Representative government gives no institutional role to the assembled people. That is what most obviously distinguishes it from the democracy of the ancient city-states. However, an analysis of the Athenian regime, the best-known example of classical democracy, shows that a further feature (one less often commented on) also separates representative democracy from so-called direct democracy. In the Athenian democracy, many important powers were not in the hands of the assembled people. Certain functions were performed by elected magistrates. But what is particularly remarkable is that most of the tasks not done by the Assembly were entrusted to citizens selected by a drawing of lots. By contrast, none of the representative governments set up in the last two centuries has ever used lot to assign even one modicum of political power, whether sovereign or executive, central or local. Representation has only been associated with the system of election, sometimes in combination with heredity (as in constitutional monarchies), but never with lot. So consistent and universal a phenomenon ought to invite attention and indeed scrutiny.
It cannot be accounted for, as can the absence of the popular assembly, by material constraints alone. To explain why representative governments grant no role to the assembly of citizens, authors usually talk about the size of modern states. It is simply not possible, in political entities so much larger and more populous than the citystates of Antiquity, to bring all the citizens together in one place to deliberate and make decisions as a body. Inevitably, therefore, the function of government is performed by a number of individualssmaller than the totality of citizens. As we have seen, the practical impossibility of gathering the whole people together was not the prime consideration motivating such founders of representative institutions as Madison or Sieyes. The fact remains that the sheer size of modern states had the effect of making it materially impracticable for the assembled people to play a part in government. Moreover, this is likely to have counted for something in the establishment of purely representative systems. On the other hand, it cannot have been the size of modern states that prompted the rejection of the lot system. Even in large, densely populated states it is technically feasible to use lot to select a small number of individuals from a bigger body. Whatever the size of that body, lot will always make it possible to extract therefrom as small a group of individuals as is required. As a method of selection, it is not impracticable; in fact, the judicial system still makes regular use of it today in constituting juries. So this exclusive recourse to election rather than lot cannot stem from purely practical constraints.
The political use of lot is virtually never thought about today.1 For a long time lot has had no place in the political culture of modern societies, and today we tend to regard it as a somewhat bizarre custom. We know, of course, that it was used in ancient Athens, and this fact is occasionally remarked upon, though chiefly in tones of amazement. In fact, that the Athenians could have adopted such a procedure seems to be the major puzzle. However, we may benefit from an inversion of the usual point of view whereby the culture of the present constitutes the center of the world. It might be better to ask: "Why do not we practice lot, and nonetheless call ourselves democrats?"
It might, of course, be objected that there is not a great deal to be learned from such a question and that the answer is obvious. Lot, it can be argued, selects anyone, no matter whom, including those with no particular aptitude for governing. It is therefore a manifestlydefective method of selection, and its disappearance requires no further explanation. This is an argument, however, in which the obviousness of the premise ought to cast doubt on the soundness of the conclusion. The Athenians, not generally regarded as unsophisticated in political matters, must have been aware that lot appointed people indiscriminately, yet they continued to use the system for two hundred years. The fact that selection by lot risks elevating unqualified citizens to public office is not a modern discovery. Incompetence in office was as much a danger in Athens as it is in present-day polities. Moreover, if Xenophon is to be believed, Socrates himself ridiculed the appointment of magistrates by lot on the grounds that no one chose ships' pilots, architects, or fluteplayers by this method.2 That means, however, that the question we should be asking is whether the Athenian democrats really did have no answer when faced with this objection. Possibly they saw advantages in lot that, all things considered, they felt outweighed this major disadvantage. Possibly, too, they had found a way of guarding against the risk of incompetence through supplementary institutional arrangements. Concerning lot, it is by no means clear that the danger of incompetence is the last word. We cannot pronounce this selection method defective and destined to disappear before we have carefully analysed how it was used in Athens and how democrats justified it.
In any case, whatever the reason lot disappeared, the crucial fact remains that Athenian democracy employed it to fill certain posts, whereas representative regimes give it no place whatsoever. The difference can hardly be without consequence on the exercise of power, the way it is distributed, and the characteristics of those who govern. The problem is identifying the consequences with any precision. So if we wish to throw light on one of the major differences between representative government and "direct" democracy, we need to compare the effects of election with those of lot.
Analyses of representative government typically contrast election with heredity. In part, such a viewpoint is justified: elected governments directly replaced hereditary governments, and there is no doubt that, in making election the chief basis of political legitimacy, the founders of our modern representative republics were above all rejecting the hereditary principle. Modern representative systems are certainly characterized by the fact that in them power is not inherited (not in essence, anyway). But what also distinguishes them, even if it receives less attention, is the complete absence of the use of lot in the assignment of political functions exercised by a restricted number of citizens. The contrast between election and lot might reveal an aspect of representative government that remains hidden so long as the hereditary system constitutes the sole point of contrast.
A study of the use of lot in Athens is in order, not only because lot is one of the distinguishing features of "direct" democracy, but also because the Athenians employed it side by side with election, which makes their institutions particularly well suited for a comparison of the two methods. Moreover, the recent publication of a superb study of Athenian democracy, remarkable in both its breadth and precision, has thrown fresh light on these points.
The Athenian democracy entrusted to citizens drawn by lot most of the functions not performed by the Popular Assembly (ekklesia). This principle applied mainly to the magistracies (archai). Of the approximately 700 magistrate posts that made up the Athenianadministration, some 600 were filled by lot. The magistracies assigned by lot (kleros) were usually collegial. The term of office was one year. A citizen was not permitted to hold a given magistracy more than once, and while he might be appointed to a number of different magistracies during his lifetime, the timetable for rendering account (no one might accede to a fresh post before having rendered account for the previous one) meant that a person could not in practice serve as a magistrate two consecutive years.