In the last major interview he gave before his death this week, Ernest Gellner told Simon Targett why, despite bloody Bosnia and divided Quebec, he still believed that ethnic nationalism would not triumph.

Le Quebec aux Quebecois!" The cry of the French Canadians was a clarion call, stirring the ethnic sentiment of five million descendants of Jacques Cartier and the New France colonists. Chanted with passionate pride, it invoked a time when Quebec was still an Eden for fur-trading francophone settlers, a time before the British general James Wolfe outwitted the Marquis de Montcalm by scaling the famously jagged cliffs overlooking the St. Lawrence River and capturing the capital city. In the end, of course, the "Oui" for independence was out-shouted by the "Non" for Canadian federalism, but the rallying cry is sure to be heard again, and the clear vitality of the Quebecois campaign gives new insistence to the central question on nationalism: is nationalism ethnic or civic? In doing so, it also calls into question the majesterial work of that crowning figure of civic nationalism: Ernest Gellner.

It is more than 30 years since Gellner, a former professor of social anthropology at Cambridge who at the time of his death from a heart attack directed the centre for the study of nationalism at George Soros's Central European University in Prague, first pronounced on the subject of nationalism, writing a celebrated essay in his second book, Thought and Change. In the early 1960s, nationalism was considered to be primordial and perennial - in Bagehot's words, "as old as history". Overturning this orthodoxy, Gellner maintained that nationalism is modern - the product not of distinctive cultural attributes but of industrialisation - and the guiding principle of political states which are socially differentiated yet unified by a "high" literate culture based on a single language.

Gellner's general interest in nationalism was rooted in his cosmopolitan background: born in Paris in 1925, raised among those he depicted as the "provincial Bohemian petit bourgeois" of Prague, educated at Oxford in the 1940s. But his academic interest, and the origin of his thesis, stemmed from his reading of Elie Kedourie's Nationalism and his study of anthropology. Kedourie, writing in 1960, contended that nationalism is recent, a novelty of Romantic thinkers. As Gellner, giving The É«ºÐÖ±²¥S what turned out to be his last major interview just before a public debate on nationalism at Warwick University last month, recalled: "His book, which I read in typescript, really shocked me out of the commonest and mistaken theory of nationalism: namely, that it is natural, that it is like hunger or gender, that it is inherent in the human psyche."

É«ºÐÖ±²¥

But Gellner did not swallow the specifics of Kedourie's argument - that nationalism was a doctrine invented by European ideologues like Kant - and that was partly because he approached the origin of nationalism from the perspective not of a political scientist but an anthropologist who had worked among the Berber tribes of Morocco. At the time, social anthropologists distinguished between the "structure" and the "culture" of a society, and for Gellner this notion "seemed to contain the clue to a valid explanation". In pre-modern, agrarian society, culture - language, accent, and so on - was richly nuanced, yet regarded as "frivolous", as "inessential" to personal identity. Structure, and an individual's place in the neatly ordered hierarchy, was everything.

By contrast, in modern, industrial society, there is an "erosion of structure", relationships become encounters - ephemeral, non-repetitive, optional - and culture, especially literacy, becomes elevated by the need for precise communication between strangers. As he once wrote: "If a man is not firmly set in a social niche, whose relationship as it were endows him with his identity, he is obliged to carry his identity with him, in his whole style of conduct and expression: in other words, his 'culture' becomes his identity." This told him that "nationality" - or "the classification of men by 'culture'" - must be modern. It also told him that any sense of "awakening", any perception of an ancient ethnic inheritance, however deeply felt, must be mistaken. "Nationalism is an expression of industrial society which romanticises pre-industrial society," he explained. "There is no relationship whatsoever between what nationalism says about itself and what is really the case."

É«ºÐÖ±²¥

These musings have become the new orthodoxy. But in the wake of the ethnic brawls of Eastern Europe and now the re-emergence of the Quebec question, this is coming under sustained attack from scholars who emphasise the importance of ethnic continuity. Leading the way is Anthony Smith, professor of sociology at the London School of Economics, and author of National Identity.

The dispute was not characterised by the personal animosity that has marked some of Gellner's academic tussles. Smith is one of Gellner's former PhD students, and Gellner said he was "proud" that Smith had become "the leading specialist on nationalism". Yet the dispute remained keenly contested, and any opportunity to debate the point was rarely missed. Just two weeks ago, Gellner flew from Prague just to renew the argument with Smith, this time before an audience at Warwick University.

There, Smith revealed that he did not think Gellner was "wrong" as such, and he resisted his mentor's gently mischievous attempts to paint him with the primordialist brush. Yet he suggested that Gellner "only tells half the story". Defining his own position as "ethno-symbolic", Smith maintained that "modern political nationalism cannot be understood without reference to earlier ethnic ties and memories", pointing out that, for instance, the story of William Wallace, written down long ago, is more than merely a mythologised appropriation of modern Scottish nationalism.

One reason for the difference between Gellner and Smith is the tangle of terminology. Gellner was suspicious of the word "ethnicity", preferring "shared culture" instead. Smith invents his own concepts, like "ethnie" for ethnic communities. Another, bigger reason is the conflict of purpose. Gellner was the great generalist, and his work amounts to an all-encompassing theory of nationalism, an anthropological model. Smith is the specialist, and his work is, as he puts it, "an approach, a perspective, not a theory".

Gellner acknowledged that there are "problems" with his theory: the Kurds seem to have a tribal and a national identity, Islam has a religious rather than a national identity, and Yugoslavia has been pulled apart by people who speak broadly the same language. But, with an engaging shrug, he suggested this is "recalcitrant evidence", almost as if there is something wrong with the evidence rather than the theory. He added, in his deep, gravelly, East European voice: "My theory fits a very great deal of the evidence: something like 80 per cent in Europe, and a lower proportion elsewhere." And, on Quebec nationalism, he maintained his theory "fits perfectly", as it corresponds to his idea of a once-backward people catching up with their rich and powerful neighbours.

É«ºÐÖ±²¥

Smith, spotlighting the shortcomings of Gellner's scheme, thinks that the origin of nationalism is still beyond the definitive theory. "I doubt," he said, "if we are in a position to offer a theory of so protean and many-sided a set of phenomena as ethnie, nations and nationalism, except at a very general level."

On this, he gets the backing of Michael Ignatieff, a cosmopolitan like Gellner, but author of a book which stresses the inescapable fact of inherited ethnic division, Blood and Belonging. If modernity has reduced the salience of inherited ethnic difference, if the content has in his words been "hollowed out", it nevertheless remains for people to rally around, and the issue then becomes what he calls "the narcissism of minor difference".

Why, for instance, do the Quebecois, who have come to share so much with contemporary anglophone Canadians, choose to separate on the basis of comparatively inconsequential cultural dissimilarity?

É«ºÐÖ±²¥

So is Gellner's hypothesis doomed? Tom Nairn, a commentator on nationalism who worked at Gellner's Prague-based centre for nationalism and is now lecturer in nationalism at Edinburgh University, thinks not - at least, not necessarily. He contends that ethnic nationalism could be a contemporary and passing phenomenon: "It may come to appear, say early next century, as primarily an episode of the 200 years from the late 18th century to the 1950s when the whole developmental process was distorted and crushed by successive forms of imperialism which aroused ethnicity as its inevitable reposte." As further evidence, he adds that "apparently hopeless cases" like South Africa, Ulster and Palestine are being resolved in terms of civic rather than ethnic nationalism, even if major setbacks like the assassination of Israel's prime minister Yitzhak Rabin demonstrate that the struggle is far from over.

Nairn's optimism was shared by Gellner. There is sometimes a feeling that the whole "ethnic versus civic" debate has long been narrowly academic, slightly contrived and slightly unworldly. Nairn acknowledges that Gellnerian modernists have been engaged in "a form of elite conversation" which has had only a small impact on the wider world. Yet, he adds, things are changing, and certainly just before his death, Gellner was ready to speculate on the future of nationalism.

An opponent of the creation of independent ethnic states - something he dismissed as "facile Woodrow Wilsonism" - he saw "the way forward through political centralisation and cultural pluralism: having cultural nationalism which concentrates on maintaining the language but does not insist on political sovereignty because the nature of modern technology is such that unless there is a strong central power there will be ecological or terrorist disaster". In particular, he recommended a shift to what he called "cantonisation", and he liked what was happening in Austria, which he said is "just the right kind of unit".

By praising the capital of the old Habsburg Empire and the birthplace of Hitler, Gellner was exhibiting his uncommon gift for controversy that lay behind the creation of a theory about nationalism which - in the teeth of recent events - might yet be proved right.

É«ºÐÖ±²¥