Ethics & International Affairs Volume 18.3 (Winter 2004/2005): Articles: Interim Imposition [Full Text]

Dec 16, 2004

The entire article is posted below, or you can download the article as it appeared in print by clicking the "download" button at the bottom of the page [PDF: 263 KB, 26 pages].

[MacArthur] said that he had issued no orders or directives, and that he had limited himself merely to suggestions. . . . He stated that it was his belief, that it was his conviction, that a constitution, no matter how good, no matter how well written, forced upon the Japanese by bayonet would last just as long as bayonets were present, and he was certain that the moment force was withdrawn and the Japanese were left to their own devices they would get rid of that constitution.

—Recorded on January 29, 1946, by Nelson T. Johnson,
Secretary-General of the Far East Commission1

In spite of the storm surrounding its first appearance, the cumbersomely named "Law of Administration for the State of Iraq for the Transitional Period" (TAL)2 has been surprisingly immune from criticism in the West since its initial signing on March 8, 2004. American officials, anxious to declare victories where they can, as well as journalists seeking newsworthiness have insisted on the more accurate and revealing term "interim constitution." Its technocratic name, designed to neutralize (or hide) its constitutional significance, may partly explain why it has received little critical attention, but a more likely explanation is that many of its readers have rightly or wrongly viewed it as offering better protections for rights, including those of minorities and women, and more safeguards against new forms of authoritarian rule than other constitutions in Islamic countries, especially those in the Arab Middle East, including Iraq's own constitutional past. Commentators are apt to overlook the imposed character of the production of the document, perhaps because they suspect that a more genuinely negotiated and consensual product would very possibly have included fewer supposed protections for rights and safeguards against dictatorship, or at least the "tyranny of the majority."

This essay has two major aims. First, I will explore the political significance of interim constitutions and offer an account of the three features they must have if they are to be successful—namely, that they serve the anti-dictatorial impulse, provide greater legitimacy to the emerging political order, and facilitate constitutional learning. Second, I will examine the Iraqi interim constitution, arguing that it lacks all three of these features, thereby negating the very purpose of having an interim constitution in the first place. I will conclude, without excessive optimism, by suggesting some policy changes that could still be undertaken in Iraq in order to serve at least one of the purposes of interim constitutions—blocking the road to future dictatorships by promoting elements of constitutionalism compatible with electoral legitimacy and majority rule.
Reasons for Adopting Interim Constitutions

It is only since the préconstitution (Loi constitutionnelle) of November 2, 1945, that helped to establish the Fourth French Republic that interim constitutions have become widely used during changes of regime. Unfortunately, because of the sad history of that republic, the denunciations of its final constitution, and the frequent abuse of interim constitutions elsewhere (particularly in the Arab world), interim documents have had a relatively bad name until recently, when they played a dramatic and positive role in the Spanish, Hungarian, and South African changes of regime.3

Before examining the features that distinguish successful from unsuccessful interim constitutions, I will summarize the reasons why they have been adopted. Interim constitutions are often enacted because there is a generally accepted political need to organize constitutional government, possibly for an extended period, even in the absence of conditions under which a genuine and fully legitimate constitution can be made. This was the rationale for the interim constitutions in Germany and Japan after World War II, though only in the former was the new constitution, the Grundgesetz, pronounced (and remained, until 1990) interim.4 In both countries foreign occupation was understood to have confiscated sovereignty, and constitution making was interpreted as a key sovereign function. In Germany, the division of the country and the inability of the Soviet zone to participate in constitution making provided additional reasons to conceive the new arrangements as merely interim until reunification.5

Perhaps the most important reason interim constitutions have been adopted is the fear of dictatorship, the expectation that in an unrestrained revolutionary process one of the sides would dominate and might either make a provisional dictatorship permanent or impose its authoritarian constitution. This can lead to the aspiration to apply the principle of constitutionalism not only to the result but also to the process of making constitutions. In the European (originally French) tradition of revolutionary constitution making, constitutional assemblies invalidate all previously constituted powers and unite in themselves the plenitude of all powers.6 Carl Schmitt rightly called these constituent assemblies "sovereign dictators"—even if he was not generally right in tracing modern dictatorship as such to this institution alone.7 An equal, or more important, threat that motivates actors to use interim constitutions is that the institution of the provisional government will be legally unchecked before the meeting of the constituent assembly that justifies its existence. Such a government has a far better chance of transforming itself into a permanent dictatorship than any assembly. An interim constitution represents a fundamental device by which the constitutional assembly and the provisional government both are subjected to quasi-constitutional rules for the duration of their tenure, possibly limiting that tenure itself in time (and punctuated by a new election).

France's so-called préconstitution,8 for example, resulted from a popular referendum that called for a new constitutional assembly that was to found a new republic.9 At the same time as ordaining a new constituent process, this referendum enacted the Loi constitutionelle of 1945, which provided clear limits to both the constitutional assembly (a rigid timetable and the requirement of a ratificatory referendum for the final document) and the provisional government (some separation of powers in order to prevent a "convention government" on the model of the revolutionary government of 1793–95).10

The anti-dictatorial impulse may have an important role to play in enacting interim fundamental laws also in cases where there is legal continuity,11 such as when the "old regime" is either a dictatorship or a highly exclusionary polity. In almost all such cases, from Spain to Hungary and South Africa, one can document the efforts of previous power holders who can no longer avoid some kind of transition from authoritarian rule to create from above a "hard democratic" or "soft authoritarian" regime in which their social, economic, personal, and even political power positions are protected or favorably transformed. Elections held under their rules are likely to produce undemocratic results, and no democratic oppositions can or should accept such imposed solutions, at least if they can help it. At the same time, democratic oppositions may be unable or unwilling to try the path of revolutionary transformation, as was indeed the case from the 1970s to 1990s. Roundtable negotiations are characteristically held to resolve this dilemma, and they typically result in interim constitutions.12 In such negotiations, the opposition can protect itself from attempts of old regime actors to maintain power, while these actors can in turn gain guarantees that a revolutionary dictatorship that would treat them as enemies of the new regime will not take place. Given the complexity and potential vulnerability of agreements that would achieve these aims, the actors generally seek ways of guaranteeing the survival of key parts of the interim constitution in the final document. Thus, while the French type of préconstitution involved only procedural limits on the constituent assembly,13 the new interim constitutions seek to impose limits on its contents, along with procedural safeguards that are likely to preserve agreements regarding contents. An example of substantive limits are the famous South African 34 Constitutional Principles, requiring the constitutional assembly to provide for constitutional government, democratic and inclusive citizenship, federalism, minority rights, as well as special rights for traditional communities. An example of procedural limits is the assignment of final constitution making to a normal parliament working under a qualified two-thirds majority amendment rule both in Hungary and in South Africa (where a document supported by simple majority could still pass if supported in a referendum by 60 percent of the voters). In both cases there is an institution safeguarding an interim constitution "worthy of defense"—a new and powerful constitutional court.

In these contexts, there are two (or more) sides that could potentially impose a constitution, and the purpose of the interim constitution is to fashion a "second best" option that each can live with since they are not able to achieve their preferred solutions. Here, too, the anti-dictatorial or anti-imposition logic leads each side to desire (however reluctantly) that constitutionalist norms be imposed both on the process and the results of constitution making.
Interim Constitutions and Legitimacy

The roundtables produced detailed constitutions that were to limit democratically elected constitution makers in terms of the substance and not just the procedures of their activity. Because they had not been democratically elected, such a significant role for the roundtables could have posed a challenge to exclusively liberal, constitutionalist justifications. Indeed, the roundtables dramatically wrestled with problems of their own legitimacy, seeking a plausible response to the deficit other than the cheerful acceptance of being in the position to impose. Where there is no democracy, a purely democratic beginning of democracy is impossible.14

The key to resolving these problems of legitimacy lies in two stages permitted by the interim constitution itself. During the first stage, normative principles other than electoral authorization and accountability must be rigorously honored to make up for the democratic deficit. The most important of these principles are plurality, publicity, and legality. Adhering to these principles provides what Jon Elster has referred to as "upstream legitimacy."15 If these principles are not adhered to, the normative integrity of the whole constituent process becomes difficult (though not impossible) to maintain. The interim constitution must then in turn organize a second stage allowing the rigorous fulfillment of the technical preconditions of democratic elections without the possibility of deformation by those organizing existing or provisional executive powers. Herein lies the "midstream legitimacy" of interim constitutions, which is supposed to compensate for inevitable weaknesses of upstream legitimacy.16 Whereas upstream legitimacy depends upon the procedures of enacting interim constitutions, midstream legitimacy is more a function of its design, including its procedures. Extreme lack of upstream legitimacy may, however, interfere with the successful functioning of even relatively good constitutional designs (and thus the generation of midstream legitimacy), and it is for this reason that I stress the importance of plurality, publicity, and legality in establishing interim constitutions.

Here I can give only a short summary of the relevant principles. By plurality I mean the requirement of drawing in as broad a range of significant political participants into the negotiating process as possible, and by using a decision rule that allows all of their voices to count. By definition, inclusion on this level cannot be perfect or complete, since it will involve some group(s) choosing (and rejecting) others as partners when none of them has been tested electorally. But it is not difficult for external observers to determine whether the main cleavages of a society are "represented" by groups having some genuine organization and support, and when in doubt their advice should certainly be relied on to complete the process of inclusion of at least the main groups. Publicity too cannot be complete or perfect if there are to be genuine negotiations. But providing for a sufficient number of public forums to present the state of the negotiations at various levels, and sufficient time for the public to absorb information about them, certainly forces actors (as in 1787 already) to adopt many positions that can be justified by using arguments that can appeal to groups across different particular interest and value constellations (in Rawls's terms, "public regarding arguments").17 Finally, legality under dictatorship may be fictional, but when participants in roundtables take law seriously it shows that their work is rule constrained and is never the product of the arbitrary wills of even a plurality of actors.18 Elements of genuine legitimacy (as well as the mutual promises and commitment of major political actors) allow for the interim constitution to limit the sovereign, unlimited powers of the constitutional legislature that would typically be only a "convention" or even an ordinary parliament, not a (classical European) constituent assembly. At the same time, the incomplete legitimacy of the interim constitution requires that the restrictions on a freely elected body be as limited as possible under the circumstances.

In a former dictatorship, one can rarely hope for ideal democratic elections in the second stage. In most cases the greatest threat to genuine and free competition are undemocratic incumbents. Interim constitutions as well as other legal arrangements, such as electoral rules, party rules, and disposition over public media, must be so devised as to inhibit the holders of executive power from manipulating, deforming, and falsifying the elections. Of course, when the threats are multiple, coming from both government and an insurrection, or from different actors in a civil war, it may be impossible to produce or enforce measures restraining the repressive apparatus of the executive. Pseudo-elections may be worse than no elections at all, and the longer elections are postponed, the more important "interim" arrangements legitimated by a complex of justifiable principles become.

An interim constitution by its nature implies a constitutional learning process involving two fundamental stages. This claim is itself based on learning from the common historical experience of many constitutions (the U.S. Constitution, the French Constitution of the Fifth Republic, the Grundgesetz) that emerged after the dramatic failure of a recent forerunner, whose problems can become the occasions for important learning experiences and corrective efforts. While the same cannot be said about minimal preconstitutions such as the one adopted in France, the more detailed interim constitutions clearly attempt to institutionalize this kind of learning. Only a detailed constitution allows learning over a wide enough range. But only when that constitution is interim can it organize learning within a single process, making it less likely that specific constitutional problems lead to revolutionary, totalizing breaks such as the one that occurred in France in 1792. If successful, they also make it less likely that experiments in less radical revision lead to replacement of all earlier achievements along with genuine problems, resulting in the replacement of the second constitution by a third, then a fourth, and so on, as occurred in France and many Latin American countries. The learning advantage of an interim constitution is that political formulas that are indispensable in the short term but very questionable in the long term (like consociationalism or rigid power sharing or great coalitions) can be included in interim constitutions, if carefully planned, without the fear of their transposition and insulation in the permanent document.19 Thus the new type of interim constitutions have allowed—as they must—a broad learning process to take place across two interlinked constitutions.

There are many constitutional problems that learning across two constitutions can address. Most obviously, there is the possibility of incoherent drafting, where one article should be changed to make it consistent with another or the rest of the constitution. More important and difficult are problems where there is no logical incompatibility but the actual functioning of rules leads to undesirable results. Here the attachment of beneficiaries to a type of malfunctioning such as dramatic and unfair overrepresentation will be a problem more difficult to handle with a normal amendment rule giving some veto powers to minorities. Conditions of democratic transition may require a very high threshold of consensus, for example through consociational devices, that within a final constitution would have a self-freezing character. In the longer term they may make a country ungovernable, and prepare the ground for future civil war, given inevitable changes in the original power and demographic constellation. The South African approach of mandating consociationalism in the first stage while not enshrining it in the second stage, allowing the freely elected assembly to move instead to constitutional protection of minorities, has been particularly successful in this regard. Only a two-stage process involving a relatively detailed interim constitution allows for this clever (temporary) use of consociationalism without provoking a revolt by the majority.

There are important prerequisites for constitutional learning to take place. Since experimentation must occur during the beginning of the process, a considerably more flexible amendment rule for the interim constitution must be established than would be appropriate for the permanent constitution in the same sociopolitical setting.20 More importantly, beyond the provisions or principles consensually agreed upon by the main participants, interim constitutions must not interfere with the learning processes involved in making the permanent constitution.21 No interim constitution could violate its own purpose more obviously than one that through its own rules both inhibited learning and tended to make rules of ratification for a permanent constitution that are destined to fail. While some interim constitutions may legitimately become permanent (as in Hungary) when a relatively open final process repeatedly fails, we face a serious abuse and attempts at self-perpetuation when the rules provided by the interim constitution are so constructed as to make the failure of the final constitution extremely likely. Finally, for an interim constitution to work as a legitimate anti-dictatorial and learning instrument, it must be enforceable (its learning must involve learning not to learn too easily!). In Hungary and South Africa at least, the provisions of the interim constitutions have been effectively enforced by strong constitutional courts.
Iraq's Interim Constitution

Detailed interim constitutions that significantly restrict a constitutional assembly or convention are most at home in negotiated or coordinated transitions or regime changes, such as those that occurred between 1976 and 1995. Such transitions generally involve roundtables, which in turn presuppose legal continuity in the midst of the severe crisis and even rupture of old regime legitimacy.22 Because the revolutionary process in Iraq involves rupture of both legality and legitimacy, it would seem a rather unlikely case for the adoption of the new type of formula (of a detailed interim constitution that substantively limits the constitutional assembly).23 The adoption of a French type of limited preconstitution, with the goal of restraining both provisional government and constituent assembly, would seem to be a more likely prospect. But it needs to be pointed out that in France the provisional government that initiated the project derived its authority from the resistance both at home and abroad, and that this obviously legitimate provisional government nevertheless sought and received popular, electoral authorization for its initial constitutional acts. Neither of these types of legitimation, revolutionary and electoral, were available to the authors that imposed the Iraqi interim constitution. Thus, under weaker initial conditions it is all the more striking that the type of interim constitution and restrictions of the constitutional assembly that were chosen in Iraq resemble more the ambition of the fully negotiated products of the late 1980s and 1990s than the relatively modest form of self-binding that occurred in France in the 1940s. Paradoxically, the process was one of imposition, but the result resembles a thoroughly negotiated settlement.

Nevertheless, an interim constitution, which may in any case be generally desirable on normative grounds, had some plausibility in a country of externally imposed revolution such as Iraq under the combined circumstances of weaker legitimacy and the particular constraints and opportunities offered by international law. Compared to indigenous revolutions carried out by an internally supported elite, Iraq's externally imposed revolution has weaker conditions of transitional legitimacy but is framed by obligations of an unbroken international legality. It is this double difference from indigenous revolutions that, curiously, made the adoption of a detailed interim constitution desirable and even likely.

Without much legitimacy, the American occupiers sought initially, as do all revolutionaries, to impose a permanent constitution. Given the conditions of legality as provided by international conventions, however, this goal could not be accomplished by an external imposition, as initially imagined. In addition, criticism of this intention was voiced from inside Iraq by the Grand Ayatollah Ali al-Sistani: "These [occupation] authorities do not have the authority to appoint the members of the constitution writing council. There must be general elections in which each eligible Iraqi can choose his representative in a constituent assembly for writing the constitution. This is to be followed by a general referendum on the constitution approved by the constituent assembly."24 An interim constitution provided the solution, but only a relatively detailed variant imposing limits on the future constituent assembly could satisfy simultaneously the political goals and the legal requirements. The question is whether under such circumstances an interim constitution of this kind could be successful without some of the political institutions that could sustain it, or whether instead it would become reduced to a purely strategic instrument that could not even succeed in constitutionalizing the process of constitution making, a goal that even minimal interim constitutions can accomplish.

In all revolutions the legitimacy of the revolutionary elite feeds on a double source: its act of liberation from a hated old regime and its promise to establish a far better new one.25 Initially only the first claim can be tested. In the case of Iraq's externally imposed revolution, however, there is both a chronic lag of legitimacy that neither indigenous revolutionary agents nor external liberators from foreign occupiers have to face. Here the liberators appear from the outset as occupiers. Their agents are agents of an occupying power. Repeated attempts have to be made therefore to manufacture legitimacy by inevitably awkward public relations, by pictorially representing Iraqi society in ethnically and religiously "balanced" consultative bodies, by trying to gain the support of respected public figures, by activating traditional structures like tribes, and by trying to create the aura of constitutionalism in an interim constitution. In the end, of course, nothing short of free elections will succeed in ensuring legitimacy. Democratic elections, however, if really free and competitive, may be won by forces most able to challenge and confront an occupation considered illegitimate. While it is relatively easy to exclude political forces from participating in various co-opted and satellite instances, it is much more difficult to exclude parties and groupings from contesting free elections if they are willing to play by the electoral rules. Thus, the one source of legitimacy that may justify the occupation as liberation could lead to its final denunciation.26 Without elections, however, the bodies that would produce interim rules of the game would have a built-in legitimacy problem. In countries such as Hungary and South Africa, it was possible to turn this disadvantage into a constitutional advantage by organizing a two-stage process. A two-stage process was eventually also adopted in Iraq, but (I will argue) without the supplementary forms of legitimacy that would be needed to sustain the process.

The external relations of states are governed by international law, and in an externally imposed revolution there is thus not the complete legal rupture that characterizes indigenous revolutions. Especially since the signing of the Hague Conventions, as affirmed by the Geneva Conventions and Security Council Resolutions 1483, 1511, and 1546, the occupation of Iraq is a legally regulated matter. Law establishes both opportunities and limits for occupying powers. So long as the occupying power is willing to play by these rules and make use of the institutions and offices of the world body (at least to some extent), this can partially compensate for missing legitimacy at the beginning of the constitution-making effort. International legality can authorize at least some of the components of the constitution-making process, including the drawing up of interim legal arrangements. It is indisputable, however, that the available form of legality, unlike the legal continuity of transitions from Spain to South Africa, also interferes with constitution making by the occupying power. This was recognized in Japan by both MacArthur's legal staff and Japanese experts.27 Specifically, international law seems to be predisposed against constitutional imposition by an occupying power and perforce its domestic agents. According to Article 43 of the 1907 Hague Convention, "The authority of the legitimate power having actually passed into the hands of the occupant, the latter shall take all steps in his power to re-establish and insure, as far as possible, public order and safety [civil life], while respecting, unless absolutely prevented, the laws in force in the country."28 According to many international lawyers this is still valid law in spite of the usual problems with enforcement and the possibility that the Security Council may permit what the treaty has banned: a significant role for the occupying power in constitution making.29 "The objective overall is not to change constitutional arrangements . . . but to protect them," according to The Responsibility to Protect. Reading these lines we cannot tell which meaning of "constitutional" the authors have in mind: regime (after all, the cause of intervention) or constitutionalism (something that did not exist in such cases and would have to be established).30 Yet even in the case of authoritarian regimes overthrown from the outside, case law seems to bear out the claim that the Hague obligation, which pertains to all regimes, not only constitutional ones, remains valid.31

In Japan, the restraints of international law were "adhered to" and the ban against external imposition was formally evaded by keeping the actual constitution-making process a carefully guarded secret and allowing the legally elected, last imperial Diet to enact formally the entirely new constitution according to the amendment rule of the old 1889 Meiji Constitution. Unlike Japan and even Afghanistan, there was, however, no parliament with traditional legitimacy in Iraq. A detailed type of interim constitution was therefore intended to supply the solution. If there was to be imposition, it could be first answered that it would be Iraqis themselves who would impose the (American appointed) Governing Council. When that answer predictably did not satisfy, it could be claimed that only an interim set of rules would be imposed. This would not change Iraq's legal order in an illegal manner, because the changes would not be permanent since the TAL leaves the task of final constitution making to a freely elected assembly.

The strategic reasons for introducing the Iraqi interim constitution are clear. And to the extent that this particular interim constitution was intended to become permanent in many respects, its creators can also rightly be accused of seeking to avoid international obligations through subterfuge. Even if all that is true, it remains an open question whether Iraq's interim constitution possesses the features that could make it a justified and successful initiative as a provisional basic law seeking to block dictatorship and promote legitimacy and learning. In the following section I will argue that on the whole it does not.
The Failure of LegitimationSeparating liberation and constitution in the manner of Arendt's concept of revolution, one could believe with some (though incomplete) justification that in Saddam's Iraq liberation had to be externally imposed. Moreover, it could also be believed in good faith that the initial imposition of the beginning of a democratic process of some kind perhaps could not have been avoided in a country like Iraq, full of undemocratic scenarios. The illegitimacy of Iraq's interim constitution is due to the manner in which it was openly imposed by the U.S.-led Coalition Provisional Authority (CPA) in early March 2003, at a time when many other scenarios were imaginable and imagined by, among others, the officials of the United Nations.32 It was a key step in a series of imposed solutions that included the establishment of the Governing Council in the summer, the November 15 "agreement" in the fall of 2003, and the choice of membership of the provisional government in May 2004.33 Not only was there lack of legitimacy, but also a spectacular process of delegitimation with the Grand Ayatollah Ali al-Sistani as its main protagonist. In the end, Sistani managed through clever maneuvering to block the legitimation of the interim constitution by international law, but only by accepting a very poorly regulated provisional government as the guardian of the process of preparing for free elections.34

The process of constitution making reappears in its structure of authority. Especially in the case of a new constitution, its legal validity is contingent upon legitimate origins. Why should this rather than some other document carry the obligation to be obeyed? The power to impose cannot alone decide this question. In the case of Iraq's interim constitution, three methods were pursued to secure valid authorization: Iraqization, Security Council authorization, and political agreement. This would seem to have been the right approach, but in the hands of the U.S. authorities each was just another alibi for imposition. At each stage, the American authorities rejected proposals for the administration of Iraq (proposed by the special representative of the secretary-general, Sergio Vieira de Mello, the director of the Office of Reconstruction and Humanitarian Assistance, General Jay Garner (Ret.), and others) that would have made the process more autonomous, pluralistic, and consultative.


It is difficult to escape the impression that Brent Scowcroft, the national security advisor of the George H. W. Bush administration, was speaking the collective mind of the government at least when he maintained that the United States could not have gone to war (especially since there were no WMDs to be found!) to deliver power in this strategic country to its ideological enemies, the clients of Iran. Thus no early elections were permissible.35 Accordingly, if the American authorities sought not a determined outcome, at the very least they wanted to have outcomes within a determined range (in particular a government that could be represented in the United States as a significant improvement on the old regime, not least because it would be open to American influence and perhaps even some kind of long-term presence.) The uncertainty that is necessary for the democratic process could not be permitted. And certainty could be achieved only if the autonomy and pluralism of the participating elites were dramatically reduced.

Whatever their intentions, the American authorities deluded themselves into thinking that it was entirely unproblematic to envision an Iraqi transition in which constitution making is the task of agents appointed by, responsible to, and accountable to them alone. The attempt to represent the imposition as having been carried out by an Iraqi authority failed. Indeed, this approach made little sense in a world accustomed to the hierarchical world of the civil law, where each legal jurisdiction gains its authority from a higher one.36 It was thus obvious to all that whether drafted by the Governing Council, its constitutional committee, or a new body co-opted by these institutions, the authority of the constitution would derive either from the CPA, which appointed these bodies, or from the body that appointed the CPA, established its jurisdiction and prerogatives, and reserved the right to revoke it at any time. The rights of the Governing Council and its offshoots are a function only of the rights of the occupying authority itself. And as we have seen, the CPA did not have the right, even in its own view, to impose upon Iraq a constitution.

At the same time the CPA never managed to actually stay out of the supposedly Iraqized process. In appearance, the interim constitution was a product of the Governing Council and its Iraqi experts. In reality, the CPA and its leader, Paul Bremer, played a decisive role in the process. When the concessions to Islamic law threatened to become embarrassing for the Americans, Bremer threatened to veto the relevant provisions. Many members of the Governing Council complained of the couplike, accelerated manner in which the whole draft was pushed through to meet American (presumably electorally conditioned) deadlines. All outside participation was minimized, and after the leaking of some very early drafts complete secrecy was preserved. The final draft was entirely unavailable until the actual signing, and was thus exposed to no criticism until the five Shiite members of the Governing Council threatened not to sign the document. This threat was itself apparently neutralized when the CPA made clear that without immediate signing, the transfer of sovereignty envisioned for June 30 (that occurred on June 28) would be in jeopardy and the nonsigners would have to take the responsibility for the continued occupation.37

Security Council Authorization

The CPA's hierarchical and operational authority had to be shored up, and this could be done only through its gaining international legal legitimacy. While there remains some doubt that even a Security Council resolution could free a country from the obligations of an occupying power under Article 43 of the Hague Conventions, there seems to be an emerging consensus among international lawyers to this effect.38 Thus it was not insignificant that while confirming the Hague Regulations in general by "reaffirming the right of the Iraqi people to freely determine their own political future," the U.S.-drafted Security Resolution 1511 seemed to accept, if ambiguously, an earlier project of constitutional imposition by the CPA through the agency of the Governing Council. The text of the resolution reads: "Welcoming the decision of the Governing Council to form a preparatory constitutional committee to prepare a constitutional conference that will draft a constitution that will embody the aspirations of the Iraqi people. . . ."39 It is important to note, however, that the same resolution does not assume that the Governing Council is sovereign. Instead, very unclearly, it states that the Governing Council "without prejudice to its further evolution, embodies the sovereignty of the state of Iraq"—which seems to mean that the Governing Council symbolizes a sovereignty that it cannot exercise, while what the CPA exercises is short of sovereign powers. This is confirmed by the long subsequent discussion of the restoration of sovereignty on June 30. Since constitution making is preeminently a sovereign function, supporters of the interim constitution felt that it was necessary to seek another Security Council resolution that explicitly authorized it and the process of its production. This was prevented, however, by Sistani, with the result that the international authorization of the TAL never took place.

Political Agreement

Legitimation is by its nature linked to the rule of some over others, and we can agree with Max Weber that neither pure imposition nor pure agreement can be the starting points of a legitimate new political order.40 Any agreement, no matter how wide, must still be imposed against members who do not agree. Two issues should be distinguished in this context: the degree of external imposition and the degree of internal pluralism. It is now rightly customary to distinguish between the imposed character of the post–World War II constitution-making process in Japan and the mainly autonomous, deliberative process in Germany, which was ultimately based on freely elected provincial assemblies.41 The case of Germany shows that it is possible to have a low degree of imposition and a high level of pluralism under foreign occupation. In Iraq, the (illogical, but self-serving) attitude of the CPA and its legal advisors seems to have been that since there is imposition anyway, the United States might as well impose its choice of a bilateral structure of negotiation and select entirely on its own the actors "representing" a society incapable of generating its own politics.

This assumption was based on a fundamental misjudgment both of the importance of pluralism and the emerging organization of Iraqi politics. Of course, the pluralism of a society can never be mirrored by political process, and pluralistic agreement always has a dimension of exclusion and imposition. Yet so long as it incorporates genuine opposition, pluralistic politics simulates and even promotes the pluralism of society. Even a partial pluralism in politics stimulates critique, accountability, and the need to justify decisions with good arguments and valid claims. Finally, while later agreements cannot take away the fact of initial imposition, they can diminish its negative impact on those who were previously excluded. These arguments are especially relevant to the constitution-making process when the rules of the political game are being decided. Political participation in such a process ought to be as broad as possible and should in principle expand. The worst-case scenario is when the pluralism of society is developing much faster than that of the framework of representation. Under such circumstances, societal pluralism leads actors to negotiate outside the formal framework of bargaining and compromise, even engaging in armed struggle. This is quite clearly what happened in Iraq.

The level of political pluralism of Iraqi society at the time of the formation of the Governing Council is debatable. But it is astonishing that since then, during the last year and a half, there has been no attempt to expand participation in spite of the fact that many new political actors emerged. Against the recommendations of the first report of the special representative of the secretary-general, Lakhdar Brahimi,42 neither the November 15 agreement nor the drafting of the interim constitution involved an extension of inclusion or participation. In the Governing Council, the principle of representation followed was an attempt to picture the ethnic, religious, and gender divisions of society—an idea more suitable to a supreme soviet than a genuine deliberative assembly. Moreover, important actors such as the Arab nationalist parties, the more radical wing of the Shia, and even many important Sunni religious groups were excluded from this body. Apparently, the only concession Vieira de Mello achieved, with his great diplomatic skill, was the inclusion of the Iraqi Communist Party. While the Kurdish members of the Governing Council do represent the major political forces of the Kurd provinces, this is less true for the Shia members taken as a whole and is not at all the case for the Sunnis. All of the latter and some of the former are exiles representing different interests and factions of the occupying authorities. The Governing Council thus had little "upstream" legitimacy in Iraq, and its scarcely disguised attempts to hold on to power before, during, and after the transfer of sovereignty under an astonishing variety of formulas further decreased this already weak legitimacy.

Can the Iraqi interim constitution's lack of upstream legitimacy be compensated for by "midstream" legitimacy? As in all cases involving a two-stage process, such substitution would depend on the successful and legitimate execution of the electoral process—provided in this case by the TAL and the electoral rule produced by UN advisors.43 There have been two major roadblocks to the achievement of electoral midstream legitimacy. One is the Sunni insurrection, which the Americans have insisted on dealing with through primarily repressive means. It is evidently not possible to hold elections where there is an open insurrection, and the chances of fully suppressing it in the Sunni triangle are very small, especially given the strategic obtuseness of the Americans, who, after having disbanded the army, left the rest of the Sunni elite more or less outside the whole process of political bargaining. Thus hard-liners and moderates in that community are continually driven together, feeding the insurrection with new recruits even as some are militarily defeated. Now, after the siege of Fallujah, the chances of an open electoral boycott will increase, even if ultimately some will participate given the possible benefits. Given the TAL's unamendable rule and Sistani's very strong pressure, postponement of elections is not an option, even if the current provisional government would love to continue to function indefinitely as such. Given the proportional representational rule with one national district, Sunni Arabs would thus be dramatically underrepresented if elections were held under anything resembling the current conditions, simply because their total vote would be depressed.44 And a constitutional assembly operating without one of the key national-religious groupings, the Sunni Arabs, could not be legitimate in the eyes of those excluded.

The second roadblock is what might be called the "Shiite paradox": the only way to stop Shia religious parties from getting a big majority in the new assembly is through fraud, the construction of a governmental list supported by the media, intimidation by the United States, or all three of these measures combined. Under such conditions, the election would not produce a legitimate result. If, however, the Shia religious parties are allowed to win in a clean election, midstream legitimacy would be achieved, but the force that would control the assembly might very well wish to repudiate the TAL. Legitimacy can thus only be generated for the process at the cost of possibly jeopardizing the TAL itself, and along with it the possibility of an assembly practicing significant self-limitation.
The Inhibition of Learning

The Iraqi interim constitution accomplishes remarkably badly the second main purpose of interim constitutions—to facilitate constitutional learning during an experimental period before a permanent constitution is instituted that is insulated against easy alteration.45 There are at present three distinct rules of change in or associated with the TAL. Of these, two (1 and 3 below) are mentioned in the document, and a third (2) follows from the language:

  1. A good part of the TAL is amendable by a vote of three-quarters of the National Assembly and the unanimous consent of the three-member presidential council (Article 3A).46 There are, however, extensive unamendable provisions: rights covered under chapter 2, the time frame of the transition as defined by the interim constitution, the powers of regions and governorates, and regulations having to do with Islam and religions. After free elections, this self-referring rule can be used to change itself, before any other part of the constitution is changed through the new rule.47
  2. Since there is no National Assembly until free elections can be held between late December and January of 2005, the interim constitution is by implication unchangeable for the period from June 28 to sometime after the elections, when the new National Assembly first meets.
  3. The TAL states only that the National Assembly "shall write a draft of the constitution of Iraq" and that this draft will be "presented for approval in a popular referendum." We don't know by what majority the assembly must agree on a draft, and if there are possible vetoes. The most convenient way of reading the text is that the draft of the permanent constitution fully replacing the interim one must be approved by 50 percent + 1 vote of the National Assembly,48 and then (as is clearly stated) by 50 percent + 1 vote of the population in a national referendum as long as two-thirds of the voters of three governorates do not vote against ratification (Articles 60 and 61). Having to do with governorates, this rule may have been meant to be unchangeable, but the restriction would be useless since the amendment rule is not enshrined.

As to the second "rule," which is not stated formally, it is clear that preventing changes during the period from June 28, 2004, to, say, January 31, 2005, means extreme rigidity for at least seven potentially crucial and difficult months. If elections were postponed this would be true even longer, except for the fact that the rule does not allow the postponement of elections for which the TAL has set a date on or before January 31, 2005 (Article 2B(2)). How problematic this rigidity is depends on the asymmetric conditions of legitimacy already discussed, as well as issues of content that I will touch on very briefly. If there is a lot that is wrong with the regulations that were hastily drawn up under the aegis of narrow political inclusion, if the current unregulated status of the provisional government proves extremely troublesome, or if conflictual actors like the Kurds and the Shiites move toward a new modus vivendi, as they very well might while still under an "electoral veil of ignorance," it is highly likely that the no amendment rule would be politically challenged by forces having more and broader support than the authors of the document had or will have.49 The same would also be true if there was a move to postpone elections after all.

Things become even more problematic in the period when amendments become possible (rule 1) and when the permanent constitution is to be drafted and ratified (rule 3). These two rules together constitute extreme consociational limits on the changing of the interim constitution. Theoretically, three-quarters of the National Assembly (with the agreement of all three members of the Presidency Council) can actually pass amendments to the interim document, except its unamendable sections, during the potentially extended period that the permanent constitution is being drafted. Practically, the representatives of any of the three ethno-religious groups (Shia Arab, Sunni Arab, Sunni Kurd) are likely to have over 25 percent of the seats, and possibly, if that convention holds, one member of the Presidency Council. Thus any of them can veto any amendment twice over. It is very likely that Arab nationalist, strongly secular, or Baghdad deputies, deputies from oil-rich regions, or whatever other combination will also have 25 percent of the votes in the assembly. Thus the possibility for vetoes of amendments would be greater than what today we could imagine by focusing on existing power constellations. Of course, some amendments may still pass through bargaining and compromise. But when an important amendment is blocked in the face of either three-quarters of the deputies and all three members of the presidium or 95 percent of the deputies, along with the president and one vice-president, a "runaway convention" that refuses to be bound by the will of the illegitimate Governing Council or the foreign CPA would become very likely.

Finally, and most seriously, the much discussed issue of a veto of a draft of a permanent constitution by three provinces or governorates would not only increase the rigidity of the overall constitutional setup but could very easily transform an interim constitution that is already very difficult to amend into an indefinite or even a quasi-permanent constitution. The vote of a simple majority passing the constitution would be too easy a requirement, since only drafts that have significant parliamentary consensus should be submitted to the population. Given the possibility of provincial vetoes, one can assume that the National Assembly would internally generate a consensual rule that could be (and probably will but should not be) as high as the 80 percent hurdle in West Germany (where the legislatures of at least five of the fourteen LŠnder could have vetoed the Grundgesetz). If, though unmentioned, there is the possibility of executive veto, its use by one member of the Presidency Council or presidium would be too high a consensus requirement, but a veto would not be unreasonable if the presidium unanimously (and perhaps a majority of two of its members) were against a draft. What is completely unreasonable is the possibility that three small provinces out of nineteen (15.8 percent) (and about 10–11 percent of the voters) would have a veto power over a constitutional draft.

Can the rules of change themselves be changed within this rather sketchy system of rules? After June 28 the provisional government in this "phase one" has no amendment powers under the interim constitution. The TAL and the TAL Annex confirm this, since they do not give any legislative powers to this "sovereign" entity. In my view one cannot plausibly deny it the revolutionary option (however unrealistic) to repudiate the TAL as a product of a foreign force with no constituent powers in international or Iraqi law, or as a result of a coerced agreement between that force and its Iraqi agent. While no one knows where the authority to do this originates (aside from well-established authoritarian practice in many countries, unfortunately), the TAL could also be set aside or suspended (in whole or in part) for the country as a whole or for specific territories. A new declaration of martial law, state of emergency, or an order of legal force by Allawi could very well have such result.

In summary, the issue of constitutional learning fares very poorly in the TAL because it is basically unamendable in phase one, partially unamendable in phase two, and its amendment rule and the ratification rule of the new constitution are extremely restrictive. In the absence of a strong and activist constitutional review, those seeking necessary constitutional change will tend to engage in radical extralegal, and legally speaking, revolutionary options (including coups and autogolpe), that threaten to undermine what was learned in the interim constitution. Was this failure of design inevitable? The answer is again a strong no. This time, however, the blame lies only indirectly (and not only) with the U.S. authorities. Quite obviously, without even knowing the details of the negotiations, it was the Kurds who sought to enshrine through various consociational vetoes (amendment rule, ratification rule, even a partial nullification of federal laws option, as in Article 54B) the federal and regional autonomy or quasi independence achieved through the negotiations leading to the TAL that in fact legalized de facto arrangements. Because it was a question of both de facto arrangements that would be defended with arms if necessary and their most stable and enthusiastic allies, the Americans supported the Kurds until almost the end (they did not accede to Kurdish demands to seek Security Council authorization for the whole rigid TAL package).50 The Kurds (and their foreign advisors), however, made the fundamental mistake of wishing to have multiple guarantees to protect the same autonomy provisions that were probably not even under any great threat from the Shiite majority. They failed to realize that they could have tried to enshrine only the parts of the interim constitution by a more difficult amendment rule, along with the amendment rule itself. As to the final constitution, they needed to agree only on some unchangeable constitutional principles protecting regional and federal autonomy, rather than holding the whole permanent constitution hostage to their veto.
The Failure of Constitutionalism?

Assuming that an interim constitution fails in its more ambitious tasks it may nevertheless succeed in imposing legal limits on constitutional assemblies and provisional governments, thus applying constitutionalism to the period and process of constitution making itself. Unfortunately, the Iraqi interim constitution will not easily achieve even this goal. This is not only because of its much weaker form of authorization than, for example, the French préconstitution of 1945, but because of its design failures. It is a combination of weak authorization (or "upstream legitimacy") and an overly ambitious design that rigidly ties the hands of the freely elected assembly, making it likely that a process of unbinding will occur once the assembly meets. This unbinding will be achieved either through agreement and formal amendment or through subversion or open repudiation. A successful majoritarian repudiation of the TAL would lead in principle to the sovereign dictatorship model of constitution making. Freely elected national (constitutional) assemblies regularly emancipate themselves from their conveners, and such a revolutionary act is much more likely when the latter are regarded as illegitimate.51

Thus, by not including authorization for the TAL in Security Council Resolution 1546, the U.S. authorities pretty much conceded Sistani's demand that the elected National Assembly not be bound by that interim constitution. Brahimi has firmly supported this view.52 He maintains that the freely elected (sovereign!) assembly cannot be bound by interim arrangements, which can become law only to the extent that the new assembly reenacts them. If this view prevails, it is reasonable to assume that the assembly could then replace both the amendment rule and the ratification rule as it sees fit, before rewriting or entirely replacing the TAL. Moreover, it could do so through a simple majority rule (unless it chooses to establish a different one, again by majority rule). Of course such a procedure would violate not only the text but also the spirit of the interim constitution, which would be entirely eviscerated if this revolutionary scenario were followed. To avoid such majoritarian implications, Brahimi quotes Bremer: "Iraqi unity requires a constitution that all of Iraq's communities can support. It is a fundamental principle of democracy that the constitution should provide for majority rule but also protect minority rights."53 This, however, refers to the desirable result of moving from consociationalism to constitutional protection of minority rights. But in the case of an unbound assembly this outcome can only be the result of voluntary self-limitation. With the interim constitution losing its binding power over the freely elected assembly, an important opportunity would be lost—even as democratic legitimacy was being regained. For the Kurds, who might thereby lose all their guarantees, these results would amount to a call for secession. In any case, this scenario would mean the complete failure of the minimal, constitutionalist rationale for having an interim constitution. If a Shiite-dominated assembly became fully unbound, the most likely constitutional project on the Iraqi horizon would be based on undiluted majoritarian popular sovereignty, with its obvious attendant dangers of religious dictatorship, civil war, or both.

Of course, the constitutional assembly could still voluntarily enact and adhere to a set of new limits for itself and the government issuing from it. These limits would be easy to break, however, and, even if they were not, the achievement of constitutionalist restraints would belong to the self-binding assembly, its majority, and above all Sistani, rather than to the TAL and its authors. Even with the best intentions, Sistani and his supporters could successfully sponsor a new set of (self-) limitations only if these could be first thoroughly renegotiated in a binding and solemn agreement of the relevant parties. Assuming (rather implausibly) some kind of highly desirable agreement among the main social forces by this time, even a legal-constitutionalist scenario without a formal repudiation of the TAL is at least imaginable. The TAL, its amendment and/or ratification rules could all be amended legally as the interim constitution itself allows in principle changing the two relevant rules. Since, due to a likely omission, neither the amendment rule (1) nor the ratification rule (3) are enshrined, or insulated against amendments, it is possible to amend either to require votes by qualified majorities by using rule 1. While it is unreasonable to assume that exactly the same minorities will vote to give up their own minority veto, and thus amend the amendment rule itself, the same is not true regarding the ratification rule. Although unlikely, it is not unimaginable, for example, that 75 percent of the assembly—which need not necessarily include any Kurdish deputies—with the support of a Kurdish vice-president chosen by two-thirds of the assembly, thus possibly by the Shiite and other nationalist deputies, would abolish the rule allowing three Kurdish provinces to veto a constitutional draft in favor of a majoritarian referendum. That Kurdish vice-president, since he would wish to stay alive, would probably want to trade his vote for a suitable agreement on federalist protections, cultural autonomy wherever Kurds live, and oil revenue distribution.

Given the possibility of either failure to produce a permanent constitution or a veto of a proposed draft by three governorates, the existence of the interim constitution may itself become a significant liability for the constitution-making process and a potential rallying point for forces seeking to maintain the status quo of the transitional period, whether that is defined in terms of extreme provincial autonomy, the power of political incumbents, or both. So far, for example, the incumbents of the Governing Council managed to survive in the Interim Government formed on June 1, in spite of Brahimi's efforts to prevent it. They remained dominant in the National Conference held in July and in the undemocratic, single-slate elections there for a National Council.54 If after well-controlled, supposedly free elections, they similarly manage to carry over their power into the new National Assembly (or at least the executive power that is based on it), they may prefer to act under the interim constitution than under any new document that could emerge out of the National Assembly. Those who choose to subvert a popular permanent constitution would not be facing a situation in which they would have to accept, in case of failure, either chaos or the open dictatorship of the constitutional assembly. These forces (as well as the voters of the three provinces) could instead rely on the survival of an undemocratic interim constitution that may in fact better coincide with their interests.

The interim constitution in fact facilitates this possibility (which would be available anyway unless a sunset clause were to have been attached to it) by providing for its own survival, if a permanent constitution were rejected, for once in sufficient and rather unambiguous detail (Article 60E). While the unsuccessful National Assembly would in that case be dissolved, a new one would be elected, and operating under the interim constitution, with all dates altered to keep to the original time frames of a one-year term, possibly extended for six more months. There is no stated limit concerning how often the same process would have to be repeated in case of repeated failure. Along with the possibility of six-month extensions of the tenure of a National Assembly (Article 60F), the interim constitution could become the framework for a system whose incumbents would be changed every eighteen months. It is impossible to know whether the actual government of the country in such a system would even operate on the provincial or regional level—which might be the preference of the forces that every eighteen months or so reject a new permanent constitution of a more centralized system.
Failures of Restraint?

If, however, the TAL can neither be successfully repudiated nor used through its own negotiated self-amendment to enact a final constitution—the result being instead total impotence and breakdown and the long-term survival of the interim arrangements—the constitutionalist restraints could not be considered to have been successful, since their purpose is to channel rather than fully inhibit action. Whether this failure of constitutionalism would mean the triumph of dictatorship would depend upon whether the TAL amounts to a set of plausible limits and restraints on the provisional government that would presumably stay in power until replaced by the national constitutional assembly.

But which provisional government would stay in power? There are three possibilities. The first is the "government of fact" that exists in the case of every revolutionary scenario but should go out of existence with the adoption of an interim constitution.55 Revolutionary governments by definition take power outside the legality of the previous regime.56 In the Iraqi instance the CPA, along with its satellite, the Governing Council, was this government of fact, and its empowerment by the Security Council under international law did not subject it to restraints that could be called constitutional. The enforcement of the Geneva Conventions is a matter left to the discretion of the occupying power itself, and its obligation not to change the laws of the land is escaped in large part by the interim constitution itself. To be sure, when the TAL came into effect, the CPA was dissolved. Nevertheless, the American-led coalition's armed forces remain under U.S. command, and the limits established by the TAL do not apply to this force.57 The government of fact—the U.S.-led coalition—continues to exist in spite of the existence of an interim constitution, and all internal acts of repression emanating from the "sovereign" government depend on its voluntary acquiescence. With respect to this provisional government the TAL has not even made the slightest attempt to provide plausible restraints at all; the rights contained therein are not rights against the actual holders of the main means of violence.

Further, the interim constitution establishes two (or more) provisional governments, not one. The first ("phase one"), regulated by the TAL and its belated Annex, is to be in power until its actual replacement by a freely elected national constitutional assembly according to rules in the original TAL (Article 2B). The second ("phase two") is to be regulated by more extensive institutions and rules provided by the TAL. The third, and so on forever, if the constitutional assembly fails to produce a ratified new constitution, leading to the election of a new one.

There are two very different structures involved. The first provisional government under the TAL Annex is clearly a dictatorship formally speaking, at least where it is backed by the requisite military power of the "government of fact."58 This is due to features of the interim constitution itself. For example, there is an almost complete dominance of the executive and no real separation of powers is provided for. The executive itself is admittedly dual: in order to issue orders with the force of law (Annex, sec. 2), the prime minister needs the approval of all three members of the Presidency Council. The new courts provided for by the TAL (along with the Supreme Court capable of constitutional review of legislation) will be established supposedly in this phase, but the latter has a weak structure to begin with and was not clearly given powers of review of the acts of the executive and, in particular, "orders with the force of law."59 (There is to be no legislation in phase one!) The National Council chosen by a National Conference in elections totally controlled by the provisional government is only consultative except for the very unlikely possibility of a two-thirds repudiation of orders of legal force (Annex, sec. 3). The only real restraint to the executive is provided by the element of federalism, which is de jure in the first phase only for the Kurdish region (Annex, sec. 2). Obviously wherever other de facto powers can check the Provisional Government, it can exercise no dictatorship. It is a peculiarity of the situation that the American government of fact and its forces struggling to regain control do so for the benefit of a dictatorship, admittedly often to the expense of even worse ones.

Further, the TAL and its Annex say nothing about emergency rules and the suspension or bypassing of their own provisions. This is a very serious matter because the TAL does have valuable rights that could benefit not only open insurrectionists but also political opponents. Thus the regulation of inevitable emergencies should have been a high priority under the given conditions. In fact an emergency order of legal force has been already promulgated (Order of Safeguarding National Security, July 6, 2004), one that to be sure declares that the TAL cannot be abrogated "in whole or part" (Article 11); all the while a variety of rights granted by that document (assembly and association, Article 13C; movement, Article 13D; demonstration, Article 13E; personal security and due process rights, Article 15B–I; property, Article 16) are potentially taken away or suspended (Article 3). States of emergency can be declared by the prime minister, with the consent of all three members of the Presidency Council, for sixty days, and can be renewed for thirty days as often as the four can agree to do so. At the time of the promulgation of the order there was not yet a National Council whose two-thirds vote could have conceivably (but not in reality) vetoed it, and the order itself gives no role to this consultative body in the declaration or renewal of emergencies.60

All this is done on the basis of the Annex to the TAL permitting the issuance of governmental "orders with the force of law that remain in effect until rescinded by future Iraqi governments," which actually violates the same Annex's ban against actions affecting the country's "destiny beyond the limited interim period." (Note that a future such order with the force of law could reverse the earlier support for the TAL just as easily!) Thus the move puts executive orders on the same level as constitutional legislation, which is the legal essence of modern dictatorship. In fact that dictatorial potential was greatly enhanced by the final regulation issued by Bremer (CPA Order #100, June 28, 2004), according to which the prime minister of the interim government inherits all the formidable and unchecked powers of the CPA (sec. 2, 1–2) and the Council of Ministers is identified as the "body vested with national legislative powers" (sec. 2, 10). While it is added that these powers are to be exercised under law (including the TAL), it was indeed one of the CPA's powers to promulgate, abrogate, or change the interim constitution.

Finally, between the TAL and its Annex there are no provisions concerning the removal of the powerful head of government—the prime minister in phase one—which is an amazing omission for at least an eight-month period and maybe more.61 Thus if the legal position of dictatorship were turned into a politically recognizable dictatorship—characterized by endless states of emergency and the postponement or open manipulation of the electoral process—very little could be done (legally speaking) to stop Allawi in the first phase. There is no court to declare his actions illegal, and no instance capable of removing him. Thus, armed with the legitimacy bestowed by Security Council Resolution 1546, the new Iraqi executive could also emancipate itself from the interim constitution, all the more easily because this could be done by simply reducing the TAL to a paper constitution of the type well known in the Arab world and especially Iraq.62 Again—draw the conclusion—this provisional government too is insufficiently constrained by the TAL.

If the elections, however, do take place as the TAL ordains (Articles 2B, 3A), the road is open in principle to the provisional government of phase two. Assuming an electoral victory for a side that wishes to abrogate the TAL, the Shiite majority, it is also logical to assume that the National Assembly would not proceed to form a new provisional government under it, thereby indicating its tacit adherence to the interim constitution. The danger would then be that the provisional government of phase one would stay in power. Because of this possibility, Allawi (who has indicated acceptance of the TAL only up to elections) may also support its repudiation. For a majority that did not want him, or his type of provisional government, or a situation of dual power and potential clash between that government and the assembly, it might then be better to choose one of three perhaps less logical options: first to allow the formation of a new provisional government and then repudiate the TAL; to repudiate it, but quickly renegotiate and repass its parts allowing the formation of a new government; or to renegotiate and to amend it legally—the best possible, though perhaps most difficult, solution.

If any of these three options were carried out successfully—and it is a very big if—then some of the real virtues of the TAL's current contents could come into play. The country would have real separation of powers between a genuinely representative National Assembly (chapter 4) and the executive based on it, with the Supreme Court, too, playing a more important role given the fact that there is legislation instead of executive "orders with legal power." Yet the governmental structure potentially yields a strong executive organized around a prime minister with parliamentary responsibility (Article 38A). The constitutionalist checks, including rights (chapter 2), surviving from the TAL could now have a real role. The federal and regional self-government provisions (chapter 8), with the exception of the annoying right of regional nullification, are workable, and in the current state of Iraq probably mandatory. Even if a new constitution were not adopted because of conflicts in the assembly (as, for example, in Hungary), and the TAL somehow survived, in this scenario Iraq could hobble on with some kind of constitutional or quasi-constitutional "provisional" government that owed its power to elections, however imperfect. It would of course be best if the main forces of the National Assembly would be able to produce a negotiated compromise not only around amending the TAL but also in drafting and enacting a genuine, new constitutional contract that would no longer have the taint of American imposition.

The various imaginable steps to constitutionalism and a provisional government under constitutional restraints remain highly unlikely. Two much less appealing scenarios are more probable. The first is that the power of incumbency will be used by the Allawi government to control elections by intimidation, falsifying results, or most likely by creating a unified governmental slate that, supported by the media and the United States, would get the preponderance of votes.63 The elections and the formation of a new provisional government would then continue the process by which the Governing Council was transmogrified as the Allawi government, and produced a co-opted National Council. Whether a National Assembly so elected would enact a new constitution or guarantee the survival of the incumbents under a continually renewed interim constitution would be immaterial: either would be nothing but a paper constitution of an oligarchic dictatorship. The second is the first scenario defeated by the mobilized Shiite majority, which would then probably bring Moqtada al-Sadr (already one of the most popular politicians in Iraq) and his followers to power rather than the Shiite SCIRI or Dawa parties in Allawi's governmental coalition (and may be still parts of his projected list). It is not easy to see such a force capable of self-limitation unless Sistani were to lead it, something that he has been very reluctant to do in spite of his repeated disappointment with the moderate Shiite politicians.64

I began writing about Iraq after an invasion I completely opposed, hoping that there was a small window of opportunity nevertheless for democratic change. I have argued here that this opening has become smaller and smaller because of the character of the U.S. imposition and its consequences. And yet the possibility of democratic change may still exist. We should therefore be on the watch concerning the Allawi government's likely efforts to convert a legal position of dictatorship into political dictatorship. And we should try to promote three things: First, the holding of free and broad-based elections in January, which also implies finding serious political "partners" in the Sunni heartland, who cannot emerge if the United States insists on the purely repressive military option. After Fallujah, the repression of the Sunni should be replaced by a far more differentiated policy. Second, the reemergence of an independent and unified Shiite movement capable of both defeating electoral manipulation, along with a governmental slate, and of self-limitation. The name of this option is, and has always been, Sistani! And, third, once a freely elected National Assembly meets, the replacement of the TAL, in two steps, one producing an interim formula for a nondictatorial provisional government, with the checks and balances as well as the governmental structure from the TAL, and another a historical compromise over the permanent constitution.

Can a disastrous policy of illegally invading and occupying a distant country without a legitimate casus belli nevertheless have some good as its unintended consequence? Yes, but one should not generally count on it. However, in the present case, the only remaining justification of the war—the establishment of a democratic and constitutional regime in Iraq—whether sincere or not, can be used by the Iraqis against their occupiers. Ironically, this justification became increasingly important as the process of external, undemocratic, authoritarian imposition unfolded, creating the performative contradiction of the American occupation regime. It is this contradiction that the Ayatollah Sistani has tried to exploit ever since his first fatwa calling for free elections, and even the interim constitution targeting electoral legitimacy is its product. A critique of the interim imposition that takes this document seriously, and calls for a much better process to arise from its internal tensions, makes political sense, because of this contradiction whose source lies deep in American history. We are likely to encounter it again, in new forms, until it is resolved, one way or the other.

1 Cited in Koseki Shoichi, The Birth of Japan's Postwar Constitution, trans. Ray A. Moore (Boulder, Colo.: Westview Press, 1997), pp. 75–76. Was it already a misrepresentation at the moment (January 29, 1946) this was uttered? It is proven by by Koseki that three days later, on February 1, Gen Courtney Whitney was already advising MacArthur to the contrary, and it was this advice that was followed. Yet the statement remains valid for Iraq today! [BACK]

2 Available at [BACK]

3 Interim constitutions in the Middle East often signified illegitimate attempts to make supposedly temporary authoritarian and/or paper constitutions permanent or semipermanent. The last glaring example was the Iraqi interim constitution of 1970 that lasted formally, though without much meaning, for thirty-four years, until the American overthrow of Saddam Hussein in 2003. See Nathan J. Brown, Constitutions in a Nonconstitutional World: Arab Basic Laws and the Prospects for Accountable Government (New York: SUNY Press, 2001), p. 70 (for Syria); p. 79 (for Egypt); and pp. 86Ð87 (for Iraq). I am grateful to this careful and serious author, on whose work I have relied here. [BACK]

4 To my knowledge it was Social Democratic lawyers during the earliest phases of German constitution making who first offered the nomenclature of "administrative statute" (instead of constitution) to indicate not only provisionality but also the inadequacy of making a constitution-like set of rules for an occupied, hence nonsovereign (and, in Germany's case, divided), country. At that time the lawyers of the three occupying governments, the United States, France, and Britain, explicitly declared their hostility to such modest language, which would devalue their own achievement, and demanded a constituent assembly producing a constitution, as in France and Italy previously. The compromise formula was the so-called Parliamentary Council drafting a Basic Law. Peter Merkl, The Origin of the West German Republic (New York: Oxford University Press, 1963), pp. 52–54. [BACK]

5Shoichi, The Birth of Japan's Postwar Constitution, pp. 33, 64, and 70, mentions the proposal for a two-stage process of constitution making in Japan that involves a provisional basic law two years before the issue came up in Germany. [BACK]

6 It was formulated this way in the early writings of Sieyès and developed with some inconsistencies through the first two French constitutional assemblies of 1789–91 and 1792–95, and most clearly that of 1848. [BACK]

7Carl Schmitt, Die Diktatur (Berlin: Duncker and Humblot, 1922); and Carl Schmitt, Verfassungslehre (Berlin: Duncker and Humblot, 1928), p. 59. [BACK]

8 See Liberation Government, "Ordonnance no 45-1836 du 17 août 1945"; available at; and Provisional Government of the Republic, "Loi constitutionnelle du 2 novembre 1945"; available at [BACK]

9 Chosen on the same ballot as the referendum, in the case of a negative answer to the first question of the referendum the assembly would have been an ordinary legislative body under the Third Republic. [BACK]

10 It is demonstrable that Charles de Gaulle deliberately fought for limits to the provisional government. The outcome was a brief, eight-article interim constitution. It was innovative, because here the interim constitution had been independently authorized by the same source as the new assembly, the popular vote. See Michael Troper, Francis Hamon, and Georges Burdeau, Droit Constitutionnel 25th ed. (Paris: LGDJ, 1997), pp. 368–74; Olivier Duhamel, Droit Constitutionnel, vol. 2 (Paris: Seuil, 2000), pp. 140–41; Jacques Godechot, ed., Les Constitutions de la France depuis 1789 (Paris: Flammarion, 1995), esp. quoting de Gaulle's press conference with its three options, pp. 358–59; and Olivier Beaud, La puissance de l'état (Paris: PUF, 1994), pp. 272–76. [BACK]

11 Legal continuity did not exist in the case of France because the overthrow of the Vichy government was "revolutionary" in the legal sense, and even the return to the Third Republic (in the case of a "no" vote on the first referendum question) would have implied a revolutionary restoration. [BACK]

12 Neither is imperative: democratically inclined incumbents, as in Spain, can enact fully competitive interim rules; informal agreements may substitute for roundtables, as in Slovenia; and internal as well as external pressure can be used to enforce legislation allowing democratic elections, as in the German Democratic Republic, where the roundtable designed a permanent constitution that was never enacted. [BACK]

13 Beaud, La puissance de l'état, p. 275. [BACK]

14 For discussion, see Andrew Arato, "Forms of Constitution Making and Theories of Democracy," in Civil Society, Constitution, and Legitimacy (Lanham, Md.: Rowman & Littlefield, 2000), pp. 229–56. [BACK]

15 Jon Elster, "Constitutional Bootstrapping in Philadelphia and Paris," in Michael Rosenfeld, ed., Constitutionalism, Identity, Difference, and Legitimacy: Theoretical Perspectives (Durham, N.C.: Duke University Press, 1994), p. 102; and Jon Elster, "Arguing and Bargaining in the Federal Convention and the Assemblée Constituante," in Raino Malnes and Arild Underdal, eds., Rationality and Institutions (Oslo: Universitetsforlaget, 1992). [BACK]

16 In contrast to Elster, who was examining single-stage processes when coining the distinction, I treat electoral legitimacy at the beginning of the second stage as "midstream," while for him this would be "upstream" with respect to the central process. In countries of interim constitution, at least recently, the forging of that constitution represents the beginning of the stream. Like Elster, I leave downstream legitimacy to final enactment and/or ratification processes, but I do not deal with it here since it is irrelevant to interim constitutions in general. [BACK]

17 See Elster, "Arguing and Bargaining." I agree that the demand for publicity at all stages would be counterproductive, dangerous, and even impossible, and yet would stress the importance of public discussion at some key junctures. [BACK]

18 In my view this way of proceeding is actually a first step in building a genuine rule of law. See Arato, "Constitution and Continuity in the East European Transitions," in Civil Society, Constitution, and Legitimacy, pp. 167–98. [BACK]

19 If these arrangements are to work, they cannot be exposed to a generally flexible amendment rule. "Sunset" provisions are able to limit the longevity of mutual guarantees that are desirable only in the initial phase of the transition. During the operation of the interim constitution, actors in a divided society can learn to interact politically and to seek guarantees that are more compatible with majority rule and the freedom of the constitutional legislature—as in South Africa, where constitutionalism replaced consociationalism. [BACK]

20 Such an amendment process, however, should not detract from the legitimacy of authorizing the interim constitution itself, as happened in Hungary when the old, non–freely elected parliament reneged on some of the roundtable agreements. See Andrew Arato and Zoltan Miklosi, "Constitution-Making in Hungary" (unpublished, 2002; available at the United States Institute of Peace or from the authors.) [BACK]

21 Indeed, the one great danger of interim constitutions even in nonauthoritarian settings is that they work too well and make themselves permanent not through the free choice of a democratic assembly but by dramatically interfering with that choice. Rien ne dure que le provisoire! (Only the provisional lasts!) it was said soon after the making of the emphatically provisional Grundgesetz that is still, fifty-five years later, Germany's valid constitution. Even when there is no interference with the freedom of a future assembly, as in Hungary, the absence of any provisions (rules, incentives, disincentives) for making the permanent constitution can lead to the interim becoming permanent. See Arato, "Refurbishing the Legitimacy of the New Regime: Constitution-Making Endgame in Hungary and Poland," in Civil Society, Constitution, and Legitimacy, pp. 199–228; and Arato and Miklosi, "Constitution-Making in Hungary." Beyond its own amendment rule and sunset provisions, it is therefore extremely important for the interim constitution to regulate in a plausible way the time frame and the procedures for making the permanent constitution. [BACK]

22 See Janos Kis, "Between Reform and Revolution," Constellations 1, no. 3 (1995), pp. 399–422; and Andrew Arato, "The Occupation of Iraq and the Difficult Transition from Dictatorship," Constellations 10, no. 3 (2003), pp. 408–24. [BACK]

23 Nevertheless, I have argued for the potential relevance of this model to Iraq in several articles, most recently in "Sistani v. Bush: Constitutional Politics in Iraq," Constellations 11, no. 2 (2004), pp. 174–92; and "Constitution-making in Iraq," Dissent (Spring 2004); available at [BACK]

24 Ali al-Husaini al-Sistani, June 25, 2003; translation by Juan Cole in Informed Comment, I have slightly altered the terminology, with his agreement. For versions of what was planned, see Noah Feldman, "Democracy, Closer Every Day," New York Times, September 24, 2003, p. A27; Steven Weisman, "Powell Gives Iraq 6 Months to Write New Constitution," New York Times, September 26, 2003, p. A1; and Patrick E. Tyler, "Iraqi Groups Badly Divided Over How to Draft a Charter," New York Times, September 30, 2003, p. A14. [BACK]

25 Hannah Arendt, On Revolution (London: Faber and Faber, 1963). [BACK]

26 This is why electoral legitimacy was so often eschewed even on the local level, even in the south, where there were no security reasons in 2003 against holding elections. See David Leigh, "General Sacked by Bush Says He Wanted Early Elections," Guardian, March 18, 2004; available at,,1171689,00.html. [BACK]

27 Shoichi, The Birth of Japan's Postwar Constitution, pp. 90ff. [BACK]

28 The Avalon Project at Yale Law School, Laws of War: Laws and Customs of War on Land (Hague II), July 29, 1899 (Washington, D.C.: Government Printing Office, 1968). [BACK]

29 See Simon Chesterman, "Occupation as Liberation: International Humanitarian Law and Regime Change," pp. 51–64, this journal. [BACK]

30 ICISS, The Responsibility to Protect (Ottawa: IDRC, 2001), p. 25. [BACK]

31 I focus only on the case involving the United States. Already in Japan in 1945 the American occupiers recognized that their self-imposed tasks were not easily consistent with international treaties outlining the goals of foreign occupation. These presupposed either outright annexation or the establishment merely of a sympathetic government to be the only goals of foreign occupation. Nevertheless, it has been documented that in Japan in 1946–47 the Hague Convention played a strong role in the formal passing of the constitutional draft in spite of the fact that the document was originally produced and imposed by MacArthur's Government Section sitting "as a constitutional convention." See Shoichi, The Birth of Japan's Postwar Constitution, pp. 79, 92–93. [BACK]

32 See Jamal Benomar, "Constitution-Making after Conflict: Lessons for Iraq," Journal of Democracy 15, no. 2 (2004), pp. 81–95. [BACK]

33 For Brahimi's thoughts on the process, see Terence Neilan, "U.N. Envoy Urges Iraqis to Give New Leaders a Chance," New York Times, June 2, 2004, online ed. "Mr. Brahimi struck a mildly surprising note when, in answer to a reporter's question, he referred to . . . L. Paul Bremer III, as 'the dictator of Iraq.' 'He has the money,' he said. 'He has the signature. Nothing happens without his agreement.'" [BACK]

34 In a letter to the UN secretary-general he wrote: "It has reached us that some are attempting to insert a mention of what they call 'The Law for the Administration of the Iraqi State in the Transitional Period' into the new UN Security Council resolution on Iraq—with the goal of lending it international legitimacy. This 'Law,' which was legislated by an unelected council in the shadow of Occupation, and with direct influence from it, binds the national parliament, which it has been decided will be elected at the beginning of the new Christian year for the purpose of passing a permanent constitution for Iraq. This matter contravenes the laws, and most children of the Iraqi people reject it." See "UN Resolution Passes Unanimously," Informed Comment,, June 9, 2004; and John F. Burns, "Shiite Ayatollah Is Warning U.N. Against Endorsing Charter Sponsored by U.S.," New York Times, March 23, 2004, p. A8. Sistani had his way, for now—SC Res. 1546 made no reference to the TAL. [BACK]

35 Sistani never bought the argument that elections were not possible before setting up the first legal provisional government. There is a lot of documentation that the Iraqi census experts agreed with him last winter, and the idea of using ration cards for registration was floated already then. [BACK]

36 See Hans Kelsen, General Theory of Law and State, trans. Anders Wedburg (Cambridge: Harvard University Press, 1945).[BACK]

37 There are also reports that the Kurdish members—unopposed by the CPA—threatened to withdraw from the whole process. It is very likely that both threats were bluffs, because President Bush had the American electorate, and the Kurds, Turkey, to worry about, if they carried them out. But even the Shia members had a lot at stake in the continued relevance of the Governing Council, and possibly in the governmental formula they expected to emerge if the interim constitution was signed. [BACK]

38 See Chesterman, "Occupation as Liberation." [BACK]

39 Note, however, that this language does not decide how the "constitutional conference" would be selected, and hence does not exclude an elected body. It speaks only of "drafting" and not of "promulgating" or "enacting" without, e.g., popular ratification. As far as the preparation of the constitutional conference was concerned, the resolution called for "national dialogue and consensus building," something that did not happen during the secretive drafting of the interim constitution. [BACK]

40 Max Weber, Economy and Society: An Outline of Interpretive Sociology (Berkeley: University of California Press, 1978). [BACK]

41 See Hideo Otake, "Two Contrasting Constitutions in the Postwar World: The Making of the Japanese and West German Constitutions," in Yoichi Higuchi, ed., Five Decades of Constitutionalism in Japanese Society (Tokyo: University of Tokyo Press, 2001), pp. 43–72. And yet, imposition worked in Japan only because MacArthur started out with a fundamental concession to the Japanese ruling elite: the survival of the "symbolic emperor" and the formal continuity of the state and the legal order. In the Iraqi case the American occupation authorities seek to impose a new arrangement without any such fundamental concession to either an electoral (as in Germany) or traditional (as in Japan) principle of legitimacy. [BACK]

42 "The Political Transition in Iraq: Report of the Fact-Finding Mission," UN Doc S/2004/140, (February 23, 2004); available at [BACK]

43 I leave aside the co-opted National Conference of July and the scandalous single-slate election organized within that body for a consultative National Council. See note 57 below. [BACK]

44 The UN's mistake lay not in choosing a proportional representation system, which is most suitable for constitutional assemblies, but rather in its choice of one national list instead of provincial ones. The latter would have provided for adequate Sunni representation even with a low turnout. [BACK]

45 Niklas Luhmann, in A Sociological Theory of Law (London: Routledge and Kegan Paul, 1985), wrote of normative learning "not to learn." This dimension is especially important for constitutions if one is to have the two-track structure of constitutionalism rightly stressed in Bruce Ackerman, We the People (Cambridge: Harvard University Press, 1991). Learning constantly would threaten to dissolve the line between constitutional and normal politics. Nevertheless, under a new constitution there must also be an opportunity to correct obvious deficiencies unanticipated by the framers, as in the case of the U.S. election of the president/vice-president on a single ballot, corrected by the 12th Amendment of 1804. In this sense, an interim constitution properly constructed extends the two-track structure to constitution making itself by providing for normal rather than extraordinary alteration for a period of time. See Andrew Arato, "Constitutional Learning," in Civil Társadalom, Forradalom es Alkotmány (Budapest: Mandátum Press, 2000); with a new and expanded version forthcoming in Theoria, where I draw on the competing perspectives of Stephen Holmes and Bruce Ackerman. [BACK]

46 In the case of ordinary laws, unanimity of the council is required for a veto (Articles 36 and 37). [BACK]

47 Since the makers of the interim constitution neglected the elementary requirement to enshrine the amendment rule if they wished to make anything else unchangeable, everything in the TAL can be changed after free elections legally, using a two-step procedure. The same mistake was made by the authors of Article V of the U.S. Constitution, but then no one knew if self-referring rules were valid or not. See the famous article of H. L. A. Hart, "Self-referring Laws?" in Essays in Jurisprudence and Philosophy (Oxford: Oxford University Press, 1983), as well as Peter Suber, The Paradox of Self-Amendment (New York: Peter Lang, 1990). [BACK]

48 I am inferring a simple majority, since nothing else is stated. The National Assembly is free to draw up its own internal procedures (Article 32A), which would have to deal with this question. If the draft constitution is interpreted as a bill, the Presidency Council would arguably have veto power over the draft before it is submitted to the people, which could mean either veto by only one, or, more likely, by all three members (Articles 36C and 37). But the constitution could be interpreted as a special law, wherein the final act of enactment referenda replace the executive that is not mentioned in this context by the very poorly drafted text. [BACK]

49 Strictly speaking from the legal point of view, an unamendable constitution can only be replaced as a whole, though politically speaking, partial illegality regarding its application (through creative interpretation, disregard of the amendment rule, etc.) is also possible. In either case the transitional legal order that interim constitutions are meant to establish and protect would be severely endangered. [BACK]

50 See Spencer Ackerman, "Iraq'd: Panic on the Streets of Kurdistan," New Republic Online, June 10, 2004; available at [BACK]

51 Jon Elster, "Arguing and Bargaining in Two Constituent Assemblies," The Storrs Lectures at Yale Law School, 1991. [BACK]

52 "[Brahimi speaking] I welcome the clarification . . . by Ambassador Bremer, who . . . stressed that Ôthe Interim Government will not have the power to do anything which cannot be undone by the elected government which takes power early next year.' The fact is that the TAL is exactly what it says it is, i.e., a transitional administrative law for the transition period. It is not a permanent Constitution. Indeed, it is not a constitution at all. The Transitional Law (or any other law adopted in the present circumstances) cannot tie the hands of the National Assembly which will be elected in January 2005 and which will have the sovereign responsibility of freely d

You may also like

AUG 9, 2023 Podcast

Nuclear Ethics for this Moment

This panel explores ethical questions surrounding nuclear weapons and builds upon a symposium published in the most recent issue of "Ethics & International Affairs."

MAY 10, 2023 Journal

Ethics & International Affairs Volume 37.1 (Spring 2023)

The editors of "Ethics & International Affairs" are pleased to present the Spring 2023 issue of the journal! The highlight of this issue is a symposium organized ...

MAY 4, 2023 Article

A New Era for "Ethics & International Affairs"

The editors of Carnegie Council's quarterly journal "Ethics & International Affairs" are proud to announce the beginning of a new era in our publishing history. Starting ...

Not translated

This content has not yet been translated into your language. You can request a translation by clicking the button below.

Request Translation