Democracy is necessarily and inevitably party government.
H. Kelsen, Vom Wesen und Wert der Democratie[In ancient Athens] There were no parties in anything like the modern sense, either among the politicians or the public.
Jones, Athenian Democracy1. Introduction
Recent research in institutional economics has augmented our understanding of the famous institutions of the direct democracy of Classical Athens (508–322 bce). Equally, studying the Athenian institutions has provided valuable insights into the use and limitations of economic methodology.Footnote 1 However, the literature has largely ignored a fundamental, and from a modern perspective, most peculiar characteristic of Athenian democratic politics, the absence of political parties. A political party is a group organized for electing candidates to office and for promoting a particular set of political principles.Footnote 2 Organizing political activity through parties solves a number of information and coordination problems in passing legislation. Contrary to representative government, in ancient Athens the central process of democracy was ‘the direct participation of ordinary people in collective self-governance’ (Ober, Reference Ober2008a: 70). What explains the absence of political parties from ancient Athens? How did the Athenian democracy resolve the information and political accountability problems inherent in collective choice without party intermediation? The present study investigates these questions.
Political parties are simultaneously organizations (that is, players within a given institutional setting) and institutions (that is, they impose rules of behaviour on members and voters alike).Footnote 3 They emerge endogenously when they promote the self-interests of rational political decision makers, politicians, policy demanders and voters. This implies that the wider institutions of collective decision making – the right to propose policy, to occupy public office and to vote – and the voting rule affect incentives to form parties. Equally, parties operate internal rules for membership, finance, election of leadership and adopting manifestos. The study focuses on an institutional explanation of the absence of parties from Athenian direct democracy.
It may be argued that the concepts of modern parties could not apply to ancient Athens because it was a small-size polity where citizens voted directly on policies. Attica is about 3,800 square kilometres, and in the 4th century the Athenian male-only citizenry numbered 30,000 (Hansen, Reference Hansen1999). Had the enfranchised citizens numbered in the hundreds of thousands, direct democracy would have been infeasible. There is an element of truth in this because in large populations spread over wide areas, debating and voting in open meetings is almost impossible. However, an explanation based on small size cannot content the theorist, as it is couched in casual observation. Athenian direct democracy was not a haphazard gathering of crowds; it had established complex collective choice mechanisms, and large numbers of citizens occupied public posts performing delegated responsibilities. It embodied a rich array of institutions whose contribution ought to be accounted for in a full understanding of the absence of parties. Athenian political leaders competed against each other to win voter support for their proposed polices. But they created neither durable associations among themselves to facilitate passage of legislation, nor formal groups to organize mass followership.
The paper proceeds as follows. To see what insights contemporary literature offers, section 2 reviews studies on the origins and reasons of existence of modern political parties. Section 3 summarizes the institutions of Athenian direct democracy and illustrates the absence of formal political parties. Section 4 investigates explanations for the absence of parties. First, it discusses views of ancient historians that citizens fearing factionalism held common preferences. Accepting however the competitive nature of Athenian politics, it then investigates the causal effects of the idiosyncratic beginnings of the Athenian democracy, the irrelevance of public office for proposing legislation, appointment to public office by lot, and direct voting on single-dimension issues. Section 5 discusses how citizens became politically informed without parties, and looks at political accountability through the courts in the absence of elections for party candidates. Section 6 concludes.
2. The rationale for political parties
Boix (Reference Boix, Boix and Stokes2007) distinguishes between two broad branches of the enormous literature on political parties: (1) sociological and political history research, which investigates party origins, types, dependence on socio-economic divisions and electoral rules (plurality or proportional representation); and (2) rational choice and especially spatial decision-making and game-theoretic models, which inquire into the reasons for the existence of political parties, the influence of the electoral system on the number of parties contesting elections, and whether the policy positions of different parties converge or diverge. The two approaches are complementary. To make sense of the history of political divisions, parties and election outcomes, an analytical framework is required. Equally, without referring to historical experiences, the game-theoretic analysis cannot explain the emergence of actual parties or the policies they advocate.Footnote 4
Modern parliaments, with the power to check the authority of kings, appeared in England in the late 17th century (evolving from the concessions made by the king to the barons described in the Magna Carta of 1215) and spread slowly to Europe and America over the 18th century. Gradually, such parliaments acquired full legislative control.Footnote 5 Elite parties originated in the early days of parliaments. Their members were local notables selected by a restricted franchise (comprising only propertied classes) with strong local ties, who did not need the support of a national organization to win election. In parliament, they formed groups sharing common interests and voted according to their consciences rather than along party lines.
Mass parties appeared in the late 19th and early 20th centuries as a result of the extension of voting rights to previously disenfranchised, poorer classes of the population and the increasing role of the central state in regulating the industrialization process and international conflicts. As parliamentary majorities grew pivotal for passing legislation, it became vital for the previously loose parliamentary groups with broadly similar policy preferences to organize into coherent, disciplined and stable units. Voters chose party platforms and voted for party-nominated candidates, who voted in parliament according to party lines.Footnote 6
Spatial decision models of parties start from the premise that voters vote instrumentally, that is, for the policy expected to maximize their net benefits. In this setting, the formation of parties solves coordination problems between legislators in parliament and asymmetric information problems afflicting voters in electoral contests (since voters may remain rationally ignorant). Legislators operating independently of each other face collective action and collective choice impossibility problems (Aldrich, Reference Aldrich1995; Jackson and Moselle, Reference Jackson and Moselle2002). The collective action problem arises when each legislator votes for bills that confer upon him the highest payoff (and rejects all others), which results in rejecting all bills for lack of majority. But when a majority of legislators forms a party binding them to cooperate, they can pass all their favourite bills, increasing their payoffs. The impossibility problem arises when legislators face multidimensional policy choices, or they have multi-peaked preferences over a single dimension. In these cases, a group of legislators forming a party can agree which policies to support, and avoid voting cycles. The platform of their preferred policies may then emerge as the winning outcome. In addition, since agreeing a coalition is likely to involve significant transaction costs, it is sensible to form a party for the long term.
Parties, as institutions that aggregate ideologically similar candidates, signal their preferences to imperfectly informed voters resolving a range of information and commitment problems. Specifically, by voting for representatives, citizens delegate decision-making powers to office holders who may be better informed and enjoy relative autonomy. This asymmetry generates principal–agent problems, where the representative-agent may renege on promised policies, and pursue its material and ideological interests to the detriment of those of the voter-principal. Party brands, whose life spans are longer than those of politicians, are valuable because they discipline candidates with short horizons to commit credibly to platforms in the long-run interests of the electorate (Alesina and Spear, Reference Alesina and Spear1988; Harrington, Reference Harrington1992), and provide low-cost information signals about the preferences of candidates (Levy, Reference Levy2004; Snyder and Ting, Reference Snyder and Ting2002). Further, politicians organized in parties exploit economies of scale in collecting information about voter preferences, publicizing manifestos and fighting elections (Aldrich, Reference Aldrich1995; Osborne and Tourky, Reference Osborne and Tourky2008).
Contrary to instrumental voting, the expressive view of voting argues that voters vote not for policy gains, but for the psychological benefits from expressing their identity and class, which are reflected by political parties (Brennan and Hamlin, Reference Brennan and Hamlin1999; Hamlin and Jennings, Reference Hamlin and Jennings2011; Hillman, Reference Hillman2010). Parties play the role of clubs offering a public good (shared ideology), for which exclusion is possible. In this light, voting for parties resolves not only the problem of information cost, but also the question of who to vote for. In a similar vein, although not using the term expressive voting, Achen and Bartels (Reference Achen and Bartels2016) demonstrate that the great majority of citizens facing pressing everyday demands pay little attention to politics and remain poorly informed. Accordingly, political behaviour and voting patterns depend on voter identity originating from attachment to social and psychological groups, making election outcomes mostly erratic reflections of the current balance of partisan loyalties.Footnote 7
Individual voters may also join parties as activists, paying dues and engaging in activities to influence policy and benefit materially from being close to elected party officials (Aldrich, Reference Aldrich1983a, Reference Aldrich1983b; Katz, Reference Katz1990).Footnote 8 Similarly, interest groups, as policy demanders, may form or back particular parties (Bawn et al., Reference Bawn, Cohen, Karol, Masket, Noel and Zaller2012; Hebert and Wagner, Reference Hebert and Wagner2018). In communities with large populations, where division of labour and knowledge are widespread and pervasive, and voters pay little attention to politics, the cost of communicating political ideas is high; this offers opportunities to interest groups and policy activists to nominate individuals who advance their policy goals, and secure rents.
In conclusion, parties, formal and durable organizations with a mass followership, resolve three (overlapping) problems: (1) coordination problems between individual politicians and legislators in parliament; (2) information problems afflicting voter decisions in elections for candidates for office; (3) pursuit of policy and ideological demands by voters. It follows that parties did not emerge in ancient Athens because the Athenians had devised alternative solutions to those problems.
3. The direct democracy of Athens
Collective decision making and appointment to office
The direct democracy of ancient Athens rose with the 508/507 reforms of Cleisthenes, an aristocrat who, after allying himself with the common people (demos), prevailed against his aristocratic opponents. His reforms confirmed citizenship rightsFootnote 9 for all adult males residing in Attica, and allocated Athenians by lottery into ten artificial ‘tribes’ (phylae) with each tribe containing a cross-section of citizens, resulting in an all-inclusive organization.Footnote 10
The Assembly of the Demos was the principal decision-making body. It debated and voted on public policy issues, war and peace, public finance, foreign policy, infrastructure projects, festivals and public honours. All Athenian males after the age of 20 had the right to attend and address the Assembly. Attendance was voluntary with the quorum of 6,000 out of a 4th-century male population of 30,000. From ten annual meetings in the 5th century, the Assembly met 40 times a year in the second half of the 4th century.Footnote 11 They voted by a show of hands (with the exception of votes for ostracism and citizenship grants where votes were recorded), and decisions were taken by simple majority. In sharp contrast to modern representative government, voting for candidates for public office was a relatively small part of the direct Athenian democracy. From 501, the Assembly elected annually, by show of hands, the Ten Generals (Strategoi), one from each tribe, who served as commanders of the army and navy and performed related domestic and external policy duties. Hansen (Reference Hansen1999: 235) writes:
A candidate from tribe I was proposed, and the people voted for or against him. The first candidate to get a majority was elected unless a named opponent to him was proposed, in which case the vote was a vote between the two of them. When no more candidates were proposed, the people proceeded to the next tribe, and so on.
‘Hands were never counted. The majority was assessed by the nine proedroi [chairmen] who made their decision on a rough estimate’ (ibid: 332). But from 440, at least one general was chosen from all tribes, so that at most nine were selected from different tribes. Thus, contrary to contemporary elected politicians who represent geographical constituencies, the Athenian generals were not representing their tribes.
Nor did the generals act like a modern cabinet of ministers, as different generals could argue for different policies. For example, in 415 Alcibiades and Nicias, although supporting different positions regarding the fight against Sparta during the Peloponnesian War, were appointed co-leaders of the Sicilian campaign. In the 5th century, the political leaders active in the Assembly were also elected generals, but that was not the case in the 4th century. The reason according to Hansen (Reference Hansen1999) was increased specialization; to deal with matters of war, military commanders had to be highly experienced military men, while orators focused on acquiring the rhetorical skills to address the Assembly.Footnote 12 In the mid-4th century, new four-year-term elected civilian offices were introduced, the treasurer of the military fund, and the board responsible for festival money (theorika).
The Athenians appointed by lot a large number of public office holders including the Council of the Five Hundred (Boule), and the panel of 6,000 jurors, who could serve in the popular courts (Heliaia). Appointment to public office by lot (sortition) from the eligible citizens was the hallmark of the ancient direct democracy, the true meaning of the rule of demos (see Tridimas, Reference Tridimas2012 and the references therein). Random selection offered equal opportunities to all citizens to hold public office, while election advantaged the rich who could afford training in public speaking and pay for liturgies (private finance of public services) that would make them popular. According to Aristotle: ‘It is thought to be democratic for the offices to be assigned by lot, for them to be elected oligarchic, and democratic for them not to have a property-qualification, oligarchic to have one’ (Politics, Book 4, 1294b, cited in Jones, Reference Jones1958). This conception of democracy is altogether different from that underpinning modern representative government, where the right to vote offers each citizen an equal opportunity to consent to what the government decides (Manin, Reference Manin1997).
The Council of the Five Hundred, set up by the reforms of Cleisthenes in 507, met at least 250 days a year, prepared the agenda for the Assembly of the Demos, oversaw the implementation of the measures passed by the Assembly and carried out the day-to-day administration of the polis. The Council brought an issue to the Assembly, either for ratification of a specific decree already passed by its members or as an open issue for discussion and vote by the Assembly. The courts were responsible for trying civil, penal and political cases, and reviewing the decisions of the Assembly. They also held to account all magistrates before taking office checking their eligibility, both during their service and after completing their term. Courts met about 200 days a year. On each court day, 2,000 jurors out of the 6,000 panel were selected by lot. After swearing the relevant oath, another lot allocated jurors to various trials. A normal jury numbered 501 with 201 as a minimum, while politically important cases could use thousands of jurors. Unlike Assembly attendees, jurors voted secretly without discussing the case and decided by simple majority.
Further, the Athenians appointed annually by lot another 600 magistrates to serve on various boards administering the city-state, religious affairs, inspection of markets and exchanges, roads and buildings.
The democratic institutions were not established whole cloth: the demos assumed control of policy making gradually in a number of steps over a long period. In 594, before the reforms of Cleisthenes, the laws of Solon enacted a variety of political and economic rights, which reduced the hold on power of a birth aristocracy. The democratic reforms culminated in 462, when the Council of Areopagus, which before the establishment of democracy oversaw laws and magistrates and conducted trials, was reduced to a homicide court for Athenian citizens. The democracy was interrupted twice. First, after a coup in 411 an oligarchy came to power where only 5,000 Athenians retained full citizenship rights, but the navy restored the democracy four months later in 410. In 404 after Athens’ defeat in the Peloponnesian War, a cruel commission of the so-called Thirty Tyrants established an oligarchy. The democracy was reinstated in 403 after a violent confrontation with the oligarchs. Various institutional reforms were then introduced. Most significantly, the assembly no longer passed laws describing ‘general norms without limit of duration’. This responsibility was granted to special boards of lawmakers (nomothetai) chosen by lot from the same panel of 6,000 jurors of the Courts (see Lyttkens et al., Reference Lyttkens, Tridimas and Lindgren2018). The Assembly retained the power to vote for decrees and decide foreign policy.
Elected generals, councillors, magistrates and jurors of the court served annual terms and received a fee for their days of service, to compensate for the opportunity cost of their time. With 30 years the minimum age for office, Hansen (Reference Hansen1999) estimates that during the 4th century the pool of eligible candidates was 20,000 citizens. No citizen could serve more than one term in his lifetime as a magistrate in the same office, with the exceptions of the Council where one could serve twice but not in successive years (but no term limits were placed on the generalship). Annual terms of service resulted in substantial rotation. Sortition implied that any citizen might hold office; rotation implied that every citizen might hold office at some time. Members of the Council, court jurors and the various administrative magistrates were ordinary citizens, amateurs without any specific training for the functions allocated to them.Footnote 13 Their tasks were designed for non-specialists ‘with reasonable intelligence and motivation’ (ibid: 244) who could then perform them successfully. On the other hand, elections took place for appointment to posts requiring leadership in military or financial matters, so that citizens could choose those inspiring confidence for their abilities. For Ober (Reference Ober2008b: 98), the Athenian democracy ‘was predicated not on the legitimacy of elected leaders but on the assumption that value is added in political decision making via the aggregation of technical and social knowledge that is widely distributed within the citizenry itself.’ He credits the institutions of direct democracy with enabling diverse and ordinary citizens to choose the policies that led to economic prosperity and military strength of Athens.Footnote 14
The absence of political parties
There was no party alteration in government the way it is understood in modern democracies. Contrary to representative democracies where the government typically initiates legislation, in Athens citizens brought issues for discussion to the Council and the Assembly. Political leaders did not propose legislation in any official capacity, but as private citizens (Hansen, Reference Hansen1999). Talented orators and rich individuals who could afford training in rhetoric were the most frequent public speakers addressing the Assembly and the courts. An orator–political leader introducing a proposal could win or lose an Assembly vote, but did not win or lose political office. Annual terms and turnover of office holders removed the possibility of developing a class of ‘career politicians’ or a specialized state bureaucracy.
The political leaders, or orators, proposed policies, argued for their adoption, confronted each other in the Assembly and the courts, and aspired to receive rewards and honours as ‘protectors of the demos’. Headlam (Reference Headlam1891: 20) wrote:
It would seem that, if elections were fought on party lines, and if their result had a serious influence on the direction of the policy, we should be able to find some one or more offices which were filled by the heads of the party, who in consequence of their election would be for a period the recognised leaders of the state.
Hammond (Reference Hammond, Boardman, Hammond, Lewis and Ostwald1988) argues that it is misleading to use party labels to characterize Athenian politics because of the immediacy of contact between political leaders and people. In the Assembly, each political leader acted individually, arguing for his preferred policy, and ‘if they formed partnerships, they did so on an ad hoc basis and readily dissolved them’ (ibid., 521). Similarly, Osborne (Reference Osborne2010: 29) concludes: ‘Athens had nothing like a party system, and although scholars occasionally suspect that a member of the Council was acting as a frontman for a more prominent individual, it seems unlikely that this happened in any systematic way.’
Hansen (Reference Hansen1999: 277–288) conjectures that at any given time the number of active orators was about 20, and in the period 403–322 there were fewer than 100. He contends that the orators did have associates, followers, friends and admirers, and as a matter of necessity they occasionally collaborated with each other to advance common interests in the Assembly and the courts. These loose associations were perhaps the closest analogues to political parties. After a careful reading of the extant sources, Hansen (Reference Hansen2014) concludes that there existed small sets of political leaders but their groupings lacked formality, durability and organizational structure to be recognized as political parties. In the ancient society, exchanges were personal and localized. Political leaders were able to control small scores of voters, but political parties, in the sense of organizations with formal rules, registered members or distinct ideological labels, were absent from all settings where voting took place (the Assembly, the Council and the courts). Men with political ambitions joined other men of influence, which made Athenian political life depend on interpersonal relationships rather than ideology or party organization (Worthington, Reference Worthington2013).
Groups of aristocrats wishing to overthrow the democracy (temporarily successful in 411 and 404) were arguably the closest that Athens experienced to informal political parties. They espoused a constitutional agenda, but were far from being political parties. In the extant sources, the word hetaireia often appears to describe groups of supporters of political leaders.Footnote 15 Hansen (Reference Hansen1999: 277–287) warns that it is best to translate the word describing the group of followers, hetaireia, or hetairikon, as ‘comrades’ association’, instead of party. Forsdyke (Reference Forsdyke2005: 85) describes hetaireia as ‘a group of wealthy and powerful men, who were not necessarily related by blood’. Similarly, Brock (Reference Brock, Bagnall, Brodersen, Champion, Erskine and Huebner2013: 3196) defines hetereia as ‘informal grouping of upper-class males who engage jointly in warfare and politics and meet socially in the symposion [drinking party]’. Hansen (Reference Hansen2014) argues that hetaireia refers to oligarchs who attempted to overthrow the democracy;Footnote 16 they were neither formal nor long-lasting organizations. He concludes that such groups were like clubs based on kinship and personal friendship.
There was always the risk that speakers addressing the Assembly could be shouted down by coordinated efforts of supporters of rival political leaders. To counteract this, a law of 346/345 prescribed that in Assembly meetings all members of one tribe, selected by lot before the session, would sit at the front of the auditorium (Hansen, Reference Hansen2014, offers details), so that no group of followers of a political leader would be able to heckle or interrupt a rival orator. Further, looking specifically at the Council of Five Hundred, Hansen (ibid) points to a law of 410/9 which assigns seats to the councillors by lot. He comments (392–393):
The reason … must have been that the Athenians wanted to thwart a tendency among the councillors to seat themselves in political groups … it is worth noting that [the law] was passed in the year when democracy was restored after the regime of the 400 and the 5000 [coup of 411]. Before the restoration the Council had been controlled by a faction of oligarchically minded citizens (Thuc. 8.66.1–2) who probably sat together in the bouleuterion. The reinstated democrats would put a stop to such practices.
We conclude that even if the presence of parties is not a binary variable but one of gradation according to the level of formal organization, durability and ideological label, it is still fair to say that at the very least, Athenian political groupings that harboured ambitions to change the constitution had no formal organization, permanent structure or detailed political objectives.
4. Explaining the absence of political parties
Homogeneous citizenry and political consensus
A small body of previous work by ancient historians attributed the absence of political parties to the relative homogeneity of the Athenian citizen body. Osborne (Reference Osborne2010) maintains that from an early age the Athenians engaged in a multitude of face-to-face activities, fostering the idea that they all espoused the same principles and practices. This strengthened a sense of community among the politically active male citizens, while simultaneous exclusion from citizenship of women, resident-aliens and slaves resulted in a more homogeneous electorate.
Ober (Reference Ober2008b) argues that the Athenians shared an overriding preference to be powerful and defend against endemic existential risks from both external and internal enemies. A Greek city-state confronted a 1:3 chance of destruction from an external enemy (ibid: 82). During the Archaic Era (750–500), Athens had to cope with various invasions and faced Megara, a hostile neighbour. In 480, just before the victory in the sea battle of Salamis, the invading Persians sacked Athens. After the 404 defeat in the Peloponnesian War, Athens had to surrender its navy and demolish the defensive Long Walls. As already described, the internal enemies overthrew the democracy twice, in 411 and in 404. Ober claims that consensus in favour of Athenian power superseded all differences leading to ‘a general lack of fixed ideological commitments of the sort that sustain a system of organized political parties’ (Reference Ober2008b: 101, original emphasis). This argument is in the tradition of a fear of factions as a threat to popular sovereignty.Footnote 17 Until the emergence of the modern representative government in the 19th century, political thinkers, wrestling with the fear that one faction may prevail over another, control government and tyrannize its opponents, searched for institutions to balance factions. Contrary to the latter, North et al. (Reference North, Wallis and Weingast2009, Chapter 6) argue that political parties as impersonal organizations made up of a multitude of factions of political and economic interests signalled the arrival of modern open access to politics and economics.
Hansen (Reference Hansen2014) attributes the absence of parties to the direct democracy per se, but does not offer a causal mechanism. He cites Swiss Landsgemeinden (annual assembly of the citizens of a canton) as a modern example of direct democracy without parties. In Landsgemeinden meetings, party politics practically dissolve; voters neither sit in groups according to party affiliation, nor do they vote on party lines. He also considers the laws against group formation by political leaders and their followers as a corroborating reason for the absence of parties.
Nevertheless, even if a general consensus about strong defence prevailed, important differences regarding how to achieve it cannot be ruled out. Differences remained about taxes to finance defence. For Jones (Reference Jones1958: 131), different socio-economic interests led to different patterns of voting:
there was a broad distinction of outlook between the propertied classes and the poor. Aristotle thought that he could discern this distinction throughout Athenian political history, and he is not likely to have been entirely wrong. In domestic affairs, it is difficult to trace it. There was no overt oligarchical party … The distinction can be discerned most clearly in foreign policy, which of course involved finance. On a number of occasions, in 396 [according to the] historian Oxyrhynchus, and on Alexander's death [according to] Diodorus, we are told that the propertied classes favoured peace or appeasement, while the poor were more bellicose. Aristophanes (in Ecclesiazusae: 197Footnote 18) declares: ‘Is there talk of equipping a fleet? The poor man says, yes, but the rich citizen and the countryman say, no.’ But this merely means that people tended to vote according to their economic interests.
Economou and Kyriazis (Reference Economou and Kyriazis2016) also show that rich and poor Athenians had different policy preferences, respectively, pro-peace and pro-war, and expansive programmes of public works offering job opportunities to the poor won majority support. Nevertheless, differences in birth, wealth and policy preferences did not give rise to political parties.
The argument that there were no parties because of ‘fear of faction’ may not explain fully the absence of parties from the antagonistic environment of ancient Athens. If parties arise endogenously in competitive politics, and political competition was present in Athens, it is essential to investigate why neither political leaders nor voters saw any benefits in founding, joining or voting for political parties.
Initial conditions
Boix (Reference Boix, Boix and Stokes2007) alludes to an important asymmetry between early and later stages of electoral politics. At the initial step, there is little information about voter preferences and the stage is fluid; as a result, early entrants may be able to shape the form of the electoral competition to suit their interests. This implies that if parties were absent from the contest for popular support at the crucial juncture of the beginning of democratization of Athens, other things being equal, they were unlikely to develop endogenously.
In ancient Greece, assemblies where the rulers, members of the aristocratic elite, addressed ordinary warriors were common from the archaic times, if not earlier, although decision-making power was in the hands of the rulers. Steadily, the ordinary warriors became economically and militarily more powerful and acquired full citizenship rights. Citizenship, a most important conceptual development in ancient Greek political thought and practice, meant that all locally born free men within a city-state had equal political rights and enjoyed legal protections, regardless of wealth, birth, education or other factors, combined with obligations to serve the community. Citizens decided by simple majority. Introduced probably in the 7th century, simple majority voting was the result of new development in hoplite warfare technology (Pitsoulis, Reference Pitsoulis2011). Counting the number of men armed with spears instead of actually fighting and awarding victory to the side with the largest tally reduced significantly the cost of conflict.
The Athenian transition to democracy, like the gradual political liberalization of the West, took place in stages. We may distinguish a number of stages, namely, the reforms of Solon, the reforms of Cleisthenes, which secured the right all (male) Athenians to attend and vote in the Assembly, and finally the extension of the right to serve in public office to the poorest classes in the 460s, introduced by the reforms of Ephialtes and Pericles.Footnote 19 No political parties formed to advance those changes. As already said, Cleisthenes prevailed against his aristocratic rivals after ‘allying himself with the common people’. Furthermore, unlike representative democracies whose electorates were divided into geographical constituencies, Cleisthenes set up Athens as a single constituency to end previous political divisions based on geographical divisions. As a result, delegates were not elected to represent local interests, and neither were proto-parties formed as alliances of local political leaders in the Assembly. Path dependence was crucial: since at the initial steps of democracy the empowered demos decided policy directly, a change towards conducting policy through delegates organized in parties became unlikely. Headlam (Reference Headlam1891: 34) makes this point implicitly: ‘The prejudice against party organizations was a necessary consequence of the principle that the demos must govern.’ We proceed to explore the institutions of direct democracy that made parties superfluous.
The irrelevance of public office to proposing legislation
Any citizen who so wished (ho voulomenos), could propose a policy, and those sufficiently concerned did so.Footnote 20 Occupying public office was irrelevant to proposing legislation. The Athenian political leaders controlled policy making because they were able to argue convincingly and win the Assembly vote, not because they were voted to office. Ober (Reference Ober2008b: 164) considers political leaders who spoke to the Assembly as ‘advisers’, ‘men possessing some expert knowledge of both the matter at hand and rhetorical technique’. In Xenophon's Memorabilia Book 3.6, Socrates illustrates to a young aspiring politician that mere eloquence would not make him a successful adviser. Instead, he needed to develop expertise in the sources and sizes of the revenues and expenses of the city-state, its military and naval strength and that of its enemies, whether the annual production of home-grown wheat could feed Athens and how much more was required. Nowhere does Socrates advise him to form a party or join existing parties. This silence is deafening. With public office inconsequential for political influence, ambitious individuals who in modern times would have joined political parties to pursue their ideological causes could do so by winning Assembly votes. These observations go a long way to explain the small number of ancient references to elections, and the complete absence of references to today's ubiquitous office of prime minister or president. Nor did ‘interest groups’ need to pursue their policy interests by supporting a party. Individuals could propose their favourite measures directly to the Assembly, or ask one of the active political leaders to promote its cause.
The right of any citizen to propose a policy opened agenda setting to all. Modern collective choice theory suggests that this effective absence of monopoly in agenda setting must have led the direct democracy to produce voting outcomes in agreement with popular preferences.Footnote 21
Appointment to public office by lot
Politicians in office have significant patronage power and may appoint their supporters to public posts. However, Athenian political leaders had no authority to make appointments to public offices because the latter were allocated by lot among the eligible (and willing) citizens, which also strongly discouraged the development of clientelistic relations between politicians-as-patrons and voters-as-clients. That is, sortition enabled citizens to occupy public office and receive a fee for service without political partisanship or personal connections with political patrons. Those interested in holding public posts had no incentive to receive nomination by political parties, for affiliation to a party could not affect the probability of holding office. Nor did serving in public office offer permanent occupation as office rotation occurred annually. Similarly, since the outcome of the lot was unpredictable, it was impossible for a political party to advance its interests by appointing supporters to public posts.
The irrelevance of holding office to proposing policy, sortition and rotation in public office diminished the incentive to set up parties to control the legislative or executive arms, which in turn discouraged corruption of magistrates. However, these factors do not explain why political leaders did not form parties as formal and binding alliances to pursue common policy objectives in the Assembly.
Single-dimension voting
Since voters care about several issues, modern elections are fought over multiple policy dimensions. As already explained, in the multi-dimension setting, parties are formed to offer combinations of policies not available in their absence; this in turn may resolve problems of cycling (Aldrich, Reference Aldrich1995) and may generate policy platforms that are preferred to the Condorcet winner in pairwise comparisons (Levy, Reference Levy2004). However, in the direct democracy of Athens, Assembly voters voted on each item debated, rather than platforms bundling several issues.
Hansen (Reference Hansen1999: 156) calculates that in the 82-year period 404–322, about 3,000 Assembly meetings took place passing about 30,000 decrees. Although detailed quantitative information is lacking, Hansen argues that the most important field of action was foreign policy, including declarations of war and peace, alliances, sending of envoys, army and navy mobilization and the financing of expeditions. Other fields included public works and their financing, religion and festivals and honorary decrees passed to recognize those contributing to Athens. With such a large number of decisions, we may presume that no bundling took place. Voting was on single dimension as Assembly goers decided separately on foreign policy, finance of public projects and so on, instead of platforms combining several of them. Deciding on each separate issue instead of platforms is a crucial qualitative advantage of direct, relative to representative, democracy.
Some issues must have been uncontroversial and easily and quickly dealt with. A few more issues were simple binary choices, as for example when granting citizenship to a non-Athenian or bestowing a public honour on an individual. In those cases, the Condorcet theorem applied.Footnote 22 No parties could have emerged under these circumstances.
Other issues however must have been debated long and hard. As is well known from public choice theory, if voting is on a single dimension and preferences are single-peaked, the median voter's ideal point wins against all other policy points in pairwise voting (Mueller, Reference Mueller2003). In a single dimension, any party comprising politicians only to the left or to the right of the median cannot win against the median. When politicians to the left and to the right are members of the same party, they cannot find a position that is preferable to the median and their coalition is unstable. It follows that if a median voter equilibrium existed, the Assembly would have reached it without the intermediation of political parties.Footnote 23 Politicians forming parties would confront costs without a commensurate increase in payoffs.
More specifically, if before voting, political leaders can identify long and predominant social divisions, they might react by forming political parties that reflect such divisions. However, if the electorate is divided along several and changing issues, formation of long-lasting political parties may not be viable. When it is possible to vote on any single issue, the number of divisions among citizens is extremely large; every single issue may generate a different coalition of interests. This implies that Assembly voting on each policy item made impossible the emergence of stable, organized groups with common long-standing interests, or coherent ideology, or aims to share the spoils of power. As a result, if a median voter equilibrium existed, the Assembly would reach it without voting for political parties; in direct democracy, political parties were superfluous.
There is corroborating evidence from modern party politics that ‘unbundling’ policy issues reduces the primacy of political parties. There may be issues, like a state's EU membership, that cause deep intra-party splits across the traditional left–right spectrum and are impossible to resolve through parliamentary party politics; such issues are often resolved by calling a referendum (Matsusaka, Reference Matsusaka2005; Tierney, Reference Tierney2012; Tridimas, Reference Tridimas2007).Footnote 24 It follows that if there are issues that cause intra-party divisions, the parties may dissolve. These theoretical considerations also formalize Hansen's (Reference Hansen2014) intuition that party politics disappear in a direct democracy.
However, the validity of the median voter equilibrium rests crucially on the assumption that voter preferences are single peaked. How often this was the case in ancient Athens is a separate inquiry for future research.
5. Voter information and political accountability without political parties
Without elections for party candidates, how did the Athenian democracy solve problems of information about policy issues and accountability of politicians?
A well-informed citizen body
Direct democracy provided many opportunities for the demos to become efficiently informed about policy issues without the intermediation of political parties. Given the short terms of service, annual rotation in office and large frequency of Assembly meetings, the citizens who served in the Council, the magistracies and the courts and attended the Assembly must have been well informed and sufficiently experienced to perform the roles expected of them by the participatory democracy. Hansen (Reference Hansen1999: 249) calculates that in the 4th century, with a population of 20,000 eligible citizens, the rule that a man could serve in the Council of Five Hundred no more than twice in his life implied that ‘over a third of all citizens over eighteen, and about two thirds of all citizens over forty, became councilors, some of them twice’. This is a staggering figure of politically active citizens in comparison to representative democracies.
Ober (Reference Ober2008b) explains that without resorting to centralized coercive arrangements the Athenians solved problems of information asymmetry and coordination successfully by relying on the dispersed knowledge of different citizens. Direct democracy, with its frequent face-to-face engagements, placed the management of the state in the hands of numerous small teams as well as larger networks of citizen-amateurs who learnt basic routines and when needed could access expert knowledge. Meeting in the Assembly, Council, courts and boards, participating in tribal, municipal and state rituals and serving in the military enabled the Athenians, both formally and informally, to share information, learn from each other, accumulate knowledge and foster innovation.
The antidosis (property exchange) procedure illustrates an ingenious way to extract and verify information in the public interest: A nominee for a trierarchy liturgy (payment for outfitting and maintaining a trireme for a year) could challenge another rich citizen either to undertake the expense or to exchange property with the nominee who would then pay. The defender could accept the liturgy, exchange property or ask a court to adjudicate. In front of hundreds-strong juries, each of the disputants tried to show their true wealth to avoid paying. The objective of antidosis was to ensure that those who actually were wealthy, rather than those who seemed to be wealthy, would pay, with the burden of proof falling on the private actors rather than the state. From a modern perspective, antidosis sought efficient public good provision, voluntary finance of public goods, and budgetary balance (Kaiser, Reference Kaiser2007).
Publicity, dissemination of information to citizens at different locations and occupations, took place through a variety of media. These included oath taking, participation in public rituals and building distinct public monuments, which became low-cost information conduits. Oath taking (by soldiers, jurors and magistrates) in front of witnesses signalled a credible commitment to act for the cause (defence, justice, etc.) Rituals are manifestations of culture, the set of beliefs about how nature and society work and the related norms of behaviour. Participation in public rituals, sacrifices, athletic and artistic competitions, built common knowledge and enhanced coordination of actions. Public monuments in prominent positions and of distinctive architecture, temples, statues, and columns (stele), acted as ‘focal points’ and reminders of shared values and history. The central agora was officially designed as an open public space for citizens to congregate, talk and do business. Statutes were inscribed on prominently displayed columns for all to see. The architecture of inward-facing circles, as in theatres and other public meeting places, allowed spectators both to observe the speaker at centre-stage and make extensive eye contact with each other, thus gaining insight on the reception of speeches.Footnote 25 Without loudspeakers or paper for writing, such means undoubtedly reduced the cost of transmitting information that built common knowledge.
It follows that the Athenians were well informed about policy issues and could make up their minds, reducing inter alia reliance on political parties. This is not to idealize them or say that they were perfectly informed. As discussed, political intermediation did take place. The ability to collect and process information and offer solutions on policy issues were essential qualities for an individual to emerge as a political leader. Further, as it is clear from forensic speeches, the orators appealed to the citizens not only on proposed policies but also on valence issues, like character and personal achievements, which also indicates lack of perfect information about policy issues.
Political accountability through the courts
In modern democracies, electing or rejecting candidates for office is an effective mechanism to discipline politicians. In view of the lack of partisan elections for office, how did the Athenians hold political leaders to account? During the 5th century, ostracism, introduced by Cleisthenes, offered such a mechanism. A special Assembly vote could send a political leader to exile for ten years, removing him from active politics, but without any additional financial sanction or other punishment. The Athenians used the mechanism sparingly, with ten attested ostracisms in the period 507–416.Footnote 26 During the 4th century, the graphe paranomon (suit for illegality) replaced ostracism. According to this procedure, any citizen could apply to the courts for annulling a decree passed by the Assembly for being contrary to the laws (general permanent rules applying to all individuals), and punish its proposer. If the court found the proposer guilty, it could hand a severe penalty, including heavy fines and the loss of political rights. The process was used with high frequency. Hansen (Reference Hansen1999: 208) claims ‘jurors must have judged a graphe paranomon something like once every month’, which, for the period 403–322 and for the administrative year divided into ten periods, implies a sum of 800.Footnote 27
As first noted by Headlam (Reference Headlam1891), lack of partisanship in elections for candidates explains to a large degree the persistent recourse of Athenian politicians to the law courts for injuring their opponents:
These political trials were really an opportunity for the expression of popular favour or distrust … And so we find that every decided change in Athenian policy is marked, not by the election of a new finance minister, or a new board of generals, but by the condemnation in the law courts of the former protector of the demos. (ibid: 36)Footnote 28
On this reading, political leaders were not necessarily brought to trial because they were guilty of illegal acts, but because the trial offered the citizenry, in its capacity as jurors, the opportunity to approve or disapprove a political leader. Further, the use of the courts in political disputes provides an additional explanation for the observations that in the 4th-century, orators rather than generals dominated politics.
It is finally important to reiterate a feedback loop from the absence of political parties to the majority voting rule. An absence of political parties negates the need to choose an electoral law, which would aggregate votes and allocate seats in the legislature to party candidates. Thus, majority voting emerges as an obvious rule to decide the election winner in an election concerning policies rather than candidates. More generally, direct participatory policy making under majority voting, sortition, accountability to the courts and the absence of political parties, comprised an integral structure and none of those institutions could operate independently of each other. Majority voting in the Assembly implied that all citizens carried the same weight in deciding issues of public interest, manifesting in practice equality of political rights and political power (respectively isonomy and isocratia, Hansen, Reference Hansen1999: 81). The latter went hand-in-hand with the right to hold public office, made effective by random selection and rotation. In turn, selection by lot obviated the intermediation of political parties and strengthened the primacy of the Assembly.
6 Conclusions
Political parties inform voters on issues of public policy, fight elections on multi-dimension policy platforms, nominate candidates for office who, when elected, choose policy. Ubiquitous in modern representative democracies, they were absent from the direct, participatory and deliberative democracy of ancient Athens. Previous scholarship attributed the absence of parties to a homogeneous citizenry who favoured congruence to antagonism. On the contrary, the present study focused on the absence of parties in competitive politics.
The premise of the Athenian democracy was the supremacy of the demos in deciding policy and equality of opportunity for citizens to hold public office. In practice these were manifested by debate and vote in the Assembly, sortition with rotation in office and scrutiny by the courts. In so far as initial circumstances condition later developments, parties were unlikely to emerge because they were absent at the launch of the democracy when the Assembly took direct control in a single constituency and adopted simple majority voting. Since citizens rather than elected officials initiated legislation, political leaders in the same office had to win the Assembly votes on policy issues rather than election to office. Selection to public office for routine administrative tasks by the luck of the draw and annual rotation diminished motives for partisan control of those appointments. Compared to modern representative democracies, these factors minimized incentives for political leaders and voters alike to coalesce in parties. However, they did not eliminate possible gains that political leaders could obtain from coordination in passing legislation. That the latter did not materialize is attributed to single-dimension voting by the Assembly using the simple majority rule. It nullified their incentives to form permanent alliances for preparing manifestos combining different policy dimensions, and for passing legislation. Finally, the participatory democracy fulfilled the role of informing citizens about policy issues, while the courts held political leaders to account. These conclusions are not a call for reforming representative government; nevertheless, they show that democracy operated successfully without political parties.
Author ORCIDs
George Tridimas, 0000-0002-3182-0322.
Acknowledgements
An earlier version of this paper was presented to the annual meeting of the European Public Choice Society in Rome, April 2018. I wish to thank Dennis Mueller, Thanos Pitsoulis and Soeren Schwuchow for useful discussions. I am also grateful to four anonymous referees for their helpful comments and suggestions. I am of course responsible for any remaining errors and omissions.
