HOW STRANGE it is not to know! The human race has spent half a million years struggling to make the universe more predictable, and we still don't know what Scotland will feel like on Friday morning. To fill that scary blank, polls and guesses fill the air like circling gulls. But Scots lie to pollsters more delicately than any other electorate in Europe. As I write, a third of the voters - an unprecedented mass, at this stage of an election - say they haven't made up their minds. Many of them are lying too. Why not? This time, it's too important a decision to share.

It could be an earthquake, after which nothing in Scotland will look the same. Or it could, though I don't think so, turn out to be one of those national flat-tyre days, another defeat adeptly snatched from the jaws of victory. There is nothing conclusive to predict with. The old historian WF Skene once made a whole book-chapter out of one enigmatic sentence in a Gaelic chronicle: "He was not King on Thursday in Kintyre." Somebody is not going to be king on Thursday in Holyrood, and I think I know who it is.

The Act of Union came into force 300 years ago on Tuesday. Some people think it will die, or enter its terminal illness, on Thursday. But that is a mistake. The union has become two things, over those centuries. The first one is a slow-growing closeness of two peoples. This has nothing to do with acts or treaties, and it endures. The other union is a document which has been read in many different ways, and it's in trouble.

The official, 1707 union has really been four different types of union in succession. Its first form was as a disabling measure, designed to complete the 1603 Union of Crowns and remove Scottish capacity to obstruct Hanoverian arrangements by tying Scotland to the will of England's parliamentary majority. Then it became an easy-going association, in which Scotland benefited enormously from the global markets and employment of the empire and was ruled by its own oligarchies with minimal London interference.

A third union, far tighter and less accommodating, appeared when the modern, interventionist British state was constructed in the 20th century. Uniformity was imposed in the cause of much-needed social justice. The state became the great provider, partly through the Scottish Office but always within guidelines for demand management and public spending laid down in Whitehall. But when Margaret Thatcher demolished the post-1945 settlement, along with what remained of traditional industries, the grandeur of the British state also shrivelled. If it was no longer the guarantor of fairness, the fount of innovation or the protector of the poor, what was this remnant for? In Scotland, that question took its own form: if the union no longer assured work and opportunity and social justice, what was the point of it?

So a fourth version of union was hastily designed, restoring a Scottish parliament that controlled some internal affairs. But the basis of this union is entirely political. Where are the grand economic and social aspirations which powered its predecessors in the 19th and 20th centuries? The purpose of devolution, in the minds of its Westminster and Downing Street architects, was no more than this: to keep the union going.

Has it enough content to be credible in this form? The signs are that it has not. The union has come to seem rather pointless even in England, where it is often perceived today as a device for giving the Scots too much money and preventing the English from doing anything about it. In Scotland, devolution looks like a much-overdue but still inadequate grant of powers that ought to be expanded. But that is not what it's meant to be. Few recognise Holyrood as only the latest of all the guises that the 1707 union has taken. But they should - it would warn them not to expect an easy extension of Holyrood's powers when Scots demand it. All these unions are made in London, and only ratified in Scotland.

So the union as a constitutional device has lost most of its old points and justifications, and is beginning to look like a husk. But that is not the only union we have. It's also an intimacy, an unofficial relationship between two peoples that is - I feel - growing slowly stronger as illusions about "Britishness" fall away. Three hundred years ago, the English truly detested the Scots. The British Museum is celebrating the union tercentenary with a book of Caricatures Of The Peoples Of The British Isles, its pages squirming with verminous, bare-arsed, shamelessly hypocritical and ambitious Scots pouring into the streets of London.

English right-wing papers today are attempting to rouse that Scotophobic feeling again, as part of their campaign to drown Gordon Brown in ridicule before he reaches the safety of Number 10. But this is not the 18th century, and the two nations are now connected by a dense root-web of family connections, friendships, common culture and mutual experience. Nothing in that sort of union would be affected by Scottish independence. The Czech and Slovak experience suggests that Alex Salmond could be right when he says that the two peoples would get on better after political separation than before.

But mention of Czechs and Slovaks leads into some of the less cheerful scenarios being imagined. Let's suppose that the SNP do win a relative majority. Suppose the LibDems abandon their unconvincing "no referendum" bargaining posture (what sort of unionist party is afraid of finding out that people don't want the union?). Suppose they form a coalition with the SNP, and an independence referendum is slated for a few years' time.

If that was all that happened, the prospect would be relatively simple. The trouble is that other things will be happening too, which will set up turmoil long before we get to drafting the referendum questions. The LibDems, even more than the SNP, are determined to demand a radical increase in the powers of the Scottish parliament (fiscal federalism and so on). The SNP are going to demand a share of oil revenue and a transfer of more currently reserved powers. But none of these things is on offer, absolutely not from Gordon Brown, and - if it were possible - even less from a British government led by David Cameron. No Westminster political leader can now afford to be seen to be making more financial and constitutional concessions to a Scottish government, and certainly not when that government is not of his own party. So there will be a head-on Edinburgh-London power struggle.

The question remains: who will win? Well, maybe nobody. The break-up of Czechoslovakia in 1993, the so-called "velvet divorce", showed that a multinational state can fall apart simply because its politicians grow tired of holding it together. The Slovak people did not demand independence, and neither did the Czechs. All they wanted was a better federation. But Czech politicians grew tired of Slovak demands for more, and decided that they would be better off without Slovakia altogether. So they announced that negotiations had failed and the union was over. (Neither people was allowed a referendum).

This could happen to Scotland. Independence could arrive, but in the shabbiest possible way. It would not come because the Scots had displayed a firm "settled will" to become a small but sovereign European nation. It would happen because a bunch of London civil servants and politicians ran out of patience with demands, saw an easier future and more promising careers without the union and invited the Scots to shut up or push off.

Another scenario is the Quebec Syndrome. This happens when the voters make a distinction between a nationalist party and its central policy. The French-speaking voters of the province eventually elected the Parti Quebecois (PQ). It stood for independence, or sometimes for "sovereignty-association" with Canada. However, when the PQ did what they were born to do and held a referendum on independence, the voters narrowly rejected it. So began a cyclical process, steadily eroding PQ morale, in which the party kept getting a mandate to govern but not a mandate to take Quebec out of Canada. As with the Czech/ Slovak Syndrome, there is evidence of the Quebec pattern in Scotland. After all, it is not unreasonable. What's wrong, from the voter's point of view, in deciding that a nationalist government will best defend our interests but that Scottish independence is a matter for hesitation? All the polls suggest this pattern is widespread in Scotland and that interest in independence is diminishing as interest in voting SNP increases.

It sounds bizarre, but it makes a sort of sense. For many years, Scots have voted for parties while rejecting their constitutional agendas (think of those periods when most supporters of full independence were Labour voters). What the Scots appear to want is beautifully pragmatic: to feel that we are running our own affairs as any normal small nation does - just that. If it can be achieved within devolution, then OK. If it requires independence, then so be it. A simple wish. But not one party has the nerve to put it into its manifesto.

These are pitfalls long ago, I think, identified by Alex Salmond. He must know that if there were to be a one-question, yes/no referendum on independence in the near future, he would almost certainly lose it. And no would mean no. But if there is a multi-question referendum, a "concession" the LibDems might artlessly buy, then there would be a good chance of the independence option winning most support - though no solution would get an absolute majority.

So independence would stay on the agenda, waiting for the right moment to be put back to the voters. And that, for the SNP, would be when London and Edinburgh come into serious political collision over "expanded devolution" and money - the moment "when England says No". Then the mood in Scotland could change sharply. Alex Salmond needs a Czech/Slovak Syndrome to avoid a Quebec Syndrome.

But this election is about the near future, about the Scottish parliament rather than distant seats at the UN. And I have been astonished to hear so many people say: "It's so hard to choose. This guy really has to get back into Holyrood, but vote for his party? No way!" Are these the same voters who said four years ago: "It's so hard to choose. They're all numpties"? People still undervalue what the parliament has achieved, but now they know a fair number of MSP candidates, and can put a face to a name on a list. Best of all, they want to choose.

Scottish democracy is taller this time round, and that's good news for the testing years ahead.