From: Gellner, Nations and Nationalism (Ithaca: Cornell University Press, 1983, pp.8-14)

One development which takes place during the agrarian epoch of human history is comparable in importance with the emergence of the state itself: the emergence of literacy and of a specialized clerical class or estate, a clerisy. Not all agrarian societies attain literacy: paraphrasing Hegel once again, we may say that at first none could read; then some could read; and eventually all can read. That, at any rate, seems to be the way in which literacy fits in with the three great ages of man. In the middle or agrarian age literacy appertains to some only. Some societies have it; and within the societies that do have it, it is always some, and never all, who can actually read.

The written word seems to enter history with the accountant and the tax collector: the earliest uses of the written sign seem often to be occasioned by the keeping of records. Once developed, however, the written word acquires other uses, legal, contractual, administrative. God himself eventually puts his covenant with humanity and his rules for the comportment of his creation in writing. Theology, legislation, litigation, administration, therapy: all engender a class of literate specialists, in alliance or more often in competition with freelance illiterate thaumaturges. In agrarian societies literacy brings forth a major chasm between the great and the little traditions (or cults). The doctrines and forms of organization of the clerisy of the great and literate cultures are highly variable, and the depth of the chasm between the great and little traditions may vary a great deal. So does the relationship of the clerisy to the state, and its own internal organization: it may be centralized or it may be loose, it may be hereditary or on the contrary constitute an open guild, and so forth.

Literacy, the establishment of a reasonably permanent and standardized script, means in effect the possibility of cultural and cognitive storage and centralization. The cognitive centralization and codification effected by a clerisy, and the political centralization which is the state, need not go hand in hand. Often they are rivals; sometimes one may capture the other; but more often, the Red and the Black, the specialists of violence and of faith, are indeed independently operating rivals, and their territories are often not coextensive.

Power and culture in the agro-literate polity

These two crucial and idiosyncratic forms of the division of labour - the centralizations of power and of culture/cognition - have profound and special implications for the typical social structure of the agro-literate polity. Their implications are best considered jointly, and they can be schematized as shown in figure l.

Figure 1 General form of the social structure of agrarian societies

In the characteristic agro-literate polity, the ruling class forms a small minority of the population, rigidly separate from the great majority of direct agricultural producers, or peasants. Generally speaking, its ideology exaggerates rather than underplays the inequality of classes and the degree of separation of the ruling stratum. This can in turn be sub-divided into a number of more specialized layers: warriors, priests, clerics, administrators, burghers. Some of these layers (for example, Christian clergy) may be non-hereditary and be re-selected in each generation, though recruitment may be closely predetermined by the other hereditary strata. The most important point, however, is this: both for the ruling stratum as a whole, and for the various sub-strata within it, there is great stress on cultural differentiation rather than on homogeneity. The more differentiated in style of all kinds the various strata are, the less friction and ambiguity there will be between them. The whole system favours horizontal lines of cultural cleavage, and it may invent and reinforce them when they are absent. Genetic and cultural differences are attributed to what were in fact merely strata differentiated by function, so as to fortify the differentiation, and endow it with authority and permanence. For instance, in early ninteenth-century Tunisia, the ruling stratum considered itself to be Turkish, though quite unable to speak that language, and in fact of very mixed ancestry and reinforced by recruits from below.

Below the horizontally stratified minority at the top, there is another world, that of the laterally separated petty communities of the lay members of the society. Here, once again, cultural differentiation is very marked, though the reasons are quite different. Small peasant communities generally live inward-turned lives, tied to the locality by economic need if not by political prescription. Even if the population of a given area starts from the same linguistic base-line - which very often is not the case - a kind of culture drift soon engenders dialectal and other differences. No-one, or almost no-one, has an interest in promoting cultural homogeneity at this social level. The state is interested in extracting taxes, maintaining the peace, and not much else, and has no interest in promoting lateral communication between its subject communities.

The clerisy may, it is true, have a measure of interest in imposing certain shared cultural norms. Some clerisies are contemptuous of and indifferent towards folk practices, while others, in the interest of monopolizing access to the sacred, to salvation, therapy and so forth, combat and actively denigrate folk culture and the freelance folk shamans who proliferate within it. But, within the general conditions prevailing in agro-literate polities, they can never really be successful. Such societies simply do not possess the means for making literacy near-universal and incorporating the broad masses of the population in a high culture, thus implementing the ideals of the clerisy. The most the clerisy can achieve is to ensure that its ideal is internalized as a valid but impracticable norm, to be respected or even revered, perhaps even aspired to in periodic outbursts of enthusiasm, but to be honoured more in the breach than in the observance in normal times.

But perhaps the central, most important fact about agro-literate society is this: almost everything in it militates against the definition of political units in terms of cultural boundaries.

In other words, had nationalism been invented in such a period its prospects of general acceptance would have been slender indeed. One might put it this way: of the two potential partners, culture and power, destined for each other according to nationalist theory, neither has much inclination for the other in the conditions prevailing in the agrarian age. Let us take each of them in turn.

Culture

Among the higher strata of agro-literate society it is clearly advantageous to stress, sharpen and accentuate the diacritical, differential, and monopolizable traits of the privileged groups. The tendency of liturgical languages to become distinct from the vernacular is very strong: it is as if literacy alone did not create enough of a barrier between cleric and layman, as if the chasm between them had to be deepened, by making the language not merely recorded in an inaccessible script, but also incomprehensible when articulated.

The establishment of horizontal cultural cleavages is not only attractive, in that it furthers the interests of the privileged and the power-holders; it is also feasible, and indeed easy. Thanks to the relative stability of agro-literate societies, sharp separations of the population into estates or castes or millets can be established and maintained without creating intolerable frictions. On the contrary, by externalizing, making absolute and underwriting inequalities, it fortifies them and makes them palatable, by endowing them with the aura of inevitability, permanence and naturalness. That which is inscribed into the nature of things and is perennial, is consequently not personally, individually offensive, not psychically intolerable.

By contrast, in an inherently mobile and unstable society the maintenance of these social dams, separating unequal levels, is intolerably difficult. The powerful currents of mobility are ever undermining them. Contrary to what Marxism has led people to expect, it is pre-industrial society which is addicted to horizontal differentiation within societies, whereas industrial society strengthens the boundaries between nations rather than those between classes.

The same tends to be true, in a different form, lower down on the social scale. Even there, preoccupation with horizontal, often subtle but locally important differentiations can be intense. But even if the local group is internally more or less homogeneous, it is most unlikely to link its own idisyncratic culture to any kind of political principle, to think in terms of a political legitimacy defined in a way which refers to the local culture. For a variety of obvious reasons, such a style of thinking is, in these conditions, most unnatural, and would indeed seem absurd to those concerned, were it explained to them. Local culture is almost invisible. The self-enclosed community tends to communicate in terms whose meaning can only be identified in context, in contrast to the relatively context-free scholasticism of the scribes. But the village patois (or shorthand or 'restricted code') has no normative or political pretensions; quite the reverse. The most it can do is identify the village of origin of anyone who opens his mounth at the local market.

In brief, cultures proliferate in this world, but its conditions do not generally encourage what might be called cultural imperialisms, the efforts of one culture or another to dominate and expand to fill out a political unit. Culture tends to be branded either horizontally (by social caste), or vertically, to define very small local communities. The factors determining political boundaries are totally distinct from those determining cultural limits. Clerisies sometimes endeavour to extend the zone of a culture, or rather, of the faith they codified for it; and states sometimes indulge in crusades, faith-endorsed aggression. But these are not the normal, pervasive conditions of agrarian society.

It is important to add that cultures in such a world proliferate in a very complex way: in many cases, it is far from clear how a given individual is to be assigned to his 'cultural background'. A Himalayan peasant, for instance, may be involved with priests and monks and shamans of several religions in different contexts at different times of the year; his caste, clan and language may link him to diverse units. The speakers of a given tribal language may, for instance, not be treated as members of it, if they happen to be of the wrong occupational caste. Life-style, occupation, language, ritual practice, may fail to be congruent. A family's economic and political survival may hinge, precisely, on the adroit manipulation and maintenance of these ambiguities, on keeping options and connections open. Its members may not have the slightest interest in, or taste for, an unambiguous, categorical self-characterization such as is nowadays associated with a putative nation, aspiring to internal homogeneity and external autonomy. In a traditional milieu an ideal of a single overriding and cultural identity makes little sense. Nepalese hill peasants often have links with a variety of religious rituals, and think in terms of caste, clan, or village (but not of nation) according to circumstance. It hardly matters whether homogeneity is preached or not. It can find little resonance.

The state in agrarian society

In these circumstances there is little incentive or opportunity for cultures to aspire to the kind of monochrome homogeneity and political pervasiveness and domination for which later, with the coming of the age of nationalism, they eventually strive. But how does the matter look from the viewpoint of the state, or, more generally, of the political unit?

Political units of the agrarian age vary enormously in size and kind. Roughly speaking, however, one can divide them into two species, or perhaps poles: local self-governing communities, and large empires. On the one hand, there are the city states, tribal segments, peasant communes and so forth, running their own affairs, with a fairly high political participation ratio (to adapt S. Andreski's useful phrase) and with only moderate inequality; and on the other, large territories controlled by a concentration of force at one point. A very characteristic political form is, of course, one which fuses these two principles: a central dominant authority co-exists with semi-autonomous local units.

The question which concerns us is whether, in our world, containing these types of unit, there are forces making for that fusion of culture and polity which is the essence of nationalism. The answer must be No. The local communities depend for their functioning on a good measure of face-to-face contact, and they cannot expand in size radically without transforming themselves out of all recognition. Hence these participatory communities seldom exhaust the culture of which they are part; they may have their local accent and customs, but these tend to be but variants of a wider inter-communicating culture containing many other similar communities. City states, for instance, seldom have a language of their own. No doubt the ancient Greeks were reasonably typical in this respect. While they possessed a vigorous awareness of their own shared culture and the contrast between it and that of all barbarians (with, incidentally, a rather low degree of horizontal cultural differentiation between Hellenes), this sense of unity had little political expression, even in aspiration, let alone in achievement. But when a pan-Hellenic polity was established under Macedonian leadership, it very rapidly grew into an empire transcending by far the bounds of Hellenism. In ancient Greece, chauvinistic though the Greeks were in their own way, there appears to have been no slogan equivalent to Ein Reich, Ein Volk, Ein Fuehrer.