words, non-223arbitrary power must be responsive to the “real” interests of those subject to it, whether they recognize them as such or not. However, this interpretation is subject to serious, and in my view decisive, objections.   First, it requires that we assume some fixed standard of real human interests -- a conception of what a good human life consists in. Yet any such standard is bound to be controversial in modern pluralistic societies consisting of diverse communities with, at points, divergent comprehensive doctrines of the good. Adopting an objective interpretation of interests would, consequently, make the very identification of instances of domination controversial. But, as Lovett argues, an account of arbitrary power that interpreted “interests” in accord with the objective view would lack the broad-based appeal any practical political ideal must have in pluralistic society.   224 A second, related objection, made forcefully by Isaiah Berlin in a somewhat different context, is that associating freedom too closely with the realization of objective 188 Ibid., 141-2222 Ibid., 141223 Ibid., 142224or “real” interests provides a too easy justification of coercion for those occupying positions of power. Having distinguished the “real” interests of individuals from their actual “empirical” preferences and aims, it becomes easy for those in power to justify coercing recalcitrant subjects in the name of their real interests: “I am then claiming that I know what they truly need better then they know it themselves,” and in consequence, “[t]his renders it easy for me to conceive of myself as coercing others for their, not my, interest.”  Taken to the extreme, such a view can serve as the justification for 225thorough-going totalitarianism, though even in its less dramatic uses it provides a too-easy defence for paternalism and intolerance.    This second objection is closely related to a third: Namely, if the interests non-arbitrary power must track are the objective interests of those subject to it, who is to determine what these objective interests are and, further, by what process or mechanism might such power be forced to track such interests? Because those subject to such power might be mistaken about their own real interests, it is not obvious how the relevant standard for regulating exercise of such power is suppose to be fixed. As the second objection suggests, the likelihood that those who occupy positions of power will arrogate the authority to make such judgements seems high. Moreover, it is unclear how, on the objective interests account, power is supposed to be forced to track the interests of those subject to it. Because those subject to such power are not necessarily the best judges of their own interests, mechanisms of democratic contestation and control would be seem to be inappropriate tracking instruments. But how else could 189 “Two Concepts,” 132-3225those in positions of power be held accountable to the interests of those subject to their power?     226 A final objection to the objective interest interpretation of arbitrary power is that it is not obviously in any better position to solve the problem of conflicting interests. That is, it seems reasonable to think that, even on an objective account of interests, the interests of one individual are bound to come into conflict with the interests of others. We will then still be without a principled criterion for determining which objective interests ought to be tracked.    2274.5.2. The ‘subjective preference’ interpretation  According to the second, “welfarist” interpretation, “interests” ought to be understood as subjectively-expressed preferences or desires. This is, in effect, the interpretation I was implicitly attributing to Pettit in the last sections: It is the avowed interests of those subject to power that must be tracked. This view avoids the first three objections to the objective interests account: It requires of non-arbitrary power that it track the self-avowed interests of those subject to it, and so does not depend on a controversial standard of objective interests. It thus also rules out the possibility of totalitarian regimes justifying coercion on grounds that they are merely serving the “real interests” of the recalcitrant individuals over whom they exercise their power. Moreover, there is a clear process by which such interests could be forced to be tracked -- i.e. democratic mechanisms.   190 I’m not entirely skeptical that this question could be given some kind of answer. Perhaps, so long as 226there is some public legal standard defining the objective interests, the rule of law and separation of powers constraints would be sufficient to hold those in positions of power to account. For instance, an independent judicial body might have the power to strike down legislation that conflicted with the legal standard of objective interests.  One could perhaps define objective interests in such a way as to rule out such conflicts in principle; 227however, it is not clear how such an account would manage to do so without sounding implausibly ad hoc. However, as we’ve seen, adopting the subjective preference interpretation presents its own difficulties. No political regime will be able to track all the preferences of each of its citizens; preferences often conflict or must compete for scarce resources. It seems then that some further principle is required to determine which interests ought to be tracked and which ought not. But there is nothing internal to the welfarist interpretation of the interest-tracking condition that can serve as such a principle.   Perhaps, though, the interest-tracking condition could be amended to incorporate such a principle. Earlier we saw that Pettit had proposed a maximization principle: i.e. that non-arbitrary power ought to be forced to track as many of the interests of those subject to it as possible. However, I argued that this view landed us in the problem of the tyranny of the majority. But perhaps some other distributive principle would fare better. One obvious suggestion would be that a non-arbitrary regime is one that was forced to give equal consideration to the preference-satisfaction of each of its subjects -- i.e. a principle of maximum equal preference-tracking.   228 However, any approach of this sort faces at least a couple problems. In the first place, it seems inconsistent with the claim of at least some neo-Roman republicans that the ideal of freedom as non-domination is one that does not depend on a prior underlying theory of justice.  The worry here is that incorporating a distributive 229principle into the very definition of arbitrary power would constitute a moralization of the concept of freedom. In Chapter 6 I will argue that the objections to moralization are ill-founded; nonetheless, it remains a problem for the neo-Roman conception of non