2.5.1 Procedures and rules for debating and deciding

Rules will be needed especially for a constituent assembly; even the regular legislature may need some special rules for making a constitution, different from its regular rules of procedure. Smaller bodies, such as constitutional commissions, may not have such complex sets of rules, but will still need rules on matters such as how it makes decisions, finances, or how many members must be present to make up a quorum for a meeting.

This section focuses particularly on constituent assemblies, and to some extent on commissions, though the issues are relevant to all bodies that engage in constitution-making tasks.

Major issues


A quorum is the number of people who must be present before discussion or a vote can take place. The idea is to prevent minorities from making decisions. Many legislatures have small quorums. Less than 50 percent would seem undesirable for a constituent assembly, which is supposed to be a true mirror of the nation. Uganda had the 50 percent quorum rule, but this proved difficult to achieve, especially as some members were ministers. In Nepal the interim constitution says that for most purposes there must be at least 25 percent of the constituent assembly members present. This is quite low; discussion could begin when only women were present, or only men, or only members of one or two castes or ethnic groups.

At least for a vote, a higher quorum should be required. If there is to be a vote that requires a special majority (two-thirds of the membership is common), then at least that number must be present. But it does mean that debate can have taken place with far fewer members present.

The quorum rule can be a source of delaying tactics. The body should clearly be well attended, but members should not be able to exploit the rule to adjourn debate in order to delay decisions. The chair ought to be able to adjourn the meeting, however, if the numbers present clearly make debate pointless.

The rules for committees could be different. They could also make it possible for the chair or convenor of a committee to decline to allow discussion to begin if, for good reason, no representatives of a particular group were present. But no group should be allowed to hold the committee for ransom by failing to turn up.

The quorum rule alone cannot be relied upon to ensure attendance. Other factors include where the body meets (are there too many distractions?), whether for other reasons the process isallowed to drag on for too long, and the effectiveness of the chair. There is also the vexed question of attendance allowances. They are not unusual, but must be properly policed or members will simply sign and go away. Publicity attached to attendance may help in some societies.

Speaking rules

Rules must recognize the freedom of speech—of all people. In some parliaments a person must “catch the speaker’s eye”—which may work against the less experienced members, and perhaps especially against women. In some assemblies the procedure has been for members to tell the chair in advance that they wish to speak on an article, and the chair will call on them in the order in which they appear on his or her list (e.g., as in Timor-Leste [2002]). Such an arrangement has the disadvantage that successive speakers do not respond to each other but simply deliver their prepared statements on the topic. Another possibility—drawn from the Kenyan National Constitutional Conference—is to provide each member with a large card on which the member’s enrolment number is printed. This is raised to indicate a wish to speak.

It is absolutely essential that there be a way to limit the length of speeches. The length of permitted speeches will probably be influenced by the size of the deliberating body, and by the stage of the process. In Uganda there was a thirty-minute limit during the initial debate—but the body comprised only 214 members. In Kenya there was a ten-minute limit, but there were 629 members. In Bolivia, during the “Vision of the Country” debate at an early stage in the process, each political group had a maximum of three hours. Committees were given two hours to report to the plenary. In debates on major motions, regional groupings had one hour each. At the decision-making stage when votes were to be taken, each member had no more than ten minutes per section to vote. And to make a motion, only two minutes were allowed.

There is a risk that extended time periods might be given only to political parties, rather than to other interest groups. It would be possible to allow other groups—which might include women, people with disabilities, or minority groups that may cross party lines, according to the particular country’s circumstances—to be given extended time.

The rules of the Indian constituent assembly provided that Hindustani (Urdu or Hindi) or English could be used, but, if the president of the constituent assembly took the view that a member could not express himself or herself in one of those languages, that member could address the constituent assembly in his or her native tongue and a summary could be provided for members in English or Hindustani. In Timor-Leste the official languages of the constituent assembly were Portuguese and the local lingua franca, Tetum, but members could express themselves in English or Indonesian (despite hostility from some to the use of Indonesian, viewed as the language of the oppressors). In Bolivia the rules of the constituent assembly provided that plenary and committee sessions would rely on interpreters and translators so that members could express themselves in their native tongues (article 52). Provisions about sign language could also be explicit.

Notice rules

Rules requiring suitable notice of any issue to be discussed help in ensuring that members do know when they should be present in order to make contributions on issues about which they feel strongly. The Bolivian rules usually required twenty-four-hour notice for documents.

Politeness and decorum rules

Also important are rules that ensure that members are addressed and listened to with respect— that other people do not use insulting language, or speak on their mobile phones in the hall, or read the newspaper or gossip with their neighbors. Rules may become rather complex and intimidating on these points, and care should be taken not to be overly formal. There may be other solutions to some problems; the parliament in Nepal has mobile phone–jamming devices.

Decision-making and voting

The majority by which decisions are to be made can be one of the most difficult aspects of the process (and has in some processes been the reason that rule making takes a long time). Different sorts of decisions may require different majorities. Majorities may be defined in various ways, including:

  • A majority of those present and voting (sometimes called a “simple majority”); if fifty people are present, and eighteen vote for a motion and twelve vote against it, the motion is carried, even though this was not a majority of all those present.
  • A majority of those present (which means that those who are present but abstain from voting are counted as having voted against); where the issue to be voted upon is whether to amend a draft constitution, this rule makes it difficult for the amendment to be accepted, which may be a desirable provision.
  • A majority of all the members, whether present or not, and which means a number greater than 50 percent of all members (sometimes called an “absolute majority”); sometimes care is taken to say that “all the members” does not include any vacant seats.
  • At least two-thirds of all the members, whether present or not (often used for constitution- making decisions).
  • At least 75 percent of all the members, whether present or not (sometimes used to amend constitutional clauses regarded as being of particularly great significance).

Note that these rules assume the normal pattern in formal decision-making bodies: that all decisions are “binary”—a member votes either “for” or “against.” Even if there are several solutions to a question, the decision-making will be structured so that each one is voted on separately.

Box 29. What is a simple majority? A debate in the Timor-Leste constituent assembly [2002]

The question was: did this mean just the largest number of those voting—or a majority of those present? The latter would mean that a person who was present but abstained was in effect voting against the motion. Members of the assembly were divided. A foreign expert favored the latter interpretation (Kathlene 1994: 565).

It is desirable for all these rules to be spelled out clearly. A note on the “simple majority” question: on the one hand, the interpretation that it requires more than half of those present does mean that a motion requires substantial support. But the rule drafters could, and should if this is what they meant, have been specific and said “a majority of those present.” The normal meaning of “simple majority” in English is “a majority of those present and voting.” “Absolute majority” would most likely have meant “a majority of those present”—as a member of the Timor-Leste constituent assembly pointed out. To allow silence to be a vote against does make it easy to abstain. When the issue is whether to amend a draft, it favors the draft.

Clear rules should specify a majority of what: “a majority of those present and voting,” “a majority of those present,” or even “a majority of all members,” making the use of the term “a simple majority” unnecessary—and indeed, it is best avoided. Rules often require that those abstaining must also indicate that they are doing so, and that the numbers abstaining be recorded. Such a rule may act as a disincentive to members abstaining, something that is often a form of abdication of responsibility.

Administrative decisions may often appropriately be made by a simple majority (“a majority of those present and voting”)—but care must be taken if these decisions might disadvantage particular groups. In many constitution-making processes the required majority is set at two- thirds of the entire body. The purpose of such a rule is to give groups within the assembly the incentive to persuade others to support them. It prevents a small minority standing in the way of a new constitution—which may be a consequence of a rule requiring a higher majority.

To require total agreement in every human endeavor is unrealistic. But a number of processes have set “consensus” as the objective—as in the South African process [1996], and also in Kenya [2005] and in Nepal [ongoing process]. “Consensus” presumably means something other than unanimity. Many people claim that seeking consensus was the traditional way of doing things in their culture. As the chair of the Uganda Constitutional Commission and (subsequently) chief justice of Uganda, Benjamin Odoki, notes, an Africa consensus involves “settling disputes by listening to everyone and taking into account all views. It is a painstaking exercise, which is most rewarding in the end because it produces no losers since all are winners, and promotes legitimacy and acceptable decisions” (Odoki 2005). In modern politics, however, consensus is often hard to achieve—even in Africa. In South Africa, the concept of “sufficient consensus” was developed; this meant that if the major parties, representing all the ethnic groups, agreed, the process could move on (but this did exclude some groups). In Nepal, consensus was defined as no vote against a provision in the constitution (regardless of how many abstained). Then there was to be a process to try to achieve agreement (through party leaders), and finally, if this proved impossible, a two-thirds majority of all the members would suffice. Consensus may be easier to achieve in small bodies, such as a commission.

Box 30. Challenges to participation of women in decision-making bodies

Cultural practices and social expectations often make it hard for women to participate fully in public bodies. One writer (Aili Mari Tripp) says, “Even in parliamentary bodies, women have difficulty being taken seriously, being listened to, and are frequently subjected to humiliating stereotypes and derogatory remarks” (Tripp 2011: 153). In Timor-Leste [2002] a female constituent assembly member raised the issue of domestic violence and was laughed at by many male members. There is also evidence of sexual harassment of women parliamentarians in some countries, including in South Africa (Geisler 2000: 618) and Uganda (Tripp 2001: 153, citing Tamale 1999), and even in Sweden (The Local 2007).

Research shows that male interruption of women speaking is more likely than any other type of harassment. In legislative committees in the United States, “women entered the discussion later, spoke less, took fewer turns, and made fewer interruptions than men” (Kathlene 1994: 565). In developing-country legislatures, parties are often weak, with patronage rather than rules being the governing principle—which makes it hard to fight for fair party procedures for women. These difficulties, which may affect other groups that have been marginalized from public life, should be borne in mind by those who draft rules.

A mechanism to handle deadlocks is almost bound to be required at some point, whether there is formal provision for it in the rules or not. (See part 2.5.2.) At different stages of a process, voting may work in different ways. If a document is being adopted for the first time, it is probable that a particular vote in favor will be required. At some stage, however, a draft already prepared by a particular process may then be presented to another body. There are basically two ways for that draft to be considered: either it is to be presumed to be the final constitution unless it is changed, or it is to be viewed as a proposal only. In Uganda the draft prepared by the constitution commission could be changed in the constituent assembly, but only if a motion to change it was moved and it passed by a two-third majority. But in Kenya [2005], the adoption of any clause of the new constitution required a two-thirds majority of the constituent assembly. In Kenya [2010], proposals to a draft constitution could be made by the parliament only by a two-thirds majority of all the members. Although there were at least that number present in the parliament, every time a vote was called members melted away—so that not a single one of 150 proposed amendments was passed, and only one was voted on.

Some constituent assemblies have made provision for secret ballots. This was possible in the Constitutional Loya Jirga in Afghanistan [2004]. In that instance there was a genuine risk of intimidation by warlords. In Nepal some members expressed concern that they might be subject to pressure, and some civil society activists were worried that members from various ethnic or caste groups and women would not have freedom to represent those groups, but would be essentially “vote banks” for their parties and their concerns. Secret voting would give members a way out of this dilemma.

Generally there are strong arguments against secret ballots, which are essentially antidemocratic: the electorate should know how the people for whom they voted carry out their responsibilities. There is also a risk in some countries of bribery; this certainly took place in the Kenyan National Constitutional Conference [2005]. A secret ballot may encourage bribery—because no one would know how individuals voted and would therefore not know whether they have been influenced by factors such as bribery. Some members might be tempted not to think hard about how they should vote, because they would not be answerable for it. Finally, most decision- making will probably take place in committees where secret voting is less feasible—either because voting will be a more open affair, or because smaller numbers will make it harder to conceal how members vote. It is unrealistic to expect political parties to agree to secret voting if they are determined to exercise control.

There are various ways of indicating a vote: a voice vote (where all those in favor simultaneously say “Yes” and then those against say “No,” and the chair judges by the volume who prevails), a show of hands, passing through division lobbies (“Yes” voters through one room and “No” voters through the other, or the equivalent of walking to the front of the hall to place a vote in an urn or box), electronic voting by pressing a button at each member’s seat, and a roll-call vote, in which the name of each person is called out and the voter must indicate his or her response out loud.

The voice vote is partially anonymous: the person sitting next to a member may know how his or her neighbour voted, but the person presiding and the public may not. This procedure was used when the Kenyan parliament adopted the proposed constitution—to go to referendum— in 2010. Though the “Aye” vote sounded overwhelming, it is not possible for anyone to say exactly who said “Aye”—and this may explain why so many members of parliament felt able to argue against the document in the referendum campaign. The roll call was used in Timor- Leste [2002]—not as a routine, but for a few issues of symbolic significance, for example in connection with the flag. Such a voting system is impractical for large constituent assemblies. This technique was used in the French constituent assembly; according to Jon Elster, “a procedure that enabled members or spectators to identify those who opposed radical measures, and to circulate lists with their names in Paris” (Elster 1993: 180).

In Timor-Leste each member had three cards and voted by holding them up: green, red, and blue (for, against, and abstain). At one point it was suggested that rather than having each group hold up its cards separately, all should hold them up together and they could be counted by three people simultaneously. This would not work with six hundred people (as in the constituent assemblies of Kenya [2005] or Nepal [ongoing process]).

One issue in constituent assemblies is: How far ought party members to be subject to a requirement to vote according to “party line”—or how extensive ought the party line to be? There are topics on which a party may genuinely have a firm position. But “whips” (a term used in some countries to refer to the party officials who exercise voting discipline) are not readily compatible with consensus—unless they are used to encourage party support for a consensus reached among leaders. Even then, if members have to be “whipped” into support, it hardly seems to indicate a national consensus.

Absence of discipline is likely to produce a more well-informed debate. In some countries it is common for matters of conscience to be a subject for a free vote (but this may produce unreasoned populist voting on certain issues). In Nepal, despite much advance discussion, the rules were silent on this—presumably leaving it to party discretion whether to apply party discipline in any particular context.


Should there be a formal possibility of cross-party alliances or caucuses? Women’s caucuses are common in legislatures around the world. Sometimes they are composed of women members from a single party, and sometimes they are cross-party. In the constitution-making process in Uganda, as Aili Mari Tripp noted, the assembly’s female delegates formed a nonpartisan women’s caucus. It offered workshops on parliamentary procedures, speech-making, and how to build a coalition; it also made sure that the women’s views and concerns were represented in a weekly radio program on the assembly’s debates (Tripp 2001: 150).

In Nepal an attempt to recognize a women’s caucus in the rules was unsuccessful, defeated by the parties’ reluctance to lose control.


The whole process of a constituent assembly should be characterized by transparency. The public should know what is about to be debated, and by whom. Papers on which discussions are to be based should be publicly available in time to permit interested parties to make submissions and get in contact with observers.

But should the sessions themselves be held in public? The South African process was open: all meetings of the constituent assembly and its bodies were accessible to the public. There are risks in openness. In the French constituent assembly the mob terrorized the delegates. Even in more orderly circumstances the presence of a certain group might overawe the members. And

there is the risk that members will be tempted to play to the gallery rather than concentrate on the business at hand. The South African parliament, when debating the constitution, had some experience of this when committees were televised. The Philadelphia convention making the United States constitution decided to sit in secret partly because of a fear that its members would be reluctant to change their minds if those minds had been revealed in public. The Spanish constituent assembly [1978] had a secrecy rule for the committees. Jon Elster, who has studied several processes, suggests that sitting in public inhibits genuine discussion by privileging stubbornness and grandstanding. On the other hand, he suggests that secrecy encourages more interest-based bargaining than deliberation. His conclusion is that committees should sit in private and the plenary in public (Elster 1995: 386).

Relations with the public and the media

The administration of the constitution-making body should be responsible for ensuring that there is a constant flow of accurate information to the public. (See part 2.3.11 on the media.) In Timor-Leste a short daily bulletin was put out in several languages. Other methods include a newsletter and a website (which would be used by a limited range of groups but would be a resource for the media and for NGOs). The rules may not go into any detail on these issues, but it would be wise to make it clear where responsibility primarily lies—though this should not preclude civil society from monitoring and publicizing the work of the body.

One way to set up a structured relationship between a constitution-making body and the public is to have accredited observers to the body. Such observers would represent organizations or interests that did not otherwise have an adequate voice in the constitution-making body. These could be professions, religious groups, ethnic or linguistic groups, or possibly certain major NGOs. They should not be too many in number—which means that they must represent substantial (in the sense of important) interest groups or issues. Narrowly focused groups’ concerns could be met by their having access to relevant thematic committees. Observers could have the right to attend all plenary and committee meetings of the assembly, but not to speak or vote. They could interact with members in an informal way, having the right to eat where the members ate. They could thus have the opportunity to persuade, but not in any way to coerce.

Rules for a commission or similar body

The rules for a constitutional commission or a similar body are likely to be simpler than those for a constituent assembly. However, drawing on some degree of common sense, and also on the rules of previous commissions, the following are, we suggest, the main points that ought to be addressed:

  • Quorum: it would be normal for a smaller body to have a higher quorum in percentage terms than a constituent assembly. This is to avoid the risk that important issues will be decided on by small numbers, and also because each member of a commission may represent different interests; more than in a large body, the absence of a single person is likely to mean that some voices cannot be heard.
  • Notification of meetings: especially when a body meets part-time, it is important that all members are given due notice of meetings.
  • Record keeping: it may more easily be overlooked in a small body that it is important to keep proper records of discussions, and ensure that they are retained in a secure place.
  • Speaking rules: it can be just as difficult to be heard in a small group as in a large one.

A small group of people may develop a good relationship, and even become friends. But this may make it harder and not easier to impose discipline on members. Rules are therefore important—even if the atmosphere means they do not have to be relied on frequently. And if the atmosphere sours—as it has in some commissions—rules that make all the members feel they have a fair chance to participate may become important.

2.5.2 Dealing with divisive issues

Contemporary constitutions are complex documents, often the products of participatory processes involving ethnic communities, religious groups, professional associations, women, youth, the disabled, and indigenous peoples, and covering what is often a wide range of policy issues. Consequently there are likely to be many divisive issues that may derail the constitution- making process. These issues may not be easily resolved by the formal procedures for decision- making. Some processes include special procedures to resolve divisive issues. Sometimes measures outside the scheme of the process may be used to resolve these issues. A particular difficulty in resolving differences is the transparent nature of many contemporary processes (a result of the high degree of public participation), which makes it hard to negotiate and compromise. The very dual nature of participatory and transparent processes makes handling divisive issues difficult.

What is a divisive issue?

Divisive issues should be distinguished from mere differences of opinion, which are inevitable when so many decisions need to be made when a constitution is drafted. These differences would be discussed and negotiated in the constitution-making body, failing which the matter would often be settled by a vote. Normally the losing party could live with the outcome.

But a divisive issue is a source of tension or even conflict among the people or the negotiating parties. One of the most common (and intractable) issues involves differences about the place of minorities and the role of diversity in state and society. In the constitution-making process in Spain after Franco, the critical issue was the conception of Spain: whether it was to be a centralized state of a unified people, or a state of autonomous regions based on ethnic diversity. With variations this has been an issue in the constitution-making process in many countries—for example, Bolivia, Fiji, Iraq, Nepal, Papua New Guinea, South Africa, and Sri Lanka. Countries have been divided over questions of political, economic, and social policy (as with differences between Maoists and other political parties in Nepal). Other divisive issues have been internal territorial boundaries, allocation of natural resources, secularism versus state religion, official languages, and systems of government. (Often that choice is between parliamentary and presidential systems.) Clearly what is considered divisive varies with the context.

What divisive issues have in common is that they relate to matters of considerable importance to the material advantages of groups, often couched in emotional terms, based on historical claims, and involving narratives of past discrimination or exploitation, or entitlements to human or group rights. Conflicting interpretations of rights or prior understandings can complicate a settlement. Divisive issues cannot be resolved in the same way as other differences, for example by a vote. There is a considerable emphasis placed on consensus in current notions of constitution-making processes, in part because many constitutions are now made in conflict or postconflict situations, where majority decisions would merely aggravate tensions instead of removing them. Divisive issues can neither be willed away nor handled in any easy way.

A divisive issue may be described as one whose nonresolution can cause fundamental problems, but that should not be resolved by majority vote, which would endanger the legitimacy of the constitution. A different dimension of divisiveness in that without its resolution, no constitution can come into being when it concerns a matter central to the constitution, such as the system of government. For example, if a two-thirds vote is necessary for a decision, and opinion is deeply divided, then a positive decision cannot be made. Sometimes an issue may be regarded as divisive in this sense, even if it is not central—if the vote of one group can lead to the rejection of the draft constitution (as can happen in a referendum). In Kenya [2010], fundamentalist Christian groups threatened to reject the draft if references to Kadhi courts were not deleted.

Some issues divide whole countries; some may be relevant to a particular interest. The former issues include disagreements about the system of government; the latter involve the application of personal laws or the availability of religious courts. They pose different threats, and different strategies may be needed to address them.

Divisive issues can arise at different stages of the process. Sometimes the very question of whether constitutional reform is necessary is divisive. In some tragic circumstances (as in Kenya and the Philippines) the matter has been “resolved” only by battles in the streets. Sometimes it may be hard to get agreement on the main objective of review. In Fiji it took six months to reach a compromise, one side demanding that the primacy of indigenous Fijians should be recognized, the other side championing Fiji’s multiethnic character.

A divisive issue can emerge during the process, for example, because of changed political circumstances. (In Kenya the change of government halfway through the process brought about a fundamental shift in the support of a key group from a parliamentary to a presidential system of government.) And spoilers may generate a contentious issue out of a proposal on which for a long time consensus may have existed.

Some divisive issues may need to be resolved quickly; they will fester if they are not. Others are best put aside for future resolution. And ways of addressing divisive issues may depend on the body charged with making decisions on the constitution. Thus different approaches are needed depending on whether the decision-making body is an expert commission, a parliament, or a broadly representative constituent assembly, and altogether different considerations apply when there is a referendum.

We now turn to some specific strategies and techniques that have been used to resolve divisive issues.

  • Some issues can be resolved by formal and symbolic recognition. Groups that have been marginalized or oppressed are often in great need of the recognition of past suffering and of their place in the new political order. Delicate negotiations may be required for the formulation to find the appropriate balance (as the example of Spain shows). Such recognition may facilitate the resolution of other issues geared more to substantive matters.
  • Deadlocks over an important issue can sometimes be resolved by a referendum. Greece and Italy resolved the contentious issue of the future of the monarchy by referendum (an option canvassed in Nepal in 2006, but effectively vetoed by the Maoists). The vexed question of the division of the Swiss canton of Bern to allow the French-speaking inhabitants of the canton a space of their own was resolved through a series of referendums soliciting the approval of interested parties (leading to the establishment of the canton of Jura). The Canadian Supreme Court has sanctioned referendum as a way to resolve the issue of Quebec’s secession. Of particular interest is the referendum in the Maldives on the choice between a presidential and a parliamentary system, the divisive potential of which emerged during the course of the proceedings of the constituent assembly. When negotiations failed, the assembly decided to refer the question to the people, and resumed its work only after the matter was settled in this way. Civil society in Kenya, where a similar issue was deeply divisive, advocated that the referendum should offer the choice between a parliament-based constitution and a president- based constitution. In both Uganda [1995] and Kenya [2005], legislation on the constitution- making process provided for referral to the people of issues that were contentious (i.e., ones that could not be resolved by a two-thirds vote). Unlike Kenya, Uganda used this provision for the question of whether to have a single or a multiparty system.
  • A referendum will yield a result, and perhaps some legitimacy for it, but it can also deepen polarity. And if the vote is close, divisions in society and politics will continue, and they will affect adversely the implementation of the constitution. In any case it is best to avoid a referendum until a serious attempt is first made to resolve the issue through negotiations.
  • It may be possible to resolve the issue by reference to public opinion. This is a likely option if the process has been participatory, people have been widely consulted, and a careful analysis of views and recommendations has been made. It has similarities to a referendum, but there are important differences: public views can be taken into account at different stages, and balances struck. Both the Ugandan and Kenyan constitutional commissions used public opinion to resolve differences. But with this method there is a considerable danger of manipulating the analysis of public opinion (which can be fragmented, and indeed may be the cause of divisions). There is also a considerable degree of subjectivity among decision-makers.
  • Some divisive issues can be negotiated by the expedient of a time limit (a “sunset clause”). The group wanting a particular provision, such as protection for some special right, can be satisfied by being given the right for a specified period. Those opposed to that protection may be reassured by the limited duration of that right. This approach was adopted at the independence of Rhodesia-Zimbabwe, where the protection of special rights for European settlers with regard to land and political representation was limited to ten years. The entire Fiji 1990 constitution was accepted by the Indo-Fijian and other minorities only on the basis that it would be reviewed within seven years of its inception (and indeed, that review led to large-scale changes). To some extent, sunset clauses have the character of transitional provisions, facilitating bargains and providing time for psychological and material adjustments.
  • Another device used to help address divisive issues is the opposite of the sunset clause— postponing the issue for resolution in the future. This may be done because the issue is extremely controversial and has the potential to break what otherwise is an overall consensus (as with the decision in Iraq to postpone the settlement of the issue of Kirkuk, a territory disputed among ethnic groups and with a troubled history of migrations and expulsions and the presence of valuable petroleum deposits). Occasionally an issue may be postponed because the members realize that information and public views are not available to make informed, practical decisions, and would not be available in time for the conclusion of the process. Usually the postponement is subject to understandings about what principle or procedure would be applied to address it.
  • Examples of postponement include the decision on whether Uganda should be a one-party (nonparty) state or a multiparty state. Iraq postponed some critical decisions on federalism because the issues were raised late in the day, and the request by the chair of the constitutional commission for an extension was rejected under pressure from the United States. In constitutional discussions on the secession and independence of Bougainville, New Caledonia, and South Sudan, the parties agreed to postpone a definitive decision by instituting a system of autonomy for a prescribed period (e.g., six or ten years), to allow time to see if autonomy met most of the concerns of the “seceding unit” (underwritten by guarantees that it would then be able to express its preference through a referendum, as an act of “self-determination”). In the case of New Caledonia (part of overseas France) and South Sudan, the referendum decision is binding, but in Bougainville it is merely advisory, triggering off a fresh round of negotiations between Bougainville and Papua New Guinea.
  • Partway between those two devices is the decision to acknowledge principles in the constitution for the ultimate resolution of the issue and leave the details and implementation for later. Often the controversy is about the details. The detailed issues may be determined with the help of expert opinion or a formal enquiry, or through a referendum (as in Spain).
  • Raising the divisive issue in the future. Closely connected with that approach is the possibility of raising the issue, whether unresolved during the process or new, in the future. The advantage of this approach is that there is no sense of closure or exclusion; the possibility that groups with specific claims may raise them in the future reconciles them to the constitution.
  • There are variations on this approach. The Spanish constitution recognizes the principle of autonomy, but the precise entitlements of regions are left to further negotiation and referendum. In Canada the opportunity to reopen an issue is largely implicit. Significantly, this approach has been made constitutional by the Supreme Court in Canada, in the decision on whether Quebec has the right to secede. A framework was devised within which the wishes of a part of the country to secede must be negotiated. Whether because of this decision or some other reason, the tension between Quebec and the rest of the provinces eased after this decision.
  • Leaving the divisive issue to other processes. A specific technique, consistent with some of the approaches mentioned above, is to leave the divisive issue to parallel or later processes, such as joint or special commissions, or the recommendations of expert group. When making its independence constitution [1975], Papua New Guinea set up a special committee, including outside experts, to recommend steps to be followed in order to set up the provincial government for extensive devolution. This approach was common in the process of decolonization, which involved specialist commissions on the electoral commission, regional or constituency boundaries, rights of minorities, and the scheme of devolution. Sometimes a matter was “resolved” by an assurance from the departing colonial authorities or the incoming government that the interests of a particular group would be addressed later. This is not a satisfactory approach in the absence of secure guarantees—there are several instances of failure to honor the undertaking.
  • Resolving the issue in the future through ordinary legislation or the judiciary. Sometimes an agreement may be reached to leave the issue to be resolved in the future, not through a constitutional process but in the ordinary working of the parliament, through legislation. Or it may be left to be resolved by the judiciary (for issues such as capital punishment, same-sex marriage, or abortion). But it is unlikely that seriously divisive issues can be resolved in this way. This may have more to do with the strategy of a drafting body, anticipating controversy in the deliberating body.
  • Using constructive ambiguity. It may be more common to reach an agreement with both (or more) sides read as supporting their position—what we call constructive ambiguity. Courts, which may be called upon to interpret the relevant provisions, may find themselves in an unenviable position; it puts unnecessary strain on them.
  • Anticipating difficulties on some issues, those who design the process may provide special procedures to resolve them. (We have already mentioned the use of referendums for this purpose.) This may be connected to the methods of voting, a specific negotiating process, or both. South Africans used the concept of “sufficient consensus,” whereby if the two major protagonists agreed on an issue, the smaller parties had to accept it. Sometimes, often informally, differences between two sides may be resolved by reference to the top leaders, who stay out of detailed negotiations among the parties. This happened in South Africa, where the buck stopped with Mandela and De Klerk, and later in Fiji. The Nepal interim constitution provides an interesting example. When an issue cannot be resolved by the constituent assembly through consensus, the matter is referred to the leaders of all parties for resolution. They have fourteen days to form a consensus, to be put to a vote in the assembly. If there is still no unanimity, another vote is taken for which the support of at least two-thirds of the members is necessary. Such a procedure leaves open the possibility that no decision will in fact be made.
  • Another example of a specific procedure is the Kenyan process [2010]. In one sense, the whole process was about “contentious issues” (essentially among political parties), which had prevented the adoption of a constitution since 2004. An elaborate process was devised, including an independent committee of experts (appointed by the parliament) interacting with both the people and the parliamentary committee on the constitution. It was for the committee of experts to determine what were contentious issues by comparing, in particular, three previous drafts. Its proposals, finalized after consulting the parliament, were to be submitted to a referendum. Implicit in this law, prepared through agreement among political parties, was the undertaking that the people would accept the legislative scheme.
  • Getting the help of third parties. Increasingly, divisive issues are resolved outside the formal framework for the constitution-making process, often with the help of third parties. We have indicated some local means for doing this. In many instances, third-party intervention is provided for by international or bilateral communities. This is particularly the case in conflict and postconflict constitution-making processes: Norway in Sri Lanka; the European Union and the United States in Sudan; eminent Africans and the African Union in Kenya; the United States and the United Nations in Iraq; the United States in Northern Ireland; the United States and the European Union in the Balkans; the United Nations in Afghanistan, Cambodia, Namibia, and Timor-Leste; and Australia and New Zealand in the Papua New Guinea/Bougainville negotiations. Sometimes the role is formal (as in Sudan, and often with the United Nations, and mostly based on United Nations Security Council resolutions). A typical form for intervention is the office of the special representative for the resolution of conflict. Such interventions often result in the marginalization of local communities and generally are contrary to the spirit and letter of the legal framework for constitution-making processes—particularly as in Iraq [2005].