In the UK, referendums are rarely held. Quite a few are offered at some distant point in the future, but governments only call them when they are confident they will win. This is what makes the drama over a referendum for Scottish independence so explosive. Referendums here are not about leaders discovering a sudden passion for direct forms of democracy. Usually they are about leaders seizing control of controversial policies.
Understandably, fragile rulers seek to control their timing and the question or questions to be placed before the electorate. If they do not believe they can win they do not hold them. Although Gordon Brown and Ed Balls applied tireless guile to stall Tony Blair's increasingly irrational desire to join the euro, Britain would not have signed up if they had been more laid back. Blair realised he was not in a position to win the referendum so he did not hold one. There were several others proposed in Labour's 1997 manifesto that never came about.
One was on electoral reform, although Blair did not hold that partly because he was worried he might win. He was never a supporter of electoral reform. The key issue for him as a leader was that he was in control of whether they took place or not, as was Harold Wilson when he brilliantly timed a vote on Britain's continuing membership of the EU in 1975 when he was certain of victory. The wily Wilson would have found a way of avoiding it if he thought he would lose. I can hear him stating mischievously if the polls had looked bad: "We need to continue renegotiating the terms of Britain's membership in Europe until we are absolutely sure they are right".
The striking element of the Scottish drama is that no leader has such control. Arguably, because of the constitutional complexities, no leader has been fully in control before. But events have moved fast in recent days. We now know almost certainly that there will be a referendum. This might seem a statement of the obvious, but, as Blair demonstrated, when leaders offer them, they do not necessarily hold them.
Now Salmond has specified the autumn of 2014 and ministers in London indicate a willingness to accept. There is little or no wriggle room for either side. Salmond has looked and sounded as confident as ever over the last few days, but by being more precise, he has lost a degree of flexibility over the timing. The Westminster government is not in control either, but then it never was. Cameron could not have imposed a date on Scotland without alienating some of those who are opposed to independence.
For once leaders have let go of some levers and the voters will hold sway on an issue, a highly volatile sequence partly because of its rarity. The closest parallel is the referendum on devolution that the Labour government felt obliged to hold in Scotland in 1979. Look what happened then. Jim Callaghan failed to secure the required majority in favour. A vote of confidence then followed that Labour lost and the party was out of power for 18 years. Leaders may like to offer referendums as a way of avoiding making a decision or as a device to unify a party. Yet they can turn on them and blow them apart.
Callaghan's 1979 referendum is a reminder of why the current so-called "process" negotiations are vital, at least as important as the campaign that will follow. The act preceding the 1979 plebiscite stipulated that 40 per cent of the entire Scottish electorate would have to vote "Yes" for the changes to be implemented. The "Yes" vote won, but on a low turnout and so failed. The "process" doomed the proposals from the beginning.
The negotiations between London and Edinburgh are therefore much more than a tedious prelude. With both Salmond and Cameron no longer in control over the timing of the referendum they can do no more than keep their fingers crossed that the immediate context of the campaign will help their cause. Much will depend on the state of the UK economy, the health of the Labour party in Scotland and in the rest of the UK, factors either wholly or largely out of the influence of Salmond and Cameron.
Of course, ideally, the immediate context should not determine the outcome as voters are making a decision that might have historic long- term consequences. But the political background at the time of the campaign is pivotal. This is one of several reasons why I am opposed to referendums.
The battle over what is placed on the ballot paper is also vital. Not surprisingly the Westminster leaders want a straight "Independence – Yes or No" question. The polls suggest they will win such a campaign, although no one knows what the mood will be in 2014. Here again is another example of leaders only wanting to hold referendums they can win, another reason why I am against them.
Salmond wants to pose an additional question about more powers for the Scottish parliament within the UK, because he is not sure he can win a vote for independence, not out of a noble, altruistic desire for all the options to be placed before the electorate. On this he may not prevail. It is for the Westminster parliament to decide on which powers it is willing to hand over to Edinburgh if Scotland remains part of the UK.
But a vote against independence in a straightforward "yes or no" ballot will not end the story. Referendums never do. Look at the continuing raging debate about Europe in spite of the massive "Yes" vote in 1975. We can expect another one on electoral reform if there is a Lab/Lib coalition after the election.
The referendum in Scotland will cause political convulsions if Salmond wins – and will not change much in the longer term if he loses. Labour's devolution settlement has unleashed a fresh appetite for nationalism, although it was designed to do the opposite. That appetite will increase whatever happens in 2014.