III. PRE-MIGRATION KOSOVO AND METOHIJA

"What's there to say ... I had hoped that it would be resolved, that the Yugoslav authorities would do something... The years went by, and nothing... I lived, worked, loved, hoped for better days - for my children... I left behind a house and land for their safety. "

(Montenegrin mother of four, age 70)

1. THE POPULATION AND SOCIETY OF THE SOCIALIST AUTONOMOUS PROVINCE OF KOSOVO

The Socialist Autonomous Province of Kosovo is situated in the southwest of the Republic of Serbia, covering an area of 10,887 square kilometers (12.5% of the territory of Serbia). In 1981, one out of every four inhabitants of Serbia lived in Kosovo, and there were 145.4 inhabitants per square kilometer (169.7 by 1987), making this the most densely populated area in Serbia and Yugoslavia, quite improportionate to its own level of economic and social development.

In many ways, including its population growth tendencies, Kosovo is different from Serbia Proper and Serbia's other constituent Autonomous Province of Vojvodina. There were differences in social, economic and demographic development four decades ago, but in many respects they have grown rather than decreased, although the general tendencies of society's development, economic growth, urbanization, deagrarianization were the same. Nowhere else in Europe can one find such enormous social, demographic and cultural differences in so small a region, within a small population and the same state and political system. We shall briefly describe some of these differences before explaining the framework within which our respondents left Kosovo.

a) Economic Development

Kosovo is the poorest and least developed part of Serbia and Yugoslavia, although a good deal of money has been poured into it. For instance, Kosovo received 30-50% of the total investments made by the Federation's Fund for Under-Developed Regions and the International Bank for Reconstruction and Development. From 1955 to 1981 the social product grew at almost the same rate as that of Serbia, from 1,658 to 8,043 million (or 4.8 times), whereas in Serbia it grew from 28,661 to 248,203 million (or 5 times over). The rate of investments was always noticeably higher in Kosovo than in Serbia or Yugoslavia, from 30 to 200%.

On the other hand, Kosovo has the lowest social product per capita, from 1,969 to 5,042 million during the 1955-1981 period, as compared to 3,974 and 15 556 million for the Republic of Serbia as a whole, with growth indexes of 2.5% and 3.9% Although major progress has been made, the social standard in Kosovo still trails behind and this is tied in with its extremely high population growth which is paralyzing the effects of societal investments and economic development.

Kosovo's economy is undergoing intensive deagrarianization, which is strongly affecting the population and unemployment From 1953 to 1981 the agricultural population dropped from 0.68 to 0.38 million, and the share of active farmers in the total active population dropped from 76.6% to 23.5%, so that today it is noticeably lower than the figures for the active population of Serbia Proper (74,8%-38,l%). This places heavy pressure on employment - from 1955 to 1981 the unemployment figure rose from 2,426 to 71,571 or 28-fold (as compared to 17-fold in Serbia Proper) The employment figure rose from 42,808 to 189,248, or 4.4 times over (as compared to three-fold in Serbia Proper), increasing the number of unemployed to every 100 employed from 5.9 to 37.8. Along with the process of deagrarianization, further heavy pressure for jobs came from the very young age structure of the population flooding the job market This pressure would have been even greater were it not for the fact that Kosovo's female population has an extremely low level of employment, 14% of working-age women, as compared to 60.3% in Serbia Proper.

b) Population, and Urbanization

With its extremely high population growth, Kosovo went from being the least densely to the most densely populated region of Serbia.

A characteristic of Kosovo's is its very dense network of small communities. The average distance between them is only 2.9 kilometers, with an average population of 1,097 per community, and 754 in rural communities. This poses major obstacles to the existence and building of post offices, schools, clinics, shops and other elements of the infrastructure, especially in communities with up to 299 inhabitants, which in 1981 numbered 379 or one out of every four.

Kosovo has a total of 26 urban communities whose population in 1981 accounted for 32.5% of the total. The level of urbanization is the lowest in the country (in the Autonomous Province of Vojvodina the urban population accounted for 54.1% and in Serbia Proper 47.8%). This sluggish urbanization from only 14.6% in 1953 is due to the exceptionally high 61% growth in the rural population from 1953 to 1981, as compared to the rest of Serbia where the rural population steadily decreased An especially salient feature of urbanization in Kosovo is the low share of the population in mixed rural-urban communities; urbanization is still concentrated in Kosovo, as compared to other parts of Serbia where it is dispersed. Hence, the relative growth of the urban population is greatest here, and increased four-fold from 1953 to 1981 (as compared to double in other regions), with countless consequences for the way of life in cities and towns.

c) Population Growth

From 1921 to 1981 Yugoslavia's population increased from 12.5 million to 22.4 million (Table 1), giving a total growth index of 178.7.

During the same period, the population of the Republic of Serbia almost doubled, from 4.8 million to 9.3 million, giving a growth index of 193.2.

During the period under consideration, the population of Kosovo doubled twice: from 432,000 to 816,000 during the 1921-1953 period, and then again from 816,000 to 1.5 million during the 1953-1981 period. Abstracting the period that includes World War II (1931-1948, wherein the 131,8 index was higher than in any other part of Yugoslavia), the annual growth rate was in excess of 2% and at times climbed to more than 3%.

In the wake of these varied growth rates, the Kosovo population's share in the total population of the Republic of Serbia kept increasing, from 8.9 in 1921 and 11.1 in 1948 to 17.0% in the last 1981 census. At the same time, the growth of the Province's population steadily increased its share in the overall growth of the Republic of Serbia, from 13.2% during the 1921-1931 period, to 27.8% in 1948-1953, and 39.1% in 1971-1981. Kosovo Province accounts for only 12.3% of the area of the Republic of Serbia and, as we said earlier, with its population density (during the 1948-1981 period) increasing from 66.6 to 145.5 inhabitants per one square kilometer, it is (since 1971) the most densely populated region in Yugoslavia, which is in total contradiction with its level of economic development.

The reason for this population explosion, which is to be found today only in some non-European countries, is the natural increment and its steady rise in absolute terms (Graph 1).

In contrast to the very sharp fall of the natural increment in other parts of the Republic of Serbia and Yugoslavia, in Kosovo in 1985 it was almost double the 1950 figure: 42,099 versus 22,231. This increase is due in part to the steady decline in the number of deaths and in the death rate, and to the very strong increase in the expected average life span (which in 1950-1985 increased from 48.6 to 66.3 years of age for men, and from 45.2 to 70.3 years of age for women). But the decline in the death rate and the Province's own socio-economic development did not touch the other component, birth, because the number of births kept growing from year to year, keeping the birth rate at an exceptionally high level, despite a certain registered decrease.

Because the number of births and absolute natural increase steadily declined in other parts of the Republic of Serbia, and in Yugoslavia as a whole. Kosovo's share therein markedly increased. In 1953 and 1981 births in Kosovo accounted for 18.5% and 31.7% respectively of the total number of births in the Republic of Serbia, and the natural increment of the population accounted for 18.5% and 60.3% respectively for those years.

Regions with a high population increment, surrounded by regions with a lower population growth rate, are usually areas of emigration and the situation in Kosovo should have encouraged more emigration to other parts of the country. Although this did happen, the number of those who left was relatively small in comparison with the growth rate of the population. The difference between growth and the natural increment of the population in Kosovo from 1953 to 1981 was 150,000, while the natural increment was 856,000. A particular characteristic of the low migration balance figure and emigration from Kosovo is its ethnic selectiveness and the huge differences in the emigration rates of certain ethnic groups within its population. This difference alone would be enough to instill changes in the ethnic make-up of the region; coupled with differences in the natural population increment of ethnic groups it brought major changes that took an entirely different turn from trends in all other parts of the country.

Kosovo's population is very young, it has a progressive age structure, and is constantly being rejuvenated, thereby setting it apart from trends in other parts of Serbia. Children under the age of 14 account for more than 40% of the total population and from 1961 to 1981 the average age dropped from 25.6 to 24.2 years of age; in other regions the share of children is low and steadily declining, the average age is high and steadily increasing (in Serbia Proper 31.2 and 35.4 years of age).

This age structure, affects the size and growth of the working age population, and will only increase the economic difficulties of finding employment.

d) Certain Features of Society in Kosovo

Outstanding progress has been made since the war in literacy and education. Until liberation from Turkish rule, the population of Kosovo was almost completely illiterate, with a 95.5% illiteracy rate among people over the age of twelve in 1921. By 1953 this figure had dropped to 58.0% and in 1981 it declined to 17.6%, but it is still the highest in Serbia and in Yugoslavia, and is being maintained not only by the illiteracy of the small older population, but by the children dropping out of compulsory schooling. Differences between men and women are striking as well: in the case of men the literacy rate has been reduced from 38% to 9.4% and in the case of women from 72.1% to 26.4%.

The institutional development of secondary and higher education in Kosovo quickly increased the educated population from 20.6% (15 and older to 22.2%), which is almost equal to that of Vojvodina Province (where the illiterate population accounts for 5.8% of the total). High illiteracy on the one hand and the rapid growth of the educated population in Kosovo on the other constitute a typical example of the expansion of education in an under-developed community. In 1981 in Kosovo, there were 140 students to every 100 inhabitants between the ages of 20 and 29 (in Serbia there were 123 and in Yugoslavia 107 students to every 100 inhabitants). The differences in the literacy of men as compared to women are very pronounced.

The population in Kosovo is one of those rare cases where there have always been more men than women, not only because of the very young age structure but also because until 1966 the death rate among women was higher than among men, especially in early childhood and during childbearing age.

The lack of education among the female population in Kosovo is coupled with the very low level of its economic activity during employable age, from 18.2% in 1953 to 14.3% with a trend that strongly departs from the trend in Serbia (in Serbia Proper the level of activity among working-age women increased from 54.9% to 60.3%). The decline in activity among the female population of Kosovo is compounded by the departure of ethnic groups that have a higher level of activity among women (Serbs, Montenegrins, Yugoslavs, Croats) and the negligible level of this activity among ethnic Albanians. The Albanian female population has no tradition of being active in farming on the family land. Hence, in 1981, women accounted for 53.6% of the active farming population in Serbia Proper, but only 8.3% in Kosovo.

The social and family status of Albanian women (a group that largely determines the traits of the population of Kosovo) is still strongly influenced by tribal custom ordinaces, collectively known as the Canon of Lek. A purchase price is paid for the bride, a dowry (in other parts of Serbia the dowry came from the bride), the woman is totally subservient to the man and has no right to decide on her marriage. "An Albanian woman has no inheritance from her parents, from either the property or the home - the Canon considers women as a surplus in the house".*

* The Canon of Lek is a collection of custom ordinances compiled and codified by Shtjefan Gjeqovi, a Catholic priest /1874-1929/. It was first printed in 1941 in Italian in Rome. The first Serbo-Croatian edition came out in 1986 in Zagreb. The quote is taken from article 20 of the third volume, "On Marriages"

Hence, for instance, Kosovo is the only part of the country where when parents divorce the children are given into the custody of the father, because the mother has no material means of support outside the family. A husband who kills his wife for infidelity is not subject to punishment; if she kills somebody, retribution is taken against not her but the male members of her parents' family.

Since the only way for a woman to have the material means of support is in her parents' or husband's family, marriage is a universal phenomenon among the adult female population. Contrary to trends in Serbian society, in Kosovo common law and even polygamous marriage is on the rise. At the end of World War II, the birth rate for illegitimate children was the lowest in the country, 1.8%, but it soon grew to 12-13% and was the highest in Serbia. However, these are children born not out of wedlock but in common-law marriages; the norms of legal marriage greatly change the woman's status in relation to that of a common-law marriage. Therefore, illegitimate children are recognized by their fathers in 60% to 70% of the cases.

There is a strong generational continuity in families; in 1981 only 14.6% lived without their children (the children had left the parents), whereas in Vojvodina, Serbia's other constituent province, this figure was 30.6%. Life outside the family ("single households") is rare in Kosovo, accounting for 2.7% of the total number of households (as compared to 16.4% in Vojvodina), and the average size of the household increased from 1948-1981 from 6.35 to 6.92 members, yet another tendency that runs contrary to the general trend of households being scaled down to the family nucleus. In 1961 households with five and more members accounted for 65.4%, and in 1981 this figure was 71.3% (households with more than seven members accounted for 39.5% and 46.3% respectively).

As can be seen, despite the vast changes that took place in the economy, education and death rate of Kosovo's population they had little or no affect on the family. This disparity points to the strong, specific influence of ethno-cultural factors which in the family nullify the effects of socio-economic development.

e) The Ethnic Make-Up of the Population

The ethnic* make-up of Kosovo's population in the more distant past is known to us from historical documents; our information about the period between the two world wars, within the frameworks of Yugoslavia, comes from indirect appraisals and ethnological-historical studies.

* By ethnic group we mean a group with a common distant origin, history, language and culture, and an awareness of communality. This definition differs from the ethnological and sociological term, but is sufficiently broad to encompass all divisions in Yugoslav science, practice and the statistical system, which use designations such as nation, nationality, minority groups and ethnic groups.

From 1921 to 1931, Atanasije Urosevic* studied the origins of the population of Kosovo and Kosovo Pomoravlje. His research pointed up the immigrant origin of the Albanian population in these parts. He registered a total of 8,461 households (families), out of which various Albanian currents accounted for 5,806 or 68.6% of the population. He listed the tribes they belonged to and places they came from.

* Atanasije Urosevic, Etnicki procesi na Kosovu tokom turske vladavine, SANU, Belgrade 1987.

The ethnic composition of present-day Kosovo Province can be assessed only indirectly for the period between the two wars, through data concerning the inhabitants' mother tongue. The 1921 Census lists 439,000 people as having Albanian for their mother tongue; in 1931 this figure was already 505,000 for the whole country, which means 280,000 to 332,000 for the Province of Kosovo, or 64% and 59% respectively. The drop in their share, despite the tangible growth in their number, is due to the arrival of the Serbian and Montenegrin population - both the return of refugees from World War I and the spontaneous and organized settlement of this sparsely populated area (in 1921 there were 41.3 people to 1 square kilometer of land).

During World War II, Yugoslavia was carved up into the German, Italian, Hungarian and Bulgarian occupation zones and most of Kosovo came within the borders of the fascist "Greater Albania". Serbs and Montenegrins again fled under the threat of direct extermination. The other aspect of change in the ethnic composition was the steady incoming stream of settlers from Albania. We do not know the numbers involved.

We can get a better idea of the postwar changes in the ethnic make-up by looking at the Population Census.

It should be said that regardless of changes in terminology from census to census (nationality, ethnic membership, undecided), the actual content of the operational statistical definition did not change; ethnic membership is registered subjectively, by personally declaring membership. Hence, changes in these declarations are possible and in Yugoslavia frequent; the reasons can be rooted in different personal and social conditions.*

* It is a known fact, for instance, that a certain number of people of German and Hungarian extraction did not declare themselves as such in the first postwar census in 1948. Regulations concerning the emigration to Turkey of ethnic Turks led to countless changes in the declarations of Moslems and Albanians, resulting in a huge increase in the number of Turks from 1948-1953. The new classification of "Yugoslav - nationally undeclared" (introduced in 1961) increased its number of 1.2 million people of different ethnic origin by 1981. The number of Romanies changed illogically from one census to the next, in the wake of pronounced changes in declaration. A separate category of changes are those due to social pressures which were exerted in some areas (including Kosovo it seems) in connection with certain censuses.

The first thing to be said is that the ethnic composition of the population is heterogeneous and changed over the years. This change moved in the opposite direction of the general Yugoslav trend. Namely, whereas all other republics showed a pronounced tendency of reducing the share of the largest group and an increase in heterogeneity, in Kosovo, the share of the largest group grew.

This ethnic homogenization after the war has proven to be a salient feature of society in the Autonomous Province of Kosovo.

Albanians account for the largest group in the population, with an exceptionally high growth rate: from 0.48 to 1.2 million during the 1948 to 1981 period, thereby increasing their share in the population from 68.5% to 77.4% (equalling the share of settler Albanian houses registered by A. Urosevic and slightly higher than the figures given by the 1921 census). The second largest group is the Serbian, whose number slightly increased during the observed period of time (with an index of 121.8 as compared to 246.0 among the Albanians), but whose share diminished from 23.6 % to only 13.2%. Montenegrins were, at the beginning of this period, the third largest group, but their number decreased and share dropped from 3.9% to only 1.7% so that in the last census they were fewer in number than the Moslems and Romanies.

As for changes in the numerical size of other ethnic groups in Kosovo, pronounced variations can be observed among the Turks, Moslems and Romanies, which cannot be explained by differences in the natural increment. At work here, as in the case of "Yugoslavs", are changes in declarations and possible migration. Special mention here is due to the small group of Croats which registered a constant growth, but which was sharply reduced by stagnation during the past decade, something that is unproportionate to this group's natural increment.

A look at the changing numbers of the three groups of Albanians, Serbs and Montenegrins during the 1948-1961 period shows that changes in their structure followed general Yugoslav lines, with a slight drop in fee share of the largest group, the Albanians: their average annual growth rate during this period was 1.29, the Serbians' was 1.32 and the Montenegrins' as high as 2.23. It should be said that changes in the number of Albanians from 1948-1953 were also affected by changes in declaration, because this same period saw a demographically impossible increase in number of Turks, in the absence of which the growth rate of the Albanian population would have been higher.*

* Assuming that the rise in the number of Albanians was diminished by 30,000 people who opted for the Turkish group, the average growth rate would have been 213, approximately the same as the growth rate of the Montenegrin group, and its share in the structure would have remained virtually unchanged.

From 1961 onwards there was a clear tendency toward increasing the share of the largest group in the population, a process of ethnic homogeneity, at a time when the opposite process of spreading ethnic heterogeneity was gaining strength. in all of Yugoslavia's republics.

With regard to the changing numbers of some ethnic groups in the eighties, migrations played a very important part in some cases, along with the specific factor of changes inpersonal declarations.

Sudden variation resulting from changes in declarations of ethnic membership are easily noticeable in the numbers of Romanies, Turks and Moslems. These variations, with far less relative importance, are also carried over to the number of Albanians, especially as the above groups are of the Moslem religion. The sudden leap in the number of Turks in 1953 is attributed to the regulations that existed at the time regarding emigration to Turkey, to the revival of the historical emigration of the Moslem population in a Christian state during the 19th and 20th centuries; the tangible decline in their number in 1961 is attributed to emigration from the country. As for "Yugoslavs", they too declined in number; in Kosovo their share in the population is the lowest in the country, as compared to their rapid growth otherwise, especially in other parts of the Republics of Serbia, Croatia and Bosnia-Herzegovina.

Assessments of the factors in the changing numbers can be made for the last decade covered by the Census.

First of all, major differences can be observed in the natural increment and its affect on population growth: it is twice as high among Albanians as among Serbs and treble the figure for Montenegrins. These differences are due, above all, to different attitudes to reproduction, and among Serbs and Montenegrins (to a smaller extent) to deformations in the age structure because of the exodus of the young population.

However, based on natural increment, the number of both Serbs and Montenegrins would increase.

The growth in the number of Albanians is dominated by an extremely high natural increment, with an average rate of 3%, and a negligible negative migration balance. The natural increment of the Serbian group was smaller than the negative migration balance, and the low increment of the Montenegrin group was twice as low as the negative migration balance. Hence the drop in these two groups' size due predominantly to migration.

By linking up growth, natural increment and the migration balance one obtains an assessment of the other factors (changed declarations of ethnic membership and possible mistakes in registration). In all three ethnic groups they had a negative affect on the growth of the population which was relatively very small among Albanians, noticeable among Montenegrins and very strong among Serbs. However, it is hard to assume so many changes of declaration among the Serbs and Montenegrins and it is more likely that this is due to various mistakes in registering the number of people who moved from Kosovo to other parts of the country.*

* In registering migration the Census starts from the last place of residence; a number of migrants moving between the republics and provinces may be lost if they again move within the territory where they have immigrated. Our analysis would need data on the place of birth and current place of residence.

A look at the migration balance regarding resettlement (1971- 1981) by the population of Kosovo shows the following:

Therefore, in Kosovo migrations are ethnically very differentiated, emigration is the main course taken, but with major differences among the ethnic groups. If we couple the number of emigrants in 1971-1981 with the number of members of the corresponding group in 1971, we get an emigration coefficient of -5.6 to every 100 inhabitants for the entire population of Kosovo Province, only -0.9 for Albanians, -18.55 for Serbs, -22.4 for Montenegrins and -18.8 for members of all other groups together.

These systematic differences registered between 1961 and 1971 mean that research on migrations must focus on their ethnic component.

f) Ethnic Changes at the Local Level

In terms of the ethnic composition of the population and changes during the 1961-1981 period*, Kosovo's communes show different degrees of heterogeneity; the number and relative share of Albanians shows a general tendency of rising. The number of Serbs and Montenegrins, meanwhile, is, with rare exceptions, declining, and their share in the total population shows a general tendency of diminishing. In addition to differences in the natural increment and migrations across the borders of the province, there are also inner migrations, with Serbs and Montenegrins tending to gather in a small circle of communes where they would be greater in number, and Albanians moving to communes where Serbs and Montenegrins were expected to move out.

* Previous censuses used, for data on ethnic composition, the administrative units valid at the time, i. e. districts, and these were not later compared with the subsequent units used. Moreover, there are no figures for 1953 and 1948 per settlement so that regrouping (which was done for the 1961 figures) is not possible. However changes in the share of Serbs, Montenegrins and Albanians were negligible during the 1948-1961 period, and thus the figures for the 1961-1981 period can be used as the basis of observation.

Ethnically homogenous communes in 1961 included Glogovac, Decani and Kacanik, where Albanians accounted for 90% and more of the population, and Leposavic where 91% of the population was Serbian. These communes show the fewest changes in ethnic composition by 1981, along with a rise in the number and share of the Albanian population, and drop in the Serbian and Montenegrin population (see appendices 8, 9, 10). In the communes of Djakovica, Orahovac, Podujevo, Srbica and Suva Reka the Albanian group accounted for 80-89% of the population in 1961 and for 90% and more in 1981. This increased the number of ethnically homogenous Albanian-populated communes in Kosovo from three to eight, and to this number one can add Vucitrn and Klina, where the Albanian population increased from 70-72% in 1961 to 85-87% in 1981 (see graphs 2 and 3).

The communes of Vitina, Pec, Prizren, Urosevac and Lipljan had in 1961 a heterogeneous ethnic make-up, with Albanians accounting for 60-69% of the population. In each of these communes, the Albanian population grew to account for 70 per cent and more, the exception being Prizren, where from 1961-1981 not only did the number of Albanians double, but the number of Turks and Moslems trebled.

In 1961, the following communes had a heterogeneous ethnic make-up, with Albanians accounting for 50-59% of the population, and Serbs and Montenegrins for a high per cent of their own: Gnjilane (35.3% Serbs), Istok (26.9% Serbs and 11.3% Montenegrins), Kosovska Kamenica (40.5% Serbs) and Pristina (33.8% Serbs and 4.0% Montenegrins). In 1961 the lowest share of Albanians was in the commune of Titova Mitrovica, 49.8%, with Serbs accounting for 40.9% and Montenegrins for 3.8%. These communes provide, relatively speaking, the strongest examples of a decline in ethnic heterogeneity and increase in the share of the Albanian population (from 64.9% in Titova Mitrovica to 76.3% in Pristina).

Lastly, in 1961 the commune of Dragas had a population consisting of 44.5% Albanians and 56.4% Moslems; by 1981 these figures were 53.2% Albanians and 46.6% Moslems.

In all these Kosovo communes, the percentage of Serbs, and Montenegrins, declined. Let us examine the changes in actual numbers.

Serbs were numerically concentrated in the Kosovo communes of Pristina, Titova Mitrovica, Urosevac, Gnjilane, Kosovska Kamenica, Leposavic, Prizren and Vitina, with a total Serbian population of 169,840 (ranging from 10 to 34 thousand). In 1981 they numbered a total of 159,312, marking a drop of 6.2 %. On the other hand, we have communes where in 1961 there were less than a 1,000 Serbs: Glogovac, Decani, Dragas and Kacanik with a total of 1,915; in 1981 that number was down to 630, marking a decline of 66.7%. In between these two groups are communes whose Serbian population declined from 55,252 to 49,548, marking a drop of 11.4%.

In 1961 Montenegrins were numerically concentrated in the communes of Pec (12,701 or 33.8% of the total number of Montenegrins in Kosovo), Pristina, Titova Mitrovica, Istok, Decani and Djakovica, numbering a total of 28,692. In 1981 their number had dropped to 22,879, a decrease of 20.2%. The number of Montenegrins in other Kosovo communes amounted to 8,896 in 1961 and 4,131 in 1981, marking a decline of 53.6%.

It should be said that there were some exceptions to the general trend toward a decline in the number of Serbs and Montenegrins. In Pristina, for instance, their number increased during the 1961-1981 period, and in Prizren their numbered remained unchanged. The number of Serbs stayed the same in the commune of Lipljan, and increased in Orahovac; it increased in Gnjilane, Pec and Mitrovica from 1961-1971 but decreased from 1971-1981. Regardless of these changes, however, their relative share in each of these communes declined. Nonetheless, these deviations from the general rule allow us to assume less frequent migrations from and to these communes within Kosovo, due to both their number therein and the different time that disturbances in ethnic relations and forms of discrimination spread.

If we look at the communes where the number of Serbs increased or remained the same during the 1961-1981 period (i.e. Pristina, Orahovac, Gnjilane, Lipljan and Prizren), we see that in 1961 they were inhabited by 35.3% of the Serbs in Kosovo, and by 1981 this number was 42.5%. In the case of Montenegrins, such communes (Pristina, Prizren, Pec, Titova Mitrovica and Djakovica) accounted for 60.9% of this ethnic group in the Province in 1961, and 76.5% in 1981.

The lower the number and relative share of Serbs and Montenegrins in the population of the communes, the sharper and sooner the decline in their number. Along with the general drop in their number is a narrowing of their territorial dispersal in the Province, and an increase in the concentration of their increasingly small number in an ever-smaller number of communes.

Deviations in certain communes (for instance, Djakovica regarding Montenegrins, Orahovac in relation to Serbs) can usually be explained by the settlement patterns within the commune, meaning in a smaller number of settlements and with these two groups combining in number.

We shall examine changes in the settlements in terms of their sum total.

In 1961 there were 147 settlements without a single Albanian inhabitant; there were just as many in 1981, although they were not necessarily the same settlements. The majority of these settlements are in the communes of Leposavic, Titova Mitrovica and Kosovska Kamenica The next group consists of 37 settlements which numbered 604 Albanians in 1961 and not a single Albanian in 1981 These were usually Albanians living in Serbian settlements whose ethnic make-up did not change. In the 139 settlements whose Albanian population decreased, all are settlements showing a general decline in their number of (usually Albanian, and less often Serbian) inhabitants, or with a complete turnabout in the number of Albanians and Moslems during the 1961-1981 period. In all other settlements the Albanian population grew.

The number of settlements without a single Serbian inhabitant increased from 410 to 670. These are usually settlements with a small number of Serbs to start with, although in some cases the population was Serbian or predominantly Serbian in 1961 and Albanian in 1981. The number of Serbs increased in 200 out of the Province's 1,439 settlements (these are 14 municipal centers, while the other settlements are ethnically Serbian or predominantly Serbian), and decreased in 600 settlements. Some of them retained the same ethnic make-up (Serbian and predominantly Serbian) while others saw major changes with the growth in the number and share of Albanians.

From 1961-1981 Montenegrins disappeared from 243 settlements, which, combined with the 760 settlements that had no Montenegrin inhabitants in 1961, gives a total of 1,003 settlements without a single Montenegrin inhabitant. Usually these were individual settlers, but often enough their number was considerable and their departure changes the ethnic composition. In settlements (48) where the number of Montenegrins increased, their share in the population did not.

There are several conclusions one can draw from looking at how the population is distributed in these types of settlements. First of all, the circle of settlements with Serbs and Montenegrins narrowed down, while the number of settlements with an increase in the Albanian population grew.

Secondly, with the dispersion of the Albanian population goes the concentration of the Serbian and Montenegrin population in an ever-smaller number of settlements. This is especially evident among the Montenegrins; in the settlements showing an increased number (and appearance), only 14 % of the total number of Kosovo's Montenegrins lived there in 1961, but 66.9% lived there in 1981. This trend is also evident among the Serbs, whose number increased from 50.7 to 64.0% of their total number in Kosovo.

Thirdly, changes in the ethnic make-up swept all parts of Kosovo as well as its small units of communes and settlements. In the communes, the Albanization process was strongest there where the number and percentage of Serbs was smaller, but in the case of settlements the biggest changes in the ethnic make-up occurred in those which had an ethnically mixed population. This meant the spreading growth of settlements with a single ethnic population, especially there where the Albanian population was most numerous, but also in a number of settlements which had a Serbian and Serbo-Montenegrin population.

Even if these trends happened to a backdrop of relatively favorable ethnic relations, they would require more detailed study, wherein the "factors of attractions: would certainly be important. However, we shall see that the main reasons for moving out were related to Kosovo itself, to discrimination against, the Serbs and Montenegrins. In addition to the extremely high natural increment of the Albanian population (leading to ethnic homogenization through natural renewal), we have the exodus of the Serbs and Montenegrins. Their departure is conditioned by social relations in Kosovo, and we cannot but wonder whether the high birth rate and sudden drop in the death rate, along with major changes in economic development and overall cultural conditions, were not also influenced by society as a means to achieve certain ends.

2. EMIGRANTS AND THEIR FAMILIES

When examining the living conditions of emigrant families before they left Kosovo and Metohija, one must first determine their personal and family circumstances at the time. When linked up with the same circumstances of the Serbian and Montenegrin groups in Kosovo Province, and Kosovo's population at large, the picture acquires a broader social framework. Subsequent comparisons of pre-emigration circumstances and circumstances dating from the time under observation will show changes in the emigrant population brought on by time and change of environment.

Only then can we move on to the main point of the study, which is to gain insight into the emigrants' ethnic, social and political conditions of life and work. Unlike the first part of the analysis, which is based on factual information, this part will largely be based on spontaneous replies by the respondents, with their statements, assessments and views. Theoretically speaking, therefore, there is the risk of subjectivizing personal histories and circumstances in Kosovo, but this does not make it any the less reliable or valid a part of the research, because there is a strong connection with the factual part of the analysis. If there are cases where the respondents' point of view altered the picture of life in Kosovo, it did so primarily in one direction: in the direction of toning down the difficulties and pressures to which they were exposed.

a) Whence They Emigrated

A look at the territorial framework of life in Kosovo and in Metohija prior to emigration is meant not only to give a description of the population but also to depict some of the broader social relationships relevant to the emigration process. Furthermore, by comparing the population in Kosovo prior to emigration and the population of Serbs and Montenegrins in the Province as a whole, one can check whether and to what extent the sample we chose and questioned in Serbia Proper can be considered sufficiently representative of the Province as well.

We shall examine these issues in terms of the type of settlement as a general division of society, in terms of the ethnic make-up of the settlement of emigration as the immediate framework, and in terms of the ethnic make-up of the commune to which the settlement belongs.

Our respondents came from 170 settlements in Kosovo, out of which (according to the 1971 census) 140 were villages, 15 towns and 15 mixed rural-urban settlements. Half the households in the sample, i.e. 253, left villages, 200 left towns, and 47 left mixed settlements (which are not numerous in the Province). Comparisons with the situation in the Province can be made not via households but via the number of household members who emigrated: 1,123 left villages, 801 left towns and 185 left mixed settlements.

The Serbian, and especially the Montenegrin population was more urbanized than the total population of the Province. A comparison of the Serbian and Montenegrin population in the Province in 1971 according to the type of settlement and structure of the sample, shows a strong concordance in the representation of the rural and urban populations, with bigger differences in representation of the population from mixed urban-rural settlements. Therefore, the structure of the sample according to the type of settlement (in terms of the share of the urban and rural population) is markedly different from these same shares in Kosovo, but largely approaches the traits of the Serbian and Montenegrin groups.

The Province of Kosovo had 22 communes, only two of which had no one represented in the sample. The two communes are Dragas and Glogovac (the sample does include people born in Glogovac who, before emigrating from Kosovo, had moved to another commune), where the number of Serbs and Montenegrins was very small.

The majority of households and household members from the sample lived in the commune of Pristina (Table 6), followed by the communes of Podujevo, Gnjilane, Kosovska Kamenica and Urosevac, Vitina and Vucitrn. The smallest number of emigrant households (virtually isolated cases) came from the communes of Decani, Orahovac, Leposavic and Srbica. The numerical order of emigrant households and emigrant household members is almost the same, because the differences in the average size of the households were small, especially when observing only emigrant household members.

The ratio between the settlement and number of Serbs and Montenegrins in Kosovo on the one hand, and their representation in the sample on the other, is examined by first assuming that the number of emigrants from the communes in the sample is primarily conditioned by the number of Serbs and Montenegrins in the total population of the commune in 1971. In that case, the territorial distribution of Serbs and Montenegrins in the Province would be balanced, or with minor differences transposed to the territorial distribution of the emigrants. This, to all intents and purposes, would mean that the conditions and factors prompting or compelling emigration existed equally in Kosovo's communes, regardless of special circumstances. (Graph 4).

According to the number of Serbs and Montenegrins registered in 1971, the communes can be divided into three groups: less than 10,000 members of the Serbian and Montenegrin nationality, between 10,000-19,000 and 20,000 and over. In this third group are five communes (Gnjilane, Urosevac, Titova Mitrovica, Pec and Pristina, with the most developed centers in the province). There were five communes numbering 10,000 to 19,000 Serbs and Montenegrins (Kosovska Kamenica, Leposavic, Lipljan, Prizren and Istok), while in the remaining 12 communes their number ranged from around 100 to 9,999.

Table 7 shows that the distribution of Serbs and Montenegrins by groups of communes according to their number differs appreciably from the population distribution of the emigrant household members. Relatively speaking, the smallest difference between Kosovo and the sample is in the communes that have the largest number of Serbs and Montenegrins, 42.1'% and 47.1% respectively. There is a bigger difference along these same lines in the group of communes with a Serbian and Montenegrin population of 10,000 to 19,000, where they accounted for 33.8% of their compatriots in Kosovo but only 26.3% of the sample. However, the group of communes with the smallest number of Serbs and Montenegrins had them accounting for 19.1% of their total number, whereas in the sample they accounted for 31.5%

Therefore, the starting premise is incorrect, because the above ratios belie it.

However, let us check this premise one more time by dividing the communes according to the percentual share of Serbs and Montenegrins in the total population. Although the number and share are interconnected, they are not the same: 48,000 Serbs and Montenegrins in the commune of Pristina accounted for 31% of its population, while 16,000 in Kosovska Kamenica accounted for 36% of that commune's population. The communes are divided into four groups according to the share of Serbs and Montenegrins together in the commune's total population in 1971. The first group consists of 9 communes (Glogovac, Decani, Dragas, Djakovica, Kacanik, Orahovac, Podujevo, Srbica, Suva Reka), with the lowest share, up to 9.9%; the second group has only two communes with a share of 10% to 19% (Vucitrn and Prizren); the third group numbers 7 communes (Vitina, Istok, Klina, Lipljan, Pec, Pristina and Urosevac); the fourth group, with the remaining 4 communes, is where the share of the Serbs and Montenegrins in the total population was 30% and more.

Here again, the premise proved incorrect, because the lower the share of the Serbs and Montenegrins in the commune's population, the more emigrant members there were in the sample, with extremely marked differences at the ends of the pole - 8.8% of Kosovo's Serbs and Montenegrins lived in the group of communes where their share in the population was the lowest, whereas in the sample they accounted for double that figure, 16.9%; half the total number of Serbs and Montenegrins in the Province lived in the four communes where they accounted for 30% and more of the population, whereas their share in the sample was less than that, i.e. 44.3%. The differences increase even more when one compares the distribution of Serbs and Montenegrins in Kosovo with the number of emigrant households, where the sample shows a very large number of families that moved out of communes where the number and share of Serbs and Montenegrins was the lowest. The larger number represented in the sample of people who left communes with a small number and low percentage of Serbs and Montenegrins is linked to the fact that during the period from 1961 to 1981 the number of Serbs and Montenegrins in these communes sharply declined.

This points to the following conclusion, which now serves as the premise for further analysis: the conditions and causes of emigration were not the same or did not have the same effect in every Kosovo commune. The negative proportion between numerical size and percentage in the Kosovo population and in the sample indicates that the number of Serbs and Montenegrins and the ethnic composition of the communes were factors, in terms of both population and society, that influenced social relations and the process of emigration.

The negative ratio between the number and percentage of Serbs and Montenegrins in the population of Kosovo's communes, and in the sample, is determined by the size of the Serbian group, as numerically the largest. However, this ratio also appears when examining Montenegrin emigrants: the fewer Montenegrins there were in the group of communes, the larger the percentage of emigrants they represented from these communes in the sample. Consequently, we have confirmation of the conclusion. It is further confirmed by another aspect of these migrations: the connection between the ethnic composition of the settlement and emigration.

The numerical predominance in the sample of families that emigrated from largely Albanian settlements, and the changes in the ethnic make-up of the settlement of emigration as registered in the population census, both say the same thing, i.e. that there is a strong trend toward emigration from settlements where Serbs and Montenegrins were smaller in number, and that most often Serbs and Montenegrins left settlements where they accounted for a minority, up to 30% of the population. Here again, numerical size proved to be an important factor of emigration, as did the ethnic composition of the settlement, which shows a major connection with living conditions in Kosovo and the conditions and causes of emigration.

b) The Households of the Emigrants

The following proved to be the most important aspects of the emigrants' households: size, family composition, possession of property and socio-professional structure.

The number of people in the household, i.e. its size, directly depends on two things: the family make-up and the number of children in the family or the marriage.

Prior to emigrating, the households numbered 2,521 members, or an average of 5.0, which was far below the average for Kosovo (6.6 in 1971 and 6.9 in 1981), but noticeably higher than the average for Serbia Proper (3.8 in 1971 and 3.4 in 1981). Even if the separation and break-up of households had nothing to do with emigration, their average size (of only emigrant members) would be bigger than in Serbia because of the bigger number of children in the family and broader family composition (excluding people who live outside of the family, in 1981 the average size of "single households" in Serbia Proper was 3.8 members). Had the households not broken up and separated before emigrating, together with the number of members in Serbia they would have averaged 6.2 members.

The bigger the household in Kosovo, the more often one encounters both categories, emigrants and those who remain behind: in households numbering 4 and less people, 12.1% had members who had emigrated, in households with 5-7 members, that figure was 34.5 % and in households with 8 members or more, it was 72%. Consequently, the separation and break-up of households most often occurred in large, especially multi-family households.

The family composition of the household is an extremely important factor that largely determines relations among the members of the household and everyday life. The predominant type nowadays is the single family household whose young grow up and leave to start their own families, breaking the generational continuity of the traditional family and household.

Prior to emigrating, half the households were actually one-family households consisting of parents and children, with only four cases where there was a single parent and children. There were 34 households where there were married couples, and only 14 single ones, all cases where young people had recently left their parental homes just before emigrating. There were 106 two-family households, usually the families of the father and the married son, and 61 of the households consisted of the extended family (usually one of the parents in the family of the married son), while 29 households were made up of three and more inner families.

This make-up differed noticeably from the situation in Kosovo. Excluding single households, one-family households accounted for 71.7% of the sample, and, in 1971, for 66.6% of the population of Kosovo, whereas three-family and bigger households accounted for 6.5% and 12% respectively. Obviously, the family composition of the emigrants was more modem than the social environment they left, given the social differences between the Serbian and Montenegrin and the Albanian population of the Province, conditioned by a lower degree of agrarianism and a higher degree of education and different cultural traditions.

The socio-professional composition of the households based on the occupations of its active members shows that half the households consisted purely of blue collar workers, and 100 were mixed, with some of the members being workers; 81 were made up of white collar workers, and 61 were mixed with some members being white collar workers; there were 74 farming households, and 47 mixed households where some of the active members were farmers. Thus, three-quarters of the polled households were non-agricultural before emigration, and one-quarter were agricultural. The de-agrarianization of households prior to emigration was higher than not only that in Kosovo but than that in Serbia Proper.

Of the mixed households, 36.1% had only one family, for white collar households this figure was 65.4%, worker and worker-white collar households 63.4%, and purely farming 60.8%. Abstracting the mixed households, the differences between the family composition of farming and non-fanning households are very small, indicating that many family traits (such as the number of children and tri-generational composition) were equally present in all the social strata of our sample. The traditional way of life was not the only factor contributing to the high percentage of multifamily non-farming households. Take housing conditions, for instance: socially-owned apartments are few and far between in Kosovo and hard to come by; the only way for a multifamily household to split up was to build one's own house or buy an apartment, but the social climate in Kosovo was not in the least conducive to such investments, and families that could and would build decided to do so outside of Kosovo. The other interpretation regarding the number of multifamily households can also be attributed to the social climate: households numbering several men, according to many of the respondents, felt safer and more protected. We shall return to this later.

Property is not only an economic indicator of the household but also a social sign of roots in the environment and inclusion in the relations of the settlement.

The poll shows that only 80 households did not own their living space, and out of that number 11 rented apartments (single people or young married couples that had separated from their parental families shortly before emigrating) while the remainder lived in socially-owned apartments, half of them in the capital of the Province. Only 6 households owned their own apartments (again in Pristina), and 414 lived in family houses.

Not counting the homestead itself, two-thirds or 339 households owned arable land, usually ranging from 2 to 5 hectares in size, but there were 58 households which owned 8 hectares and more. The average size of the holding was 4.9 hectares, i.e. 1,670 hectares of arable land in toto.

There are fewer farming and mixed-farming households in the sample than those that own arable land. For instance, 220 non-farming households owned land, probably a family inheritance, indicating their agricultural origin and a possible extra source of income. In the group of blue collar households, 74.7% owned land, in the group of white-and-blue collar households that figure was 54.6% and in the group of white-collar households it was 47% (Table 8). In purely farming households, 93.3% owned land; in two cases the households consisted of landless farmers, and in three cases all the land had been confiscated or bought up for other purposes, such as for the building of facilities. Instances of partial confiscation or purchasing of land were more numerous. Especially noteworthy is the fact that among land-owning households the smallest average holdings were to be found among purely farming households, just 3.9 hectares as compared to 4.4 for blue collar, 5.2 for blue-and-white-collar and 8.2 hectares for farming-blue-collar households. One might be prone to say that the size of the property influenced the social mobility of the descendents in the farming family, as well as their schooling, and that the majority of non-farming heads of family trace their roots to farming families.

As can be seen from Table 8, the sources of income are quite different from the occupations of the active members of the household: in terms of occupation, 14.8% were fanning, 9.8% were mixed and 75.8% were non-farming households, whereas in terms of the sources of income, 14.8% were farming, 54% mixed and only 31.2% non-farming households, (assuming, of course, that the property brought in an income).

In trying to assess the overall importance of the professional and social composition of the analyzed households before emigration, we can say that these are households that are strongly affected by the economic and social transformation of the Autonomous Province of Kosovo. However, although their professional structure is varied and primarily non-agricultural, these households remain tied to the land and enjoyed the advantages offered in Yugoslavia by a mixed economy. With few exceptions, these households owned the roofs over their heads, which, at a time when there is a housing crisis, constitutes a major economic and social advantage. Linking up these economic facts with those about the family make-up and size of the household, we can say that this is a relatively stable group of households which was able to effectively combine the effects of economic and social transformation with the advantages of their previous situation and inheritance.

c) The Members of the Household

In this study, the traits of the household members can be divided into those of general significance (such as age, sex, education, employment) and those of specific significance in relation to this research (such as nationality, origin and knowledge of the Albanian language). What follows is an analysis of both groups of traits, relating to both the emigrant and non-emigrant members of households that moved.

In 1971 the average age of the household members in the sample was 24.4 for those who had emigrated, and 26.4 for those who had not. The average age of the total population of Kosovo was 23.1, as compared to Serbia Proper where it was 36.4. The sample taken from the poll was on the average, older than the population of the environment they had left, but younger than the population of the environment they were to join.

The economic activity of the households in Kosovo prior to emigration and their division into emigrant and non-emigrant parts, was at a considerably higher level than it was for Kosovo's overall population (with 34.7% active in 1961, 28 % in 1971 and only 22.6% in 1981.) This can be attributed to the above-mentioned age profile, and to the higher level of activity by Serbian and Montenegrin women than that of the overall female population of Kosovo.

With regard to the occupations of the active population, blue collar workers accounted for more than half of the sample, although they accounted for only 26.4% of the total active population of Kosovo in 1971. White collar workers accounted for 31.7 % of the sample, as compared to 14.4% of Kosovo's population. Unfortunately, there are no data on nationality vis-a-vis occupation, which would have allowed us to see the extent to which the composition of the sample corresponds with that of the Serbian and Montenegrin group in Kosovo, and to check the hypothesis that certain segments of the social stratification were not equally swept up by the process of emigration. The low share of farmers in the sample could be due to the deagrarianization of the Serbian and particularly Montenegrin population, or to the fact that farmers, especially from Serbian or mostly Serbian settlements, found it harder to pick-up and leave. It should be said that later analyses will show that farmers in mixed and predominantly Albanian settlements were exposed to the harshest forms of pressure to move out.

Characteristic differences emerge from a comparison of those members who emigrated and those who did not. Members who did not emigrate with the household were not only older on the average, but also included a markedly higher share of farmers. This is understandable because it was the older generation that largely stayed behind while young members and children were the ones who left.

Out of the 848 active non-farmers among the household members in Kosovo, 539 or 65.5% were employed before emigration, and one third of the labor force was unemployed. This means that to every 100 members employed there were 70 looking for jobs, which is a much higher degree of unemployment than in The total population of Kosovo in 1971 (where out of 107 thousand employed, 22 thousand were unemployed, meaning there were 21 unemployed to every 100 employed), and even in 1981 (72 thousand unemployed to 188 thousand employed, or 38 unemployed to every 100 employed).

In other words, before emigration the households had an abundance of members looking for jobs, and they were mostly the ones who would later emigrate. The situation was not the same in every household, because 145 did not have a single employed member, 208 had only one and 147 had several members employed.

Before emigration, 382 people from 243 households were looking for work. Out of that figure 108 people from 71 households found jobs and 274 did not, although it is not unusual to wait eight to ten years for a Job. Moreover, noticeably more jobs were found during the 1965-1975 period than a decade later. Added to this is the fact that many of the Jobless did not even try to find work in Kosovo because they knew that they would not get it, especially the young who had completed their schooling close to the date of the household s emigration.

Hence, job-hunting emerges as a frequent reason for emigrating, and at first glance one might think that unemployment, as a classical reason for migration, was an instrumental factor here as well. But, the fact that the jobless figure in the households from the sample was higher than in the total population of Kosovo, that the difficulties of finding a job were largely connected with nationality, that many of the emigrants quit jobs they had, and that there was the whole combination of ethnic relations in Kosovo at both the general level and within firms (all of which compelled people to emigrate), indicates that one must not be hasty with one's conclusions.

The nationality of the household members was, of course, defined by the objective of the research and by the sample itself, but among the Serbs and Montenegrins were a certain number of other nationalities, mostly wives of Croatian, Macedonian, Moslem, and Yugoslav nationality as well as one Albanian. It should be added that several households were of Romany origin, but thought of themselves as Serbs and declared themselves accordingly.

There were 1, 763 Serbs among the emigrants, 350 Serbs among those who did not emigrate with the household, 346 Montenegrins in the I former and 52 Montenegrins in the latter category. The other nationalities I viewed together had 10 people among the emigrants, and none among the non-emigrant members of the households.

Montenegrins accounted for 15.8% of the sample, in other words there were 18.8 Montenegrins to every 100 Serbs. This ratio is bigger than it is in the total population of Kosovo: in 1961 it was 16.6 and in 1981 only 12.9 Clearly, the Montenegrin, part of Kosovo's population was relatively harder hit by emigration than the Serbian.

The homeland of the emigrating population is very important in any migration, especially when ethnic factors and conditions are also involved.

Prior to emigrating, Kosovo-born members of the households accounted for 85.6%, and settlers for 14.4%. These are lower figures than for the population of Kosovo in 1971, when there were only 4.3% that had moved there from other parts of the country.

In our sample, 61.3% of the household members were born in the place of emigration, whereas in the overall Kosovo population in 1971 they accounted for 79.6% (largely because of the high percentage of children). Hence, the sample had a higher overall degree of territorial mobility and settlement than the total population of Kosovo, and also a higher degree of mobility within Kosovo Province. While 8% of the household members had moved within their commune of birth, and 16.3% within the communes of Kosovo, in the total population of Kosovo Province these figures amounted to 4.7% and 11.4% respectively. The total Serbian and Montenegrin population in Kosovo again had a higher degree of territorial mobility. In 1971, 63.8% of the Serbian group and only 38.1 % of the Montenegrin group lived in their place of birth, as compared to the afore-mentioned 79.6% of the total population of Kosovo Province.

The homeland of the household can be established via the head of the household, with 415 Kosovo-born and 17% having come to settle there, which is almost the same degree as when looking at the total number of household members. More than half of the non-Kosovo-born household heads came from Serbia Proper, and a fifth from Montenegro. The most frequent reason for coming was that their parents had moved there, they had found jobs there or been transferred, they had bought land, and in 14 cases they had moved there after World War I. Fifty-four household heads had moved to Kosovo in the period between World Wars I and II, 31 had come after World War II, accounting for two-thirds and one-third of the total number of settlers respectively. This means that the average life span for non-Kosovo born household heads who moved there between the two world wars ranges from 40 to 50 years until the moment of emigration.

The number of household heads' fathers born in Kosovo is lower, 286 or 57.2%; 26 never lived in Kosovo and 188 or 37.2% had moved there. Of those fathers who had moved to Kosovo, 164 had come between the two world wars and only 12.7% after World War II, pointing to the very long habitation of Kosovo on the part of both immigrant fathers and their sons. The fathers came from a somewhat wider area than in the case of immigrant household heads. They immigrated to Kosovo, first as part of resettlement, second because they had bought land there. These two reasons together accounted for three-quarters of the immigrant fathers. Hence, the arrival of the older generation was tied in with agriculture, the principal branch of economy at the time, with the abolishment of feudal property relations which had existed in Kosovo until the creation of Yugoslavia, and with a lower population density than Serbia Proper had in 1921.

Knowing the language of the other group, in this case Albanian which is spoken by the majority of Kosovo's population, is, as in any ethnically heterogeneous environment, imperative for direct communication among people of different nationalities. Given the constitutionally guaranteed equality of languages, knowledge of Albanian was not, normatively speaking, a condition for communication between individuals and institutions, but in actual fact that is just what it is, as we shall see later on.

Out of the household members, 342 knew Albanian well or very well, 385 had a smattering of Albanian or knew it poorly, which together accounts for 34.4 % of the household members capable of communicating directly in both languages, at least in some small way. However, since it is hard to imagine children knowing another language without it being through direct contact - and that did not exist in Kosovo because as early as kindergarten the children are divided according to language and nationality - it would seem more appropriate to determine the degree of knowing the other language by looking at those members who were 15 and older at the time of observation. Considering the average interval between emigration and observation, this would mean isolating primarily pre-school children, in which case 41.5% knew Albanian. Unfortunately, we lack comparative data on whether and to what extent the adult Albanian population knew Serbo-Croatian, although we can assume that it was not noticeably higher, because Albanian women account for 45.2% of the population over the age of 15.

An extensive analysis of the answers clearly points to two things. First, knowledge of Albanian largely applies to men and is rare among women, which is understandable given the status women have in the t family and Kosovo society. Female children and women are mostly con-pined to the family circle, especially in the Albanian group; this does not allow for mixing between girls and women of different nationalities and are is virtually no mixing between boys and girls, especially if they are of different nationalities. Mixing and playing together in childhood and when young is of crucial importance for learning the other language of the given environment.

Second, knowledge of the Albanian language is rare among the young, but considerably more frequent among the older and oldest generations. This suggests, and is confirmed by other responses in the study, that direct, day-to-day relations and communication between Serbs and Montenegrins on the one hand, and Albanians on the other, were more intensive and comprehensive in the recent past than they are today. In any event, knowledge of Albanian is more frequent among generations that were born and raised between the two world wars, (i.e. when this language was not constitutionally guaranteed equality it was less used at the institutional level) than among generations that grew up under socialism, when Albanian acquired all rights and was a subject taught in elementary and secondary school.

Knowledge of Albanian is clearly tied in with the ethnic make-up of the settlement and commune in Kosovo. Among households that emigrated from Serbian settlements that stayed Serbian, only one-third, 34 8% had a knowledge of Albanian, in the group of mostly Serbian but with a mixed ethnic make-up that percentage was 62% and in the group of Albanian and largely Albanian settlements, 72.2% of the emigrant households had at least one member who knew Albanian.

This is one of the few findings in our research that coincides with the general rule regarding the relationship between language and ethnic number, a rule that says that the language of the other group is often known by the less numerous group in the settlement or commune. Knowing the language of the other group often stems from ethnic relations in the past, it is an important condition for communicating in the here and now, and should make that communication easier It should follow, therefore, that the most harmonious and least perturbed ethnic relations are to be found where Serbs and Montenegrins accounted for a low share of the population, because this is where knowledge of Albanian is most frequent. But, the threats to and emigration of the Serbian and Montenegrin population run completely counter to this rule, because they were most threatened there where they were few in number. The only thing is that discrimination via language, as a form of discrimination in general, was less effective in this category.

3. DISTURBED ETHNIC RELATIONS

a. A General Assessment of Relations with the Albanian Population

"I've seen it all, and I don't want to remember it.. ."

(A retired worker who moved out of Pristina)

Asked about what relations with Albanians were like in the place from which they emigrated, a small number of respondents gave the following sort of extremist answers "Beyond descriptions, "It's unimaginable for anyone who hasn't lived there". However, the majority of answers covered a wide range of descriptions, most of which reflect pressure and discrimination.

The first step taken in the analysis is a general assessment of relations between the non-Albanian and Albanian population given by the respondents to the question. Only 8.6% described relations as "good", whereas 70.6 % said they were "not good", "poor" or "very poor". Considering the broad interval of time covered by the settlement of the households included in the poll, some of the answers that described relations as "good" or as "separated, without relations", were probably tied in with the experience of earlier settlers.

The explanation for the assessment points to not only the quality but also the content of the relations, thereby defining them still further. More than 70% of the respondents gave such an explanation (362 cases), often with two or more explanations, so that the number of replies was 795.

Assessments that describe relations as "good" would correspond with explanations linked to "visits, friendships, spending time together". However, although as many as 43 respondents described relations as "good", only 12 cited "visits, friendships, spending time together" as the explanation. This indicates that under "good" the respondents also included relations where there was in fact avoidance, separation or simply the absence of direct conflicts between members of the household and the environment.

Only 2.4% of the households cite "visits, friendship, spending time together" with the Albanian population in the settlement of emigration. One tenth of all the explanations refers to "non-involvement, avoidance", implying on both sides, because the voluntary separation of Serbs and Montenegrins was not enough to protect them from the pressures being brought to bear on them. As many as 87.7% of the answers explaining relations with the Albanian population in the settlement indicate some sort of discrimination. Even "non-involvement, avoidances is the result of the start of discrimination.

The most frequently mentioned description of poor relations is indirect verbal pressure"; 30% of the cases cited this. In second place is "material damage" (22%). In short, more than half of the explanations of relations are tied in with "direct verbal pressure" and "material damage". On the other hand, more than one-fourth of all the explanations cite some kind of physical violence: "fights, physical risks", "aggression against children and women", "violence by the organs of authority (police)" and "physical assaults, injuries, rape and attempted rape".

There is a multifold connection between the ethnic structure of the settlement (the share of Serbs and Montenegrins) and different aspects of the emigration process. It can also be observed when looking at the general assessment and explanation of relations with the Albanian population according to communes, grouped in terms of the share of Serbs and Montenegrins.

The indexes of association with the general assessment of relations show that the descriptions of "good" and "separation" appear least often among households that emigrated from communes with the lowest share of Serbs and Montenegrins, and increase as their share in the population grows. On the other hand, descriptions ranging from "not good" to "very poor" become relatively more frequent as the number of Serbs and Montenegrins becomes smaller.

These same connecting lines exist with the explanations of the assessments. Non-involvement, avoidance and good relations appear noticeably more often among emigrants from communes with 30% and more Serbs and Montenegrins than in the sample, but are very infrequent among households that emigrated from communes with the lowest share of Serbs and Montenegrins; the explanations most frequently given in these communes have to do with acts of discrimination. The strongest connection between the share of Serbs and Montenegrins and the type of discrimination is in the case of "non-involvement, avoidance" in communes where the segregation resorted to by Serbs and Montenegrins as a way of self-protection is actually the only possible way, there where number enables relative self-sufficiency and protection.

Grouping the communes is a necessary generalization because major differences often exist within them. For instance, explanations for assessments of relations with the Albanian population in the settlement show that the situation in Podujevo is considerably worse than in Suva Reka, in communes with 9.9% Serbs and Montenegrins. It is hard to make a qualitative comparison in other communes of this group because of the small number of emigrant households, although one does notice, for instance, that all the households that emigrated from Kacanik had suffered from material damages. In the group with 10%-19%, the situation was less unfavorable in Prizren than in Vucitrn, judging by the respondents' answers. In the group of communes where Serbs and Montenegrins accounted for 20 to 29% of the population in 1971, Vitina stands out: this commune accounts for half the answers that list "friendship, visits, spending time together", but on the other hand, it cites the worst form of relations more often than other communes in this group. The communes of Istok and Lipljan are characterized by separation, segregation and lighter forms of discrimination (psychological pressure and direct verbal pressure), and they could be joined by Pristina. Lastly, the answers for Pec and Urosevac show that the gravest forms of discrimination account for a strong share. In the final group of communes with Serbs and Montenegrins accounting for 30% and more (abstracting Leposavic), Kosovska Kamenica had fewer severe forms than Mitrovica where psychological and verbal pressures were strong, while relations were disturbed most in the commune of Gnjilane. Invaluable for our research and its credibility would be the possibility of comparing these statements by our respondents with some social registration of different relations and disturbances in Kosovo's communes, but these sources were not available to us.

The above findings clearly indicate that relations between Serbs and Montenegrins on the one hand and Albanians on the other become all the worse when Serbs and Montenegrins are less represented in the ethnic structure of the commune. Judging by the indexes of association, the critical line for a major shift in relations is representation accounting for 20% to 29%. This means that the drop in the share of Serbs and Montenegrins in the ethnic structure of the communes directly affected the deterioration of relations. All individual factors of this change (such as the decline in the birth rate of Serbs and Montenegrins and rise in that of Albanians, the emigration of Serbs and Montenegrins and settlement of Albanians) worked to aggravate relations between the Serbian and Montenegrin population on the one hand and the Albanian on the other. In this case, then, number is a kind of key to understanding discrimination.

A comprehensive analysis of replies concerning relations in the settlement usually reveals that very unpleasant experiences are presented in these sentences:

- Relations in the settlement were unbearable, there was a constant feeling of fear, mistrust, of being unsafe, especially for children, and the certainty that things would never get better.

In some cases, seemingly good relations in the settlement belied intolerance.

- Relations looked as though they were good but in fact they were bad.

Relations in the settlement are usually described as "non-existent", "unbearable", "bad", with words like "intolerance", "problems".

An in-depth analysis of the answers sheds more light on certain aspects of relations in the settlement. For the primary groups the settlement is the territorial framework and therefore relations within it are primarily if not exclusively of concern to relations in the primary groups, outside institutions.

Certainly the main primary groups through which direct discrimination can be examined are those of friends and neighbors. The frequency of friendly ties between members of different ethnic groups certainly diminishes the existence of discrimination Most of the answers about relations in the settlement do not mention spending time and being friends with Albanians. One gets the impression that this is something so unimaginable to the respondents that they do not think it necessary even to mention the non-existence of friendship and companionship between the non-Albanian and Albanian populations. Hence, the following kind of answer is very rare:

- Serbian and Albanian families spent a minimum amount of time together. It was a public secret that the corso was divided. Young people hardly mixed at all.

It is interesting, however, that among the many negative experiences cited there are also instances of friendship with Albanians. As a result, one and the same family can have both very positive and very negative experiences, as in the case of a family where the father was beaten up in the street for no reason by some Albanians, and the son said:

- There were also other examples: after emigrating my father was in a traffic accident and was hospitalized. Two days later he was visited by a neighbor; an old Albanian and his son came and stayed until my father was released from the hospital.

Individual cases of friendship are mostly tied in with good relations in the settlement at large, the ethnic structure and earlier emigration from Kosovo. For instance, a household head who emigrated in 1966 from Suva Reka where, he estimated, 70% of the population at the time was Albanian and 30% Serbian, says that relations were:

- Good at the time. People mixed and invited one another to their celebrations. Life was peaceful, there was no fear of Albanians.

Positive experience stemming from friendships is often tied in with the traditional, long-standing existence of good relations between the nationalities inhabiting the settlement:

- Relations were good. In Kosovska Vitina there were no major disorders even during the most recent demonstrations, and during the war we lived nicely with the Albanians. We took care of one another. Only one man from our area got killed, whereas only Serbs lived in Serbia and they killed each other.

This second example shows an awareness that discrimination can be selective, so that not all families in the same settlement were equally exposed to discrimination. The criteria of selectiveness varied: the reputation and influence of the household head (a doctor or judge in a small commune, for instance), the number of men heading the family and circle of relations, the possession of weapons in the household and readiness to use them in self-defense against what were once friendships with Albanians.

- Until 1965-1966 relations were tolerable, but from then on a distinction was made between old-time Serbs and Serbs who had settled there between the two wars. The former were more or less respected, at least by the older Albanians, but the settlers were even then exposed to different kinds of pressure. After the 1981 demonstrations there was no difference between the non-respect and molesting of old-time Serbs and that of settlers. Some thirty years ago relations were tolerable, and there were even apparent friendships which evaporated after the demonstrations.

One respondent, for instance, said that his Albanian neighbor had warned him before the 1981 demonstrations to steer clear, not to go alone, because he was on a list of Serbs and Montenegrins who were to be eliminated, and the building in which the household head worked was planned to be blown sky-high The respondent claims that many people on that list disappeared or were "accidently" hit by cars and died. Another respondent says that his Albanian friends would warn him before every riot so that he could lie low. But, public displays of friendship toward Serbs and Montenegrins were dangerous for the Albanians themselves. One household head related how an Albanian neighbor had stopped his house from being stoned but had later had trouble with his compatriots and the authorities. In another instance, when some Serbs were fired from their jobs, their friend, an Albanian, was fired too, Just because he had been their friend, said another respondent The most drastic example of the dangers risked by Albanians who stood up to discrimination against their neighbors was that of an Albanian from a predominantly Serbian settlement. He reported to the authorities that a group of armed Albanians was near the village. A few days later he was ambushed and killed. It is no wonder then that friendships were suspended, evaporated, were reduced to "Good day", if that.

- All our Albanian friends turned away from us.

The selectiveness of discrimination can be discussed not only in terms of being differently exposed to and endangered by it, but also in terms of the different vehicles of discrimination.

Many of the answers include explicit claims that the younger generation of Albanians spearheaded the conflict.

- My parents live in a village near Pristina. They are on good terms with the old Albanians even today. All these disturbances, pressures, violence are incited by young and middle-aged people.

Some of the answers point to yet another kind of selectiveness among the spearheads of discrimination, based not on divisions along generational lines but on the autochthony of the Albanian population. In this respect the respondents draw a clear distinction between old-tuner Kosovo Albanians and Albanian immigrants (from Albania):

- I have to admit that the better educated and more cultured Albanian inhabitants behaved correctly. We had good neighbors, and my brother had good friends. The old-timers are good people, the immigrants are primitive, all they know how to do is fight, they reach for their knives at the drop of a hat. My father used to be friends with the Albanians, they were like brothers, true friends, they attended each other's weddings and slavas.

The absence of conflict between members of different ethnic groups in the settlement can be due to their physical separation, where they do not mix.

- Relations weren't bad, because we didn't have much to do with one another. Serbian houses were in one part of the village, and Albanian in the other. Now and then there would be arguments.

Such separation may have led to the absence of discrimination if there was none to start with, but not if discrimination existed there anyway. The inhabitants of purely Serbian settlements, for instance, were often threatened by their surroundings, especially on arable land near the district border, or when travelling. It is not uncustomary for children in Serbian villages to complete their schooling in the village, instead of continuing their education in the neighboring settlement if need be, or for them to be sent to relatives in Serbia to continue their education.

Segregation is often a transitional form between good and bad relations among members of different ethnic groups. Because of its visibility, it can be a good alarm bell for the deterioration of inter-ethnic relations. Besides housing divisions in the settlement, the most drastic example of segregation is dividing the corso, the street down the middle. Unlike the many historical, cultural and economic factors that work to form relatively separate parts of the settlement in which members of different ethnic groups live, division of the corso depends exclusively on inter-ethnic relations within a given relatively short time period. Division of the corso clearly signals a qualitative decline in inter-ethnic relations. One respondent observed that the deterioration of inter-ethnic relations occurred:

- after the 1966 demonstrations in Pristina, when the corso in Pristina was divided and divisions began everywhere. Another respondent says:

The division of the corso went hand in hand with the division of cafes, pastery shops and other public meeting places.

- People go only to Serbian cafes. On the corso, the ones walk on one side and the others on the other side of the street. Judging by the answers, with time the streets in most of the settlements became dangerous, "hostile". To walk down the street raised fears, was risky and that meant reducing freedom of movement.

Unpleasant incidents in the street ranged from avoidance and being ignored to fist-fights and explosions:

- When an Albanian was alone (during an encounter in the street), then he would say hello, but if there were two or more, then they wouldn't say a word.

Especially at risk in the street were women, one of the sadder signs of discrimination:

- I've gotten into a million fights with Albanians. They molested our women colleagues in school, on the street, in the bus. Their guys "rode" our girls in the middle of the bus. How can you just sit and watch something like that?

As discrimination increased with the passage of time, more and more Serbs began "retreating":

- There was no camaraderie, everything was reduced to just saying "Hello" and even that was only when we had to. If an Albanian said anything to a Serb, the Serb had to keep his mouth shut and so Serbs increasingly retreated and got used to all sorts of things.

There is no doubt that many Serbs felt humiliated by this need to avoid conflict and restrict their own movements, and even by their own emigration.

b. The Time of Deteriorating Relations

"Until 1965-1966, we boys played football together with the Albanians, then they wanted us to play "Serbs versus Albanians" and after that we never mixed again, we always separated".

(HS worker, emigrant from Titova Mitrovica)

One gets a fuller picture of relations in the settlement by analysing how they changed. A comparison of relations in the settlement at the time of emigration by the household and thirty years earlier shows that as many as two-thirds of the respondents feel relations were better before.

Although few of the respondents explained what they meant by "better", based on those who did, we can say that this in most cases referred to the existence of friendships and spending time together. However, one out of every twelve respondents who said relations were better, spoke about covert antagonism and individual conflicts. Certainly the most important finding is that not one respondent felt that relations in the settlement were worse thirty years earlier than at the time of emigration. This confirms the general trend of deterioration in human relations. And this assessment becomes all the more significant when one remembers the existence of one of the dominant official objectives of social development - the strengthening of brotherhood and unity. According to the respondents, relations in Kosovo developed in the opposite direction of that objective.

Assessments of when relations deteriorated and pressures and discrimination against Serbs and Montenegrins began to emerge and spread, places it at the time of preparations for and the outbreak of the "first counter-revolutions, as one of the respondents called the demonstrations by Albanian nationalists in 1968, almost two decades earlier.

Out of 405 households that pinpointed the year when relations deteriorated (in some instances descriptively, "right after the war", "after Rankovic was dismissed", "after the Brioni Plenum", "after they got a flag and Kosmet's name was changed", all of which were translated into their corresponding dates), only 56, or one out of seven, placed this deterioration after 1968. It should be said that until 1981 and the massive escalation of Albanian nationalism, the broader public outside the province did not know about the state of ethnic relations in Kosovo.

A qualitative analysis of the answers shows that the deterioration of relations in the settlement is largely associated with historical and political events:

- Some fifteen years ago, you could walk through town with no problem. After Rankovic's dismissal, Albanians simply ran rampant. The organs of authority supported the Albanians and beat up Serbs and Montenegrins at every opportunity. The courts tried to convict as many Serbs as possible. The public information media in the province did a cosmetic job and made it look as if the situation in Kosovo was better than anywhere else in Yugoslavia.

A certain number of respondents associated the deterioration in inter-ethnic relations with changes in the ethnic structure of the leadership in Kosovo.

- 1967-68 changed the national make-up of the leading authorities.

- After the dismissals in 1966, all Serbs were systematically removed from posts of leadership.

Sudden changes in relations are also associated with the 1968 and 1981 demonstrations by Albanian chauvinists.

- There were major demonstrations here in Podujevo in 1968, and if the army hadn't stepped in, all the Serbian houses would have been destroyed. There were serious excesses, Serbs were beaten up, there were even murders. The Albanians were armed to the teeth, with knives, rifles... they raped women and girls. We didn't dare walk in the streets day or night, and we kept axes by the door.

The general deterioration of relations was reflected in the relations of the primary groups as well. One married couple, both university graduates who moved away from Pristina, said:

- Relations between Serbs and Albanians in Pristina have been deteriorating for years and turning into acquaintanceships with a polite "how do you do". At the slightest provocation, friendships were ended.

After analysing the answers about the deterioration o