When Russia invaded Georgia and then recognised South Ossetia and Abkhazia as independent states, was it following the precedent of Kosovo? Interviewed by the BBC after the announcement of recognition, Dmitry Medvedev, the Russian President, said: "We did it as some other states did in the case of Kosovo." But in that interview he also called Georgia a "special case" that could not be compared to Kosovo, and on the same day the Russian Foreign Minister, Sergei Lavrov, said there was no parallel with the Kosovan case. The very next day, however, Medvedev published an article in the Financial Times citing the Kosovan precedent.

At first sight, this resembles the confusion of miscreant schoolboys who have not yet got their story straight. All kinds of justifications have been thrown into the argument, with little attempt at coherence: protecting Russia's borders, defending Russian passport holders, halting "genocide", and so on. But where the references to Kosovo are concerned, the confusion is surely deliberate.

On the one hand, the idea that there is a single principle, valid for both Kosovo and South Ossetia, is played down: that way, Russia can stick to its existing policy of refusing (for its own geopolitical reasons) to recognise Kosovo. But on the other hand, the K-word is invoked just frequently enough to make Westerners feel uneasy: do they bear some responsibility for the Georgian crisis (they are meant to wonder), and are they applying double standards? In this, at least, the Russian tactic has been quite successful.

Invoking the Kosovan precedent can mean either or both of two claims: that Russia's military action in Georgia resembles Nato's intervention in Kosovo; and that Russia's recognition of South Ossetia and Abkhazia resembles the recognition of Kosovan independence by Western governments earlier this year. The first of these claims need not detain us long.

The Nato intervention in March 1999 took place only after prolonged international diplomatic efforts had failed. Yugoslav forces had already driven more than 400,000 people from their homes, many of which they destroyed by shelling and arson. The UN Security Council had passed a resolution (1199), invoking Chapter VII of its charter (the "enforcement" provisions), demanding a withdrawal of the forces "used for civilian repression". A large observer mission from the Organisation for Security & Co--operation in Europe, which had operated inside Kosovo, had reported further massacres of civilians. Western governments, which had spent more than $20bn coping with the previous war in Bosnia and its tide of refugees, had both genuine humanitarian concerns and a real interest in preventing the spread of instability in the wider region. When Nato started bombing, it was to make Yugoslav forces withdraw from Kosovo; when they agreed to do so, it stopped. Only someone who has been entirely without access to news reports could imagine that Russia's actions have followed anything like this pattern.

But what about the claim that Russian recognition of South Ossetia and Abkhazia follows the Kosovan example? Here the argument can seem more plausible - at least to those who have never understood the legal and constitutional basis for Kosovo's independence. And those people include, unfortunately, most Western politicians and diplomats. Asked to defend their recognition of Kosovo, they have either plumped for a general argument about the principle of self-determination or insisted that Kosovo is a "special case" (thanks to the exceptional nature of the Serbian repression and mass expulsions), which does not follow or create any precedent. Each of these arguments may have some value, but neither suffices to build a solid legal and constitutional justification.

Yet such a justification does exist. Kosovo has a solid claim to independence, based on a well established precedent: the precedent of Slovenia, Croatia, Macedonia and Bosnia. Western governments have been surprisingly bad at explaining this. The only person who has stated it clearly is the President of Croatia, Stipe Mesi?, who has pointed out that the underlying justification is the same for both Kosovo and Croatia: both were constituent units of the former Yugoslavia.

But first a little historical background. The territory of Kosovo was part of the Ottoman Empire for just over 450 years, until it was conquered by Serb and Montenegrin forces in 1912. (Serbs would say that it was "liberated"; but they formed less than 25 per cent of the Kosovo population at that time.) After the conquest, it remained occupied territory; it was not legally incorporated into Serbia. It was then conquered again in the First World War, and finally absorbed into a Yugoslav kingdom in 1918. For the rest of the 20^th century, with one major interruption (the Second World War), it was always part of a Yugoslav state. It was never simply part of a Serbian state - until the summer of 2006, when it was treated, for the first time in modern history, as part of a sovereign Serbia. (When Kosovo gained its independence in February of this year, it had been treated as simply part of Serbia for less than 20 months.)

Under the Titoist constitution of 1974 - the final version of the Yugoslav constitution, before the illegal changes forced through by Slobodan Miloševi? - there were eight units: six republics and two provinces. Kosovo was one of the provinces, with a dual status, which was defined in a rather contradictory way. On the one hand, Kosovo was described as part of Serbia; on the other, many clauses gave Kosovo all the essential rights and functions of a republic within the federal system. It had its own parliament, government, bank, territorial defence force and so on; it had the right to issue its own constitution; it even had the right to sign international agreements; and it was represented on all the main federal bodies directly - not as a part of Serbia. To say that Kosovo had a dual status in the constitution might make it sound as if it was half part of Serbia and half part of the federal system, but the reality was more like 5 per cent and 95 per cent. It was part of Serbia in some very limited and theoretical ways, and directly part of the federation in almost every practical way.

In 1991-2 the European Union asked a commission of constitutional lawyers to advise it on the legal aspects of the break-up of Yugoslavia. This commission, chaired by Robert Badinter, found that the Yugoslav federation had "dissolved"; and the use of that term is important. There is a great difference between what happens when one or more units "secede" from a federation and what happens when the entire federation dissolves. In the case of secession, there is a continuing constitutional and legal order: branches may drop off the tree, but the tree trunk still exists. In the case of dissolution, there is no more tree: every unit can go its separate way.

This finding has huge significance for Kosovo, even though Kosovo itself was not discussed by the Badinter commission. (It is sometimes said that the commission decreed that only the six republics, and not the two provinces, had the right to become independent, but this is not true; the Badinter commission existed only to answer specific queries put to it, and no such query was ever put.) For the essential question is this: when a federation dissolves, what does it dissolve into? And the answer must be: into its federal units.

This issue has sometimes been confused by people who say that, under that constitution, only the six republics had the right to secede, not the two provinces. That statement is not really correct, but in any case a constitutional "right to secede" is not relevant here. Such a right is something that can be asserted in the constitutional court, and that cannot happen when the entire federal state dissolves. Then there is no "secession", no constitutional court and no continuing legal order to be enforced.

How can we decide, at that moment, which units have a right to emerge as independent states from the former federation? The answer must be that we analyse how the units of the federal state have operated, from a descriptive or functional point of view. Considered in this way, Kosovo had clearly been a unit of the federation, exercising governmental powers on a par with the other units; therefore it too should have had the right to seek independence when that federation dissolved.

Those who say that the independence of Kosovo violates the Helsinki principles about creating new international borders should ask themselves whether those principles were violated when Slovenia and Croatia became independent. Certainly, the independence of those states was contested by Serbia at the time. But the international community easily accepted that what was happening was not the creation of completely new borders but the upgrading of existing borders - the borders of federal units - to international status.

The only important difference here between Slovenia and Kosovo is that the rights of the former were recognised quickly by the international community while those of the latter were not. For purely pragmatic political reasons, Western governments continued to treat Kosovo as if it were just part of the so-called federal republic set up by Serbia and Montenegro after the break-up of Yugoslavia. And again, for pragmatic reasons, they drifted into the position of treating Kosovo as merely an internal part of Serbia - in theory, though not in practice - after the summer of 2006. But the underlying justification of Kosovo's claim to independence remains, and cannot be overturned merely because Western governments made a mistaken interpretation of the facts. Those countries that still deny recognition to Kosovo are only prolonging that mistake.

How, then, does this precedent apply to South Ossetia and Abkhazia? The short answer is that it does not. In the Communist period, Abkhazia was an "autonomous republic" within Georgia, with some special rights of self--government; South Ossetia had a lower status, as an "autonomous province". But Georgia as a whole was not, and is not, a federation; and it is certainly not a federation that has undergone dissolution. The Soviet Union did dissolve, but Abkhazia and South Ossetia had not been operating directly as units of the Soviet federal system. There is nothing here equivalent to the emergence of Kosovo from the dissolution of the former Yugoslavia.

Possibly, a case could be made for Abkhazia's having had a right to independence on the basis of three little-known laws passed under Gorbachev in April 1990. These laws said that the "autonomous republics" within the member states of the Soviet Union had the same rights of self--determination as the member states themselves. So it could be claimed that, during the short period in which the USSR continued to exist, the "autonomous republics" also had the right to seek independence. This right existed, however, only within the (very recent) constitutional law of the Soviet Union, and would have no force once that union had dissolved - a dissolution that did in fact take place before the Abkhaz declaration of independence.

In any case, Moscow is unlikely to use this argument. For it is the fundamental legal argument for the independence, from Russia, of Chechnya (which did assert its claim in good time); and it was on the brutal suppression of the Chechens that Vladimir Putin rose to power. There is also an essential difference between the Chechen and Abkhaz cases. In Chechnya the majority of the population were Chechens, who did indeed support independence; in Abkhazia, the Abkhaz themselves were a small minority, and if they are more than 50 per cent today - which is still doubtful - it is only as a result of the large-scale ethnic cleansing of Georgians from their homes. Even if Abkhazia did once have some Soviet legal right to seek independence, the Abkhaz could never have attained it by democratic means.

Instead they have attained, with Russian help, a limbo status that will not turn into genuine independence. And that, of course, is what Moscow wants: the indefinite prolongation of a crisis, creating permanent trouble for, and leverage against, Georgia. Western recognition of Kosovo was designed to end limbo status and to create a permanent solution from which Serbia would also benefit in the long run, becoming a normal state and entering the EU.

Here too, then, there is a fundamental dissimilarity: Russia's aim is to perpetuate a problem, not to solve it. Paradoxical though it sounds, it is surely true to say that Russia has recognised South Ossetia and Abkhazia precisely because it does not expect other countries to do so.