Downloads: 2 available

Available in English

The negotiations in practice (1999)

Download Article
Mark Durkan provides a detailed account of the negotiating process, the formats and tactics, and how it was all brought together at the end.

On 30 March, Downing Street faxed its draft for Strand One to Hume and Trimble but to no one else. Anger was expressed in the strongest possible terms. The objection was not just the manner of delivery of such paper. 

Mark Durkan

The formal negotiating process that led to the Belfast Agreement on 10 April 1998 was unusually long and for long periods appeared to be going nowhere, but they also provided many opportunities to try different mechanisms to overcome obstacles when the talks became bogged down. Mark Durkan provides a detailed account of the negotiating process, the formats and tactics, and how it was all brought together at the end.


The formal negotiating process that led to the Belfast Agreement on 10 April 1998 was unusually long. The twenty-two months it lasted from 10 June 1996 provided many opportunities to try different mechanisms and assess how they helped negotiations to move forward. The mechanisms were not new in themselves but the participants were occasionally able to use them in innovative ways to overcome obstacles when the talks became bogged down. At other times individual parties persisted in supporting specific procedures which seemed to protect their interests and as a result for long periods little progress was made and the talks appeared to be going nowhere.

Procedural disputes

The talks got off to an acrimonious and inauspicious start with sessions dominated by the issue of rules of procedure. The unionist parties were resentful that the British and Irish governments had published a set of ground rules on 16 April and appointed the three independent chairs without consulting them. They argued that this set the precedent for the governments to control the talks. Consequently the first day was characterised by delay and uncertainty as the unionist UUP, DUP and UKUP parties rejected both the ground rules and the appointment, in particular, of former US Senator George Mitchell as the senior independent chair. The arguments regarding the rules were more than just pedantic points about procedural preferences. They were about the working ethos of the process and its political potential.

Doubt and suspicion were heightened by the unfamiliar format for negotiations which was designed to ensure the involvement of up to ten parties as mandated in the election of May 1996, including those with even very marginal electoral support. Provision was made for Sinn Féin’s inclusion in the negotiations at the invitation of the Secretary of State for Northern Ireland on the basis of a reinstatement of the IRA ceasefire.

There was anger and embarrassment as the independent co-chairs were not allowed to assume their duties. Media attention focused on unionist resistance to the appointment of Mitchell and their insinuations about his intended role. However, he and his colleagues were installed when the UUP, having made its point that it would resist any suggestion that proposals could be imposed on the parties, rested its objection. In return the two governments accepted that the talks participants would negotiate and agree rules of procedure for the negotiations themselves.

Those preliminary negotiations were conducted without agreed rules, with Mitchell effectively managing the proprieties with the confidence of all the parties. At this stage no time constraints were imposed on contributions in plenary meetings, and many arguments were long-winded and repetitive. It was inevitable that some parties would labour their attitudes to the ‘peace process’ and background events in this induction period of the negotiating process.

The independent chairs facilitated these negotiations more actively than they facilitated the substantive negotiations. They sought proposals from parties, provided comparative tables to guide discussion and used bilateral meetings to ease things forward. As total compromise did not emerge voluntarily, Mitchell ended up shuttling and brokering with his own compromise drafts on the still disputed points. Under the rules for the rest of the negotiations that he helped to establish, Mitchell did not play as direct or decisive a mediation role again.

The arguments revolved around the role and remit of the independent chairs; how comprehensive the agenda could be; whether or not parties had the right to raise issues free from veto; the principle of inclusion of parties ‘related’ to paramilitary groups on ceasefire; whether or not decommissioning was to be a ‘term of entry’ for Sinn Féin; and the construction of a mechanism to secure agreed decisions where consensus proved impossible. The formula adopted was ‘sufficient consensus’ which stated that a decision could be adopted if a majority of the representatives of each community, unionist and nationalist, were in agreement.

The adoption of agreed rules represented a more significant achievement than was recognised. It followed a very difficult summer with crises over Orange Order (the Protestant society linked to the UUP) parades and the resulting resignation of the SDLP from the unionist-dominated Forum, the elected body which had been established to provide a pool from which the negotiating teams were drawn.

The manner in which the rules were adopted is worthy of note. Even with the parties’ own adjustments and compromises and the mediation of the independent chairs there was still no unanimity over all areas of the rules. However, the DUP accepted a procedure under which the section of the rules on decision-making by sufficient consensus would be voted on first. This method was then applied to the other sections of the rules which meant that the DUP could record its opposition but not block procedures which it did not like – what Peter Robinson of the DUP termed a ‘good loser’ approach.

The format applied

A crucial part of the initial discussions was the arranging of a suitable environment for negotiations, including the division of issues between different meetings, the representation of parties, and the structuring of the meeting room. Since the Brooke talks in 1991 it had been accepted that there were three sets of relationships which should be dealt with in separate strands: Strand One dealing with internal relationships within Northern Ireland; Strand Two with relations within the island of Ireland; and Strand Three with UK–Irish relations. Consequently five ‘formats’ were named in the Rules of Procedure: Plenary, the three Strands, and a Business Committee. Under Rule 5 liaison sub-committees on confidence building and decommissioning were eventually established in order to separate out topics which threatened to undermine the ability of the parties to work together.

The Plenary format (chaired by Mitchell) as well as Strand Two and the Business Committee (both chaired by General de Chastelain) involved all the parties and both governments. Strand One (chaired by the British government) did not include the Irish government. Strand Three negotiations were between the two governments with arrangements for consultations with the other parties. To an outsider these arrangements may seem pedantic and immaterial, but the parties were attempting to establish a working environment that reflected their positions. For example, the unionists attempted to merge Strand Two and Strand Three so that they would participate in the discussions between the British and Irish governments while the SDLP and the governments wanted to retain the concept of separate and equal consideration of each Strand.

The term ‘format’ is also used to describe the form of the negotiations arrangements which deal with the numbers involved and the style and shape of meetings. UK legislation had provided for the larger parties to have three negotiating seats each and the smaller parties two, with all parties entitled to three people in support. The two governments each had three negotiating seats and five people in support. The three independent chairs also had their support staff (up to five in practice) and there were civil service note takers working under the direction of the independent chairs. This meant that formal negotiating meetings were relatively large. With the DUP and UKUP walkout on the entry of Sinn Féin there ended up being sixty-eight at the plenary sessions.

In these circumstances there was not really a negotiating table but a chamber with a large square of tables, two rows of seats and a lot of space in the middle. Contributions had to be by microphone for acoustical as well as procedural reasons.

The ‘top’ side of the square was occupied by the independent chairs flanked by the British and Irish governments. The parties were seated in alphabetical order and the chair, for a long time, took submissions from the parties in alphabetical rotation, always starting with the Alliance Party. For the purposes of the talks the DUP used its full title, the Ulster Democratic Unionist Party, thereby securing a later turn in a round.

Too many parties

Problems did not derive from the number of parties involved and unproductive bilateral meetings show that having a smaller number of parties would not necessarily have improved communication. Indeed earlier talks involving four parties suffered from a similar syndrome.

The large number of parties in fact helped communications on several counts. The negotiations were more inclusive, with the widest span of democratic representation ever involved. This helped to differentiate the process from previous failed initiatives and underlined that ‘ownership of the process’ was not just for the larger parties. The multi-party negotiations were in keeping with the likely composition of any new elected institutions which were expected to comprise more than just the bigger parties. The larger number of parties helped with the problem of language in the negotiations. In previous negotiations parties had tended to be very precious about their own terminology and quite jealous or suspicious about the language of others. In practice the larger number of parties generated a wider range of language and induced more flexible attitudes to innovative and variable terminology.

Controlling the agenda

The fact that the Opening Plenary ran from June 1996 until September 1997 could not be blamed on the format of the negotiations but on the attempts of parties to advance their aims by controlling the topics for discussion. Such considerations lay behind the unionist insistence that no substantive negotiations could begin until they were satisfied on the issue of decommissioning. An alternative proposal that the Opening Plenary should only ‘address’ rather than ‘resolve’ the issue was not accepted.

The following weeks and months saw the strongest manifestation of ‘variable geometry’ – the concept of flexibility in the shape and size of meetings. Bilateral discussions between the UUP and the SDLP and later trilateral discussions involving the Alliance managed to produce a working paper on decommissioning on which there was, even according to UUP sources, ninety per cent agreement. At the same time bilateral discussions had produced agreement between the SDLP and the UUP on a possible Comprehensive Agenda for the substantive negotiations which was designed to reassure each party about the intentions of the other. The Alliance Party was also willing to support this proposal. When the Opening Plenary finally concluded on 24 September 1997 the Procedural Motion contained the SDLP/UUP/ Alliance proposed Comprehensive Agenda and relied heavily on the ‘unagreed’ decommissioning paper.

The fact that a Westminster election was pending seemed to further inhibit some parties for the first few months of 1997. Aware of a likely change of government, the UUP, along with other unionist parties, now stressed the differentiation between ‘decommissioning’ and ‘the terms of entry for Sinn Féin’. Tedious and pedantic as the differentiation seemed to some, it indicated that David Trimble was trying to establish a justification for remaining in talks in the event of the admission of Sinn Féin after a second IRA ceasefire.

The ‘sufficient consensus’ rule caused a significant problem in the way it was applied in the Opening Plenary. The independent chair was reluctant to convene meetings or set business without sufficient consensus for this course of action. This made it impossible even to discuss issues on the basis of ‘taking note without voting’. Extending the sufficient consensus requirement to even relatively mundane matters of procedural management served to compound the sense of gridlock and heighten the frustration.


[[{"type":"media","view_mode":"media_original","fid":"1349","attributes":{"height":573,"width":566,"class":"media-image media-element file-media-original"},"link_text":null}]]

Surfing the agenda

One shared tactic to avoid difficulties in dealing with specific substantive issues was to review widely all the issues or ‘surf the agenda’. When the ‘substantive negotiations’ finally started in October 1997, the chairs decided to take each of the Strand One and Strand Two agenda items in turn, normally meeting one day per Strand each week. The intention was to allow parties to air and share all the issues that mattered to them and to gain some measure of the approach being taken by others. Parties were invited to submit papers on each item on which they were entitled to speak in the large negotiating format.

The parties’ contributions and exchanges centred more on what they were not prepared to compromise rather than on what they were willing to accept. With limited written submissions from some, conflicting historical analyses and a shortage of actual proposals this ‘surfing the agenda’ exercise had little substantive value. It did, however, mark some sort of graduation in the negotiation process with parties starting to get a sense of how serious others were about given issues. Some significant points were made and with some passion. The importance of good listening was now more apparent and relevant – at least to some.

Experience in smaller formats

Everyone agreed on the need to get the negotiations to move up a gear or two but not on how to achieve this. Proposals tended to be about getting stuck into an issue or locking on to one Strand to get some substance established. It was not just the agenda that mattered but format, with all parties seeming to agree on the limited value of the big room sessions.

During the long Opening Plenary bilateral and multilateral meetings were the favoured channel for real negotiations. However, these also had a questionable success rate. Successive bilateral discussions with all the parties in turn were hardly stimulating. Impressionistic accounts of other parties’ positions or intentions were exchanged at second and third hand and different parties focused on quite different things in the same round of bilateral or multilateral discussions. These ineffectual meetings usually took place as a way of trying to avoid sterile sessions in the large format.
In 1997 a smaller Plenary sub-group format was established with two representatives from each party to identify key issues for resolution and to propose an appropriate format for dealing with them. A new room was created adjoining the conference room to accommodate this smaller format. It was not free from tension or recrimination and had no guarantee of enhanced understanding.
Some private press briefings by government officials misleadingly heralded this exercise as ‘drawing up heads of agreement’ – something that parties were hardly minded to do going into a three-week Christmas recess. Based on these discussions the independent chair offered a paper outlining the main issues which needed to be addressed but, predictably, parties had various sensitivities and reverted to their own proposals in the subsequent discussions.
In deference to the collegiate spirit which was supposed to inform this smaller format, SDLP negotiator Seamus Mallon once asked if there might be a coffee break of fifteen minutes ‘so that a couple of us could talk with a view to offering some compromise on this’. The two SDLP and UUP representatives stayed on in the meeting room and managed to reach some agreement that was given strong encouragement by both governments. Sinn Féin reacted adversely and the Irish government declined to support the SDLP/UUP compromise. At one point Sinn Féin complained that people had asked for a coffee break but had not even gone for coffee while the SDLP and UUP delegates were pointing to their coffee cups still on the table.
Engineering ‘sufficient consensus’ was still difficult. In this context it was also inappropriate. The SDLP view, made clear at the time, was that any such reconstitution of the agenda should be agreed by consensus to avoid creating unnecessary and unhelpful ‘winners and losers’. A ‘note not vote’ option which allowed the tabling of papers on key issues which participants could note without prejudice but would not have to vote on was rejected. The UUP began to insist that this sub-group format could take no decisions anyway.
The seriousness of the situation was brought home to participants. At the last Plenary before Christmas the independent chair cautioned participants that they were going into probably the last recess of the negotiations. The SDLP warned that the failure to produce a working list of the key issues and agree a format would be an even greater problem when negotiations resumed. The warning that ‘the calendar will not stay neutral just because we have called a recess’ was borne out by subsequent statements and events when the Secretary of State and some party leaders went into the Maze Prison to discuss the process with sceptical loyalist paramilitaries. At this point the prospects for progress seemed to be fading fast.

Establishing engagement

In early 1998 signs of ‘chemistry’ became evident as parties and personalities got back to each other’s points or ideas in other than negative tones. It is not entirely clear why this was so but a number of developments may have helped. Responding to the difficulties during and before the recess, the two premiers in January prepared their own ‘heads of agreement’ paper similar to the ‘key issues’ drafts on the table before the recess. The aim was to give the resuming negotiations some measure of focus as the proposed May deadline came closer. 

At this stage the smaller negotiating format was used but with only two assistants allowed in support of the two representatives per party. Some parties fielded different negotiators according to the issue under discussion with the effect of giving mixed messages at times. Parties canvassed their various proposals, concepts and models but, more importantly, with some tactical exceptions, they explored each other’s ideas, seeking further explanations or offering explanations for their own reservations or objections. The SDLP and UUP eventually had language proofing sessions where they could identify and explain for each other their sensitivities about the sort of terms they were respectively using or likely to use. These sessions lapsed into discussions of substance but this was hardly a problem in the circumstances. 

Sessions in London and Dublin (three days in each city) in this period were dominated by the exclusion of the UDP in London and Sinn Féin in Dublin because of breaches in the ceasefires. Although the intended benefits of a change of scenery were not realised, one of the most significant and poignant personal contributions of the talks process was made in London by Reg Empey of the UUP when he struggled to show some appreciation of the attitudes and calibre of the republican negotiators, and asked for something similar for unionism. The genuineness of the terms in which he spoke had an impact on the other parties that they tried to reciprocate. In bilateral talks, as well as rehearsing arguments on the issues, parties were sharing constituency sensitivities and exchanging thoughts on how the negotiations could be advanced to a productive climax.

The government-ordained 9 April deadline was needed. With the deadline in place parties began to press on the issues which had not been substantively pursued in the negotiations to date. They stressed issues and ideas of particular importance to them such as policing, prisoners, civic forum, decision-making and safeguards, and the UUP returned to the decommissioning question. It was recognised that policing could not be properly addressed in that negotiating context and that some form of commission should deal with it. Predictably, most parties went into public relations overdrive in this period which added to the impression of intractability.

However, what was needed was reflection on what was being said and on options that would as far as possible meet the requirements of the parties. For example, an informal ‘homework club’ evolved within the SDLP in which a small group of party members met to discuss options, test them against party requirements and against those of other parties, redraft them for better acceptability and workability and even produce alternative versions to have in reserve for the conclusive negotiations. SDLP party member Seamus Mallon told this group at the time to ‘let the others draft their press releases, we have to draft an agreement’.

The rapid development of proposals at this stage is exemplified by the emergence, only weeks before the deadline, of the concept of a binary office of First and Deputy First Minister, in which all functions would be exercised jointly. There was also concern that key decisions in the proposed assembly should have cross-community support. This led to a proposal for the adoption of ‘parallel consent’ under which a majority of both unionists and nationalists would be required for such decisions.

Bringing it all together

The real challenge now was how to bring the negotiations, imperfect as they were, from all the different Strands into one arena so that one agreement could be decisively negotiated. An interim deadline of 3 April 1998 was set for the independent chairs to produce a paper that would be a draft agreement for final negotiation. Mitchell held further private meetings with individual parties to explore their proposals before completing the draft. At this stage he was doubtful about the inclusion of specific ideas which might not obtain broad support, and parties were concerned that the draft might not reflect their ideas. Following its meeting with Mitchell, the SDLP was somewhat worried that he was not encouraging about its Strand One proposals on ‘parallel consent’ and other related safeguards in relation to decision-making and their proposal for a joint office of First and Deputy First Minister. He made it clear that, as with Strand Three, the Strand Two section of his paper was being prepared by the two governments.

Worry about Mitchell’s likely approach was replaced by profound concern and resentment when, on 30 March, Downing Street faxed its draft for Strand One to Hume and Trimble but to no one else. Anger was expressed in the strongest possible terms. There was no question as far as the SDLP was concerned of negotiating with a fax machine. The objection was not just the manner of delivery of such a paper. Its content reflected very little of the party’s thinking.
The sense that this was no way to manage serious negotiations was strengthened by the fact that Mitchell knew nothing of such drafts until told of them by the SDLP. When a subsequent Downing Street draft was sent on 2 April, a corrected or revised version followed within a couple of hours; and even Paul Murphy, the British Minister of State at the negotiations, and his officials were unaware of this when the SDLP went to discuss concerns with him.
The 3 April deadline for the Mitchell paper was missed because the governments had not finalised their Strand Two paper. Mitchell wanted to circulate the paper minus Strand Two but the other parties advised strongly against this for a variety of reasons but most importantly because the proposals needed to be considered as a complete package. While waiting for the Mitchell paper on 3 April and fending off Downing Street’s faxes, parties continued their own bilateral discussions. On the morning of 3 April the UUP and SDLP held a bilateral meeting to address some major issues of difference. The UUP was concerned about the idea of an executive model in Strand One as well as Strand Two as a whole and at this point the SDLP outlined its concept of the joint office of First and Deputy First Minister which had already been developed in the ‘homework club’. The UUP expressed working interest in the concept but declined the offer of a paper. Other contentious points included decision-making safeguards such as ‘parallel consent’. These were also points of difficulty for the Alliance Party. It was only in the final week that the SDLP came up with the conclusive answer to these points with the idea of electing the First and Deputy First Minister jointly by ‘parallel consent’.
When the Mitchell paper was circulated on the night of 6 April, only three days before the deadline for the end of the talks, there was an intense negative reaction from unionist parties and the Alliance Party to the Strand Two paper because they found the proposed cross-border links were too strong. Privately the SDLP, the only nationalist party negotiating in Strand One, was disappointed and worried about the Strand One proposals because only marginal consideration had been given to their ideas. When the two premiers arrived the UUP and Sinn Féin concentrated on Strand Two but negotiations also continued on prisoners, policing, equality, human rights as well as the other Strands.
There was therefore no collective or round-table negotiation as such. In all probability no one participant, not even the independent chair or either prime minister, knew everything that was going on throughout this period. It was not perfect traffic management but it beat the gridlock. Instead of the stalemate of a zero-sum game we now had a complex matrix of negotiations. Some participants were surprised – not all pleasantly – when they saw the aggregate outcome in the proposed final Agreement on 10 April. The Plenary session to formally table the Agreement was delayed as a result of discussion in the UUP delegation on whether to accept a package which did not resolve all their concerns. Meetings between party leaders and the prime ministers continued as problems were cross-checked and supplementary issues raised. Rumours and varying moods went round the building like waves as people were tired after a sleepless night, hungry, with no food in the building, and relieved, surprised, exhilarated or humbled with the anticipation of agreement.
Those final two weeks had seen the sort of constant and concentrated negotiation that was crucial. The role of the two prime ministers had been decisive in terms of negotiation and brokerage, but the role of the parties should not be underestimated. The patience and sensitivity of the independent chairs – not just George Mitchell but Harri Holkeri and John de Chastelain too, helped to facilitate all of this. Mitchell’s sense of purpose was matched by his sense of patient caution. He had all along talked of himself as ‘Humpty Dumpty – I can only jump once’. His instinct for the right time was acute, as demonstrated by his decision to call the plenary session at fifteen minutes’ notice on the evening of Good Friday.
  • The negotiations will operate on the basis of consensus. If it appears to the Chairman, after a reasonable period of discussion, that there is no unanimity, he may follow one or more of the courses of action set out below: (a) consult with the participants, with a view to putting forward a solution that he believes will secure agreement; and/or, (b) invite the participants to set up a broadly representative working group, including nominees of both Governments (or of the British Government alone in the case of Strand One issues); and/or, (c) obtain the agreement of the participants to refer the matter to an agreed group of experts for advice, requesting a report within an agreed specified period.
  • It will also be open to the participants, acting solely by agreement and only at their instigation, and subject to the provisions set out in paragraph 37 below, to refer the matter to the forum for consideration, requesting a report.
  • If, after the participants have considered any further proposals arising from one or more of the courses of action set out above, there is no unanimity on a particular proposition, the Chairman may propose that negotiations proceed on the basis that the proposition has been approved by sufficient consensus, as defined in paragraph 34.
  • Sufficient consensus may apply to all decisions taken in any format.
  • A proposition would be deemed to have sufficient consensus when supported by participating political parties which, taken together, shall have obtained a majority of the valid votes cast in the elections held on 30 May 1996 and which between them represent a majority of both the unionist and nationalist communities in Northern Ireland respectively and which also constitute a majority of the participating political parties. With the exception of Strand One, both Governments must endorse the particular proposition for it to be deemed to have sufficient consensus.
  • Where it does not prove possible to achieve either unanimity or sufficient consensus, the Chairman will work, in consultation with the participants, to remove obstacles to the reaching of agreement, or, with the agreement of the participants, may decide to move on to the next item on the agenda.
  • The negotiations will proceed on the principle that nothing will be finally agreed in any format until everything is agreed in the negotiations as a whole. Subject to this principle, it would nevertheless be possible, solely on the basis of agreement among the participants, to proceed on the assumption of contingent agreement on any individual aspect of the negotiations.
(From the Rules of Procedure 29 July 1996)

Issue editor