Nations & Nationalism

Memory and modernity:
reflections on Ernest Gellner's theory of nationalism

by Anthony D. Smith, 21 March 1996
The Ernest Gellner Memorial Lecture

[see also Nations & Nationalism - Warwick Debate: Anthony D. Smith & Ernest Gellner]

This is a sad and strange occasion. I have been asked to stand in the place where my teacher, Ernest Gellner, was to stand today and to continue an unfinished dialogue which we have been conducting for much of our scholarly lives. I have been asked to speak to you about Ernest's theory of nationalism, the great issue of the modern world with which he grappled all his life and to which he made so unique and profound a contribution.

Though I had heard Ernest lecture in 1964 and 1965, it was only when he agreed to supervise my Ph.D. thesis in 1966 that I came into close contact with him on a regular basis. Since that time, Ernest has been in my thoughts as a teacher and scholar, and above all as a pioneer of the sociological study of our common passionate interest in nationalism. (And in our last meetings, both of us shared the hope of seeing a sister Institute dedicated to the study of nationalism being created at LSE, similar to that which Ernest directed at the Central European University in Prague. But that hope seems unlikely to be realised).

Though I had read Words and Things at Oxford, my first real encounter with Ernest's thought was with his second book, Thought and Change (1964), especially the chapter on nationalism. This chapter has largely set the terms of subsequent debate in the field. From this encounter, and my subsequent work under Ernest's supervision, I took away four fundamental lessons in the study of nationalism.

The first was the centrality of nationalism for an understanding of the modern world. The fact that Ernest took up the issues of nationalism in the 1960s and that he kept returning to them, when most social scientists were interested in Marxism, functionalism, phenomenology, indeed everything but nationalism, and the fact that he established a Centre for the Study of Nationalism in Prague devoted to research and latterly teaching in this field, demonstrates how thoroughly he appreciated the power, ubiquity and durability of nationalism.

The second lesson that I learnt from Ernest was the need to appreciate the sheer complexity, the protean elusiveness, of the phenomena that were gathered together under the rubric of 'nationalism'. This is why he insisted on comparative analysis, and on the need to formulate typologies that do justice to the complexities of nationalism.

A third lesson was the sociological reality of nations and nationalism. Unlike many latterday scholars for whom the nation is a cultural artefact and nationalism a discourse, Ernest insisted on the structural embeddedness of nations and nationalism. Hence, the need, as he saw it, to use sociological concepts and methods to provide an understanding of this most complex of phenomena. This meant of course jettisoning nationalism's own account of itself, as an awakening of the slumbering but primordial nation through the kiss of nationalist Prince Charmings; and, instead, grasping nationalism as the necessary outcome of a particular kind of social structure and culture.

The final lesson was the hardest, the one on which I have stumbled most. Nations as well as nationalisms, Ernest argued, are wholly modern. They are not only recent, dating from the period of the French Revolution or a bit earlier, they are also novel, the products of 'modernity' that whole nexus of processes that went into the making of the West over the last four centuries, including capitalism, industrialism, urbanisation, the bureaucratic state and secularisation.

It was this final lesson that was at issue in our last encounter at Warwick, when, at the invitation of Edward Mortimer and the university, Ernest and I debated the origins and functions of nationalism. That debate, just twelve days before his tragically early death, was entitled: 'The nation: real or imagined?' But since we both agreed that nations, like buildings or works of art, are created - albeit over generations - and are therefore both real and imagined, the question became the different but perhaps more important one of the relationship between nations and their putative pasts. It is, after all, difficult to see how a purely cultural artefact could inspire the loyalty and self-sacrifice of countless people. On the other hand, the primordialist picture of natural nations, of nations inscribed in the natural order, was equally unacceptable. So the question then became: where do nations come from?

Do nations have navels?
Or, in Ernest's words, 'Do nations have navels'?

This question was the title that Ernest gave me for this Nations and Nationalism lecture, as we stood on the platform in Coventry after the Warwick debate. Hence my problem, and my title. Fortunately, Warwick University kindly supplied me with a transcript of Ernest's reply to my opening statement in that debate. Let me quote the bit about national navels, on which he hoped to elaborate. Speaking of the dividing line between modernists and primordialists, Ernest asked for the kind of evidence that would decide whether nations had pasts that matter, or whether the world and nations with them was created about the end of the eighteenth century, 'and nothing before that makes the slightest difference to the issues we face'. 'Was mankind', he asked, 'the creator of Adam and did it slowly evolve?' The evidence that was debated at the time this issue was alive revolved around the question: did Adam have or did he not have a navel? Now, it's a very crucial question, you see. No, no you may fall about laughing, but obviously if Adam was created by God at a certain date, let's say 4,003 BC, obviously I mean it's a prima facie first reaction that he didn't have a navel, so to say, because Adam did not go through the process by which people acquire navels. Therefore we do know what will decide whether the world is very old and mankind evolved or whether the world was created about 6,000 years ago. Namely, all we need to know is whether Adam had a navel or not. The question I'm going to address myself to of course is do nations have navels or do they not?

My main case for modernism that I'm trying to highlight in this debate, is that on the whole the ethnic, the cultural national community, which is such an important part of Anthony's case, is rather like the navel. Some nations have it and some don't, and in any case, it's inessential (Gellner 1995, 1-2). What in a way Anthony is saying is that he is anti-creationist and we have this plethora of navels and they are essential, as he said, and this I think is the crux of the case between him and me.

In my opening statement at Warwick, I had argued that the modernist standpoint which Ernest embraced - the idea that nations are products of modernisation and could not have existed before the advent of modernity - told only half the story (Smith 1995b)

Well, if it tells half the story [Ernest quipped], that is for me enough, because it means that the additional bits of the story in the other half are redundant. He may not have meant it this way, but if the modernist theory accounts for half or 60% or 40% or 30% of the nations, this is good for me. (Ibid., 2)

Well, as Ernest knew, I certainly didn't mean it that way. But, from his standpoint, it would actually suffice if there were just one case of a nation being accounted for by modernism, for modernism to be true. And he produced his case: the Estonians. This is what he had to say about them: At the beginning of the nineteenth century they didn't even have a name for themselves. They were just referred to as people who lived on the land, as opposed to German or Swedish burghers and aristocrats and Russian administrators. They had no ethnonym. They were just a category without any ethnic self-consciousness. Since then they've been brilliantly successful in creating a vibrant culture. (Ibid., 2) And he went on to praise this 'very vital and vibrant culture', which is so vividly displayed in the Ethnographic Museum in Tartu with its 100,000 objects, one for every 10 Estonians, claiming that it was created by 'the kind of modernist process which I then generalise for nationalism and nations in general'. And Ernest returned to the Estonians, at the end of his opening statement, when he tried to list the factors that may help us to predict which potential nations or cultural categories will assert themselves, a question I had posed in my opening statement: Now obviously it does matter to predict which nations will assert themselves, which potential nations, which cultural categories will assert themselves and which will not. I would say it is inherent in the situation that you cannot tell. You can indicate certain factors. Size is an obvious one, very small cultural groups give up. Continuity is another one, but not an essential one. Some diasporic communities have very effectively asserted themselves. Size, continuity, existence of symbolism, are impor- tant, but again the Estonians created nationalism ex nihilo in the course of the nineteenth century. (Ibid., 3)

I could quibble here, and say that the issue was not whether the Estonians created nationalism ex nihilo in the nineteenth century, but whether the Estonian nation was created by the Estonian nationalists ex nihilo. And while we would both agree that Estonian nationalism, indeed any nationalism, was modern, where Ernest and I would differ is whether the nations that nationalism creates are wholly modern creations ex nihilo. Ernest returned to this question, when he disagreed with my reading of the classical legacy of modern Greece, but admitted that there is some continuity with Byzantium or at any rate with the clerical organisation left behind by Byzantium certainly, but sometimes there is and sometimes there isn't [continuity]. So I would say in general there is a certain amount of navel about, but not everywhere and on the whole it's not important. It's not like the cycles of respiration, blood circulation or food digestion which Adam would have to have in order to live at the moment of creation. (Ibid., 3)

Now here lies the rub. If we pursue the analogy, we recall that God created Adam, fashioning his body and then breathing life into it. Not even the most megalomaniac nationalist has claimed quite that power. They have, of course, seen themselves as awakeners; but the body of the nation merely slumbered, it was not without life. Should we confer on nationalists that divine power, to create ex nihilo?

Of course, Ernest wants to confer that power through nationalism ultimately on modernity, on the growth society, on industrialism and its cultural prerequisites. For Ernest, the genealogy of the nation is located in the requirements of modernity, not the heritage of pre-modern pasts. Ernest is claiming that nations have no parents, no pedigree, except the needs of modern society. Those needs can only be met by a mass, public, literate, specialised and academy-supervised culture, a 'high culture', preferably in a specific language which allows context-free communication. A 'high culture' is the only cement for a modern, mobile, industrial society; and this is the only kind of society open to us today.

For Ernest, the world was irreversibly transformed by a cluster of economic and scientific changes since the seventeenth century. Traditional agro-literate societies were increasingly replaced by growth-oriented, mobile, industrial societies. The rise of high cultures and nations is a consequence of the mobility and anonymity of modern society and of the semantic, non- physical nature of modern work. Today what really matters is not kingship or land or faith, but education into and membership of a high culture community, that is, a nation (Gellner 1983, ch. 2).

So, just as Pallas Athene sprang fully armed from the head of Zeus, without parents, so nations emerged fully-fledged from the requirements of modernity. If nations did have navels, they were purely ornamental.

But, can we derive nations tout court from the needs of modernity? To be fair, it isn't modernity that directly creates nations. To quote Ernest's original formulation: '[Nationalism] invents nations where they do not exist, - but it does need some preexisting differentiating marks to work on, even if ... these are purely negative' (Gellner 1964, 168). The same sequence is restated in his later book: "It is nationalism which engenders nations, and not the other way round." Admittedly, nationalism uses the preexisting, historically inherited proliferation of cultures or cultural wealth, though it uses them very selectively and it most often transforms them radically. Dead languages can be revived, traditions invented, quite fictitious pristine purities restored. But this culturally creative, fanciful, positively inventive aspect of nationalist ardour ought not to allow anyone to conclude, erroneously, that nationalism is a contingent, artificial, ideological invention ... (Gellner 1983, 55-56)

This is a crucial passage, but it is by no means an isolated one. Throughout his writings on nationalism, Ernest keeps returning to the idea that nationalisms frequently make use of the past, albeit very selectively. This reveals an ambivalence at the heart of his theory, one highlighted by the word 'admittedly' in the passage I have just quoted. It is this ambivalence that I wish to explore, because here, I believe, lies the main limitation of all 'modernist' theories of nationalism, including Ernest's. I want to examine this ambivalence under three headings: the parentage or genealogy of nations, the question of cultural continuity and transforma- tion, and the role of collective memory.

The genealogy of nations

As far as the genealogy of nations is concerned, Ernest is saying two things. Nations are navel-less, they don't have parents; and even if they did, it's irrelevant. Nations begin de novo, in a brave new industrial world.

One might start by asking which of these positions Ernest really claims. If some nations had navels, they had ancestors. We could then try to compare the navel-less, ancestor-less nations with the nations that had navels and ancestors, to see how each class of nations was faring. That is an interesting empirical question. But, if having ancestors is a priori irrelevant, then why should even some nationalisms make use of 'their' pasts? Note, it is not any past. For my nation, your past will not do. It has to be 'my' past, or pasts, or more usually, some of my pasts. But why return to the past at all? If the past is irrelevant to the needs of a modern society, then why does any nationalism bother to return to some sort of 'past'? Is this just a delusion, a matter of false consciousness? That is a position Ernest would, I believe, strongly deny, but he does not really explore the issue.

The other answer often given to this question is that elites, or people in general, have to return to tradition and ancestry to legitimate the new type of industrial-capitalist society and control the changes it must undergo. But that only begs the question as to why elites or people in general feel the need to refer back to 'their' ancestral traditions, or 'invent' ones that are aligned with these older traditions. Can it be because they are still quite powerful, and many people still operate in terms of these traditions, however irrelevant they may seem to some clites and to the theorist of modernity? In other words, many people appear not only to believe they have navels; they believe in the reality of the situation which gave them navels, and which their navels symbolise. In short, they believe they have collective parents, and these parents are in important ways relevant to their present situation (Matossian 1962; Hobsbawm and Ranger 1983, Introduc- tion and ch. 7).

This belief is not entirely unfounded. Historically, the members of a community can point to a considerable amount of evidence to support their belief in the genealogy of nations. They can refer to documents and artefacts which bear out their belief that many present-day industrial or industria- lising societies from England and France to Russia and Poland, from Japan and Korea to America and Mexico, are closely related to, indeed grew out of, past communities with which they identify. Members can point to the fact that, despite the many transformations they have undergone, their nations continue to share with past communities such features as a proper name, a rough territory, a language, some artistic styles, sets of myths and symbols, traditions of heroes and heroines, memories of golden ages, and the like. In other words, they conceive of their nation, despite all these changes, as 'stemming from' older communities of historic culture with whom they share myths of descent and common memories, including links with a homeland (cf. Johnson 1992).

We do not have to accept the ideology of nationalism itself, with its romantic belief in the awakening of the nation, its mission and destiny, to realise that we cannot fully grasp the rise and character of so many modern nations unless we explore their historical antecedents, and the continuing influence of those antecedents in the modern epoch. Ernest's modernism tells us how a modern nation operates, indeed must operate, in the modern, industrial age. But it cannot tell us which nations will emerge where, and why these nations rather than others.

To return to an example which Ernest used in the Warwick debate: modern Israel, he argued, is furnished with all the cultural equipment needed in the modern world: a literate, mass public education system, a common modernised language, a modern system of communications and legal system, in short, a 'high culture' of the kind required by the mobile, anonymous society which industrialism creates. In a state like Israel, where immigrants from over seventy lands and many cultures have been ingathered, this sort of standardised, public, unifying, 'high culture' is all the more necessary. To cope with the challenges of modernity, which is what any society must do if it is to survive today, you require a 'high culture'. In the modern world, the culture and religion of the past is at best irrelevant, at worst an impediment (cf. Friedmann 1967; Vital 1990, chs. 5-6).

But equally, in my view, this example demonstrates that we cannot hope to explain the rise and character of modern nations solely in terms of the requirements of modernity. Even that arch-modernist Theodor Herzl conceived of Israel as a haven for an ancient diaspora people, a Judenslaat, a state of and for unassimilable Jews, taking up where the last independent Jewish state, the Hasmonean state, had left off, in Zion. It was this ultimately religious and political vision, rather than the needs of modernity, that inspired and mobilised many diaspora Jews to become Zionists and take the arduous road to Palestine; and it was a vision that assumed a genealogy and an ancient pedigree and name for a nation-to-be, one that addressed, as does every nationalism, a designated and particular 'people'. To assert, with another modernist, Erie Hobsbawm, that there is simply no connection between the age-old Jewish yearnings and pilgrimages to Zion and the modern ingathering of Jewish exiles into Palestine, is to miss, not only the element of ethnic ascription, but also the whole aspect of popular motivation and collective self-understanding which is essential to the success of any nationalism. This is what I meant when I argued that modernism can tell us only half the story. It tells us in general why there have to be nations and nationalism in the modern world; it does not tell us what those nations will be, or where they will emerge, or. why so many people are prepared to die for them. Nor does it tell us much about the character of particular nationalisms, whom they address, and whether they are religious or secular, conservative or radical, civic or ethnic - issues that are vital both for the participants and their victims, and for a scholarly understanding of nations and nationalism (Hobsbawm 1990, ch. 1; Wistrich 1995).

I am suggesting, then, that to understand modern nations and nation- alism, we have to explore not only the processes and requirements of modernity, but also the genealogies of nations. In fact, we have to explore the impact of the processes of modernisation on those genealogies, and the way in which they give rise to selections and transformations by each generation of preexisting ethnic ties and of the ethnic traditions they have inherited.

Now, we may admit that in the case of the nations I have cited, it makes sense to explore their genealogies. But, what of modern nations that have lost their parents or never had them, or are not quite sure who their parents were? This poses considerable problems for nationalists attempting to create nations. It is certainly one reason for the enormous popularity of the Kalevala with the Finns, and the Kalevipoeg with the Estonians. (Yes, the Estonians did have a navel, after all. As a leading historian of Estonia, Toivo Raun, writing of the Estonian national revival of the 1860s, put it: 'Among the Estonian population, the importance of Kalevipoeg was not so much literary - it took decades to reach a wide audience - as it was symbolic, affirming the historical existence of the Estonian nation' (Raun 1987, 56, 76; cf. Branch 1985, Introduction).)

Both epics traced the descent of the Finns and Estonians to Iron Age culture-communities, and thereby provided these dispossessed and subject peoples with a sense of their dignity through native ancestry and an ancient and heroic ethnic past. In this way, they confirmed the worldwide belief in the virtues of national geneologies. To dismiss this by attributing it to the ubiquitous influence of nationalism again begs the question of why so many people have been mobilised on the basis of this particular belief in the genealogy of nations. Besides, nationalists have usually managed to find some historical antecedents for their nations-to-be, albeit often embellished and exaggerated, and this suggests that there are mechanisms at work which ensure some connection and even continuity between the modern nation and one or more pasts. To two of these mechanisms I now turn.

Cultural change and continuity

The first lies in the field of culture, and it provides us with a second focus for exploring the ambivalence in Ernest's and other versions of modernism.

For Ernest, modernity introduces a radical cultural break. This has two aspects. The first is underlined in the theorem which underpins his early formulation of modernism. That theorem states that in pre-modern societies, culture reinforces structure, whereas in modern societies, culture replaces structure. By this Ernest meant that kinship roles organised social life in simple, traditional societies, and symbols, myths, traditions and codes reinforced and expressed that kinship structure. Modern society, with the possible exception of bureaucracy, has no such structure. Instead, it has a common culture. In the polyglot, anonymous city, where most encounters are ephemeral, people can only relate to each other through context-free communication. This requires a common culture in preferably a common language. The precondition of membership in such a society and of citizenship in the state is literacy. Today, by necessity, 'we are all clerks' (Gellner 1964, ch. 7).

In a later article and in his Nations and Nationalism, Ernest focused more upon the changed nature of work and the generic training required for a niobile, industrial society. To train a mobile workforce and citizenry to inaster the techniques of semantic work, modern societies require a new kind of education system. For this, Ernest coined the term lexo-socialisa- tion'. In the old, agro-literate society, rote learning at one's mother's knee or in the village school sufficed. In a modern, industrialising or industrial society, external, state-imposed, standardised, mass schooling was needed to create the literate and technically sophisticated workforce, necessary to man the industrial machine. And the teachers, too, had to be specialised educational personnel, able to service the new literate 'high culture' which characterises and defines modern nations (Geliner 1973; 1982; 1983, ch. 3; cf. also 1994, ch. 3).

This concept of a 'high culture' became the key to Ernest's later theory of nationalism. In an interesting section of Nations and Nationalism, Ernest contrasts the 'high' culture of modern societies with the 'low' cultures of agro-literate societies. A 'high' culture, as we have seen, is a literate, sophisticated culture, serviced by specialised educational personnel and taught formally in mass, public, standardised and academy-supervised institutions of learning. It is a highly cultivated or 'garden' culture. A 'low' culture, by contrast, is wild, spontaneous, undirected and unsupervised. These are the cultures that readily spring up, unbidden, in societies where the great mass of the population are food-producers servicing the needs of tiny specialised elites - clerisies, aristocracies, merchants and the like - who are almost completely cut off socially and culturally from the peasant masses. In such a society, there is neither need nor room for nations and nationalisms, since the many 'low' cultures of the peasants are local and ,almost invisible'. Thus, in agro-literate societies, in Ernest's words: 'Culture tends to be branded either horizontally (by social caste), or vertically, to define very small local communities' (Gellner 1983, 16-17).

Now, for Ernest, all these 'low' cultures are doomed. They are cut off, like so many umbilical cords, because they are simply irrelevant in an impersonal, mobile modern society. If they are remembered at all, it is only through some symbols, in the same way that navels remind us of our origins. Nationalism, Ernest claims, is basically a product of modernity. It is, he says,

essentially, the general imposition of a high culture on society, where previously low Cultures had taken up the lives of the majority, and in some cases the totality, of the Population ... it is the establishment of an anonymous, impersonal society, with mutually substitutable atomized individuals, held together above all by a shared Culture of this kind, in place of a previous complex structure of local groups, sustained by folk cultures reproduced locally and idiosyncratically by the micro- groups themselves. That is what really happens. (Gellner 1983, 57)

Nothing could be clearer. The many, old 'low' cultures vanish. They are replaced by a single, new 'high' culture, or 'nation'. This is the true meaning of nationalism.

But there are two problems here, of which Ernest was well aware. Some 'low' cultures are not severed. Instead, they become 'high' cultures. The Finns and the Estonians clearly fall into this category, as do many of the cultures of the other smaller, subject peoples of Eastern Europe and the former Soviet Union. The other problem is that certain old elite cultures become 'high' cultures. The literary cultures of the Jews, the Armenians and the Greeks clearly fall under this heading, as do several of the cultures of Western peoples like the Catalans, Scots and French. Awareness of the difficulties posed for modernism by both these problems is an important source of its ambivalence (Gellner 1994, 37-44).

How do 'low' cultures become 'high' cultures? Why does Estonian win out over German, Swedish and Russian cultures in Estonia, and Finnish over Swedish and Russian cultures in Finland? Both these cultures were local, popular, largely confined to the peasants, at least at first. Why do these 'Ruritanians' become conscious of their local folk cultures and seek to turn what were 'low' cultures into 'high' ones for the nation-to-be?

Or were they really such 'low' cultures? And is the contrast between 'low' and 'high' cultures as sharp as Ernest alleges? In the case of Estonia, we know of Estonian language religious texts during the Reformation; and certainly by the seventeenth century, with the establishment of the University of Tartu and later Forselius' school system, the basis of a literate Estonian culture emerged a century and a half before the arrival of the Romantic movement in the Baltic states in the mid-nineteenth century. True, Germans and Swedes made the running, but a native Estonian poem of 1708 lamenting the miseries of the Great Northern War between Peter the Great and the Swedes, fought over Estonian lands, reveals a growing Estonian consciousness. Moreover, written Estonian can even be found in the thirteenth century Chronicle of Henry of Livonia recording the German conquest of Estonia in the face of much resistance. All this suggests that the transition to an Estonian 'high' culture was much more gradual and long- drawn-out than a modernist account would suggest (Raun 1987, 57, 76).

If this is the case with a so-called 'low' culture such as the Estonian, it is likely to prove even more true of old, literate, specialist-supported and therefore 'high' cultures like those of the French, the Arabs, the Jews and the Greeks. True, the languages and cultures of these peoples had to be 1modernised' to cope with modern conditions: they had to be simplified, standardised, secularised and expanded to cover all sorts of undreamt-of phenomena and novel concepts, and embrace all classes and regions of the nation-to-be. But the old 'high' caste-literate cultures were not scrapped and replaced; they were adapted, purified, enlarged and diffused, often through self-conscious cultural reformist movements. Sometimes, as in modern Greece, this involved a measure of compromise with its pasts, between a popular Byzantine Orthodox heritage and a classicising Athenian language and culture. In this case, the recovery of ancient Greek texts and sculptures did create considerable preoccupation with Periclean Athens among the C,reek-speaking intelligentsia, but it had constantly to compete with the more popular memories of Byzantium carried by an Orthodox liturgy and congregation (Frazee 1969; but cf. Kitromilides 1989).

What 1 am arguing here is that most modern languages and cultures are not 'invented': they are connected to, and often continuous with, much older cultures which the modernising nationalists adapt and standardise. By Ernest's criteria, many of these older languages and cultures were 'high' cultures. But, even where they were 'low' (or 'lower'), spontaneous, popular cultures, they could become the basis for a subsequent 'high culture'. Ernest hints at this when he speaks of Ruritanians in the metropolis of megalomania who, faced with the problems of labour migration and bureaucracy, soon come to understand the difference between dealing with a co-national, 'one understanding and sympathising with their culture, and someone hostile to it. This very concrete experience taught them to be aware of their culture, and to love it (or, indeed, to wish to be rid of it).' In other words, it is the old 'low' culture to which they cling, or not, as the case may be. And it is the old 'low' culture which, far from being cut off and thrown away, will soon become the modern 'high' taught culture, albeit for several hundred thousands or millions of people (Gellner 1983, 61).

There are many examples of this cultural connectedness and continuity amid change, and we need to remind ourselves that cultural continuity is not the same as cultural fixity. Take the realm of language development. The English and French languages evolved over many centuries, with several admixtures of other languages, yet we can trace lines of development which reveal their underlying continuity. Alternatively, there is a conscious reform of language and culture, as occurred with the Turkic languages under the impulse of the jadid educational movement of Ismail Bey Gasprinski, or with Hebrew through the modernising reforms of Eliezer Ben-Yehudah. In the latter case, the differences between biblical and modern Hebrew are considerable; yet modern Hebrew is clearly based upon, and developed from, biblical Hebrew (Zenkovsky 1953; Fishman 1968; Rickard 1974; Edwards 1985).

In terms of names, territorial attachments and myths of origin, too, there are striking connections and continuities, despite changes of cultural contents over time. This is especially true of island cultures like Japan, with its relative continuity of territory, identity and origin myths. But it can also be found in mixed cultures like that of Mexico, whose modern cultural nationalists have sought to recover and reappropriate some aspects of the pre-Colombian, mainly Aztec, past. Of course, it can be argued that the very need to recover the past is evidence for discontinuity. There certainly had been discontinuity, especially after Hernan Cortes' invasion. But, among the many indigenous ethnies of Mexico, the old cultures still live in varying degrees and guises, to be used as partial models and disseminated through the mass, public education system to the mestizo majority (Franco 1970; Lehmann 1982; Florescano 1992).

Collective memory and modern nations

This leads us directly to the final focus of modernist ambivalence, namely, the part played by collective memory in the formation of nations. Collective memories form another major link with an ethnic past or pasts. Ernest was very conscious of the role of memory in creating nations, if only because, like Renan, he emphasised the importance of national amnesia and getting one's history wrong for the maintenance of national solidarity. But there was no systematic attempt in his work to deal with the problems posed by shared memories of a collective past (Gellner 1982).

For Renan, memories were constitutive of the nation. The nation is built on shared memories of joy and suffering, and above all of collective sacrifices. Hence the importance of battles, defeats no less than victories, for mobilising and unifying ethnies and nations a1ll too evident in such sensitive areas of national conflict as Bosnia and Palestine (Renan 1882).

Memory, of course, can be easily manipulated. Witness the sudden surge of feeling over the mosque built on the temple of Ram at Ayodhya in India, or the post-war Israeli cult of Masada, a formerly obscure episode and half- forgotten fortress on the Dead Sea. Besides, we need to distinguish between genuine folk memories, and the more official, documented or excavated records of an often heroic past (Billig 1995, ch. 2).

Despite these caveats, shared historical memories play a vital role in modern nationalism. The question is: how far can the modernist theory of nationalism accommodate them? There are, 1 think, two problems here. The first is that the 'nation' which modernist theories of nationalism conceive as the object of explanation, is divested of 'identity'. It is either conflated with the state, to become the 'nation-state', or it is equated, as in Ernest's theory, with a modern 'high' culture, to become a more or less stable configuration of objective traits like language and customs in a large, anonymous, unmediated and co-cultural unit. Now the nation does, indeed, have some 'objective' attributes like a name, a demarcated territory and a common economy. But equally important are its more subjective properties such as a fund of distinctive myths and memories, as well as elements of a common mass culture. This means that we must take into account the perceptions, sentiments and activities of its members in the definition of national identity. The cultivation of shared memories constitutes a vital element of this nation-defining activity (Gellner 1964, ch. 7; cf. Grosby 1991).

The second reason why modernist theories give little space to the role of collective memories is their tendency to rely on purely structural explana- tions. With the exception of Benedict Anderson's analysis of the re- presentation of national images, most modernists trace the origins, rise and course of nations and nationalism to the consequences of (uneven) capitalism, industrialism, militarism, the bureaucratic state, or class conflict, or combinations of these. Where the role of ideas is also admitted, the origins of nationalism are ascribed to the influence of secularism, the 1Enlightenrnent and sometimes Romanticism. Only in this last movement is there any room for a consideration of the role of collective memory, but Romanticism is usually treated, if at all, as a secondary, even residual, explanatory factor (Nairn 1977, ch. 2; cf. Kedourie 1960).

I think that we can overcome these limitations and build into Ernest's framework a fuller account of the role of shared memories, if we marry his insistence that nationalisms create nations to the ethno-symbolic resources that they must use if they are to succeed. Take the vistas opened up by the emerging disciplines of archaeology and history. The excavations of Great Zimbabwe with its Elliptical Temple, of Teotihuacan on the central Mexican plateau, and of the tomb of Tutankhamun in Egypt, created no continuity between the modern nations of Zimbabwe, Mexico and Egypt and 'their' presumed ancient or medieval ethnic pasts. What they did was to suggest, in some cases establish, connections with distant and glorious periods, or 'golden ages', of communal history, thereby extending the collective self-imaginings and shared memories of their members back in time through a reconstructed past, and conferring a sense of dignity and authenticity on their citizens. It is modern citizens who need and reconstruct an heroic ethnic past; but once reconstructed, that past exerts its own power of definition through ancestry and shared, albeit taught, memory (Cham- berlin 1979; Ades 1989; Gershoni and Jankowski 1986; Smith 1995a).

The 'territorialisation of memory' provides another example of the power of shared rememberings. By this I mean the ways in which shared memories become attached to particular terrains, and over time forge delimited 'homelands'. The term 'homeland' suggests an ancestral territory, one which has become communalised through shared memories of collected experi- ences. The ancestral land is the place where, in the shared memories of its inhabitants, the great events that formed the nation took place; the place where the heroes, saints and sages of the community from which the nation later developed lived and worked, and the place where the forefathers and mothers are buried. This last element is particularly important. It ties each family to the homeland through memories of the last resting-places of their ancestors, and it sanctifies the homeland by creating its sacred sites and Popular pilgrimages (Smith 1986, ch. 8 and 1996).

Memory, then, is bound to place, a special place, a homeland. It is also crucial to identity. In fact, one might almost say: no memory, no identity; no identity, no nation. That is why nationalists must rediscover and appropriate shared memories of the past. Identification with a past is the key to creating the nation, because only by 'remembering the past' can a Collective identity come into being. The very act of remembering together, of commemorating some event or hero, creates a bond between citizens whose self-interest often brings them into conflict. Hence the constant need to reawaken public memories, to engage in commemorative rites and remembrance ceremonies, especially for those who gave their lives for the community; and to tie those memories to the homeland through daily routines and 'flagging' (Billig 1995, chs. 3-4).

Collective memories, then, are active components in the creation and reproduction of nations. Whether they are familial and unmediated, as often occurs in sub-Saharan Africa, or mediated and public, a construct of elites enacted in rites and ceremonies, and recalled in epics and chronicles, flags and anthems, shared memories are necessary for the formation of nations. States may be established without recourse to memory and remembering. But nations require shared memories to give their often heterogenous citizenry a common habitat, a source of pride and dignity, and a common destiny. Indeed, if we define the nation as a named human population sharing an historic territory, common myths and memories, a mass, public culture, a common economy and common legal rights and duties for all members, shared memories are required by definition. Without them, the subjective element, the sense of being part of a nation, would be absent. There 'could be no passionate identification by individual citizens with a particular 'nation', only a generalised calculating loyalty to the state (Mazrui 1985; Smith 1991, ch. 1; cf. Viroli 1995).

We can go further. If the modern nation is, in large part, a creation of nationalism, as Ernest argued, then there are three vital elements of the nation which, in my view, depend on the role of collective memory. The first is the drive for regeneration which is based on memories of a golden age, or golden ages. This is the idealised former age of great splendour, power and glory, intellectual or artistic creativity, or religiosity and sanctity. It is the age of the community's exemplars - its saints and sages, poets and heroes, artists and explorers - the ideal against which to measure the present, usually lamentable, state of the nation, and spur to emulation for successive generations. The memory of the golden age signifies the possibility and hope of national regeneration.

A second element is the sense of collective mission and national destiny. There is no nationalism, and few present-day nations, that do not proclaim some special mission and unique destiny. But a sense of collective mission presupposes shared memories of a past or pasts in which the nation was entrusted with that mission, and which shaped a unique community as the vehicle for the development and reproduction of 'irreplaceable culture values'. Similarly, a sense of national destiny presupposes a well remem- bered past, a history of a unique trajectory along which 'we' are destined to travel. Without such memories, without rituals of commemoration, the nation would have no distinctive task or future, and hence no raison d'itre (see Weber 1947; Smith 1992).

The third vital component is a sense of national authenticity, and this 400 is closely bound up with shared memories. What is or is not mine, ,What is or is not distinctive, representative, or original, is closely tied to questions of remembering and forgetting. What is 'inauthentic' is, in part, ,What is alien to popular consciousness and folk memory. What is original and 'ours' is that which has been hallowed by the shared memories of 'the people'. The acceptance of the Kalevala as Finland's national epic owed much to the survival and resonance among the peasants of the Kasrelian folk ballads on which Elias Lonnrot based his modern compilation, even if the ,memories' contained in that epic were less than historical (Branch 1985, Introduction).

Together, these nationalist concepts of regeneration and the golden age, mission and destiny, authenticity and folk culture, all presuppose the influence of shared memories of a collective past, however distorted or dimly remembered. And, since it is nationalism that largely creates the modem nation, the modern nation must be built on shared memories of some past or pasts which can mobilise and unite its members.

Conclusion

Nihil ex nihilo. Nothing comes from nothing. Ernest called himself a 'creationist', but he attributed the sudden birth of nationalist humanity to a process, the process of modernisation which, like the biblical creation process of 6,000 years ago, was sudden and discontinuous. And modernisa- tion keeps bringing nations into being, suddenly, explosively.

My view, on the contrary, Ernest termed 'evolutionist', indeed 'primordi- alist'. I hope I have made it clear that I, in no sense, subscribe to any of the forms of 'primordialism'. Nations are modern, as is nationalism, even when their members think they are very old and even when they are in part created out of pre-modern cultures and memories. They have not been there all the time. it is possible that something like modern nations emerged here and there in the ancient and medieval worlds. That is at least an open question, requiring more research. But, in general, nations are modern.

Can my position be called 'evolutionist' in opposition to Ernest's creationism'? Not in any strong sense of that term. There is too much discontinuity and change between pre-modern and modern communities to warrant the conclusion that modern nations are the product of slow, gradual, incremental growth from rude beginnings. But, in a weaker sense, there is considerable evidence that modern nations are connected with earlier ethnic categories and communities and created out of preexisting origin myths, ethnic cultures and shared memories; and that those nations with a vivid, widespread sense of an ethnic past, are likely to be more unified and distinctive than those which lack that sense (see Armstrong 1982; Smith 1986).

It is important to stress here that pre-modern ethnies are not nations, whether in Ernest's definition of the nation, or mine. They generally lack a clearly demarcated territory which their members occupy, equal legal rights and duties for all members, and a public, mass culture. What they do have, and what they bequeath, albeit selectively, to modern nations, is a fund of myths, symbols, values and shared memories, some distinctive customs and traditions, a general location, and sometimes a proper name. Without these shared memories and traditions, myths and symbols, the basis for creating a nation is tenuous and the task herculean.

Of course, there are exceptions to the rule. Some islands, like Trinidad or Mauritius, emptied of their original inhabitants, may gradually be forged, not without conflict, into unified and distinctive nations through the conscious creation or use of overarching myths and traditions, memories and symbols. The process of ethno-genesis, after all, goes on all the time, along with, and as part of, the creation of new nations. The same process may also be taking place in the former Italian province, and now independent state, of Eritrea with its two religions and nine language groups. Nevertheless, these exceptions only go to show that the widely accepted model of the unified and distinctive nation is derived from the many nations with a dominant ethnic past, and that, where such a past is lacking, the task of creating a modern nation - as opposed to a state - is very much harder (Cliffe 1989; Eriksen 1993).

This brings me to perhaps the most fundamental difference between my approach and that of Ernest Gellner. For Ernest, it is possible and desirable to have a general theory of nationalism, one that derives from the postulates of modernity. For myself, no such general theory is possible. Though I prefer a certain kind of approach, which may be termed 'ethno-symbolist', I feel that the differences between nationalisms across periods and continents are too great to be embraced by a single Euclidean theory. For such a theory can never tell us, as Ernest admitted, which are the nations-to-be and why they have this or that distinctive character and trajectory.

At the same time, such is the force and sweep of Ernest's own theory that nobody can fail to be convinced of the centrality and ubiquity of nations and nationalism for the world we live in. Ernest has reveafed the sociological foundations of our world of nations and shown us why nationalism must remain a vital and enduring force in the contemporary world. His originality consists in demonstrating why the link between culture and politics is so intrinsic to the modern world and why it must generate so much passion. As a result, Ernest was not among the many who foresaw an early supersession of nations and nationalism, although he was more optimistic about the diminution of its fires in affluent, democratic states. This is because he thought that the imperatives of industrialism and mass education would in the end override the power of shared memories of great events and ancient or recent antagonisms. Of this I am not so sure. The past cannot be swept away so easily.

So: to paraphrase Rousseau, a nation must have a navel, and if they have lot got one, we must start by inventing one. And it is because nations have na,vels, and because those navels, and the memories and traditions, myths and symbols they represent, mean so much to the people that have them, that we are so unlikely to see the early transcendence of nations and nationalism.