It is an academic, personal and political honour to give the ninth John Whyte memorial lecture.footnote It is an academic honour because John Whyte was the most dispassionate analyst of our conflict—and so is a hard act to follow. Interpreting Northern Ireland still conveys his marvellous gifts of clarity and concision in exposition. It is a personal honour because, together with Ernest Gellner, he was the mentor who had the greatest influence on me as a young lecturer. Lastly, it is a political honour. John Whyte worried whether social scientific research on Northern Ireland was worthwhile. Nevertheless, he contributed extensively to public deliberation in defiance of his occasional despair on this matter. He would have been pleased at the extent to which social science, including political science, can be discerned in the making and nature of the Agreement.

The Agreement of 10 April 1998, ratified in referenda in both parts of Ireland on 22 May 1998, is a major achievement, both for its negotiators and for the peoples of Ireland and Britain. To make it, many politicians, officials, paramilitaries, and ordinary citizens had been through trials by ordeal. It emerged from a political desert whose only landmarks were failed ‘initiatives’. Yet the Agreement that emerged from that desert has no agreed name. It carries no person’s name, British or Irish or American, and the names of no roles, be they Prime Ministers,Taoisigh, Secretaries of State, Foreign Ministers, or Party Leaders. Some know it by the place it was made, as the Belfast Agreement, or, more controversially, as the Stormont Agreement. But it was not signed by all of its supporters in the final negotiating chambers, and it was made in many places: in Dublin, London and Washington; in smaller cities, towns and villages; and in airports, aeroplanes, and unofficial ‘communications’. Some just know it by its date: the 10 April 1998 Agreement, or the Good Friday Agreement. The former seems too limited, while the latter, gives too much credit to Christianity—both as a source of resolution, and as a cause of conflict.

It is also known as the British-Irish Agreement, after the peoples who confirmed it in referenda in both parts of Ireland—though, strictly speaking, only the British in Ireland as well as the Irish in Ireland were asked to ratify it. I prefer to call it the British-Irish Agreement. This name reflects an important fact: the Agreement is the fulfilment of a previous Agreement, the Anglo–Irish Agreement. But we Irish and British know that much resides in names, and, to avoid giving any further offence to anyone’s sensibilities, I will refer simply to the Agreement.footnote1

What kind of institutional Agreement is it? The answer for a student of political science is that it is a consociational agreement, that is, a political arrangement that meets all four of the criteria laid down by Arend Lijphart:

A consociation is an association of communities—in this case the communities are British unionist, Irish nationalist, and others. A consociation can be created without any explicit consociational theory to guide it—indeed that has often happened.footnote3 More often, consociations are the equilibrium outcomes of bargains or pacts between the political leaders of ethnic or religious communities. This Agreement is the product of tacit and explicit consociational thought,footnote4 and of bargaining, or of what is sometimes called ‘pacting’.

But the Agreement is not just consociational, and it departs from Lijphart’s prescriptions in important respects that have practical implications for Northern Ireland and for regulating ethnic and national conflict elsewhere: it has important external dimensions; it was made with national, and not just ethnic or religious communities; and it has been endorsed by (most of) the leaders and (most of) the led. Indeed, I suspect it is the first consociational pact to have been immediately popularly endorsed by referendum. To be formulaic: the Agreement envisages an internal consociation built within overarching con/federal institutions; it has imaginative elements of co-sovereignty; it promises a novel model of ‘double protection’; and it rests on a bargain derived from diametrically conflicting hopes about its likely long-run outcome, but that may not destabilize it. One supplement must be added to this very lengthy formula. The Agreement is vulnerable both to post-Agreement bargaining and to legalism. Let me justify this phrasing.

At the heart of any consociational arrangement is executive power-sharing. The Agreement creates a dual premiership. Indeed, it can be argued that it establishes two quasi-presidential figures in a devolved Northern Assembly: a First Minister and a Deputy First Minister. They have presidential characteristics because, once elected, it is almost impossible to depose them until the next general election—presidentialism means an executive that cannot be destroyed by an assembly except through impeachment, and, in future, I maintain that it will be extremely difficult for the Northern Assembly to remove its dual premiers. Let me make this clear through a current illustration. Even if David Trimble’s party colleagues were to vote unanimously to depose him from the leadership of the Ulster Unionist Party (uup), he could not be forced to resign his position as First Minister. If he did not wish to go, he could only be deposed if enough nationalists colluded with enough unionists to enforce it—but, to do that, nationalists in the Assembly would have to bring down their own Deputy First Minister. This possibility exists because the First Minister and the Deputy First Minister are elected together by the parallel consent procedure (see Insert I). This procedure requires them to have the support of fifty per cent of registered nationalists and unionists as well as a majority of the Assembly. Critically, this rule gives very strong incentives to unionists and nationalists to nominate a candidate for one of these positions that is acceptable to at least a majority of the other bloc’s members in the Assembly. So, even if in the future Gerry Adams leads Sinn Féin into surpassing the sdlp in seats won in the Assembly, unionists will be able to bloc his nomination as Deputy Chief Minister. Likewise, nationalists can veto an unacceptable hard-line unionist. In the first elections for these posts, pro-Agreement unionists in the uup and the Progressive Unionist Party voted solidly for the combination of David Trimble of the uup and Seamus Mallon of the sdlp. Naturally, so did the sdlp. Sinn Féin deliberately abstained to avoid the First and Deputy First Ministers being chosen by more nationalists than by unionists—an outcome that might have endangered Trimble’s status with the unionist public, and a sign of Sinn Féin’s maturing avoidance of provocation.