THE IMPORTANCE OF OBSERVATION, CLASSIFICATION AND DESCRIPTION IN THE CONSTRUCTION OF THE ETHNIC IDENTITY OF BUNJEVCI FROM BACKA (1851-1910) Aleksandar VUKIC,' Mario BARAn COBISS 1.01 ABSTRACT The Importance of Observation, Classification and Description in the Construction of the Ethnic Identity of Bunjevci from Bačka (1851-1910) This paper considers the influence of ideological, scientific-ethnographic and census classification schemes and descriptions on the formation of the ethnic identity of Bunjevci from Bačka during the period from 1851 to 1910. Starting from the standpoint of a theory of social systems and institutions, the influences of observational apparatuses, observations, classifications, and descriptions on the construction of ethnic identity are discussed. It is shown that transformations in the social order lead to changes in the construction principles of perception and classification schemes. Accordingly, the interaction of commonsensical, ideological, scientific, and official statistical classification schemes and descriptions in the modern social order created an ethno-nationalistic discourse that strongly affected the formation of the ethnic identity of Bunjevci from Bačka. KEY WORDS: Bunjevci, classification, social system, population censuses, identity IZVLEČEK Pomen opazovanja, klasificiranja in opisovanja nastajanja etnične identitete v Bunjevcih v Bački (1851-1910) Članek obravnava vpliv ideoloških, znanstveno-etnografskih in popisnih klasifikacijskih metod in opisov nastajanja etnične identitete v Bunjevcih v Bački v obdobju med letoma 1851 in 1910. Vplive opazovalnih orodij, opazovanj, klasifikacij in opisov na nastanek etnične identitete obravnava z vidika teorije družbenih sistemov in institucij. Izkaže se, da spremembe družbenega reda vodijo v spremembe principov oblikovanja zaznavnih in klasifikacijskih sistemov. Skladno s tem je interakcija zdravorazumskih, ideoloških, znanstvenih in uradnih statističnih klasifikacijskih metod in opisov v sodobnem družbenem redu povzročila etnonacionalistični diskurz, ki je močno vplival na nastajanje etnične identitete v Bunjevcih. KLJUČNE BESEDE: Bunjevci, klasificiranje, socialni sistem, popis prebivalstva, identiteta I PhD in Sociology, Research Associate, Institute for Migration and Ethnic Studies, Trg S. Radica 3, HR-10000 Zagreb, Croatia; aleksandarvuki@yahoo.com. II BA in Sociology and History, Research Assistant, Institute for Migration and Ethnic Studies, Trg S. Radica 3, HR-10000 Zagreb, Croatia; PhD candidate, Department of Sociology, Faculty of Humanities and Social Sciences, University of Zagreb; mario.bara@imin.hr. INTRODUCTION Over the last thirty years, ethnic groups, ethnic relations, nations and nationalism have become one of the central focal points of social research.1 Most frequently discussed are the relations of ethnic groups and nations, their primordiality and modernity, substantiality, and construction.2 However, the systemic social approach of this paper deals primarily with the question of how the mechanisms of social and psychological systems produce ethnic structures. If one accepts Luhmann's (1990) principle, whereby systems are constructed and maintained through processes of differentiation and cataloguing, then special attention must be devoted to systems of classification of human collectives. Modern debates on the "nature" of classification schemes have brought to focus two main approaches: while the first sees classification schemes as a reflection of objective ethnic structure, the second focuses on the construc-tivist3 influence of classification schemes on their subject - the ethnic structure (Kertzer, Arel 2002). This paper begins with the premise that ethnic structure is a product of the permeation of internal and external classification schemes. The distinction between internal and external definitions allows the ethnic identity to be observed at several different levels within a single analytical framework. According to Jenkins, the application of internal and external categorizations and classifications is a dialectical process. External categorizations and classifications are at the same time essential dimensions of internal definitions.4 The experience of categorizing and classifying can reinforce an existing collective identity by providing resistance, reacting, and strengthening boundaries (Jenkins 1996: 23). Outer classification schemes - in the form of commonsensical, scientific, and ideological discourse as well as official censuses - are an expression of the reflexive nature of the modern social order, which reproduces itself while at the same time creating various discourses of itself. Discourses, i.e. descriptions and self-descriptions, are closely connected to the ways in which societies are structured and organized. They are what form personal and collective identities. Furthermore, according to Foucault (1976: 21-23), each episteme that prevails in a specific social system is closely associated with the power structure. It is not, as Weber (1976) claims, primarily associated with the right to enforce power, but is tied to knowledge, which is power over others, the power to label, categorize, and classify others. Modern western institutions and the discourses associated with them have created a modern classification scheme at the centre of which is a new understanding of social groups and individuals. Foucault (1976: 25-26) illustrates his conception of power/knowledge with an example of how the term "population" was constructed in the 18th century. The term became necessary in the context of the state's control over its citizens, taxation, and the regulation of health and housing requirements. Up to this point, those living under the rule of a monarch were referred to as "common people," "plebs," or "loyal subjects," but the notion of a country with its "population" was new and had different implications. First of all, the question of numbers (which was associated with the power of the state) came into focus, as well as the question of the economic resources needed to sustain the needs of everyone. A new scientific and quantitative discourse was also developed - social statistics - which is an expression of modern efforts of the state to tackle the issues of managing and controlling its population.5 1 The theoretical framework used for this study was developed by A. Vukic (see Vukic 2008). 2 Detailed descriptions of more recent theories and studies of ethnicity and nation can be found in the works of Katunaric (2003), Poutignat, Streiff-Fenart (1997), and Smith (2003). 3 In sociology, the year 1983 is considered to be a turning point for the constructivist approach to ethnicity, nation, and nationalism. That year, important works by E. Gellner, B. Anderson and E. Hobsbawm were published: their constructivism "ends the so-called essentialism of historical phenomena and conceptualizes them as classifications of the human spirit and their categories" (Wehler 2005: 9-10). 4 The process of identifying "us" demands that "they" need to be different and thus separated from "us", to be in contrast with "us"; group categorization will likely flow, at least in part, through the positive or negative categorization of others. 5 The International Statistical Congresses in Brussels in 1853 and Vienna in 1857 are acknowledged as the first In this sense, the classification schemes of the censuses played a significant role in the construction of the identities of the different ethnic groups of the Habsburg Monarchy. This notion is incorporated in the main premise of this paper, which is based on the following insights: a) the modern social order is a reflexive order which is created, maintained, and developed on the basis of information collected about itself and its environment, b) the modern social order performs its basic functions by creating an economic, sociological, political, and statistical discourse of itself, c) the result of this is the construction of terms such as society, market, and population, whose purpose is to manage the population and control it through a new scheme of denoting, categorizing, and classifying, d) the power of denoting and classifying objects is a prerequisite of control over the symbolic order within a social system; this power will have practical effects on the way an ethnos is defined - it determines who is granted all of the rights (benefits) in a given state or political community, i.e. political nation, e) in the observed historical period, the effects of the power of classification were ensured through ever-increasing intervention by the governing state apparatus and science in the everyday life of its residents. ETHNIC GROUPS/INSTITUTIONS IN SOCIAL SYSTEMS THEORY According to the postulates of modern systems theory, which treats evolution primarily as the process of development of knowledge, communication, and consciousness, ethnicity is viewed as one of the forms of organization of human populations developed through evolution, whose basic function is to maintain and produce knowledge (Luhmann 2001; Parsons 1951). In such a theoretical framework, progress and change in the structure and identities of ethnic groups are not the result of their cultural differences. Cultural differences are not inherently a characteristic of the phenomenon (ethnos). What happens is a cognitive process in which observers produce diversity by differentiating (Bateson 1979: 99). Consequently, this paper focuses its attention on the process of internal and external differentiations that enable the observer to select and sort objects into classes and categories. The internal mechanisms of selection are (ethnic) institutions, while the external mechanisms are the classification schemes of the population. If the boundary between a social system and its surroundings is taken as a constitutive factor, then the selective purpose of institutions becomes significant.6 By the selection mechanisms of categorizing-classifying, institutions merge the motives for doing and anticipating, thus developing social structures through creating types of behaviours. One of the oldest institutions is the ethnic institution (Parsons 1978). Through the evolution of attempts to standardize census classifications and categorizations. Only very slowly did the category "spoken language" become, in spite of resistance, one of the key questions on census forms (Arel 2002: 94-95). Also, for the first time, statisticians at these congresses began to debate the meaning of the term "nationality" and, for statistics, its practical definition and operationalization. It was not until the Congress in London in 1860 that language became one of the key criteria in defining ethnicity. 6 Institutions take credit for making routine decisions, solving common problems, and doing much of the routine thinking on behalf of individuals. Institutional structures can be viewed as forms of informational complexity. Past experience is set within the institutional rules and acts as a guide for what to expect in the future. For social systems, feedback between obligations and expectations is crucial; it is the mechanism of connecting the past and the future which is constitutive for maintaining action-communication systems (Luhmann 2001a). social systems it has been the main space-time framework for most members of the human species. Ethnic groups/institutions "work" in accordance with a simple binary coding principle. The basic coding operation is performed through the process of distinction along the "us"-versus-"them" axis that draws a line between a system and its environment. All further manipulations with "objects" in the environment are left to operative programs derived from the main code. The identity of ethnic groups is developed from the dialectic of inner coding and outer classifications. According to Isajiw (1990), the orientation and motivation of individuals in an "ethnic situation" is determined by the cognitive, affective, and moral programs of ethnic institutions. This is especially true in social contexts in which ethnic categories and classification schemes have great strength as a principle of social organization and stratification (Epstein 1978). In these situations ethnic groups code each other and accomplish a certain level of consistency and coordination in their actions, thereby developing intricate systems of social classification. One significant and well studied element of these systems is stereotypes. Stereotypes of oneself and of others can be observed as selection mechanisms by which the complexity of reality is reduced. On the other hand, in a complex multiethnic situation stereotypes are programs for determining ethnic closeness or distance. They allow individuals to keep their identity "fluid", and to include some of "them" in their own group or to distance them even more from their own boundaries. PRINCIPLES OF CONSTRUCTING CLASSIFICATION SHEMES The principles according to which classification schemes are constructed are derived from types of considerations and discourses specific to different social orders. When a social order is changed, so are the principles of constructing classification schemes. From the standpoint of social systems theory, classification schemes are institutionalized forms that steer action and communication in one of several potential directions. The possible directions of consideration, functioning, and behaviour depend on the form of the media of exchange of information. Hunting-gathering societies are based on the verbal transfer of information; traditional societies on manuscripts; and modern societies on the printed word.7 Through the evolution of the social order from hunting and agrarian societies to traditional societies and from traditional to modern societies, three distinct systems of classification can be identified. The "primitive" classifications of hunting and agrarian societies are based on the principle of distinguishing nature from culture/society. For Lévi-Strauss, the operation of classifying through binary coding is a congenital attribute of the human mind.8 If this premise is accepted, it is very easy to get from these basic operations inherent to each individual to simple, but for human culture constitutive, classification schemes such as Lévi-Strauss's well-known nature vs. culture / male vs. female / good vs. evil. In hunting and agrarian societies, each member of an ethnic group must learn to differentiate among his fellow members according to their social status. This systemic operation is done through the discrimination/coding of the inner and the outer. "I" cannot exist without many types of "we": my family, my clan, my tribe. In this classification scheme, we belong to the culture, and they belong to nature. 7 Debates on epistemological and communicational cuts that induce changes to the forms of communication in social orders and psychological systems can be found in works of Havelock (2003) and Luhmann (2001). 8 According to Leach's interpretation of Lévi-Strauss's theory, "verbal categories hold within them a mechanism by which universal structural characteristics of the human mind transform into universal characteristics of human culture. (...) If these universal characteristics do exist, they have to be considered congenital on a certain very deep level. Then they can be viewed as patterns that have embedded themselves into the human psyche during the evolution of mankind. The creation of categories follows a similar universal natural path. The human mind is structurally predetermined to develop categories of a specific type in a specific way" (Leach 1982: 46). In traditional social orders, the classification criterion is the place that a certain social group possesses in space or in a hierarchical system. Social identity in traditional systems is initially constructed from the concept of place, and not just place of origin, but an entire set of institutionalized relationships based on this concept. A great part of an individual's personality as a member of an ethnic group is derived from a certain idea of place, locality, which provides the social context for relations with members of other ethnic groups. Consequently, to be connected to a certain place does not only mean having a point of origin - it means having social roots and traits specific to the type of person a certain individual is (Rosen 1984: 23). Modern systems of classification are based on the principle of institutionalized social positions and connecting functions. While in traditional social systems the basis for classification is most often connected to a certain place, in modern societies classifications are usually derived from an abstract concept of "population", interpreted as a gathering of all individuals living in a state's territory. According to principles of socially accepted classification schemes, people differ in the colour of their skin, gender, age, educational, religion, ethnicity, or professional status. In its structural mergence with science at the beginning of 19th century, the state imposed itself as the "highest instance" for denoting, categorizing and classifying the population. Continuing from Foucault's (1979) studies of how "subjects are constituted," Hacking (1986) presents the process of the "invention of people" through denoting (labelling) and mechanisms of coercion which ensure those labels are interiorized. Working on 19th century statistical data on deviate behaviours and control over deviant individuals, Hacking (1990: 3) shows that the modern ways of denoting have constructed "many more types of people than the world has ever before seen." The practice of counting and classifying has by itself created a multitude of new divisions. As quickly as the newly invented statistical, ethnic, medical, or gender categories appeared, people - spontaneously or by coercion - accepted the attributes given to them and lived in accordance with them. However, at the same time the opposite process took place - using the method of logical abstraction it merged local ethnic groups into territorial ethnic communities or, if the conditions were favourable, into nations. SCIENTIFIC AND IDEOLOGICAL CLASSIFICATION SCHEMES AND CONSTRUCTION OF ETHNIC IDENTITIES IN THE HABSBURG MONARCHY Pre-modern and modern censuses used various classification schemes. The population of the Habsburg Monarchy was registered according to class, religious, language, territorial and ethnic criteria. These criteria were all strongly affiliated with transformations of the social order during early modernization occurring at the time of the censuses. Censuses show that the classification schemes in traditional orders categorized and registered the population according to class and religious background. From a religious standpoint, the apparatus of classification-categorization recognizes Roman Catholics, Greek Catholics, Orthodox, Lutherans, Calvin-ists, and Jews. From the standpoint of classes, the classification scheme recognizes clergy, aristocrats, citizens and peasants. While in some censuses members of the first two classes were entirely excluded, some omitted only the female aristocrats. Censuses were regular only for the residents of the Habsburg Military Frontier so that the military contingents could be defined.9 For this reason, records on the population of the Monarchy are for the most part disorganized and impressionistic. 9 Ethnic groups at the Military Frontier included the following: Uskoks of Senj, Krmpocani of the Littoral, Bunjevci, Vlachs, "schismatics", in the Slavonian part of the Military Frontier Serbs and Rascians, and (the most numerous) Vlachs. The category "Croat" did not exist; only Catholics or "pribezi" (fugitives) existed (Kaser 1997: 184-185). The first classification schemes from the Romantic era which aspired to organize this world in the making, as did the one created by Pavel Safarik in 1826, used language as the fundamental differential criterion of nations. But faced with many difficulties in differentiating South Slavic ethnic groups, Safarik had to modify his classification scheme to include additional criteria as well. On the basis of language propinquity, Slavs account for the most general category in his classification scheme. Slavs are divided into East, West and South, which are then further divided into "language tribes" - Croats, Serbs, Slovenes, and a residual and not clearly defined category of "others". Among Croats Safarik includes both Kajkavian Croats and the Orthodox population in Croatia, while under "others" he categorises Catholic "Slavo-Serbs" and Bosnian Catholics. Bulgarians, Serbs in Hungary, Serbs in Turkey, Bosnians, Montenegrins, Slavonians and Dalmatians are categorized as Serbs. Safarik's classification scheme certainly affected the perceptions of numerous representatives of the cultural elite in the Habsburg Monarchy in many ways (Gross 1985: 49). Ljudevit Gaj, for instance, presents his classification scheme of ethnic groups among South Slavs in an article entitled "Our Nation," published in his daily newspaper Novine horvatzke in 1835 (Sidak 1981: 198). Starting with the language criterion, Gaj easily identifies Bulgarians, Slovenes and Serbs (from the Principality of Serbia, and Hungary) on the periphery of his imaginary Great Illyria. However, the central area remains rather vague. For the most part, now starting from the territorial criterion, Gaj differentiates eight distinct "branches": Slovenes, Croats, Slavonians, Dalmatians, Bosnians, Montenegrins, Serbs, and Bulgarians. Gaj saw most of the population between Slovenia and the Principality of Serbia as undoubtedly being of Croatian descent. Among them he also categorizes Croats in Hungary, Banat and Baranya (Gaj 1835: 235). He disapproved of particular movements that identified ethnicities such as Bezjaci Kajkavians, Vlachs, Sokci, Bunjevci,10 etc., as opposed to a single Slavic nation of "Illyrians" (Gaj 1835). One can get a fairly good idea of how many people belonged to the different ethnic groups, which at that time were still only partially integrated, from ethnographic studies of ethnic groups based on church records carried out at that time - specifically, evaluations made by "experts" as well as the field research carried out in 1846 under the leadership of Karl Czoernig, director of the state office of administrative statistics in Vienna. According to Czoernig, who harboured ambitions in the field of scientific ethnography, the basis for ethnographic classification schemes should be a complex combination of characteristics related to language, religion, and historical mentality together with the territorial-ecologi-calconditions under which people lived (Arel 2002: 95). Based on numerous studies, Czoernig and his associates came up with various different estimates of the number of people belonging to specific ethnic groups in the Monarchy (1856: 26-28). More interesting than the estimate itself, however, is the classification scheme applied to these nations. Croats in the Monarchy were divided into the following groups: Slovene-Croats in the area of Civil Croatia and the Military Frontier, Serbo-Croats at the Military Frontier, Croats in Carniola, and Croats in Istria and the Kvarner islands. Furthermore, there was an additional category: members of Croatian "language islands" - i.e., pockets of Croatian speakers in Austria, Moravia, Hungary and Vojvodina. A special category consisted of Croats in the military service. A "hybrid" classification scheme was applied to Serbs as well. It combined ethnic, language, and territorial markers together with the category of social role. As a result, according to Czoernig, Serbs in Dalmatia were divided into Morlachs, Ragusans, and Bokeljs, followed by Serbs settled on the Dalmatian coast and islands. Serbs in Vojvodina and Banat were divided into orthodox Serbs, Sokci, and Bunjevci. One classification scheme that played a significant role in efforts to discuss the language and 10 Bunjevci (pl. of Bunjevac) are an ethnic group who live mostly in the Backa region of northern Serbia (province of Vojvodina) and the Baja region of southern Hungary. Other groups of Bunjevci live in Croatia (Lika and the Croatian Littoral) and in Bosnia and Herzegovina. Most Bunjevci consider themselves to be Croats. Serbia is the only country that recognizes Bunjevci as a separate national minority. However, only some members of their community actually declare themselves as Bunjevci, while the rest declare themselves as Croats. identity (from a scientific-ideological standpoint) of ethnic groups in Croatia, Bosnia and Herzegovina and South Hungary is the one developed by Vuk Karadžic (1849). Karadžic developed his classification scheme under the influence of Safarik's theoretical conclusions and Czoernig's research. Unlike earlier classification schemes, Karadžic applies only the language criteria. His central argument is that the Germans, just like the Hungarians, although belonging to different religions, speak the same language and are therefore one nation. Following the same logic, the South Slavic languages are divided into four dialects: Štokavian, spoken by Serbs, Bulgarian, spoken by Bulgarians, Čakavian, spoken by Croats, and Kajkavian, spoken by Slovenes. Whatever his motives for this classification scheme were, its political, ideological, and communicational effects were of great significance for the development of the discourse on ethnic groups among the urban elite of that time (Jelavich 1992: 25-26). Ante Starcevic's (1971: 46-50) response was motivated by political goals, but it was not, as is often stated, only a distorted and hypertrophied reflection of Karadžic's scheme. Different criteria lay at the basis of his classification scheme. Starčevic believed that the core principle for classification was the "spirit" or "being" of a nation and its state-legislative tradition. For Starčevic, language and religion did not hold significant meaning for the determination of ethnic identity. Hence according to his scheme there are only two nations in the Balkans: Croats and Bulgarians, who have no relation to Slavs. According to Starčevic, Serbs simply did not exist, while Slovenes were "highland Croats" (Gross 1973: 15-53). The scheme that greatly affected the Croatian intellectual elite was the Illyrian-South Slavic classification scheme. Unlike the static, structural schemes so far described, the Illyrian argument had a pronounced developmental-dynamic character. On the basis of language criteria, it began by asserting that a common language of South Slavs existed in ancient history. Due to their divergent history over the centuries, the South Slavs differentiated into Croats, Slovenes, Serbs, and Bulgarians. According to the Illyrianists, at some time in the future, these nations would unite into a common Yugoslav nation (Jelavich 1992: 30-31). Finally, especially in the centres of power of the Habsburg Monarchy, there were classification schemes based on "racial" criteria11 (Taylor 1990). These classification schemes were the implicit starting point for the census policies. Within the Monarchy, the classification scheme "recognized" five races: Slavs, Germans, Hungarians, Romans and Israelites. While Hungarians and Germans each formed a compact statistical category, Slavs and Romans were for ideological and political reasons separated into "language tribes" (Vukic 2008). The Slavs were divided into Czechs, Moravians, Slovaks, Croats, Serbs, Slovenes, Poles, etc., while the Romans were divided into Italians, Greeks, Macedonians, Vlachs, etc. OFFICIAL STATISTICAL CLASSIFICATION SCHEMES AND THEIR INFLUENCE ON THE IDENTITY OF BUNJEVCI FROM BAČKA From the standpoint of modernization, neo-absolutism represented progress in comparison to earlier eras, primarily due to the introduction of modern legislation, judiciary, census-taking methods, fiscal system, and a provisional land registry. The highly complex ethnic makeup of Hungary, like that of other parts of the Monarchy, often made it difficult for external observers and census takers to determine the ethnic relatedness/remote-ness of the various ethnic groups, especially among the South Slavs. Comparing the Habsburg and the Austro-Hungarian censuses, one can recognize a conscious intention on the part of the state apparatus to fragmentize ethnic groups in order (by using the mechanisms of labelling and classifying) to prevent 11 Except for Jews, the classification principle in the "racial" scheme was membership of a specific large language group (Slavs, Germans, etc.). their integration on a "proto-national basis". Ethnic groups which we conditionally and retrospectively identify as Croatian migrated to Hungarian territory over several centuries from various areas of origin. As a result, in later centuries census takers were in doubt as to how to define the different groups of Slavs who had settled there, having migrated from what are now Croatia and Bosnia-Herzegovina. Hence, they were frequently inconsistent in their registries, mixing up regional and ethnic names. But what certainly contributed was the language and ethnic kinship of the Slavic populations being counted. By the end of the 19th century, historic, ethnographic, and ethno-linguistic studies defined the spatial representations and differences between the different ethnic groups (Barth 2009). The population accepted these categories ascribed to them either spontaneously or under coercion by the state apparatus (schooling system, state administration, institutions). By internalizing the ascribed census categories, individuals and groups developed identities that distanced them from their origin. Classification perplexity in the censuses of the Habsburg Monarchy Modern statistics appeared when the state could no longer rely on population data provided by the clergy based on registers of births, marriages, and deaths. Records in lexicons and encyclopaedias on the South Slavic Catholic population in South Hungary most often speak of Dalmatians, but the eth-nonyms Illyrians and Rascians appear as well. Apart from the classification category Dalmatians, which was most often used by Bunjevci from Backa to identify themselves (Antunovics 1858), the categories Sokci and Bunjevci were used as well, usually for the same population. In his Statistics of Hungary for Bacs-Bodrog County, Elek Fenyes (1842) mentions a population by the name of Sokci under which he includes Bunjevci as a subgroup. The necessity of the modernization of the post-revolutionary Habsburg Monarchy imposed the need for a comprehensive census of its population. The census of 1850/1851 was the first to be simultaneously carried out throughout almost the entire country. The Military Frontier was the only territory to be omitted from the census. The census only recorded ethnicity and had no record of language. However, from the standpoint of modern methodology it holds no relevance. It is particularly indicative that most respondents did not know/understand how to answer the question regarding their ethnicity, and in addition there was substantial confusion among the census takers themselves. Despite the numerous inaccuracies, the state, statisticians, and scientists relied on the collected data in later attempts to classify the population. In the Serbian Voivodeship and the Temes Banat, the following census categories were applied to inhabitants originating from Croatia, Slavonia, Dalmatia, and Bosnia: Dalmatians, Bunjevci, Sokci, and Croats (Gross 1985: 52). Due to inconsistencies in the classification of the above-mentioned population, in subsequent censuses and studies there was considerable crossover between categories. In Elek Fenyes's geography of Hungary, for the population he previously labelled as Sokci (in an 1842 statistical study), he uses the broader term Dalmatians when referring to the same settlements (1851). For political reasons, specifically the repression of national movements, in the next census (1857), the question about ethnicity was omitted. This time, the population was classified according to religion (Roman Catholics, Greek Catholics, Orthodox, Jews, etc.). The very next year (1858) the Geographical Lexicon of the Kingdom of Hungary and the Serbian Voivodeship and Temes Banat predominantly uses the term Dalmatians for all Croatian settlers (Bunjevci as well as Sokci). In the same lexicon, Hornyansky (1858) states that "Croats or Dalmatians" live in Bacs-Bodrog County, providing only the total number of residents. The ethnonym Croat in documents and censuses regarding Backa, Banat, and Syrmia during the 19th century was mostly used for individuals from Civil Croatia and the Croatian Military Frontier, who mostly settled during the 18th and the early 19th century. On the other hand, the terms Dalmatians, Croats, Bunjevci, or Illyrians were all used for those who settled between the 13th and the 17th century (although the name Dalmatians prevailed by the mid 19th century). The imperial authorities in Temesvar issued a decree in 1860 to local political establishments in municipalities, districts, and counties to use the official censuses of 1850 and 1857 as a basis for conducting a new population census of the Serbian Voivodeship and the Temes Banat using the criteria of religion and ethnicity. Based on the collected data, the "state accountancy" created a new official census. According to the published summary results, Croats, Bunjevci and Šokci were categorised as a single population group, while Dalmatians were omitted as a census category. For instance, if the data from local census units (settlements, municipalities and counties) is analyzed, Croats were registered in Baja County. At the same time, in Sombor and Subotica people of the same ethnic background were registered under a different census category - Bunjevci. In other counties the census category Croats was used much more often than Šokci (Census 1860: 104-148). On the Monarchy level, although no records on ethnic structure were published in the census of 1857, the state administration for its own purposes operated with data that also included ethnic determinants. However, much confusion occurred on this level as well. With the strengthening of the Hungarian national movement and the formation of the dual Austro-Hungarian Monarchy in 1867, a new moment in classification of the population appeared. The official bodies of Hungary slowly pushed aside the Dalmatian name and started emphasizing the ethnonym Bunjevci to a greater extent, using it as a census category as well. The reason for this was of a political nature, since the Croatian National Revival was at its peak in Dalmatia at the time, which meant that when it came to Bunjevci, the Dalmatian label could serve as a clear marker of their Croatian descent. The ethnonym Bunjevci, which has coexisted with the ethnonym Dalmatians since the times when Croatian groups settled in the Danube region, was internalized by the settlers partly spontaneously and partly through state authorities' mechanisms of repression. Thus it was under this name that the identity of the group would continue to develop and strengthen and under which the national revival would be set in motion. For all activities within the national revival, the Bunjevci cultural elite from Bačka chose this ethnonym in an effort to remain neutral, not only toward the Hungarian authorities, but toward both the Croatian and Serbian nations as well. From a political standpoint it seemed to be too early to support the Croatian national revival because of the role Croatia played in ending the Hungarian revolution (1848-49), in which Hungary wanted to free itself from Austrian domination. The identity of the Bunjevci from Bačka was very fluid and susceptible to change. It varied from being exclusively Bunjevci to politically leaning toward the Hungarian, Yugoslav, and eventually Croatian idea. When it came to selecting either one of the national labels existing in Bačka (Croatian or Serbian), representatives of the Bunjevci believed that in the given circumstances it was better to remain neutral, for any choice could additionally hinder their position. Subsequent censuses up to 1890 show the number of Croats and Serbs in Hungary only as a total. Although separated from that point onwards, it was without any consistency. When it came to census categories such as Dalmats, Bunjevci, or Šokci, they were often registered in some places as Croats and in others as Serbs - even in the same county. Across the area of those counties covering present-day Vojvodina, the censuses from 1900 and 1910 registered the South Slavic Catholic population according to six census categories: Croats, Dalmats, Bunjevci, Šokci, Karaševci, and "others". It is noticeable that in these censuses, Dalmatians from some settlements were treated as a unique category separate from Bunjevci or Šokci, while in others they were registered as the majority population. As of the 1910 census, the census category Dalmatians was almost completely lost. Contributing to the claim of non-systematic and faulty methodology is the fact that within some settlements inhabitants who were undoubtedly of the same ethnicity were divided into Dalmatians and Bunjevci; Dalmatians and Croats; Bunjevci and Croats; Croats and Šokci; and finally, Bunjevci and Šokci. Following up on the results of the 1900 census, Hungarian ethnologist and anthropogeographer Geza Czirbusz (1902: 503) divided the Croatian inhabitants in Bačka into several groups. According to her, Dalmats, Bunjevci, and Šokci were one kindred population divided across different settlements. In addition, fifteen settlements were listed as having mixed populations of Dalmats, Bunjevci, and Šokci. The transitions in self-identification from Croats into Bunjevci, Šokci, or Dalmats; Bunjevci into Dal- matians and Croats; Sokci into Dalmatians and Croats; and so forth, as well as transitions from one census category to another, were facilitated by confessional and language (dialectal) relatedness. All of the above-mentioned population groups in Backa spoke the Stokavian Ikavian dialect, and their affiliation with the Catholic Church clearly separated them from the Orthodox Serbs. Since the censuses were based on the spoken language of the population, the groups in question could not specify any substantial differences between their "languages" ("Illyrian", "Croatian", "Dalmatian", "Slavonian", "Bunjevac" or "Sokac"). In addition to this, all of the above-mentioned Stokavian Ikavian "languages" were considered by these groups to be their own. In the debates during the 1870s in the revival newspaper Bunjevacko sokacke novine on the written language that should be used by Bunjevci and Sokci, most commentators believed that the Croatian literary language should be introduced. Clarifying on which of the Croatian languages should be accepted as a language standard, one commentator wrote: "When books speak of a Croatian language, this is in fact the language we have up till now called the Illyrian, or Dalmatian, or even the Slavonian language" (Vujevic 1870: 326). From the 1880s onwards, it is evident that Bunjevci increasingly identified themselves under the Croatian label. Starting with the representatives of the cultural elite, this spread through other levels of society. Foreign affairs, with the later weakening and final demise of the Monarchy, played an important role in establishing the Croatian identity. Prior to, during, and following the First World War, as the eth-nonym Croat rose in the self-identification pyramid of Bunjevci and Sokci, the name Dalmat completely vanished. The classification schemes and the "ethnic combinations" that came out of them remained influential for decades to come. It is thus evident that the perception of ethnic groups and the data on them in a certain area depends in part on the nature of the categorisation-classification apparatus. What is even more important, however, is that the very existence of classification influences the mind's focusing on identity and ethnic borders. The following table clearly presents the structure and changes of the classification schemes discussed in the paper: Chart 1: The structure and changes of the classification schemes (1785-1857) Type of Classifier Year the Subject of the Principle of Divisional Categories Sub-categories classification scheme was classification classification class scheme developed scheme Official Feudal gover- 1785-1787 Members of Class, Class, Nobles and Nobles, State government nment admini- classes Religion Confession Commons; officials, Clergy, statistical stration Christians/ Jews Merchants, Peasants Official Feudal 1805 Commoners, Religion Commo- Catholics, government government Men ners, Orthodox, Cal- statistical administration Men vinists, Jews, Lutherans Scientific Safarik 1826 Population of the Balkans Language Native speakers Croats, Slovenes, Serbs Kajkavian Croats, Orthodox Croats, Bulgarians, Montenegrins... Scientificide- Gaj 1835 South Slavic Language, South Slavs Croats, Bulga- Bunjevci, Sokci, ological population Territory rians, Serbs, Slovenes Slavonians, Dalmatians, Bosnians, Montenegrins Type of Classifier Year the Subject of the Principle of Divisional Categories Sub-categories classification scheme was classification classification class scheme developed scheme Scientific Fényes 1842 Population of Language, Ethnic Croats, Serbs, Croats, Serbs, Hungary Religion, Mentality, Territory groups Germans, Hungarians, Czechs... Montenegrins, Sokci, Bunjevci, Kliments (Albanian clan) Scientific Czoernig 1847 Population of Language, Ethnic Croats, Serbs, Croats, Serbs, the Monarchy Religion, Mentality, Territory groups Germans, Hungarians, Czechs. Morlachs, Bokeljs, Croats in military service. Scientificide- Karadzic 1836 South Slavic Language South Slavs Dialects: Caka- Croats, Bulga- ological published in 1849 population vian, Bulgarian, Stokavian, Kajkavian rians, Serbs (Bosnians, Bunjevci, Sokci, Latins), Slovenes Scientificide- Illyrianists 1835-1850 South Slavic Language South Slavs Croats, Slo- ological population venes, Serbs, Bulgarians Official Modern 1851 Population Ethnicity, Population Hungarians, Dalmatians, government government of Serbian Religion Germans, Bunjevci, statistical administration Voivodeship Serbs, Croats Sokci. Ideological Ideologists of 1842-1914 Population Cultural Races in the Slavs, Croats, Serbs, the Monarchy of the Monarchy heritage Monarchy Germans, Romans, Hungarians, Jews Poles, Italians, Greeks, Vlachs... Ideological Starcevic 1852 South Slavic Mentality Balkan Croats, "Highlander" population and state legislation nations Bulgarian Croats Official Modern 1857 Population Religion Population Roman Cath., government state of Serbian Greek Cath., statistical administration Voivodeship Orthodox... CONCLUSION Interaction between inner and outer classification schemes in times of modernization had a great effect on the formation of ethnic identities and consequently on the formation of the ethnic structure of the Habsburg Monarchy during the observed period. The power of the classification schemes greatly affected the creation of a new symbolic order whose functioning is necessary for the maintenance of boundaries and identities of modern social systems. Unlike the symbolic orders of the traditional type, whose main characteristic is consistency, symbolic orders of the modern type are characterized by functional inter-changeability of their elements. While the traditional symbolic order can be more easily depicted as a symbolic universe of discourses and meanings, the modern symbolic order manifests itself as a mulitverse - a world of meaning that is constantly reconstructed from the standpoint of an element's position in the symbolic network. In so doing, every standpoint can be taken as equally correct and constitutive for the modern symbolic order. The ideological, official-census, and scientific classification schemes of the population discussed in this paper had several functions in the symbolic order of the Habsburg Monarchy: a) classification schemes become a means of deriving a different way of observing, and hence understanding, the world, b) the categorized population becomes an object of various ideological, scientific, demographic and economic manipulations, c) the population becomes a resource (economical, military, political, ideological) in the hands of the rational state apparatus, d) considerations on the numbers as well as the religious, ethnic and other markers of the society become an important communication theme of social systems in the process of modernization, e) the statistical results of censuses become powerful tools in determining ethnic borders as well as the integration and political mobilization of an ethnos for the achievement of political goals. The strength of classification schemes is also visible in the processes of forming modern ethnic identities on the territory of Backa settled by Bunjevci and other ethnic groups. During the observed period, scientific-ideological classifications encouraged thought on the common identity of all ethnic groups speaking the Stokavian dialect, while the official classification schemes worked towards the fragmentation of these groups. Still, the very nature of social statistics, as a modern way of observing based on precise numerical indicators served, on one hand, the process of transforming some ethnic categories into ethnic communities, and on the other, the process of erasing some ethnic categories from official classification schemes. This happened with the ethnic category Bunjevci during the period of socialist Yugoslavia. However, Bunjevci in Backa reappear as an ethnic category in the Republic of Serbia in the 1991 census, the 2002 census, and the 2011 census, confirming the initial premise that there is a correlation between classification (perception) schemes of state apparatuses and the development of ethnic identities and communities. REFERENCES Antunovics, Jözsef (1858). A szabadkai dalmata nepszokasok, rövid törteneti ismertetessel. Hazank 1/1-2, 203-215. Arel, Dominique (2002). Language Categories in Censuses: Backward - or Forward - looking? Census and Identity: The Politics of Race, Ethnicity and Language in a National Censuses (eds. David Kertzer, Dominique Arel). Cambridge: Cambridge University Press, 92-120. Barth, Daniel (2009). On the Border of Western and Eastern Christianity. The Coexistence and Conflicts of Religions in Southern Hungarian Bacska in the 18th Century. Acta Ethnologica Danubiana 10-11, 53-60. Bateson, Gregory (1979). Mind and Nature: A Necessary Unity. New York: Dutton. Census 1860 (1863). Stanovništvo Srbske Vojvodine po veri i narodnosti (po zvaničnom izveštaju popisa od 1860. godine). Letopis srbski za 1864, 104-148. Czirbusz, Geza (1902). Magyarorszag aXX. evszazad elejen. Temesvar: Polatsek. Czoernig, Karl (1856). Die Vertheilung der Voelkerstaemme und deren Gruppen in der österreichischen Monarchie (Sprachgraenzen und Sprachinseln) sammt einer statistisch-ethnographischen Ubersicht. Wien. Epstein, Arnold Leonard (1978). Ethos and Identity. London: Tavistok. Fenyes, Elek (1842) Magyarorszag statistikaja. 1. Pesten: Trattner-Karolyi Tulajdona. Fenyes, Elek (1851). Magyarorszag geographiai szotara mellyben minden varos, falu es puszta körülme-nyesen leiratik. 1-2. Pesten: Nyomatott Kozma Vazulnal. Foucault, Michel (1976). The History of Sexuality. Harmondsfort: Penguin. Foucault, Michel (1979). Discipline and Punish: the Birth of the Prison. Harmondsforth: Penguin. Gaj, Ljudevit (1835). Naš narod. Danica Horvatska, Slavonska i Dalmatinska 1/34, 234-236. Gross, Mirjana (1973). Povijestpravaške ideologije. Zagreb: Institut za hrvatsku povijest. Gross, Mirjana (1985). Počeci moderne Hrvatske. Zagreb: Globus. sc Hacking, Ian (1986). Making up People. Reconstructing Individualism (ed. Thomas C. Heller). Stanford: University Press, 222-236. Hacking, Ian (1990). The Taming of Chance. Cambridge: University Press. Havelock, Eric (2003). Muza uči pisati. Zagreb: AGM. Hornyansky, Victor (1858). Geographisches Lexikon des Königreichs Ungarn und der serbischen Woiwodschaft mit dem temescher Banate. Pest: Verlag von Gustav Heckenast. Isajiw, Wsevolod W. (1990). Ethnic Identity Retention, Ethnic Identity and Equality: Varieties of Experience in a Canadian City (eds. Raymond Breton, Warren Kalbach, Wsevolod Isajiw). Toronto: University of Toronto Press, 34-91. Jelavich, Charles (1992). Južnoslavenski nacionalizmi. Zagreb: Nakladni zavod Globus. Jenkins, Richard (1996). Social Identity. London, New York: Routledge. Karadžic, Vuk S. (1849). Kovčežič za istoriju, jezik i običaje Srba sva tri zakona. Beč: Štamparija jermenskoga manastira. Kaser, Karl (1997). Slobodan seljak i vojnik. Zagreb: Naprijed. Katunaric, Vjeran (2003). Spornazajednica. Zagreb: Jesenski i Turk/HSD.s. Kertzer, David, Arel, Dominique (eds.) (2002). Census and Identity: The Politics of Race, Ethnicity and Language in a National Censuses. Cambridge: Cambridge University Press. Leach, Edmund (1982). Levi-Strauss. Beograd: BIGZ. Lévi-Strauss, Claude (1969). Totemism. London etc.: Penguin Books. Luhmann, Niklas (1990). Essays of Self-Reference. New York: Columbia University Press. Luhmann, Niklas (2001). Društveni sistemi. S. Karlovci, Novi Sad: Izdavačka knjižarnica Zorana Stojanovica. Luhmann, Niklas (2001a). Znanost društva. Zagreb: Politička kultura. Parsons, Talcott (1951). The Social System. New York: Free Press. Parsons, Talcott (1978). Action Theory and the Human Condition. London/New York: Free Press. Poutignat, Philippe, Streiff-Fenart, Jocelyne (1997). Teorije o etnicitetu. Beograd: Čigoja štampa. Rosen, Lawrence (1984). Bargaining for Reality: The Construction of Social Relations in a Muslim Community. Chicago: University of Chicago Press. Smith, Anthony D. (2003). Nacionalizam imodernizam. Zagreb: Politička misao. Starčevic, Ante (1971). Političkispisi. Zagreb: Znanje. Šidak, Jaroslav (1981). Ilirski pokret. Društveni razvoj u Hrvatskoj od 16. do početka 20 stolječa (ed. Mirjana Gross). Zagreb: Globus, 191-230. Taylor, Allan John Percival (1990). Habsburška monarhija. Zagreb: Znanje. Weber, Max (1976). Privreda i društvo. Beograd: Prosveta. Wehler, Hans-Urlich (2005). Nacionalizam. Zagreb: Jesenski i Turk. Vujevic, Stipan (1870). Bunjevačke i šokačke novine 1/1-42, 326. Vukic, Aleksandar (2008). Moc klasifikacije: klasifikacijske sheme i konstrukcija etničke strukture u Ban-skoj Hrvatskoj i Slavoniji (1785.-1860.). Migracijske i etničke teme 24/1-2, 31-48.