Constitution Making Under Occupation: The Politics of Imposed Revolution in Iraq

We’re going to follow two parallel tracks: the Governance Team will continue to work on details with the Arabs on the [I]GC while I tackle the difficult issues directly with the Kurds. Then all the parties will get together to hammer out an interim constitution that would withstand the stresses of sovereignty beset by a stubborn insurgency. And we need to do all this by March 1….

—L. Paul (“Jerry”) Bremer, late December 2003,

according to My Year in Iraq

In this chapter,¹ I will argue that at the heart of the interim constitution, the product of the first stage of Iraq’s constitution-making process, was a state bargain. This idea is a clarification, not an abandonment, of my earlier stress on imposed constitution making, which others who once disagreed with me have since made their own.² The bargaining in question was highly exclusionary, more so than even political participation in occupied Iraq, and the exclusion was imposed. The results of the bargain would never have survived the various levels of negotiation and could not have been ultimately insulated by a very difficult rule of change had it not been for constantly renewed threats, in effect acts of force, on the part of the American occupiers. On the other hand, it is also true that what was imposed was not the Americans’ own initial preference but was the result of a genuine bargain with one agent, the only agent they treated as an equal, the Kurdish parties that controlled the Kurdistan Regional Government and who were more or less completely united during these negotiations.

Below, I will contest the notion that the bargain was a historic compromise between Arab and Kurdish positions. The idea of a genuine American-Kurdish bargain, where the CPA negotiated in effect for Arab Iraq, may imply something like that, so I will respond to any possible confusion and criticism in advance. The position the Americans began with in the negotiations was indeed very close to Arab civic or postnationalist positions (the terms will be explained below). But they abandoned that perspective relatively early, and the deal they made was not a fair compromise between the initial positions. It could still be represented as a fair American-Kurdish bargain—after all, the two sides, understanding all the circumstances and power factors, entered into it freely—but it was not a fair Arab-Kurdish bargain, because Arab positions were abandoned, in their name, by others who would not have to live with most of the consequences. Indeed, this was done without using the considerable threat potential of the American government and the CPA, the factor Arabs in the process relied on to the extent they accepted the two-sided structure of bargaining in the first place.

Thus it would be misleading to treat the central phase of the process of making the Transitional Administrative Law, the process of bargaining over the territorial structure of the state, as mere imposition. It was and remained imposition vis-à-vis the Arabs, but it also involved bargaining with the Kurds. The transformation of the Arab-Kurd relationship into a purely strategic one on which no stable new state structure could be based was one important consequence of this asymmetrical way of proceeding.

In general, we must assume the presence of all three forms of coming to a collective decision in constitutional negotiations. Undoubtedly, imposition and bargaining both involve threats and the willingness not to carry them out in return for concessions. But it is worthwhile to distinguish the two categories. Imposition is relevant to the extent that (1) an actor’s credible threats cannot be met by effective counterthreats, (2) threats play a much greater role than promises, (3) the bargaining relation becomes monological rather than genuinely interactive, and (4) the result involves no exchange of concessions. In fact, remembering a warning from Max Weber, insisting only on the inevitable presence of both imposition and agreement,³ we should always keep in mind all three terms. Persuasion (“agreement”) must be present in order to keep one’s own side together, the “friend” portion of Carl Schmitt’s famous couplet, which would become too unstable and prone to defection if based solely on interest, fear, or having a common enemy, or even all three together. Imposition is always present, because no two or more sides are ever completely equal, and the stronger always gets to impose to some extent. As long as there is voting on a final draft, and there should be, the winners impose at least part of their constitution on the losers. The same is true when two or more sides make a compromise that excludes a third or fourth on whom the constitution or at least a part of it is then imposed by a majority or qualified majority. It is equally unthinkable finally that different sides should be able to persuade one another on all issues and that there would be no need for compromise. But it may be difficult to imagine a legitimate process where there is no persuasion at all (at least implicit persuasion concerning the fairness of the procedure itself) if compromise processes are to have any success. There are indeed many issues that lack a single legitimate solution, and there is no normative reason to expect one of the two sides in every debate to be persuaded. But only persuasion can lead to an interactive framework that could be the basis of fair bargaining, and only the latter allows parties to generate a minimum of trust, if not in each other then in the framework, and to regard their compromises as more than simply strategic and temporary.⁴

Thus there is imposition, persuasive arguing, and bargaining in all successful negotiations, especially including constitution-making processes. Of course, the weight of each element need not be the same. If Elster’s emphases concerning 1787 and 1789–1791 could be rightly put on persuasive arguing and bargaining, in Iraq, according to my underlying hypothesis, everything shifted to imposition and bargaining. Ultimately, there were two reasons for this: the amount of force available to one primary actor, the U.S. government, and the repeated and continued insistence of that actor to accomplish constitution-making tasks according to rigid, artificial, and accelerated timetables. The presence of open force and an apparent lack of time make the use of persuasive arguments a highly implausible way of advancing one’s interests in a negotiating situation. I doubt that at the Iraqi venues of negotiation there was a great deal of arguing in the sense of attempts at mutual persuasion based on principles, but it is difficult to tell given the dearth of records and credible testimony.⁵ At some of the venues, as I will show, we can assume the presence of persuasion based on participation in a common struggle and the obligations that would arise from that. Even here, the instrumental use of public-regarding arguments was the thing that must have been feared the most, namely that the supposedly weaker party will go public with a story of trust and betrayal. But the overall negotiation process had little relationship to publics other than an engagement in the most crass and transparent public-relations operations. Indeed, again following Elster, both imposition and bargaining were often presented to the press in public-regarding forms (and even imposition was masked as genuine bargaining). Under Iraqi conditions of trust, however, hypocrisy rarely had the desired result. Very likely, on the contrary, even genuine public-regarding claims inevitably appeared to be hypocritical. This is a serious matter, because coming to agreement by using public-regarding arguments that authentically could be presented to the outside as such is an extremely important element of the legitimacy of a constitution-making process. It is also part of the “glue” that makes the actual bargains something more than merely strategic ones that could be renounced at the slightest excuse or opportunity.⁶

I will on the whole avoid evaluating the few claims of justice on the part of the actors, which mostly dealt with past injuries and their proper contemporary and future institutional redress. As far as I am concerned, all of the sides have suffered enough by now, and many though by no means all of the arrangements they seek (for example, a postnational or civic-national state on the part of some Arabs and bi- or multinational arrangements in the case of many Kurds, if they accept “Iraq” at all) could all be made compatible with the demands of justice, even if in different ways, as long as they were promoted in liberal constitutionalist versions. In my view, there is no single just solution to the problem of defining the demos or demoi of a divided society over a given territory, but it is not the goal of this work to demonstrate that rather obvious normative claim. The various solutions to this and other problems that were to concern the Iraqi constitution makers were, however, greatly tied to the past and present structure of inherited memories, ideologies, interests, and power positions of the various actors. The question throughout the process of constitution making was whether these memories and ideologies allowed actors with diverging interests and power positions to compromise their different ideas about institutional solutions. While it is possible but by no means certain that the particular memories and ideologies made mutual persuasion unlikely, in my view they certainly would have allowed principled compromise solutions if a fair bargaining framework had been provided. In relation to the four pressing issues I will discuss, state formation, government structure, rules of constitutional change, and the relationship between state and religion, there is enough to indicate the outlines of where second-best solutions could have been (and, in the last case, were actually found to an extent). The reason why this did not happen in all four cases and for the overall constitutional package was not first and foremost because of the failure of persuasion, which probably never had a chance, but because of the triumph of imposition and its timetables over fair bargaining.

That at least is my thesis in this chapter. To demonstrate it, I will first try to consider the venues and actors in the processes of making the TAL. Then I go on to discuss, as ideal types, the positions the main actors could draw on regarding questions of state and government formation and where the possible intellectual lines of compromise between them lay. Next, I will describe the actual process of the making of the TAL, moving through the venues and making the case for the centrality-of-the-state bargain amid the various phases. Then, switching perspectives and looking at the TAL itself, I consider the three most important areas of constitution making, where there were sharp divergences of positions, to evaluate whether the outcome should be understood as a historic compromise or ultimately the imposition of the perspective of one side. Finally, after considering the deep legitimacy problems of the TAL and its creation, I will consider Sistani’s final battle against the interim constitution and the provocative but inaccurate suggestion that the Grand Ayatollah actually managed to invalidate, and not just delegitimate, the TAL. I end with a discussion of the failure of state reconstruction in the TAL.

The Transitional Administrative Law was made in four venues, and people who assume that it was one or the other that produced the whole thing are mistaken. In chronological order, but definitely not in order of importance, these were, first, the ten-member Drafting Committee of the Iraqi Governing Council founded in December 2003, under the chairmanship of Adnan Pachachi,⁷ which may have actually dominated the process very early and produced at least one draft in January⁸ but later was reduced to a clerical function. This group, in terms of its power to do anything, was the least important, and we know the least about how it worked. It may very possibly be the case that here initially attempts were made to make principled arguments for positions. It seems, however, that when serious disagreements manifested themselves, the Pachachi Drafting Committee hopelessly deadlocked.⁹ The second venue, in order of importance, was the Interim Governing Council itself. It is hard to say when exactly this body began to discuss TAL drafts and amendments, probably very late if Bremer’s recollection can be trusted.¹⁰ But if the Crisis Group is right, its members attended the drafting committee earlier, and the line between the two bodies was fluid.¹¹ Equally important, the IGC did not have a constitutional secretariat or expert staff to preprocess the issues and were entirely dependent on the other venues to prepare the discussions, materials, and drafts for them, and this was a very serious weakness. There was only one exception to this, the issue of state and religion, where many of the members had comparative knowledge and both settled and sharply differing views. Here the sources indicate, as I will show, that there was genuine discussion and give and take in this body, with the political principals playing direct roles, and even the great power holders, Bremer from within and, the sources say,¹² Sistani from without, paid close attention and exercised influence. In my view, however, this issue was construed in largely symbolic terms on whose outcome, as I will show, very little was to hinge in the end, except that one side, the Shi’ite clerics, were to sacrifice a lot of negotiating capital over it.

In my view, two other venues were more important. It is commonly said that American experts drew up the TAL, which was originally written in English, not Arabic.¹³ This is also how one of the insiders, Larry Diamond, seems to describe it, with two qualifications. First, the American drafters did work on a draft submitted to them by the Pachachi Drafting Committee, and second, they included two important expatriate Iraqi (nonconstitutional) lawyers, Feisal Istrabadi and Salem Chalabi, both members of that committee, with the former beginning his service by translating and rewriting the Pachachi draft.¹⁴ Thus there was considerable overlap between what I will call, using Bremer’s language, the Governance Team and the Drafting Committee.¹⁵ Altogether, the Governance Team, an informal subcommittee of the Governance Office of the CPA (the latter name aping MacArthur’s Government Section, which sat as a “constitutional convention”), had five members, according to Diamond: himself; Istrabadi; Chalabi; a British foreign service officer, Irfan Siddiq; and an American political appointee, Roman Martinez.¹⁶ Very likely others including Scott Carpenter, Meghan O’Sullivan, and unnamed lawyers and bureaucrats went in and out of the group. But there was no constitutional lawyer.¹⁷ According to Diamond, they worked tirelessly for many weeks.¹⁸ From all descriptions, that could not be said of the other three venues, so it is likely that the actual drafting did occur mostly here. However, drafting should not be confused with making, and the Drafting Committee should not be confused with the overlapping Governance Team, which is what Chandrasekaran seems to do, with Chalabi, Istrabadi, and Diamond as his main informants.¹⁹ Thus even if the five-member Governance Team “made” most of the TAL—which it did not, as I will try to show regarding the essentials—this would not make it an Iraqi, nonimposed product.²⁰

From Diamond’s own description, it seems to me that the basically American Governance Team had its actual political influence in the area of governmental structure. While it seems that some members of the group had very strong, well-formed opinions on some subjects, for example Istrabadi in his opposition to any ethnic federalism and consociationalism and Martinez in his opposition to judicial (constitutional) review, these ideas were easily eliminated not as much by their internal debates as by the political and ideological trend of the general proceedings. It was when technical solutions were sought to previously made political decisions that the drafters had some freedom, even if at times they wound up (as in the case of the veto powers of the members of the Presidency Council) going against, whether deliberately or not, the original political intentions of actors who may not have understood what the “experts” actually did. Thus they had considerable power, and they had this power on rather less symbolic issues than the plenary of the IGC as a whole.

Where the five-member Governance Team (and especially the Iraqi Drafting Committee) had no power at all was on questions of state structure.²¹ “The federalism issue was temporarily quarantined while Bremer and other top CPA officials negotiated directly with the Kurds.”²² That negotiation was to last until the bitter end, and was to be no mere “conversation among friends.” According to Paul Bremer (whether or not these words were actually said), a two-track strategy was decided on early in the game: “We are going to follow two parallel tracks: the Governance Team [his use of this term may have been somewhat different in terms of its personal composition than mine] will continue to work on details with the Arabs on the [I]GC while I tackle the difficult issues directly with the Kurds.”²³ These trips to Kurdistan continued from January 2, 2004, through at least the middle of February,²⁴ and there were discussions with the Kurds as a caucus in Baghdad²⁵ and with them alone up until the last day before there was a final agreement on the TAL. One such meeting was also held with the “Shi’a House.”²⁶ The “I” in the Bremer quotation, of course, was not quite accurate, even if it revealed a weakness that the Kurds or their advisers discovered how to utilize.²⁷ Bremer went to Kurdistan a number of times, with either Ron Blackwill,²⁸ the British diplomat Jeremy Greenstock, or two young aides, Martinez or O’Sullivan,²⁹ but there is little sign that he ever took anyone along with even the slightest expertise on issues such as federalism, natural-resource allocation, the conversion of the militias, and governmental structure. On their home turf in Kurdistan, in Erbil or Salahuddin as the case may be, Massoud Barzani and Sami Abd-al-Rahman of the KDP and Jalal Talabani and Barham Salih of the PUK, themselves well experienced in governmental-institutional matters (unlike their U.S. counterparts), had great reserves of expertise to draw on, including a very talented group of American, Irish, Canadian, British, and Kurdish exile experts in law, political science, and negotiation.³⁰

As I have already suggested, these Kurdish negotiations were the most important. This was true first because only here did power and issue significance come together. Elsewhere, the participants of the other venues did not have the ultimate power to decide (the draft committee and the Governance Team) or were not given the time and supporting expertise to really discuss the fundamental issues (the IGC as a whole). The Kurds knew exactly over which issues there could be no compromise, but beyond that they were willing to be quite flexible, in particular regarding Kirkuk and the distribution of oil resources. In return, they expected two things and got three, the last the most important. First, whatever compromise was going to be made in Erbil, it was going to be the compromise, and the Americans were expected to impose it on the rest of the IGC.³¹ The Kurds themselves were not similarly bound. They could and did try—and this was the second thing they expected—to turn their concessions on federalism into gains on the governmental structure being drafted by the five-person governance subcommittee. And they could and did ask for entirely new things in the final short plenary mode, and they expected the Americans still to support them substantively and procedurally. These actions radiated out from the state negotiations and made the Kurds the dominant force next to the Americans in the process as a whole.

And that outcome was perhaps foreordained by the structure of the negotiations over the state structure. It was only here that Bremer and the CPA treated another actor, or two united actors, as an equal, in a genuine bargaining situation without an attempt to impose or to use threats that could not countered (with respect to the Kurds, the threat would be that of an American or Turkish invasion). While I was not there and cannot say for sure, I would not be surprised if the two sides used real arguments to persuade each other on the basis of common interests, common earlier support for each other, and even shared values, along with, most likely, common opposition to the Islamists, or “Black Turbans,” as the Kurds referred to the Shi’ite clerics. With the historical mistrust of the Kurds and the likely prejudices of the Americans, this could have been a context in which some trust was built.

Much more important was the fact that whereas all other negotiations were multilateral, if highly exclusionary, this process was bilateral. The Americans, in other words, accepted the most fundamental premise: Kurdistan was one and Iraq was one, and the two were negotiating their federation and not, as Galbraith supposedly explained, the ways and means of the devolution of power in a united state, in the form of a new autonomy.³² The operative phrase for the Kurds became “voluntary union between the Kurdish and Arab peoples,” which had already appeared in an article by Massoud Barzani on December 21, 2003.³³ Of course Iraq, not having a de facto state, could not negotiate with a quasi-state on an equal basis, and this is why Bremer took it upon himself to deal with the Kurds while “mere technicians” did some drafting with the Arabs. Bremer himself claims that all this was “suggested by several Arabs on the IGC.” Diamond says that it was Pachachi’s idea³⁴ to send the CPA boss to Kurdistan, and if so, the elder Sunni statesman must have assumed that only the American leadership had the power and authority to negotiate with the Kurds.³⁵ Indeed, when (apparently meaninglessly) selected Arab participants were invited to join a meeting in Erbil or when the Drafting Committee went there for a session, nothing much was accomplished.³⁶

But what was at stake was incredibly serious and went beyond occasional Arab presence here or there.³⁷ The Kurds were consistent supporters of a selected rather than elected body to draw up a constitution, and they consistently opposed early elections. This was hardly because their parties needed time to prepare themselves, since they were politically the best organized and most experienced in Iraq.³⁸ What was anathema to them for very obvious demographic reasons was a sovereign body, elected by a one-person-one-vote principle, drawing up a constitution for the whole of Iraq on the basis of even a qualified majority decision. They immediately grasped the significance of the November 15, 2003, agreement, and Jalal Talabani was entirely right to sign it from a Kurdish point of view, despite later criticisms he received for accepting language vaguely having to do with territorial eighteen-province federalism.³⁹ That issue could be and was dealt with later; the choice of the negotiating forum was far more important!

What Talabani or his advisers grasped, unlike their Arab counterparts most likely except for Sistani, was the significance of the IGC under the CPA producing an unamendable interim constitution that would significantly preempt and structure the work of the constituent assembly and the final constitution. But even an interim constitution and its negotiation involved hidden dangers for them. If the process was constructed fairly and inclusively, the Kurds would be only one-fifth of the forces present. Their military strength would be neutralized by the referee, the United States, and could not be used as a threat. Secession was a threat but not a fatal one, unless they could take Kirkuk and the oilfields with them, and the Americans had reoccupied that part of Iraq after its early Kurdish conquest in 2003.⁴⁰ Secession also risked deep problems with Turkey, especially with Kirkuk involved. To be sure, the structure of the IGC was not fair to begin with. With Sunni exclusion,⁴¹ Kurdish strength in the IGC was greater than it would have been in a truly representative co-opted body. But their view on nationality, state structure, and governmental institutions was a distinctly minority view, especially initially. They would have received concessions, but the tendency would have been to grant them cultural autonomy in the context of eighteen-province federalism. So it was crucially important for them but unacceptable to everyone else involved to change the format from a round table of, say, four major and a number of minor participants (major: Americans, Kurds, religious Shi’ites, and secular Shi’ites; minor: religious Sunnis, independents, and other ethnic groups) to a figuratively two-sided table of Kurd-Arab negotiations. This was not possible because it clearly would have incited the resistance of all those who opposed the sectarian redefinition and possible division of the country. But it was equally good to get the same structure via a separate set of negotiations in which the Americans represented Arab Iraq; in fact, as it turned out, it was much better, because of the unexpected cooperative attitude or weakness of this substitute partner.

Once the premise of a special bilateral venue was accepted, it would have contradicted the negotiating situation itself to ask the Kurds to give up just those things that led the Americans to accept them as an almost equal partner.⁴² Letting the Kurds keep those things, however, made them entirely unequal to all the other Iraqi participants in the negotiations, and it required the Americans to enforce precisely this inequality.

Finally, it was important that those who would have objected most vociferously were kept far from the most important negotiations. This was certainly true for the relevant members of the IGC and the Drafting Committee. Equally or even more important, just at the time that Bremer and the Kurds were bargaining the most intensely, the return of L. Brahimi to Iraq was also being negotiated, and he struck his compromise with Sistani around the time the Americans finalized their state bargain with the Kurds. Because of the demands of international law, UN officials were almost unanimous in their opposition to negotiating and especially altering state territorial structure under conditions of an occupation, and to Brahimi and Benomar, being liberal and secular Arabs, the idea of a division of Iraq on an ethnic basis was hardly appealing.⁴³ They were, however, not part of the crucial negotiations with Talabani and Barzani, though of course in February they could have been included. What they knew of the emerging deal is hard to reconstruct. In effect, however, they were offering a bargain to Sistani based on delaying the free election to the constituent assembly at a time when the very significance of that assembly was being reduced by a deal concerning the state that, having been arranged before the elections, would thus be one less thing over which Iraq’s elected representatives would have decision-making powers. No wonder that Brahimi felt cheated and undermined afterward.⁴⁴

But the UN could in no way reverse or even modify the result of the state deal. Only Sistani could attempt to do that, and in the end he failed as well.

Having destroyed the Iraqi state in its territorial and organizational integrity and having contributed to the division of the state’s people on ethnic grounds, the Americans clearly understood that part of the constitution-making process would have to do with state making. Most of their bilateral discussions with Kurds were focused on this issue, and it was these negotiations in Erbil and in the end in Baghdad that decided the question regarding at least the territorial structure of the state. They also touched on, less inconclusively, the related question of the possession of armed forces within the state.⁴⁵ While the reconstitution of the state’s people as two nations along with protected nationalities was also discussed in the bilateral talks, these matters were handled fairly consensually by the other three venues. Given the territorial structure negotiated by the Kurds and Americans, however, it is difficult to believe that the issue of binationalism (language rights in particular) was as open as it may have seemed to some participants. To understand the connection, it may be worthwhile to briefly sketch, even if as ideal types, the main positions present in the controversy. These positions have particular representatives in Iraq, but they cut across parties and in their pure form do not represent party positions.

The ethnic nationalist position ascribed to many Kurds, usually without further qualification, sometimes to the Kurdish “street” and rarely to a specific individual, periodically surfaces in the statements of main leaders and even some foreign advocates.⁴⁶ It is based on the more or less correct historical premise that Iraq’s originally patched-together territory has always been the homeland of two major “nations,” both quite recent imagined communities, of which one, the Kurds, have never accepted attempts to “Arabize” the whole territory, despite repression, assimilation attempts, ethnic cleansing, and forced deportation. According to the Kurd ethnic nationalist, it is both a matter of justice and unfulfilled historical promises of the great powers that twenty-five to thirty million Kurds, the largest nation in the world without a state, receive their own nation-state.⁴⁷ For the ethnic nationalist Kurd, there “always” was a Kurd entity in the sense of a people and a territory, and all it has been missing despite relevant promises was a state organization covering the whole territory and administering the whole people.⁴⁸ For the ethnic nationalist, here as elsewhere, ultimately there is a choice only between (1) Arabic (and Turkish) ethnic nationalism along with the repression and even elimination of the Kurds as a people and (2) Kurdish ethnic nationalism. There is no room here for multiple identities, for example, a Kurdish national one within an Iraqi civil “nation.”

Given their history in Iraq, this position sees only three acceptable institutional solutions for its aspirations. In order of desirability, the first would be a greater Kurdish state incorporating also the Kurds of Turkey, Iran, and Syria, an option everyone regards as impossible in the short and middle term. The second would be a smaller Kurdish state carved out from Iraq, encompassing both the three governorates plus the fragments of two others in the present Kurdistan region and all Kurd-majority areas or even Kurd-plurality areas in Iraq, including Kirkuk. This option would be undesirable without the added territories and may be currently impossible with such additions, so there is also a third formula, a binational state with a more or less independent Kurdistan within it. This is seen by the ethnic nationalist as a huge concession from the point of view of his or her first and second options, a mere third best, and in return there is a definite expectation of the territorial expansion of the Kurdish part of the binational structure to include all Kurdish majority and even plurality areas, including Kirkuk province and Kirkuk city. As we have seen, in this view of things, such a confederation or federation (not really a state) should be negotiated bilaterally between Kurds and Arabs. Ideally, it would have two symmetric parts, one Kurdish, one Arab. But the organization of each should be up to its own constitution, and in theory it would be acceptable that the Arab part organize itself subfederally, in terms of, say, thirteen to fifteen provinces or governorates. Thus an asymmetric structure on a Canadian model (the Arab part would not have its own regional government, only the federation government; the Kurds would have two governments) would be acceptable, depending on the status of the whole and the powers given to the parts (great) and the whole (very limited).⁴⁹ As long as each part—that is, first and foremost, Kurdistan—retains a veto over all constitutional legislation, foreign and military policy, and possibly all national decisions of any importance, even a three-part organization (Kurdistan, a Shia region, and a Sunni region) or a five-part one would be acceptable. The stress is on the veto power.

The ethnic nationalist perspective to the extent it accepts Iraq at all sees it as a treaty organization, a confederation, or a federation (not federal state) only somewhat more centralized, if an asymmetric structure is conceded. At the same time, the ethnic nationalist will tend to accept the empirically false and logically somewhat incoherent idea of Brendan O’Leary (himself a liberal nationalist) that there is some kind of deep structural link between three dimensions: ethnically defined federations, powerful units with weak centers, and power sharing in the center.⁵⁰ And, in fact, the linkage is more logical from the ethnic nationalist point of view than otherwise, though it is not completely clear why someone who wants to separate would still want to rule the unit one is separating from. However, the motivation is fairly understandable. If the ethnic nationalist cannot have his independent state, he will want a three-fold guarantee against the “state” that is conceded to the federation, which will make the unit itself a quasi-state. Note, however, that in this version power sharing means a device to weaken the central federation by a system of rigid vetoes.

There is no question that this relatively well represented position in the IGC and the Drafting Committee in terms of individuals lacked a power base and may have appeared extreme and ideological.⁵¹ Its advocates accepted Kurdish claims that pan-Arab nationalism has resulted in great oppression of and crimes against the Kurds. But replacing an ethnic national state with a binational state or by ethnically based federalism in their minds would be no solution. When Arabs such as K. Makiya make the argument, they often use an Israeli analogy. In Iraq, like Israel, ethnic definitions of the state or of units of the state would be in the end incompatible with democracy, because those not part of the relevant ethnic group would be lesser citizens of the state or the unit.⁵² Iraq (and its units) should be a state (and units) of all its (and their) citizens, and they should all receive the same rights and obligations as citizens of Iraq. Federalism is here favored not as a way of instituting special identities but as another set of checks and balances against arbitrary government (“separation of powers” and “protection of minorities”), but federal units organized on an ethnic basis could themselves become small-scale but equally potent threats to individual and minority rights as could an ethnically defined national state.⁵³ Ethnic federalism would lead to the complete ethnicization and, in Iraq, the sectarianization of politics, which would threaten the survival of any kind of statehood. One answer is therefore comprehensive separation and division of powers, where all the different branches and levels—central and decentralized—control, monitor, and correct one another. Another is fiscal federalism (central control and equitable sharing of a large share of the resources to the units) complementing the territorial one that splits up ethnic fiefdoms.⁵⁴

There are two possible versions of the civic model, a postnationalist and a republican nationalist one. The difference has to do with the thickness of the nonethnic collective identity that is affirmed and its resulting openness (of the postnationalist) or suspicion (by the republican nationalist) toward claims of subethnic nationalisms when restricted to demands for cultural autonomy. It could be said that while the postnationalist could live with a “state nation” concept as proposed by Linz and Stepan, involving multiple identities and multicultural rights, the republican nationalist still imagines a “nation state” but with all requisite individual rights.⁵⁵ It is very hard to say where Iraqi liberal advocates on the Drafting Committee would have defined themselves in relation to these two positions. Most likely, the republican conception may have seemed a little too close to Arab nationalism (see below) and was not considered to be worth advocating under the circumstances and given the way the IGC was constituted. But it is also possible that all civic nationalists were also committed or strategic multiculturalists, given the atmosphere of our times, especially in the patron country.

However that may have been, although advocates and even states may not be entirely coherent on these matters, a whole variety of federalist arrangements are compatible with the two versions, granting stronger or weaker powers to the units and to the whole, which could be a centralized or a decentralized state with provinces, a federal state, or a federation, but not a confederation or an asymmetric federation with one “federacy” that has a confederal, treaty-based status. Of course the model allows for the possibility, as in India, of a highly (though diminishing over time) centralized federal state, with some or even most of its territorially defined units having an ethnic majority, but the units themselves not being ethnically defined and therefore a consolidation of units on merely ethnic grounds not easily permitted.⁵⁶ Disincentives to such consolidation⁵⁷ can be established, or they would need to involve constitutional amendments if not outright bans. Amendment rules would involve participation by federal state organs and legislatures or electorates of the units, according to various possible quantitative proportions, but outright vetoes would not be allowed. The question of whose powers are enumerated and whose would be the reserved powers need not be solved in the American way, for example, and current theorists tend to favor enumerating both sets of powers as well as shared powers. With respect to Iraq, this position in its postnationalist and some republican versions would affirm or allow a federalism based on eighteen provinces, with significant powers devolved to them. Evidently, it would have to accept the fact that each province would have an ethnic majority or plurality. The postnationalist at least could offer cultural autonomy to all the main nationalities of the country, which could go so far as establishing two official and several other protected languages, allow public use of several above a certain threshold of population, establish institutions of higher learning in at least Kurdish and Arabic and other schools in all protected languages, and so on. But the constitution in the postnational version would not define the state as that of one or of two nations, and it may not mention the word nation at all. Or alternately, in a more republican version (coming closer to Iraqi nationalism, see below), it could define the Iraqi nation in terms of all of its citizens, whatever their ethnicity, religion, or gender.

The liberal nationalists are clearly different from both the ethnic nationalist and the republican nationalist positions, because they think in terms of the possibility of multiple identities. They would not go so far as to adopt the category of “state nation” for Iraq. Thus liberal nationalists postulate the possibility of two or even more national identities within a single non-national state identity, for example “Iraqi identity.”⁵⁸ Its advocates say they are liberals but “not difference-blind liberals.” The position even prefers a relatively thin civic definition of nation, but it argues in that case that in Iraq two such nonethnic definitions are needed, one Arab and one Kurd, or possibly three, one Iraqi, one Arab, and one Kurd.⁵⁹ The argument for this cannot be theoretical, since theoretically either one (Iraqi) or three such identities fulfill the same civic purpose, and indeed the two-part Arab-Kurd division threatens to reethnicize the civic conception. (One could be Iraqi and Arab, Iraqi and Kurd, or just Iraqi, but never Arab and Kurd.) The reason for making their particular choice (aside from latent ethnic nationalist sympathies) is historical: the bitter experience of perhaps the largest “nation” without a state, the Kurds, at the hands of Arab nationalism requires separation if any common state framework is to be preserved. Here the example of Israel is used in quite a different way than it is by Makiya: genocide helped produce a Jewish state, not integration-minded Jews.⁶⁰ That is a fact, but so are the consequences, good and bad. Makiya’s point regarding some of the outcomes of an ethnic definition of the state in Israel is not thereby diminished, especially to an Arab audience, but apparently that is not to whom the liberal nationalist is speaking.

Admittedly the model, unlike that of the ethnic nationalist, goes beyond the leading Israeli paradigms today. While the Kurds, like the Israelis, simply cannot and will not trust Arabs in any framework that does not make them the center of a distinct statelike formation, that formation can however “federate” on the basis of mutual need and interest with an Arab statelike entity called Iraq (or in the least nationalist version, even “in Iraq”). Thus for the liberal nationalist too the maintenance of the Kurdistan region is not negotiable, and its geographical extension is a highly desirable goal. While important Kurdish politicians such as Mahmoud Othman and Hoshyar Zebari seem to hold fairly consistent versions of this position, it has been developed in the greatest detail by several foreign advisors, who use it to reply to the charges made by the civic nationalists and postnationalists against the ethnic nationalist perspective. However, the distinctness of the position when compared to ethnic nationalism comes into question. Understanding the Kurdish perspective as a civic rather than ethnic nationalist one, O’Leary and Salih defend the position against charges that Kurdish nationalism could be as repressive over a smaller territory as Arab nationalism was over a larger one.⁶¹ The historical experience of Kurdistan since 1991 partly bears out their claim, although Kurdish rule in Kirkuk more recently has left a lot to be desired as well.⁶² But the real question concerns the future rather than the past. Here even the foreign advisors of the Kurds differ, and there is much more willingness on the part of some (such as O’Leary) to see the guarantees for individuals and minorities in a future Kurdistan also in terms of state-and federation-wide relations of checks and balances. Others (Galbraith) push for a kind of confederal status that leaves Kurdistan, when expanded in territory, part of a larger integration in name only.

There are, I think, important strategic differences between Kurdish perspectives, differences that roughly correspond to the ethnic and liberal nationalist positions even if inconsistently. They do in the case of O’Leary, the most consistent thinker on the side of the Kurds. The liberal nationalist also trades Kirkuk only because he must. But, despite the fact that he may accept some kind of link between ethnically based federalism (strong powers for the regions and power sharing), he is willing to see these devices as functionally interchangeable protections for the region and the ethnic group.⁶³ He wants power sharing not only to weaken the federal state but to retain a strong federal state for the purposes deemed legitimate or shared. Thus he should be more willing than the ethnic nationalist to trade some powers of his ethnically defined region, because he wants a somewhat stronger organization than mere treaty-based integration.⁶⁴ In return for what is given up, the liberal nationalist seeks to have greater powers in the federation, which as a minority the Kurds can get reliably through consociational structures of power sharing and less reliably through a federal chamber either suitably, that is proportionally, or, in their favor, disproportionally organized (but not according to a system of eighteen governorates of which the Kurds have three, or one-sixth, as opposed to their one-fifth share of the population). In order to constrain his own ethnos, the liberal nationalist can even affirm a strong constitutional judiciary with a strong Kurdish presence.⁶⁵ Here the problem is only that enforcement mechanisms in other federal states or federations involve strong executives, federal interventions, federal police forces, strong federal military, and the like—the very institutions the liberal nationalist opposes creating, along with his ethnic nationalist colleague. Finally, logically, seeking a stronger integration than a treaty organization would suggest an amendment rule perhaps between those of the civic postnationalist and the ethnic nationalists’ veto, but here too the actual positions of liberal nationalists and the ethnic nationalists on the Kurdish side tend to become indistinguishable.

Ethnic nationalism has obviously existed among Arabs too, as pan-Arabism or Arab nationalism, and it has always interpreted and fought for Iraq as a Sunni Arab country that should grant cultural rights to Kurds and Shi’ite Arabs very reluctantly and sparingly—and territorial rights not at all. The perspective is thus constitutionally strongly centralistic, in line with Iraqi traditions (at times compromised vis-à-vis the Kurds). It was entirely excluded from the negotiating process, despite its continuing political popularity in Iraq among Sunnis. What is obvious is that even if the perspective had been present its chances of winning would have been nil. Its exclusion nonetheless had some consequences. Paradoxically, had Arab nationalists been present, it would have been easier to see the postnationalist position as a possible basis of compromise. Otherwise the IGC would have had to face the internal prospect of a polarization on the lines of ethnic nationalisms, followed by the inevitable breakup of the country itself, which at the time most of the members would not have relished. The best way to avoid this would have been to eliminate the question of national definition from the discussion altogether. At the very least, the postnational or a compatible Iraqi national position could have been strengthened. Thus the exclusion of an undesirable perspective both weakened its role and shifted the whole intellectual balance of the discussions. Still, centralism had its advocates among the Shi’ite clergy, the potential beneficiaries. This, however, was a very weak position because among the exiles and their American sponsors “federalism” was always accepted as a magic mantra that could not be compromised. Moreover, the Shi’ites too have suffered because of Arab nationalism, from which they were excluded as “Persians” or because they follow a supposedly non-Arab version of Islam. The form of nationalism that has periodically appealed at least to more secular Shi’ites (and many Kurds, during the Bar Sidqi experiment, with the Communists, and under Kassem) was “Iraqi” nationalism, which was a nonliberal version of the civic republican form; according to this, Iraq was an Islamic nation of all its ethnic and religious groups.⁶⁶ Perhaps understanding their relatively narrow majority, they were not interested in explicitly redefining Iraq as a Shi’ite Arab nation. But they were nevertheless allergic to liberal postnationalist or even civic republican definitions of the political community, which to them in any version reeked of secularist, antireligious biases and traditions well known in the region.⁶⁷ In any case, they were both too distracted and divided to strongly represent Iraqi nationalism against the Kurdish gambit in either its ethnic or liberal version. As I will show, identity issues for them came to be focused on the role of Islam in the state, which was rather irrelevant constitutionally because here local control rather than constitutional formulae was going to decide things. But they did not seem to care too much whether Iraq was Arab or Arab and Kurd or just Iraqi as long as it was Islamic. Second, the more politically savvy among them early on saw the differences among alternative federal proposals that were functionally equivalent to the Kurds, namely whether the “rest of Iraq” would be one or two or four regions in a symmetric structure or merely fifteen provinces in an overall asymmetric one. And depending on party and geographic constituency, they either opposed regions altogether or supported them because they too wanted one or two of them in oil-rich areas with large Shi’ite majority populations. Until the final round of negotiations, their own choices or their divisions excluded them from this part of the discussion, and the basic agreement concerning the Iraqi state was thus made without them. When it was crowned with a ratification rule that protected the new arrangements, they woke up much too late.

Purely intellectually, the weakness of positions 4, 5, and 6 in the process (and especially the absence of 6, which would have been the most aggressive) tended to make Kurdish liberal nationalism rather than postnationalism the natural basis of compromise. Moreover, because of their harsh disagreement over Islam, secular nationalists and postnationalists and Shi’ite Iraqi nationalists had trouble finding common ground; in fact, each side seemed to be more comfortable with the Kurds, who were flexible on the big symbolic issues. They were Islamic but not fundamentalists. “Socially” (and such issues influenced even the highly sophisticated UN delegation), that put them between the liberals and the “Black Turbans,” who did not as a result see that as far as the main issue was concerned they were together on the opposite side of the Kurds. Aside from the structure of bargaining, the intellectual positions of the Kurds, especially the ones who did not argue on the grounds of ethnic nationalism, could begin to occupy a middle ground. Two things should be noted. Organizationally, as I already explained, there were two fundamental choices (rather than three): (1) a bargaining situation involving two parties, Iraq and the Kurds (a two-sided table), or (2) one in which the Kurds would be one actor among several (round table). On this point the two Kurdish positions were in agreement, and in fact both adopted the ethnic nationalist premise of “Iraq as a voluntary union.” Even more importantly, they managed to have their way. Just as the American negotiators, who originally started out with simple but reasonable versions of the postnationalist federal conception (eighteen-province, genuine federalism with full cultural autonomy), were brought into the Kurdish framework by accepting the negotiating model, the same process also blended the two Kurdish perspectives even closer together. To an extent, the ethnic nationalist position became the long-range objective if not the bottom line, and the liberal nationalist governmental schemes and statements of values became either public-relations tools or strategic bargaining ploys. Even if this is a slight exaggeration, the weakness of the opposing perspective pushed the supposedly compromise position, the Kurdish liberal nationalist one, in a more ethnic direction. Here are my hypotheses concerning the intellectual causal scheme that, in itself of course, could not have been decisive: the American exclusion of the Sunni nationalists led to a Kurdish center in the spectrum of positions. The weakness of the Shi’ite Iraqi nationalists and the secular postnationalists, who could not form a real alliance, made the other extreme, the Kurdish nationalists, stronger, and this made the center, the Kurdish liberals, move toward their positions.

Political outcomes, however, do not take place first and foremost on the level of ideas. Undoubtedly, the surprisingly weak pressure of the Americans on behalf of a civic postnationalist perspective they were expected to favor was to be an even more fundamental reason for what was to happen, and it is to this dynamic that I will now turn.

Looking at all the negotiating venues and their products, it becomes clearer how in the end a Kurdish amalgam of ethnic and liberal nationalist elements triumphed. I will here focus only on questions of state structure and political identity. In the Drafting Committee led by Adnan Pachachi, the civic nationalists and postnationalists were apparently strong, mainly because of the role of the chairman. Whatever Pachachi himself represented during his first governmental role (before 1968), he was now an Iraqi (rather than Arab) secular nationalist⁶⁸ opposed to ethnic federalism, provincial control of oil resources, and forms of power sharing based on ethnicity.⁶⁹ He worked mostly through Feisal Istrabadi, who held similar views and who as a drafter and advocate seemed to have been quite accomplished and impressive.⁷⁰ I do not know if at this level persuasive arguments of the type we see in Makiya’s well-known piece were attempted on behalf of civic postnationalism or Iraqi nationalism. Given the strong convictions of the chair and his top representative on this ten-person committee, there were certainly some frank exchanges of views. Whether or not there was any serious attempt to persuade, strong threats were certainly used on both sides, using external instances to pressure the drafters.

In December, according to Diamond, “Washington” informed the Kurds that they would have to give up their region and their regional government and would have to accept a federalism based on eighteen provinces. This was certainly based on a common opposition in the Pentagon, State Department, and White House to ethnic federalism, which they saw as a formula for the unacceptable breakup of Iraq.⁷¹ This converges with both O’Leary’s view that the U.S. authorities interpreted the November 15 Agreement in the sense of territorial or eighteen-province federalism (though this is not really supported by the agreement’s rather vague wording) and with evidence that the Kurdish signer J. Talabani came under strong pressure in Kurdistan for even signing the document.⁷² In reality, the agreement was vague on the question of federalism and did not exclude a Kurdish region or a special status for the Kurds.⁷³ In any case, the Kurds responded with very strong threats of their own, including secession.⁷⁴ On December 20, 2003, the five Kurdish leaders on the IGC submitted a draft “bill” in which they outlined their vision of federalism.⁷⁵ In Barzani’s summary, directed against both Iraqi and “foreign” interlocutors, the main principles were the voluntary union of two peoples (Iraqi and Arab used interchangeably for the other side), no surrender of anything in the existing situation including the Kurdistan Regional Government, the rejection of the separation of the Kurdish governorates from one another and at least to this extent province-based federalism, and the inclusion of all other “Kurdish areas” in Kurdistan, including Kirkuk.⁷⁶ Finally, the Kurdish members of the Drafting Committee (among them Fersat Ahmad) submitted a list of specific demands, based on “draft constitutions” adopted by the Kurdistan National Assembly in October 2002, for inclusion in the TAL. Among them:

the establishment of a federal Kurdish region, recognition of Kurds as one of the two main nationalities of Iraq, recognition of Kurdish as an official national language alongside Arabic, recognition of the Kurdish (regional) flag and anthem, reversal of Arabization in mixed areas and a highly evolved form of decentralization that would give Kurds a significant degree of autonomy and control over resources in their federal region. Proposed language concerning non-Kurdish matters proved relatively noncontroversial but everything having to do with Kurdish aspirations led to stalemate in the committee, which operated by consensus.⁷⁷

At this point, the different venues interact. All sources indicate that Bremer went to Erbil to break the deadlock in the Drafting Committee.⁷⁸ But he did not really get any results until later in the month (January 7, at the earliest, but that was only an outline conceding the existence of a Kurdistan Region), since the deadlock at first continued in Erbil, as we will see. Meanwhile, obviously not unaffected by the American-Kurdish negotiations, and with some possible help from the Governance Team, which was focused on other issues given Bremer’s instructions to remain hands-off, it is nevertheless possible to argue that the Drafting Committee produced a compromise (or at least “amalgamation,” according to the Crisis Group)⁷⁹ of its own, based on its earliest draft (which we do not have) and Kurdish submissions. According to Diamond,⁸⁰ during the first days of January there was already a draft he could read, written by an unnamed advisor of Pachachi and translated as well as redrafted by Istrabadi. This text has never been published, but from Diamond’s critical description we can tell that it strongly resembled what has been called the Pachachi draft, which was published in Arabic on February 1.⁸¹ It did not seem to contain a formula on federalism, however—at least Diamond does not mention such a thing. This can be taken in terms of the emerging formula for all of the early versions: leave the status quo as it is for the transition period and let or even require the elected constituent assembly decide most but not all the issues linked to the ultimate meaning of federalism in Iraq. The one issue I think the drafts did decide (possibly and even likely because of Bremer’s deal on January 7) was that there would be a territory called Kurdistan beyond provincial federalism if that was the scheme ultimately chosen. And this was a concession to the Kurds, even if it fell well short of the idea of a voluntary union. I will rely on the one published version to make the argument and will use O’Leary’s commentary as a partial corrective.

The Pachachi Draft seems amateurish next to the finished TAL, but that is in retrospect a very misleading impression. It has been rightly said to be “deliberately short on detail” and this, I think, was a matter of both democratic and international legal principle. Given the legitimacy problems of the CPA and the IGC, due to Sistani’s demands for a freely elected assembly, the idea of significantly binding a body with much higher legitimacy than the drafters of the TAL may and should have seemed unacceptable. Given international law, the idea of seriously transforming Iraq’s state structure and indeed irreversibly altering its regime under and by the authority of a foreign occupation also could and should have seemed unacceptable. But it was exactly the latter that was sought by the Kurds, who rightly recognized that a detailed interim constitution would help create or legitimate facts on the ground that would be very difficult to reverse later in a permanent constitution. Thus all the Kurdish proposals were highly detailed in all matters that concerned them.⁸² This invited “compromise” solutions that would be themselves highly detailed, as the TAL eventually was, and therefore this structural issue of the type of interim constitution represented not a compromise but the adoption of the Kurdish preference of neutralizing the constituent assembly reluctantly conceded to Sistani, one elected on the basis of a one-person-one-vote principle.

According to the Pachachi Draft published on February 1, 2004, Iraq is an independent and sovereign state with a democratic, parliamentary, pluralist, and federal “system,” but it is neither said to be Arab, or Arab and Kurd, or multinational, or any other kind of “national” (art. 3). In line with a postnationalist conception, or because no decision was possible, “nation” is not mentioned in the draft. Language is, and it is, in line with a more republican model, Arabic (art. 3). Regarding both these provisions, it is said that in the transitional period the current special status (regarding statehood) of Kurdistan and current special situation (regarding language) in the territory of Kurdistan shall be respected. The name Kirkuk is not mentioned in the draft, but respect for current special status would not in any way include Kirkuk city and Kirkuk province in Kurdistan, which after being captured by the Kurds from Saddam’s forces have been, unlike Kurdistan, occupied by the Americans—at least formally. Thus the draft does not meet any Kurdish demands regarding Kirkuk nor does it promise that the constituent assembly would even deal with this question. For the transition period, federalism is discussed only indirectly, with references to the eighteen provinces (one of which is Kirkuk) and the applicability of the TAL itself in all of them, and by outlining the powers of central government: foreign policy; defense; guarding of borders; peace and war; monetary, currency, and development policy; public budget; and citizenship affairs (art. 3). However, as this draft contains some ironclad principles for the drafting of the permanent constitution, one of these is relevant to federalism: the final constitution must include “a democratic, pluralist federal system including a unified Iraq and organizing the relationship between the territory of Kurdistan and the central government.” While this language again does not include Kirkuk, in one respect at least this version of the TAL would seem to preempt the elected constituent assembly: there would have to be in whatever federal formula chosen a place for a territory called Kurdistan, and not just three or four Kurdish majority provinces or governorates (art. 42). However, it is not clear who would control adherence to the constitutional principles, and whatever formula emerged in the final constitution would have to be approved in a referendum that without further elaboration seems to be a countrywide vote requiring only majority approval.

O’Leary is working with either a later version, a different translation, or a somewhat arbitrary interpretation of the Pachachi Draft.⁸³ That is why, surprisingly, he detects a harder-line position on the Arab side than I have. Either his polemical reading style, the translation, or the then current state of American-Kurdish negotiations could be responsible. There are crucial differences between our readings. The control of integrated armed forces and natural resources seem to be new, and could have been included at the behest of Bremer, to counter relevant Kurdish demands. The declaration of the Kurdistan Regional Government as a “subordinate level of government” seems to be new, but instead of a denigration, this seems to have been rather an acceptance of the Kurdish demand that the KRG would not be abolished. I note that the supremacy of the Federal State is not as O’Leary thinks “a wholly antifederal” but only a “wholly anticonfederal” mode of thought. There has been such supremacy in the United States, (West) Germany, India, arguably the European Union, and even Canada, though not recognized by those in Quebec who seek a confederal status. Other features of the two drafts regarding federalism seem to be the same, though interestingly O’Leary does not mention what would have been for him strongly in the “minus” column, Arabic as the official language, and, on the “plus” side, the limitation of the constitutional assembly by a constitutional principle that seems to enshrine the territorial integrity of Kurdistan. Perhaps these elements were now gone. There was a trend among the drafters of the TAL, because of the legitimation problems of the whole process, to eliminate constitutional principles binding the constitutional assembly in the South African manner. The powers of the constituent assembly were, however, still there, and O’Leary explicitly mentions the (implicitly or explicitly?) majoritarian ratification rule for the final constitution.

I think it would be a great mistake to see the early drafts emerging from the IGC Drafting Committee as the basis for the eventual compromise with Kurd positions. Of course, one can put various proposals next to each other, extrapolate some such relationship, and argue that the TAL created something like the Pachachi model in Iraq and somewhat modified versions of Kurd proposals for Kurdistan.⁸⁴ This very dualism, however, was itself the heart of the Kurdish proposal and therefore should not be understood as some kind of compromise. Whatever real bargaining occurred after the Drafting Committee deadlocked was between Bremer and the Kurds, and thus if anyone compromised it was them, around their own positions, which in the case of the CPA were constantly shifting. Initially at this venue too there was deadlock. But the point of this negotiation from the very beginning was that neither side wanted to act on or even fully articulate its most potent threats (which would have been fundamentally unequal), and as a result there was genuine give and take. After the January 2 “acrimonious and unproductive session,” when each side presented its hard-line position, Bremer returned in seventy-two hours with Ron Black-will, and the two sides made an effort to avoid taking inflexible positions.⁸⁵ Bremer seemed to understand the history that had led to the special position of the Kurds, appreciated their military alliance, and supported their demand for federalism, but only within a unified Iraq. Moreover, he rejected settling the most difficult questions (for example, Kirkuk) in an interim document, and said the United States would not accept a federalism based on ethnicity, which he very rightly recognized as “a central feature of the Kurd’s draft.”⁸⁶ Thus the Kurds gained very little at this point. But Bremer realized at the same time (“Barzani remained silent”) that they were not going to simply give in to the American positions. He “left Irbil with a sense of apprehension.”⁸⁷ According to Diamond, before they left, they decided to defer the question of ethnic versus provincial federalism and concentrated on the powers of government. The idea was that if the Kurds agreed to a strong central government with sufficient protections, maybe they could have their ethnic federalism after all. I think in the end the opposite happened, but the strategy made a great deal of sense. With the most important issues left undecided, the Kurds were willing to entertain the possibility of a relatively strong central government with exclusive control over defense, oil and water resources, fiscal and monetary policy, and borders.⁸⁸

By January 7, the Kurds got Bremer to return to the theme he was avoiding. In return for conceding a relatively strong central government, they wanted him to accept that they could not retreat to where they were before the war in 2003, and therefore accept the principle of a “voluntary union with Iraq.” This time what was always their truly fundamental principle was not completely rejected.⁸⁹ They also got Bremer to accept some modest action on Kirkuk and the establishment of a property-claims commission to adjudicate longstanding disputes due to the policies of Arabization. At the same time, the leaders of the CPA heard but did not (yet) give in on other Kurd demands, notably the nullification of federal laws, the retention of the Peshmerga, the banning of federal troops from Kurdistan, and a binational definition of the state. Nevertheless, in Diamond’s not implausible view at these January meetings, which continued through the month sometimes with and sometimes without Bremer, a “historic bargain” was struck between the CPA and the Kurds, based on the tradeoff of significant central powers and the preservation of a unified Kurdistan region with far greater powers than the eighteen provinces (or rather fifteen, because the three in Kurdistan would have no powers). Advocates of the Kurds do not see matters this way. According to Galbraith, on January 27, 2004, Barzani and Talabani met alone with Bremer, a disastrous mistake on the part of the Kurds, as he earlier explained.⁹⁰ At that meeting, the three seemed to have agreed to a formula that sounds like Diamond’s “historic bargain”: Kurdistan as a federal unit and a central government with great powers, including military and judiciary powers. The Americans thought they had an agreement, but according to Galbraith the Kurds later claimed they did not. It is difficult to know whom to believe. Given the power difference, one would have to go with the Americans, ordinarily, because they could insist on the deal being honored. But these were not ordinary times, and the United States did not have an ordinary government during them. In any case, even Galbraith says “fortunately” it was the Americans themselves who abandoned the deal.

At this point, the accounts of Diamond and Galbraith merge. According to Galbraith, it was on February 6 that Bremer informed Barzani that all references to the Kurdistan Regional Government would have to be struck from the TAL at the insistence of the White House, along with Kurdish as a second official language.⁹¹ While this set of events is interestingly missing from Bremer’s otherwise pretty complete memoirs, they are confirmed from the other side by Diamond. Chandrasekaran specifically mentions Rice and Wolfowitz as both being behind the order to Bremer.⁹² The Kurds, feeling double crossed (a different story than Galbraith’s, because this would indicate initial adherence to the January 27 agreement), retreated to an extreme and even more truculent position. Now even moderate leaders, including the remarkable Mahmoud Othman, joined the chorus of more extreme demands.⁹³

The so-called Kurdistan Chapter, submitted for inclusion in the TAL on February 13 but first discussed on February 10, 2004, was clearly a unified Kurdish response to the new American position. It was neither an initial bargaining position nor primarily a response to a Pachachi Draft, as O’Leary and the Crisis Group claim.⁹⁴ Galbraith, who along with O’Leary seems to have been one of the authors of the Kurdistan Chapter, makes a much more convincing case for the chronology and the politics of this proposal, though it is surprising that these sophisticated operatives, working so closely together, have not gotten their story straight.⁹⁵ For the moment, I wish to only summarize this proposal and consider its details in comparison with the final TAL arrangements themselves.⁹⁶ Only a few matters would be the province of the federal government; otherwise Kurdistan’s laws would be supreme in the region. Kurdistan would have its own army and own its oil resources, but it would not manage fields currently in operation. Iraqi troops could enter Kurdistan and taxes could be collected there only with the permission of the Kurdistan national assembly. The permanent constitution would apply in Kurdistan only if approved by a majority of its voters.⁹⁷

We have only Galbraith’s testimony for what happened at the next negotiating session, with Bremer back in Erbil. That he supposedly insisted on the January 27 agreement is hard to believe, since the White House repudiated it. But anything is possible; perhaps he already got his bosses to backtrack, since they in reality only had the option either to come up with a credible set of threats or make concessions. He refused to discuss the “Kurdistan Chapter.” Then or at a subsequent session, he got the Kurds to give up recognition of the Peshmerga and agree to its formal dissolution (which the Kurds would certainly not go through with) and to federal control of resources in the TAL.⁹⁸ Meanwhile, having persuaded Washington of the necessity of preserving the Kurdistan Regional Government, this key institution was again conceded to the Kurds, along with Kurdish as an official language and some partial measures for reversing the Arabization of Kirkuk before the final settlement of this question.⁹⁹ Note that a key result was, despite many strong words by Bremer to the contrary, an asymmetrical form of federalism that established the most powerful unit on entirely ethnic grounds. The other parts of Iraq would be under the system of provinces; only Kurds would possess a powerful regional government to protect their interests. All others would only have one government, the federal one, and how strong that would be was again made dependent on the Kurds. Somewhere along the line—I cannot tell when—a new balance sheet was drawn (most likely by the American experts of the CPA, or the experts of the KRG, or somehow the two together) among the powers of the federal state and that of the region, and, amazingly, the right of nullifying (amending, not applying) federal laws in all but a few enumerated areas was conceded to the Kurds as well. It is because of this structure that people began to speak about a “historic bargain” or a “historic compromise.” To me it looks rather like a compromise well tilted in the direction of the Kurds, but that outcome, already prefigured in the bilateral structure of the negotiations, would not be fully visible until somewhat later, when the Kurdish-CPA bargain was processed through the body officially responsible for enacting the TAL, the full IGC.

Consider the position of the IGC. Its own Drafting Committee, having recommended the bare outlines of a provincial federalism but conceding some undetermined special status for Kurdistan as a constitutional principle, left the actual work of producing a model (in accordance with both democracy and international law) to a freely elected constituent assembly that would be no longer under foreign occupation. Now they were suddenly given a “compromise” model of asymmetrical provincial-ethnic federalism hammered out in a state bargain by the Kurds and Americans. Did they have the power to resist this deal, especially when the strongest opponents of ethnic nationalism, the secular postnationalists, were weak on the council, and when the Iraqi nationalists among the Shi’ites were distracted by religious issues? One response, the line of least resistance, was that the Drafting Committee and especially the American Governance Team proceeded apace to fashion a TAL consistent with the basic idea behind the bargain. Indeed, as I will show, under the false assumption propagated by the Kurds that it was they who had surrendered the most because of the signature issues of “Kirkuk” and “oil,” a governmental structure was negotiated that would have made sense only if the powers of the federal units were much weaker, in line with the very early strategy of the CPA.¹⁰⁰

But then three new things happened. The Kurds decided to insist on the “Kurdistan Chapter” as a formal submission to the TAL, despite the fact that they did not get Bremer’s assent to this. Some of the Shi’ites became attracted to a symmetric model of purely ethnic federalism. And third, everyone became sidetracked by the mainly, but not exclusively, symbolic issue of the relationship of Islam to the state.

The Kurdish challenge came first, inviting the Shi’ite move on federalism as a response. What was originally a response to Bremer’s repudiation of a prior agreement now became a stake in Kurd-Arab negotiations, renewing the possibility of deadlock. And this released the Arab, mainly Sunni and secular members of the IGC, who now roundly attacked ethnic federalism and the danger of dividing Iraq.¹⁰¹ This time, however, a deadlock did not occur, for two reasons. One is that Bremer, though he did not formally accept the Kurdish laundry list not previously agreed to, given his bargain with the Kurds, also did not renew his own earlier strong stand against ethnic federalism in any way; thus in effect he switched sides. And a new and unpredicted division occurred on the Arab and even the Shi’a side: some of the Shi’ites, mainly around SCIRI, became interested (or revealed their interest) in an entirely ethnic-based, symmetric form of federalism where they could “have everything the Kurds have.”¹⁰² While the Kurdish idea of a voluntary union made most sense with a two-member Kurd-Arab treaty organization or confederation, it could be made compatible with an asymmetric structure negotiated with Bremer as well as three-member or five-member “federations” or “confederations,”¹⁰³ always with the proviso that the Kurds would retain a veto over all constitutional changes (preferably for all Iraq but at least as they concerned Kurdistan).

Once it became clear that only Kurds were getting a potentially strong regional government to protect their interests, in a setting where the federal government might be weak, the disadvantages of an asymmetrical structure for the Arabs became clear. This was especially true for some of the religious Shi’ites, who had a primarily southern base and who, like the Kurds, could hope to control large oil resources in their territory, whatever early arrangements for resources were to say on the matter. One could have assumed that the religious Shi’ites had the greatest interest in a strong majoritarian central government, because they could control it. This was so if they were united, but they were not. Regions with oil had different interests than regions without it, and the attitude to Iraqi nationalism was also different between the various parties (the SCIRI, Da’wa, and, outside the IGC, the Sadrists, to whom one did not want to lose the oil-poor urban vote). Moreover, now the Americans were helping to remove Kurdistan from central authority, and the negotiating process began to favor power sharing, as I will show, and thus a weak government no one would exclusively control. Thus having what the Kurds had made a lot more material sense, especially for those for whom symbolic issues of identity were increasingly articulated on a transnational rather than national level, concerned with Islam rather than nation. It may even be that this turn from Iraqi nationalism made their demands on behalf of Islam, the other possible center of identity, all the more vociferous. But more likely it was the other way around. Interest in the transnational allowed them to focus more on the homogeneous region than the heterogeneous state.

When demands surfaced from SCIRI (the Supreme Council for the Islamic Revolution in Iraq) for the ability of other provinces to form regions, the Kurds were, most likely, happy to oblige. They did not propose the idea, but among acceptable solutions this was for them marginally better than an asymmetric structure that might unite rest of Iraq, centrally organized against them, depending on the type of voting. A symmetric, ethnically based federation or confederation could effectively be the same as an outright two-unit confederation, where a veto by one is automatic.¹⁰⁴ More immediately, they were getting new and unexpected allies and supporters for the regional idea, their opposition was being split, and the TAL would be even more restrictive with respect to the freedom of action of the constitutional assembly. Indirectly, if the Shi’ites dropped their focus on the federal government, it would be more likely that the latter would be weak. Only the secular nationalists and the Iraqi nationalists among the Shi’ites, who may have recognized that the goals of Sistani were being undermined, objected.¹⁰⁵ Given the relative disinterest of the Kurds and the Americans concerning this particular outcome, genuine negotiations and compromise did follow, with the surprise ending of a separate meeting between the Kurds and the Shi’ites, on February 28, where Kurdish support for new potential regions (which they did not mind) and for some formula acceptable to the Shi’ites on Islam (more on this below) was traded for Shi’ite support for two official languages in the state.¹⁰⁶ It was probably because of such a deal, and not merely because of a concern in the IGC for greater equality among parts of Iraq, that it was finalized that the TAL (after the initial elimination of this provision) wound up allowing new region formation in the transition period, but only for three provinces, and with the permission of the majority of the national assembly, which was perhaps unlikely but hardly impossible for say three Shi’ite regions (and one Sunni region) if they wished to do so (a large Sunni region would be another matter, or a united, oil-rich Shi’ite region!).¹⁰⁷ None of this, however, would have happened if the asymmetrical structure was not first separately negotiated and then treated as a given by the CPA. At any rate, now the possibility of an Iraq composed of quasi-independent regions was established, and equally relevantly, the Kurds almost until the very end had a free ride home. But that was because the IGC, including the religious Shi’ites, still did not realize that while compromising on some details, they were also well on the way to accepting the deeper Kurdish scheme for the new Iraq as a voluntary union.

Finally and fatefully, the organized group on the IGC with the greatest popular backing, the Shi’a House, which had been rather uninvolved in most of the constitutional negotiations,¹⁰⁸ now mightily distracted itself in the middle of the critical, final possible negotiations over the state structure and wasted the remaining time (given the highly artificial American deadlines) to solve a largely symbolic issue in a largely symbolic way. This was the problem of Islam as an official state religion, and it was negotiated largely within the IGC as a whole. Admittedly, this issue was intertwined with the very important question of a Personal Status Law, which in principle would have a very significant effect on the lives of men and women all over Iraq, determining whether questions of marriage, divorce, custody, and even some property matters would come under religious or civil courts. But while the latter question was being resolved, an inordinate amount of attention was paid to wholly symbolic issues mattering far less.

It is probably true that the overall discussions concerning Islam and the state probably represented the most genuine, open-ended, bargained segment of the negotiations over the TAL, with several actors exerting strong pressure both inside and outside the formal negotiations, and to some, this particular part of the negotiations was evidence that the interim constitution was not ultimately an American-imposed one, or at least was not imposed unidirectionally, whatever that may mean.¹⁰⁹ To “neoconservative and evangelical voices within the U.S. political sphere” and Senators Santorum and Brownback, Iraq’s constitution had to be secular, even though all the Iraqi forces on the IGC, including the Iraqi Communist party, noted that every constitution in the Arab world made Islam the official state religion.

As a matter of fact, as the same author well knows, Bremer indeed helped to “impose” by his standing veto threat the withdrawal of a potentially new status law passed by the IGC on December 29, 2003, as resolution 137, which (voiding a law from 1959) was to make the Shari’a the foundation of family and civil law. The new regulation was to have been incorporated both in a so-called Personal Status Law and in the TAL, but neither happened. As far as I can reconstruct, during the first half of January this proposal was strongly challenged in the streets of Baghdad by a series of women’s demonstrations ranging from hundreds to thousands of participants. Undoubtedly, the anger of the demonstrators was fueled by the dramatic worsening of the conditions of women in many parts of the country, the return of honor killings, the enforcement of strict Islamic dress, firings from many jobs, and reprisals and threats against female political activists.¹¹⁰ While up to eighty-five groups were said to participate,¹¹¹ the leading role was played by two civil society organizations: the OWFI (the Organization of Women’s Freedom in Iraq)¹¹² and the Women’s Alliance for a Democratic Iraq (WAFDI).¹¹³

It was above all the larger and more radical demonstrations led by the OWFI that began the process that in the end led the IGC to revote and repeal resolution 137, although Bremer’s veto threat also must have played a major role. That threat may have been influenced by the appeals of the Women’s Alliance, and it is even likelier that the requirement that one-quarter of the National Assembly be female was written into the TAL as a consequence of their demand. Assuming that the figure represented a compromise between two factions of the IGC,¹¹⁴ Bremer’s changing one-quarter to every third person on electoral lists when writing the electoral law was probably influenced by demands addressed directly to him.¹¹⁵ On the electoral guarantees for women the Shi’ites were surprisingly pliant—or not so surprisingly, given their later ability to control their women MPs.¹¹⁶ It was different regarding the question of the status law. When on February 27, 2003, on the “surprise” (to whom?) motion of a secular Shi’ite female member of the IGC (obviously prompted by Pachachi, cleared by Bremer, and backed by a crowd of women allowed into the chamber) resolution 137 was withdrawn, prompting a walkout by religious Shi’ite members and their allies, there was outrage, even though the idea of getting a new status law into the TAL was probably already dead.¹¹⁷ My evidence: it was not even once again raised by Shi’ite members in negotiations over the TAL. However, the event was almost certainly timed if not engineered as a response to a Shi’ite consultation, the day before, with the Ayatollah Sistani, concerning the question of Islam in the TAL, and the Shi’ites interpreted it as an frontal attack.¹¹⁸

Thus the walkout, prompted probably as much by high-handed strategy as substance, was very real, and it sought, among other things—much too late in the game—formal bilateral negotiations with the CPA, like the Kurds had for a long time.¹¹⁹ When this was granted, there were only two issues left to negotiate: making sure the Shi’a could have in principle what the Kurds had, namely the right to form regions in oil-rich areas and the symbolic issue concerning Islam. With the first demand quickly granted given the prearranged support of the Kurds,¹²⁰ all energy was expended on the second. And here the CPA was unwilling or unable to act entirely on its own. Initially the question was whether Islam was to be “the” or “a” source of legislation. The first idea was clearly nonsensical for the very modern constitution the TAL was going to be, and the second did not require of the lawmaker much of anything. In the case of passing the first formula, everything would depend on the kind of constitutional court Iraq would have to enforce it. For the TAL, the Drafting Committee and the Governance Team had already settled on a mixed framework, despite some fundamentalist antijudicial views of one American member (oddly enough for a civil-law country), namely a “Federal Supreme Court” of nine (!) members (in reality a constitutional tribunal that shared the powers of a highest court of appeal with the Federal Court of Cassation: TAL art. 44b) who had to be picked according to difficult consensus requirements by a three-member Presidency Council from nominees drawn from higher and lower secular juridical councils. This body could rule laws and lower legal acts of any type unconstitutional, but it is doubtful if it ever would on the basis that Islam was not the source of the act in question, whatever that is even supposed to mean, beyond perhaps natural-law principles of justice (TAL art. 44, 2–3).

According to N. Feldman, Shi’ite Islamists, because of unfortunate promises made to Brownback and Santorum by Bremer, which “reeked of imposed constitutionalism,” upped the ante. They now demanded a new formulation: no law should be able to “contradict Islam.”¹²¹ Most likely, the Shi’ite Islamists simply realized that it was easier to show in a court a contradiction between a specific law and a specific Islamic tenet than to demonstrate that some law, any law likely to be proposed, did not have some relation to some kind of Islamic rule, practice, pedigree, or interpretation. This is pretty much the picture I get out of Bremer’s and Diamond’s converging but not identical retelling of the events. When, instead of returning to the IGC plenary, the Shi’ite House finally forced Bremer to negotiate with it directly on February 28, it got him to at least entertain the possibility of the new wording, as long as a totally redundant proposal concerning no law being able to contradict the table of rights was added as well.¹²² This was then sold as a compromise, having received the “green light” from Secretary Rice at Camp David and “cleared in advance with the Ayatollah Sistani” in Najaf.

So it went. I think if Sistani told the Shi’ite politicians anything, he may very well have told them that the new wording would be more effective anyway. But even the new formulation was going to achieve little without the right kind of body to enforce it, one that the Islamists were not going to get in the TAL at any rate, and there is no evidence that they were even asking for it. Still, if Sistani really was involved, the matter was now enormously important in their eyes on a symbolic level. The fact is, the matter was equally important for the secular side in the negotiations, and the Kurds who did not care one way or the other apparently joined them. Thus there was yet another fight in the IGC in the final hours before Bremer’s March 1 deadline, with the Kurds supporting the Shi’ites, who would not change anything but in the end agreed to add yet another meaningless phrase: legislation could not contradict “democratic values.”¹²³ The end result was actually a verbal compromise, if a quite meaningless one: not only laws contradicting Islam and the rights provided for in the TAL were to be forbidden in the transitional period (art. 7), but also laws contradicting the principles of democracy, whatever those principles were deemed to be. This, of course, was an entirely superfluous and redundant addition, since legal acts contrary to democracy or enumerated rights would be unconstitutional under other paragraphs of the TAL. The whole provision was moreover useless to either side until a Federal Supreme Court was created, which never was to happen, however “unconstitutionally,” during most of the life of the TAL. The most important thing about this whole incident was the amount of time wasted on it, the fact that it divided the secular (civic) and religious (Iraqi) nationalists on a very emotional issue, and, as we will see, that it involved so much time, energy, and divisiveness in the last crucial hours of negotiation over the TAL. And this is the case because, as the Shi’ites would realize very soon but still too late, something much more important was happening at the same time.

The American-Kurdish bargaining over the state did not stop with the special sessions in Kurdistan. The Kurds were supremely conscious of one flaw in all that they had achieved: a freely elected constitutional assembly could take it all back. Thus it was very important in their eyes that, of all their proposals, one was clearly not accepted, nor was its inclusion in the TAL seriously considered, namely, the provision of the “Kurdistan Chapter” (art. 5) that would allow the permanent constitution or any law replacing the TAL to apply to Kurdistan only if approved in a referendum of the region.¹²⁴ This was not a potential veto of the permanent constitution but, in line with the idea of voluntary union, the affirmation of one entity to agree to the replacement of one state treaty by another according to its own sovereign or quasi-sovereign decision. The issue was not raised as far as I can see during the February 28 Shi’ite-Kurdish talks, because the talks very likely would have fallen apart if they had. At any rate, Bremer does not mention it. I will initially follow his very revealing testimony, which is as interesting for its revelations as for its symptomatic omissions.

After the Shi’ite walkout was diffused, at 2 a.m. on February 29, according to Bremer, the Kurds suddenly presented a two-page set of demands, including demands for money, the legalization of the Peshmerga and, most crucially, “the right to veto the ratification of the constitution.” Since Bremer says these were the issues presented three weeks earlier, I am assuming that the last provision was still referring to a veto that would apply to Kurdistan only. It is at this point that Bremer supposedly convened the Kurds “in a small, dark work room,” complained about last-minute demands, and said they threatened (not he threatened, as Galbraith says, but the result is the same) the Kurd-U.S. special relationship.¹²⁵ It is this that Galbraith considers “more brutal” treatment, when very likely he was one of the people to help the Kurds understand the possible utilization of the weakness of the Americans generated by their own artificial deadlines.¹²⁶

Despite his complaining, Bremer nevertheless decided to sit down with the Kurds alone, before the IGC plenary, “to get agreements on the demands they had sprung the night before”¹²⁷ and, unsurprisingly, given the same structure and the lack of time on the American but not Kurdish side, the results were again the same. The Kurds, appealing to American political and public-relations interests, expressed their fear of a constitution that the “black turbans” (the Shia clerics) would write. Then they¹²⁸ proposed a new idea: ratification of the constitution by simple majority, as long as two-thirds of the voters of three provinces did not reject it. Bremer is quite right (along with almost everyone else) to interpret this as a Kurdish veto,¹²⁹ but he did not seem to notice or remark that this veto pertained to the ratification of the constitution for Iraq as a whole and not just for its validity in Kurdistan. Thus it went beyond even the veto contained in the “Kurdistan Chapter.” He may have been misled by the fact that in this formula voting would be by governorates and not by the region as a whole. The Kurds were thus asking for formally less but substantively more than before. It was a brilliant move.

According to Galbraith, the Kurds made the acceptance of the provision a condition of their supporting the TAL, a document that he repeatedly alleges represented a bad compromise for them.¹³⁰ I will try to show below that with the default provision already built into the TAL in case of the failure of ratification, this could not have been true whatever anybody said. The ratification veto would be no gain for the Kurds over the TAL without it, because all it would get them in case they hated the new constitution was the preservation of the TAL itself. Thus what they were trying to do in reality was enshrine their old gains and create the foundation for additional ones.¹³¹ But if this is true, then the claim that they would not accept the TAL, involving great gains, unless they got even more, was perhaps a huge bluff. Nevertheless, it is quite likely that the Kurds did tell Bremer that without the ratification rule they would not accept the TAL. If they did, given his self-imposed deadline and his own desire to limit the black turbans in advance, Bremer fell for what may have been only a negotiating ploy. He promised to consult Washington about the proposal and got Condi Rice’s agreement at 3 p.m., February 29, the last day.

The Kurds were told the good news at 5:15 p.m. At this point, in Bremer’s recollections something strange happens. According to him, between 6:45 p.m. and 4:20 a.m. (March 1), the plenary of the IGC met. “Much of the text was agreed on the day before,” and thus it could not have contained the ratification rule. For much of the night the question of Islam in the TAL was discussed, and the compromise—including its attendant redundancies discussed above—was finalized. This happened by midnight. After that the compromises were easy. At 2:15, Bremer managed to get the idea of reconvening anew quashed. There was a short debate, also quashed, on security arrangements with the United States. Finally, a unanimous vote, at 4:20 a.m., on the TAL as a whole was held. There was not one word of discussion on the central question of a dramatically new ratification rule.¹³²

Diamond, probably incorrectly, puts the IGC all-night plenary on March 3, but he supplies the missing details.¹³³ According to him, on February 29 Bremer was not able to finish as planned and was forced to allow “a seventy-two-hour marathon struggle to conclude the interim constitution.”¹³⁴ The decisive event described or rather omitted by Bremer supposedly came on the night of March 3, rather than February 29. At 4 a.m., an hour that corresponds to Bremer’s timeline, when all else was done, Massoud Barzani offered an amendment to TAL Article 61c that contained the new ratification rule. Without anyone really considering or certainly discussing what they voted for, the IGC adopted the change and the TAL was finished.¹³⁵

According to another version, SCIRI’s Adel Mahdi did raise an objection to the ratification rule, saying that Sistani would never accept it, but when asked for an alternative he could not provide it and instead asked for a day’s delay, which Bremer would not grant.¹³⁶ It is then that the text was unanimously accepted. This version does not make the event any different than Diamond’s description, though it makes voting for the TAL as it stood even more questionable on the part of the Shi’a House. A third version explains the voting by arguing that it was Mahdi himself, as the “Kurds’ advocate,” who persuaded the other Shi’ites, “who had not done their homework,” to go along with the plan, arguing that they could not afford to alienate either the Kurds or the CPA.¹³⁷ In any case, over the next day or two, when the Shi’ite leaders and especially Sistani’s circle realized the full meaning of what they had done, they strenuously and vociferously objected to the procedure and to the ratification rule, but too late. The Kurds would not budge, because now they had what they wanted, and Bremer refused to reopen any negotiations dealing with the TAL. The ratification rule was successfully imposed on the IGC as a whole—though that was not to be the end of the story.

This may be the place to ask why the Americans, the world’s one and only superpower, proved so weak with respect to the Kurds. It cannot be, as we have seen, because their positions were identical. Initially at least, Washington understood the demand for a Kurdistan Region in terms of ethnic nationalism, to which all sections of the U.S. government were opposed; moreover, they foresaw some of the disastrous consequences of going down that road.¹³⁸ So in this context at least we cannot speak of ignorance, only a lack of will.

First, there was a question of the bargaining strategy. This was evidently delegated to Bremer and his team, who were inexperienced and outgunned by the Kurds and their advisors. The CPA team never understood the significance of the two-sided structure, which they never should have granted, and this in itself must have cost them some concessions. Bremer may have chosen his approach because of his repeatedly documented desires both to personally micromanage and to do it all very quickly. The first desire, anchored deeply in his personality, he could not satisfy given the messy IGC format in Baghdad, while the second was imposed on him by the Republicans, with an eye to the U.S. electoral timetables. The two factors partially explain why so much was conceded. The experienced leaders of the Kurds were not ashamed to rely on real experts, and they had much more time—as far as they were concerned, they had been waiting since the treaty of Sèvres. As their advisors must have explained to Talabani and Barzani, the Americans meant Bush, and Bush had to have results quickly if he wanted to be reelected. But even so, threats of secession or of not signing any version of the TAL were all pretty hollow, and at the very least Bremer could have called the Kurds’ bluffs or held out longer when his bosses tried to do so.

There had to be other things also at work. Consider the fact that while the Sunni-led insurrection was already unfolding, the Shi’ites led by Sistani mounted a series of tough constitutional challenges that forced the Americans to back down or compromise. Not only in the war against Saddam but in all these later fights, the Americans and the Kurds were on the same side. They also shared more or less their opposition to the Islamization of the new Iraq, whether this fear was real or imagined. For Bremer, these issues and fights were intensely personal, and his distrust and dislike of Sistani seems to have greatly increased over time. Was it therefore the Americans’ far greater trust in the Kurds than in any of the Arab factions that made them their prisoner, or was it a common desire to stop the “black turbans,” the friends of Iran, from inheriting the prize the war was about to give them—or did the Americans, in classical imperial fashion, actually aim at a weak Iraqi state open to influence and penetration?

The several failed attempts from Washington to deprive the Kurds of their Kurdistan Regional Government, whose potentially negative consequences for the integrity of Iraq as a whole were obvious, is consistent with the first or second but not the third interpretation. Even if initially American positions on the structure of the state were opposed to those of the Kurds and implicitly at least were closer to those of what seems to be Sistani’s Iraqi nationalist position, that initial preference was altered precisely because of the Americans’ very serious conflict with Sistani. The Kurds were able to successfully represent a position according to which even a relatively united Iraqi federal state (which is what territorial federalism implied) would easily fall under the domination of the majority and thus the “black turbans.” In order to prevent such an outcome (which probably also would have been seen as a victory for Iran in the region), the American negotiators first softened their stand and then accepted the basic Kurdish demands, toward which the negotiating format was already predisposed. Finally, when it came to the Kurdistan veto, the Americans were predisposed to accept this simply because of their own desire to limit a constituent assembly they would not be able to control as easily as they could the makers of the TAL. Very likely the change of form from an ethnic veto to a three-governorate model, whoever thought of this move, was also helpful in gaining the incredibly rapid acceptance by Bremer and Rice. That was partly a function of time, but given the radicality of the proposal, we have to assume that the experts on the National Security Council decided that the combination of rules of change corresponded to the American interest in controlling the Shi’ite majority.

With the eventual convergence of interests between the Americans and the Kurds, it seems to me nevertheless very important finally that the Americans could not occupy Kurdistan, because of the Turkish refusal to allow an invasion from the north, and thus did not have any obvious military means to force the Kurds to do anything.¹³⁹ This led to the mistaken view that they had to be treated completely differently than those who were supposedly under their full physical control and “tutelage.” Thus the origins of the two-sided negotiations, which came to resemble a pact between two holders of state power in two independent units, conceding the most fundamental point to the Kurds right from the outset, may have had more of an effect than even Bremer’s very real mistakes in strategy. And, given such a bargaining structure and the Bush administration’s time constraints, the Kurds’ ability to outwait the Americans always proved decisive. But whatever the reasons, once the deal emerged among Bremer, Barzani, and Talabani, to put the matter crudely, it became their consensual position, and the rest of the IGC or the drafting committee of five was in no position to do much about it. At best what could still be done, if inconclusively, was to generalize some of the concessions given to the Kurds to the Shi’ites as well, with provisions such as the ability of any three provinces, potentially, to form a region, and the firm intention to hold on to other militias under the same cover given to the Peshmerga. These concessions only confirmed the fears of those who thought that conceding an ethnically defined, asymmetrical federal structure to the Kurds would keep Iraq on the path of state destruction rather than lead to a model under which the state could be rebuilt.

A final note on this point: had there been an Iraqi state and an Iraqi army, the negotiations with the Kurds could not have taken the form they did. In that case, most likely the Americans would not have been bargaining in the place of Arab Iraq, and even in the unlikely case that the Americans had been, they would have had a very powerful bargaining chip that the Kurds could disregard only at their own peril. Whatever compromise was struck, it have been struck taking this factor into account.

Another way to examine the process is from the point of view of the result, the Transitional Administrative Law itself. Ultimately, even if the process failed from a normative point of view, had the interim constitution itself nevertheless embodied a “historic compromise,” as some of its defenders claim, it could have contributed to its own effectiveness and legitimacy during the course of its roughly year-and-a-half existence. Here I wish to examine the question of to what extent the interim constitution represented such a substantive compromise, with respect to three areas where the TAL did have great deal of material constitutional relevance during that period: the territorial dimension of the state structure, the model of government, and the rules of constitutional change. I will not repeat my discussion of the relations of state and church and the significant symbolic compromises made, but rather I will focus on where local power rather than the TAL’s prescriptions were to decide, for example, what kind of status law Iraqi citizens were going to live under. Many readers, especially those who interpret all constitutions in terms of their tables of rights, will find it strange that I will not pay any attention to the impressive and well-drafted table of constitutional rights in the TAL. This I do for several reasons. First, during the life of the TAL no apparatus of courts (until the end, almost no constitutional court) was created that could imaginably enforce such rights against the “state.” Second, I consider it a laughable proposition that such protections were enacted while they were not meant for a moment to apply to the main actual holders of the means of violence, the military forces under coalition command.¹⁴⁰ Third, and most important, rights against the state and rights to state outputs are relevant only if there is a state, and the TAL failed, as I will show, in its attempted reconstruction of any plausible state in terms of territory, people, or form of organization of the exclusive means of violence. Of course, the rights of the TAL also did not, do not, and could not apply to the decentralized forces holding the means of violence in Iraq, the insurgencies and the militia, the latter having infiltrated the new army and the police, and thus it is hard to see what the relevance of constitutional rights is supposed to be, outside of a symbolic and public-relations one. Let me note finally that given my interest here in negotiation and compromise I detected no debate, conflict, or need for compromise concerning the table of rights outside of some technical suggestions made by American participants to Arab drafters. No one thought it worthwhile to spend a lot of time on this; it was permissible to let Mr. F. Istrabadi, using American, international, and Middle Eastern precedents, just go ahead and draft.¹⁴¹ To his credit, he seems to have wished for the extension of rights protection to the American forces, which was equivalent, however, to wishing the occupation to end, because the United States has never made such guarantees of putting its forces abroad under a foreign constitution and foreign courts, even if they actually existed and were viable. I do not discount the symbolic and programmatic importance of asserting the rights of the TAL, but I fear that they now are also infected by the interim constitution’s illegitimate origins. The future possible role of these rights in Iraq and the Middle East depends much more on the solution of the other problems that were treated in the TAL and the permanent constitution. It is to them I now turn.

The TAL’s solution is the asymmetric structure composed of Kurdistan, still containing formally at least three provinces and two fragments, and fifteen other provinces, with only Kurdistan having a regional government. But as I already detailed, any three provinces could form a region if their electorates and the National Assembly approved (art. 53c). Thus a bridge was provided to a symmetrical system of a federation of ethnically based regions. To be sure, this bridge explicitly and doubly excluded Kirkuk (and Baghdad too) for the transitional period as far as Kurdistan, already composed of three regions, was concerned. But that in general was the result of the CPA-Kurdish bargain, the status quo plus solution, that led to the TAL delaying the final resolution of the territorial question of the disputed province and the city for the final constitution. The adjective plus was now really earned (it is unclear to me when these provisions were agreed on and why there was no strong Arab response),¹⁴² because the TAL undertook a series of measures that had the effect of beginning the reversal of the Saddam-era ethnic cleansing, by altering the property and demographic structure of especially Kirkuk city, the latter not only by resettling Kurdish expellees, which is understandable, but also by expelling individuals (read Arabs) “newly introduced” (art. 58a 1 and 2). Here the TAL went well beyond the “Kurdistan Chapter,” which was silent on all Kirkuk-related questions. In line with that spirit, while the resolution of the final territorial status of Kirkuk was left to the permanent constitution, this would now mean a period after the completion of the measures dealing with demographic and property transfers. Only then would the relevant population, reconstructed by these measures, be finally consulted.

As to the division of powers, the TAL takes over the formula of the “Kurdistan Chapter” (art. 1, sec. 3) of assigning all unenumerated powers in Kurdistan to the Kurdistan Regional Government except those powers where the “Provisional Government of Iraq,” now called the Iraqi Transitional Government, has exclusive (enumerated) competence.¹⁴³ But the TAL again goes beyond the “Kurdistan Chapter” in one respect, by giving a nullification right, “the right to amend the application” of all federal laws, with the exception of laws within the exclusive competence of the federal government and decisions of federal courts in the areas of that exclusive competence (TAL art. 54b). Where the TAL did not satisfy maximal Kurdish demands was in its enumeration of the very short list of exclusive federal competence, which included the uncontroversial areas of foreign policy, fiscal policy, monetary policy, weights and measures, citizenship, and immigration; the very controversial areas of defense, securing borders, and resources (national management and ownership); and the somewhat controversial area of public communications (TAL art. 25). The “Kurdistan Chapter” sought a recognition of the Peshmerga and its transformation into a Kurdistan national guard (art. 2) as well as the “non-deployment of other Iraqi armed forces in Kurdistan.” These were not granted, and by implication, Kurdistan’s borders too would be guarded by federal forces. Moreover, the TAL expressly required the dissolution of all militias (art. 27b), and in the eyes of the Americans and Arabs the Peshmerga was a militia. Note, however, that from the Kurdish point of view the Peshmerga was Kurdistan’s national army and not a militia, and there was never the slightest intention to dissolve them or cede the defense of Kurdistan’s borders to any other force.¹⁴⁴ As to oil, the Kurds were not in a similar bargaining position. Present resources were around Kirkuk, which they did not control militarily. Moreover, they wished to battle the impression that they wanted the city for its oil.¹⁴⁵ It was not even completely clear (the Kurds said) if they would gain or lose from national ownership, given the greater reserves of the south.¹⁴⁶ Thus they sought a formula in the “Kurdistan Chapter” (art. 3) that would grant Kurdistan ownership, but management of current fields could still be left in federal hands as long as the region received its just share of the proceeds. The TAL did not grant this mixed formula, for the moment keeping the ownership and management of all fields and not only the management of current fields in federal hands, promising only some affirmative action to areas previously unjustly deprived of the proceeds (art. 25e).

There are various ways of speaking about this package concerning the territorial structure of the state. Galbraith seems to treat it as a defeat of Kurdish positions (“Bremer got most of what he wanted in the Transitional Administrative Law”)¹⁴⁷ but then goes on to describe the far more essential things conceded to the Kurds. All the later outcry, when the TAL seemed to be under the threat of abrogation, is wholly inconsistent with this perspective. Others more convincingly speak of a historic compromise¹⁴⁸ or of the TAL as a compromise between strongly divergent Arab and Kurdish proposals.¹⁴⁹ Undoubtedly in terms of contents, the matter can be looked upon this way: both sides got some but not all the things they wanted, leaving aside the matter stressed by Galbraith that the Kurds never intended to or could be made to fully deliver on their side.¹⁵⁰ What is deeply unconvincing is to say that the compromise consists of the asymmetrical model itself, with Arabs getting the type of state they wanted for Arab Iraq, a centralized federation, and the Kurds getting what has been called a “federacy” for Kurdistan.¹⁵¹

The asymmetrical model itself was an expression of the fundamental Kurdish drive to understand the new structure as a voluntary union. Let Arab Iraq be organized one way and Kurdistan be organized the other way, and we can arrange a few remaining common affairs by something like a treaty. That was the position of the Kurds before the beginning of the bilateral negotiations with the Americans (and this was conceded by the CPA by the structure of those negotiations, though never in so many words), and now the TAL projected a state structure much more compatible with this assumption of a voluntary union than with any conceivable notion of a coherent state organization. Granted, there was already a way of making it coherent in the TAL, at the risk of its disintegration.

The new term “federacy” expresses a quasi-confederal enclave in what is otherwise supposed to be, but in my view cannot be, a true federation and especially a federal state. At issue here first and foremost is not how the parts (Arab Iraq and Kurdistan, the supposed two halves of the compromise) are organized but how they are related. O’Leary has defined “confederation” as that type of “federal” system where the units retain sovereignty.¹⁵² Less legalistically, we might say that a confederation is a system where the political center has no one area where it is relatively autonomous in making policy with respect to the whole country, while in a federal state both center and units have autonomous policymaking powers.¹⁵³ Kurdistan is supposedly a federacy, not a unit of a confederation, because in international law at the very least it is not sovereign; Iraq as a whole is so recognized. Moreover, the Iraqi government does have or seems to have autonomous policymaking powers. A federacy, O’Leary says, is semisovereign, but that term is entirely meaningless.¹⁵⁴ And here is where formal understandings of sovereignty lead us completely astray. Kurdistan has its own army, under its own control, and is unoccupied. Iraq does not, and it is. Kurdistan has its own constitution and its own government; Arab Iraq does not, only the federal government does, that is, the government of Arab Iraq. This is, of course, the very asymmetrical arrangement itself, but more is involved than that. As for the TAL, it has given very few powers to the central government that are enforceable in Kurdistan, but it has given all unenumerated powers and rights of nullification to the KRG. On the crucial question of the constituent power, Kurdistan can veto any constitution that would apply to it and indeed to the rest of Iraq. It is the only such veto from the ethnic point of view. The Shi’ites have the majority, that is, more than a veto, and the Sunnis have less than veto power, as we will see when we examine both the amendment and the ratification rules. Thus while the whole is not a treaty organization formally, the Kurdish “federacy” comes close to being a unit of a confederal state or confederation whose constitution is actually a treaty that can be made and altered only with the full consent of the contracting party, a “voluntary union.” The reason why Iraq does not appear to be a confederation formally is because its “federal” government does have autonomous powers. But in reality it has them only with respect to Arab Iraq, and this is what allows Kurdistan to be a federacy, also formally indicated by its powers and veto rights.

According to O’Leary, when there is a federacy within a federation, neither side has the right to unilaterally alter the division of powers between the units and the federation—that is, the constitution that entrenches these—thus the implication is that there is neither federal constitutional supremacy as in a federal state nor the right of secession as in a confederation. Today, however, the unilateral right of secession is no longer the right of members of treaty organizations, whether or not we want to retain the term “confederation” for a treaty organization where such a right is spelled out.¹⁵⁵ Thus a federacy still can and does signify membership in a kind of treaty organization. The peculiarity of Iraq is that this is a one-sided relationship according to the TAL, because the Kurds, who have to agree to all changes of the constitution and thus can consider it a treaty from their point of view, alone have the requisite autonomy of the federal government, even as they participate in both governments in Iraq, regional and federal. No other part of Iraq has the same autonomy, at least formally speaking, because in reality the federal government is an extremely weak one with respect to all parts of the country. Reality is more symmetrical than the formal relationships. Despite the formal asymmetry, the structure encourages the three major ethnic-religious groups to treat the constitution as a treaty, though the Sunnis do not alone dispose of a veto power over it. They too, however, like the other two ethnic groups, have the actual power to control large parts of the country.

Thus we do not have to go to Arab authors to get a wholly different evaluation of the TAL than “Kurdish defeat” or “historic compromise.” Amazingly enough, sometimes in the very same texts a few pages later, advocates of the Kurdish cause go on to say that what is much more true is that the same TAL “marked an overall defeat of the integrationists, the national, centralist, and majoritarian federalists of Iraq.”¹⁵⁶ An overall defeat is very different than a great compromise. But the reason why the first formula is right and the second wrong can be understood only if we look at the other parts of the TAL, especially its governmental, amending, and ratificatory formulas.

From both the Kurdish ethnic and the liberal nationalist positions, powerful regions needed to be linked to Iraqi government through models of governmental power sharing, preferably of the consociational type. In my view, both empirically (think of India) and logically the links between the three dimensions (ethnically based federalism, strong powers for the units, and power sharing in the center) are tenuous, though there are many cases admittedly when all three vary together.¹⁵⁷ The most I would concede is that ethnically based federalism in a context of strong divisions may require the additional protection of either strong unit powers or participation in the government of the center. Being protected twice over can be at times necessary but can also be superfluous. It can even be the source of strong new resentments, when a minority nationality that has successfully separated itself nevertheless controls the lives of the rest through guaranteed participation and vetoes.

The motives of the two major Kurdish positions in seeking power sharing could be slightly different to the extent that the ethnic nationalist seeks only to weaken while the liberal nationalist also actually wishes to participate in “Iraq.” Thus typically federalist forms of interlocking the region and Iraq could be more interesting for the liberal nationalist, if the proper formula for a second chamber could be found, a difficult task given the provincial organization of the rest of Iraq. For the TAL at least, the Kurds thus sought consociationalism. This was all the more so¹⁵⁸ because the TAL did not grant their region all the powers they wanted, and thus they needed more powers in the transitional federal government. But perhaps because they did not all seek it for exactly the same reason (weakening versus participation), here their bargaining power was somewhat diminished. It was also diminished because the governmental structure, unlike the state structure, was produced by American drafters with some prejudices for strong government (the Governance Team) in interaction with the IGC, rather than in the two-sided format between Bremer and the Kurdish leaders.

Why did the latter not hold out for their preferred negotiating structure with respect to the institutions of the federal government? I can only offer some hypotheses here. First, this issue was far less important. Implicitly, the Kurds too must have realized that they were now seeking a third level of protection for their ethnic federalism. But the issue had some import, and the creation of a purely majoritarian central government for Iraq could have represented a great de facto danger to regional independence won de jure. Second, it is also very possible, though I have no proof, that since Bremer was going to impose the results of the state bargain on the IGC, it was understood explicitly or implicitly that some kind of balance between imposition and bargaining had to be kept if the Shi’ite majority was not to bolt the process. Thus either Bremer refused to give the Kurds more or the Kurds and their advisors knew where the limits of the possible lay and thus did not even ask to include the government institutions in the two-sided negotiations. Most likely, the latter possibility actually happened, because there is no trace in Kurdish proposals of a governmental structure for Iraq. We know, however, that they eventually did seek to recoup some losses on this level and that they also believed that when it came to the final constitution, they could trade back some prior concessions on governmental institutions that made Iraq potentially ungovernable for even more regional independence than was provided in the TAL and for movement on Kirkuk.¹⁵⁹

We know how allergic Pachachi, the chair of the Drafting Committee, and Istrabadi, the active link between that body and the Governance Team, were to power sharing, especially on an ethnic basis.¹⁶⁰ The latter even wanted to introduce a second parliamentary chamber early, with incentives against region formation. This proposal was totally unacceptable to the Kurds, who could not accept such a body on a provincial basis at all, and in any case it was not common for constitutional assemblies. We also know that American advisors such as Diamond strongly (and rightly) pushed for a parliamentary rather than a presidential form of government, which is not as easily open to power sharing. Formally speaking, this side left a strong impression on the TAL. Iraq’s transitional government both before and after elections was to be parliamentary, and after the election of the single-chamber Transitional National (Constituent) Assembly the prime minister and his ministers (individually) had to get and retain the assembly’s confidence by a (simple) majority vote (TAL art. 38a). The prime minister could also dismiss his ministers on his own, with the majority of the National Assembly behind him (TAL art. 41). But this parliamentary institution was surrounded by a series of partially presidentialist and quasi-consociational structures. There would be a three-person Presidency Council, first selected (TAL annex) and after free elections elected by the National Assembly, on a single slate, by a two-thirds vote (TAL art. 36a). While it remained unspoken, everyone understood that this had to mean conventionally one Shi’ite Arab, one Kurd, and one Sunni Arab member, even though a Shi’ite deal with one of the other groups could possibly subvert that imagined convention. The Presidency Council, having to act unanimously, would then select the prime minister, and this meant that a president or vice president of any of the three groups could veto the choice (TAL art. 38a). Unless a candidate presented a power-sharing cabinet, he could be vetoed. The same would happen after a vote of no confidence in a prime minister, when a new government had to be appointed, which would make the parliamentary majority think twice about whether to use the instrument at all (as of this writing three years into Iraqi “parliamentarism” there has not been even one such vote, not to speak of removal). Even the prime minister’s removal of ministers runs into the problem of reappointment by full consensus: it could not be used to transform his government in a majoritarian direction, as I once thought. Only if the three “presidents” could not agree for two weeks would parliament be able to nominate and elect a prime minister and a cabinet, but only by two-thirds vote. This might overcome the veto of one group, but not of two, and it certainly is not a majoritarian or parliamentary device.

I believe that the presidents, each of them, were also meant to have veto power over legislation. But the postulate that they have to act as a unit in this case leads to the possibility of “a veto of the veto” by any of the three members. Was this result intended? Was it a result of a compromise? Was the trick of substituting one negative concept (“may veto”) for a positive one (“must sign” or “must approve”) (TAL art. 37) the revenge of a majoritarian drafter on a consociationalist bargainer, unnoticed by the experts of the Kurdistan Regional Government? Or did the liberal nationalist experts notice but forget to mention the fact to their ethnic nationalist colleagues? When the amendment rule was crafted (TAL art. 3), the drafters did know how to write the positive phrase “unanimous approval of the Presidency Council,” but in that case, with three-fourths of the assembly concurring, the device was redundant. The fact is that “the veto of the veto” somewhat vitiated the consociational character of the structure and would have made the transitional assembly more capable of legislating by simple majority (!) if it really wished to do so (which in fact it did not). The supposed power, a kind of impeachment after “due process,” of the Presidency Council to dismiss the prime minister or ministers also runs into a veto of such a decision by a single member (TAL art. 41).

On the face of it then, the governmental structure is a bargained compromise between majoritarian parliamentary, consociationalist, and vaguely presidentialist forms, even if it is hard to know just how it was negotiated. Again, however, the structure does not indicate what the Kurds claim, namely that the Arabs, or more exactly the Shi’ites, got a strong, majoritarian government for the Arab part of Iraq while the Kurds got safeguards to preserve their freedom in Kurdistan. While the latter half of the sentence is undoubtedly true, the former is not. In the process of getting some consociationalist safeguards that would make it hard to form a government that would attack the rights and powers of the Kurds, the rest of Iraq got both a mixed and weak structure. Given a plurality of parties, it would be hard to form a government under it, and the government formed would contain too many built-in vetoes. It would be very hard to replace an ineffective government. And as we will see later, it would be almost impossible to change the constitutional structure that caused all this except in the direction of making the same built-in problems even more serious.

Reasonable people can and will disagree concerning the type of compromise between Arab and Kurdish positions that went into the state and governmental structure of the TAL. One’s judgment will inevitably depend on what one takes to be the real initial positions of each side, and that is not fully knowable because these are always disguised for negotiating purposes. If we assumed, for example, that the Kurds really wanted to secede, there would be merit to the position that they sacrificed the most in the TAL, which does not even grant them the right of unilateral secession. I of course make the contrary assumption and think they must have been really surprised at how easily the world’s last remaining superpower conceded their fundamental understanding of Iraq as a voluntary union, which was against its own interests as most of its policymakers initially understood it. My case cannot be made foolproof by referring to the bargaining structure in general nor even to the bargains themselves as they pertain to the state and governmental structure. That is because there was some (I think mainly two-sided) bargaining, there were some (I think too few) concessions to Arab positions in these two areas, and it may all come down to what one chooses to emphasize and sometimes which sources one believes.¹⁶¹ In my view, the issue can be fairly definitively decided when we turn to the topic of constitutional change and the deeply related one of constitutional identity. Here the key rule has been imposed, I will argue, producing a constitutional identity that fully incorporates the Kurdish view of a “voluntary union” and enshrines it against future legal alteration.

The interim constitution was to govern Iraq for a transitional period, but more importantly it was also to govern the making of the permanent constitution. Here the political reasons for adopting this instrument and its historical meaning fully coincide. The idea is both to apply constitutionalism to the transition period and to constitution making itself, both processes being open to the possibility of dictatorial concentrations of authority.¹⁶² While it is the governmental and state structure that regulate the transitional period, it is through its rules of change that the interim constitution regulates the making of the permanent constitution.¹⁶³ These rules of change are not just the ratification rule, whose imposition I already discussed, but three rules, the amendment rule, the ratification rule itself, and what I called the “failsafe rule,” which O’Leary and his colleagues have called, even more correctly, the “default rule.”¹⁶⁴ Let me treat them in that order and then evaluate the package in relationship to the problem of constitutional identity in the TAL.

Contrary to the intentions of the American framers of the November 15 Agreement, who probably sought to maximize thereby their own input into the permanent constitution, the TAL was in the end not made unamendable. I should emphasize again that interim constitutions seek to accomplish two apparently contradictory things: to impose constitutionalist restraints on government and to allow learning during the period of political transition.¹⁶⁵ Formally speaking, an unamendable interim constitution would sacrifice the second goal for the sake of the first. Indeed, contrary to a published view of Feisal Istrabadi, this duality would generally call for amendment rules somewhat less difficult than those of permanent constitutions.¹⁶⁶ Because of such considerations, and not because of a conflict and compromise in the IGC, the TAL abandoned the initial prejudice of its planners on this point, but incompletely and partially. In fact, all of its amendment rules belong to the most difficult set of types among the world’s constitutions, and they are significantly more difficult than the amendment rules of the interim constitutions I know of, including South Africa’s, a country en route to being a federal state, a condition generally associated with difficult amendment rules. They “reflect” not so much the initial American intention but more the establishment of a Kurdish quasi-confederal enclave or a federacy in Iraq. Whatever their reason, these rules were nevertheless a serious impediment to legal constitutional learning and an implicit invitation to illegal learning.

There were in fact four distinct amendment rules in or associated with the TAL. Of these, the first and second below were mentioned in the document, the third follows from the language, and the fourth, soon moot, was only implied by the conditions of its authorization:

1. A good part of the TAL was amendable by a vote of three-fourths of the National Assembly and the unanimous consent of the three-member Presidential Council (art. 3a), who were conventionally assumed to be one Shi’ite Arab, one Kurd, and one Sunni Arab.

2. However, there were extensive unamendable provisions according to the same article (3a): rights covered under chapter 2, the timeframe of the transition as defined by the interim constitution, the powers of regions and governorates, and regulations having to do with Islam specifically and religions in general. Regarding both 1 and 2, after free elections, this self-referring rule could be used to change itself before any other part of the constitution could have been changed through the new rule.¹⁶⁷ Politically, such a “revision of the revision” is always explosive and rarely wise.

3. Since there would be no National Assembly until free elections could be held between late December and January 2005, the interim constitution was by implication unchangeable for the period from June 1 to some time after the elections, when the new National Assembly first met and chose a Presidential Council, as required by 3a.

4. What was not mentioned but was probably inevitable was that the CPA itself as the sole source of authority could have changed any part of the interim constitution before its full entry into force, that is, between March 8 and June 30. Since the CPA never acted in a constitutional capacity on its own, the latter date was probably June 1, when in fact the IGC used the very last opportunity to add the short annex.

Rule 4 was soon moot. As to the third “rule,” also unstated formally, it requires no analysis to show that unchangeability during the period between June 30 and, say, January 31 meant extreme rigidity for at least seven potentially crucial and difficult months. How bad an idea this was may be a matter relevant only to future constitution makers, because there were no important projects to revise the TAL in that period. Strictly speaking from the legal point of view, an unamendable constitution can only be replaced in its entirety, though of course politically speaking partial illegality regarding its application (through creative interpretation, disregard of the amendment rule, and so on) is also possible. In either case, the transitional legal order that interim constitutions are meant to establish and protect would be severely endangered. As we will see, it was largely the latter (disregard of the amendment rule) that occurred in Iraq.

The situation only gets worse during the period when amendments become possible (rule 1), because now there was a freely elected assembly that could run into governability problems and could still not easily modify the rules that caused them. Again, the rigidity of the rule encouraged full abrogation or illegality (and their anticipation as well, as preemptive moves) during a potentially extended period, especially long if the TAL were preserved because of the failure of constitution making. The formal amendment rule constituted an extreme consociational limit (as opposed to the confederal limit of the ratification rule) on the changing of the interim constitution. Theoretically, three-fourths of the National Assembly and the agreement of all three members of the Presidential Council could actually pass amendments to the interim document (except its unamendable sections) in the potentially extended period during which the permanent constitution was being drafted. Practically, the representatives of any of the three ethnic-religious (Shi’a Arab, Sunni Arab, and Kurd) groups were likely to have over one-quarter of the seats and possibly, if the convention held, one member of the presidential council. Thus any of them could veto any amendment twice over. It is very possible that Iraqi nationalist deputies, or strongly secular deputies, or deputies from Baghdad, or deputies from oil-rich regions, or any other combination could also have one-fourth of the votes in the assembly. Thus the possibility for vetoes of amendments could be greater than the framers of the rule initially imagined. Of course, some amendments might still pass through bargaining and compromise. But when in the face of a little less than three-fourths of the deputies and let us say all three members of the presidium, or, alternately, 95 percent of the deputies along with the president and one vice president, an important amendment could be blocked, the likelihood of a “runaway convention” that refused to be bound by the will of the illegitimate Governing Council or the foreign CPA would be great indeed. Or alternately, the provision encouraged unconstitutional amendments made by lower legal acts that could not be reviewed if a Federal Supreme Court was not set up or, alternately, could be passed with such a court’s connivance. Again, all these solutions would encourage a culture of “illegality” on the governmental level and preemptively so on the regional one.

All the same, the amendment rule was for potentially a short period, and the design of the institutions already described was not so disastrous that it would have to be used repeatedly. Their importance is great only in relation to the ratification rule, which, unlike the amendment rule, was imposed in an entirely indefensible manner and was greatly resented by almost everyone except its Kurdish proponents and beneficiaries. The amendment rule meant that either this rule could not be changed at all (because it related to the powers of regions) or could be changed only if the Kurdish parliamentary faction and a Kurdish member of the Presidency Council both agreed, the second being relevant in case of a split in that faction. The same amendment rule guaranteed that the TAL as a default position would not be changed, thus strengthening the threat of nonratification. Thus the three rules discussed here very much belong together. Without this amendment rule, the following two rules would be useless to their beneficiaries, because without it a new constitution could be adopted as a single amending act.

1. 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.” It was not said by what vote the assembly had to agree on a draft. The most convenient way of reading the text was that the draft of the permanent constitution fully replacing the interim one had to be approved by 50 percent plus one vote of the National Assembly,¹⁶⁸ and then, as clearly stated, by 50 percent plus one of the population as a whole in a national referendum, as long as two-thirds of the voters of three governorates did not vote against ratification (art. 60–61a–c). Having to do with governorates, this rule may have been meant to be unchangeable, but that restriction would be useless, since the amendment rule was not enshrined.

2. In case ratification failed (61e) or the National Assembly failed to produce a permanent constitution initially by August 15, 2005, without changing that deadline through constitutional amendment, the National Assembly was to be dissolved, new elections called, and Iraq was to continue to operate under the TAL with all deadlines changed to keep the making of the new constitution within one year. Since nothing else was to be changed, I interpret this provision as depicting a scenario that could have happened over and over again, making the TAL the default or failsafe constitution of the whole process.¹⁶⁹

The first rule will be called the ratification rule, and the second the default rule.¹⁷⁰ I consider the issue, discussed above, of whether the Sunnis could have used the ratification rule to their advantage no longer worth debating, because even O’Leary and his colleagues contradict themselves (or rather, they contradict O’Leary and Galbraith) on this point and because the actual referendum bore out the view of most participants, from Bremer to Sistani.¹⁷¹ Sunnis had nothing to do with introducing the measure and ramming it through as a veritable coup against the rest of the IGC. Since they did not like the TAL, they had no interest in a veto that would make the TAL a default position. Thus the ratification rule did nothing else (at least originally) than give an absolute veto to the Kurdistan Region (acting through the three Kurdish provinces) over a constitution passed by the majority, which could be either an absolute majority or a very high, qualified majority of the National Assembly. Whether it was ironed out between Bremer and the Kurds or introduced by the Kurds only after Bremer rejected a veto for the Kurdistan Region as a whole, the three-province structure was a very formal concession to the Americans. The Kurds, who were intent in eliminating their governorates (regarded as “Trojan horses”) and their administrations, now could not do so until all the referenda on the final constitution were over.¹⁷² Three votes were needed to defeat the constitution, whatever the overall vote of the region was. But the voting difference here was always based on an illusion, a pure formality. There was no chance that a three-province vote and a Kurdistan Region vote would produce different results. The rulers of the region, who were united on Iraqi constitutional issues and who knew how to operate by democratic centralism, controlled both.

Thus regardless of its formal concession to provincial federalism, the rule treats the creation of a new constitution like the making of a treaty between two partners, which needs the complete agreement of the two as wholes and not just some of their population (one-third of each Kurdish province could agree with say, 100 percent of the rest of Iraq, and still the constitution would fail). A federal law (such as U.S. Article V, an amendment rule rather than a ratification rule) could have involved the passing of the constitution by a high, qualified majority, and its ratification (like Articles VII and V) by nine-thirteenths or three-fourths of the provinces.¹⁷³ In appearance, this is what TAL article 61c does, since if the majority were joined by just one of the three Kurdish provinces the constitution would pass. But this would have been like asking South Carolina, North Carolina, or Georgia of the original thirteen states to vote for a constitution banning slavery, an illusory hope regarding even far less momentous issues. As long as the two Kurdish parties stuck together, the uniformity of voting in the three provinces would be guaranteed (and if they did not stick together, their federalist and confederalist aspirations would be finished anyway). Thus the rule was and was meant to be a quasi-confederal rule for a treaty organization or at least for an asymmetrical structure that has a “federacy,” that is, a confederal enclave attached to it by something like a treaty. With respect to the ratification rule of the “Kurdistan Chapter,” which was frankly and openly treatylike, the new rule gave the Kurdish parties everything the original rule had and more, because now they could actually deprive the rest of Iraq of having a new constitution. This was going to be an important source of power when it came to negotiating the final constitution, as O’Leary and his colleagues openly foretold.

The amendment rule and the ratification rule one attaches to a constitution gives important clues to how the rulemakers regard that constitution. The more rigid the rules of change, the more attached the rulemakers are to the constitution. The rules discussed here indicate the high regard the Kurds had for the TAL and belie the interpretation that they regarded it as a defeat and only accepted it because of the ratification rule itself. The proof is the default rule. In case the Kurds were forced to defeat the new constitution, they could always go back to the TAL and its rigid amendment rule. To their opponents, that would be a threat, but to them it had to sound like a pretty good fallback position—“BATNA,” as O’Leary and his colleagues called it. In fact, they already seemed to regard the TAL as pretty close to a treaty protected by a consociational amendment rule and a quasi-confederal ratification rule. That puts them in an admirable position, as I have explained, and I am well supported: “the Kurds could gain more autonomy under any new arrangements but would never have to accept less.”¹⁷⁴

I used to think that the confederal ratification rule and the default rule were as inimical to learning as the amendment rule.¹⁷⁵ But as I now see it, they do allow learning, but in a predetermined (that is, always more confederal) direction. If anyone wants anything new put in the permanent constitution, say a new status law, a new structure for the courts, or a new law for region formation, and wants to make sure it will actually be enacted, they can have it, but each time only in return for a new concession for the Kurds, who did not in the TAL get everything they desired from the point of view of regional autonomy—though they got the next best thing: a constitutionally built-in procedure for eventually getting it all.¹⁷⁶

Rules of constitutional change are not just rules. For some, they indicate the ultimate locus of sovereignty, a perhaps too metaphysical way of looking at the matter, especially where the relevant provision, for example, Article V of the U.S. Constitution, would indicate a rather frozen document.¹⁷⁷ Undoubtedly, however, a fundamental power or dimension of sovereignty is here involved: the ability of a political community to fashion its own fundamental rules. Whatever a constitution says, the amendment rule and the ratification rule control a process through which everything could in principle be otherwise. But that is not all. Aside from powers of sovereignty, what is at stake is also the entity or subject to which sovereignty is imputed,¹⁷⁸ the identity of what Americans call “we the people,” which, according to Sujit Choudhry, is the (symbolic) identity of the political community.¹⁷⁹ As he explains, the amendment rules and constitution-making rules, by determining which individuals can participate in the most fundamental form of boundary determining and potentially radical decision making, determine which individuals and communities can participate in political decision making. If a fundamental rule of change gives individual members of the legislature and individual voters the exclusive role, then the political community is defined as unitary in line with what I called the postnationalist or civic-republican nationalist conception. If a fundamental rule, on the contrary, gives veto power to individual communities already organized in a highly autonomous or independent manner, then it defines the country as the voluntary union of those communities, as the ethnic nationalist wishes. Finally, if a rule would compromise between these options and produce a synthesis where qualified majorities based on both individual and unit voting could make or change the constitution, but without vetoes, then it would come close to the multiple-identities model proposed by the liberal nationalist (and perhaps also be acceptable to the postnationalist). For Iraq, one such rule would have been a constitution-making procedure requiring the assent of two-thirds of parliament and three-fourths of the governorates.¹⁸⁰ But rule 61c together with 61e and 61g of the TAL were not this rule but rather ones in line with the aspirations of the ethnic nationalists. It finalized accordingly what the latter group sought: a definition of Iraq as a voluntary union of two ethnically defined peoples.

In several articles, Choudhry persuasively argues both that the amendment rule/ratification rule are not neutral rules and that there is no neutral process to generate them. Not only do they presuppose a political community and a political identity, but the rule according to which they would be made also would, and so on with infinite regress. I extended this argument with respect to procedure, when speaking about the shape of the table. A two-sided table (Kurds versus Iraq) presupposed the model of voluntary union; a round table with the Kurds as one participant at best allowed only devolution. Ultimately, such choices can be made only according to power relations as filtered through ideologies. But a legitimate way of making the choice would assume prior informal bargaining among the relevant actors themselves. The problem in Iraq was that the choice was made by only one of the parties together with the agents of the external, occupying power. And then on top of it all, that external power turned out to bargain incredibly weakly, even in the two-sided structure of negotiations. When the relevant ratification rule was introduced at the last minute, the U.S. managers accepted it in a wholly unthinking way and went so far as to manipulate the whole meeting so as to make sure the most important item in the TAL would be passed without any discussion at all. Thus the American occupier, without reflection perhaps but decisively all the same, imposed not only an interim constitution but a well-enshrined constitutional identity on Iraq.

I believe I may have been one of the first to argue that TAL would have deep legitimacy problems due to the exclusionary and imposed nature of its origins, and now several well-known interpreters have come to affirm that perspective.¹⁸¹ I did not, however, claim that it would lose its validity in the narrow legal sense, a position that since then has also found its advocates. My argument was originally based on interrelated domestic legal (instances with lower legitimacy trying to bind freely elected assemblies with higher, democratic legitimacy), international legal (Hague Prohibitions imperfectly overcome by UN Security Council authorization), political-sociological (the exclusionary structure of bodies that participated in process, especially with respect to organized sectors of the Arab Sunni part of the population), and procedural (the role of open American imposition in the process) dimensions.¹⁸²

To this list has now been added the significant issue of the absence any public openness or consultation before or during the making of the TAL, a consideration I must admit I omitted because I thought it too utopian under the circumstances. But Diamond’s analysis shows that to some extent it was possible to organize a subsequent sales or public-relations effort for the TAL, one that could only lead to anger and resentment of course, and this means that, yes, even in Iraq, the process could have been much more open and public.¹⁸³ But this, of course, would have been incompatible with what really was going on at the actual venues of negotiation, which was not merely an elite bargain needed to be initially shielded from the public but involved the active exclusion of some elites, the humiliation of others, the letting of still others play with the symbolic issues that happened to concern them the most, and the exaggerated role of the occupying power and one domestic actor among all others, the latter because it was the only one with a military capability, due to acts of the occupying power. None of this was fit for public discussion or even consumption. But the absence of the latter certainly contributed to the legitimacy problem, which is amazingly enough frankly admitted by the advocates of the greatest beneficiaries, the Kurds.¹⁸⁴

Legitimation problems do not automatically turn into a loss of legitimacy. For this to happen, there must be an important social actor that will plausibly challenge the old claims of legitimation and offer a serious alternative. This actor was once again the Grand Ayatollah Sistani. I cannot tell whether he was equally concerned with the issue of Islam in the state as were his supposed representatives in the IGC. I have my doubts, because on this issue he seemed to always support various compromise formulas—not that they really mattered. On the contrary, when he found out about the final text of the TAL, the day after it was rammed through, his opposition to it was nearly complete, even if he concentrated on two provisions only, article 61c, with the Kurdish three-province veto, and the consociational three-person structure of the Presidency Council, where he may have thought that one member had a veto (accurately, for government formation, but erroneously, for legislation).¹⁸⁵ The fight is entirely misconstrued, I think, if we see it as a challenge of the Kurdish demand “for minority rights” on the basis of the Shi’ite “quest for majority rule.” As advocates of the Kurds recognize, at issue were two structures of the state, one with and one without a Kurdish “federacy” with veto powers over all constitutional matters. Between such an asymmetrical federation and a purely majoritarian democracy there were a lot of alternatives, including federations involving all manner of minority rights, and there is no justification in claiming or implying that Sistani was against them all. All we know for sure is that he had to consider article 61c an abrogation of the concession of the freely elected constitutional “convention” made to him on November 15 and accepted by him in the compromise with Brahimi in the form of a European (and Iranian, Indian, and Iraqi) type of constituent assembly. Would he have rejected a solution that limited this assembly by any procedural rules at all? We will never know, because the three-province veto was a very extreme limitation, by one national minority, which gave its representatives a very high level of control over the assembly.¹⁸⁶ It would have been totally inconsistent with his previous modus operandi had he accepted it.

The Kurdish response to the challenge was predictably tough and unyielding; Sistani could not threaten them directly. The CPA’s response was irrational, uncompromising, and probably driven by timetables. It was the Shi’ites who chose a very sophisticated double strategy, to sign and not sign the TAL at the same time, which was to have very negative consequences for the constitutional development.¹⁸⁷ They probably opted for this because they both wanted to get to their elections as scheduled in the TAL and to retain the rights of the constituent assembly when elected, as against the restriction of the TAL. In brief, they wanted to “have their TAL and eat it too.”

The Shi’ite members of the IGC were to be in charge of the first, TAL-friendly dimension of the process. After demonstrably not signing the TAL on March 5, hugely embarrassing the Americans, they signed on the next day—and then announced that they considered it undemocratic, they had signed only to preserve the unity of the country (!), and that they would work to have it amended.¹⁸⁸ Sistani was evidently going to be responsible for the second dimension of the process, which involved working to delegitimate the TAL. On the same day (according to Juan Cole’s translation, on March 9), he released a fatwa: “any law prepared for the transitional period will not gain legitimacy except after it is endorsed by an elected national assembly. Additionally, this law places obstacles in the path of reaching a permanent constitution for the country that maintains its unity and the rights of its sons of all ethnicities and sects.”

Note that Sistani did not clearly say the law was null and void and that it would be wrong to obey it, nor that it could not be confirmed and repassed by the freely elected assembly if that assembly so wished. What he was disputing was legitimacy (which could only be democratic in his eyes) and not legality, whether or not he clearly articulated this distinction. After all, political actors supposedly in close consultation with him signed the law even if they then appended reservations, including the intention to amend it. But it is certainly true that he strongly believed that the freely elected assembly could not be bound by the TAL unless it chose to reaffirm it, and this could be (would have to be, in fact) done initially according to its own rules rather than the TAL’s.¹⁸⁹

As to legitimacy, he was concerned especially with an international-law legitimation of the TAL, which could conceivably bind the elected assembly, and thus he took immediate steps to block it. With the elimination of the caucus scheme, the TAL no longer had a method for choosing a government. “There was no plan B.”¹⁹⁰ The job, the American authorities reluctantly agreed, would have to be done in consultation by Ambassador Brahimi, based on his earlier success with Sistani. Sistani was not opposed to this, but he hoped to use it as leverage against the TAL, all the more so since he too had had some success with Brahimi before. In a March 19 letter to Brahimi,¹⁹¹ two weeks after the signing of the TAL and shortly before Brahimi was to come to Iraq once again, Sistani specifically mentioned the TAL’s nullifying the usefulness of the free elections conceded to him, and he warned Brahimi of total noncooperation with his visit if the United Nations in any way legitimated the TAL.¹⁹² He was not complaining about the concessions his side was forced to make about Islam¹⁹³ but rather about the state and governmental structure conceded to the Kurds and enshrined by a series of limitations on the constitutional assembly: namely, the rules of change. There was in fact nothing else in the TAL that limited that assembly, since the earlier tentative constitutional principles in the Pachachi draft were gone. But Sistani was evidently smart enough to understand that these particular rules of change in fact enshrined the default position of the Kurds, the imposed TAL, which involved their quasi-confederal enclave and the consociational structures of government that could not be changed without Kurdish consent, whatever the freely elected assembly and the immense majority of the people of Iraq wished. His concern was thus state and governmental structures, not minority rights more generally, unless we think that it is a minority’s inalienable right to have a “federacy” and “consociational” democracy. His fears of these, beyond the loss of power for the assembly, had to do with the breakup of Iraq and general ungovernability, and these were hardly unreasonable fears in the given region in relation to an ethnically defined confederal enclave (think of Bangladesh) or consociationalism (think of Lebanon).

With Brahimi, Sistani found a partner ready to share his fears concerning the making of fundamental decisions before the country had a legitimate elected government.¹⁹⁴ We know that UN officials did not like the way the TAL was being produced, and though they did not share Sistani’s majoritarian aspirations, they could have been happy to promise to deemphasize the interim constitution. We do not know what Brahimi promised to Sistani, and what relevant promises he extracted from the U.S. government, but we do know that they all fully cooperated in the choosing of the interim executive, where Sistani did not get his first choice for prime minister but at least got a candidate acceptable to him. It was actually Brahimi who wound up hugely disappointed here, because he was not able to construct a government of technocrats and was forced to basically transmute the old IGC into a new Interim Executive. He blamed Bremer, the “dictator of Iraq,”¹⁹⁵ for not supporting him, and it seems pretty clear that Bremer did get the candidate, Ayad Allawi, that he and his bosses wanted, as opposed to a “black turban,” which would have looked bad for the U.S. elections and who supposedly would never say “thank you” to President Bush!

Few people noticed that while Sistani made no great stand for the black turbans (supposedly because he could not choose between SCIRI and Da’wa, which is ridiculous), in the process he had to follow the TAL’s scheme for governmental offices and drop his objections to the three-person presidency, at least for the time being. The whole process strengthened the TAL as the legal framework on which they operated. Nevertheless, Sistani continued to battle for the freedom of his constitutional assembly, and thus against the legitimation of a document that would bind it. Thus when the government was formed in a way that was not his first preference but was at least acceptable to him, he renewed his campaign vis-à-vis the United Nations. First, he got Brahimi, a true gentleman, to deliver on what could have been an earlier promise, and he pronounced the TAL to be an interim document that could of course not bind the freely elected constitutional assembly, which was exactly what the CPA, the Kurds, and the document itself were trying to do.¹⁹⁶ Then, in case the message concerning the TAL was getting lost because of his own earlier cooperation, on May 7 Sistani addressed the following letter to the UN Secretary-General:

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” [i.e., the TAL] 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. For this reason, any attempt to bestow legitimacy on it through mentioning it in the UN resolution would be considered an action contrary to the will of the Iraqi people and a harbinger of grave consequences.

What the grave consequences might be was already relatively clear, as the intifada of Moqtada al Sadr was gathering strength. It was crucial that Sistani did not support any such option—or even large peaceful demonstrations—against the interim arrangements, now very close to the U.S. elections. Unlike Brahimi, it seems, Sistani fully understood the vulnerabilities of the Americans. So did the Kurds, who felt their superpower patrons wavering. Thus they too decided to play the same game: In a June 4 letter to President Bush, Masuod Barzani and Jalal Talabani specifically asked that

The Transitional Administrative Law (TAL) be incorporated into the new UN Security Council Resolution or otherwise recognized as law binding on the transitional government, both before and after elections. If the TAL is abrogated, the Kurdistan Regional Government will have no choice but to refrain from participating in the central government and its institutions, not to take part in the national elections, and to bar representatives of the central Government from Kurdistan.¹⁹⁷

Four days later, UN SC Res. 1546 approved the formation of a “sovereign” Interim Government of Iraq, its assumption of full authority (“transfer of sovereignty”) by June 30 with the end of the occupation and the CPA, a timetable (proposed by the TAL) for elections, and the formation of a Transitional National Assembly and Government, which would draft a permanent [sic] constitution leading to a constitutionally elected government by December 31, 2005. But 1546 did not refer to the TAL. Thus Sistani rather than the Kurds succeeded in this final round: the TAL did not receive international legitimation by the UN Security Council, and it can be said that in the end the U.S. government did not fully support its product. Nevertheless, 1546 did not detail an alternative method and scenario for government formation, moving toward elections, and writing a constitution, concepts it affirmed, leaving open only one of two possible alternatives. The more likely one was that these matters would still be regulated by the TAL, whose scenario was being followed, and certainly this is what the American government assumed. But a very narrow way of reading 1546 was also compatible with the idea of a provisional government, now established, that could fill each item with content as it went along as long as it adhered to the timetable, which was now enacted in international law (assuming the very contestable priority of international over domestic law in a formally nonoccupied country). To some extent then, arguably, 1546 left it up to the Interim Executive itself to decide whether Iraq’s transition would be regulated by a classic quasi-revolutionary provisional government or a genuine interim constitution. Allawi’s government then decided this question in favor of the latter, but without encroaching on the freely elected assembly’s prerogatives. Sistani won his last battle, but the straightjacket for the constitutional assembly was not thereby removed. Nor were most of the other results very positive.

Some supporters of the Kurds argue (on rather self-interested grounds, as we will see) that the TAL was indeed abrogated as Barzani and Talabani warned.¹⁹⁸ The most sustained version of this argument was made by Peter Galbraith, even if he later toned it down. Accordingly, under UN SC Res. 1483 Iraq was under “belligerent occupation”: “Occupying powers are not allowed to make permanent, or irreversible, changes in an occupied country.[¹⁹⁹] Occupying powers cannot cede territory [my emphasis], sell assets, or make permanent law. Accordingly, all law made by the Coalition Provisional Authority (CPA) expired when the occupation ended on June 28.”²⁰⁰ The United States could have secured Security Council authorization for lawmaking before making the TAL, but it did not. It could have secured Security Council authorization for the TAL (and other CPA-passed laws) afterward, before the end of the occupation, but it failed to do so. Not doing so signifies rookie mistakes on the part of Bremer and his inexperienced team, in the first instance, and an abandonment of the TAL by the U.S. administration in the second. Ergo, the TAL is legally null and void.²⁰¹ Relying on Galbraith’s claims about what went wrong, the same argument about the rights of occupying powers are less precisely repeated by O’Leary.²⁰² In his view, the U.S. government had two choices to save the TAL (and CPA legislation): either get UN recognition or set up an interim government that could recognize these laws. It did neither, and therefore Iraq was left with no “formal interim constitution, and in consequence no recognized ground rules for the negotiation and ratification of the permanent constitution.”

Written in September 2004, the last lines were surprising ones to write from the Kurdish point of view, but there was a method behind the apparent madness: if they are not bound, we are not bound; what is sauce for the goose is sauce for the gander. Kurdistan will apply those provisions of the TAL that it approves and ignore those it does not.²⁰³ And then in retrospect we are told that “the Kurds never implemented the provisions of the TAL they did not like. They never gave up control of their international borders with Iran and Turkey and continued to develop their oil resources without reference to Baghdad. Of course, they kept the peshmerga.”²⁰⁴

This self-righteous attitude justifying illegality and (from their own point of view) treaty violation is built on a house of cards. I was one of the first to discuss the deep legitimacy problems of the TAL, its delegitimation by Sistani, and the possibility therefore of a repudiation by the constitutional assembly, which, by the way, would have been a legitimate but revolutionary repudiation that no UN Security Council resolution could have blocked. This, however, is by no means the same as to claim the absence of legal validity. Certainly, Sistani’s protest, no matter how legitimate, was an entirely political one, which could acquire legal force only if a public law organ, say the Interim Government or the Transitional National Assembly, repudiated the TAL. Though clever, the international-law argument made by Galbraith is not foolproof. To say a body cannot legislate does not mean that its laws, if they have already been followed, are null and void.²⁰⁵ After June 28, for example, the de-Baathified civil servants or the dismissed military officers were still without their positions (because of CPA orders 1 and 2). They could not just return to work, and it would have taken another legal decision of some kind, a court order or a new piece of legislation, to give their jobs back to them, even under Iraqi legal conditions, such as they were. Of course these orders, later modified, remained in effect. Conversely, though they got their jobs under the TAL, I. Allawi, his ministers, and the Presidency Council could not suddenly lose them because of the supposed abrogation of the law on which they stood. And so on with all the many laws of the CPA, including the Electoral Law (CPA Order 96, June 7, 2004) and the Electoral Commissions Law (CPA Order 92, May 31, 2004), which were used to regulate the elections of January 2005. Thus the TAL rightly pronounces a necessary legal fact, even if a self-referring one, that all these orders remain in effect until repealed by or derogated from by some other law (TAL art. 26c).

Moreover, the terms “permanent, or irreversible, changes in an occupied country” do not apply in a completely obvious way to an interim constitution or transitional administrative law. Whatever we may think of the lawyers of the CPA and the U.S. government, they understood and worked with the Hague requirements in Iraq no less than their predecessors in Japan and found in each case a different way to disguise or legalize their imposition and their desire to make permanent change under the guise of transitional ones.²⁰⁶ If we are to focus on where the TAL actually violated international law, it would have to be on the question of state structure and its insulation through the amendment and ratification rules. In the face of repeated warnings and criticisms from UN sources, the TAL did make changes here that would be difficult to reverse and made sure through the rules of amendment and ratification that they could not be reversed even when the occupation ended. Thus it is understandable that the Kurds wanted international legal recognition for American illegal actions in which they were deeply implicated. But to the extent the rules now existed, it was nevertheless strange that it was its beneficiaries who now claimed they were invalid and that they were free to follow them or not as they pleased.

Admittedly, the Kurds could claim that they were responding to Sistani’s challenge to have his constitutional assembly abrogate the TAL, even if preemptively. Anticipating repudiation, they repudiated. The situation is the same with the now famous preventive or supposedly “preemptive” war. In fact, they were simply committing illegalities. But that is not all. They also assumed that the TAL was in effect and thereby recognized it. They did not follow through the threats contained in the Barzani and Talabani letters. They assumed their positions in the Iraqi interim government, accepted the financial grants due to them, fully participated in elections according to CPA electoral rules(!), counted on the provisions regarding Kirkuk being carried out, and never forgot about their amendment and ratification rules, through which they could get many more concessions in the future, all under the TAL. They had their TAL and were feasting on it too. This could probably be said of the Shi’ites too, who kept and developed their militias and undoubtedly introduced religious-status law wherever they controlled local government to a sufficient extent. But illegalities, even when committed by provincial or regional governments, should not be confused with abrogation. Only a national body on the same level as the authors of the TAL could be said to abrogate it. And this never happened, as weak as the interim constitution’s legitimacy may have been.

First and foremost, after taking office, the TAL was a matter for the Allawi government to uphold. It is hard to understand both O’Leary’s claim that the Americans should have set up an interim government capable of recognizing the TAL (which would have to be a government also capable of not recognizing or abrogating it) and his (inconsistent) complaint that Allawi, despite the Kurds’ relevant demand, refused to legislatively enact it. As both O’Leary and Galbraith admit, Allawi did agree to abide by it until the first meeting of the freely elected Transitional (constitutional) National Assembly, but this limitation only meant that he did not feel he had the right to preempt the relevant decision of that body nor the power to challenge Sistani (supported by Brahimi) on this score.²⁰⁷ As both O’Leary and Galbraith should have known, Allawi could not reenact it as a genuine piece of legislation, simply because he had no legislature.²⁰⁸ But because he could declare (executive) orders with the force of law (annex, sec. 2), he actually did confirm the TAL in one of these, his Order Safeguarding National Security (July 6, 2004), which was about the most he could do.²⁰⁹ Thus in this respect, he gave little justification for any other group’s wholesale violation of agreements incorporated in the TAL or even for disregarding substantive provisions, all while adhering to the schedules contained in the TAL.²¹⁰

There is only one justification for what the Kurds did, but it is a serious one. The Shi’ites were the first to reduce their relation to the TAL to a double, strategic one, inviting others in effect to follow them. Kurdish spokesmen (for example, Barem Saleh) have rightly pointed out that the TAL was unanimously approved by the Shi’ite leadership in the IGC. Thus it is disingenuous to say, as did Mowaffak al-Rubaie, one of the signers, “that you cannot control the will of the people [in the constituent assembly] … whatever they will do, they will do,” when the same parties hoped to control the popular majority in the very same assembly.²¹¹ The same criticism would be invalid with respect to the initial agreement the night TAL article 61c, the ratification rule, was introduced and hammered through by Bremer without discussion. But the two days of formal signing were another matter. The Shi’ites first refused and then proceeded to sign after extended deliberation, understanding the consequences. That they announced and even “attached” reservations is immaterial. They signed. And they signed precisely with a dual strategy in mind. They wanted the TAL to deliver for them what it could (elections) while remaining free to go beyond it when they got what they wanted. Now the Kurds were about to adopt the same double attitude, with the difference that in their case partial repudiation could only be in the form of “illegality,” whereas the Shi’ites, when in control of the constituent assembly, could repudiate the TAL by establishing a new legality. But that difference could be seen in terms of the type of power each side had, not as an ultimate normative difference.

Nevertheless, I believe the Shi’ites had a better reason to adopt a double attitude to the TAL, even if the Kurds could be hardly expected to understand that. That reason had to do with negotiating and bargaining within an ultimately imposed structure, under the gun, where the Americans were in the position to use the continuation of the formal occupation, the restoration of formal sovereignty, and having democratic elections as bargaining chips. This put the Shi’ite leadership in a double bind almost every step of the way, and they were continually reminded by Bremer of the examples of the Shi’ite electoral boycotts in British Iraq, if they needed to be reminded of what happens when one chooses the alternative of simply withdrawing from a process seen as badly deformed by an occupation. Back in the 1920s, the Shi’ite mujtahids were to lose political influence for many decades as a result of choosing the clear and honorable path of resistance. As they now must have seen, they could neither withdraw nor not withdraw, even when they were very severely provoked, for instance with the adoption of 61c. Signing and keeping options open was a rational response to a very deformed political process, even if in retrospect the results proved devastating.

The Kurds understandably interpreted the TAL signing as an agreement among the signers. The Shi’ite double attitude was one addressed to the Kurds, as far as the Kurds were concerned, threatening what the Kurds had achieved—and the Kurds conveniently forgot how they achieved it. Thus to counter it, the Kurds assumed a double attitude as well. Despite their threats, they entered the Interim Government and played along with the TAL on the “federal” level. But on the level of their (con)federacy, the Kurds illegally continued to solidify all their positions as a quasi-state. All this is undeniable, since their defenders say so themselves, though denying any illegality, since the TAL was supposedly null and void. But this position is untenable, as I have argued, and in the light of subsequent developments those who have argued that the TAL was null and void have backed away from that assertion. The illegalities and their results are there to stay.

What the Kurds and their defenders wound up admitting is that the state bargain, the heart of the TAL, the only real bargain in it, a kind of treaty between a superpower and the quasi-state of Kurdistan, broke down before the ink was dry. The United States, the superpower, after having destroyed the Iraqi state, did not manage to use the constitutional process, in its first stage at least, to reconstruct that state in its geographical-territorial capacity. It negotiated an asymmetrical political structure of a federation for Iraq and a confederal enclave or “federacy” for Kurdistan, and overall “stateness” was to inhere in a very few powers the federation was to retain in and over that confederacy. Those powers were now recovered by Kurdistan, with the very partial exception of foreign affairs (controlled in fact by the United States), leaving a fractured territorial entity in place.

If that left the rest of Iraq a federal state, as the Kurds sometimes argue, perhaps state formation could still be pronounced a success. But the concessions to the Kurds in the TAL and the illegalities they committed despite the TAL were among the reasons why this was also not possible, and to his credit the Ayatollah Sistani was aware of this set of problems too. Let us recall that his objections were not only to the confederal but also to the consociational features of the TAL. In fact, consociationism, by no means a logical necessity if there is an ethnically based confederation, produced very a weak government for Arab Iraq, allowing all sorts of Kurdish vetoes over it (constituting or replacing government, amending the constitution). Even more significantly, as Sistani explicitly said (by then it was no longer a prediction), it sectarianized or ethnicized politics, with each group now looking to imitate the Kurds’ success in gaining regional and veto rights. Here the second dimension of stateness, that of having some kind of unified status for the citizenry (the “people” of the state, not to be confused with the ethnic or even the republican idea of “nation”), was profoundly endangered, despite the TAL’s provisions for a unified Iraqi citizenship. Finally, and deeply related, the concession to the Peshmerga, while certainly not the only cause, made the survival of other militias unchallengeable. And with militias in control, the third (and perhaps most immediate) dimension of statehood, already severely compromised by CPA orders 1 and 2, the organizational dimension that produces a monopoly over violence, was shipwrecked. A so-called Iraqi army and police could be built, but these would be based on militia units serving en masse, in organized fashion, keeping their primary allegiances, with the consequence that the population could encounter “foreign” and very much hated militias in “Iraqi” uniforms.

All the other supposed achievements of the TAL become irrelevant in terms of this failure of state making. If there is no state, government as the political organization that is supposed to control such an entity cannot function. At most, as early pluralists have thought, it becomes one competing organization among many, local and regional, civil and military, secular and religious. If there is no state, there are no state and confessional relations, and religious authorities will occupy the space of jurisdiction that government is unable to operate in. If there is no state, there are no rights against the state nor are there rights the state can positively guarantee. In failing at its most fundamental task, state rebuilding, the TAL failed at them all—with one exception. While it could not provide a serious constitution for the interim period, it did provide a framework for negotiating a supposedly permanent constitution. Its rules of change, which were a very important source of its failures, also turned out to be the most successful in achieving their purpose—if at the cost of more devastation. Despite claims of its abrogation, the TAL remained a highly constraining blueprint for Iraq’s subsequent constitutional process.