EPT 6_4 6/29/07 12:30 PM Page 465 article EJPT Size and Virtue Francisco Herreros European Journal of Political Theory Unidad de Políticas Comparadas © SAGE Publications Ltd, Los Angeles, London, New Delhi and Singapore issn 1474-8851, 6(4) 465–484 [DOI: 10.1177/1474885107080651] a b s t r a c t : The importance of the size of a political community for the development of civic virtue has usually been related to the advantages of small size in the possibility of direct democracy and the fulfilment of the classical ideal of freedom as govern and being governed by turn. While these are important variables for the development of civic virtue, in this article it is argued that small size also matters because it allows the development of civic virtue by a reputation-building mechanism. The correlate of this argument is that as the political community grows in size, this mechanism turns increasingly unfeasible. However, the article also claims that certain institutional devices for the spread of information about people’s preferences can help the development of civic virtue even in big republics. This argument is illustrated with the example of the Roman censorship, an institution that flourished during the Roman Republican period. k e y w o r d s : assurance game, civic virtue, collective action, prisoners’ dilemma, republicanism, size Introduction The notion of civic virtue is central to the set of ideas that make up the corpus of republicanism. Civic virtue has been defined as ‘the disposition to further public over private good in action and deliberation’,1 and is considered by various republican authors as a necessary quality for both leaders and ordinary citizens in order to attain a well-ordered republic devoted to the common good and free of corruption. Since the rise of liberalism and the progressive demise of the classical republican thought, the notion of civic virtue has been increasingly considered a thing of the past, an old-fashioned concept too demanding for the modern world. Most famously, this idea was advanced by Benjamin Constant, who considered that the abolition of slavery deprived the free population of the leisure necessary to take care of public affairs in the scale required by the ancient notion of civic virtue.2 A Contact address: Francisco Herreros, Unidad de Políticas Comparadas/CSIC, Alfonso XII 18 5º, Madrid 28014 Email: herreros@ceacs.march.es 465 EPT 6_4 6/29/07 12:30 PM Page 466 European Journal of Political Theory 6(4) similar argument had been previously advanced by Montesquieu: the attachment of modern men to private interests, especially those related to commerce, was incompatible with the display of the same degree of self-sacrifice as the ancients.3 More recent works consider, for example, that ‘public virtue asks too great an abstraction from self . . . to have any chance of being successfully cultivated’.4 These pessimistic assessments of the plausibility of civic virtue in the modern world are, most likely, based in an extreme version of the idea of public virtue, a version that requires absolute sacrifice of private interest on behalf of the common good. Less extreme versions require just occasional sacrifices of the citizens’ private interests to help achieve the public interest, and they would be probably more suitable for the modern world.5 In this article I will argue that the problems of the plausibility and exact contents of the notion of civic virtue in the modern world are not just related to the impossibility of devoting as much time to the public affairs as the citizens of ancient republics as Athens did but is also a problem of size. I will argue that while in small republics civic virtue could be developed through reputation building, in bigger republics the functioning of that mechanism was much more problematic and, from a certain turning point, the size problem made it unfeasible. However, I will also argue that certain institutions developed in the ancient world, especially the Roman censorship, could be understood as a means to adapt the reputation mechanism to a larger republic. While the excessive intrusiveness of this institution in people’s private lives makes its automatic translation to the modern world undesirable, some of its features offer some keys for the designing of institutions for the promotion of civic-spirited behaviour. The article will be structured as follows. In the next section, I will develop the reputation mechanism and the problems it poses in large republics. In the third section, I will argue that, under certain circumstances, the reputation mechanism could still work in large republics. I will illustrate this idea with the example of the Roman censorship. The Development of Civic Virtue through Reputation-Building in Small Republics 466 As we have seen, civic virtue entails a certain amount of sacrifice of private interest in behalf of the common good of the republic. The amount of this selfsacrifice varies among authors. According to Montesquieu, civic virtue is a form of self-renouncing, self-sacrifice of private interest in behalf of the interest of the country.6 According to other authors, although civic virtue requires the equalizing of the private interest with the common good, this is a rational decision, at least in the long run. Rousseau, for example, considered the social contract as a rational decision to the extent that submitting to the general will defines a formal common interest that is a necessary condition for pursuing private interests.7 If we consider civic virtue as a form of cooperative preferences (whether conditional EPT 6_4 6/29/07 12:30 PM Page 467 Herreros: Size and Virtue or unconditional), it could also be reasonable to consider that there are some benefits in mutual cooperation that cannot be reached if each citizen behaves as a free-rider.8 This idea is captured, among others, by Tocqueville, when he says that obedience to the law in a democratic republic is rational because social cooperation is useful to each individual,9 and by Machiavelli, when he claims that the citizens love the republic and are capable of virtuous deeds because they realize that the republic is the foundation of their liberty, security and prosperity.10 Other interpretations of civic virtue consider it a condition for human happiness. This means that the virtuous citizen, who acts according to the public interest, is also happy: being virtuous is a precondition to being happy and free. According to Aristotle, for example, the happiness of each man is the same as the happiness of the polis and happiness is the exercise of virtue. The participation of citizens in the public life of the polis is the only way to update the rational nature of man and this, in turn, is the adequate framework for happiness.11 Finally, there are other interpretations of civic virtue that, while acknowledging that civic virtue entails a certain amount of sacrifice for the common good, consider nonetheless that it does not amount to complete disregarding of one’s private interest. We can include among these interpretations Tocqueville’s concept of ‘enlightened selfinterest’: small sacrifices of private interests, self-control, plus a certain amount of far-sightedness, given that small sacrifices in the short run on behalf of social cooperation can lead to higher individual benefits in the long run.12 The idea that public virtue entails only small sacrifices and is compatible with the pursuing of private interests is, in fact, coherent with the actual practice in ancient republics like Athens, where citizens had also a private sphere and where political activities only took a small fraction of the citizens’ time.13 How can this civic virtue be promoted? That is, how to create citizens who in their preferences take into account the common good of the political community? One solution is promoting civic virtue through education. This idea was defended by many authors in the republican tradition, among others, by civic humanists like Latini,14 by Rousseau15 and by American revolutionaries like John Adams16 and Thomas Jefferson.17 Another possible solution is the use of coercion: forcing people to behave in a virtuous manner. That is, using the punitive powers of the republic to avoid people acting as free-riders. This coercive solution is advocated, for example, by Machiavelli,18 Algernon Sidney19 and Rousseau.20 It implies the application of legal sanctions to corrupt or non-virtuous behaviour, as an incentive for people to behave in a virtuous manner. Another form of selective incentive for promoting virtue, this time positive, was the one or one-and-a-half drakmas given to the first 6000 participants in the Athenian assembly in the age of Demosthenes.21 There is, however, another form of promoting virtuous preferences beyond education and positive and negative selective incentives. This form is related to the size of the community. I will argue in the rest of this section that civic virtue can be developed in small republics through the building of reputation. However, 467 EPT 6_4 6/29/07 12:30 PM Page 468 European Journal of Political Theory 6(4) 468 I will also argue that, as the size of the republic increases, the development of civic virtue through the building of reputation becomes more and more unfeasible. Of course, the problem of size as an obstacle to the development of civic virtue, direct democracy or, in other terms, the development of an idea of liberty based on the notion that each citizen governs and is governed by turn, has been advanced by many authors. The idea is that ancient democracies were small in size and this allowed legislative power to be exercised by all citizens in the assembly but, in modern nation states, high numbers of citizens make representation inevitable. For the ancient Greeks, the polity should be small in territory and population: this enhanced the opportunities for participation in and control of the government.22 Rousseau famously considered that the state had to be neither too big to be well governed, nor too small to be self-sufficient.23 Among the American revolutionaries, Federalists like Madison, as well as most Antifederalists, considered that democracy was not possible in big republics. The Antifederalist Brutus, for example, claimed that Greece and Rome remained free while their territories were small. When they conquered new territories, they turned into tyrannies.24 However, the idea here is not that in ancient republics a small population and territory promoted civic virtue because they allowed direct participation in the legislative process as well as the possibility of governing and being governed by turn. This was probably crucial for the development of civic virtue in ancient republics like Athens, but size could be important for civic virtue for another reason: small size allowed the development of reputation of public-spiritedness and this, in turn, allowed widespread cooperation in the political community. In order to make this argument clear, we should first return to our virtuous citizen. He has to decide whether to cooperate or not in his political community. In this sense, the political community can be seen as a cooperative enterprise, where citizens can decide to cooperate or to be free-riders. According to Dagger, the idea that the political community is a cooperative association is an implication of Rousseau’s concept of the general will.25 This same idea is implicit in other republican authors. Cicero, for example, considered that the republic is an association of people under the same law.26 Thomas Paine claimed explicitly that the political community is a national association constituted for the attainment of the common good.27 Finally, according to Tocqueville, the political community allows people to access the fruits of social cooperation,28 implying that the political community was a cooperative enterprise. Therefore, we could say that for republican authors the common good is something that can only be attained through social cooperation.29 Faced with the decision of whether to cooperate or not in this cooperative enterprise, the preferences of the virtuous citizen can be of at least two types: he can be an unconditional or a conditional cooperator. The first preference of an unconditional cooperator is always to cooperate, regardless of his beliefs about other people’s preferences. If cooperation is, for example, an unconditional moral EPT 6_4 6/29/07 12:30 PM Page 469 Herreros: Size and Virtue duty, then a citizen could cooperate for moral reasons, regardless of the consequences of his action. We can think of a conditional cooperator, following Dagger, as an agent whose preferences are those of an assurance game.30 In an assurance game, the first preference of all the players is to cooperate. However, if they think that the other players are not going to reciprocate, they will prefer not to cooperate. They want mutual cooperation, but they do not want to be cheated. Whereas an unconditional cooperator will always cooperate, a conditional cooperator will cooperate as long as the other players also cooperate. These preferences are different to those of a prisoner’s dilemma, where the dominant strategy for each of the players is not to cooperate, regardless of the other player’s action. If the political community is conformed by unconditional cooperators, there will not be a problem of collective action. The solution will always be cooperative. For example, consider a game between virtuous citizens in the sense of Aristotle, citizens whose happiness depends on the achievement of a cooperative equilibrium. In this kind of game, both players have a dominant strategy: always to cooperate. The equilibrium, therefore, will be mutual cooperation.31 However, unconditional cooperation is probably a too demanding and unrealistic version of civic virtue. Therefore, I will assume that virtuous players are instead conditional cooperators, players with preferences of an assurance game (A). To explain how the size of the community can transform people’s preferences, instead of assuming only one possible type of preferences, I will assume that there is a second type of players, with prisoner’s dilemma preferences (PD). How can small size affect the transformation of PD players into A players? I will explain it using two simple models of a collective action problem. Suppose a political community with N players that can be two types: type A and type PD. I will discuss two situations where size could matter in reaching a cooperative solution and transforming the preferences of A types into PD types. In the first situation we have a big N. It is a big republic where most people do not know each other. The game is depicted in Figure 1. In the game, there are two players. Player 1 is type A, a conditional cooperator, whereas player 2 can be two types: A or PD. An initial move by Nature determines whether player 2 is type A, with probability α∈ [0,1], or type PD with probability 1 – α. Given that the republic is big and people do not know each other, player 1 does not know for sure which is the type of player 2. This is reflected in the game by the dotted lines between the two nodes of player 1. She can only guess who her partner is and that guess is captured in probabilities α and α – 1. Each of the players can either cooperate (C) or not cooperate (D). If both cooperate, they both obtain a payoff of b, with b > 0. If just one of them cooperates, the cooperator obtains a payoff of –b and the non-cooperator a payoff of a, with a > b. The differences between the types of player 2 are denoted by parameter δ. For a type A player when player 1 decides to cooperate and she does not reciprocate, her payoffs are reduced by this parameter δ, with δ > a. This is the psychological cost a conditional cooperator suffers when she breaches an agreement. 469 EPT 6_4 6/29/07 12:30 PM Page 470 European Journal of Political Theory 6(4) N α 1–α 1 1 D C C b, b 2 2 2 D –b, a – δ C a – δ, –b D C D 0, 0 C b, b 2 D –b, a C a – δ, –b D 0, 0 Figure 1. Cooperation in a large republic A possible subgame perfect equilibrium of the game is one in which player 2 cooperates when he is an A type and does not cooperate when he is a PD type. Taking into account these strategies of player 2, player 1 will cooperate when α≥ 470 b 2b – a + δ This means that the probability of player 1 cooperating depends, to a great extent, on the initial probability she assigns to player 2 of being a type A. If she trusts enough that player 2 is a conditional cooperator, then she will cooperate. There is another possible subgame perfect equilibrium, in which both types of player 2 decide not to cooperate. In this case, player 1 will not cooperate either. This equilibrium captures the fact that player 1 is a conditional cooperator: she will cooperate as long as the other players reciprocate. If this is not the case, she prefers not to cooperate. What are the consequences of these equilibria for the transformation of the preferences of the players? In the second equilibrium, both types of player 2 obtain the same payoffs (0, the status quo ante). Player 1 also receives a 0 payoff. Therefore, as every player receives the same expected payoff, there will not be a crowding-out of types in subsequent games. Consider the first situation where type A reciprocates player 1’s trust but type PD does not reciprocate. The expected payoffs of type A are b, whereas the expected payoffs of type PD are higher, a. This means that preferences of the type A will be crowded out in successive games. The change of preferences is not EPT 6_4 6/29/07 12:30 PM Page 471 Herreros: Size and Virtue 1 D C 2 0, 0 C b, b D –b, a – δ Figure 2. Cooperation in a small republic towards more people with preferences of an assurance game but towards people with more prisoner’s dilemma preferences. The only possibility for preferences of the type PD to be crowded out is if the game is repeatedly played between the same two players. In this case, as it is well known, there is a possible strategy where mutual cooperation is achieved. This strategy is tit-for-tat (begin by cooperating and then doing what the other player has done in the previous round of the game).32 For this combination of tit-for-tat strategies to hold, it is just necessary that both players value the future enough. That is, they have to be interested in future encounters with the other player and, as a consequence, they will want to build a reputation of being cooperators. In our game, however, as people do not know each other, a PD player 2 will prefer to switch partners in successive games, looking for other victims to cheat. He will not be interested in building a reputation of being a cooperator. Can things be different if the republic is small? I think that if the republic is small enough for all people to know each other, the outcome will indeed be different. In a small republic where all people know each other, the shape of the game would be much simpler. The game in a small political community is depicted in Figure 2. In this game player 2 can also be two types: A and PD. The difference between both is captured in parameter δ. If player 2 is an A type, δ ≥ a, whereas if player 2 is a PD type, δ = 0. The crucial difference with the previous game is that player 1 knows which type player 2 is. She knows for sure if she is dealing with an A player or a PD player. If player 2 is a PD type, it is easy to see by backward induction that the subgame perfect equilibrium is D, D: both players will not cooperate. If player 2 is an A type, the equilibrium is C, C: both players will cooperate. What will be the consequences of this outcome for player 2’s preferences? Notice that in this case the expected payoffs of a PD type (0, the status quo ante) are lower than the expected payoffs of an A type (b). This means that PD preferences will be crowded out in a small republic. Being a conditional cooperator always provides 471 EPT 6_4 6/29/07 12:30 PM Page 472 European Journal of Political Theory 6(4) 472 higher expected payoffs and, therefore, citizens will be interested in building a reputation of altruism. This conclusion is consistent with one of the characteristics that some republican authors considered as an advantage of small republics: in small republics people know each other well. Ancient Greeks considered that smallness ‘made it possible for every citizen to know every other, to estimate his qualities, to understand his problems’.33 Aristotle claimed that the despot will always try to avoid people knowing each other well, because this is the basis of mutual trust and cooperation.34 Rousseau also claimed that in a big state, where most people do not know each other, virtue is ignored and vice is unpunished.35 John Stuart Mill’s idea of the open ballot is also related to this. He considered that in stable and small neighbourhoods and where everybody knew everybody, people could supervise each other’s behaviour and each citizen would care about the opinion of his neighbour. The publicity of the vote generated incentives in terms of individual reputation to decide responsibly, to feel a sort of moral obligation towards the community.36 Therefore, size matters for civic virtue. In small republics civic virtue is advanced not only because direct participation in the legislative is more plausible and each citizen has a good chance of governing and being governed by turns. These conditions are no doubt crucial for the development of civic virtue in small republics but there is also another advantage of smallness: when all the people know each other, and assuming that the population initially includes a share of conditional and unconditional cooperators, prisoner’s dilemma preferences will be crowded out. That is, all people will be conditional cooperators, because their expected payoffs in this case will be systematically higher than if their preferences are those of a prisoner’s dilemma. The correlate of this conclusion is that in big republics civic virtue is not likely to be developed, at least through the mechanism here advanced. As we have seen in the game of Figure 1, when people do not know each other, conditional cooperators will be exploited. As a consequence, conditional altruism will be crowded out. In a sense, in this scenario conditional cooperators will be wise following Machiavelli’s maxim: ‘a man that wants to be good in all situations will be ruined among many bad’.37 A possible solution, already mentioned, is to coerce people into acting according to the common good. The Antifederalists claimed, for example, that big republics (like the United States after the 1787 Constitution) could only be governed through coercion.38 ‘Coercion’ means in this context the application of legal sanctions toward people who act against the common good of the republic. The reputation mechanism in small republics can imply another sort of coercion. An individual could be forced to comply with a virtuous behaviour by fear of social sanctions such as shame or ostracism. These sanctions, however, do not imply legal coercion. The question is whether compliance with the common good in big republics can only be achieved using legal coercion, even a modicum amount of it. Is the EPT 6_4 6/29/07 12:30 PM Page 473 Herreros: Size and Virtue reputation-building mechanism implausible in big republics? In the next section I will claim that under certain circumstances virtue can be, at least theoretically, promoted through reputation-building even in big republics. I will illustrate this idea using the Roman censorship. This example will allow me to illustrate how a certain institutional mechanism for the spreading of information about other people’s preferences can reduce the effects of size on the development of civic virtue. Virtue through Reputation-Building in Big Republics: the Roman Censorship The censorship was one of the highest magistracies of the Roman Republic. It was created in 443 BC as a patrician magistracy, until 351 BC, when Gaius Marcius Rutilus was appointed the first plebeian censor, although it seems that the patrician influence in this office was predominant until it was taken over by the emperor after 22 BC.39 There were always two censors and, given the importance of the office, they were usually chosen among ex-consuls. According to Livy, the censor was the guarantor of traditional customs and social norms, took decisions about the honour and dishonour of senators and equites and about the legitimacy of public and private places, and took charge of Rome’s finances.40 It was generally regarded as the highest dignity in the state, with the exception of the dictatorship. One of the main differences of the censorship compared to the other Roman magistracies was the length of the office. Magistrates were chosen in Rome for very short periods. One of the highest magistracies under the Republic, the consulship, was, for example, held only for one year. Livy considered the short duration of the consulship the main guarantee of freedom in republican Rome.41 By contrast, censors were chosen originally for a period of five years. However, in 433 BC the dictator Mamercus Aemilius Mamercinus limited the duration of the office to 18 months.42 The duties of the censors were the reading of the Senate’s list (lectio senatus), the register of the citizens and of their property (census), the administration of the finances of the state and the management of public works and the keeping of the public morals (Regimem Morum).43 The last of those functions, the keeping of public morals, was considered by many Romans, like Livy, the main function of the censorship. Cicero also considered the concern with the mores, especially the mores of senators, the main responsibility of censors.44 This responsibility consisted in the punishment of moral faults that were not included in any law. Those moral faults included, for example, abuse of children, lack of respect towards parents, prodigality, greed, luxury, ill-treatment towards slaves, violation of oaths (as mentioned by Cicero in De Officiis)45 and, in general, any action against the traditional Roman character, ethics and habits. In all these cases, censors could apply special edicts (edicta censo- 473 EPT 6_4 6/29/07 12:30 PM Page 474 European Journal of Political Theory 6(4) 474 ria) in which they could, for example, impose taxes against luxury, exclude people from the senatorial roll or from a tribe, or they could publish a note (nota censoria) that was a signal of dishonour.46 The nota censoria was applied in many cases to immoral actions that were not illegal: that is, it was applied to actions that were not included within the reach of the positive laws of the Republic. The consequence of the nota censoria was what Cicero called ignominia.47 This ignominia was a signal of dishonour and, generally, it did not imply any other punishment, such as depriving a magistrate of his office or disqualifying a citizen to obtain a magistracy. Dishonour was, in many cases, the only punishment, as we can find in the following example of the application of the nota censoria in Livy. After the Second Punic War, the censors tried to punish the various vices that flourished after the catastrophic Roman defeat in Cannae. Those people that tried to escape from Rome after the defeat, for fear of Hannibal’s army, received a nota censoria, as did those who tried to avoid military service during the war against Hannibal. In Livy’s example, the nota censoria included various material punishments, like, in some cases, exclusion from the tribes, but, most importantly, the nota censoria gave dishonour to those who received it.48 The remarks and reproofs of the censors concerning morals were made public, and in that fact lay their great effect. As the historian of the Roman censorship Jaakko Soulahti claims, although ‘the censors were not in a position to inflict actual material punition on the person under reproach, the public shame that went with the censure was in itself a heavy punishment’.49 In this section I will argue that the censorship, especially the special edict or nota censoria that was applied by censors to those who acted against the Roman mores, was an institution that could be interpreted as a way to overcome the problem of size in the development of civic virtue. It could be seen as an institution designed to fill the knowledge gap about citizens in big republics. Rome was indeed a big republic, at least much bigger than the classical ancient democracy, Athens. It seems that in 509 BC, when the Etruscan monarchy was overthrown and the Roman Republic was founded, the population of Rome was about 130,000 men, women and children, although other estimates increase that figure to about 390,000.50 During the Republic the population rose to nearly one million adult males in the census of 70 BC, at the end of the Republican period.51 In comparison, the size of Athens during the democratic period stood at 30,000 citizens at most.52 Besides, that population referred in the census included not only those living in the city of Rome itself, but all Roman citizens, most of them living in the Italian countryside outside the city of Rome. It was therefore a big republic with a population extending over a wide area. From the point of view of the development of civic virtue through a reputation-building process, the city of Rome resembled more the situation of the game of Figure 1, where citizens ignored other people’s type, than the situation of the game of Figure 2, where citizens live in a small republic and have enough information about each other’s type. EPT 6_4 6/29/07 12:30 PM Page 475 Herreros: Size and Virtue Under these circumstances, civic virtue could indeed be promoted by a variety of measures but, most probably, not through a reputation-building mechanism, at least not in the absence of an institutional device that could transmit to all players information about other players’ type. This role could be played by the nota censoria, the censorial mark that was applied to those that acted against public morals. As we have seen, this censorial mark did not necessarily imply legal punishment. Some of the actions that could be the object of the censorial mark were not indeed illegal. They included many actions in the private life of individuals, as for example living in celibacy when a person ought to be married to provide the state with citizens and soldiers, ill-treatment of one’s wife or children, a luxurious mode of living, an extravagant expenditure of money, or cruelty towards slaves and clients. The censorial mark implied dishonour, because an infamous private conduct was made public. Every citizen could have access to the special edict that established a nota censoria against a Roman citizen. This censorial mark could be interpreted, in this sense, as an institution that provided citizens with relatively cheap information about other citizens’ type. Some pieces of information provided by a censorial mark could be directly related to the devotion of the citizen towards the common good. The example mentioned of a nota censoria in Livy’s History was clearly connected to activities contrary to the common good of the republic. Those citizens that tried to leave Rome for fear of Hannibal’s army were clearly placing their own safety before the safety of the Republic. The other examples of actions that could be objects of a censorial mark are not at first sight as directly related to activities against the common good. They are, as we have seen, activities appertaining to the citizen’s private sphere. However, they could also be connected to the idea of civic virtue. In a sense, all of those activities reveal a lack of control of one’s passions. Ill-treatment of children, wives or slaves, or an extravagant or luxurious way of life can be taken as examples of lack of self-control. And civic virtue entails, among other things, self-control of passions. According to Aristotle, for example, the virtuous man is free because there are no internal obstacles that frustrate his will.53 He considered moderation and good sense two characteristics of virtue.54 In a similar way, the Romans included prudence and temperance among the cardinal virtues.55 All these qualities were certainly related to what the Roman censorship understood as a virtuous life. The examples of misconduct listed undoubtedly meant that the citizen was devoted to the fulfilment of his private passions instead of to the attainment of the common good. The condemnation of an extravagantly luxurious way of life is a good example of this. Many republican authors, from the Italian Renaissance to some of the founders of the American Republic, systematically condemned the uncontrolled pursuit of riches as a corrupt behaviour, and advocated instead frugality and temperance. The Italian humanists, like Latini, considered the classical cardinal virtues as requisites of a life according to the common good of 475 EPT 6_4 6/29/07 12:30 PM Page 476 European Journal of Political Theory 6(4) the republic. For some of them, the pursuit of private wealth was incompatible with civic virtue. This is the case, according to Skinner, with Machiavelli and Guicciardini.56 Montesquieu considered that civic virtue implies self-control of passions, and this amounts to frugality and love of equality.57 We can find this same idea in Rousseau: the common good is incompatible with an extravagant desire for wealth.58 Finally, the condemnation of luxury as a symptom of a social illness is ritually found in the radical Whigs of 18th-century England59 and, to a certain extent, in some of the founders of the American Republic, who advocated the classical version of Roman virtues interpreted as frugality, simplicity, temperance and fortitude.60 To the extent that the information contained in a censorial mark about the private behaviour of a Roman citizen could be interpreted as a cheap signal of how much he was devoted to the common good, the structure of the game of Figure 1 could be altered. As we have seen, in that game, a citizen with preferences of an assurance game, that is, a conditional cooperator, interacts with another citizen chosen at random among a very big pool of citizens. He does not know whether the second player is also a conditional cooperator or has prisoner’s dilemma preferences. This lack of information leads assurance game preferences to be crowded out. Now the game can be structured as follows. Player 1 checks out if player 2 has been included in a nota censoria. If he has been included, the game ends. If not, player 1 offers cooperation to player 2. Both types of player 2 decide if they reciprocate with cooperation or not. 476 This allows transformation of the game in Figure 1, the game of cooperation in big republics, into something more similar to the game in Figure 2, the game of cooperation in small republics. If the institution of censorship provides reliable information about the other players’ type, the expected outcome of this game is that player 1 will only cooperate with those players that have not been included in a censorial mark. Those included in a censorial mark, therefore, will obtain payoffs systematically lower than the rest of the citizens. As a consequence, prisoner’s dilemma preferences will be crowded out. In other terms, people will try to avoid being included in a censorial mark. By being included in the nota censoria, they risk being excluded from cooperation with other players. What type of cooperation? We can interpret the cooperation offered (or not) by player 1 in the second stage of the game as cooperation for a private exchange, or cooperation in activities more directly related to the public sphere. In both cases, the inclusion in the censorial mark is costly for player 2. Consider for example a citizen who decided to abandon Rome after the Legions’ defeat in Cannae. This is clearly a type of behaviour contrary to the common good of the Republic, although probably not illegal. After the war, this citizen is included in a censorial mark that is made public. The inclusion in the censorial mark precludes him from participation in any collective initiative in the public sphere because the other participants will EPT 6_4 6/29/07 12:30 PM Page 477 Herreros: Size and Virtue not trust him. He has proved to be non-virtuous, this information is common knowledge and, therefore, he is socially ostracized. But the costs that the citizen has to bear can also be extended to the private sphere. He is perceived to be narrowly self-interested in the public sphere. And this perception could spill over to the private sphere. If player 1 sees that player 2 has been included in a censorial mark, and the information provided in the censorial mark shows that he has behaved in an egotistic way towards the political community, player 1 may think that it is risky to cooperate with player 2 in the private sphere. That is, if player 1 offers a private exchange to player 2, he risks facing the same narrowly selfinterested behaviour that player 2 showed in the public sphere. Thus, player 2 faces two types of potential costs by behaving in a non-virtuous manner, as he loses future payoffs in both public and private exchanges. Therefore, he will try to act in a virtuous manner. This interpretation of the Roman censorship, especially of the censorial mark, is similar to another example of a mechanism for reduction of free-riding in situations where there is a big number of citizens and it is not possible to know the type of each one. This mechanism, as analysed by Milgrom, North and Weingast, was the medieval law merchant.61 The Lex Mercatoria was a mechanism designed to cope with problems derived from the expansion of trade. As trading communities grew larger, it became harder within each community for merchants to monitor one another’s behaviour. The new situation allowed people to act as free-riders. The law merchant was based on the presence of a specialized actor, a ‘judge’ or ‘law merchant’, who serves as a repository of information. Each party could visit the judge prior to finalizing a contract. At that time, the parties could ask the judge for the records of previous judgements about the other player. Knowing the other player’s trading history, each could decide whether to finish a contract with him or not. The problem of an expanding community of merchants was, in a way, similar to the problems of an expanded political community. Both situations allowed people to cheat on each other, to behave as free-riders. Both the law merchant in the medieval fairs, like the Champagne Fairs analysed by Milgrom, North and Weingast, and the Roman censorship were institutional mechanism to cope with the gap of information that the players experienced. Whereas the Lex Mercatoria promoted a reputation of honesty in private exchanges, the institution of censorship promoted a different kind of reputation, that of being civic-spirited. However, as I have argued before, the inclusion in the censorial mark could also have costs in private exchanges for the person included, because people could consider an egotistical behaviour in the public sphere as a signal of narrowly self-interested preferences also in private exchanges. The Roman censorship could be interpreted, therefore, as a mechanism that made possible the development of civic virtue through reputation-building even in big republics. The importance of censors for the maintenance of virtue in ancient Rome was acknowledged by Roman authors like Cicero and Livy, but also 477 EPT 6_4 6/29/07 12:30 PM Page 478 European Journal of Political Theory 6(4) 478 by modern republican authors, like Montesquieu and Rousseau. Montesquieu considered the censors the safeguards of the Republic against all vices. He correctly distinguished between the law that punishes crimes and the role of the censors, who act against behaviour that cannot be qualified as criminal. It is interesting to note that according to Montesquieu an important feature of the role of the censors was the publicity of their punishments.62 Rousseau considered the censorship a crucial institution for the maintenance of virtue: it sustained the ancient customs, avoiding corruption and vice.63 Therefore, was the censorship an adequate mechanism for the maintenance of virtue in big republics? As a mechanism for providing relatively cheap information about the players’ type, it had to fulfil certain additional requisites in order to work properly. First, in the sequence of the game as outlined before I have assumed that player 1 does not face any transaction costs when checking if player 2 has been included or not in a censorial mark. Actually, this checking always entails certain costs, at the very least some opportunity costs. Therefore, the mechanism of the censorship works best the higher is the publicity of its procedures, especially the publicity of the censorial marks, as Montesquieu correctly saw. Second, for the institution to provide reliable information it had to function in a non-arbitrary way. People had to trust in the ability of censors to detect nonvirtuous behaviour and they had to trust in their honesty. There were indeed certain informal requisites that seemed related to the reputation for honesty of Roman censors. Cicero, for example, considered that the office demanded very great resolution and very sound judgement.64 Although, strictly speaking, it does not seem to have been an office of the first rank, it was a post of dignity and its holders were supposed to be men of repute, nearly always ex-consuls. It was regarded as the crown of a successful official career.65 This does not seem to be a legal prerogative of the office but an informal understanding of the requisites necessary for the office to work properly.66 However, according to the historian W.E. Heitland, the censorship entailed practical irresponsibility in the sense that the censors could not be tried for their actions during the term of their office. It seems, therefore, that in the exercise of their power censors were regulated solely by their own views of duty. This is the most serious problem of the institution of censorship in the Roman Republic: its lack of accountability. This lack of accountability made the good selection of the censors the most important safeguard against arbitrariness. Even if they had been selected correctly, the censors had to be reliable in the other sense: citizens had to trust in their capacity to detect non-virtuous behaviour. It seems in that sense that the censors were excellently suited to their task of watching the morals of the citizens. For the census they had to acquire very detailed information on the property and family relations of each person, and this allowed them to weight their capacity to pay taxes, their sense of the economy, and their qualities as citizens.67 However, the lack of accountability would probably affect people’s trust in the reliability of the information provided by the censors. Given that they were virtually unaccountable, citizens might think, EPT 6_4 6/29/07 12:30 PM Page 479 Herreros: Size and Virtue for example, that the censors could be tempted to act arbitrarily against personal enemies or otherwise be bribed to conceal information about certain nonvirtuous behaviour. These were serious problems in the actual working of Roman censorship that hampered its capacity to act as a mechanism for the development of civic virtue. Even if the censors were really accountable to the public, another possible objection to the role of censorship as a mechanism for the development of civic virtue through reputation-building was that the cooperative equilibrium achieved through this mechanism would not be free. That is, people would be compelled to behave virtuously by the fear of being socially ostracized if they behaved in a non-virtuous manner and this behaviour was made public by the censors. This is a fair objection. Consider it, nonetheless, from the point of view of the structure of the game as presented in this section. Assume that the information provided by the nota censoria in the first stage of the game is reliable (for example, because we have a robust mechanism for censor’s accountability). In this game we have a conditional cooperator, who cooperates freely (on the condition that the other player reciprocates) and a second player with prisoner’s dilemma preferences. If the institution of the censorship works correctly, the second player will be included in a nota censoria that reveals his type to the first player. This means that the first player will decide not to offer cooperation to the second player in the second stage of the game. In order to avoid being ostracized and lose the benefits of cooperation with other citizens, the second player will try not to be included in a nota censoria. He can of course bribe the censor in order not to be included in the nota censoria, but given that we have assumed that the censorship works correctly, his only real possibility of not being included in the nota censoria is refraining from cheating his fellow citizens. The second player cooperates out of fear of being socially sanctioned. This does not necessarily mean that he is not choosing freely to cooperate. He can still decide not to reciprocate cooperation. His sanction if he decides so will simply be that the other citizens will not trust him in subsequent encounters and he will lose the potential payoffs of future cooperation. That is, his non-cooperative decision has costs but this does not mean that it would not be a free decision. A more decisive objection to the institution of the censorship, or more concretely to its applicability in the modern world, is that it is regrettably much too intrusive in people’s private lives. This was in fact one of the criticisms directed against the Roman censors by, for example, Benjamin Constant: they had an inquisitive monitoring of personal lives.68 The role given by Rousseau to censors in a well-ordered republic in his Du Contrat social has been also criticized on that same account.69 In this sense, it is an institution incompatible with individual freedom. The historian W.E. Heitland cites as an example of the ‘arbitrary and inquisitorial nature of the censorial power’ the expulsion of an ex-consul of great distinction from the Senate roll on the ground of luxurious extravagance, after it had been found that he kept ten pounds of silver plate.70 As we have seen, the 479 EPT 6_4 6/29/07 12:30 PM Page 480 European Journal of Political Theory 6(4) intrusion of the censorship in the private lives of Roman citizens was justified because one of the features of a virtuous citizen was self-control of his own passions. Classical authors saw behaviours like adultery or luxurious extravagances as clear examples of acts on behalf of one’s self-interests. Clearly this monitoring of private behaviour is undesirable in the modern world. Nonetheless, this does not necessarily mean that the institution of the censorship does not have anything to say about how to promote civic virtue in big republics. The crucial aspects of this institution were, first, that it provided citizens with information about other people’s preferences and, second, that the publicity of that information affected those people’s reputation. An alternative possibility to the censorship as designed in Republican Rome is the publicity of only those actions with clear public content. These would exclude behaviour such as adultery, religious practices (or lack of them) or luxurious extravagances, which are part of the private life of the individual. It would include, for example, the publication of the names of people who have not paid their taxes, a clear free-rider behaviour in a cooperative enterprise like the political community. In this case, to evade one’s taxes cannot be considered a matter of the private sphere of the individual but directly related to the common good of the Republic. One possible objection to this idea is that in the modern world people are not concerned at all with a reputation of being civicminded. However, I think that there is a good case to claim that people are not interested in being perceived as narrowly self-interested. If people publicly express self-interested behaviour in the public sphere (as, for example, evading taxes), they face the risk of being also perceived as self-interested in private affairs and risk, therefore, a loss of trustworthiness in private exchanges. There is some empirical evidence that people indeed do not want to be perceived as narrowly self-interested. Take for example Fiskin and Luskin’s ‘deliberative polls’.71 These polls were conducted on a random sample of some relevant population. They are invited for a deliberative weekend, where they discuss political issues. One of their results was that people refrained in public from defending arguments that could be perceived as narrowly self-interested. They tried to defend arguments compatible with the common good. There are also many attitude surveys that conclude that people want very much not to be perceived as narrowly selfinterested.72 If this is the case, institutions that make public un-cooperative behaviour in the public sphere could certainly help the development of civic behaviour. Conclusion 480 In this article I have argued that there is a relation between size and the development of civic virtue. Traditionally, this relation has been assumed to be connected to the plausibility in small republics of direct participation and the fulfilment of the classical notion of freedom as governing and being governed by turns, and its implausibility in larger ones. While this effect of size on civic virtue is a crucial one, EPT 6_4 6/29/07 12:30 PM Page 481 Herreros: Size and Virtue my idea here has been that small size allows another possibility for the development of civic virtue, understood as being a conditional cooperation or, more technically, an agent with preferences of an assurance game instead of prisoner’s dilemma preferences. This possibility is connected to the building of reputation. I have shown that in big republics, and in the absence of institutional mechanisms for the spread of information about people’s preferences, preferences of an assurance game will likely be crowded out by prisoner’s dilemma preferences. By contrast, in small republics where there is perfect information about each others’ preferences, prisoner’s dilemma preferences will instead be crowded out. I have further argued that, even if the plausibility of a reputation-building mechanism is made increasingly unfeasible as the size of the republic grows, there is a possibility for that mechanism to hold if adequate institutions are in place. In a big republic like ancient Rome was, this mechanism was the censorship. One of the devices in the hands of the censors was the nota censoria, a mark of dishonour. Providing that certain conditions held, like publicity of the censorial marks and censors’ impartiality and competence, this mark could give citizens relatively cheap and reliable information about other players’ type. The example of the Roman censorship shows that the problem of size in the development of civic virtue can be overcome with the aid of an adequate institutional mechanism. However, it is not at all clear that this solution is free of problems. More concretely, the Roman censorship, although interesting for dismissing the idea that virtue is only possible in small republics, does raise issue of accountability and privacy. In any case, the example of the Roman censorship only demonstrates the role that adequate institutions can have in promoting civic virtue in big republics. Size matters for civic virtue but institutional creativity can also have a say in the development of virtuous preferences, even in big republics. Notes 1. Richard Dagger (1997) Civic Virtues: Rights, Citizenship and Republican Liberalism, p. 14. New York and Oxford: Oxford University Press. 2. Benjamin Constant (XXXX) ‘The Liberty of the Ancients Compared with that of the Moderns’, in Constant. Political Writings, ed. Biancamaria Fontana, pp. 314–15. Cambridge: Cambridge University Press. 3. See Maurizio Viroli (1995) For Love of Country: An Essay on Patriotism and Nationalism, p. 71. New York: Oxford University Press. Viroli (ibid. pp. 43–4) considers that the same argument was also advanced by Guicciardini. Regarding Montesquieu’s vision of civic virtue as an anachronism, see also Iseult Honohan (2002) Civic Republicanism, pp. 81–2. New York: Routledge. 4. Shelley Burtt (1993) ‘The Politics of Virtue Today: A Critique and a Proposal’, American Political Science Review 87: 360–8. See also Shelley Burtt (1990) ‘The Good Citizen’s Psyche: On the Psychology of Civic Virtue’, Polity 23: 23–38. 5. Dagger (n. 1), p. 100. See also Maurizio Viroli (1999) Repubblicanesimo, pp. 65–6. Rome: Editori Laterza. 481 EPT 6_4 6/29/07 12:30 PM Page 482 European Journal of Political Theory 6(4) 482 6. Montesquieu (1987) Del espíritu de las leyes, pp. 21, 29. Madrid: Tecnos. 7. Maurizio Viroli (1988) Jean-Jacques Rousseau and the ‘Well-Ordered Society’, pp. 124–7. Cambridge: Cambridge University Press. 8. In that sense, the conflict between the dominant strategy and the Pareto-optimal strategy in a Prisoners’ Dilemma can be applied to Rousseau’s distinction between ‘the will of all’ and the ‘general will’. See W.G. Runciman and Amartya K. Sen (1965) ‘Games, Justice and the General Will’, Mind 74: 554–62. A similar idea in Joshua Cohen (1986) ‘Reflections on Rousseau: Autonomy and Democracy’, Philosophy and Public Affairs 15(3): 275–97. 9. Alexis de Tocqueville (1995) La democracia en América, p. 62. Madrid: Alianza Editorial. 10. See Viroli (n. 3), p. 73. Machiavelli also considered that in the republics the common good is easier to attain than in other political regimes, because the common good is the good of the majority of the population. Therefore, according to Machiavelli, for the majority of the population there is not a contradiction between private interests and common good. Machiavelli (1996) Discursos sobre la primera década de Tito Livio, p. 186. Madrid: Alianza Editorial. 11. Aristotle (1997) Política, 1324a, 1332a, tr. Julián Marías and María Araújo. Madrid: Centro de Estudios Políticos y Constitucionales. See also Antoni Domènech (1989) De la ética a la política, pp. 83–4. Barcelona: Crítica. 12. Tocqueville (n. 9), pp. 109–10. 13. Mogens Herman Hansen (1999) The Athenian Democracy in the Age of Demosthenes, pp. 62, 318. Norman, OK: University of Oklahoma Press. In this sense, Cicero is an example of a classical republicanism that did not advocate the adjuration of self-interest, but only to act as ‘reasonable human beings, pursuing our own advantage in an enlightened and moderate manner’. Neal Wood (1991) Cicero’s Social and Political Thought, p. 77. Berkeley: University of California Press. 14. Quentin Skinner (1978) The Foundations of Modern Political Thought, vol. 1, The Renaissance, p. 88. Cambridge: Cambridge University Press. 15. Honohan (n. 3), p. 95. 16. David McCullough (2002) John Adams, p. 223. New York: Touchstone. 17. See Thomas Jefferson (1984) Writings, pp. 365–73. New York: The Library of America. See also Garret Ward Sheldon (1991) The Political Philosophy of Thomas Jefferson, p. 16. Baltimore: Johns Hopkins University Press. 18. See Machiavelli (n. 10), p. 104. 19. Algernon Sidney (1996) Court Maxims, p. 196. Cambridge: Cambridge University Press. 20. Jean-Jacques Rousseau (2004) Du Contrat social, p. 186. Paris: Gallimard. 21. Hansen (n. 13), p. 150. 22. Robert A. Dahl and Edward R. Tufte (1973) Size and Democracy: The Politics of the Smaller European Democracies. Stanford, CA: Stanford University Press. 23. Rousseau (n. 20), p. 208. 24. James Madison, Alexander Hamilton and John Jay (1987) The Federalist Papers, p. 141. London: Penguin. W.B. Allen and Gordon Lloyd (eds) (2002) The Essential Antifederalist, pp. 27, 96, 104, 111, 122, 171, 181. New York: Rowman & Littlefield. 25. Dagger (n. 1), p. 89. See also James Miller (1984) Rousseau: Dreamer of Democracy, pp. 61, 182. New Haven and London: Yale University Press. 26. Cicero (2000) Sobre la república [De Re publica], tr. Álvaro D’Ors, 1. 39. Madrid: Gredos. 27. Thomas Paine (1995) Rights of Man, pp. 218, 251. Oxford: Oxford University Press. See also Thomas Paine (1995) Common Sense, p. 6. Oxford: Oxford University Press. 28. Tocqueville (n. 9), p. 62. 29. See Honohan (n. 3), p. 151. EPT 6_4 6/29/07 12:30 PM Page 483 Herreros: Size and Virtue 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. Dagger (n. 1), p. 112. Domènech (n. 11), pp. 92–3. Robert Axelrod (1984) The Evolution of Cooperation. New York: Basic Books. Dahl and Tufte (n. 22), p. 5. Aristotle (n. 11), 1313b. Rousseau (n. 20), p. 209. See Nadia Urbinati (2002) Mill on Democracy: From the Athenian Polis to Representative Government, pp. 104–22. Chicago: University of Chicago Press. Machiavelli (1992) El Príncipe, p. 83. Madrid: Alianza Editorial. Gordon S. Word (1987) The Creation of the American Republic. 1776–1787, p. 500. Williamsburg, VA: University of North Carolina Press. See Robert Vincent Cram (1940) ‘The Roman Censors’, Harvard Studies in Classical Philology 51: 71–110. Mommsen also considered the censorship as a mostly patrician magistracy. See Theodor Mommsen (1983) Historia de Roma, book 3, pp. 336–337. Madrid: Turner; [1856]. The same idea in Arthur Rosenberg (1926) Historia de la República Romana, p. 55. Madrid: Revista de Occidente. Livy (2001) Historia de Roma desde su fundación [Ab Urbe Condita], tr. Jose Antonio Villar Vidal, 4. 8. 1–3. Madrid: Gredos. Ibid. 2. 1. 7–8. Ibid. 4. 24. W.E. Heitland (1909) The Roman Republic, pp. 87–8. New York: Greenwood Press. S.I. Kovaliov (1973) Historia de Roma, p. 125. Madrid: Akal. Alan E. Astin (1985) ‘Cicero and the Censorship’, Classical Philology 80(3): 233–9. Cicero (2001) Sobre los deberes [De Officiis], tr. José Guillén Cabañero, 3.111. Madrid: Alianza Editorial. Kovaliov (n. 43), p. 125. Cicero (n. 26), 4. 6. Livy (n. 40), 24. 18. 2–10. Jaakko Suolahti (1963) The Roman Censors: A Study on Social Structure, p. 49. Helsinki: Suomalainen Tiedeakatemia. Lorne H. Ward (1990) ‘Roman Population, Territory, Tribe, City and Army Size from the Republic Founding to the Veientane War, 509 BC–400 BC’, American Journal of Philology 111(1): 5–39. Elio Lo Cascio (1994) ‘The Size of the Roman Population: Beloch and the Meaning of the Augustan Census Figures’, Journal of Roman Studies 84: 23–40. Hansen (n. 13), p. 60. Of course, the differences between Republican Rome and democratic Athens were not only those of size. The degree of political participation of the demos was much higher in Athens than in Rome. Rome was never considered as a democracy at all. Polybius considered it the major example of a mixed constitution, but it would probably be more accurate to consider republican Rome as a type of oligarchy. However, my point here is just that, being a much larger political community than Athens, the problems associated with the reputation mechanism for the building of civic virtue should be logically much more severe in Rome than in Athens. Domènech (n. 11), p. 88. Aristotle (2000) Retórica, 1366b. Madrid: Gredos. Cicero (n. 45), p. 66. Skinner (n. 14), pp. 42–7, 163. See also J.G.A. Pocock (1975) The Machiavellian Moment, p. 135. Princeton: Princeton University Press. Felix Gilbert (1965) Machiavelli and Guicciardini. Politics and History in Sixteenth Century Florence. Princeton: Princeton University Press. 483 EPT 6_4 6/29/07 12:30 PM Page 484 European Journal of Political Theory 6(4) 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 484 Montesquieu (n. 6), p. 33. See Viroli (n. 7), p. 161. Word (n. 38), p. 52. See Carl J. Richards (1994) The Founders and the Classics. Greece, Rome and the American Enlightenment, p. 185. Cambridge, MA: Harvard University Press. Paul R. Milgrom, Douglass C. North and Barry R. Weingast (1990) ‘The Role of Institutions in the Revival of Trade: The Law Merchant, Private Judges, and the Champagne Fairs’, Economics and Politics 2(1): 1–23. Montesquieu (n. 6), p. 18. Rousseau (n. 20). Astin (n. 44), p. 235. Heitland (n. 43), p. 88. Ibid. p. 328. Suolahti (n. 49), pp. 48–9. Constant (n. 2), p. 311. See Viroli (n. 7), p. 206. Heitland (n. 43), p. 121. See James Fishkin and Robert C. Luskin (2000) ‘The Quest for Deliberative Democracy’, in Michael Saward (ed.) Democratic Innovation. Deliberation, Representation and Association, pp. 17–28. London: Routledge. For example, there is extensive empirical material in the economic voting literature that does not find evidence for the assumption of the theory that voters are primarily guided by self-interest. See Leif Lewin (1991) Self-Interest and Public Interest in Western Politics. Oxford: Oxford University Press.