|
|
LGI / Resources / Ethnic relationsNATIONAL AND SOCIAL INCLUSION/EXCLUSION IN FORMER YUGOSLAVIA BEFORE THE WAR OF 1991 1 • 2 • 3 • 4 • 5 • 6 • 7 • 8 • 9Vratu{a-@unji}, Vera The paper explores possibilities of comparative research of the relationship between national and social inclusion/exclusion few years before the violent destruction of the former Yugoslavia. Through tragic interplay of internal and external causes, it became the privileged ground just for such investigation. Research empirical foundation is the 1989-1990 data base of the Consortium of Yugoslav social sciences’ institutes. The answers to the selected questions were used as indicators of national and social distance and inclusion/exclusion. Answer modalities to five questions ranging from complete agreement to complete disagreement with the attribution of importance to national component in social intercourse, were pondered and scaled. Respondents classified according to their self-identification with one of the nine most numerous ethnic, confessional and national groups in former Yugoslavia, were ranked according to frequency of the particular answer modalities’ choice, implicating different degrees of inclusionist or exclusionist orientation towards co-nationals and others. The existence of a ranking pattern is discovered. Albanians, Slovenes and Macedonians found themselves without exception on one of the first three places near the exclusionist extreme of all five scales. Yugoslavs, Hungarians and Muslims, with few exceptions, found themselves on one of the top three places near the scales’ inclusionist extreme. In the middle of the five scales, respondents Croats, Serbs and Montenegrins most often interchanged on the places closer to the inclusionist or the exclusionist scales’ extreme, depending on the particular question’s content. Underlying structural and historical factors of the observed pattern were analyzed. The main finding of the paper is that the propensity to express the feeling of national and social inclusion/exclusion by respondents was not influenced exclusively by their national identification as such. The concrete cultural, social, economic and political situation in their living environment, especially the minority/majority status and the place in social division of labor, proved to be the important co-determinants as well. Exceptions from the observed regularities are studied as indicators of "trouble spots". They are methodologically fruitful tool for inter-ethnic conflict potential diagnosis, identification of the main social actors and even prediction. The 1981 census data, the last comprehensive data available for entire former Yugoslavia, were used to construct the representation index of different national groups in the most desired occupations. They were compared with survey data in order to determine validity of generalizations from the sample to the population at large. Theoretical and methodological framework of the study Introduction This paper represents the continuation of the study on the possibilities and limits of comparative secondary analysis of survey data for the understanding and explanation of the former Yugoslavia’s violent breakup. "Balkanization" into five states (for the time being, at least), made this bygone multi-national federation into a privileged ground for comparative research of national and social inclusion/exclusion. Slash between the last two concepts with opposite meanings is appropriate as a reminder of the reflexive and contradictory character of social relationships through which not only the individual, but also the collective self-identity is formed. The picture that one social group has of itself is mediated through the picture that others have about it. Those that feel themselves to be looked down upon and excluded by others, react often in self-defense by depreciating and excluding those that according to their feeling reject them. Resulting incorporation of some people as "insiders", implies by definition exclusion of other people as "outsiders". This phenomenon has often been labelled as ethnocentrism. Questions It is well known tenent of social psyhology that individual’s inclination to act and think in certain manner is influenced, among other factors, by objective characteristics of their life situation. In this paper I would like to find the answer to the following questions: 1) Are there differences between ethnic groups in former Yugoslavia in the extent to which they favour to interact with each other; or conversely the extent to which they prefer to avoid social interchange? 2) Are there differences within ethnic groups in the extent to which they prefer to avoid ethnic intercommunication, according to structural characteristics such as their social position? This paper focuses only on the examination of the existence and intensity of the correlation between respondents’ national belonging and their place in the social division of labor, on the one hand, and their tendency to have an supporting or rejecting attitude toward inter-ethnic social relationships, on the other. Data and measurements Empirical foundation of the research is once more the data base from the research ˘ ˘ Social structure and the quality of living˘ ˘ carried out by the Consortium of Yugoslav social sciences’ institutes in 1989-1990. This was actually the first and the last all-Yugoslav research project. Data for Slovenia were gathered somewhat later than in other five former Yugoslav Republics and two Provinces. This fact reflected the atmosphere of progressive disintegration of cooperation between different Yugoslav peoples inhabiting its socio-political and administrative units. Affected were all spheres of social activity, from politics and economics to culture and science. Because of this, data for all administrative units are not completely comparable. In spite of these shortfalls, I have included all data in this research believing that it is better to have imperfectly comparable data than no data at all. The 1989/90 stratified sample was fairly representative in terms of ethnic/national composition of former Yugoslavia. In the 1981 census, the last comprehensive statistics available for entire former Yugoslavia (Albanians refused to be registered in 1991), there were 2.6% Montenegrins, 19.7% Croats, 6.0% Macedonians, 8.0 Muslims, 7.8 Slovenes, 36.3 Serbs, 7.7% Albanians, 1.9% Hungarians, 5.4% Yugoslavs, and 4.6% of other nationalities in entire population. When we compare this data with the column "nationality" in the Tables presented in the Appendix, we can see that less numerous ethnic and national groups were over-represented (Montenegrins by 3.6%, Yugoslavs by 1.5%, Muslims by 1.1%, Slovenes by 1.0%, Macedonians by 0.8%, Albanians by 0.5%), while more numerous ones were under-represented (Serbs by 3.0% and Croats by 1.8%). This can be explained by the over-proportional enlargement of the samples for smaller administrative units like Montenegro and Kosovo and Metohiya in order to make possible the data analysis within them taken separately. Under-representation of Hungarians by 0.8% is due to the "mistake in the realization of the planned sample for Voyvodina (Hod`i}, Alija, 1991: 425) How much are we justified to generalize survey findings concerning social bearers of the exclusionist orientation onto the general population? In order to answer this question we must examine closer the representativeness of the sample through comparison of survey data and census data. Survey data presented in Table 6 and census data presented in Table 7 point out to the existence of characteristic differences between members of different nationalities regarding their position in socio-professional structure and therefore also in social power distribution. In the category of unemployed or "active population not exercising an occupation", Slovenes have the lowest index of participation and Albanians the highest (41 : 265, Table 6; 45:187, Table 7). This finding can be explained by the fact that Slovenes are geographically concentrated in the industrially most developed parts of the former Yugoslavia with the lowest rate of unemployment, while Albanians are concentrated the most in K&M, industrially the least developed region in former Yugoslavia. In spite of considerable differences in concrete figures, the rank of other nationalities’ members according to their participation indexes’ in the jobless group are similar in both Tables and correspond as a rule to the overall economic situation in R/P where the particular nationality is concentrated the most. It is interesting to point out that according to both sources Yugoslavs are together with Montenegrins, Macedonians and Muslims overrepresented also in the category of unemployed. In the category peasants Muslims are not at all over-represented as it is suggested by survey findings, (214), but are in fact under-represented in it (77). Even from survey data in Table 6, however, it is obvious that greater percentage of Serbs are peasants than it is the case by Muslims. This can be elucidated through historical fact that Serbs were less urbanized than Turks and Islamized Serbs and Croats, who were privileged during the Osmanly occupation and lived in administrative centers. Rank of other nationalities is similar in both Tables, and is positively correlated with the degree of industrial development of respective R/P. Yugoslavs are according to both data sources among the least represented in the category of peasants. In the occupational category "workers", only respondents Muslims and Slovenians are significantly over-represented in it according to both survey data (126 and 128), and census data (136 and 127). Having in mind that census data are not subdivided according to qualifications, it is not possible to check unexpected and un-probable survey finding that Muslims are the most over-represented in the category of more qualified workers, (143) and that Slovenians are among the most under-represented in this category of work places (72), immediately after Albanians (63). Historical and economic reasons for this doubt are already mentioned. The findings for the category of artisans and private businesspersons from the sample (respondents Croats, Motenegrins and Macedonians the most under represented with indexes around 70, and Slovenians the most overrepresented with index of 205) could not be compared and thus checked against the corresponding category in census data. The 1981 census, namely, did not include ownership as the social status determination criterion of rising importance, even though the "overflow" of state and social resources onto the private accounts of economic, political and cultural elite in the foreign banks had begun already in the middle of sixties after the economic reform of 1965. It is true that at the time very small portion of the work force was officially employed in the private sector. Numerous journal articles and unsystematized first hand experience suggested however that "informal sector" and "gray economy" were thriving. Insistence only on the "social sector" in census data therefore resulted in underrepresentation of all nationalities, but especially Albanians in the categories of "trade" and "services’ workers" (indexes 68 and 61). The 1989/90 survey sample did not contribute enough to the correction of the census data flaws for it failed to include sufficing number of respondents representing small, middle and big private entrepreneurs especially among Hungarians and Macedonians (just 3 and 8 respondents). . : In the category of officials, Yugoslavs and Montenegrins are among the most over-represented according to both data sources (from first to third rank with indexes 177, 124 in Table 6, and 173,147 in Table 7). Hungarians are not over-represented in it as careless observer might conclude from the Table 6 (127), if he did not notice that just 14 respondents Hungarians fall into this category. In reality index of Hungarians’ participation in "officials" amounted only to 70 according to census data (Table 7), just 14 index points more than for Albanians. Survey data on the category of experts employed in industry are corroborated by census data on chemists, physicists, technicians, technologues, traffic experts and natural scientists as far as the over-representation of Slovenians and Yugoslavs, on the one hand, and under-representation of Hungarians and Muslims in this category is concerned. This time just two respondents Hungarians and 15 respondents Muslims falling into the category of experts employed in industry were enough to suggest the correct conclusion on their underrepresentation in this category (73 and 65 in Table 6, 55 and 68 in Table 7). Surely, it is impossible to guess when it is allowed to generalize survey results in spite of insufficient number of respondents that can be classified in particular category. The safest thing to do is therefore to control the survey data whenever it is possible through census data. One of the most serious difficulties connected with the use and secondary analysis of other researchers’ survey data, is the absence of the possibility to influence the creation of research instruments. They are mostly constructed for entirely different exploration interests (quality of living in this case). Aware of this problem, I have nevertheless chosen among available survey questions those that seemed to be valid instruments for the examination of the relationship between the ethnic and social inclusion/exclusion. Among the available 300 items, the answers to the following questions were used as indicators of readiness to socialize with members of other ethnic groups or to keep away from them on the contrary: I. How much is for You personally important the national belonging while choosing the spouse? II. With members of which nation among the cited ones would You like to work with the most? III. Do You agree and to what extent with the statement: A person can feel completely safe only when he/she lives in the milieu where exists the majority of its nation’s members?; IV. Do You agree and to what extent with the statement: Among nations good cooperation can be materialized, but never also the complete trust?. Statements concerning the importance of national belonging in everyday interactions were becoming potentially controversial theme at the time of research in 1989/90 because the creation of independent national states by elites of particular ethnic groups having demographic majority in the realm of corresponding territorial units was in full swing. It implied the loss of the status of the "constitutive nation" for members of minority ethnic group in the given territorial unit. In the atmosphere of heightened inter-ethnic tensions, even the best trained interviewers could not completely melt away distrust among the members of local ethnic minority while answering such disputable questions. The uncertain reliability and validity of these research instruments must be therefore kept in mind while deciphering the survey results. The scales of inclusion/exclusion were constructed by using answer modalities to chosen questions, ranging between complete refusal and complete acceptance of the importance of national belonging for social intercourse, as quasi ponders. Scale scores or the average answer modalities for each nationality of respondents in entire former Yugoslavia were calculated as the ratio between the sum of frequencies for each answer modality multiplied by 1,2,3 or 4, depending on the answer modalities’ formulation, on the one hand, and the total number of that nation’s members in the sample, on the other (see Tables 1,2,3,4 and 5 in the Appendix). The category of national belonging in this research is defined through self-identification of respondents. They could choose between 18 names of different nationalities or add their own name. The category of the position in social division of labor is operationalized on the basis of the occupations’ typology of the Federal Bureau of Statistics. It is elaborated by taking into account the qualifications of the respondents occupying various work places (see Table 6). Thus it largely incorporates in itself several other important factors like educational attainment, social prestige, political power. Even respondents’ age (in the category of others are represented the most retired people and students) and the place of residence in the city or in the countryside (peasants or industrial workers) can be deduced from data on the "work place". As can be seen through the comparison of cited 1981 census data (Table 7) and 1989/90 research data from (Table 6), the category of management personnel was over-proportionally represented in stratified samples. This was done on purpose to enable closer examination of the economic and political elite’s attitudes. While interpreting the research results, it should be also kept in mind that in the samples the categories of peasants and workers were under-proportionally represented, especially in the sample for the so called Central Serbia without Provinces Voyvodina and Kosovo and Metohiya. At the same time, the "other" category was over-represented. Survey data are therefore deficient when they are classified according to work place of respondents. In the sample for Slovenia, for instance, exist only two political functionaries, while the entire sample for Serbia is smaller than it should be, on the one hand, but 15.2% of respondents, again in the sample for Slovenia, fall into the category "without answer", on the other, so that respondents non-Slovenians contribute to this last category only 6.9% in entire sample for former Yugoslavia. In order to check the hypothesis that place in the international and local division of labor can contribute to the explanation of differences in inclusionist/exclusionist orientation of the given nationalities' members toward members of other nationalities, the representation index of particular ethnic groups in specific socio-professional categories was constructed. It is calculated as the ratio between the percentage participation of particular ethnic group’s members in a specific occupation group, and the percentage participation of the same ethnic group’s members in the sample (survey data) or in active population (census data). Because of deficiencies of survey data, more reliable basis for the calculation of the participation index of different nationalities in various occupations are therefore 1981 census data (see Table 7). It must be kept in mind, however, that during more than eight years that passed between the census and the survey research, some changes in ocupational structure could have taken place. Absence of sufficient number of respondents of different nationalities in all socio-professional categories in the samples for particular R/Ps necessary for valid statistical conclusions, forced me to include in the analyzes the answers to selected questions of all respondents that identify themselves as members of the particular nationality in entire former Yugoslavia. Having in mind geographic concentration of these nationalities in particular R/P where they have absolute or relative demographic majority, it is still possible to deduce some inferences as to which socio-professional group of respondents within each nationality contributed the most to the expression of exclusionist attitudes toward members of other nationalities in specific R/P. It must be emphasized, however, that even this decision to join all respondents of the same nationality irrespective of the region where they live was not sufficient in the case of small nationality groups (only 158 respondents Hungarians for instance), or in the case of particular occupational categories (sub-groups of workers or experts for example, or technicians and very specific and interesting category of security personnel) even in the bigger samples. In such cases recoding of data was done. In the cases where conditions for valid statistical conclusions were not met (at least 20 or more respondents for the occupation category), data in Tables from 1a to 5a are put into parentheses as a warning against rash generalizations. Considering that variables in this research are nominal, not having numeric properties of meaningful distance between the values, zero point, relative size or symmetry, it is understandable that the tables of contingency and ranking were used as main tools of statistical description of the state of examined relationships at the time of research. Cramer’s variant of the Chi-square-based coefficient of contingency, that can attain the maximum of 1 for tables of any dimension, thus making different tables comparable, is used as the main measure of association between national belonging and place in the social division of labor, on the one hand, with national and social exclusion, as indicated by readiness to express the attitude implying national and social (dis)affection, (in)security and (mis)trust, on the other. It is complemented by the Goodman and Kruskal’s symmetric and asymmetric Lambda variant of the proportional reduction of prediction error measurement. Longitudinal data on the changes of the respondents’ attitudes toward the importance of national belonging through time not being available, it was only possible to resort to the use of comparative method through space. Such contrasting can also contribute to the checking of the reliability and validity of the chosen research instruments. Whenever there existed sufficient number of respondents for valid statistical conclusions, the answers of the members of the particular ethnic group in the situation of local demographic majority, were compared with the answers of the same ethnic group’s members being in the situation of local minority. Research results The main research findings will be exposed separately for each question-indicator, going from those characterizing the state of interethnic closeness or distance in the most significant primary social relationships, toward the ones pointing out to the degree of inclusionist and exclusionist orientation among respondents identifying with different nationalities in wider social milieu and in the global society. Question I. "How much is for You personally important the national belonging while choosing the spouse?" Survey data presented in the Table 1. confirm the hypothesis that there exists statistically significant connection between respondents’ national affiliation and their readiness to enter into matrimony with members of other nationalities (Cramer’s V .249). Four fifths of respondents Albanians and around one third of respondents Macedonians and Slovenians (81.8%, 38.2% and 28.9% respectively) answered that the nationality of the spouse was very important. Taking into account their percentage participation in entire sample, it comes out that respondents belonging to these three nationalities were over-proportionally represented in the extreme exclusionist answer modality (indexes of representation 333, 156 and 118 respectively). From the Table 1 it is also visible that around four fifths of respondents Hungarians and Yugoslavs (72.2% and 83.3%) answered that the nationality of the spouse was not important. They were the most over-represented in the extreme inclusionist answer modality (indexes 155 and 181 respectively). It is therefore not surprising that on the scale of inclusion/exclusion ranging from 1 to 4, respondents Albanians, Macedonians, Slovenians attained the highest scores (3.63, 2.74, and 2.49 respectively), whilst respondents Hungarians and Yugoslavs achieved the smallest scores (1.44 and 1.27 respectively). Other examined nationalities had below-average scores, amounting to 2.16 for the whole sample: Muslims (2.14), Montenegrins (2.04), Croats (1.98) and Serbs (1.94) in descending order. Such ranking can be explained in part by the strong correlation between national, religious and confessional identification of respondents. First ranked respondents Albanians, namely, declared to be persuaded believers in 25.3% of cases and identified themselves with Islamic confession in more than nine tenth of cases (93.1%) (Vratu{a-@unji}, Vera, 1996c: Table 1, p. 498). Islamic confession is known to regulate more strictly conjugal relationships of the believers than other confessions (it is forbidden to Muslim women to marry "infidel" man for instance). Fourth rank of respondents Muslims (in the meaning of nationality according to the constitutional law of 1971), could be explicated by the fact that they identified less often with Muslim confession (76.2%) and less often declared to be persuaded believers (18.2%) than Albanians. The lowest (ninth) ranked respondents Yugoslavs on the other hand, defined as nationally not-determined in population censuses, most often declared not to belong to any confession (63.0%) and claimed to be persuaded believers in only 1.4% of cases (Vratu{a-@unji}, Vera, 1996c: Ibid). Distinct confessional and religious self-identification of respondents affiliated to different nationalities, however, can not account for the finding that respondents Slovenians and Hungarians, that both identified with the same confession (Catholicism) in four fifths of cases (80,7% and 82.3%) and declared to be persuaded believers in one sixths of cases (14.5% and 15.2%), found themselves on the opposite ends of the inclusion/exclusion scale. A look at the Table 9. column No.I, containing the distribution of the extreme exclusionist oriented answer modality to the first question classified according to the nationality of respondents and to the Republic/Province (R/P) in which they live, suggests one of the possible additional explanations. It indicates that irrespective of their national affiliation, respondents that live in the situation of the demographic minority in a given R/P, have a tendency to answer that the nationality of the spouse is "very important" less often than the members of local majority nationality and their co-nationals in "mother" R/P. Thus respondents Serbs in Croatia, for instance, opted for the extreme exclusionist answer modality in 3.3% of cases, more than five times less often than respondents Croats (17.5%) and more than six times less often than respondents Serbs in Serbia (20.0%). Respondents Muslims in Croatia have attributed great importance to the nationality of the spouse less often (11.9%) than respondents Croats in Croatia, respondents Montenegrins in Montenegro (12.7%:15.2%) and respondents Muslims in Bosnia and Herzegovina (B&H) (18.6%) as well. Thus they implicitly expressed readiness to assimilate into the majority nationality or to enlarge the group of nationally not-determined Yugoslavs through inter-ethnic marriages. The choice of exclusionist answer modality by respondents members of majority nationalities, on the contrary, could be seen as the expression of their tendency to retain numeric preponderance on the basis of nationally homogenous marriages. Small number of respondents Slovenians and Macedonians in the samples of other R/P does not allow to check whether members of these nationalities remain so much preoccupied with the nationality of the spouse (29.4% and 40.0%) and consequently with the ethnic "cleanliness" of the posterity. when they reside outside their "mother" republics. In Voyvodina, with the smallest percentage of average agreement with the importance of spouse’s nationality (9%), respondents Croats and Hungarians lagged behind local respondents Serbs in this respect just by two insignificant percentages (9.0%, 8.7% and 11.2% respectively). Noticeable deviation from this rule are the cases of respondents Croats in B&H, Muslims in Serbia, Serbs in Montenegro and Albanians in Montenegro and Macedonia, who declared that the nationality of the spouse was "very important" approximately two times more often (28.8%, 46.4%, 31.4%, 33.3% and 81.0%) than respondents Muslims, Serbs, Motenegrins and Macedonians having relative or absolute demographic majority in respective republics (18.6%, 20.0%, 15.2%, 13.3% and 39.6%, Table 9). Their exclusionist answers indicate the tendency to close up within the circle of co-nationals when the primary relationships are concerned. Such attitude is the result of complex feedback in interethnic relations, both positive action to preserve one’s national identity. and reaction to the feeling of being excluded. Already mentioned explanation pointing out to the specificity of Islamic marital norms is obviously not sufficient to account for the differences in attitudes between respondents identifying with Christianizm. Thus Macedonians, who according to their allegation very little believed in religious teachings (4.4%), opted for the extreme exclusionist answer modality twice as often (39.6%) as respondents Serbs (20.0) who declared to be persuaded believers more often than Macedonians (8.5%) (Vratu{a-@unji}, Vera, 1996c: Ibid). The explanation of these differences in the attribution of importance to national dimension in primary social intercourse should be therefore sought also in wider historical context and in the actual social power relationships both between different nationalities and within the same ones in a particular R/P at the time of research. Tried and true way to begin the examination of inter- and intra-national social power relationships in a given pluri-ethnic society is to analyze general life situation of its members as influenced by their place in social hierarchy. Data in Table 1a, containing the occurrences of the answer modality stating that the nationality of the spouse is "very important" classified by respondents’ nationality and workplace, confirms the thesis that respondents’ place in social division of labor and their attitudes. toward the importance of spouse’s nationality are related (Cramer’s V variates between the lowest .102 and .115 in the case of respondents Yugoslavs and Serbs, and the highest .230 and .190 in the case of respondents Macedonians and Montenegrins). Respondents of all nationalities active in occupations near the bottom of the social ladder, especially respondents peasants, tied to natural rhythms and tradition by their way of living, as well as workers of lower qualifications, tend to be over-represented in exclusionist oriented answer modalities (index 146 and 138) and thus less opened toward members of other nationalities than those near the top of this ladder. Data corroborates the expectation that exclusionist orientation is the least manifested among respondents professional politicians, "responsible" for at least verbal support to the official legitimization ideology of "unity and brotherhood" between different nationalities in former Yugoslavia (index 44). A number of exceptions to this rule exist, however, that should be pointed out. From the Table 1a it is visible that respondents managers Albanians and Montenegrins, specialist employed in administration Muslims, security personnel Croats and Yugoslavs, have slightly more often, or at least as often, chosen exclusionist answer modality to this question.(79.1% and 13.0%, 19.5%, 21.7% and 3.7% respectively) than their co-nationals higher qualified workers and peasants (78.0% and 13.1, 17.8%, 18.0% and 3.2% respectively). 1 • 2 • 3 • 4 • 5 • 6 • 7 • 8 • 9LGI / Resources / Ethnic relations |