• 1 Introduction

  Among the many admirable people I have known over my career in academia, John Braithwaite commands a special place. He has invariably advanced our understanding of the topics he has investigated. He has done groundbreaking work of this kind across a number of disciplines and domains. And he has always shaped his research agenda by a keen sense of where he is likely to do most good. In displaying all these features at once, he has put himself in a class of his own: sui generis, beyond compare.
  John has made a deep impression on my own work in areas ranging from the economy of esteem to criminal justice, to regulatory theory, institutional design and global justice, to democratic and political theory. The deepest mark he made was in nurturing me to think about issues of justice in a fashion that we came to see as a contemporary version of the long republican tradition; hence the subtitle – A Republican theory of criminal justice – of Not just deserts, a book we wrote together in 1989 (Braithwaite & Pettit, 1990).
  Since the republican way of thinking has shaped the work we have each done over the past thirty years, it seems appropriate to devote this little essay to a topic in that area. The focus of the journal in which it appears is restorative justice, a project dear to John’s heart, and indeed to mine (McGeer & Pettit, 2015). But John himself sees the project as part of the wider concern with republican justice (Braithwaite & Parker, 1999; Braithwaite & Pettit, 2000). And so, I hope I may be forgiven for concentrating on a topic in republican theory more generally.
  I concentrate on an aspect of republican political thought to which John has made a striking and influential contribution (Braithwaite, 1997), developing insights nurtured and applied in the formidable range of his normative, regulatory and peace-related studies (see e.g. Ayres & Braithwaite, 1992; Braithwaite, 2000; Braithwaite & Drahos, 2000; Braithwaite, Charlesworth & Soares, 2012). That topic is the separation and sharing of powers encoded in the republican idea of a mixed constitution.
  The republican tradition dates back to the Roman Republic, when figures like Polybius and Cicero identified its main elements, arguing that Rome exemplified all.¹x In this brief introduction to republican thought, I draw on my 1990 book with John and my later study, entitled Republicanism: a theory of freedom and government (Pettit, 1997). For a complementary account published at the same time by someone whose work we both drew upon, see Skinner (1998). For a more up-to-date overview of republican research, see Lovett and Pettit (2009). It went into the background of institutional and philosophical thinking with the rise of the Empire but was resurrected in the independent trading cities that emerged in Northern Italy in the high middle ages; these cities identified themselves as republics and prided themselves on their independent status and on the independent status enjoyed by their (often, alas, elite) citizens. The core ideas of the tradition were articulated by Niccolo Machiavelli in the early sixteenth century in his Discourses on Livy’s histories and were influential in shaping the Polish Republic of the Nobles in that century, the Dutch Republic that emerged in the seventeenth century and the English Republic of the 1640s and 1650s.
  Despite the restoration of the king in 1660, republican ideas became the staple of political thought in England and the areas it controlled in the eighteenth century, with James Harrington and Algernon Sydney being taken as important theorists of the tradition. It assumed different forms, however, some more conservative, some quite radical. The radical strand was dominant in critiques of the status quo such as the anonymous Cato’s letters, which appeared in the 1720s, and made its greatest impact in the American War of Independence of the 1770s. The republican ideas behind that revolt were memorably rehearsed in the Federalist papers of 1787, in which James Madison, Alexander Hamilton and John Jay defended the constitution devised in that year for the new republic.
  There are, broadly, three main ideas in the republican tradition, which faded in the nineteenth century in the face of the rise of classical liberal thought. One is the idea that the freedom of citizens in making personal choices requires that they not be subject to the power of any master or dominus, even a wholly good-willed one, and that the state can give citizens equal freedom as non-domination by protecting and empowering them equally under the law. The second is the idea that if the state is not itself to be a dominating presence in the lives of its citizens, it should be organised under a mixed constitution that separates and shares the powers of shaping and imposing law across different bodies, including bodies sometimes elected by the people; and the third is that if this mixed constitution is to survive and thrive, then citizens had better be vigilant and virtuous in interrogating and challenging those in power: they had better have a contestatory profile.
  Here I concentrate on the separating and sharing of power that has had such prominence across John’s writings. After a brief introduction to the idea, I distinguish between a radical and a moderate form that it may assume at the highest level of political life, one presidential the other parliamentary, and then I argue in favour of the moderate, parliamentary kind. This turns out to be an argument in favour of the Australian system of government rather than the American, and I hope it may seem like a fit offering with which to honour John Braithwaite’s extraordinary contribution; this, after all, has put antipodean studies at the forefront of international social science.

• 2 The mixed constitution

  The idea of the mixed constitution idea was prominent in the work of Polybius (2011, Bk 6) in the second century BCE, and it was strongly endorsed by other, later figures in the long republican tradition. These include Niccolo Machiavelli (1997) in the late medieval period, James Harrington (1992) during the seventeenth-century English Republic, and the authors of the Federalist papers in revolutionary America at the end of the eighteenth century (Madison, Hamilton & Jay, 1987). These figures were all institutional innovators, and they each took the mixed constitution under a variant interpretation of their own.
  Organised as an agent with a monopoly over the exercise of coercion within the territory that it polices, the state is capable of doing much good but also much harm. Every political theory seeks to identify the good it can do – in the republican case, the good of enabling people to be their own undominated men and women – but equally, every theory should look to how the state can be constrained so that those in office do not abuse the power to further that good. By John’s account, and by the republican tradition in general, the state should have a good deal of power; it should not shrink, for example, to the restricted dimensions of the neoliberal night watchman. But the power it has over people’s lives should not be dominating. It should not be a form of power that is exercised on the basis of an unconstrained arbitrium or will; it should not be arbitrary or discretionary in that distinctive sense (Pettit, 1997).
  How to constrain the state, so as to render it non-arbitrary? How to constrain it so that it acts, broadly, under a discipline imposed and endorsed by its people, not at the discretion of those in office, or their friends? The republican response has always been to suggest that it should be required to operate via a coordinated network of subagencies – modules – each with a characteristic power and responsibility of its own. Power should be dispersed across a variety of centres, each responsive in its own way to popular concerns, and it should be exercised under constraints that ensure the coherence of those agencies, letting the law speak with a single voice. These are the guiding ideas behind every version of the mixed constitution (Lieberman, 2006).²x A trope often used to characterise the mixed constitution was that it involves a compromise between the three classical models of democracy, aristocracy and monarchy. But this was not taken literally at any point, even by Polybius: thus, for him, the consuls represented monarchical power, even though there were always two, not one, and they held office only for a year.
  The original model of that constitution was provided by republican Rome, as Polybius saw it; he was a Greek who went to Rome as a hostage but stayed on to be one of its most enthusiastic commentators. Members of the elite Senate, who held or would have held public office, could propose any measure for legislation. And it would then pass into law if it was supported at a meeting of one of the relatively inclusive or popular bodies to which it might be referred: usually, to the tribal or to the centuriate assembly of citizens. When it came to implementing the laws, this was done by annually elected public officials, with the advice of the Senate, where in each kind of executive office, power was divided between a number of individuals: at the top of the hierarchy, for example, between two consuls. And when it came to adjudication, this was done by bodies of citizens, sometimes as many as two hundred, assembling to hear the case. The guiding idea behind the arrangement is caught in a medieval Italian principle: where many councils exist, there is the safety of the people; ubi multa consilia, ibi salus populi (Waley, 1988: 39-40).
  What sorts of devices might be employed for modularising or decentralising power? Among procedural devices, as we might call them, three stand out. One is a separation of power between judicial authorities and legislative-administrative ones, and possibly between legislative and administrative authorities as well. It requires those separate bodies to coordinate in determining the precise content of any law that is to pass muster over the longer term. This separation of powers was given particular prominence in the eighteenth century by the Baron de Montesquieu (1989) in his Spirit of the laws.
  The second device, more central in the early versions, imposes a sharing of power among different individuals or bodies in the exercise of one or more of these roles. This is exemplified in the way different electoral and legislative bodies shared power in Rome, as it is in the later arrangement for a bicameral legislature with an upper and lower house. It might also be taken to be illustrated by a court system that operates under the law of precedent.
  The third device, prominent in more recent political formations, is the outsourcing of one or another form of power to independent agencies that exercise an effective, if not always conclusive, influence over what is done. Such agencies include informational authorities like the bureau of statistics, a source of economic data or a press office; invested authorities like an electoral commission or a central bank; or invigilating authorities like inspectorates or commissions of inquiry (Tucker, 2018). The body to which power is outsourced will generally be appointed by the outsourcing power: in effect, the legislature or administration. But it will be more or less independent, depending on how far the outsourcing power can resist its judgments, whether formally or just as a matter of effective politics.
  Apart from relying on procedural devices of separating, sharing and outsourcing power, the modular ideal of the mixed constitution normally involves three broadly democratic devices, as we would describe them today.³x The notion of democracy in the Greek world was that of a system in which the people or demos had a great deal of power or kratos (Ober, 2008). It became associated with a voting system only in the work of modern absolutists – enthusiasts for a powerful monarchy – like Jean Bodin in France and Thomas Hobbes in England; they cast democracy as a participatory system in which every full citizen votes on every public issue. Because democracy in that sense was infeasible by most lights – Bodin and Hobbes would have welcomed this – no one in the modern republican tradition would have embraced measures described as democratic; it was only in the mid-nineteenth century, that the word regained a positive valence, becoming associated with representative systems. If constitutional provisions are generally only removable by special recourse to democratic decision-making – say, a referendum or something of the kind – this enforces a sharing of legislative power with the electorate. And if people play two standard democratic roles in electing people to government, and in contesting what government proposes or actually does – say, in the courts, the media or the streets – that deepens the modularity of a system.
  Such democratic devices are not a novelty in the republican inventory of mixed constitutional arrangements. For example, the electoral and contestatory roles of ordinary people were of particular importance in the Roman republican system (Arena, 2012; Millar 1998). The election of officials – and indeed the making of laws – was in the hands of popular assemblies. And the capacity of ordinary people to challenge officialdom, usually by assembling in the forum, was identified by commentators like Machiavelli as one of the most characteristic and important features of the regime.

• 3 Decentralisation, radical and moderate

  In summary, then, most advanced democracies today decentralise the powers of shaping and imposing law, in six distinct ways: More traditionally, by separating judicial from legislative and administrative powers; by forcing a sharing of power in one or more of these three areas; and by requiring the legislative or administrative powers to outsource some of their duties in more or less independent agencies. And democratically, by allowing the people a legislative voice in referenda; by enabling people to determine electorally who holds the main offices in government; and by enabling them individually or in smaller grouping to contest government policies and laws, whether in the courts, the media or the streets.
  There are many possible variations in the modularity or decentralisation of a political system, as already indicated in the last section. But there is one dimension of variation that is particularly worth noting. This is the sort of variation that occurs between a presidential system like that of the United States and a parliamentary system like that associated with Westminster. A useful model of parliamentary democracy is provided by Australia, especially if we want to compare it with the presidential system of the United States. In Australia, as in America, there are two elected houses of parliament, a written constitution, a system of judicial review, a federation of relatively independent states and the same sort of electoral, generally non-proportional regime. Like the United States, it is certainly a decentralised state, displaying the three procedural and the three democratic devices mentioned. And like the United States, indeed, it is decentralised in another dimension to the extent that it is a federation of relatively independent states. But still, as we shall see, it is not modular or decentralised in the same radical degree.
  The reason for giving the presidential-parliamentary divide attention is to combat the common assumption that the mixed or decentralised constitution is specific to the American presidential system and that parliamentary systems are not decentralised in the same sense; that they install a single will in power, usually dignified as the sovereign will of the people. The first task is to clarify the differences between the ways in which they are each decentralised. Having clarified the distinction, we can then look at which mode of modularity has greater appeal.

• 4 Parliamentary democracy in Australia

  The main difference between the American and Australian systems, to focus on those exemplars, stems from the different ways in which the head of the executive or administration is chosen; once chosen, that person has a good deal of discretion in selecting other members of the administrative team. In the American system, the head of the executive – the President – is elected by the people just as members of the legislature are elected and has an independent, electoral source of authority. In the Australian, the chief executive – in this case, the Prime Minister – is elected by the members of the legislature from among their ranks and does not have an independent form of electoral authorisation (Linz & Valenzuela, 1994).
  Given the fact that the Prime Minister depends for retaining power on retaining majority support in the legislature, there must be members within the lower house – a party or coalition of parties – who are committed to making common cause with one another and with the chief executive, at least over the term of the parliament. And that means that the Prime Minister is at the head of a group that has to be corporately organised as an agent. This corporate organisation is the party or coalition in power – the government, in common parlance – and it is bound to have a high degree of control, not just over the executive that it keeps in power, but also over the legislature. Being a fixed majority, and having reason to make common cause with one another, the members will be able to decide on what laws to propose and pass.⁴x Although private members may have the right to propose bills in parliament, the government can almost always ensure that they do not get a hearing or that their bills do not get accepted.
  This control means that in the parliamentary system any proposed law or any proposed law-based policy will be authored in most cases by the corporate party in power. But authorship is only one form of control. Consider how the control over what is published in a newspaper or journal or book is not the author’s alone. It is shared with those in the position of an editor who can reject or revise what has been submitted, demand that certain additions or changes be made or at least impose pressures designed to elicit amendment. The polity inevitably involves the same mix of authorial and editorial control. And the distinctive thing about the parliamentary system is that there is a single, relatively unified author of laws and policies – the corporate party or population in power – there are a number of sites where editors or would-be editors can appear.
  Editorial efforts may be sponsored and supported by the opposition in parliament, of course. But they will generally be pursued outside legislature, say, by the various statutory authorities, including the courts, established under the constitution or under established convention; by public-interest movements or associations formed among the population; or by initiatives on the part of groups or individuals in the courts, in the media, or on the streets. The authorial role will be centralised within the legislature – and in one party or coalition in the legislature – while the editorial role will be distributed outside.

• 5 Presidential democracy in the United States

  It takes only a little reflection to see that the authorial-editorial divide is bound to assume a very different shape in the presidential democracy of the United States. In this case, there is no likelihood of a single party or coalition of parties assuming a corporate shape under the leadership of the President, since members of the legislature will not have to make common cause in order to keep the administration in power. Members of the same party, perhaps with some individuals from another party, may converge in the attempt to pass certain laws together. But they will each have individual incentives, usually associated local interests or lobby groups, that may keep them at loggerheads and render it difficult to make or indeed repeal laws. This is going to be particularly so, of course, when, as in normal times, no single party controls both houses of Congress and the Presidency.
  In this situation, it is more or less inevitable that now one legislator or set of legislators, now another, will propose measures for legislation in the hope of being able to muster majority support in both houses and win the consent of the President. And it is inevitable that they will face editorial challenges, not just outside the legislature, but also within. The result, which experience bears out, is that in this system the authorial role can be assumed, although not always with much prospect of success, by any members of the legislature, not just by the majority party. And the result, equally, is that any authorial initiative that gains traction among legislators can be edited beyond recognition inside the legislature itself as well as outside. The authorial role will be distributed across the members of the legislature, not centralised in one party or coalition; and the editorial role will be distributed both inside and outside the legislature. Where the author-editor relationship is one-many in the parliamentary system, it assumes a many-many character in the presidential.

• 6 The contrast between the systems

  These comments should make clear that where the presidential system exemplified by the United States is massively modular or decentralised, the parliamentary system exemplified by Australia is only moderately so. In the presidential system, there are many authors and many editors and so many distinct centres that have to be integrated in the making of decisions: many veto-points, as it is often put. In the parliamentary system there are also many editors or would-be editors. But there is only one effective author: the party in government.
  But before pressing on, we should note that in one respect this presidential-parliamentary contrast is overdrawn. The moderation in a democratic parliamentary system will be reduced – the system will approach the presidential extreme – insofar as the electoral system is designed so that the mix of views in the legislature and executive tends to be proportionate to the mix in the population as a whole. Proportionality would be very poor if, for example, there were single-seat majority constituencies, since a party with 20 per cent support might be defeated in every one of them. Proportionality would be much increased, with that party gaining roughly 20 per cent of the legislature, if there were an electoral system that operated, as in Israel, with a single constituency covering the country as a whole.
  If a parliamentary democracy is to be moderately, as distinct from radically, decentralised, then there must be a constraint on proportionality. In an extremely proportional system, there are likely to be many smaller parties. And with an increase in the number of parties there is likely to be a tendency to multiply the number of potential authors of law and policy within the parliament – and, correspondingly, the number of potential editors – especially if the government has the fixed support only of a minority: if it has to rely on the support of others outside that minority to get any law through.

• 7 Three major advantages of keeping modularity moderate

  In a radically modular regime like the presidential system of the United States – or like the highly proportional, parliamentary system in Israel – there are manifest incentives in place that are likely to lead to what we may call negligence. As those proposing a bill seek to get a majority on side in each house of Congress, and to make sure the President will not veto it, there is a powerful incentive to add more and more clauses to win support for the proposal. And this incentive can drive legislative proposals – omnibus bills, as they are called – that are so long and so complex that their coherence with other laws and with the constitution is often in doubt. In short, it can motivate a negligence among legislators as to what they are proposing and, as often happens, a reliance on the courts to sort out the mess. It is no accident that the judicial review of legislation plays such a prominent role in the United States.
  But as negligence on this front is a danger in a presidential system, so too is gridlock. As those voting on a bill are exposed to the influence of local or lobby pressure or tempted by a chance to exploit a bargaining advantage, they may be led to veto a measure that is important for the polity as a whole. We have seen this happen even with a measure as significant as honouring a sovereign debt. Since no particular agent, individual or corporate, will have to wear the responsibility for failing to deliver – everyone can blame someone else – they may each reject the perspective of the whole in favour of a more self-seeking viewpoint, thereby generating gridlock in policymaking.
  The dangers of negligence and gridlock are much less threatening in a relatively non-proportional, parliamentary system like Australia’s. With only one effective author of law, it is going to be easier in such a system to avoid these problems. The party in power will be able to get legislation passed and to make sure that it is up to par in virtues like clarity and coherence. After all, it will command a fixed majority in the legislature, or something close to such a majority, and not have to find a majority for each individual case. And this being a matter of common awareness, that party will be subject to reputational incentives to display that ability and prove itself a conscientious, effective government.
  The third respect in which the relatively non-proportional, parliamentary system promises to do better than the presidential is in providing the electorate with a good sense, in advance of an election, of the sort of laws that those they elect will put in place. Since laws are made in a presidential system on the basis, in each case, of the backroom dealmaking required to garner majority support for the bill, no individual or party going to the polls will be able to make a firm promise about what legislation will be put in place, if they are elected. And so the electorate will have to cast their votes on the basis of affiliation with very general principles, if principles divide the parties, or more plausibly on the basis of a tribal sort of identification with person or party.
  Things can be much better in a parliamentary system like that in Australia. Each potential government in a non-proportional parliamentary system, be it a party or coalition of parties, will be able to formulate a credible programme of legislation in advance. And, this again being a matter of common awareness, the party or coalition will have an incentive to propose such a programme and to seek to live up to it, should it win government. Thus, the voters will have a good idea of the legislative programmes between which they are choosing. The laws and policies that eventually materialise will not be the product of backroom, post-election deals, or at least not to the same extent. Most of the deals that are done will materialise before the election, appearing in the electoral programmes of the rival groups, and will be subject to the assessment of the voting public.

• 8 Three incidental advantages of keeping modularity moderate

  Apart from these three salient advantages, there are also three more incidental benefits that a parliamentary system promises. A first benefit is that if a single party, or a small coalition of parties, holds power, as in a parliamentary system, it is going to be more difficult for private-interest lobbies to influence what it does; they will not be able to focus, as in a radically modular system, on covertly pressuring or persuading marginal members who can push things their way. The mining company that seeks to influence government policy will find it harder – harder but not, alas, impossible – to persuade a party room to do its bidding than to persuade the few that can give them majority support in congress.
  A second, incidental advantage of the parliamentary system is that local voters and activists will have little or no chance of driving national policy by considerations of purely local advantage or interest, since representatives will have to vote, as a matter of common awareness, with their party: they will have to keep faith with the governing majority, and not let more sectional concerns drive their vote. In his famous 1774 speech to the electors of Bristol, Edmund Burke warned against the possibility of a legislature degenerating into a ‘congress of ambassadors’ where the members each seek to advance a distinct interest and pay little attention to the good of the whole. The parliamentary system would guard against that danger, the presidential would make it all too likely.
  A third, incidental advantage of the parliamentary system is that representatives will be motivated, as the members of an effective government, to vote with a view to the best outcome that looks to be achievable. They will be less tempted than in the presidential system by the prospect of striking an appealing, expressive posture – say, a posture that will give them the high moral ground – at a potential cost to the welfare of people overall (Brennan & Lomasky, 1993).
  Take an issue such as whether prostitution should be decriminalised and assume for argument’s sake that decriminalisation does make good sense: this, say, because criminalisation puts prostitutes beyond the reach of public protection and is liable to foster organised crime and generate funds for bribing the police. In which sort of system is decriminalisation more likely to occur? My response, borne out in some measure by actual differences between Australia and the United States, is: the moderately modular system. A government that operates as a corporate agent will be able to weigh the pros and cons, register the case for decriminalisation and act as a body in favour of decriminalisation: this, at any rate, in the absence of organised opposition, say by a church. But when legislation depends on individual lawmakers joining an ad hoc majority to support it, each is likely to be more intent on not attracting a name for favouring prostitution than on seeing that the common interest is well served overall.

• 9 Conclusion

  Throughout his work, including his work on criminal justice, John Braithwaite has always argued for a strong but non-dominating polity: a state capable of realising the sort of good he champions without threatening to impose on citizens in the manner of a master. I hope that the thoughts aired in this short tribute will be congenial to him, holding out the prospect of keeping government modular and unthreatening while enabling it to get things done. I offer them, for what they are worth, as a tribute to a good friend, a great person and an outstanding scholar. I count myself blessed that I have been able to work and think – and to live and make merry – within his circle.

References

• Arena, V. (2012). Libertas and the practice of politics in the late Roman Republic. Cambridge: Cambridge University Press.

• Ayres, I. & Braithwaite, J. (1992). Responsive regulation. Oxford: Oxford University Press.

• Braithwaite, J. (1997). On speaking softly and carrying big sticks: neglected dimensions of a republican separation of powers. University of Toronto Law Journal, 47, 305-361. doi: 10.2307/825973.

• Braithwaite, J. (2000). The new regulatory state and the transformation of criminology. British Journal of Criminology, 40, 222-238. doi: 10.1093/bjc/40.2.222.

• Braithwaite, J., Charlesworth, H. & Soares, A. (2012). Networked governance of freedom and tyranny: peace in East Timor. Canberra: ANU Press.

• Braithwaite, J. & Drahos, P. (2000). Global business regulation. Cambridge: Cambridge University Press.

• Braithwaite, J. & Parker, C. (1999). Restorative justice is republican justice. In G. Bazemore & L. Walgrave (eds.), Restorative juvenile justice (pp. 103-125). New York: Willow Tree Press.

• Braithwaite, J. & Pettit, P. (1990). Not just deserts: a republican theory of criminal justice. Oxford: Oxford University Press.

• Braithwaite, J. & Pettit, P. (2000). Republicanism and restorative justice: an explanatory and normative connection. In J. Braithwaite & H. Strang (eds.), Restorative justice: from philosophy to practice (pp. 145-163). Burlington: Ashgate.

• Brennan, G. & Lomasky, L. (1993). Democracy and decision: the pure theory of electoral preference. Oxford: Oxford University Press.

• Harrington, J. (1992). The Commonwealth of Oceana and a system of politics. Cambridge: Cambridge University Press.

• Lieberman, D. (2006). The mixed constitution and the common law. In M. Goldie & R. Wokler (eds.), The Cambridge history of eighteenth-century political thought (pp. 317-346). Cambridge: Cambridge University Press.

• Linz, J.J. & Valenzuela, A. (eds.). (1994). The failure of presidential democracy. Baltimore: Johns Hopkins University Press.

• Lovett, F. & Pettit, P. (2009). Neo-republicanism: a normative and institutional research program. Annual Review of Political Science, 12, 18-29. doi: 10.1146/annurev.polisci.12.040907.120952.

• Machiavelli, N. (1997). Discourses on Livy. Oxford: Oxford University Press.

• Madison, J., Hamilton, A. & Jay, J. (1987). The federalist papers. Harmondsworth: Penguin.

• McGeer, V. & Pettit, P. (2015). The desirability and feasibility of restorative justice. Restorative Justice: An International Journal, 3(3), 325-341. doi: 10.1080/20504721.2015.1110885.

• Millar, F. (1998). The crowd in Rome in the Late Republic. Ann Arbor: University of Michigan Press.

• Montesquieu, C.d.S. (1989). The spirit of the laws. Cambridge: Cambridge University Press.

• Ober, J. (2008). The original meaning of ‘Democracy’. Constellations, 15, 3-9. doi: 10.1111/j.1467-8675.2008.00471.x.

• Pettit, P. (1997). Republicanism: a theory of freedom and government. Oxford: Oxford University Press.

• Polybius (2011). The histories (vol. 3). Cambridge: Harvard University Press.

• Skinner, Q. (1998). Liberty before liberalism. Cambridge: Cambridge University Press.

• Tucker, P. (2018). Unelected power: the quest for legitimacy in central banking and the regulatory state. Princeton: Princeton University Press.

• Waley, D. (1988). The Italian city-republics (3rd ed.). London: Longman.

Noten

• * Contact author: ppettit@princeton.edu.

• 1 In this brief introduction to republican thought, I draw on my 1990 book with John and my later study, entitled Republicanism: a theory of freedom and government (Pettit, 1997). For a complementary account published at the same time by someone whose work we both drew upon, see Skinner (1998). For a more up-to-date overview of republican research, see Lovett and Pettit (2009).

• 2 A trope often used to characterise the mixed constitution was that it involves a compromise between the three classical models of democracy, aristocracy and monarchy. But this was not taken literally at any point, even by Polybius: thus, for him, the consuls represented monarchical power, even though there were always two, not one, and they held office only for a year.

• 3 The notion of democracy in the Greek world was that of a system in which the people or demos had a great deal of power or kratos (Ober, 2008). It became associated with a voting system only in the work of modern absolutists – enthusiasts for a powerful monarchy – like Jean Bodin in France and Thomas Hobbes in England; they cast democracy as a participatory system in which every full citizen votes on every public issue. Because democracy in that sense was infeasible by most lights – Bodin and Hobbes would have welcomed this – no one in the modern republican tradition would have embraced measures described as democratic; it was only in the mid-nineteenth century, that the word regained a positive valence, becoming associated with representative systems.

• 4 Although private members may have the right to propose bills in parliament, the government can almost always ensure that they do not get a hearing or that their bills do not get accepted.