This publication is copyright Malachi O'Doherty 1998 and is included on the CAIN site by permission of Blackstaff Press Limited and the author. You may not edit, adapt, or redistribute changed versions of this for other than your personal use without express written permission. Redistribution for commercial purposes is not permitted.

One night in May 1995 in the Felons club, in west Belfast, an IRA man called Bobby Storey directed me into a corner for a quiet word. Bobby Storey is about six foot four and I am five foot two so I wasn't about to argue. He was a human tower, and the hand he pressed against my chest was nearly as big as my chest itself. I have been asked many times if I am not afraid of being duffed up by the IRA for all the rude things I have said about them. I generally expect they won't respond to me like that, but at that moment I was contemplating that this might be a breach with tradition.

'I am not going to hit you, but I want to tell you that you are a slug,' Storey said.

I started to reply but he asked me to apply my energies instead to a consideration of the merits of his statement. 'Shut your fucking mouth,' he said.

After I had read it, I searched the available literature for information on Bobby Storey. Bowyer Bell (1993) wrote that the police nicknamed Storey 'Brain Surgeon'. Normally this would be a way of saying he was stupid. He's not stupid. BS, Bobby Storey, Brain Surgeon. The initials matched, and presumably the nickname appealed to some policeman's sense of humour.

Bowyer Bell was given access for his research to all levels of the IRA by its army council, but this detail seems to have come directly from the Royal Ulster Constabulary (RUC). How else would Bowyer Bell have known a code name used in police communications? Bobby Storey, I presumed, would have read the book himself, or at least flicked through the index for references to himself, and learned what was in the public domain about him. Apparently he hadn't.

My point was to argue that the ceasefire that we depended on was itself dependent on people like Brain Surgeon. That was the reality we lived with. How did we like that? I wanted to ask if this was an inevitable part of any peace process, that we would become dependent on the disciplinary machinery within paramilitary groups themselves for public order.

This idea had been mooted some years before, in September 1988, by Frank Wright and John Lampen (Fortnight 265). They contrived a sketch in which various parties, including Sinn Féin, and the British and Irish governments, discussed the possible steps towards a British announcement of intent to withdraw from Northern Ireland, and towards a ceasefire by the IRA. In the sketch, Sinn Féin complained that making an IRA ceasefire a precondition for Sinn Féin being given a place at talks imposed on them the responsibility for curtailing or disowning all republican violence. Wright and Lampen were dealing with political pragmatism, however, not with the moral question of how a state or its people might become dependent on law-breakers, and potentially on law-breaking, to maintain the peace.

That question arose again in May 1997, when John Alderdice of the Alliance Party argued for the removal of loyalist parties from talks because of violence coming from groups that those loyalists claimed they were not linked to. Billy Hutchinson of one of those loyalist parties, the Progressive Unionist Party, reminded a BBC audience (Good Morning Ulster, 27 May 1997) that, the year before, Lord Alderdice had been similarly critical of the loyalists when they had issued threats to try to keep loyalist dissidents in line. The implication seemed to be: either the Alliance Party should accept that loyalists had to police their ceasefire, or it could not penalise them for failing to do so.

A ceasefire is a military operation. The authority of military leaders is required to maintain it. Those leaders may be people who are enemies of the state; they may be people who have ordered or conducted the murders of hundreds of people in terms that we would wholly abhor, but when they have taken on the job of containing their own murderous followers, through the authority they have acquired over those people by leading them to murder, then our interests in them are reversed. We are concerned now that they should succeed.

Brain Surgeon was working for us now.

'You got all that shit from the police,' he said.
'No, I read it in a book.'
'Shut up. I have got the text of what you wrote.'
'I know, I sent it to you.

A few weeks later another journalist was talking to one of Storey's friends and asked him why Storey had leaned so heavily on me. Storey's friend said, 'I guess Bobby doesn't like Malachi thinking he is free to come and go in this area when he knocks the Provies so much. Coming from the area, people would expect him to have a bit more understanding.' I have heard this before. It tires me. It presumes that anyone who grew up around the IRA, or with the IRA around them, would learn to see things their way. Either they would cease to rest their political convictions on moral principles, in the same way we all did when we relied on Brain Surgeon to police the ceasefire, or they would find moral certainties of a new kind, that would allow them to stand shoulder to shoulder with the killers. Either they would discover that there were no rights and wrongs and that IRA members were at least no more reprehensible than anyone else, or they would find a greater evil in the state than it normally displays in middle-class areas, and would understand how the options of republicans were limited to the tragic but self-sacrificing course they had taken.

I hadn't bought either of these propositions. There was an obvious paradox for me in that it was usually apologists for the IRA rather than the IRA themselves who argued that they had no choice but to conduct themselves as they did. It became part of the common lore that these were people who had suffered, been humiliated and angered, and who were striking back out of their hurt and rage at the British army or the RUC. Rarely did republicans make such a case for themselves. They didn't want to be pitied as victims; they wanted to be regarded as politically astute and purposeful. So they rode on a support base that had a contrary image of them to the one they indulged themselves.

The meeting I had gone to in the Felons club, that night, showed just how reasoned and methodical republicans were in their approach to politics. In a club like this one, which you could only be a full member of if you had served a prison sentence for the republican cause, members of Sinn Féin and the IRA spoke openly, expecting to be answered only by people who were broadly in sympathy with them. Republicans were asserting their commitment to their cause and weighing the political demands of the moment against their need to conserve their resolve. They were going to be pragmatic in their dealings with others, but they were going to stick firmly to basic principles. They were going to form political allegiances to serve their own advantage, but they were not going to let others set the pace of change for them. Nor were they going to accept anything that smacked of defeat.

The main speaker at the meeting was Gerry Kelly, there to present a lecture to the Bobby Sands Discussion Group. Kelly was a leading member of the Sinn Féin and IRA talks team that had been meeting with government officials and ministers for six months now. A former jailbreaker who had been extradited from Amsterdam with Brendan McFarlane, Kelly had a reputation for sitting in on those meetings with the British and saying nothing. People said he was scary.

His lecture that night was like an exposition of holy writ. He took the famous lines from Bobby Sands the hunger striker: 'Everyone, whether republican or not, has his or her own part to play.' They are mundane words, the sort of thing a team coach would use, but Kelly derived from them the authority of the republican movement to work alongside constitutional nationalists in the SDLP. Spoken by Sands, the words sounded as if they were meant to extend campaign inclusiveness to the weak and the weary. They said, You don't have to be a street fighter or a political wizard to take part, because every little bit of effort helps. They said also: This campaign is bigger than the republican struggle, so even those who disagree with us should get involved for the sake of saving lives. We are doing God's work here, and even if you can't measure up to our commitment you should give us some little bit of help. Kelly was using the words to explain something that Sands could never, in his day, have envisaged. At the end of the hunger strikes, with Sands and nine other hunger strikers dead, the IRA prisoners described the SDLP as 'imperialist lickspittles' who had 'occupied their time trying to make political gain by attacking those who did genuinely endeavour to end the issue honourably' (Campbell, McKeown and O'Hagan [eds.], Nor Meekly Serve My Time, 1994). But now, in 1995, Sands's words offered the key to explaining to republicans that working with the SDLP need not amount to a sellout.

Kelly spoke also about the talks with British officials around the demand for a decommissioning of IRA weapons. He said that he would be going back up to Stormont, the talks venue, again the following week. He would be putting on that coat again, he said. That got a laugh. He always wore a trench coat reminiscent of the traditional image of an IRA man, and the joke and its response acknowledged that he and his audience enjoyed the symbolism. There would, he said, be no decommissioning, of any kind. That was his frank promise to the audience.

Several questions were raised from the floor, some of them by people anxious that republicans were diluting their old tradition, and Kelly did his best to assure people that nothing of the kind was happening. I was looking for some sense that republicans were preparing the minds of their followers for the inevitability of a political compromise, but I got no sense of that at all. I knew other journalists there who believed the ceasefire was the end of the IRA's campaign. They believed that the people of the Falls Road and other republican areas were being deceived by the IRA about the scale of the real political prospects before them; that the republicans' new political allies in the SDLP and the Irish government had got it right, but that nobody had had the heart to admit openly that the party was over. As far as I could see, people were being given an unrealistic sense of what was available to them, and would be outraged with those who blocked the way to it, rather than with those who had fostered their high expectations.

Maybe republican leaders had accepted by the time of the 1994 ceasefire that the campaign had run its course, but had found a new vitality and a new value to the violence as the peace process advanced. The bomb at Canary Wharf which in February 1996 ended the ceasefire had far more impact than any bomb before it, and the force of that impact derived from the shattered hopes of millions. No bomb before had ever darkened the hearts of so many people. It was the peace prior to it that had made it so powerful.

There were sceptics in the audience that night, but they were not doubting the will of the IRA to sustain a ceasefire; they were doubting that the current strategy would produce a British withdrawal. None of these sceptics seemed deterred by the fact that Brain Surgeon, the policeman of the ceasefire, who had been named in the Sunday Times, was present at the meeting, drinking with friends at a neighbouring table. Whatever dark thoughts troubled his mind about the journalist a few tables away, many people there felt free to air their doubts, perhaps for the first time, before the republican dignitary on the platform. Was this strategy taking us to a British withdrawal? Could the SDLP be trusted not to ditch us? Would the people be humiliated? If I had been an IRA leader in that hall I might have been angered by all these theorists in a bar, debating whether I should go on risking my neck. Bobby Storey there had just been released from a long sentence, and was still on licence. Eighteen months later he would be arrested for allegedly bombing Thiepval barracks, an attack which brought the IRA campaign back home to Northern Ireland. Was he not wondering why he should be throwing his own life away, let alone so many others', for these critics of the movement? There was some organic connection between these speakers and the IRA, so that republicans sympathised with the fears of speakers from the floor, and speakers from the floor easily assumed the right to require further violence from the IRA, if they thought that was in order.

Not that the audience could be perceived as a strong moral pressure on the IRA to stick to its guns rather than settle terms short of a united Ireland. It was hardly representative of anything more than a sympathetic base. People with opinions strongly divergent from those of the republican movement were not expected to be there, certainly not in significant numbers, though there were often journalists or members of other political groups at meetings like this. Some republicans, like Bobby Storey, presumably thought that critics of the republican movement should be told to stay away. That was why he had me in the corner. What right had I to be there if I was only going to go away and criticise?

Republicans have mostly accepted that they have no power over some people. Their control is geographically circumscribed. What Storey did in attacking me was exceptional behaviour. For instance, once I was recording interviews in Newry near a republican parade against crime in the area. It was a transparent propaganda stunt. Two men who had been ordered out of the town under threat of a kneecapping had taken sanctuary in the cathedral. The republicans were out to show that they held the moral seniority. I moved with a tape recorder among the women supporting the parade, but no one would speak to me because a tall thin man in a brown leather jacket hovered over me wherever I went.

'You are preventing me from doing my job,' I said.
He bent over and yelled at me: 'You are an anti-republican.'
'Since when was that a crime?'

There was no point in trying this on with Bobby Storey. He would have hit me hard if I had answered him back, at least that's what he said he would do, and I believed him. So I stood back against the wall, and held my peace while he told me what rubbish I was. Someone walked past and glanced at me to see what was going on, then either thought better of intervening or trusted that if matters were in Bobby's hands then all was as it should be. Bobby told me that I was taking information about him from the police. I was one of a group of slugs, and he named two others, who were in the pockets of the police. Then he told me he wasn't going to hit me, but he wanted me to know what he thought of me. I soon did.

And I felt small and humiliated. Intimidation works: it makes you timid. I have known this since my schooldays. Anyone who ever said, 'I'll wipe that silly grin off your face, O'Doherty', usually did. Violence awakens the self-accuser. It sets your mind to absolving the abuser and to finding flaws in yourself that explain his behaviour. And in that state of mind - as you do - I asked myself if I was wrong. I don't feel that every turn of my life that directed me away from joining the IRA is to my credit. I have been pulled from my house and kicked down through the garden by a soldier. I saw that soldier standing in our living room screaming back at my mother who was screaming at him, she in her nightie, he in a uniform with a rifle and a blackened face. I had been stupid. I was drunk and blew a whistle out the window to alert the neighbourhood that this man and his men were crawling through our garden. I should have had more sense. He kicked the door in. Why did I not resolve to kill him or his kind afterwards? Was it because I was wiser and could grasp the larger context, or was it just because I had a job to go to the next day? Was it because I had a perverse and squirming obsequious streak in me that would own up to my own awfulness and defer to authority rather than accept the burden of knowing that some people should have their power taken away from them?

After my encounter with Storey, I walked down the Falls Road feeling I had hacked at my own roots, deserted my own people to violence and disgrace, even joined in berating them, all for a failure to remember where I came from and to identify with what they had suffered.

Next day I pulled myself together. Isn't it awesome how a fright will shake your convictions?

The trouble with guns is that there is such a limited number of things you can do with them. In choosing to employ the same tools for their campaign as their predecessors had used, republicans from 1970 onward confined themselves to a political methodology that had only a narrow potential. The challenge for them was to extend that potential. They were creative and successful in this. They discovered that armed protest can provide a lot more than a physical shove against an enemy, and that it can actually be used to close off the options of your enemy - and even those of your political neighbours. It can challenge them to apply their minds to a problem that you want resolved, even when that problem has no similar urgency for them.

By such devices, military weaponry becomes a political tool, but its use brings a huge political cost. Republicans say they have been demonised, though they could not have reasonably expected to be loved when they were killing people and destroying property. They were able, at times, to present a charming facade, to win friends in the media and abroad, to inspire an astonishing sufferance, at times, of their brutality, so that severe critics would accept the proposition that they had little choice; but through the 1980s they reached the limit of their potential to convert voters to their cause, and change insisted itself upon them.

Ultimately, violence was damaging. Once the organisation started pouring energy into extending the influence of Sinn Féin, there emerged a conflict between the method and the aspiration. Killing people would cost votes. That was never more clearly admitted than in April 1997, when the IRA called a de facto local ceasefire to facilitate Sinn Féin's election campaign. In doing that the IRA risked abandoning two valued principles. The first of these was automaticity - to take a word from the American writer on military strategy Schelling - meaning the presentation of violence as a virtual force of nature, uncontrollable in the end by anyone other than those who create the conditions that - you say - create it, in this case the British. The second principle sacrificed by the April 1997 ceasefire was distinction, meaning the contrived distinction between the IRA as the makers of violence and Sinn Féin as the explainers of it.

Strategically, automaticity is the principle that your forces cannot be controlled or bartered with. This makes them much more dangerous. A human force is made as unthinking as a landmine. Its violence has the inevitability of the workings of a machine or the spontaneity of a force of nature. Just as the Russians, during the Cold War, were to be persuaded that all-out nuclear war would inevitably follow any advance of their tanks into West Germany, IRA violence was presented as an inevitable product of political circumstances, which could not be curtailed or reasoned with. Automaticity is something that your enemy has to believe in, even when you don't wholly believe in it yourself. Essentially it is a form of deterrence, though in the case of the IRA campaign, the thing to be deterred has already happened, the British are here, and the violence has to be presented as an inevitable consequence.

Another modern example of automaticity, as a strategically valuable conceit, is the conviction that you die if you mess with the SAS, because they shoot first and ask questions later. That widespread belief, whether true or not, is an asset to the SAS. To waive it, by insisting that the SAS will always observe the strictures of the civil law, would be to limit their effectiveness. The three IRA members shot dead by the SAS in Gibraltar in 1988 were to be regarded as having been killed as if by a natural law as implacable as electricity. The knowledge that the SAS were so ruthless, and would be excused their excesses, would deter further IRA missions a lot more effectively than any suggestion that the SAS were thinking, considerate people who would tailor each ambush to the conditions of the time.

The erosion of limits set by the United States in Iraq since the Gulf War shows up the strategic danger of surrendering automaticity. Your limits can be tested and stretched. Similarly the IRA, in showing itself ready in April 1997 to curtail its own threat for political advantage, was taking such a strategic risk; it was showing that it was potentially amenable to bartering in changed political conditions. It was also compromising the political argument of Sinn Féin, that violence arises from political conditions that it alone has no power to change. The sacrifices in principle made to enable that ceasefire can be taken as indications of the importance of it to the IRA. The latter was not going to risk jeopardising the credibility of its implacable resolve if it was not going to get some advantage in return. Success in the elections was clearly such a valued goal that the IRA was prepared to make costly sacrifices to help achieve it.

Sinn Féin argues that the violence of the IRA is a product of circumstances, and that only the correction of those circumstances can end it. The importance of this principle, strategically, is that it focuses the minds of others on creating political measures to assuage that violence. If the violence can in fact be switched on and off to suit electoral purposes, it clearly isn't generated by political conditions, it clearly isn't the passion of the aggrieved in action; rather, it is the tool of a thinking movement.

The distinction between Sinn Féin and the IRA that is insisted upon by Sinn Féin is compromised because the violence is stopped as a favour to Sinn Féin, presumably at the request of Sinn Féin, and probably with terms and time limits agreed with Sinn Féin. This distinction has been of vital importance to Sinn Féin because throughout the 1 990s the position of the us, British and Irish governments has been that there could be no question of Sinn Féin entering negotiations to secure a political settlement to the conflict without an IRA ceasefire. The Sinn Féin response to this was to insist on its distinction from the IRA and to claim the right to be involved in negotiations purely on the strength of the party's electoral mandate. But if that mandate was won with the help of the IRA, then the plausibility of the distinction between the two movements was lost, and the very mandate that Sinn Féin said won them the right to enter talks was shown to be contaminated.

The election of 1997 showed that the Provisionals had overcome some of the contradictions of the 'long war' strategy, that they really had a new game to play, and that they were playing this new game well. Sinn Féin's share of the vote in Northern Ireland rose from 10 per cent to 16 per cent. Their electoral standing increased dramatically, reflecting the fact that the party had won international respect for its efforts at peace making. Behind this was the real prospect that Sinn Féin would overtake the SDLP, now a party of ageing men. Seamus Mallon had had heart problems, Hume was clearly exhausted and overwrought. The party's only two other MPs, Joe Hendron and Eddie McGrady, were sceptical about the whole peace process anyway.

Hume's endorsement of Adams as a committed peacemaker, as a collaborator in a shared project and as a man who could be trusted and respected could not be withdrawn in time to prevent Adams increasing his vote against Hendron in West Belfast on the back of it. Had the 1997 election resulted, as it might well have done, in Sinn Féin being the leading nationalist party in Northern Ireland, all prospects of a partitionist devolved settlement would have died. Politics would have been reduced to a simple, uncluttered argument over whether the region was British or Irish. This has to be regarded as such a prize for republicans that the entire peace effort may plausibly be assumed to have been directed at achieving it.

The year of elections showed the strongest signs yet of demographic change affecting the balance between parties in Northern Ireland. The academic Brendan O'Leary interpreted this in the Irish Times of 2 July 1997, arguing that unionism had forever lost the prospect of once again dominating Northern Ireland, that the Affiance Party would hold the balance of power in the coming decades, and that the IRA campaign was now immoral and unnecessary because nationalists were being removed from the danger of being dominated by a unionist bloc.

In articles and speeches after the 1994 ceasefire, Tim Pat Coogan had also argued from demographic trends that the war for nationalists was effectively over. This entailed some assumptions that few republicans would be happy with. They know that not all nationalists, as voting cultural Catholics are usually called, actually want a united Ireland. The Derry journalist Frank Curran spoke of this in a 1996 interview with me for the UTV series No Offence. He thinks that Catholics will bide their time and make their decision when they see what the options are. A republican might derive sectarian satisfaction from the humiliation of unionism, but would not see any prospect of a united Ireland while there remained a chance that the middle ground might knit into an allegiance against republicanism.

Garret FitzGerald returned to the demography arguments in the Irish Times in August 1997, concluding that the likely outcome was not nationalist supremacy but realisation of the need for mutual respect between evenly balanced communities. Like O'Leary, FitzGerald saw the demographic trend as pointing towards the politics of reconciliation rather than victory for either ethnic bloc.

In the 1997 elections, constitutional nationalism nearly lost out to republicanism. Through the peace initiative, it had appeared to validate republicanism as a junior partner that was deluded only about means. If republicanism was the youthful energetic wing of nationalism, thought many Catholics, what good reason was there not to vote for it and contribute thereby to the project that John Hume had started? Constitutional nationalism appeared to have no coherent project of its own, other than to put manners on Sinn Féin. John Hume was seen as a heroic, self-sacrificing man who had obviously worked himself to exhaustion for peace. Those who had procrastinated long enough almost to destroy his health - Sinn Féin - were paradoxically regarded as allies of Hume, frustrated only by British intransigence and the unionist veto.

Sinn Féin's standing was further helped by the issue of electoral pacts, which surfaced once more in the run-up to the election. The essential unity of nationalism could be demonstrated by electoral pacts, and those who refused such pacts could be presented as traitors to their tribe. But if nationalism was to be a unified ethnic force, then the lesser part of it was under moral pressure to take the leadership of the larger. Sinn Féin slipped through that challenge by wrong-footing the SDLP into appearing to be the side that was least interested in a pact. Had the IRA not resumed its campaign in February 1996, Sinn Féin would have entered the election of spring 1997 as junior partners in a coherent nationalist ethnic political coalition. But without a ceasefire, the SDLP could not make a pact with them. That turned out to be to Sinn Féin's advantage, for the SDLP could then be presented as the party which had betrayed northern Catholics. The SDLP, it could be argued, would rather let the DUP's Willie McCrea win Mid-Ulster than let Sinn Féin's Martin McGuinness have a free run.

Catholics in Northern Ireland asked unreflectively: If unionists can unite against nationalists, then why can't nationalists unite against unionists? No one had explained to them why politics should not be simply an interethnic contest, least of all John Hume, who had already defined it as such. Ironically, if the day ever comes when constitutional nationalism makes a deal with unionism in the middle ground, it will be the republicans, if they are still in the minority, who will be diverging from the tribe if they reject it.

For two decades up to the 1994 ceasefire, the IRA maintained a level rate of murder and sabotage, as a form of political protest against an internal settlement in Northern Ireland. This may have been driven by simple political absolutism, in which nothing but the whole goal would suffice. It may have been a means of raising the ceiling on compromise, so that, though ultimately republicans might have to do a deal, the final compromise for which they would be able to trade an end to violence would be one that was much closer to their ultimate goal. In the meantime they successfully invalidated the prospects of powersharing or any other internal arrangements. Until they reached the point at which they were ready to trade, the violence served to remind anyone tempted to consider working for such a settlement that there was little point in their trying. Even a political agreement that was endorsed by 80 per cent of the population, if such were possible, would not bring an end to violence.

Republicans would talk us all into a clear linkage between two things that were not necessarily part of each other: constitutional agreement and peace. What satisfied most people in both communities might not be enough to satisfy republicans. Linking agreement and peace together would put all parties under a moral onus to find, not the best compromise between their different positions, but an agreement that could include Sinn Féin. Like children at a party organising games, it would make little difference what most people wanted if the huffy brat who owned the ball couldn't be happy too.

For two decades the veto of violence operated. Experimentation with political initiatives designed to assuage the wounded sense of Irish identity that lay behind the violence stopped with the Anglo-Irish Agreement of 1985. That agreement had been open to three-yearly review, but the reviews did not come. The British government had worked out that the best counter to the republican veto on settlements was to be sparing with initiatives. Then the IRA's violent protest would have nothing to connect with. The production of murder and sabotage would continue with little variation, but would seem increasingly pointless.

There were moments, as on Remembrance Day 1987, when eleven people were slaughtered by a bomb explosion at the cenotaph in Enniskillen, when the campaign seemed blindly barbarous and wholly untenable. It continued, however. What would happen, I once asked Alex Maskey, leader of Sinn Féin in Belfast City Council, if the IRA stopped? 'We would go back to Stormont,' he said. 'It would just be Orange rule again.' But even in a devolved government, I said, the whole balance of parties would be different now. It would be nothing like it was. 'We would see no difference.'

Things might have changed enough for many. They had not changed enough for him, and it was his allies alone who would decide when the line had been crossed at which it was safe to relax the veto. That would be when they were sure there would be no internal settlement at all, of any kind. What prospect was there, then, for a middle-ground settlement appeasing republicans and bringing them on board? Logically, very little.

The consent principle, which had been coined within the Joint Declaration of the British and Irish governments in December 1993, guaranteed the constitution of Northern Ireland against change that was not approved by the greater number of the people there. Republicans called this 'the Unionist veto'. Gerry Adams had said to me in an interview:

I don't think it is up to the British to define self-determination for the Irish people, that's the first thing. I think we can then define it as a whole in any way we want, we can exercise it in any way that we want. I think one of the problems with the declaration is that from the British point of view it addresses the issue which is significant and welcome but then it appears to say 'you have the right to self-determination but ...' and it goes on to qualify it, and my view is that you can't have half a right or a third of a right. You either have the right or not. We of course have that right, that's denied us by external impediment, by British interference; how we exercise that right without internal or external impediment is just a matter of common sense.

The acceptance of the Joint Declaration by the SDLP and the Ulster Unionists was the first major step towards the creation of a middle-ground consensus. In that sense it was a huge problem for republicans. Republicans were getting suggestions from the government, from John Hume, and from David Trimble in small hints, that a political deal was possible. It could, however, only be a deal that knitted together the middle ground in Northern Ireland. Hume and Trimble were also having meetings to discuss the looming crisis at Drumcree through the spring of 1997. Republicans must have understood that talks that proceeded on the basis of the consent principle would have no chance of producing a united Ireland. They had known that since 1993, and rejected the consent principle because it blocked the way to a united Ireland. The fact that in 1994 they called a ceasefire, then, even with the consent principle in place, seemed to hint that they had decided they could in fact live with that principle. It implied that they had recognised and accepted that the only tenable settlement was one in which Hume and Trimble created powersharing between them.

But that is what the armed campaign of the IRA had sought continually to prevent. They were seriously tested on whether this apparent change was real, because the greater stumbling block had been the demand that they decommission, that is, give up, weapons. That demand was endorsed by the report of the international body chaired by Senator George Mitchell, who was later to chair the talks process itself.

In the mid-1990s, people believed in the good intentions of Sinn Féin because they chose to be optimistic. For two decades pessimism had been the unerring guide to Northern Ireland politics. People now wanted to believe in something good happening. Many were overwhelmed by the personal charm and evident conviction of Gerry Adams. Though he often came across as ill-tempered, fumbling and unfocused in media interviews, and always evaded the core questions about the relationship between Sinn Féin and the IRA, at a personal level he won sceptics over. Suddenly they could see why he had to be cagey. Hadn't he to move slowly and bring the whole movement with him? How could he move from a negotiating position before negotiations started and tell us frankly what he would settle for? Wasn't he at risk personally from the hard men of the IRA if we drew him out too far? Excuses were made for Adams on the basis of hopeful conjecture.

The commitment of Gerry Adams to finding peace was endorsed by John Hume, but many others were impressed by Adams too and took him to mean something by his words which he never actually said, which was that the IRA were ready to settle for less than a British withdrawal, so long as it was agreed by the Irish people as a whole, their view expressed through the contrivance of simultaneous referenda North and South.

What is striking about much of what has been written already on the peace initiative of the republicans is that though that initiative is open to very different subjective readings, many writers have chosen to make optimistic readings from the information available. Mallie and McKittrick, for example, point to a number of negative signs in 1994 which shook the faith of many who believed that a ceasefire was coming. They included Father Alex Reid, who had been a channel of communication for republicans, and who began to doubt whether republicans were serious. The spate of attacks on loyalists in the summer of 1994 is cited as one of the negative indicators, and Mallie and McKittrick record a suggestion that the attacks were a stratagem to wrong-foot the loyalists into continuing their violence after an IRA ceasefire. The security forces would then turn on the loyalists and the world would see them as the real evil. That is one plausible reading, but there are others. Might not the attacks just have been an attempt to provoke increased sectarian violence so that a ceasefire option could be waived? On what basis does an observer opt for one reading of events over another?

Similarly the massive attacks on London are treated as possible evidence that the IRA had begun to think it might drive the British out of Ireland by inflicting untenable damage on the British capital, but Mallie and McKittrick simply conclude that the IRA must have dismissed this idea. Why did these writers not consider seriously that the big new bombs were an integral part of the new strategy? They made a choice to believe in the ceasefire in the face of serious evidence against the sincerity of the IRA. In the end their choice presumably came down to subjective impression.

The tendency to interpret the vague language of republicans optimistically was evident again in the days before the IRA's restoration of its ceasefire on 21 July 1997. It can be seen in much of the Irish News's coverage of Sinn Féin statements. For example, a front-page story on 17 July 1997 claimed that Sinn Féin had relaxed its demand for Irish unity. Reporter Michael O'Toole wrote: 'Sinn Féin has given the broadest indication yet that it would be prepared to accept a political agreement which fell short of a united Ireland.' Already, O'Toole is acknowledging that Sinn Féin only ever hints at accommodation, but never makes its position explicit. The hint in this case is a line from a long statement from Gerry Adams, published in an inside page, in which he says that Sinn Féin would press for a 'renegotiation of the union', which O'Toole interprets for us, on his own initiative, as meaning that Sinn Féin will not be insisting on an abandonment of the Union. In fact, the statement by Gerry Adams focuses mostly on nonnegotiable concessions that Sinn Féin is demanding of the British government, concessions that it believes should be granted regardless of the talks process. Only in the final paragraphs does Adams allude to the talks themselves. There he says that Sinn Féin will urge the Irish government to press for Irish unity, and that the party itself will work for 'political, economic and democratic transformation of this island'.

Hyped as a 'marked departure' from 'thirty-two-county' rhetoric, the statement itself is equally amenable to being read as an entrenchment of the Sinn Féin position, as an insistence on major political change outside the context of the talks process, including the disarming of the RUC, progress towards improving conditions for and releasing of political prisoners, and equality of status for the Irish language. (Later, republicans claimed that these optimistic readings of their position amounted to a dirty tricks campaign against them.) The SDLP, by contrast, had been arguing that reform of the RUC would come after a political settlement.

Another example is worth citing of the remarkable optimism of observers of the peace process. Part of the lore of the 1994 peacemaking is that it hinged at the end on whether or not veteran republican Joe Cahill would be given a visa to enter the United States. (The panic to persuade Bill Clinton to issue the visa is outlined by McKittrick and Mallie.) This seems incomprehensible. Would the IRA at the point of decision in August 1994 really have carried on over such a thing? Imagine telling a policeman's wife that her husband had to be shot because Nancy Soderberg opposed giving Joe Cahill a visa. Yet, that is what Jean Kennedy Smith told Soderberg, what Alex Reid told Albert Reynolds, what most observers of the peace process accept. These people believe simultaneously that republicans are peacemakers and also that they would kill for the right of Joe Cahill to a us visa. They appear not to notice the paradox.

One of the problems for those who produce violence but seek to negotiate an end to it is, perhaps, that the focus of the violence shifts from the big aspiration of a united Ireland to the lesser aspirations of steps in their peace process. So republicans have threatened to maintain violence simply to insist on the right not to decommission weapons; but the implied threat that they would continue to kill and risk the lives even of their own members, just to get Joe Cahill a visa, seems implausible.

Those writers who accept such an implausible idea are perhaps going further than they ought in attempting to understand the problems of the republican movement.

The problem, of course, is trying to interpret republican intentions when indications point almost simultaneously in contrary directions. This dual-purpose character to the political agenda of republicans is routine, but was rarely as distinct and puzzling as in July 1997. Then, republican politics seemed to be moving towards war on behalf of the community and also towards political talks with unionism on a compromise settlement. What was amazing about that month was the speed of events, and the fact that the prevailing spirit of the beginning was so comprehensively superseded at the end. Many writers, primarily Conor Cruise O'Brien, had predicted that the republican movement would use street protest to bring about warfare on the streets, to seek out a renewed legitimisation for their campaign. By the end of the first week of July 1997, that assessment seemed well validated, and the calamity of unprecedented civil unrest seemed imminent. Two weeks later, the calamity having been averted, the IRA declared a new ceasefire, on precisely the terms it had previously said it would. Its preconditions had been met, and it kept its word.

Two incompatible tendencies worked themselves out over the same period. The one tended towards chaos, the other towards a ceasefire. The stage at which communal warfare seemed inevitable was when republicans and others were preparing to oppose Orange parades planned for routes that passed through Catholic areas of Belfast, Derry and several small towns. The previous year, the whole of Northern Ireland had been engulfed in protest and riot, after the police blocked the way of an Orange parade from Drumcree towards the Catholic Garvaghy Road. The Orangemen had mustered so much disorder that the police simply had to back down and turn their batons on the Catholic residents of Garvaghy to clear a way for the parade. At the beginning of July 1997 no apparent means existed for averting a repeat. Whatever the decision taken by the government, it seemed inevitable that huge disorder would follow. If the RUC barred the way of the Drumcree parade, as they had in 1996, a similar muster of Orangemen all over Northern Ireland was likely; if it did not, then rioting by Catholics would follow. The Catholics would resist the second contentious parade of the week, through the Ormeau Road and, once humiliated at Garvaghy, would prepare greater resistance. What was most dangerous was that tens of thousands of Orangemen would be marching in Belfast on the same day. There could be a riot on an unprecedented scale.

The question of the contending rights of Orangemen to march through Catholic areas and of residents' groups to stop them was one that had the power to create a renewed validation for the republican cause. Had the scale of trouble predicted actually occurred, then such a validation might have followed. Republicans, however, played the game less ambitiously, apparently simply to prove that a nationalist community veto operated effectively. Just as Protestants mustering in numbers had the power to block change, as demonstrated so many times before, Catholics now showed they had the same power. That is to say, republicans seem to have been content not to push Northern Ireland over the edge for a regeneration of armed struggle, but instead to use street protest to contend civil rights issues. The question is whether that implied an overall lowering of their sights, or a complementary strategy. Would republicans direct their energies into street protest for intermediate goals as an alternative to working for a united Ireland, or would the intermediate struggles simply be a way of inching forward towards the primary goal?

The armed struggle had become tedious over the years, through the eighties and into the nineties, and blocking an internal settlement in itself did not give republicans any tangible achievement to show their followers. Street protest did. And it offered the anomalous image of Gerry Adams on the Andersonstown Road recovering hijacked vehicles and personally returning them to their owners. Whether Adams could have prevented chaos, had the contest over parades run its course with neither side backing down, is very doubtful. He must then be credited with supporting a massive gamble by the residents' groups and their republican supporters, and presumably with having ideas on how to manage the calamity if it occurred.

The potential value of a confrontation over parades was that it could serve the regeneration of republicanism and the legitimation of armed struggle, as similar calamities had done in the past. Republicanism received its fillips from events that appeared to confirm its analysis, or events that raised such emotion among Catholics that the inconsistencies in the republican analysis could be overlooked. Undoubtedly, the early trouble was exacerbated by social and political factors. Frank Curran has pointed to the six months mandatory sentence for rioting as an important one. Hundreds of young men were sent to prison on the basis of evidence from soldiers that had little or no local credibility. This familiarised law-abiding families with the insides of prisons and drew them into an allegiance with other families, including republicans, for whom imprisonment was no disgrace.

There were often such fillips available to republicans. The very frequency of them argues for the unviability of Northern Ireland. The republicans did not always have to rely on their own energies and imagination to produce such crises, though often they did. The chief example was the hunger strikes campaign, another was the peace process itself. Both had the power to project the British and unsympathetic Irish nationalists as the real obstacles to progress. In the same way that the Easter Rising was a revolt against Redmond, and the Hamas campaign is a revolt against Arafat and the PLO, the IRA campaign is a revolt against John Hume. At the end of the hunger strikes, the IRA dismissed the SDLP as 'imperialist lickspittles'.

These upheavals also have the power to call home to the tribe those Catholics who normally back the SDLP. After the RUC forced an Orange parade down the Garvaghy Road in Portadown in July 1997, the SDLP swung round to the radical position of demanding the rerouting of and even a moratorium on parades, whereas before they had wanted negotiation and compromise. The moral tug of an issue like parades radicalises the whole of nationalism, and it is Sinn Féin that benefits. In the quieter times that follow, the SDLP softens its position and is seen to have sold out.

Part of the explanation of the SDLP's change of position over Orange marches lies in the confrontations that took place over the 1996 parade. When the RUC gave way before Orange protest at Drumcree in that summer of 1996 and reversed a decision to halt the Orange parade along the Garvaghy Road, it seemed to many Catholics to have thrown away whatever credibility it had established as an impartial, nonsectarian force. The Orange Order had refused the instruction from the police that they should stay off the Garvaghy Road, where their insistence on marching was leading to increasing disruption of public order by both sides. The Orange Order got its way by force of numbers. Republicans argued that this showed that the police force had not changed fundamentally since 1969. It would still do what the Ulster Unionist Party or the Orange Order told it to do.

The credibility of this argument in the face of the facts was a serious embarrassment to middle-class Catholics who wanted to participate in the Northern Ireland state and live at peace with Protestants. Now it seemed that everything their cousins and siblings in Andersonstown and Ardoyne had said about the police was true after all.

The next round of the quarrel between the Orange Order and the residents of the Garvaghy Road was due in the first week of July 1997. This was anticipated a full year in advance, and the expectation was that the police would either confront the Orange Order once again but this time not give way before protest, however great, or once again they would clear the Garvaghy Road for an Orange parade, in which case the disaffection of Catholics from the police would be complete and irreversible. Suddenly it seemed as if the entire problem of Northern Ireland had coalesced into a single issue of territorial rights on a stretch of road less than a mile long. Republicanism and Orangeism cling to anachronistic mythologies, the one that Ireland must fulfil its destiny as a single jurisdiction, the other that democracy can only be assured by a British Protestant monarch forestalling the encroachment of Vatican authority. Both are evident nonsense when phrased so simply, but the contention was between people of the different traditions insisting on rights, and the matter was serious enough to threaten violence over the whole region.

The Secretary of State for Northern Ireland Mo Mowlam said that the government would work to find agreement between the residents of the Garvaghy Road and the Orange Order. It failed to do this, and its integrity in the project came under scrutiny when a document leaked after the 1997 parade was allowed through showed that the British government had been seeking all along to find some way of getting 'Orange feet onto the Garvaghy Road'. What had not been clear was that the government would actually decline to make a decision on the merits of the contending claims of the Orange Order and the residents' group. That was to be the job of a parades commission which had not yet been legally authorised to make such rulings. In the meantime, the decision would fall to the police, and they would make it on the basis of public order considerations. If opposing the Orangemen was likely to produce greater violence than clearing the protesters off the road, then the Orangemen would not be opposed. Put simply, the police were given the freedom to choose the easier of the two contenders to deal with, without having to consider the rights and wrongs of the dispute.

So they assembled behind riot shields on the Garvaghy Road in the early hours of Sunday, 6 July 1997, ploughed the protesters off the road, and sealed the residents into side streets to make way for a thick phalanx of trudging Orangemen. The RUC came in new riot gear that included balaclavas worn behind visors. They moved along the road, shoving protesters forward. As angry men from the area lunged themselves at the police shields, some were batoned about the head.

The immediate result of this was a consolidation of Catholic opinion against the decision. John Hume, who had worked for compromise on previous contested parades, and on this one, now argued that the right of Orangemen to march a contested road was not equivalent to the right of residents not to be imposed upon, because the greater physical inconvenience was the one facing the residents. Orangemen compelled to walk a road parallel to the one of their choice were hardly suffering the same inconvenience as residents forbidden access to the main road past the end of their street and forced to hold mass in a side street for want of being free to cross the road to church.

The case for compromise had been based on a model of the opposing groups' rights as competing and equal. It failed when it became clear that the demands of the residents and the Orangemen were wholly exclusive of each other and not equally balanced. The residents seemed to have a greater right, because they suffered the greater imposition for the sake of the rights of the Orangemen.

The government's decision not to adjudicate on the points of principle separating the Garvaghy Road residents and the Orange Order left open the question of which party had the greater right in the issue. A problem with the 1997 decision to let the Orangemen through is that it could only work once. If the default from making an agreement was always going to be public order, then each side could make its own calculation on which way things would go if they declined to agree. The one most likely to get its way would have no incentive to make an agreement. The one least likely to get its way could prepare itself to reap what political advantages were available.

The Orangemen knew after Garvaghy Road 1997 that they had no incentive to negotiate over the Ormeau Road because, for want of agreement, the public order consideration prevailed, and since they posed the greater threat, they would get their way, if they insisted on it. Mowlam's insistence that a decision on the Ormeau Road Twelfth march be made either by agreement or on public order grounds seemed very risky. With the Twelfth approaching and the danger of the whole Belfast parade mustering at the lower Ormeau in support of the local parade if it was stopped, there was a real danger of unprecedented street violence. Faced with that prospect, the Orange Order withdrew its Ormeau Road parade and three others in contended areas. Northern Ireland had been saved from its greatest threat not by government but by the Orange Order, which had realised that the cost of standing its ground might be a calamity for Northern Ireland.

The government had chosen to work for agreement, and for want of it to fall back on the public order principle. If the competing rights of Orangemen and nationalists weren't equal, however, this fall-back principle produced the risk of inherent injustice. If the state was unjust, then protest against the state was warranted, and arguably, force against the state was legitimate. That had been the inference drawn by republicans, and thousands who flocked to join them, in 1969. The graffiti and placards that followed Drumcree 1996 had said, 'Back to '69'. The argument from republicans was that nothing had changed since then. With such a vivid demonstration of the division in society, it was more plausible than ever that the violence of the IRA was a symptom of a fundamental malaise. Perhaps we had even been distracted from that malaise by the horror of the campaign. Perhaps we had concentrated too much after all on condemning that campaign and had overlooked the gravity of the underlying problem. What seemed implied by the confrontations over parades was that Northern Ireland was a society that would be tearing itself apart anyway, even if the bombers and gunmen stayed at home. It wasn't the intransigence of the IRA that threatened to take us all into a civil war, but old-fashioned territorial sectarianism.

There were some who argued that the issues were the same, that it was the IRA who directed the residents' groups. I had seen an indication of that myself at a residents' meeting in the lower Ormeau area in 1995, when a rousing mob of men at the back of the hall cheered for every speaker in favour of confrontation and held silence after every speaker against it. Gerard Rice, who chaired the meeting, refused to allow a vote and went out to tell the media that there had been a unanimous consensus in favour of blocking the next day's Orange parade. Two years later, in areas where compromise was reached with Orangemen, the local nationalists and clergy were often clear that the deciding factor was the low level of involvement of Sinn Féin. These were areas where the marching lodge was made up of people who lived in or close to the streets they wished to march through, who shopped in the same shops that the Catholics of those streets shopped in. Charles Kenwell is a storekeeper in Dromore, County Tyrone. In 1997 he met with a group of residents convened by the Church of Ireland and Catholic clergy of the town. The group was representative of all political parties and several other organisations. The Sinn Féin presence at the meeting was low, but proportionate. Its objections to the parade passing through Church Street on the Twelfth were overruled by the other Catholics present.

The emergence of the parades issue in the mid-1990s was fired by a number of things. Orangemen marching down the Ormeau Road in Belfast in July 1992 had horrified local residents by taunting them with five-finger salutes while walking past the local bookmaker's. Five Catholics had been shot dead there by loyalist paramilitaries in February 1992. In Portadown, where the parades were frequently accompanied by rioting, local Jesuits had organised protests that stopped short of blocking the road. Every year through the mid-eighties and into the nineties, protesters would have a tea party in the middle of the road before the parade came along, and they would consent to being carried prone to the footpath by policemen.

At the Ormeau Road, in Portadown and in Derry as well, and in other areas too, protesters changed tactics and organised to block roads and force a choice on the police between redirecting the parades or clearing the roads by force. There are several indications that this was a strategy devised by Sinn Féin. The three residents' groups involved were all led by former republican prisoners. Yet, Sinn Féin had not always had a strong party line on parades. A Social Attitudes survey in 1990 had asked people if they thought that Protestant parades were too closely or too loosely policed. There was sufficient diversity of opinion among Sinn Féin voters to suggest that they were free to make up their own minds on the issue. Twenty-seven per cent of Sinn Féin voters canvassed said that controls on Protestant parades were used too little, while eighteen per cent said they were used too much. There was clearly no party line on the question at that time. It seems reasonable, therefore, to speculate that a policy decision was made within the party some time after 1990.

The argument that protest groups were directed by Sinn Féin was discounted for practical purposes by the Secretary of State, Mo Mowlam. She chose to deal with the residents' groups as if they were wholly representative of the areas from which they came. Sinn Féin now represented more than 16 per cent of the electorate and had very substantial majorities in some areas. The Orange decision not to negotiate with people who were members of Sinn Féin or former prisoners could not hold up against democratic standards when this amounted, as it did in some areas, to refusing to deal with anyone but representatives of minorities. It also seemed inconsistent with the fact that unionists had to deal with Sinn Féin councillors in local government.

Yet the suspicion of the Orangemen was that they were being forced to deal with Sinn Féin directly, working to its own agenda, and not just with people who happened to be members of it. They suspected that those members of Sinn Féin who were leading residents' groups were taking their guidance not from the other members and their communities but from the party itself. The Orange Order and other sceptics were sure that the opposition to parades was part of a larger conspiracy. The focus of accusations of conspiracy at Garvaghy Road was Breandán Mac Cionnaith, who was seen as a republican with a record of violence. He had aligned himself with similar protests in Derry and Belfast. His partner in the coalition at Garvaghy Road was the Jesuit priest Father Eamon Stack. Father Stack defined the role of the protest not as a challenge to the Orange Order, but as a challenge to the state to treat Protestants and Catholics with equal respect. He believed that if he could hold the coalition together, in the face of criticism, that would create the best chance of a peaceful end to the crisis. He said that the coalition had decided to offer such slight resistance to the parade from Drumcree church that, should the police choose to force the march through, they would be able to sweep the coalition's resistance aside. The point would then have been illustrated dramatically that the police favoured the Orangemen, but no one would have been hurt.

Gerry Adams appears to have endorsed the conspiracy theory of parades protest himself in words attributed to him in an edition of the RTE current affairs programme Prime Time broadcast in spring 1997. Adams had been addressing a meeting of republicans in a closed session in Athboy, in the Republic, the previous autumn and had challenged those present, 'Do you think Drumcree happened by accident?' Adams said that three years of hard work by activists had produced the massive stand-off at Drumcree in 1996, which had resulted in the reversal of a decision by the RUC, from confronting the Orangemen to beating Catholics off the road to escort them through. Adams said that these were the sort of 'scene changes' (whatever that means) that the party had to exploit. Privately some members of the Garvaghy Road residents' coalition were furious. 'That's just Adams shooting his mouth off,' said one. In fact, Adams's words do suggest someone asserting an implausible point, speaking to people inclined not to believe him, as if he was trying to persuade them that the peace process and street politics were producing results.

However conspiratorial the origins of some of the residents' groups, the issue they raised attracted greater support as the tensions between the two sides increased and as the policing of the disputes became more violent. This echoed the period of the civil rights protests in the late sixties and the increasing disaffection of Catholics with the police at that time. The simple law such events followed seemed to be: the worse things get the worse they get. Violence simply produced more violence in an irresistible spiral. Heavy- handed policing, however necessary in terms of the conditions of the day, would always annoy people sufficiently to ensure that the next confrontation would be harder to deal with.

The insistence of Orangemen on marching through Catholic areas despite the opposition of residents made it appear as if it was the right to annoy those residents that was most important to them. When police action to remove protesters led to people being injured by plastic bullets and batons, the simple impression made was that the state would deploy force to preserve the rights of Orangemen over the rights of Catholics. It wasn't that simple. In fact, the police stopped nearly all parades going through the lower Ormeau Road. They took their guidance, however, from public order legislation, which required them to make their decision on a parade on the basis of which course would lead to least violence. That meant that the larger parades along the Ormeau, on the Twelfth of July each year, were forced through. The message of that to Catholics was that when Orangemen mustered or threatened to muster, they got their way.

This process was familiar to any observer of the Northern Ireland Troubles. An issue acquired legitimacy and support as it was repressed. The anger that an issue can mobilise at first may be very little, but once heads have been cracked on the streets the same issue can rally much stronger commitment. Right to the end, there would be Catholics who would not go along with this plan to block roads against Orange parades. On RTE radio, on the Friday morning before the 1997 Drumcree parade, Catholics interviewed in the centre of Portadown said that they thought that the parade should go ahead, and that the people of the road should simply go into their houses and ignore it. Others could see that if the issue was brought to the point of confrontation, attitudes would harden on both sides, and forces on both sides would then be easier to rally and harder to deal with.

The government was in a genuine double bind which it genuinely wanted out of, but neither the Garvaghy Road residents' coalition nor the Orange Order would concede significant ground. A long delay in announcing a decision on the parade allowed both sides a chance to relax their positions, right up to the very end, and neither took it. This was foreseeable. They had made their contrary positions clear well in advance. Equally foreseeable were the ways in which the images of the day would be read. To Catholics the sight of police in riot gear with shields shoving protesters off the road would mean simply that the Orange Order still had the power to summon the full weight of the state to its aid. Mowlam said the decision was dictated by circumstances, and in the interests of public safety. Ronnie Flanagan, the police chief constable, said it was clear that violence would have been directed against Catholics if the parade had not been allowed to go ahead.

One of the Jesuits at Garvaghy Road said that the inevitable result of the decision to let the parade go through would be the humiliation and disillusionment of the people of the area. It would be much harder now to persuade them that a peaceful way forward was possible. The miscalculation of the Jesuits had been that a peaceful way of stopping Orange parades through their area was possible. It wasn't.

A conspiratorial reading of events would say that republicans inspired this confrontation and reaped an affirmation from it of their assessment that peaceful politics would not bring justice for Catholics. A more sympathetic reading would say that republicans had been trying to prove that peaceful protest could bring change, or at least that they were conducting an experiment to see if it would. Either way, the conclusion would have been the same, that peaceful protest had failed and that Catholics' rights were expendable for a bit of peace and quiet. In which case the republican response would inevitably be: Well, we'll see about that.

The republicans did not invent the division between Catholics and the Orange Order, and they had stayed out of this quarrel for decades. Catholics had learned to live with Orange parades. That may have entailed a denial of their own sense of intimidation. They always knew, under their skin, that too much trouble would follow the disrupting of these parades for it to be worthwhile. They sensed that the Orange Order was too big and too strong. I remember as a child watching an Orange parade through Belfast city centre. Two young soldiers stood near me, unarmed but in uniform. A man in a black suit broke from the parade to reprimand the soldiers. He ordered them to stand to attention, and they did. The Orange Order had seemed to be in charge of more than its own members.

According to IRA defector Sean O'Callaghan, the strategy of the IRA was to bring loyalists into confrontation with the state, or at least to wrong-foot them into being the greater offenders, and perhaps to demolish the legitimacy of the demand that republicans sign up to the Mitchell Principles committing them to peaceful negotiation (by showing that unionists flouted them too, but were kept in the talks because of their numbers). If that's what republicans were up to, then that strategy was effectively countered by the decision to let the Drumcree parade go ahead: paradoxically, the effect of favouring the Orange Order was to put an onus on unionists to behave themselves and to accept later restraints. By winning, the Orange Order was losing. As soon as the members of the Orange Order set one foot on the Garvaghy Road it owed Mo Mowlam a very big favour.

The Orangemen got their parade down the Garvaghy Road in 1997. They decided days later to waive four other parades which till then had been just as important to them, including their parade down the Ormeau Road on l2 July. They took great credit for that decision, though many complained that they had taken it only because they were threatened by the mustering Catholics on the other side of the Ormeau Bridge. The police had told them that they could not provide enough protection for the parades. The Orange Order had had one very strong hand to play, but only one, and the trouble was that it was too strong. It could redirect the main Belfast parade of nearly thirty thousand people to the Ormeau Road, to demand the right to walk through. There were strong hints that this is precisely what it intended to do. The police advised Orange leaders that loss of life on a large scale would be unavoidable if they did that.

In holding firm to its insistence on marching the Garvaghy Road and winning the government's cooperation, the Orange Order had thrown away a potential moral advantage. Afterwards, it could be said against it that it had accepted a victory won by a massive threat against the state and that it had given nationalists an excuse for violence. At the same time it had discarded the opportunity to set an example of tolerance. This was not a good starting point for unionist entry into talks on the future of Northern Ireland. A conspiracy against the Orange Order might have just failed, but the real untenability of the Orange position remained clear, perhaps clearer. Furthermore, it was not the Orange Order that had confronted the supposed republican conspiracy to foment disorder and defeated it, but government and the RUC. This was a bit like a rich father bailing out a delinquent son; he would be seen abroad as the brat's chief ally, but when the boy was safely home, the real price of saving his hide would be made clear.

The strategic way for republicans to have responded to circumstances in which the Orange Order had discarded a moral point would have been to restrain all violence from their own side. A government in disapproving form with the Orange Order, and owed a favour, would have perhaps have had more time for peaceful republicans. Adams addressed a crowd on the Falls Road on the afternoon that the Garvaghy Road march was forced through. He said that, as someone who had lived all his life on the Falls Road, he saw no reason why the people of the area should suffer in the protest, that there should be no violence that wasn't 'consistent with a coherent strategy'. Republicans had clearly anticipated well how things would go and had prepared their response. It appears to have depended to the end on the Orangemen being foolish enough to try to enforce their demands, for when the Orangemen instead backed down, the republicans seemed wrong-footed and unready to articulate a response.

The lessons of the Garvaghy Road crisis were revolutionary. The Orange Order found that its only weapon when threatened was too big to be usable. It had a nuclear option, in that it could mobilise the whole Belfast parade against the lower Ormeau, or it could rally thousands at Drumcree if obstructed there. The potential of either move was too calamitous for the tactic to be permissible. It was faced with a strategic problem that the United States would understand; it can obliterate all life on the planet, if it chooses to, but it couldn't restore order in Somalia.

Republicans had been demanding parity of esteem as the central plank of the new republican politics, but they seemed ill-equipped to recognise it when they won it. The republican complaint was that the unionists operated a veto on change, but the stand-down of the Orange Order indicated that a nationalist veto functioned too. The point had already been made by Mowlam (and the Conservative Secretary of State Sir Patrick Mayhew before her) that political change would require the consent of both communities. The contention over parades had provided a practical illustration that when Protestants insisted on having their way they inevitably set limits to their own freedom.

There was another sense in which the dispute illustrated the potential of republican protest. The implications of this were horrific, but arose directly from the nature of the decision taken over the Garvaghy Road parade. The RUC chief constable, Ronnie Flanagan, said that he made his decision on the basis of the relative threats from each side. That meant that if nationalists could provide a counter-threat greater than that of Orangeism they would, logically, get their way. The terms on which the decision was taken presented the IRA with the first opportunity for it to meaningfully provide tactical defence for a Catholic community. Drumcree presented the IRA with a possible new way of understanding Catholic defence; it need no longer be strictly the mythical defence of Catholic lives against attack, but the defence of Catholic interest by threat. There was now a practical means of deploying the threat of violence in the interests of the Catholic community. A new legacy was available to them to replace the legacy and lessons of August 1969 if that was what they wanted to make of it. The implied message of Flanagan's decision was that the IRA or others in the nationalist community could avert the danger of future impositions by presenting an even greater threat than that posed by Orangeism. The very criteria deployed by the police were an invitation to both sides to compete to provide the greater threat. It hardly seems creditable that a modern government would give insurgents such a message, but the implications of the decision taken at Garvaghy Road in 1997 were as simple and obvious as that.

In 1996 the threat posed by nationalists, and weighed in the balance against the Orange Order, was that the IRA campaign, already resumed in England, would extend to Northern Ireland. By 1997 Gerry Adams was saying that prospects of a ceasefire might hinge on the Garvaghy Road decision. To observers, this seemed to have little plausibility, because it was the sort of thing that was routinely said. The IRA was not going to change its plans over a parade, any more than it had been willing to over, say, the release of Lee Clegg, the paratrooper convicted of the murder of Belfast joyrider Karen Reilly. Outrage would produce riots and destruction, but there would always be a clear distinction between the IRA's agenda and the enactment of community anger through rioting on the streets. The threat of a resumption of violence by the IRA was far greater than the threat from Orangeism, because it offered years of violence against a week of it. By 1997 the IRA was assumed to have committed itself to a resumed campaign, and few believed that it might be persuaded to change its ways by a decision in favour of nationalists on the Garvaghy Road. Those who understood the background machinations for a ceasefire, and how close they were to success, understood the IRA well enough to know that they would make their decision on the basis of the demands they had already made, and not in response to passing crises.

The potential was clear for the IRA to use its violence to enforce occasional demands in the Catholic interest rather than as a sustained campaign against constitutional compromise. If it wanted to bully the government into interim concessions, it could do that. A campaign of violence in defence of Catholic interests would be very different from the long war against an internal settlement. It would have to be occasional rather than consistent. It would respond to conditions as they changed, rather than attack randomly like a force of nature. If violence was pegged to a single irresolvable condition like the British presence, it would have to be seen as nonnegotiable on any other terms. If it was to press for Catholic interests, it would have to present itself as negotiable. The threat, outweighing the Orange threat, would have to be seen as dependent on the decision to be taken, so that the state could know that when it favoured the Catholic interest the threat would be withdrawn. If violence was going to follow, whatever the state did, there would be no incentive to bargain with it.

The instrument of Catholic defence of this kind has usually been rioting and destruction, presenting itself as communal outrage. Paradoxically it does more to damage Catholic communities. This is how it has always been. Rioting on the streets followed the release of Lee Clegg, as it followed the Garvaghy Road decision. The core campaign of the IRA seemed not to be responding directly to these things. (An old example of force used creatively by the IRA in the defence of interests was the killing of prison officers during the 1980-81 hunger strikes campaign, which carried the message that if prison officers ceased applying harsh measures inside the prisons they would be spared. Strategically in IRA terms, it was reformist rather than revolutionary.)

Sinn Féin's campaign against Orange parades was arguably part of a new strategy, away from armed struggle towards more constructive campaigning for the Catholic interest through social action. Those who dismissed the parades protests as simply a republican plot were not only overlooking the real groundswell of support gathering round the issue, but were failing to consider what the nature of the plot itself might be. Had Mo Mowlam banned the Drumcree parade at the instigation of nationalists, she would have signalled to them that there was a potential for progress through social action rather than violence. That in itself would not of course have been a satisfactory reason for banning the parade, but it would have been no worse than the reason for which it was allowed to go ahead.

One legacy of the decision to force the parade through the Garvaghy Road was that the republicans found an opportunity to show political maturity. Gerry Adams, who had three weeks before been slithering out of commenting meaningfully on the IRA murder of two policemen in Lurgan, was now calling for restraint. And the SDLP and Sinn Féin were once again sharing a common purpose, only Sinn Féin seemed the party more able to pursue it. Republicans now had a civil rights agenda and were plausible champions of a people that the state was content to walk over when it suited their practical interests. If there was a lesson in all this for republicans it was that they need never fear that the government would deny them for all time the opportunity to present themselves as the vigorous champions of Catholic rights. There would always be something to fight for, and there would always be popular support to mobilise. They would continue to grow as a party while they had such issues to fight.

The experiences of the marching season point to the existence of a nationalist community veto. They showed that whereas the Orange Order had power to dictate its will, that power was cumbersome and unusable, and they showed that republicans, when they campaigned on community issues, could rally increased political support and draw in support from the SDLP. There was little indication at the end of the crisis that republicans were impressed with these discoveries, or that they would rest their future political tactics on them to the exclusion of the campaign for a British withdrawal. Yet a ceasefire was only days away. Just a month before the Twelfth, it had seemed obvious that the Provisionals were intent on war.

It was one of the sunniest days of the year. Two policemen, Roland Graham and David Johnston, walked back towards their station along Church Walk in Lurgan. Two or three men, said to have been wearing women's wigs, ran up behind them and shot them both in the head. They slumped dead together onto the pavement. For six hours their bodies lay where they had fallen, shielded from public view by a furniture van. Forensic staff came and photographed them and gathered evidence. Reporters a few yards away waited for statements from politicians and a senior police officer. Children skited about excitedly and asked anyone who would pay attention to them if they had seen the bodies at all. A hot but ordinary day dragged on.

There is a strange contrast between the atmosphere at a scene like that and the condensed version of it reproduced by the media. You almost don't believe that you were there, when you see it all again on television. That's because it didn't feel so urgent at the time. The woman police officer, who was probably a friend of the dead men, stands with a clipboard at the corner of the street, directing passers-by away from the bodies, taking abuse from some of them for her civility. The reporters are anxious at first to get the details, then relax and wait, then get bored, then begin to exchange jokes with each other. They are to go away and tell the world how significant this moment was, in which two men died, but already it is routine to them.

David McKittrick, writing in the Independent the next day (17 June 1997), said: 'the killings have confounded the analysis of almost everyone, including both the opponents of republicans and their sympathisers'. McKittrick was one of the optimists. His analysis, as developed in the book on the peace process, A Fight for Peace, that he co-authored with Eamonn Mallie, was built around a belief that the IRA was working out a wholly unarmed political approach. The IRA document outlining the rationale for such a strategy was called the 'TUAS' document. This was being read to mean Totally Unarmed Strategy. Sceptics, after the ending of the first ceasefire of the peace process, in February 1996, were interpreting it to mean Tactical Use of Armed Struggle, which seemed to fit better with both the acronym and the facts. The murders in Lurgan, however, seemed to fit neither notion of what republicans were up to. Certainly this was not unarmed politics, but in what sense was it a tactical deployment of violence? That theory suggested that republicans would work their way into a political process and try to manipulate it in stages. They would use salami strategy, and wear away at the stipulations laid down on their involvement. This double murder seemed more like a reckless act of murderous passion which could only provoke censure and retaliation. The British government was already speaking, through officials, to Sinn Féin, and was already meeting the Sinn Féin terms for a ceasefire. Just three days before the shooting, these officials had told Martin McGuinness that the government had conceded the core Sinn Féin demand for a time frame for talks. The murders didn't seem to fit with any theory that attributed any kind of new refinement of strategy to the IRA. Why hit out at the precise time at which your enemy is giving way to your demands, and make it harder for him to concede further without looking like a weakling?

Constables Graham and Johnston knew, as we all knew, that there was no IRA ceasefire in place. The Provisionals had suspended their violence for the elections but had signalled its resumption when they planted a van bomb on a mountain road near Poleglass two weeks before the Lurgan shooting. Most of the media took up the IRA's own language and called that bomb a land mine, though it was no different from any other booby trap of its kind. The IRA said it had been abandoned because of civilian activity in the area. They would leave it to the police to decommission it. The police stayed away in case the message was simply a ruse to draw them in. Children set alight to the bomb and it flared like a giant roman candle. Sinn Féin condemned the police negligence. The day before, they had been picketing the same stretch of road because of the risk to children from traffic. Now they had no comment to make on the IRA's decision to leave a bomb where children played. In Sinn Féin's eyes, the only people to blame for the danger those children had been subjected to were the people whom the bomb had been designed to kill.

In Derry the previous week, an IRA sniper had opened fire on an army car. The IRA claimed to have hit someone inside it. The police and army said no one had been injured. Rumours said an inexperienced gunman, or gunwoman, had lost balance after a premature first shot and had not recovered in time to take accurate aim. The intention had been clear, however, and constables Graham and Johnston knew well that they were at risk.

There were grounds for optimism, however. Anyone reading the politics of the moment would have thought this an unlikely time for the IRA to strike. It was mid-June. The inter-party talks had resumed at Stormont. Tony Blair was new to government with a massive majority and had declared his commitment to getting a solution to the Northern Ireland problem. The Secretary of State, Mo Mowlam, as well as conceding the republicans' demand for a time frame for talks, had said that the train would leave the station, with or without republicans on board. She had even said that if they were on that train it might leave without unionists, if they couldn't stomach talking with Sinn Féin. Fianna Fáil leader Bertie Ahern had just secured his place as taoiseach in the new Dáil. Everything was in place, or was close to being in place, for a more sympathetic handling of republicans. It was decision time for the IRA. Most of what their new political allies and their declared strategy had sought for them had been achieved, or been brought within reach.

Republicans had a problem with Blair, but it was one of interpretation. Blair had come over to Belfast and said he did not envisage a united Ireland within the lifetime of anyone in the crowd he addressed. He was declaring his personal reading of the realities of Northern Ireland, and underlining the consent principle contained in the Downing Street Declaration, the joint statement by the British prime minister and the taoiseach in December 1993 which laid down the basic principles of the peace process. Republicans put an ominously negative spin on this, interpreting the remarks as a declaration of British government policy to rule in occupied Ireland for another ninety years. Had they been disposed to giving Blair space to be creative they would not, presumably, have misrepresented his words so eagerly.

The killings of the constables closed down a series of meetings between Sinn Féin leaders and British government officials. Blair read the shootings as a rejection. Like the old sages of the Upanishads rejecting definitions of reality with Not this, Not this, republicans were still spurning the most imaginative efforts of governments and political leaders to contrive a dignified course into negotiations for them. Shooting Roland Graham and David Johnston appeared calculated to abort the whole peace process and to close off the prospects of Sinn Féin getting onto the talks 'train'. It brought into question the commitment of republicans to peace making in any terms at all, other than victory. It appeared to have been just a ham-fisted continuation of the war, with no specific political significance other than to remind the government and everyone else that there was no ceasefire in place yet, perhaps - if the thing was thought out at all - so that it would be all the more valued when it came.

The shootings suggested that armed politics was still being used in the same way it had been used through the 1980s, not to facilitate a settlement but to prevent one. The Lurgan killings seemed to signal not that something had changed, but that, as yet, nothing had changed. The republican explanation was that these things happen when there is a conflict in place, as if by some process of subterranean effervescence that only they understood. The rest of the country might be in shock and the United States support base might be outraged, but this wasn't to be seen as politics, they implied, but as chemistry. When republicans aren't getting their way, these things happen, and there isn't much that can be done about it other than for government and others to go and work harder to find a way of meeting republican needs.

Shock has a limited life span in Northern Ireland. Within two weeks, government officials were taking phone calls from Sinn Féin again and writing position papers for them. Business had resumed.

After the Lurgan shootings, it began to appear as if republicans could not be wooed into talks. In fact, a ceasefire would come immediately after reservations expressed by Sinn Féin had been cleared up to their satisfaction. The outstanding obstacle to talks, so far as the IRA was concerned, was the demand that weapons be decommissioned in advance. The government was still working on ways to assure republicans that this demand would not actually obstruct the course of the talks themselves. Republicans had succeeded in persuading many political figures that the demand for decommissioning of any weaponry at all before a final settlement could not be met. Yet to proceed with an armed campaign, in opposition to this demand for decommissioning, would smack of a bloody-minded insistence on fighting for nothing more than the right to retain those weapons. That is what was happening. Republicans were now being criticised by friends like Bernadette McAliskey and former allies like Ruairí Ó Brádaigh for risking and taking life merely to secure a negotiating position. They didn't want, however, to be sucked into a prolonged talks process that would smother them and get them nothing. Yet it was hard to see how any talks process would get them any more for nationalists than the SDLP could secure alone.

When a ceasefire was called a month after the murders, what was most surprising about it was its integrity. There had been a clear list of demands put to government since the previous October. John Hume had carried them himself to John Major. He said at the time: 'I put peace on the table and they rejected it.' Tony Blair met those demands, and the republicans kept their word.

There were other pressures on them, but these need not be regarded as having been decisive. After the shootings in Lurgan, many, including the us friends of republicans, began to doubt whether republicans were serious or not. Congressman Peter King, a personal friend of Gerry Adams, for the first time called for a urn-lateral ceasefire by the IRA. Editorials in the New York newspaper the Irish Voice declared total bewilderment, and discussions on the republican bulletin boards on the Internet reflected the same confusion and distaste. Republican supporters across the world who normally spent several hours a day by their computers ready to defend the movement against criticism simply disappeared for that week.

The timing of the declaration of a ceasefire, when it came, seemed calculated to damage the Ulster Unionist Party leader David Trimble and the whole unionist family, but it is not like republicans to move so nimbly. The 'long war' is a lumbering process. The talks process was approaching a vote on government proposals for the decommissioning of weapons, which Trimble and others saw as a waiving of the issue. He was threatening to walk out of the talks if the proposals were not firmed up to insist on the actual physical handing over of weapons, as a precondition to Sinn Féin staying in talks. Trimble was not going to get his way, but he was in danger of taking the blame now for scuppering the talks at the very moment at which the IRA had made the process look most hopeful. What argued against the view that the IRA had simply called an opportunistic ceasefire, in order to wind Trimble, however, was that it came within the government's timetable. Any later than the end of July, and Sinn Féin would not qualify for entry into talks. It came soon after the government's final reassurance that the decommissioning question would not hold up those talks, and it allowed Sinn Féin to influence the debate on the decommissioning proposals.

Some commentators, like John Waters in the Irish Times, remarked on the significance of the ceasefire starting 526 days after the Canary Wharf bomb, which had come 526 days after the start of the first ceasefire. The coincidence seemed to be a message from the IRA, perhaps a declaration that they pick their timings with care, a hint that they will be dictated to by symmetry before they'll be dictated to by governments. Who knows? It allows people to believe that the IRA are working to a recondite agenda. Maybe they are.

The fact is that they were getting what they had asked for. Meaningful symmetry and the chance to throw unionism into confusion seem simply to have been an added bonus. After the ceasefire declaration, it was possible to scrutinise the documents sent to Sinn Féin before and after the Lurgan killings and to see the development of the government's position. It was after Lurgan that the decision was made by the British government not to let the decommissioning question hold up talks.

A letter, or aide-mémoire, sent three days before the Lurgan murders outlined the time frame and also addressed decommissioning:

The government has always made it clear that it wants to resolve this rapidly to the satisfaction of the participants so that it does not block the substantive political negotiations. Realistically, this can only be on the basis of implementing all aspects of the Mitchell report.

What was missing from the aide-mémoire was a clear promise that the decommissioning issue would not hold up talks. That was what constables Graham and Johnston died for, and it was granted to Sinn Féin in a further letter, from Mo Mowlam to Martin McGuinness, on 9 July. Dr Mowlam's letter quoted and beefed out assurances that the two governments had made within days after the Lurgan murders: 'Both the British and Irish Governments share the view that ..... voluntary and mutual decommissioning can be achieved only in the context of progress in comprehensive and inclusive political negotiations". It is in this context that both governments "... acknowledge a particular responsibility to carry the process forward with energy and determination so as to build confidence without blocking the negotitations".' Clearly if decommissioning could only happen in the context of progress, it could not come at the start of the talks. That was the promise the republicans were waiting for. If unionists were demanding that the IRA surrender weapons, an Independent Commission would take up their complaint and seek a response to it, but the talks would not stop. Republicans might come under serious pressure to begin disarming during the talks, but their absolute refusal to do so would not stall progress towards completion of the talks process.

The point that the question of decommissioning would not hold up talks was reinforced by a reassurance from Mo Mowlam that if the parties failed to agree within a time frame, the two governments would continue to work for agreement anyway. This meant, effectively, that if the unionists wanted to waste the period of the talks by demanding IRA guns, they would simply squander their prospects of influencing the terms of the referendum to be put to the people of Northern Ireland by the government. The crucial words were:

[The government] cannot give a guarantee of a successful outcome because that will require agreement and consent among both unionists and nationalists, as well as both Governments. There are many difficult issues, any of which if not addressed in good faith and resolved satisfactorily, could hold back overall agreement. But both Governments will be working to overcome obstacles to agreement and, if these negotiations do not succeed despite their best efforts, they will together continue to pursue rapid progress to an overall agreed settlement acceptable to both unionists and nationalists.

The irony of the republican position when the prospects of entry into talks arose was that the very guns that had been used to prevent an internal settlement became the main obstacle to it when they were silenced. The guns did not even have to be fired in order to serve the purpose they had served for so long; they had only to exist. In fact, they didn't even have to exist either, so long as sufficient numbers of people believed that they did. Yet at this moment of extraordinary tactical refinement the republicans were saying that they wanted into talks and that they would accept the outcome of those talks. All they required of the other parties was that they should ignore the guns. The trouble with guns is that it isn't easy to ignore them. Unionists argued that the threat implied in the retention of guns would influence the course of negotiations unfairly, would pollute the democratic process. They were taking a possible political advantage in skirting talks on the excuse that they could not negotiate with people who held arms, but they had a point too. The question over the republican refusal to allow for that point is whether they ever expected the other parties to deal with them while they retained their weapons, or if they were really secretly hoping that the unionists would take fright and refuse. Were the silenced guns of the ceasefires effectively doing the same job for republicans as they had been doing before?

Arguments over guns would continue, if that was what the unionists wanted, but that would not harm republicans. They had nothing to lose by the talks failing to address constitutional issues and getting bogged down in trying to produce disarmament. All that would show was that unionists were unwilling to negotiate. In the end the governments would frame the terms for a referendum on a new devolved settlement, and that was likely to be something that would suit nationalists better than it would suit unionists. Further, if it failed, that would remove devolution as a tenable settlement and challenge the governments to start thinking about joint authority.

Unionists had said they would not negotiate with guns 'on the table, under the table or outside the door'. In saying this they were repeating a phrase coined by John Hume, and taunting him with the charge that he had changed his own mind on this. Republicans had persuaded Hume and the two governments that decommissioning would be simply impossible, because it would betoken surrender. It was more complicated than that. Decommissioning would be a symbolic declaration of the unity of the IRA and Sinn Féin. That is something that was worth great efforts by governments to achieve, and by republicans to avert. If the IRA could be persuaded to give up weapons by those who bartered with Sinn Féin, it would have admitted that it was the same movement. It would also have admitted that republican violence was negotiable. Where might that end, but with the whole IRA campaign being talked down? Violence was being presented by Sinn Féin as a virtual force of nature, which they could show others how to assuage, not as a concerted political campaign that was amenable to reason.

The political value of a refusal to decommission was that it locked unionists onto a single issue and made them appear obsessive and unreasonable. How was it, wondered nationalists of all shades, that everyone else could see that decommissioning amounted to surrender and was therefore too much to ask, but unionists could not see this? It was a significant tactical achievement by republicans to win so many people over to thinking in this way, and it worked to their advantage. The demand for weapons to be given up, even as tokens of good faith, was refused not because it is impossible -nothing is impossible - but to make life difficult for others.

There were strong indications that massive quantities of explosive had been decommissioned in the past, for practical reasons. Several large bombs were discarded in Belfast in the run-up to the Downing Street Declaration, and no such bombs were ever detonated in Belfast again. That suggests that the IRA was dumping gear. It made perfect sense for them to do that rather than hold onto it. Why risk someone getting caught with bags of fertiliser mix that were not going to be used?

By refusing, for political reasons, to decommission, the IRA would eventually either bring the unionists into talks on the understanding that the IRA was a legitimate army, or the war would go on. Hume says it is more important to decommission mind-sets, but that is exactly what the unionists were trying to do; they were trying to get a declaration from the IRA that they had no further attachment to weaponry. They knew as well as anyone that there was no serious prospect of the IRA being completely disarmed, or at least most of those who led the Ulster Unionist Party knew this. The fact that the gesture would have been only a token was often cited by nationalists as evidence that it was meaningless, when it was precisely as a token that it would have been valued.

The arguments for and against decommissioning were frequently framed in terms of the physical value of weaponry in war. The IRA, said one side of the argument, could not be expected to give up its weapons when it had not actually lost the war. The British had failed to defeat their enemy, and were in the position of having to settle terms, essentially to sue for peace with an army that had fought them to a standstill. Conor Gearty, writing in the Guardian, drew an analogy with the PLO which, far from being disarmed by the Israelis, had been assisted in expanding its weaponry so that it could control its territory and police dissident elements.

This was not war in that sense, however. No one was going to concede territory to the control of the IRA. The British, while interested in finding a settlement to the satisfaction of the majority in Northern Ireland, had not been brought to the table by force to concede a constitutional stand-down. It had the option of continuing to police the disorder in Northern Ireland if it chose to. The actual significance of the weapons was political rather than military. The IRA had chosen to discuss the decommissioning demand only in terms that saw it as an undesirable and unearned surrender. Why would they deplete their fighting potential when the dispute was not yet settled? Why, some argued, would they risk leaving the Catholic population undefended before they had won guarantees of its safety?

But the IRA would not have been depleted as a military force if it had decommissioned the same number of weapons in a year as it actually expended anyway. The political advantage of refusing to decommission those weapons must have been reckoned to be greater than the political advantage of letting them go. That superior political advantage lay in forcing a reduction of the conditions for Sinn Féin's entry into talks and imposing the precondition on others that a tactical ceasefire was an acceptable starting point for negotiations. The IRA was successful in that project.

The IRA effectively defeated the British principle that a ceasefire had to be a clear declaration of an end to violence, and got the British to accept what John Hume articulated once himself, that a tactical ceasefire is better than no ceasefire at all. In this they created more room for themselves to operate. It became possible for republicans to speak more openly about the prospects of the second ceasefire ending if others didn't accede to political change.

Thus, following journalists' reports that the ceasefire was to be reviewed after four months, Sinn Féin's Máirtín Ó Muilleoir wrote thus in the Andersonstown News of 26 July 1997:

As for the shock horror IRA November review of its ceasefire, the reality is that the republicans need to review their ceasefire not every four months but every week. How else is the IRA to protect the peace? How else is it to consider ways in which the ceasefire could be underpinned? Ordinary nationalists, determined to defend the peace, demand consistent monitoring by the IRA of its ceasefire.

The conclusion I am led to by the arguments I have outlined in this book is that republicanism is an inappropriate response to the conditions it claims to have been fostered by, and it is one that exacerbates sectarianism and obstructs compromise. It is a creation of both conspiracy and social and political conditions. There was nothing inevitable about working-class Catholic disaffection framing itself into a republican agenda. There were other movements that might have prevailed, with greater prospects for reconciliation and reform. Some of these were armed groups that actually accepted the Provisional analysis that defence was needed, and these were swept away in the political rivalries of the period, so that a stark distinction emerged between just the SDLP and Sinn Féin.

A community of republican conviction that was unimaginable at the beginning of the Troubles now seems so intractable that it cannot be bartered away. In addition, there is an Irish identity in areas like west Belfast that is stronger than it ever was before. Gerry Adams says he can buy his newspaper and get his hair cut in Irish. Yet one of the most telling things about that first generation of modern republicans is their anglicised names. It was later generations who took the Gaelic form; that would have been simply eccentric in 1970.

The civil rights ambition of that time was that social revolution could happen without reference to the old republican template. It seemed a plausible dream then, but it proved hopeless. Republicanism was sufficiently alive to direct the course of the calamity that followed and also to sustain its hierarchy through the upheaval itself, to establish for many the prevailing version of what had happened. The republicans were strong enough to knit their present into the country's past, and to revive the old problem, of how to get off the hook of republican absolutism. A republican definition of the problem ties you to the problem itself so tightly that you can never solve it.

It produces a community of experience, experience that affirms the original vision. So the hunger strikes had to happen - because republicans insisted on being regarded as legitimate soldiers of a long-defeated government. The British then were the enemy, for many who were not republicans, not because they would not concede that legitimacy to republicans, but because they would let men die rather than give them their own clothes.

The point of this book has been to challenge some of the basic myths about republicanism. I have argued that the IRA campaign has not been defensive, in the sense of protecting the lives of people within its support community, and that it has not been offensive, in the sense of being part of a territorial war. I have argued that the armed struggle has in fact been a form of political protest, effectively a veto on an internal settlement. It has often been described as the expression of the hurt of a community, but I think I have shown also that the community within which the IRA moves has a diversity of political opinions and that the organic links between people and the movement, through sympathy with the prisoners and the dead, do not actually translate into a commitment to republican politics, but often involve an acquiescence in things that are not approved of at all.

Eamonn McCann has suggested that we are asking too much of the IRA in demanding that they simply stop their campaign: that amounts to asking them to stand down as the legitimate army of the Irish Republic declared by Pearse and validated by the general election of 1918. That is the problem; that the IRA cannot simply reverse its motivation. The Provisionals came into existence to be true to the old principles and in opposition to all who compromised them. There is nowhere now for them to go that does not compel them to admit that they were wrong all along. Perhaps John Hume was willing to at least endorse their past, in rewriting the history of August 1969, as the price of getting them to stop now, but that entails a lie that is offensive to both republicans and their critics.

The Provisionals, if they really want to stop, are caught in two ways. They have to reconcile the new politics with past action which said that compromise was betrayal. And, ironically, having long argued that the Irish people had a moral right to take up arms to enforce a claim for British withdrawal they are left with no grounds for objecting to other paramilitary groups taking up what they have put down. The Continuity Army Council and the INLA are doing just that. Unless the Provisionals argue that the exclusive right to fight for Ireland lay with themselves alone, they can have no grounds for complaint about others bombing on.

The peace process has been dominated by a scepticism among most parties that the Provisionals can really be making such a radical transition, not just from purist republicanism enforced by violence to pragmatic negotiation, but from forestalling a compromise to embracing one.

Initially, from the British side, the peace process was organised procrastination. The republicans wanted to be in talks within three months and were, but these were not the constitutional negotiations with the British that they had expected. On the night before the first trip to Stormont in December 1994, Gerry Adams presented the talks team to the media outside the Sinn Féin offices on the Falls Road. These were the negotiators. This was a historic moment, the return to that moment in which Michael Collins had faltered. The striking things about that night were that so few people were there, and that Sinn Féin really appeared to believe that a historic moment had come.

'What's your fallback position, when the British say no?' I asked Gerry Adams.
'Do you think the Union is safe?' he said, as if the very idea was ridiculous.

John Hume's theory demands that the British government must ultimately go back to the mistakes of 1912 and 1974 and confront the unionists, and break their veto over constitutional change. He has repeated this theory several times, particularly in his book Personal Views. For Hume, this means a confrontation at the point of constitutional reform. It does not include the idea that Orange-men should be confronted at Drumcree over their right to march. In fact, he has worked for compromise to prevent such confrontation, and spoken of the civil rights of Orangemen. Hume's confrontation only happens when a unionist leadership has reached agreement and attempts are made to overthrow it through mass mobilisation. It is that mobilised threat to agreement that must be confronted, and successfully withstood, for the first time.

There is no similar challenge to nationalists, other than that they stop the violence. Nor is there any clear shared sense among nationalists of where they are going. Unionists know what they want. They want the Union and the border, and they want to stop the growth of the nationalist advantage over them. Nationalists do not seem to know what they want, and they have the luxury of not yet having to say what they want. The moment of truth for them will come when demographic change creates the opportunity of a nationalist majority making a clear decision on the future, whether to go into a united Ireland or not, for instance. Until that day arrives, and it may still be a very long way off, they simply don't have to declare themselves. To unionists, who are constantly being challenged on what they will accept, nationalists seem, therefore, to be riding in the wake of republicans, leaving it to history to open greater opportunities for them. By leaving it to republicans to be the cutting edge of nationalism, and declaring no preference for any other outcome themselves, they squander their hopes.

True, nationalists would argue that there is no visionary unionist on the other side to meet them in a coalition against republicanism. There is no one declaring a love of reconciliation or powersharing. But that is not the whole of the problem for them. There is no actual parity between the positions of unionists and of nationalists. Unionists have to declare their limits every day, nationalists never have to declare them at all. The unionist sin is that they are so unwilling to relax their demands and trade, the nationalist sin is that they are unwilling to declare their demands and trade.

The IRA is not at the centre of the dispute between nationalism and unionism. Addressing that ruptured relationship does not necessarily mean appeasing the demands of the IRA. Skewing the process of reconciliation in the direction of appeasing the IRA may actually make the problem worse.

Sinn Féin regards the unionists as being under a moral onus to help bring IRA violence to an end. In July 1997, after it was clear that Sinn Féin would be admitted to talks, Mitchel McLaughlin said that there was a danger of us sliding back to the old violence if unionists didn't take their responsibilities for peacemaking seriously. John Hume sees it in the same terms. This could only be so if the violence of the IRA was a necessary or inevitable result of unionist behaviour or the Northern Irish past. This is the significance of the Hume argument that all trouble comes out of Bombay Street. It is essentially a nationalist vision in which the unionists are in the wrong. They can hardly be expected to see it that way themselves. This is the result of seeing the problem as a conflict requiring conflict resolution. It is possible to see that there has been huge injustice in Northern Ireland without holding it responsible for IRA violence. It is also important to see that the IRA campaign was honed to be an efficient party political instrument for preventing reconciliation, rather than delude ourselves that it is merely a symptom of the pain of excluded people.

I do not want my children and theirs to be taught in their schools that the IRA campaign was a necessary phase in the readjustment of the constitutional anomaly created in 1921. I will tell them myself that it was a wasteful and horrid business. But the price of ending it may be a reassessment of our history to allow for it. We will argue that the need for it has passed, rather than that there was never a need for it.