[Elster, Jon (1991), Arguing and bargaining in the Federal Convention and the Assemblée Constituante, Working Paper - University of Chicago, August 1991]
[This text was found on the Net. Since I have not seen the original document I cannot guarantee the correct reproduction of pargraphs, French letters and italics. There were no page numbers.]

[ Books | Articles | Reviews | Index | The main menu ]

Arguing and Bargaining

in the Federal Convention and the Assemblée Constituante

WORKING PAPER NUMBER 4 Center for the Study of Constitutionalism in Eastern Europe

Jon Elster
University of Chicago
August 1991


The topic of the paper is the nature of the debates that took place in the Federal Convention in Philadelphia in 1787 and the Assemblée Constituante in Paris 1789-91.1 I shall not be concerned with the outcomes of the debates, i.e. with the constitutional documents that emerged. Instead, I focus on the style of debating. To what extent can the debates be described as a form of bargaining, oriented towards a compromise between different interest groups? And to what extent were the participants engaged in ar gument and discussion, constrained by ideals of impartiality and consistency?

The Federal Convention for a considerable part revolved around bargaining and compromise. The Constituent Assembly had very little bargaining, if any at all. Instead, its members argued consistently in terms of the common good. This difference was due partly to the nature of the issues, partly to the very different settings in which the debates took place, and partly to differences in political philosophy.

I am not claiming that the French were motivated by loftier ideals or were somehow more rational and disinterested. To argue for and against claims of this sort, one would have to go beyond what was said in the assemblies, and look at what the participants said elsewhere and, above all, at their actions. I am concerned with their words, not with their thoughts. But sometimes words matter as much as thoughts. One difference between open debate and closed conclaves is that the public setting forces the participants to argue in terms of the general good: certain arguments from self-interest do not stand the light of day.2 This does not imply that a genuine consensus will be reached. As wonderfully shown in Tocqueville's Souvenirs, assembly debates can also lead to polarization and to false consensus. The debates of 1789-91 illustrate both these perversions of public debate. But even if the participants only pay lip- service to the common good, the outcome will be different. When voting is public, they have to put their vote where their mouth is. And when debates are public, their utterings are constrained by the need to argue in general, disinterested terms.

Although the paper ostensibly is about the late eighteenth century, it was written very much with the late twentieth century in mind. The political transitions and constitutional debates that are currently taking place in Eastern Europe and the Soviet Union have a number of features in common with the American and French experiences. In a final section I briefly allude to some of the similarities. Needless to say, as all historical analogies they are partial and limited, but used with caution they may suggest distinctions to be made and hypotheses to be explored.

Arguing and bargaining

Let me first spell out the distinction between the two modes of communication, and first note that I am concerned here only with communications that are constrained by the need to make a decision within a specified time limit. A constituent assembly, for instance, cannot behave as if it were a philosophical seminar, in which argument and counterargument can be pursued indefinitely without any need to reach a conclusion.

To argue is to engage in communication for the purpose of persuading an opponent, i.e. to make the other change beliefs about factual or normative matters. In such discussions, the only thing that is supposed to count is "the power of the better argument". The parties are not allowed to appeal to their superior material resources. The factual or normative statements asserted in a process of arguing are made with a claim to being valid. For factual statements, validity means the same as truth. For normative statements, the notion of validity is more controversial. For the present purposes, I shall focus on impartiality as a necessary condition for normative validity.

To bargain is to engage in communication for the purpose of forcing or inducing the opponent to accept one's claim. To achieve this end, bargainers rely on threats and promises that will have to be executed outside of the assembly itself.3 Bargaining power does not derive from the "power of the better argument", but from material resources, manpower and the like. Statements asserted in a process of bargaining are made with a claim to being credible, in the sense that the bargainers must try to make their opponents believe that the threats or promises would actually be carried out.

Threats and promises involve post-communication behavior. The need for credibility may also induce specific forms of pre-communication behavior. Before the actual bargaining begins, the parties may take certain actions - precommit themselves - for the purpose of lending credibility to threats. Routine precommitment strategies include the choice of people well-known for their intransigeance as negotiators; the imposition of bound mandates on the negotiators; the use of delaying devices such as the obligatory referral of proposals to one's constituency; and the public announcement of demands in ways that make it hard to step down without loss of prestige. More exotic devices can also be imagined.4

The process of rational argument can also involve pre-discussion choices. Since universal participation is often impracticable, selection and representation procedures have to be established. Since the pursuit of consensus may take too much time, some voting procedure usually has to be adopted. Clearly, the representatives cannot themselves decide how they are to be selected. Typically, in the process that decides how representatives are to be elected or selected it is also decided how they are to decide once assembled. The representatives may, however, decide to overrule that prior decision (see below).

Most instances of bargaining also contain elements of argumentation. In collective wage negotiations, for instance, much of the time is spent arguing about factual matters, such as trends in the purchasing power of the workers or the financial health of the firm. Normative arguments are also used, as when workers claim that their moderation in bad times entitles them to a larger share of the gains in good times.5 Yet typically normative arguments serve a different function in such contexts, that of lending credibility to threats that would otherwise not be believable. Usually, threats are credible only if it would be in the interest of the threatener to carry them out if the occasion should arise. However, individuals who are motivated by strong perceptions of fairness are sometimes willing to hurt themselves rather than be treated unfairly. Hence they can credibly threaten to break off negotiations even when it is clear that they would do worse on their own than on the terms proposed by the opponent.

Conversely, many apparently rational discussions take place behind a hidden threat that one or more of the participants might leave the table or the assembly. In countries where the government has the power to dissolve parliament, this threat is constantly present. Even when it is rarely invoked and still more rarely used, the threat may have a considerable impact on the proceedings. Also, members of the parliamentary opposition may be influenced by implicit or explicit promises that their constituencies will receive government contracts if they vote for the government's proposal.

Sometimes, it may be difficult to decide whether a given communication is part of an argument or a move in a bargaining process. In particular, it is not always clear whether a given statement is meant as a threat or as a warning. To influence the beliefs of the other party, it is sometimes useful to utter a warning to the effect that if the opponent does such-and-such, so-and-so will happen. In principle, it is easy to distinguish threats from warnings. In practice, there is considerable ambiguity. A trade union leader may tell the management negotiator that "Unless our demands are met, I will be unable to keep my members from striking". Formally, the statement is issued as a warning - an assertion about certain objective facts outside anyone's control. In reality, it is common knowledge to the parties that the strike propensity is to a large extent within the control of the union leader. When a threat is disguised as a warning, we could also and equivalently say that a process of bargaining is disguised as a process of arguing or rational discussion.

In actual communicative situations one will usually find elements of both arguing and bargaining. And in some situations it may not be clear whether we are dealing with the one or the other. These facts do not, however, destroy the distinction. Arguing and bargaining have to be seen as two ideal-typical modes of communication, which are rarely instantiated in their pure form.

We may distinguish between opportunistic and non-opportunistic statements. An opportunistic factual statement is a misrepresentation of facts for self-serving purposes. In bargaining contexts, such misrepresentations routinely occur when the parties make statements about their own preferences. A party who can successfully disguise his impatience, for instance, will enhance his ability to make credible threats and be more likely to get a good agreement. Similar misrepresentations of fact take place in argumentative situations. Each party will typically stress the factual aspects of the situation that support its own position, and neglect others. Other parties may in turn be counted on to expose this bias, so that in the end a more balanced picture emerges.6

An opportunistic normative statement is a self-interested claim dressed up as an impartial one. In collective wage bargaining, for instance, appeals to principles of distributive justice are often made opportunistically. The number of plausible-sounding norms of fairness is so large that most groups will be able to find some norm that corresponds, at least roughly, to their self-interest. (In fact, rough correspondence is usually better than perfect correspondence, since it may enhance the efficacy of the normative argument by defusing the suspicion that the norm is used for self-serving purposes.) In political argumentation, conservatives may appeal to liberty and fundamental rights when defending the interests of private business, while socialists may appeal to ideals of social justice when defending the interests of their constituents. But there is an important difference between arguing and bargaining in this respect. In bargaining contexts, one may appeal to impartial normative arguments, to convince the opponent or to lend credibility to threats and warnings, but one does not have to do so. Bargaining often proceeds without reference to any motives beyond self-interest. In argumentative situations, one has to phrase one's argument in impartial terms, as if one were arguing for the public good and not for one's own self-interest. To persuade others that society ought to be organized so as to accommodate a particular set of private interests one must either back the claim by a credible threat or argue that this arrangement will in fact serve a more impartial end that commands general acceptance. One cannot say, "I want this outcome because it is good for me", and expect others to be persuaded. There simply is no language game of that kind.

I do not mean to suggest that all impartial normative arguments are opportunistic, or that all motives ultimately are self-interested. An argument to the effect that environmental policies have to take account of the interests of future generations is likely to be genuinely impartial and disinterested. Arguments in favor of the mentally handicapped or of third-world countries are also likely to have an impartial basis. And even when a group argues for a policy that does favor itself, it may do so for other than opportunistic reasons. When women claimed the right to vote, they could ground their claim in widely accepted standards of fairness. My point is not, therefore, that self- interest is rampant. Rather my claim is that when people are moved by self-interest, the context of rational discussion forces them to disguise it as something else.

In addition to the norm of impartiality that constrains argumentative communication, such debates are subject to a norm of consistency. Any given normative argument may, in any given situation, be used in an opportunistic manner. But once it has been used, the speaker is stuck with it. If he wants to retain his persuasive powers, he cannot simply disregard the argument if a later situation should arise in which it works against his self-interest. A participant who is seen as choosing norms a la carte, and discarding them whenever they work against him, will undermine himself in the long run.

The joint impact of the constraints of impartiality and consistency can be considerable. They force speakers to deploy arguments that not only contain no explicit reference to their self-interest, but may actually go against it. And when the decision procedure is public, a further norm of consistency - between words and behavior - forces people to vote for the proposal whose merit they have been arguing. The final decision may not in any recognizable sense reflect the public interest, but nor is there any reason to believe that it will always reflect a compromise between the private interests of the parties.

Philadelphia and Paris

The differences between the two assemblies are more striking and far-reaching than the similarities. I begin, nevertheless, with the latter.

The two assemblies took place at about the same time. They had many intellectual ancestors in common. The orators use many of the same historical precedents in arguing for or against various institutional arrange-ments. The British political system, in particular, is constantly invoked in both cases. Several of the issues on the agenda were similar: bicameralism, executive veto, the link of suffrage to property, how the constitution itself was to be revised. Also, as discussed more fully below, both assemblies faced the problem of how far they could deviate from their mandates and claim to represent the nation as a whole rather than the states or estates that had sent them. The most important difference is that in Paris the first task of the assembly was to destroy the old regime, including the institutions that supported two of the three deliberating estates. After the abolition of feudal privileges and the confiscation of church property, the Revolution in France became a "winner-take-all" process which gave very little room for compromise. Some exceptions are noted below, but the characterization seems roughly correct. In Philadelphia, by contrast, the Confederation remained very much a political reality throughout the proceedings. Consequently, the bootstrapping procedures used in the Convention (see below) were much less radical than those of the Constituent Assembly.

A second important difference is that the American events more naturally lend themselves to a bargaining interpretation, because the actors were geographically distinct states rather than functionally defined estates. I shall return to this point later. Here it is sufficient to point out that while it is clear what it means for a state to threaten to withdraw from a federation, it is harder to understand what it would mean for, say, the nobility, to threaten to withdraw from society. Yet bargaining among estates cannot be excluded. A Church under attack can threaten to withhold services that are of immense importance to the citizens and that no other institution can provide.

The American assembly did not work under pressure. Although most of the delegates thought that the Confederation was in need of reform, there was agreement that the situation was less urgent than it had been when the Confederation had originally been formed. Indeed, James Wilson argued that the Convention, precisely because it took place under less urgent circumstances, could be used to redress some of the unfair decisions that had been made in 1776.7 In Paris, by contrast, the work of the assembly took place under menace and pressure first from the king and later from the Parisian crowds.8 Also, the financial situation was desperate.

The American debates took place behind closed doors, and the proceedings were kept secret. The secrecy could be kept because the group of delegates was small and the Convention itself lasted less than four months. The French debates were intensely public; there were more than a thousand delegates; and the proceedings stretched out for more than two years. These facts of space, time and number inevitably set the tone for the debates. The American discussions, in particular, are virtually free of the moralizing cant - invoking either the goodness and wisdom of the king or similarly infallible virtues in the people - which characterizes all but a very few of the French delegates.

Finally, the American constitution lasted - the French did not. Since 1791 up to the present France has had about a dozen constitutions, whereas the United States have kept the constitution made in Philadelphia, amended only within the framework laid down in that document itself. Since the successors of the 1791 constitution have also been short-lived, the circumstances in which it was made may not provide the full explanation of its lack of staying power. There may be more general elements in French political culture that enter in to the explanation of its failure as well as of that of later constitutions. But it seems plausible that part of the explanation could be the radical procedures adopted by the Constituent Assembly. As we shall see, it did not adopt for itself any of the flywheels it recommended to stabilize later legislatures, nor any other cooling or delaying devices.

In summary, it is evident that the conditions for reaching agreement and sticking to it were much more propitious in Philadelphia than in Paris. An overdrawn way of stating the contrast could be that in France the debates led to a false and unstable consensus, made under the pressure of events and the radical clubs, whereas in the United States the outcome was a viable compromise, hammered out by pragmatic negotiators. But the contrast is too stark. In both assemblies, there were genuine discussions that led to genuine consensus.


Constitutions can be changed or even made within a constitutional framework.9 This was not the case in either of the two assemblies. Instead, they had to make up their own rules to a considerable extent. In the French case, everything was essentially created ex nihilo. In the American case, the procedure by which the Convention was established did to some extent constrain what the delegates could do and propose, but they also had a great deal of latitude in defining their own mandate.

To get the question properly in focus we may distinguish between four sets of procedures.
First, there are the procedures used in sending delegates to the assembly.
Second, there are the procedures used by the assembly in its deliberations.
Thirdly, there are the procedures proposed by the assembly for future legislatures.
Finally, there are the procedures proposed by the assembly for ratification of these proposals.
These procedures can be most clearly contrasted in the American case.

The delegates were sent by the State legislatures, some of them with instructions about how to vote on specific issues. Thus the delegates represented their states, not the American people.

Voting in the Convention itself proceeded along similar lines. Although the Pennsylvanians wanted to refuse the smaller states an equal vote in the Convention, their proposal was never put on the table.10 When a committee was formed to forge a compromise on the second chamber, James Wilson "objected to the committee because it would decide according to that very rule of voting which was opposed on one side"11, but to no avail.

Yet this procedure could not in itself ensure that the outcome would be equal representation in the senate. The large states could and did use their bargaining power to try to get a different outcome. They failed, but not because the rules of voting in the Convention made equal representation a foregone conclusion. Later, I offer an alternative explanation of the failure.

Finally, the ratification procedure stipulated that the constitution had to be approved by conventions in nine of the thirteen states. Since this procedure corresponded to the Articles of Confederation, it might be seen as reflecting the spirit of that document rather than the spirit of the document which was to be ratified. On the other hand, the proposal to have the ratification done by special conventions rather than by the State legislatures reflected the spirit of the new document. Instead of sending the constitution through a cumbersome two-step process of ratification by the legislature, the framers thought it preferable to have a direct appeal to the people.

Another aspect of the issue was whether the Convention had the authority to propose the radical changes embodied in the constitution. Over and over again, the argument was made that the delegates were authorized only to proposes changes in the Articles of Confederation, not to create a wholly new system.12 The advocates for a radical change had two replies. James Wilson said that "He conceived himself authorized to conclude nothing, but to be liberty to propose any thing."13 George Mason argued that "In certain seasons of public danger it is commendable to exceed power"; Edmund Randolph, similarly, "was not scrupulous on the point of power".14

These issues also surface in the French case. The delegates to the Constituent Assembly decided to ignore their instructions altogether, and consider themselves the representatives of the nation rather than of their estates. The problem of voting by estate or by head is analogous to the problem of equal or proportional voting by the States in the Convention, but it was resolved in the opposite manner. In their discussions, therefore, the delegates were a single body, deciding by majority vote. Some of them thought, however, that future legislatures had to be bicameral. Mounier, for instance, argued that the degree of strength created by a single-chamber assembly, while appropriate for making a constitution, would be inappropriate in a legislative assembly.15 Similarly, Mirabeau argued that the King should have veto in the constitution, but not over the constitution itself.16 Summarizing both points, Clermont-Tonnerre observed that the "three- headed hydra" - king, first chamber and second chamber - that the constitution should create could not itself have created a constitution.17

The Constituent assembly discussed whether to have the Constitution ratified by the people, e.g. by the first legislature following the Constituent Assembly, but in the end made no proposal for popular ratification. The constitution was submitted to the King, who accepted it. The French delegates did not feel bound by their mandates, but accorded themselves the right to discuss whatever issues they wanted to take up. Lally-Tollendal made a half-hearted attempt, somewhat along the lines of Wilson's argument cited earlier, to reconcile the autonomy of the delegators with those of the delegated: "Il faut étre entiérement constitués pour faire des lois, mais il suffit d'étre assemblés pour faire des projets de lois."18 But the fine distinction proved pointless, since the delegates just went ahead and decided that they were "entirely constituted".

The French case presents an important aspect not found in the American one. The Assemblée Constituante functioned also as an ordinary legislative assembly. That arrangement may be undesirable, since a main task of a constitutional assembly is to strike the proper balance of power between the legislative and the executive branches of government. A legislative assembly that writes itself into the central place in the constitution may not enjoy sufficient legitimacy. The members of the French Assemblée Constituante were acutely aware of this problem, and solved it by voting themselves ineligible for the first legislative assembly. Today, this decision - proposed by Robespierre - is generally acknowledged as a disastrous piece of populist overkill. It did, however, reflect a genuine problem.

The problem of bootstrap-pulling is an aspect of the problem of legitimacy. Procedural legitimacy can be upstream or downstream, by delegation or by ratification. (The two can interact, since those who delegate the power to propose changes often want to approve them.) Some strategic aspects of the constitutional process follow from the need to compromise with legitimacy-conferring institutions. The Federal Convention could not ignore the States. The Constituent Assembly could not ignore the King. At the same time, framers may try to secure legitimacy by the soundness of their proposals. This was the strategy adopted by the leading delegates in Philadelphia: to work within the framework of the Confederation, yet bend and stretch it as much as possible to promote a better mode of government. A roughly analogous scenario might have occurred in France, had it not been for the King's flight to Varennes.

Bargaining, uncertainty and impartiality

At one point during the Philadelphia debates, Gouverneur Morris argued that "If we can't agree on a rule that will be just at this time, how can we expect to find one that will be just in all times to come."19 Perhaps we should stand this observation on its head. It may be easier to create a constitution for the future than to find one that commands agreement in the present. The basic fact of uncertainty that makes it hard to legislate for the future may also in other and more important respects make that task easier.

The future circumstances and needs of society cannot be known with any degree of precision. Hence, for instance, George Mason's doubts about the motion to deny Congress the right to emit bills on the credit of the U.S.: "Though he had a mortal hatred to paper money, yet as he could not foresee all emergences, he was unwilling to tie the hands of the Legislature. He observed that the late war could not have been carried on, had such a prohibition existed."20 Yet in another context Mason argues persuasively that uncertainty can create consensus:
We ought to attend to the rights of every class of people. He had often wondered at the indifference of the superior classes of society to this dictate of humanity and policy, considering that however affluent their circumstances, or elevated their situations, might be, the course of a few years, not only might but certainly would distribute their posteriority through the lowest classes of society. Every selfish motive therefore, every family attachment, ought to recommend such as system of policy as would provide no less carefully for the rights and happiness of the lowest than of the highest orders of citizens.21
Such veil-of-ignorance arguments 22 were also used in other contexts. Thus Gouverneur Morris argued that "State attachments, and State importance have been the bane of this Country. We cannot annihilate; but we may perhaps take out the teeth of the serpents. He wished our ideas to be enlarged to the true interest of man, instead of being circumscribed within the narrow compass of a particular spot. And after all how little can be the motive yielded by selfishness for such a policy. Who can say whether he himself, much less whether his children, will the next year be an inhabitant of this or that state."23

This argument refers to the thirteen states then in existence, but it was also used to cover the accession of future states. Against Gerry's proposal to "limit the number of new states to be admitted into the Union, in such a manner, that they should never be able to outnumber the Atlantic States"24, Sherman replied that "we are providing for our posterity, for our children and grand children, who would be as likely to be citizens of new Western states as of the old states".25

These veil-of-ignorance arguments are based on uncertainty and the long-term interest of families. They form, therefore, an intermediate category between self-interested bargaining and disinterested reasoning about the common good. On the one hand, there is no scope for bargaining when interests coincide, as they do behind the veil of ignorance. On the other hand, veil-of-ignorance arguments based on uncertainty should not be confused with arguments from impartiality. Mason's argument for granting new states equal states belongs to this latter category: "Strong objections have been drawn from the danger to the Atlantic interests from new Western states. Ought we to sacrifice what we know to be right in itself, lest it should prove favorable to states which are not yet in existence. If the Western states are to be admitted into the Union as they arise, they must, he would repeat, be treated as equals, and subjected to no degrading discrimination."26

For the remainder of this paper, I shall disregard the intermediate category, and focus on the extremes of bargaining among parties with conflicting short-term interests and impartial, rational discussion about the common good. In the American case, I shall assume that delegates bargain on behalf of their states, not on behalf of their personal interests.27 Since this is a paper about the nature of the debates and not about ultimate motivations, this assumption seems justified. Although the closed doors of the Philadelphia assembly made it possible to argue explicitly in terms of the interests of states, the doors were not closed enough to allow personal self-interest to come into the open.

The distinction between bargaining and rational discussion can be related to a distinction between the transitional effects and the equilibrium effects of a new constitution. It is tempting to see the goal of rational discussion as that of creating the best constitution sub specie aeternitatis, i.e. the one with the best equilibrium properties. This view is misguided, however. The goal ought to be that of finding the best possible development from now on.28 Extreme focus on equilibrium properties can lead to unjustified imposition of suffering on the generations that have the bad luck of living in the transitional period.

In the Constituent Assembly, Malouet stands out in referring to "tous les hommes paisibles qui ne veulent point sacrifier leur repos et la génération actuelle aux générations futures".29 The welfare of the current generation ought to count more heavily than that of any single future generation, not because of time discounting per se, but because of the uncertainty that will always surround conjectures about long-term effects. But even were there no uncertainty at all, the current generation has rights that block the intertemporal utilitarian calculus. For these reasons, rational discussion ought to take account of transitional effects and not only of equilibrium properties of the new system.

Other transitional effects may be relevant for reasons of upstream and downstream legitimacy. If impartial discussion about the common good led to consensus on a constitution that stood no chance of being adopted, it would be an instance of rationalism, not of rationality. In the Federal Convention, the debate over ratification illustrates this point. As observed by Rufus King, "A Convention being a single house, the adoption may be more easily carried thro' it than thro' the Legislatures where there are several branches. The Legislatures also being to lose power, will be most likely to raise objections."30 King and Wilson 31 used the latter fact as an argument for a different procedure of ratification, not involving the State legislatures. Charles Pinckney, however, argued the other way around. The first branch of the national legislature, he argued, should be elected by the state legislatures and not by the people, among other reasons because "the legislatures would be less likely to promote the adoption of the new government, if they were to be excluded from all share in it."32

I shall focus, however, on the transitional effects that are related to bargaining. Any specific constitutional design will affect the interests of currently existing groups. Even if all expect to gain, as I shall assume, each arrangement will have special benefits for some groups. This is the general problem of bargaining: how to divide the benefits from cooperation? How to choose one among several Pareto- improvements? In the long-run, the differential impact is attenuated by uncertainty, as I explained earlier. In the short run, however, gains can differ strongly.

The interaction between bargaining arguments and arguments from impartiality is too complex to be considered here, especially since it is often not clear whether arguments from the common good are not really group interest arguments in disguise. My task here is to isolate the two strands as they appear in the debates, not to consider how well they reflect different types of motivations or how they interact to produce the final outcome. One challenge must be faced, however.33 Given the need for the framers to have the constitution ratified or, more generally, accepted by interest groups in society, why would the final outcome differ from that which would be produced by pure bargaining? In other words, are not the pragmatic constraints on impartiality so strong as to remove all bite from the idea of rational discussion among disinterested agents?

Let us assume, for simplicity, that there is no element of bargaining in the assembly. The framers are exclusively motivated by the common good. Let us call their first-best outcome A. Let us assume, moreover, that if they had been exclusively concerned with bargaining on behalf of economic and political interest groups, outcome B would have been reached. The question is whether the constraints on ratification or, more generally, acceptance of the proposed constitution, will force the framers to propose B or something very close to B, or whether they have some latitude to move a considerable way towards A. If the situation is very urgent, because of the threat of war, a financial crisis or need for reconstruction, the framers may be able to impose what is from their point of view a very good second-best solution. The costs of sending the delegates back to the conference table or of breaking off talks altogether would be prohibitive. A second reason for believing in the autonomy of the constituent assembly is that in general nobody knows exactly or even approximately what B looks like. If the proposed constitution provides a minimal degree of satisfaction to all parties, no party may feel certain that it would have been able to get more by using credible threats. Thus the framers can play on transaction costs as well as uncertainty to carve out a space for impartiality.

Rational argumentation

I shall distinguish between consequentialist and non- consequentialist arguments. Both were important in the constituent assemblies. Appeals to consequences abound, as do rights-based arguments. Proceeding from this distinction, I shall discuss two sets of issues. First, how are rights and consequences set against each other? Second, what is the structure of consequentialist reasoning?

Issues of representation and suffrage can be used to illustrate the trade-off between rights and outcomes. John Dickinson argued that any scheme that would give some states no representation in the Senate would be "unfair".34 Madison argued that any deviation from proportional representation was "unjust".35 To reconcile the two claims from justice, the smallest state could be given one representative, and the larger ones a proportionally larger number. This would, however, give a very large Senate, which would, in Madison's eyes, be "inexpedient".36 One way of describing the system that was finally chosen would be to say that expediency and Dickinson's conception of justice won out against Madison's conception. Another way of describing it, to which I turn below, would be to say that the bargaining power of the larger states proved to be no larger than that of the small states.

The issue of suffrage similarly had two aspects. On the one hand, there was a concern for outcomes, i.e. with good decision-making. In both assemblies, property restrictions on suffrage were justified on various consequentialist grounds. In his speech on the suffrage, Barnave singled out three arguments for property qualifications. They tend to promote, if not invariably at least usually, lumiéres, intért ˆ la chose publique, and indéependance de la fortune.37 At the Federal Convention, Gouverneur Morris argued that property rather than tax-paying should be the criterion for the right to vote. If the poor had the vote, the rich would buy their votes, and a an aristocracy of wealth would be set up.38

On the other hand, there was a concern with the right to participate. The right was perceived not as unconditional, but as contingent on contribution to one's country. The idea was not only "No taxation without representation", but equally "No representation without taxation". Siéyes, for instance, made a distinction between active and passive citizens: "Les femmes, du moins dans l'état actuel, les enfants, les étrangers, et tous ceux qui ne contribueraient en rien ˆ soutenir l'établissement public, ne doi vent point influer activement sur la chose publique. Tous peuvent jouir des avantages de la société, mais ceux-lˆ seuls qui contribuent ˆ l'établissement public, sont comme les vrais actionnaires de la grande entreprise sociale."39

Contribution does not only include paying taxes, but also defending one's country. In Philadelphia, Benjamin Franklin insisted strongly on this idea: "in time of war a country owed much to the lower class of citizens. Our late war was an instance of what they could suffer and perform. If denied the right of suffrage it would debase their spirit and detach them from the interest of the country".40 In Paris, Clermont- Tonnerre made a brief reference to the fact that Jews serve in the national militia, to refute one of several arguments for denying them citizenship and, a fortiori, the right to vote.41

Various voting schemes can be seen as attempting to incorporate all of these elements, instrumental as well as rights-based. Thus on August 28 1789 the Constituent Assembly set up a three-tiered system that was designed to ensure universal participation as well as concern for the quality of decision.42 At the lowest level, there were weak tax-paying qualifications for being an active citizen, including the right to vote in the primary assemblies. These assemblies elected a corps of electors who in turn chose the deputies to the national assembly. To be eligible as elector, relatively strong income qualifications had to be satisfied. To be eligible as deputy, it was enough to satisfy the weak qualification. Here, the (conditional) right to participate is embodied in the first and third stages, and the instrumentalist requirements of competence at the second stage.

I now turn to the structure of the consequentialist arguments. They turn largely on factual issues of two kinds. First, how will society develop over the foreseeable future? Second and most important, what are the causal relations between institutional environment and individual behavior? I shall discuss them in that order.

In the debates in the Constituent Assembly I have not come across a single reference to probable future developments of the economy or demography of the country. In the Federal Convention, both Madison and Gouverneur Morris draw support for their views on suffrage from socio-economic extrapolations. Madison was worried about requiring landed property for members of Congress. Looking back, he observed that "It had often happened that men who had acquired landed property on credit, got into the legislatures with a view of promoting an unjust protection against their creditors".43 Looking to the future, he asserted that "The three principal classes into which our citizens were divisible, were the landed, the commercial, and the manufacturing. The 2d and 3rd class bear as yet a small proportion to the first. The proportion, however, will daily increase. We see in the populous countries in Europe now, what we shall be hereafter."44 From this he concluded that the non-landed interests ought not to be excluded from the Legislature.

In his views on the suffrage, cited above, Gouverneur Morris draws on a similar analysis. "We should not confine our attention to the present moment. The time is not distant when this country will abound with mechanics and manufacturers who will receive their bread from their employers."45 As mentioned, he drew the apparently paradoxical conclusion that since the number of citizens without landed property could be expected to increase, it was all the more important that they not be given the vote. He agreed with Madison, however, in that landed property should not be a condition of eligibility. In fact, the two of them "concurred... in thinking that qualifications in the electors would be much more effectual than in the elected".46 The resemblance with the French scheme discussed earlier should be noted.

These predictions, however, are much less important than the general causal propositions asserted time and again in the two assemblies. The members dabbled in psychology, economics, sociology and political science to an extraordinary extent. Some of their arguments seem fairly sound, while others are highly speculative. Generally speaking, political institutions can have two kinds of effects on behavior. The first belongs to economics and public choice: given individual motivations, institutions can set up incentives for individuals to act in socially desirable or undesirable ways. The second effect falls within the scope of psychology: individual motivations may themselves be affected by the institutional environment. The more robust arguments fall mainly in the first category, whereas the arguments of the second kind are more fragile and conjectural.

A general assumption made in the debates at the Federal Convention is that officials will act in the interests of their constituents if and only if this coincides with their own self-interest. Wilson, Gouverneur Morris and Madison all make this point.47 It is not clear whether the claim reflects a substantive theory of human nature or, as in Hume 48, simply a worst-case assumption adopted on grounds of prudence. If the latter, the apparently cynical assumption is actually tainted with rosy optimism, since there are worse things than self-interest. Envy, for instance, can induce people to act against their self- interest, cutting off their nose to spite their face.49 In the debates at the Convention there are a few passages that sound as if they are assuming envious motivations 50, but by and large the general hypothesis seems to be that people act in their narrowly conceived material interest.

Madison's fear that indebted landowners might adopt legislation favoring creditors is a typical example of incentive-effect reasoning. Another example, also from Madison, occurs in his comments on a proposal that in voting for the president, each elector should have two votes, one of which at least should be cast for a candidate not from his own state. Madison "thought something valuable might be made of the suggestion... The only objection which occurred was that each citizen after having given his vote for his favorite fellow citizen would throw away his second vote on some obscure citizen of another State, in order to ensure the object of his first choice."51

A further Madisonian example refers to the advantage of having the president elected by a set of electors specifically chosen for this purpose: "As the electors would be chosen for the occasion, would meet at once, and proceed immediately to an appointment, there would be very little opportunity for cabal, or opportunity".52

A final example, also from Madison, is less convincing. Arguing against selection of the executive by the legislature, he writes that "the candidate would intrigue with the legislature, would derive his appointment from the predominant faction, and be apt to render his administration subservient to its views".53 But it is not clear that a candidate's promise to favor his electors would be credible. Although Madison seems to subscribe to the maxim "Let the king beware of the kingmaker", the opposite view, "Let the kingmaker beware of the king", would seem just as appropriate.

Many other examples from the Convention could have been cited, but these give the general flavor of the arguments. Straightforward incentive reasoning of this kind is much rarer in the Constituent Assembly, mainly, no doubt, because the French delegates were totally lacking in political experience. Although the French representatives, like their American counterparts, draw on examples from ancient and modern history, they lacked the practical knowledge of politics that the Americans had acquired in Congress or in the State legislatures. Although the speeches contain many general warnings about the risk of the legislative or the executive usurping power, there are few specific analyses of the kind cited in the previous paragraph. To be sure, there are exceptions. I shall cite two of them, both taken from the report by Bergasse on the organization of the judiciary.

Bergasse beings by asking, "Comment, par l'institution mme des formes destinées ˆ procurer la conviction des coupables, parviendrez vous ˆ faire na”tre la confiance dans le coeur de l'homme injustement accusé?" One of the procedures he proposes for enhancing confidence in the judiciary is the following:
La confiance na”tra, si le magistrat qui applique la loi est distingué du magistrat qui met sous la puissance de la loi, c'est-ˆ-dire du magistrat qui décréte l'accusé. ...Tant que le magistrat qui décréte sera le mme que celui qui juge, vous aurez toujours ˆ craindre, s'il a décrété sur de faux soupons, son amour-propre ou sa prévention ne le portent ˆ justifier, par une condamnation inique, un décret injustement lancé.54
The argument, in other words, is that the unjustly accused man has less to fear if he knows that the judge has invested nothing of his "amour-propre" in the decision to prosecute. The argument is not that others might corrupt the judge, but that he might be corrupted by his own pride and refusal to admit that he had made a mistake.

In discussing the organization of the police, Bergasse suggests that justices of peace, unlike other judges, should be chosen by popular election for a short period. Anticipating the objection that such judges might abuse their position, he asserts that
Tant qu'un homme ne dispose que du pouvoir d'un moment, et que, destiné ˆ rentrer dans la classe ordinaire des citoyens il sent qu'il ne peut accro”tre ce pouvoir sans se nuire ˆ lui-mme lorsqu'il n'en disposera pas, il n'est pas ˆ craindre qu'il en abuse, et qu'il fasse server ˆ ses passions particuliéres une autorité qui, dans la suite, employée par un autre, pourrait si facilement lui devenir funeste.55
The argument can be read in one of two ways. (i) If the judge abuses his power, the increased powers of his office will make it easier for his successor to punish him. (ii) If he abuses his power, his successor will do so too, and that might harm him. On either reading, especially the second, it lacks the robust realism of the Americans, who assumed that officials "make hay when the sun shines"56.

The other causal issue concerns the impact of institutions on the motivation of individuals. As a broad generalization, I think it is true to say that this concern is more prominent in the French assembly. The American delegates did not think that the citizens could be reformed by the constitution. At best, cleverly designed institutions could force them to consider their enlightened self-interest, rather than yielding to the impulses of the moment and the passions of the crowd. This was the purpose of the various delaying and cooling devices that figure centrally in both debates.

The French, however, went further. They believed in the virtues of public deliberation as a means for changing preferences, rather than merely aggregating them.57 The argument is most clearly stated by Siéyes: "Il ne s'agit pas ici de recenser un scrutin démocratique, mais de proposer, d'écouter, de se concerter, de modifier son avis, enfin, de former en commun une volonté commune".58 Arguing along similar lines, several members of the Constituent Assembly attacked the idea that revisions of the constitution could take place at the initiative of local assemblies and individual petitions.59 A preference that has not been purified by public discussion is not a true preference ("un véritable voeu")60. Representative institutions were sharply distinguished from mere democracy, "ce qu'il y a dans la nature de plus odieux, de plus subversif, de plus nuisible au peuple lui-mme".61 The ideal of rational discussion thus enters twice into the debates of the Constituent Assembly. The professed goal of the members was to create legislative institutions that would foster consensus rather than vote counting and horse trading. And by the nature of that ideal, such institutions could only be established by similarly consensual means.

Constitutional bargaining

In this section I attempt to take the notion of constitutional bargaining in a (somewhat) non-metaphorical sense, by interpreting the debates in the light of modern bargaining theory.62 For reasons briefly indicated above, I shall rely exclusively on the Federal Convention. There were no doubt elements of bargaining and compromise in the Constituent Assembly. The abolition of feudal privileges was tempered by compensations which may have been subject to some bargaining. The single-chamber system with a suspensive veto for the King can be seen as a compromise between the full "three-headed hydra" and Siéyes's proposal of an entirely autonomous assembly. But the debates themselves do not, to my knowledge, indicate any such dealings.

Two bargaining compromises stand out in the Convention: the rule that a slave should be counted as three fifths of a free man for purposes of representation, and the mixed congressional system with proportional representation in one house and equal representation in the other. I shall limit myself to the latter issue, which was much more explicitly discussed in bargaining terms.

First, however, I need to say something about bargaining theory in general. Bargaining concerns the division of the benefits from cooperation, compared to a state-of-nature alternative. In the case of bargaining among separate states, this alternative is just the ordinary international order. In the case of bargaining among estates, among political parties or between civil and military institutions, the non-cooperative alternative is a state of permanent guerilla warfare.

Bargaining theory has two distinct branches. The most developed is two-person bargaining theory, as applied, for instance, to wage bargaining between capital and labor. The more general n-person theory is qualitatively different, since it has to take account of the possibility of sub-coalitions. In bargaining over government formation, for instance, the grand coalition that includes all parties almost never forms. And even when the grand coalition does form, as it did in the United States in 1788, the threat of sub-coalition formation can play a vital role in shaping the terms of agreement. For various reasons, n-person coalition problems are less tractable than two-person bargaining. We do not really have a good theoretical understanding of coalition formation. We can, however, use two-person theory to throw some light on the n-person case.

In the Federal Convention, we may talk as if the large states formed one actor and the small states another, and use two- person theory to understand the nature of the bargaining between them. But there were many other, cross-cutting divisions among the states that rivaled with the size issue as potential foci for coalition formation, the North-South division being the most important. Hence a full account would have to specify the payoff structure for all possible subcoalitions of the thirteen states, and to propose a theory that, on the basis of these payoffs, predicts that the grand coalition will form and the terms on which it will form. The first task is impossible for practical reasons, and the second is unresolved at the time of writing. This is why the following discussion offers at best a somewhat non-metaphorical use of bargaining theory.

Let us assume that the parties to the bargaining are rational and, more specifically, that they act to maximize material rewards. In that case, the outcome of bargaining is largely shaped by two factors. On the one hand, the outcome is constrained by the outside options of the parties, i.e. by the rewards they would obtain if the bargaining broke down and a permanent state of non-cooperation were to obtain. A rational agent will not accept an outcome that is worse than his outside option. On the other hand , the outcome will be determined by the inside options of the parties, i.e. by the resources available to them while the negotiations are going on. Typically, the inside options are inferior to the outside options. An agent will do better for himself if he breaks off the negotiation altogether and redeploys his resources than if he keeps negotiating while his resources are idling. Yet the differences in inside options may matter more for the outcome, because they determine which threats are credible and which are not.

Rather than elaborate on the formal definitions, let me explain the idea of outside and inside options by an illustration from wage bargaining. For the workers, the outside option is set by the wage they could obtain elsewhere or the level of unemployment benefits. Their inside option is set by the size of their strike fund any whatever other support they might receive during a strike. The outside option of the firm is set by the resale value of the plant, while its inside option is determined by fixed costs, inventory size and the like. Outside options constrain the wage agreement: neither the workers nor the firm will accept an outcome that is inferior to what they could get on their own. Inside options determine the credibility of threats, and hence how much the parties can get over and above their outside options. Intuitively, a party to the bargaining has an edge if his resources enable him to hold out longer than the other party.

This proposition throws light on Wilson's argument that the equality of states in the Confederation was due to "the urgent circumstances of the time"63 or to "necessity"64, and that the Convention ought to adopt proportional representation since "the situation of things is now a little altered".65 In a time of national danger, time is of the essence. No single state can better afford to hold out than any other; hence bargaining power is equalized. In periods of comparative calm, the larger and more self-sufficient states regain their natural bargaining advantage.

On two occasions Gouverneur Morris used a similar argument with respect to future contingencies. Arguing against the proposal that all bills for raising money should originate in the first branch, he referred to the following scenario: "Suppose an enemy at the door, and money instantly and absolutely necessary for repelling him, may not the popular branch avail itself of this duress, to extort concessions from the Senate destructive of the constitution itself?"66

This argument is more dubious than Wilson's , since the threat of the popular branch to withhold money would hardly be credible. On a later occasion a proposal was made to have a census at regular intervals for the purposes of adjusting representation. According to Madison's notes, "Mr.Govr. Morris opposed it as fettering the legislature too much. Advantage might be taken of in time of war or the apprehension of it, by new states to extort particular favors."67 This argument also seems implausible, for similar reasons.

Note, however, the difference between Wilson and Morris. Wilson, when reminding the smaller states that the circumstances which lent them bargaining power have changed, is himself engaged in bargaining. Morris is engaged in rational discussion, arguing that since certain future parties may use their bargaining power to obtain unfair concessions, they ought not to be given the rights from which that bargaining power would be derived. In fact, Morris explicitly asserts that he does not want to bargain. According to Madison's notes, "He wished gentlemen to extend their views beyond the present moment of time; beyond the narrow limits of place from which they derive their political origin. If he were to believe some things which he had heard, he should suppose that we were assembled to truck and bargain for our particular states. He cannot descend to think that any gentlemen are really actuated by these views. We must look forward to the effects of what we do."68 Yet in the very same intervention Morris shows himself to be very much engaged in bargaining, brandishing the threat that "if persuasion does not unite [the country], the sword will".69

Early in the convention Wilson made a brief observation ("ad terrorem", as Madison says in his notes) "to the idea of not suffering a disposition in the plurality of states to confederate anew on better principles, to be defeated by the inconsiderate or selfish opposition of a few states. He hoped the provision for ratifying would be put on such a footing as to admit of such a partial union, with a door open for the accession of the rest."70 The remark, which opens for the formation of subcoalitions, is left dangling. The next major intervention along similar lines is by Madison. Addressing himself to the smaller states who want equal representation,
He begged them to consider the situation in which they would remain in case their pertinacious adherence to an inadmissible plan should prevent the adoption of any plan. The contemplation of such an event was painful; but it would be prudent to submit to the task of examining it at a distance, that the means of escaping it might be the more readily embraced. Let the union of the States be dissolved and one of two consequences must happen. Either the states must remain individually independent and sovereign; or two or more confederacies must be formed among them. In the first event would the small states be more secure against the ambition and power of their larger neighbours, than they would be under a general Government pervading with equal energy every part of the empire, and having an equal interest in protecting every part against every other part? In the second, can the smaller expect that their larger neighbours would confederate with them on the principle of the present confederacy, which gives to each member an equal suffrage; or that they would exact less severe concessions from the smaller states than are proposed in the scheme [of proportional representation]?71
Here Madison is characterizing various outside options of the small states, without explicitly saying anything about those of the larger states. Nathaniel Gorham supplemented the argument in this respect:
The states as now confederated have no doubt a right to refuse to be consolidated, or to be formed into any new system. But he wished the small states which seemed most ready to object, to consider which are to give up most, they or the larger ones. He conceived that a rupture of the Union would be an event unhappy for all, but surely the large states would be least unable to take care of themselves, and to make connections with one another. The weak therefore were most interested in establishing some general system for maintaining order.72
Taken together, these passages claim that the small states are in a weak bargaining position because their outside options are much worse than those of the large states. This view corresponds to the theories of bargaining that were prevalent from 1950 until 1980.73 According to these theories, the bargaining agreement that will be reached is not simply constrained by the disagreement outcomes (the outside options), but actually determined by those outcomes, in a sense to be explained. Consider, for instance, the particular theory which says that the outcome of two-party bargaining will be one in which the gains to each party compared to the disagreement outcome are proportional to the maximal gains it could have obtained had the other party only received its disagreement payoff.


In the diagram, the area enclosed by the two axes and the curve are the feasible outcomes. Suppose that A is the outside option or disagreement outcome. The theory just stated predicts that the agreement will be C. If the disagreement outcome moves to B, making the second party's outside option worse, the agreement changes to D, also making the second party worse off. Other theories, which predict different outcomes, are also responsive to changes in the outside options, over and above their constraining e ffect. The passages from Madison and Gorham can similarly be understood as saying that because of their inferior disagreement outcomes the small states lack the bargaining power to get equal representation.

The argument sounds persuasive; yet the small states got their way. Could it be because Madison and Gorham were using the wrong bargaining theory, which places all emphasis on outside options and none at all on inside options? Wilson, as we saw, invokes inside options in his reference to urgency and necessity. From this line of reasoning we might conclude that the large states had an edge in 1787 because they could afford to hold out longer than the small states. Yet this argument was never explicitly made

Nor are there any signs that the delegates from the small states were more impatient to reach a conclusion than the large states. In fact, the duration of the Convention would hardly matter for bargaining strength unless the parties expected that it might go on for years, which they did not. Thus Wilson's argument, although correct in principle, had no bite in practice.

An alternative explanation of the victory of the small states could be that their representatives believed so strongly in their cause that they would rather suffer a loss than being treated unfairly; or, more realistically perhaps, that they were able to convey a belief of this kind to the representatives of the larger states.

We might ask, however, whether equal and proportional representation were the only two possibilities. Could not the large states have forced the small states to agree to an intermediary solution, in which they would have more senators than proportionality would give them, yet less than full equality would provide? Proposals of this kind were actually discussed at various times, although never in a very serious way. I can think of two reasons why they did not succeed.

First, equal representation and proportional representation are the only obvious focal points, with a natural salience that singles them out for special attention. Once proportional representation was ruled out, equality held out the best hope for agreement. Second, equal representation was also advanced on grounds of justice. To the extent that states - not only individuals - were seen to have rights, equality of representation was the inescapable conclusion. Bargaining power, psychological salience and conceptions of justice all enter into the explanation of the compromise in Philadelphia.

The effects of publicity

A crucial procedural difference between the two assemblies was, as noted, that one was open and the other closed. In this perspective, we may consider two different effects of public proceedings.

On the one hand, publicity, if properly structured, can have the desirable effects mentioned earlier. By forcing the speakers to argue in terms of the common good, it can "launder" preferences of their self-interested components.74 By contrast, closed proceedings tend to foster bargaining and explicit recognition of self-interest. The agreement that is eventually reached may reflect bargaining power, not the force of argument. In a constitution that is to last for centuries, a bias of this kind can do a great deal of harm.

On the other hand, publicity makes it difficult for speakers to back down from their positions. Mason noted that "a record of the opinions of members would be an obstacle to a change of them on conviction".75 A fortiori, the same would be true of public proceedings. This effect in turn induces another: people may use publicity as a precommitment device. This is well known from wage bargaining: both employers' associations and unions tend to make public demands so as to raise the stakes and make it more credible that they will not back down. This is the polarization effect referred to earlier. Also, speakers may use the public as a resonance board to terrorize their opponents into submission. This is the false consensus effect. These effects may prevent agreement from ever being reached, or, if an agreement is reached, cause it to have a bias in favor of the party who can command greater support in the audience. These considerations point to the need for very careful design of the constitutional setting itself. To design good institutions, the constituent assembly must itself be well designed.

Conclusion: Implications for Eastern Europe?

Although the paper so far has been concerned partly with general arguments, partly with the late eighteenth century, it was written very much with the late twentieth century in mind. The political transitions and constitutional debates that are currently taking place in Eastern Europe and the Soviet Union have a number of features in common with the American and French experiences. Let me suggest, very briefly, some of these similarities.

In Paris, as I said earlier (citing Clermont-Tonnerre), the constitutional debates took place under the double extra- constitutional pressures of the King and the Parisian mob. In Eastern Europe, the transitions to democracy, notably the Round Table Talks in East Germany, Poland and Hungary, took place under the double threat of Russian intervention and mass demonstrations. The more recent constitutional debates are also taking place under extra-parliamentary constraints, but not the same ones. On the one hand, authorities in the former Communist countries are constrained by their need for Western credit. On the other hand, they do not want to take measures that might provoke mass demonstrations or street violence. Some leaders will issue warnings that unless their motions are adopted, credits will not be forthcoming. (But the warning may actually be a threat, if these leaders can influence the policy of the lending countries.) Others will issue threats that unless their proposals are voted, mass strikes will paralyze the country. (But the threat may actually be a warning, if the masses are genuinely uncontrolable.)

In Philadelphia, the threat of extra-constitutional force came from the possibility of secession. Spokesmen of both small and large states at various times threatened to break out of the confederacy. This situation has close parallels in several countries in Eastern Europe, not to speak of the Soviet Union. In Yugoslavia, the issue of secession looms increasingly large. Note, however, that six republics can form coalitions in several hundred ways. Although some of these can be excluded a priori, because of traditional hostility, the number of remaining possibilities is dauntingly large. It is possible that bargaining theory might serve to exclude some of the remaining combinations; and to predict, at least in a qualitative manner, the terms on which coalitions will be formed. In the Czech and Slovak Republic, the issue of proportional versus equal representation of the republics - the issue over which the Federal Convention was almost torn apart - is highly salient. Although the Czechs outnumber the Slovaks two to one, the Slovaks demand and may well succeed in obtaining parity of power in many important federal organs. At the moment of writing, however, the possibility that there will not be any the Czech and Slovak federation at all, at most a loose confederacy, is widely discussed.

The question of the legitimacy and legitimacy-conferring capacity of the constituent assemblies arose both in the French and the American context, as we saw earlier. These issues, too, are present in Eastern Europe, most acutely perhaps in Poland. Here, the parliament that is also serving as constituent assembly was not the result of free elections, but of the hybrid electoral system that emerged from the Round Table Talks between Solidarity and the Communist Party. In Bulgaria the legitimacy of the Communist-dominated parliament is highly contested. In many countries, the tendency is for parliament also to be the constituent assembly. As I mentioned earlier, this duality creates an unfortunate, but perhaps unavoidable, tendency to reinforce the place of the legislative branch of government. A variant of the same problem exists in Hungary. Here, recognition of the fact that the governmental apparatus is too contaminated by the past to be trustworthy has led to a constitutional bias in favor of the legislative. But this creates another dilemma. Either the constitution is supposed to last indefinitely; and in that case future generations will be saddled with a system designed for the period of transition. Or the constitution is supposed to be merely transitory; and in that case it will not have its intended beneficial effects on private long-term planning, for which stability and predictability are main desiderata.

The two eighteenth century experiences suggest an uncomfortable choice between having the constitution drafted by constitutional subcommittees of parliament meeting in closed sessions, and having it drafted by constitutional assemblies meeting in plenary sessions. The goal would be to design a mixed procedure, which would prevent both horsetrading and demagoguery from dominating the debates.

A final, very tentative lesson from the late eighteenth century might be the following. The Americans designed their constitution on the worst-case assumption that human motivations will remain self-interested. The French believed that they could design a constitution that would change human motivations for the better. These two ideas need not be incompatible. But to design a constitution that will work only if it succeeds in changing motivations, is to court disaster.

* I thank Carlos Waisman and other seminar participants in the Center for Iberian and Latin American Studies at the University of California (San Diego) for valuable comments after a talk that eventually turned into this paper. I am also grateful to Bernard Manin, Russell Hardin, Stephen Holmes, Raino Malnes, Adam Przeworski and Cass Sunstein for their comments on an earlier draft.

1 The paper is based on M.Farrand (ed.), The Records of the Federal Convention, vols. I and II, New Haven: Yale University Press 1966. and on the speeches collected in F.Furet et R.Halévi (eds.), Orateurs de la Révolution franaise, vol.I: Les constituants, Paris Gallimard 1989, occasionally supplemented by the Archives Parlementaires.

2 Cp. Ch.I.5 of my Sour Grapes, Cambridge University Press 1983.

3 In the following I single out threats rather than promises as the main form of communication in bargaining contexts. For a discussion of the difference between these two ways of influencing the opponent see pp.272 ff. of my The Cement of Society, Cambridge University Press 1989.

4 The locus classicus is T.C.Schelling, The Strategy of Conflict, Cambridge, Mass.: Harvard University Press 1963.

5 See Ch.6 of The Cement of Society.

6 I owe this view of deliberation to Bernard Manin.

7 Farrand, vol.I, pp.179, 343. William Paterson (ibid., p.251) retorted that even granting that the original agreement was unfair, it is not clear that "the donor [can] resume his gift without the consent of the donee".

8 See notably the speak by Clermont-Tonnerre of 31.8.1789 (Les constituants, p.251-52 ) in which he urges the assembly to resist the pressure of the people as they had withstood that of the king.

9 In actual cases, wholly new constitutions are rarely made within a pre-existing constitutional framework. In theory, however, this might happen. Thus in the Constituent Assembly one delegate distinguished between the the need to create "un moyen de réformer partiellement la Constitution" and the need to have "un moyen légal de changer entiérement la Constitution" (Archives Parlementaires, vol.XXX, p.97). The recent rewriting of the Hungarian constitution was a case of complete change through patchwork incrementalism.

10 Paterson at one point asks rhetorically, "If a proportional representation be right, why do we not vote so here?" (Farrand, vol.I, p.250).

11 Farrand, vol.I, p.515.

12 E.g.Farrand, vol.I, pp.34, 249, 250.

13 Ibid., p.253.

14 Ibid., pp. 346, 255.

15 1.9.1789 (Les constituants, p.881).

16 1.9.1789 Les constituants, p.676).

17 4.9.1789 (Archives Parlementaires, vol.VIII, p.574).

18 29.5.1789 (Les constituants, p.350).

19 Farrand, vol.I, p.583.

20 Farrand, vol.II, p.309. The argument provides a perfect illustration of the idea that self-binding has dangers as well as advantages. For discussion, see my Solomonic Judgements, Cambridge University Press 1989, p.198-99.

21 Farrand, vol., p.49.

22 The arguments invoke a real veil of ignorance, not the hypothetical veil that is used to define "the original position" in Rawls's Theory of Justice. Rawls's argument is much closer to what I here call impartiality.

23 Farrand, vol.I, p.530.

24 Farrand, vol.II, p.3.

25 Ibid. Gouverneur Morris did not apply veil-of-ignorance reasoning to the new states. "Provision be made to prevent the maritime states from being hereafter outvoted by [the new states]. He thought this might easily be done by irrevocably fixing the number of representatives which the Atlantic states should respectively have, and the number which each new state will have. This would not be unjust, as the Western settlers would previously know the conditions on which they were to possess their l ands." (Farrand, vol.I, p.533.)

26 Farrand, vol.I, p.579.

27 Hence I shall disregard most of the issues raised in Charles Beard, An Economic Interpretation of the Constitution of the United States, reprinted with a new Introduction by Forrest McDonald, New York: The Free Press 1986.

28 I am indebted to Serge Kolm for this formulation.

29 26.2.1791 (Les constituants, p.492).

30 Farrand, vol.I, p.123.

31 Farrand, vol.I, p.379.

32 Farrand, vol.I, p.132. Similar fears were expressed by Ellsworth (ibid., p.374).

33 This responds to an argument made by Gary Jacobson.

34 Farrand, vol.I, p.159.

35 Farrand, vol.I, p.151.

36 Ibid.

37 11.8.1789 (Les constituants, p.44).

38 Farrand, vol.I, p.512; vol.II, p.202.

39 20-21 7. 1789 (Les constituants, p.1014).

40 Farrand, vol.II, p.p.210.

41 23.12,1789 (Les constituants, p.244).

42 See the speak by Barnave cited above, as well as the editorial comments in Les constituants, p.1194.

43 Farrand, vol.II, p.123.

44 Farrand, vol.II, p.124. A different and less insightful class analysis was proposed by Charles Pinckney, who distinguished between the landed interest, the monied interest and commerce (Farrand vol.I, p.401).

45 Farrand, vol.II, p.202.

46 Farrand, vol.II, p.124.

47 Farrand, vol.II, pp.31-32, 104, 250.

48 Essays: Moral, Political and Literary, Oxford University Press 1963, p.40.

49 See my "Envy in social life", in R.Zeckhauser (ed.), Strategic Reflections on Human Behavior, forthcoming from M.I.T.Press.

50 Farrand vol.I, pp.72, 181.

51 Farrand, vol.II, p.114.

52 Farrand, vol.II, p.110-11.

53 Farrand, vol.II, p.109.

54 17.8.1789 (Les constituants, p.114)

55 Op.cit., p.117.

56 See for instance Farrand, vol.II, p.33.

57 For this distinction, see my "The market and the forum", in J.Elster and A.Hylland (eds.), Foundations of Social Choice Theory, Cambridge University Press 1986.

58 7.9.1789 (Les constituants, p.1026). Italics added.

59 d'André 30.8.1791 (Archives Parlementaires, vol.XXX, p.68- 69) ; Barnave 31.8.1791 (Les constituants, p.54).

60 Barnave, loc.cit.

61 Ibid.

62 For a non-technical exposition of bargaining theory, see Ch.2 of The Cement of Society.

63 Farrand, vol.I, p.179.

64 Farrand, vol.I, p.343.

65 Ibid.

66 Farrand, vol.I, p.545.

67 Farrand, vol.I, p.571.

68 Farrand, vol.I, p.529.

69 Farrand, vol.I, p.530. Later (ibid., p.532) Hugh Williamson asserted that Morris's threat was only intended as a warning. Similarly, Gunning Bedford first issued a threat on behalf of the small states (ibid.,p.492) and later restated it as a mere warning (ibid., p.531).

70 Farrand, vol.I, p.123.

71 Farrand, vol.I, p.320-21.

72 Farrand, vol.I, p.462.

73 For brief descriptions of this theory and the theory that replaced it, see again Ch.2 of The Cement of Society.

74 Se R.Goodin, "Laundering preferences", in Elster and Hylland (eds.), op.cit.

75 Farrand, vol.I, p.10.

[Elster, Jon (1991), Arguing and bargaining in the Federal Convention and the Assemblée Constituante, Working Paper - University of Chicago, August 1991]
[This text was found on the Net. Since I have not seen the original document I cannot guarantee the correct reproduction of pargraphs, French letters and italics. There were no page numbers.]

[ Books | Articles | Reviews | Index | The main menu ]