Demography and the Origins of the Yugoslav Civil War

E. A. Hammel

Departments of Anthropology and Demography

University of California, Berkeley

Anthropology Today 9 (1): 4-9, Feb 1993 Royal Anthropological Institute of Great Britain and Ireland

My ultimate interest in this paper is to outline some of the deep historical circumstances leading to the current Yugoslav civil war. Having spent much of my professional life developing an empathy for the peoples of that former country, it is not a subject that I approach with equanimity or scholarly detachment, although I will here pretend to those virtues.[1] See map of Balkan regions. My subsidiary interest is to cast these circumstances in a framework that may contribute to the general topic of this session, the demography of cultural replacement. To those ends, I concentrate on migration and perceived ethnicity in the Balkans, and the crystallization of political blocks based on ethnic identification. I go on to speculate more widely about the importance of political and symbolic processes to demography.

The history of the Indo-European peoples is one of massive migration encompassed in the German term, Völkerwanderung. Impelled perhaps by population pressures behind them, they came into Europe and especially the Balkans as proto-Greek or Greek tribes about 4,000 or more years ago, followed by Italic and then by Celtic peoples perhaps a thousand years later, by Germanic tribes a few centuries after that, then by the Slavs around 700-500 AD, then by the Ottoman Turks around 1300 AD, and most recently by Germans who have advanced several times since the 17th century, only to retreat (but who may do so again). This general account obscures not only some extraordinary empirical features, such as the double rebound of the Goths, the wideranging depradations of the Vandals through Iberia and the Maghreb to the Italian peninsula, of the Vikings all over Europe including down to the Black Sea, and the general Drang nach Osten from the 9th Century onward, but also some varieties of migration I would like to distinguish theoretically.[2] That work, fundamentally in human geography, has been augmented mostly by linguistic analysis of dialect distributions.

Here I can only summarize. When the Slavs came into the Balkans between the 5th and 7th centuries AD the interior of the peninsula was already occupied by Illyrian peoples, and the littoral by Messapian and Italic speakers. The Illyrians had been under the influence of either the Eastern or the Western Roman Empires, and in some regions were strongly Latinized, especially along the Adriatic and in its immediate Dinaric hinterland. Just as the germanic Anglo-Saxons had called the Romanized Celts Welsch and their land Wales or Cornwall, or the germanic tribes someday to be French had called them Walloon, or the Goths had called the Dacians Wallachian, so the Goths of the Balkans and following them the Slavs called these Illyrians and other Latinized populations Vlach.[3] They were pushed by the Slavs into the uplands where they subsisted as shepherds, or into fortified Adriatic towns like Dubrovnik or Split, and the term Vlach came to mean in the Serbian Middle Ages generically a shepherd, or a "Roman" (namely a townsman speaking Vulgar Latin). But the secular movement of Slavs into Illyrian lands did not simply push the Illyrians up or or out. There was much cyclical migration from the crest of the Dinarics to the Adriatic on one side and to the great river plains of the interior and much chaotic movement both by Illyrians and Slavs. The Illyrians, who by the Middle Ages were identified (rightly or wrongly) as Albanians, were not eliminated but changed by this cultural chafing and were substantially slavicized by those Slavs who had moved into the uplands to be shepherds, themselves coming to be called Vlachs. The Vlachs were a distinct social and economic class. In the mediaeval law code of the Emperor Stefan Dusan, agricultural Serbian serfs were forbidden to marry Vlachs, which means they sometimes did. In the tax records of the monastery of Decani in the 14th Century, Slavic Vlach and Albanian pastoral villages (katuni) are both differentiated from villages of Slavic agricultural serfs, artisans, and fishermen. The social organization of the Slavic Vlach villages was intermediate between that of the Slavic agricultural serfs and that of the Albanians.[4] Some of the now Slavic tribes of Montenegro are almost surely of Albanian, thus of Illyrian origin. In much the same way, the Slavic populations flowed out onto the Peljesac peninsula north of Dubrovnik and thence into the Adriatic islands, slavicizing the Romanic inhabitants.

In this early migration an incoming, foreign group moved along the river valleys, pushing the previous residents into the uplands, followed them then into those uplands, and merged with them. This is cultural replacement by inundation, not just by ejection or elimination. It was assisted in a major way by the cyclical nature of transhumance and the chaos of ordinary life, bringing these people into continual, and not always inimical contact. Even in my own time, the Slavic inhabitants of the coast of Hercegovina south of Dubrovnik were bilingual in Slavic and Romance, and their knowledge of Vulgar Latin, now Italian, need not have been only recent, or dependent just on prior Venetian dominance; it is more likely it stretched into classical antiquity. They still engaged in regular, seasonal sea commerce with villagers on the Italian coast, and there was no indication that that this intermittent smuggling was recent either.[5] Similar examples of bilingualism can be found in the mechanics of transhumance to the Adriatic and the Aegean. Chaos was no small factor, either. In the classic area of Slavic pastoralism, the Montenegrin Alps (Crnogorska Brda), one can collect genealogies as deep as 14 generations, and the explanation of the founding of a lineage is more often than not, "Our people came here because of blood", meaning the founder had fled from a blood feud in the lowlands.[6]

It is important to understand that the Slavic movement along the major river valleys was at least by the 9th Century if not earlier a movement first in response to empire and then one of empire and not of simple tribal groups. The Slovenes and Croats had been pressed south and east by the Germans since Charlemagne, finally to be cut off from the Central Slavs by the Germanic Drang nach Osten and the Magyar penetration from the East. The Croats had their own distinct feudal organization by the 12th Century and were vassals either of the Hungarians or the Austrians then and thereafter. The Serbs were pushed north and west by the Bulgars and by the Byzantines from the 9th Century onward, then became important rivals to Byzantium from the 12th Century in the mediaeval Serbian Empire. Imperial pressure from the south and east intensified with the Ottoman penetration into Europe beginning in the late 13th Century. Major migrational streams were set in motion both by the German/Magyar pincers and the Turkish advance, and these were, in our typology, secular. The whole Slavic population of the southeastern Balkans was shifted decidedly northwest, some identifiably Serbian lineages settling as far north as Maribor close to the Austrian border. Yet there were important cyclical and random migrations imposed on these streams. In the 16th Century Turkish census of the Vlachs of Smederevo, a substantial proportion of villages were noted as abandoned, whether on account of military operations or patterns of slash and burn agriculture or long-cycle transhumance we do not know.

I focus mostly on the Turkish advance and its effects. Large segments of the Slavic population either evacuated the river valleys or were islamicized and remained. Almost all of the Albanian and the majority of the Bosnian population that did not flee adopted Islam. The cities and towns were almost exclusively Muslim, partly Turkish, partly Albanian, and partly Slavic. Only in Macedonia and Bosnia were there substantial numbers of Muslim Slav peasants. The plains of Metohija and Kosovo experienced a major influx of islamicized Albanians, and the Serbs began a long flight from that region that led directly to one of the major tensions of the current war. In most of Serbia itself the major river valleys were depopulated and claimed by almost impenetrable oak forests. The evacuating Christian population, in some cases Roman Catholic but mostly Orthodox, moved partly into the uplands as had the Illyrians before them, intensifying the slavicization of the Albanians, but also toward the borders of the Austrian and Hungarian lands. As the Turks moved northward over the two centuries after the battle of Kosovo, they pushed ahead of them a no-man's land called serhat, occupied by pastoral Slavic Vlachs, some Catholic but mostly Orthodox. A mirror image of this zone lay on the Austrian side, where it was called krajina in Serbo-Croatian or Militärgrenze in German, and in both zones the Vlachs came to serve in an organized way as military serfs. They moved chaotically, not only raiding their brethren across the border but rebelling against Turk or German and defecting from time to time to the other side. By the 16th Century this zone on the Austrian side stabilized in a reverse S-curve that ran north from Knin then east around the shoulder of northwest Bosnia, then north from the confluence of the Sava and Una Rivers to the Drava, protecting Croatia and the Venetian gulf from the Ottomans. It abandoned Slavonia to the Turks after the Hungarian defeat at Mohacs in 1526, and the quarter-million inhabitants of Slavonia came to be mostly Turks, islamicized Slavs, and Orthodox Vlachs. After the Turkish defeat before Vienna in 1683 and the subsequent reconquest of Slavonia, almost 200,000 of its quarter-million inhabitants fled the advancing Austrians and were replaced by a return flow of Catholic immigrants, mostly Croats, some Hungarians, some Germans. The Austrian advance drove almost to the field of Kosovo where the Serbs had been vanquished four centuries earlier, attempted to raise a Serb rebellion, failed, and fell back to the Sava. By some accounts as many as 80,000 out of perhaps three-quarters of a million Serbs fled with them, settling in eastern Slavonia and the Hungarian plain. These are not inconsiderable migrations.

The next major migrational process came with the spasmodic decay of the Turkish Empire. As it released its grip on Serbia, especially between 1800 and 1868, the upland pastoralists drained back down into the internal river valleys and the Adriatic littoral like meltwater in the Spring and seldom to the regions from which they had come. Cvijic notes that in the core area of Serbia (Sumadija), about 80% of the population in the early 20th Century was immigrant, and similar ratios prevailed throughout the peninsula. The upland Slavs were not the only ones to move as the Turks retreated. Catholic Albanians also flowed down the slopes. Deteriorating conditions in Kosovo, Metohija, and Macedonia under the crumbling Ottoman Empire led to substantial outmigration of many of the residents, especially Serbs who were being encouraged to resettle in Serbia by favorable homestead and tax policies in the newly emergent kingdom, and after 1878 there was a major flow of Muslim Albanians into the evacuated areas. After the Balkan Wars of 1912-13 and World War I there were substantial withdrawals of Slavic and other Muslims into Bulgaria and Turkey, and of course further to the south the population exchanges of Greek and Turkish populations are well known.

Thus we see that the Slavic population was in part first pressed into the uplands, imposing a secular change on the existing transhumant experience. It was also shifted to the north and west, where it was compressed and scrambled against the Germanic wall, then released both southeast and downhill as the Turks retreated, with reverse northward flows either in consequence of collapsed rebellion or, later, deteriorating economic and political conditions and attractive settlement policies in the new Serbian state.

What did this migrational process do to ethnic identification and its cultural symbols? The three primary elements of ethnic identification in the Balkans are kinship, language, and religion. They are not neatly related. Ethnic identification is summarized in labels, such as Serb, Croat, Muslim -- these especially in the context of the Civil War. Kinship is reckoned shallowly and rather bilaterally among most Muslims, deeply and more patrilineally among Orthdox, and something in between among Catholics. It was not difficult to record genealogies 14 generations deep in upland Serbia and Montenegro as recently as the 1960s but hard to do so more than 3 generations deep among Bosnian Muslims. Thus, the efficacy of kinship in defining ethnic groups varies. Especially among the Orthodox, consciousness of kinship ties to populations in areas of origin is strong.

Language is a tricky criterion. The Slavic speakers are sharply distinct from Albanians, Hungarians, Germans, Turks, Greeks and others. However, a large proportion of the local population in any area through which a major linguistic boundary runs are bilingual. Among the Slavic speakers, linguistic differentiation is gradual, in a dialect continuum from northwest to southeast; the Slavic languages have differentiated less than the Germanic or Romance languages, and mutual intelligibility is quite high. Only minute attention to dialect detail makes ethnic symbolization possible. This dialect continuum has been segmented by internationally imposed political boundaries and the centralizing efforts of core states, and the intellectuals of such states have sometimes been busy erecting linguisting boundaries to serve nationalist interests. In general, the linguistic divisions are based on the particular word we gloss as the interrogative pronoun, "what?" (sto, ca, kaj), and the rendering of the unstable "jat'" vowel of Late Common Slavic as "i", "ije" or "je", or "e". Even these isoglosses are not neatly distributed. Without going into detail, I note that the northwest shift of Slavs fleeing the Turks drove a wedge of Montenegrin and Herzegovinian dialect up through Bosnia into Croatia-Slavonia, separating speakers who had once formed a band running across Bosnia and Slavonia from the Adriatic to the Drava.

Religion is the most public and the most commonly invoked criterion of ethnicity. The religious history of the region is complicated. The Slavs were Christianized in the 10th Century by the efforts of the SS Cyril and Methodius, Macedonian monks who developed two alphabets for the translation of the Scripture into what we now know as Old Church Slavonic, a dialect very close to Old Common Slavic. They began their work with the central Slavs in Moravia, where German monks had failed before them. The essentially protestant nature of their linguistic efforts was a thorn in the side of Rome and a symbol of the emerging schism in the Church, which was formalized within little more than a century. The original alphabet, the Glagolithic, persisted in Dalmatia and Bosnia where it came to have a regional symbolic quality, replaced by Latinic in the Catholic church and by Cyrillic, the second of the two alphabets, in the Orthodox. In Dalmatia Glagolithic was used in the Protestant Reformation. The quality of political separatism evinced in the Reformation was manifested more exotically in the Bogumil heresy in Bosnia from about the 11th to the 15th Centuries. Bogumilism, a dualistic Manichean Christianity, originated in Bulgaria in the 10th Century in an Orthodox context and spread throughout the Balkans, and may have been ancestral (or at least fraternal) to the Albigensian heresy in France and the Paterene heresy in Italy as well. It was extirpated in mediaeval Serbia by Emperor Stefan Dusan who vacillated between Rome and Byzantium, finally accepting the latter, but flourished in Bosnia where it provided a neutral ground between the two contending churches that symbolized the Eastern and Western Roman Empires. It is claimed to have been popular first among the peasantry and then adopted by the Bosnian nobility. With the arrival of the Turks, all of the Bogumils seem to have converted to Islam, led perhaps by the nobility who were able to preserve their feudal privileges by becoming vassals of the Sultan. Islam, like Bogumilism, afforded a refuge from both contending Christian empires.[7]

These three dimensions of ethnicity: kinship, language, and religion, crosscut. Cvijic notes Catholic populations whose ancestry lies in Orthodox areas and who maintain kinship ties with families who are Orthodox. There are Catholic Serbs in Dubrovnik who celebrate that distinctively Orthodox feast of the household saint called the slava. The Catholic inhabitants of Konavlje south of Dubrovnik refer to the Orthodox as od stare ruke (of the old hand), suggesting their own prior membership in the faith.[8] There were at one time some Protestant Slavs, principally in Slovenia, to some extent in Dalmatia, but most of all in Slavonia under the Turks where there was no interference from Catholic bishops, but the Counter-Reformation erased them from the local religious map. There are Albanians of all three faiths: Orthodox, Catholic, and Muslim. Hungarians and Germans in the region are either Catholic or Protestant. Religion thus does not define ethnicity across major language divisions; no Catholic Croat claims common ethnicity with a Catholic Hungarian. On the other hand religion divides language communities into endogamous subsets, some of which are taken as identifiable ethnic groups. For example, Catholic and Muslim Albanians recognize that they are Albanians, but of different faiths. On the other hand, Catholic and Orthodox Slavs do not recognize common ethnicity; no Croat peasant claims co-ethnicity with Serb peasants, and neither with Muslim Slavs, even if they speak virtually identical dialects.[9]

Out of this kaleidoscope emerge the politically relevant ethnic groups that we see opposed in the Balkans today and quintessentially in Bosnia and Herzegovina. While the elements of kinship, language, and religion are the symbolic characteristics of ethnic membership, they fail to define the ethnic groups in any consistent or historical way. Croats, Serbs, and Muslim Slavs in Bosnia speak dialects that are only narrowly distinguishable. The dialect of the Bosnian Serbs is closer to that of most of the Croats of the region than it is to the Serbian of the core of Serbia. Similarly, the dialect of most Bosnian Croats is closer to that of the Serbs of the region than it is to that of the Croats of northern Dalmatia or the core of Croatia. The symbol that they use to differentiate themselves is religion, but religion fails in that task outside the region (for example with the Catholic Serbs of Dubrovnik).

These contending ethnic groups are clusters of symbols that are the detritus of imperial history and that are currently mobilized by political organizers to make impermeable boundaries where such boundaries did not exist before. It is not a new process. Under the Austrians, the Slavs of the Military Border were originally thought of all as Vlachs, whether Catholic or Orthodox; only later under their most Catholic majesties was there pressure to make the boundary of religion at least as wide as that of the State. Exactly that homogenizing instinct to achieve a congruence of political borders and symbolic qualities led Franjo Tudjman[10] and the Croatian Democratic Party to strip the krajina Serbs of their cultural distinctiveness and privileges, granted to them by the Communist Party as a way to prevent the rise of mini-nation states. The attempt to limit symbolic expression had the same result it did under Maria Theresa and Joseph II -- armed rebellion. This ethnic cleansing of a resident population is now met by ethnic cleansing of territories through the mechanisms of terror and expulsion as Serbia seeks to extend its boundaries to match the old Ottoman ones in Bosnia. Serbia is not above the first strategy, either, as it limits the use of Albanian in official and educational discourse, bent on completion of the slavicization of Illyrians that resulted in much of Montenegro.[11]

In this microcosm I see lessons for us in the interpretation of anthropological history. The first lesson is that imperial adventures make a difference. The process of ethnic cleansing begins when cultural and especially religious homogeneity is required to ensure political obedience. This first lesson is certainly obvious after the historical formation of the state, but may have been important before it, as well. The second lesson is that all of this was less of a problem before Johann Gottfried Herder and especially in the Balkans before Woodrow Wilson, in his pursuit of a particular imperialist adventure, namely the incapacitation of the German Reich, legitimized the ethnic nation-state and confused its creation with democracy. Democracy, free markets, ethnic self-determination, and general well being continue to be confused. What is missing in the historical picture since about 1800 is the idea of citizenship by right of residence rather than by right of blood or the adoption of the symbols of blood kinship.[12] What we see now as the cultural results of the migrational streams that are themselves started by imperialist adventures are the twin pressures of expulsion and conformity, the choice for populations to convert or flee.[13] It is hard to think of an anthropological subject during the last 6,000 years that was isolated from imperialist machinations, beginning with the emergence of city states in the Near East. The difference today is only in the level of the armament. The underlying processes are the same.

Let me try to push this kind of argument further, in a more general demographic sense than migrational process itself. One of the great puzzles of human history is the interrelationship between population growth and technological change, addressed by Gordon Childe, Ester Boserup, Thomas Malthus, Mark Cohen, and others. One view is that fortuitous improvements in subsistence through food production permitted population increase; here, the technological shift is thought to be exogenous. Another is that fotuitous population increase obliged human groups to intensify their exploitation of the environment; here, the population increase is thought to be exogenous. Deeper causes for the first (technological shift) are sometimes suggested in dessication or other environmental changes, but there are few suggestions of deeper cause for the second (population increase), unless they lie in Malthus' ideas of inherent excess fertility and improvidence ("the passion between the sexes").

I propose instead, in the spirit of this paper, a political reason that accords well with modern theories of pronatalism. It would be difficult to imagine populations of the Paleolithic that were not connected by trade networks and that were not in some sense hierarchical. Certainly by the European Epipaleolithic there is reason to suppose some lack of pure bilaterality in trade patterns, so that trade comes to look more like tribute or at least unequal, status-loaded exchange. Just as a thought experiment, I would like to push that notion further back in time and suggest, at the limits of credibility, that there are no truly and permanently egalitarian trade systems but that all are characterized by some level of profit-taking or, if you like, exploitation. Since the items traded early on were mostly extracted from the natural environment without elaboration, the provision of trade goods was labor-intensive. Children could easily have been important contributors to the family economy under such circumstances; it takes no stretch of the imagination to suppose them picking up amber on Baltic beaches or crawling into narrow holes in flint mines. Where the labor value of children is important to families, fertility is usually higher than it might otherwise be. Practices that might otherwise limit population (such as infanticide) would be expensive and might be suppressed. Thus population under such circumstances might increase more rapidly than otherwise. We can imagine such a scenario even under egalitarian conditions, as for example two washerwomen with helpful children might increase their fertility to take in more and more of eachother's washing. But I prefer the nonegalitarian scenario, as my example suggests.

This last speculative excursion is only meant to press us harder to recognize the fundamental qualities of political relations and the manipulation of behavior that goes into the maintenance of such relations. Status competition, fueled by trade, may have stimulated population growth and thus technological changes that permitted still higher population levels. The growth of political organization to accommodate these otherwise unmanageable aggregates has led us to the control of symbols, thoughts, identity and loyalty that lie at the roots of erupting conflict as imperial control decays. It is an old subject, from Wittfogel and before, but perhaps with a slight twist and a deeper reach. It also leads us to a solution in the Balkan crisis and those others that will surely erupt in the meltdown of the Cold War, one achieved only partially, haltingly, and sadly as yet incompletely in some Western countries, a commitment to equality before the law, regardless of one's symbolic shields.