In 1991, as the conflict in Northern Ireland raged, I was part of a small group of what might be called 'intellectuals' with various positions in civil society (organised labour, NGO sector, academia, journalism) who met to agonise over the eternal question: "What is to be done?" We hit upon the idea of inviting a commission of outsiders to come and hear from ordinary people about life in Northern Ireland and to make recommendations on new ways of tackling what seemed to be an intractable problem. We had a strong economic and social agenda, not entirely divorced from the constitutional question about our Irishness and/or Britishness. Initiative 92, as the project became known, gave birth to the first halting steps of a new form of civil society engagement with what became known as the Northern Ireland peace process.

We approached Torkel Opsahl, a Norwegian human rights lawyer, to chair the commission and invited others to serve alongside him. The same group managed the organisational framework, the fundraising and the promotional aspects of the commission's work. This involved taking the initiative into the lanes, streets and by-ways of Northern Ireland to hear what local people felt and to nudge hitherto recalcitrant political blocs to engage in some sort of process beyond violence. The initiative culminated in the compilation of a report that was published, launched and disseminated among the political parties and the wider public.

We all agreed that we must not stop with the publication of a book and therefore arranged for a one-year extension to the project to disseminate and animate the results through an extensive follow-up programme. Of course the first IRA ceasefire of 31 August 1994 was not a direct consequence of the process we initiated. But with the benefit of hindsight, many observers pinpointed our contribution to creating an atmosphere of greater participation in debate, easing the situation and softening the edges of the conflict.

I remember feeling alienated as a civil society practitioner when the talks process chaired by Senator George Mitchell began; the talks were taking place behind closed doors (however understandably), but I wanted to help the process along by assisting with explanation, communication and elaboration of the key principles of any accommodation then being negotiated. Another role soon opened up. The UK government of John Major had promised a 'triple lock' before any negotiated agreement could take effect: the parties to the negotiation must agree, the two supervising governments (the UK and Ireland) must agree and then the people must agree through a referendum held simultaneously on both sides of the Irish border. The referendum created the opportunity for civil society players to organise a "YES" Campaign. So we did and thereby contributed to the 81 per cent turnout – massive by UK and Irish standards – and the 71.2 per cent vote in favour of the agreement.

Yet the resoundingly endorsed Belfast Agreement, signed on Good Friday 1998, was not the end of the story. Again blessed by hindsight, we can see that it was only the end of the beginning. Its implementation was – and remains – critically contested, again requiring the engagement of players other than elected political representatives to help 'oil the wheels' of the process. And so we are still involved as observers, commentators, activists, trade unionists, business people and NGOs.

This personal vignette shows three distinct phases in my modest contribution to recent events in Northern Ireland: preparation for peace, the formal negotiations and implementation/consolidation. In Northern Ireland, the drive toward negotiations came principally from internal actors. Simultaneously, however, external pressures from the United States – sometimes stimulated by the influential Irish diaspora – and the European Union added urgency to the dynamic. The principal political parties, as selected by the electorate, were responsible for negotiating an agreement. Civil society's role was to help prepare society for change.

This process reveals some elements in the developmental sequencing of political participation by the public in a wider peace process that I would like to explore in more detail, drawing on experiences in Northern Ireland and elsewhere. Hindsight often makes it possible to chart a linear progression between these phases. At the time, however, it may often feel more like a zig-zag, as initiatives break down, interested parties position themselves in a way that offends others, fears and apprehensions increase the contested territory – and often spark violence – and a curious, if not confused, electorate express their apprehension or alienation from the process.

If civil society organisations and a broader proportion of the overall public are sufficiently prepared to engage in peacemaking, it can both create a climate conducive to negotiations and help to ensure that the social infrastructure is developed for their voices to be heard at formal peace talks. As the Northern Ireland example shows, there are typically combinations of internal and external influences that encourage representatives of armed combatant groups to come to the negotiating table. These influences interact with civil society roles in complex ways. The role of external brokers and mediators can be helpful but only if the process is itself 'owned', sooner or later, within the conflict area. Here there is a neat balance between the role of catalyst and the role of guest facilitator.

Often external forces have an overt role. Sometimes they come from countries with considerable influence on the conflict. In Northern Ireland this was exemplified by US Senator George Mitchell, economic envoy and then chair of the formal talks process. There are also governments whose diplomatic corps are well known for their impartial observers and mediators. For example, the 'Oslo Process' is a term used both by Israelis and Palestinians and by Guatemalans. In both cases, the Norwegian government supported civil society facilitated 'talks-about-talks' between representatives of the combatant groups leading to a substantive negotiation process. In Mali, the Norwegians provided support when Norwegian Church Aid was asked to help with the inter-community meetings that eventually brought peace.

In other cases, internal actors are more important. In South Africa, the immense mobilising capacity of the United Democratic Front allied to Nelson Mandela's ANC was a powerful force motivating the transition from apartheid. A joint business and church led initiative enabled the 1991 National Peace Accord process to be firmly rooted in civil society but connected at the same time to the key power-brokers. In the Philippines, the agenda for peace talks emerged following a July 1992 initiative by newly-elected President Fidel V. Ramos to create a National Unification Commission (NUC) mandated to hold consultations with all concerned sectors at the provincial, regional and national levels. The NUC worked together with local actors so that the consultations at the local and regional levels were led by civil society, both practically and symbolically, rather than the government. Often called the 'laboratory of peace', the community of Mogotes, Colombia established a unique Municipal Constituent Assembly using classic 'bottom-up' mobilisation techniques. In the context of ongoing and widespread violence, community members initiated a peaceful process of political change at the local level. Mogotes is now a 'zone of peace' that has inspired hundreds of others throughout the country.

These examples indicate that the mechanisms for public participation in peace processes can be extremely important. Yet they do not occur unless people make them happen. This typically involves a substantial degree of both advocacy, to ensure that their voices are heard, and mobilisation to generate the capacity to create opportunities – whether proactively or reactively. The preparatory phase is thus a time for mobilising voices, formulating substantive agendas, designing processes and developing a popular constituency of interest to support and engage in conflict resolution. Sequencing, however, is critical. For example, consulting civil society on the contents of the peace agreement after it has been negotiated and signed can be difficult and sometimes counterproductive, especially if substantive input is no longer sought or possible. Power-brokers may seek to marginalise civil society voices after initial consultations so as to dominate decision-making and reconsolidate their control in the post-agreement period; caution must be exercised to ensure they do not dictate the process.

In addition to preparing processes, it is important to emphasise the significance of preparing people so that they feel comfortable and are able to participate fully and make effective contributions. Alienation and frustration can otherwise set in very quickly. Participation may be impaired if the environment is exclusionary because of over-reliance on unfamiliar procedural rules, for example, or a formal style that intimidates those with less experience.

There are a number of dilemmas regarding who sits at the negotiating table: representatives of the armed groups? Political parties? organised civil society? This question can become a knotty problem. Are politicians not representatives of the people? Are they not entitled to negotiate the best deal and then sell it to their constituencies of interest? How can civil society, essentially and inherently representative of particular interests, play a helpful and supportive role?

In Northern Ireland the multi-party talks leading to the Belfast Agreement took place behind closed doors. Although the public was aware of the process, there were few official channels to allow input into the content of the agreement being drafted. This created a degree of concern amongst organised civil society and also created a barrier to 'bringing along' the public in support of the Agreement. This was exacerbated by selective leaks amongst the actual participating parties to 'spin' their side of any particular argument in order to gain political advantage. This made the task of selling the agreement to the public in the subsequent referendum, scheduled for only six weeks after the Agreement was signed, all the more difficult.

By contrast, the South African model took an alternative route to write a new constitution. The Constitutional Assembly deployed a multi-track approach. The negotiations deployed a representative decision-making model with elected party representatives negotiating on behalf of the members of political parties, which held them accountable. This was complemented by broad public consultation on the contents of the new constitution so that everyone would feel that they had an opportunity to articulate their perspectives, interests and ideals. There were high levels of awareness about the process and a broad feeling of ownership over the constitution that finally emerged.

Sometimes the overlap between civil society and the formal political process can be seen as an opportunity. For example, to ensure that women would be present at the negotiating table in Northern Ireland, a cross-community group of women from civil society formed the Northern Ireland Women's Coalition to contest the elections and won a place at the talks. Alternately, engagement in official politics can be a threat to civil society's effectiveness, as was the case when Guatemala's Civil Society Assembly (ASC) was weakened at a critical moment by the exodus of civil society leaders to assume new political roles.

Here, the distinction between organised voices and wider public opinion may be helpful in differentiating the levels and tiers of involvement. Even within organisations, some members or groupings may feel disenfranchised or overlooked, stimulating them to work outside their authorised organisational structures. Many will also use the press, such as the letters columns of newspapers, to create the mirage of participation.

Sometimes, public participation can be critical when a formal process falters or breaks down. Civil society activists can nudge key actors back into a peace process, as was seen in Guatemala, Mali and the Philippines. They can also help to create or maintain a climate conducive to negotiations. In South Africa the local and regional peace committee structures underneath the National Peace Accord were able to help stabilise the situation in April 1993 after the assassination of the ANC activist Chris Hani. NPA structures also contributed to the creation of conditions where a negotiated political transition could take place, by facilitating communication and modelling non-violent approaches to the myriad manifestations of the political crisis.

Senator George Mitchell, after signing off on the Belfast Agreement, is said to have observed: "poor souls – now the real trouble begins!" He was observing that regardless of the difficulty of the negotiations, the implementation often can be even more difficult. It is important to build a broad constituency of support for the process, not just for the piece of paper that was negotiated.

Those who want to derail the agreed package often can do so through selected acts of violence or political vandalism. Hard-line factions within armed groups may feel aggrieved that their principles were diluted to achieve the necessary accommodations and those left outside the process can spot the moment to disrupt it yet further. The challenge can be to encourage them to realise that they are better off inside the process than in the wilderness seeking to damage from outside – or to ensure that they are so far outside that they can do it no serious harm.

Marrack Goulding observes the difficulties of an active civil society role in the post-settlement phase if the institutional capacities are weak. A salutary lesson can be drawn from Northern Ireland, where the agreement mandated that a civil society voice could be institutionalised in the new governing Assembly through a formally structured Civic Forum. Although agreed by the participants at the talks, incorporated into the agreement and endorsed by the people in the referendum of May 1998, the Civic Forum was stillborn. It was effectively starved of resources by the 'real' politicians and hampered by a cumbersome electoral college mechanism of appointment. It has failed to excite interest or notice: organised civil society works around it, the politicians ignore it, and the press do not trouble with it.

At the time of writing in autumn 2002, the Northern Ireland process has run into further difficulties with a breach of trust between the four main political parties operating devolved governance under the 1998 Agreement; a crisis that led to the suspension of the Northern Ireland Assembly by the British Government. As a fierce 'blame game' takes place, the public feel generally disillusioned and alienated from the process. Perhaps there should have been structural mechanisms to bind the communities into the new governing system; perhaps the political parties succumbed to temptation in seizing the spoils of agreement for themselves, without reflecting on how its roots must extend into local communities in order to underpin the fragile settlement.

It seems that unless the people are closely involved, through their representative organisations and through organised public opinion, any agreement can merely become a dry letter. The power of the people is beautifully exemplified by Mogotes, where the population organised non-violent direct action to force the resignation of their corrupt mayor. In Mali, political negotiations were not in themselves able to provide the foundations for peace. It was only when communities took responsibility for resolving the conflicts affecting their regions that peace was achieved and the conflict transformed. The conclusion here is that 'bottom-up' and 'top-down' travel in different directions and cannot be neatly corralled together. The jigsaw of relationships requires sensitivity and flexibility in order to give appropriate place to each and dampen neither's independence or autonomy. To do so effectively is a challenge for all governments and civil societies.