CHARLEMAGNE J 1984 "Bridging the Cultural Gap" Unesco Courier
October (Special Issue on "The Gypsies") pp.15-17 BELL P 1984 Peasants in Socialist Transition: Life in a Collectiuised
Hungarian Village University of California Press, Berkely ERDÖS K 1960 "Le probléme Tsigane en Hongrie " Etudes Tsiganes
Vol.3 pp.1-10 FÉL E and HOFER T 1969 Proper Peasants: Traditional Life in a
Hungarian Village Viking Fund Publications in Anthropology, No. 46,
Aldine, Chicago GELLNER E 1983 Nations and Nationalism Blackwell, Oxford GUY W 1975 "Ways of Looking at Rom: The Case of Czechoslovakia " in
Rehfisch F ed. Gypsies, Tinkers and Other Travellers Academic Press,
London, pp. 201-229 De HEUSCH L 1966 A la découverte des Tsiganes: Une expedition de
reconnaissance Ed. de LTnstitut de Sociologie de L'Université Libre,
Bruxelles KAMINSKI I-M 1980 The State of Ambiguity, Studies of Gypsy Refugees
Anthropological Research, Gothenburg OKELY J 1983 The Traveller-Gypsies Cambridge University Press,
Cambridge PIASERE L 1984 Mare Roma: Categories Humaines et Structure Sociale
Une Contribution a L'ethnologie Tsigane These pour le Doctorat du 3.ěme
Cycle, University of Paris STEWART M 1989 '"True Speech': Song and the Moral Order of a
Hungarian Vlach Gypsy Community" Man Vol 24, pp 79-102 STEWART M 1993 "Gypsies, the work ethic and Hungarian Socialism" in
Hann C M ed. Socialism Routledge, London STEWART M forthcoming The Time of the Gypsies: poverty, cultural
identity and resistance to proletarianisation in socialist Hungary t            Westview, Boulder
f?.' SUTHERLAND A 1975 Gypsies: The Hidden Americans Tavistock, London |) TOMKA M 1984 "The Gypsy Craftsmen of Europe" Unesco Courier October f'-'          (Special Issue on "The Gypsies") pp.15-17
f         WILLIAMS P 1984 Marriage Tsigane: une ceremonie de fiancailles chez les
Rom de Paris L'Harmattan/Selaf, Paris
oiiapier /   üong performance: a model for social interaction among Vlach Gypsies in South-eastern Hungary Iren Kertész-Wilkinson, Lecturer in Ethnomusicology, Goldsmiths' College, London
The Vlach Gypsies or Rom, as they refer to themselves, form an identifiable social category among the three main groups of Hungarian Gypsies, the others being the Beash, who speak Romanian and Hungarian, and the Romungri, or older-established Hungarian Gypsy groups. The Vlach Gypsies are marked by three major cultural features: their own language, Romani, their own song repertoire and their own style of performing songs and dance. All three features are essential to any public event and thus to Vlach identity. This is reflected in the folk view that every 'true' Rom must be able to sing, dance or tell folk tales in the Romani way, at least to an acceptable degree.
Song and dance performances of the Hungarian Vlach Gypsies embody a range of symbol systems which constitute Rom identity. These performances externalise the Vlach Gypsy inner social structure through which the place and role of individuals and their interactions can be enacted and, albeit temporarily, re-structured. They also create an explicitly acknowledged boundary between themselves and non-Gypsy or other Gypsy groups.
The meaning of musical performances as a cultural phenomenon and a social fact cannot be understood without examining the social contexts in which musical activities take place. Such contextual investigations are all the more indispensable in cases such as that of Hungarian Vlach Gypsies where - in distinction to many other societies including the peasant cultures of eastern European within which they are embedded - more or less the same repertoire is performed in a wide range of contexts. In such circumstances it is misleading to assign a single meaning to song and dance performances. This is especially so because, as symbolic systems, they already contain multiple possible interpretations and therefore social context merely adds a further dimension to that complexity.
97    j
Social contexts of cultural manifestations have generally been grouped into two main categories: public and private. Of the two, the public contexts have gained greater attention from social scientists and ethnographers because, as Richard Baumann (1978:280) points out, they are:
the most prominent performance contexts within a community. They art. as a njie. scheduled event?, rbiincted in s-ettmg. clearly bounded, and widely public, involving the most highly formalised performance forms and accomplished performers of the community."
However, as Baumann continues:
"The most challenging job that faces the student of performance is establishing the continuity between the noticeable and public performance of cultural performances, and the spontaneous^ unscheduled, optional performance contexts of everyday life." The study of performance, as Ruth Finnegan (cited in Baumann 1978) reminds us, is by no means simple or:
"... altogether straightforward, partly because this interest has been relatively undeveloped till recently compared to textual analysis, partly because of the unlimited number of possible components in performance given the variety of cultural, historical and generic conventions and the varying ways these are developed."
In my own field of Hungarian Vlach Gypsy song studies the contextual approach has been largely neglected until recently. Most researchers, following the model set by Hungarian folk music research, have had as their primary goal the collection of large samples of data in order to establish the characteristic features of a particular style before focusing on in-depth studies. This extensive survey method limited the time that was spent in any specific community and thus inevitably led to a neglect of contextual studies of performances.
Among the few who have changed this path is the English anthropologist, Michael Stewart, who conducted in-depth research of a north Hungarian Vlach Gypsy community in the mid-1980s. His anthropological background and the model of his teacher, Maurice Bloch (1974), led Stewart, rightly, to consider performance events as forms of ritualistic and symbolic communication in which singing and dance play essential roles. In the spirit of the structuralist approach his study brought out various binary oppositions of Vlach Gypsy life: between everyday speech and ritualistic performances of song, relationships between Gypsies and non-Gypsies and male-female gender oppositions, all of which Stewart showed, to some extent, have their parallels in the musical performance. Stewart's study is undoubtedly a significant development in Hungarian Vlach Gypsy song research. He was the first to study performance of the songs as an enactment of social life and to realise that their significance to the 'insiders' lay in their offering a vehicle for telling the 'truth'.
I
That entails a selection of values that are expressed in the song text which, according to Stew(art's study, must be in Romani language and sung to a Vlach Gypsy melody.
Stewart's approach, however, is flawed by being limited to a single context of Performance, the male-only mulatsogo. and by viewing sons as essentially only a verbal expression. This constrains hin: to the correspondingly limited conclusion that performance serves -r.e purpose of expressing a traditional form of authority through which Mach Gypsy men wish to establish equality between one another whilst asserting their domination over women. This echoes Blochs verdict on the function of song and dance performance among the Merina of Madagascar as well as in other parts of the world. Ultimately, Stewart may have detracted from the undoubted value of his work by failing to consider whether his a priori model fitted the full diversity of the culture to which he applied it. As Ruth Finnegan (1992:100) pointed out in relation to the variability of social situations of performance, one has to be:
"... alert to this variety rather than just presupposing one model (most likely projected from personal preconceptions)."
IndeeoVStewart mentions that there are other events, such as weddings, during which women are present, but he omits them from his study on the ground that the folk terminology does not name them mulatsago (Stewart 1987:145). Whilst no one would argue about the importance of folk concepts in establishing valid categories, we must be wary about accepting them as the sole determinants. A consideration of all contexts may reveal aspects of folk classifications that in turn hide multiple functions of performances. Concepts and functions of performance are indeed significant aspects of any ethnomusicological and anthropological studies that enlarge our knowledge about human societies. Nevertheless, the aims of these two disciplines are also to show disparities which need to be recognised. As Blacking (1989:61) has noted:
"Studies of music and dance have been both helped and seriously hindered by the application of ideas and methods drawn from history and the social sciences, chiefly because they cannot account for the crucial fact that individual aesthetic experience frequently transcends time, place and social constraints. It has indeed been useful to examine the varieties of human creativity within historical and cultural framework but at the same time, the concepts of history and the social sciences can shift the emphasis away from the content of the music and dance and concentrate on their uses and their extra-musical and extra-dance meanings to those who use them, rather than on their special qualities and musical experiences... Using socio-political concepts to explain musical variety implies that artistic imagination is simply part of the superstructure of social life, determined by economic and political formations: musical variety is therefore seen as the product of political and ethnic differentiation rather than of a special kind of human creativity."
Romani culture and Gypsy identity
In order to study and understand musical performance as experience and as human creativity, many ethnomusicologists have urged the need for a special approach, namely, participation in music-making (for example Baily 1984; Blacking in Howard 1991; Feld 1982; Hood 1960; Kertész-Wilkmson 1994). Some anthropologists have offered similar advice. For example, Colin Turnbull (1990:77) has noted that:
"Few of us, for instance, pay nearly enough attention to the use of the senses of smell, taste, and touch, or to non-verbal sounds. Yet these, particularly sound, are often key elements in ritual and other religious behaviour. Sound alone provides a royal road to the liminal condition, and the feelings it evokes can be highly significant for the fieldworker if he allows them, if not sound itself, to move him both physically and emotionally, instead of confining himself to an exclusively conscious, rational framework." Turnbull (1990,p.79) argues later that:
"The use of subjectivity, of total (including emotional, spiritual) participation ... is not unlike use of rational process by which we recognise without any discomfort that things are seldom, if ever, what they seem to be, and set about arriving-at a more accurate, ŕ" Tnore complete knowledge of what they are by reconsidering them * ,-jn light of other information, form different perspectives, and so
forth. Anthropologists go to exceptional lengths, in fact, not to ^iiccept things for what they seem to be, but those lengths are entirely intellectual and, all too often, spectacularly acrobatic but ultimately meaningless."
This type of approach opens up the possibility of surmounting the limitations that result from regarding song and musical performance mainly on a verbal level, which is my second main criticism against Stewart's work. The researcher's physical as well as emotional participation in performance may result in the discovery of many processes of communication that take place at a non-verbal level. We may find that relationships between participants are not simply reproductive or counterbalancing social relationships but specific interactions or alliances that are based on the outcome of kinetic, affective and aesthetic experiences.
This chapter will show how Hungarian Vlach Gypsy song performance may gain different meanings according to a variety of social context. My aim is to show not so much how musical performance reveals a different reality from the "mask of appearance" of everyday reality, as argued by Michael Stewart (1989:79) but how, through a study of multiple contexts, we can "discriminate between the common mask [of musical performance! and the complex variations which it conceals" (Cohen 1984:74 [my parenthesis)). Because music operates simultaneously at different levels, not only can it accommodate the highly social and highly individual within the framework of a single context, it can also offer, as a form of symbolic expression, a perfect medium to enact the conflicting or paradoxical values and emotions which serve to affirm and re-structure social interactions.
Song performance: a model for social interaction
My analysis will concentrate on the repertoire and performance rules of both the slow and dance song genres of the Vlach Gypsies within the dialectical movement of public/communal, the less formal and the private/ individual contexts.
Social and economic life of the two communities under study
The fieldwork on which this chapter is based was carried out during two extended visits to Hungary. The first of the visits, from August 1987 to January 1988, was mainly among the Vlach Gypsies of Community A. It was in the course of this first period that I decided to make Community B the locality for the second phase which was completed between July 1988 and January 1989.
The two communities are situated in Békés County, in the south-east corner of Hungary, an area that lies between the Transylvanian mountains and the Great Hungarian Plain within easy reach of the Romanian and Serbian borders. It is an ethnically mixed area with significant populations of various Gypsy groups (Vlach Rom, Romungri, Beash), and Slovaks, Romanians and Serbs interspersed among the Hungarian majority. This diversity "is experienced most obviously through the various languages one hears when travelling on the trains and buses, still the main means of transport between the county's settlements. The climate is continental with hot summers and often bitterly cold winters with heavy snowfalls. The hot weather and the. rich soil favour agriculture. Besides grain crops, the cooperative farms of this region specialise in growing onions, paprikas and melons. Food-processing and the manufacture of shoes, clothing and bricks provide most of the limited opportunities for work in light industry.
In both communities seasonal work is an important means of subsistence. During springtime the Rom collect edible snails whilst in the summer they pluck geese, pick blueberries and cut sorghum. In the late summer and autumn these can be supplemented by gathering corn cobs left in the fields by the harvesting machines; some of these are sold, the rest are kept to feed the horses during the winter. With the exception of plucking geese, the work is carried out by entire family groups working together as teams. They work hard for four or five hours a day, sometimes more, but in keeping with the Rom tradition of not separating work and enjoyment, they treat it as a family outing with jokes and songs livening up every moment. The democratic nature of Vlach Rom society finds expression in the equal division of rewards for labour. Children get their fair share but they often spend it on household goods to help their mothers.
The combination of various factors, such as lack of heavy industry in this area, no land to join co-operatives, the general poor health of Vlach Gypsies and a gradual shift in Hungarian government policy on provision of social allowances and security for those who live below the official poverty line, has meant that in the south-east of Hungary most Rom men, after a longer or shorter period of state employment, revert to their traditional occupation of horse-dealing or set up their own private business. In accordance with Rom tradition, it is up to women to support the family on a day-to-day basis through whatever jobs the local economy offers. In Community A some women worked either in a factory making socks and tights or in poultry factories. Others received a state disability pension on grounds of poor health and/or child allowances which served to provide for
101
ultui
the families' basic necessities including food, clothing and household requisites. The local council, including its chairman, was sympathetic to Gypsies and actively assisted improvements in the settlement as well as providing various supplementary welfare allowances including grants for outfitting the children at the start of the school year and for purchasing fuel during winter.
The council of Community B, by contrast, was hostile towards Gypsies. It not only resorted to every possible pretext for refusing assistance, which included setting absurd conditions for providing loans for improvement or purchase of properties, it also imposed punitive fines on Gypsies caught collecting wood on common lands. Nevertheless, the Rom women were in demand among Hungarians of the village and nearby towns for various occasional jobs such as domestic cleaning, white-washing walls and gathering onions, garlic or poppy-seed for which they were given goods in exchange or some money.
The difference between the two Vlach Rom groups is apparent in their economic and social structure. Whilst the Rom of Community A have higher incomes that advantage is offset by the fact that they must buy almost everything in the shops or from Hungarian neighbours. The minimal role of welfare assistance in Community B is compensated Tor by a continuing demand from Hungarians for traditional domestic help from the Gypsies. That makes the Vlach Rom of Community B more resourceful in finding means of subsistence but in the process they end up even more marginalised in the eyes not only of the Hungarian population but even of other Rom groups which tend to look down on them for their poverty.
As regards social structure, Community A comprises three core families who are allied through marriage. Through the loss of many kin members to neighbouring towns over recent decades, it is more open than Community B in that Rom men from outside are accepted for marriage. Furthermore, the main kin group in Community A is regarded as of higher standing in the Rom hierarchy of'nations' than its equivalent in Community B. The Rom of Community B compensate for this by maintaining strong unity as a closed cognate group.
The inner structure of the two communities
The Vlach Gypsies in the two communities I studied have structured their inner life on the purity principle to create boundaries between gender and generation which also encompass classifications of space, time, food, clothes, animals and the universe. In general, these correspond more or less closely with the findings reported for other Romani groups by Ann Sutherland (1975), Judith Okely (1983) and Michael Stewart (1987, 1989) although there are certain notable differences in where the line of acceptability is drawn. Amongst the Rom I studied those lines were often subject to individual interpretation. Thus, for example, my hostess's daughter in Community B allowed her favourite kitten to sleep in the room despite the fact that the grooming habits of dogs and cats render them 'unclean' as a rule. An even more extreme case was a puppy that my hostess in Community B had herself breast-fed and which grew up to be treated as a trustworthy and, so to speak, 'Romanised' member of the family.
Contrary to Stewart's observation among the northern Hungarian "'Vlach Rom, the Rom in the south-east make distinctions between various
qgfjgreSSřTT^ sp^jrTy-^^
^^^^mmmm^mm

classes of non-Gypsies or gajo. The term gajo [Romani - herafter RJ is applied to those with whom they would not mix socially, accept food or clothes from and often referred to as "dirty peasants". Non-Gypsies with whom they can eat, accept clothes from or even marry are considered to be 'genteeľ, úri [Hungarian - herafter H]. This is similar to the findings of Andzrej Mirga (1986) in the Tatra mountains of Poland where non-Gypsies who respect the Rom are accepted as patjan [RJ or 'good persons' worthy of being invited to sit at the same table, manus [R] 'non-Gypsy man' or raj [RJ, 'genteel' non-Gypsies. The converse applies to other Rom. Although in one sense they are always regarded as superior to the non-Rom they are placed within a subtle hierarchical division. Every family or individual will find another who is in some way inferior, either because "their women are whores" or because "they are dirty Rom" whilst those in a higher 'nation' can be belittled for being "mean". This implies that the rules, although in one sense static and rigid as abstractions, are also flexible enough to be expanded to incorporate precisely those aspects which are normally excluded as 'other' and 'outside'or, alternatively, to be narrowed to exclude aspects tBUt-Would be incorporated in another context.
Some expressions of the public /private dichotomy in folk views It must be emphasised that 'public' and 'private' should be understood here in the Vlach Rom cultural context. For example, privacy in the western sense does not exist amongst the Vlach Rom. Families often sleep in the same room and visitors can call at any time, even during the night. Everyone knows almost everything about one another. Food and household utensils are often shared and letters opened. The notion of'between you and me' may involve another half-dozen people who are also present but, being cognate to the speaker, would be regarded as one with her or him. Nevertheless, at times of conflict the knowledge of any past breach of Romani rules may be used to gain advantage over another. Thus Vlach Rom social life underlines the primacy of public and communal values, in contrast to the stress laid on privacy in western life. It is important for the following argument to understand that the continuum of Vlach Rom social contexts could be perceived as one that ranges merely from the widely public to the less public. Nevertheless, certain aspects of life - mainly those associated with pollution such as physical love-making or certain illnesses -must remain private knowledge. These subjects are not talked about in public, especially by women in front of men, although men may occasionally boast about their sexual life in a competitive spirit regardless who is present. (Women would talk about private matters only among reliable kin).
The folk view indicates a difference between public and less formal events either explicitly or implicitly through various instructions as to how one should behave in public. A woman might say, for instance, "I will not argue with my husband in front of strange Rom" whilst a man might tell his son, "You must learn, my little boy, how to make merriment nicely, singing well, using your hands to clap, to click your fingers, to dance and beat up your wife if she misbehaves!" I have even heard a child say, "There are times when children must leave the adults to enjoy their own merry-making."
In most cases the nature of a gathering is indicated by the type of food and drink that is served. On grand occasions, such as weddings and funeral rituals, food is compulsory. At weddings a variety of dishes, including
103
.    Romani culture and Gypsy identity
.    desserts, is prepared whilst at a funeral rite the deceased's favourite dish will be offered along with the traditional Rom stewed lamb with paprika. In some instances a relative's dream will reveal the dead person wishes. For remembrance rituals held six weeks and one year after the funeral the Rom who attend eat a meal cooked by a non-Gypsy woman whilst specially invited non-Gypsy guests have a different meal cooked by a Rom woman. Since the latter meal is held to contain the soul of the deceased, any Rom would be horrified to partake of it, even by accident. Whilst a wide range of drinks is available for both events, for a funeral wake male guests must bring also a bottle of retjija [R] or spirit. For calendar festivities like Christmas stuffed cabbage (sarma [R]) and various cakes (kalaco [R]) are made whilst the men drink beer and the women sweet liqueurs. On less formal occasions the refreshments may consist solely of beer for all.
Repertoire as a model for social boundaries
In their song repertoire the Vlach Rom differentiate between two main genres: 'slow' or 'listening' songs and dance songs. The Rom of both communities where I studied label slow songs as mulatoso, meaning revelling or merry-making, or hallgatoso, meaning listening to or quiet. (The same label 'listening song' applies to slow songs among the'Polish Highlanders and in Romania in areas where the Gypsies had an obvious influence on the local musical culture.) These etymologically Hungarian terms have gradually replaced the formerly used Romani term, loki djili, which also means slow song. The mulatoso-hallgatoso category compromises both Vlach Rom songs, the románe djili [R], and Hungarian ones, called either magyar nóta [H] or ungriko djili [R]. The latter are mainly nineteenth century compositions in imitation folk style but also include some Hungarian folk songs. Singers make a distinction between Hungarian and Vlach songs but they do not differentiate between Hungarian folk song and magyar nóta within their Hungarian repertoire. However, Hungarian songs heard on the radio or television are divided into magyar nóta and folk songs (i.e. those that they themselves do not sing).
Stewart (1989) has proposed the crucial insiders' requirement that ;    Vlach slow songs must be performed in Romani. While this is generally true, from an analytical point of view, language cannot be the sole determinant as I have come across many instances where the Vlach Rom use Hungarian in romanes djili and, conversely, Romani in Hungarian songs (see also Hajdu 1958:23; Szegö 1977:272). For the Vlach Rom, melody also plays an important part in determining the nature of the songs, despite the fact that their verbal analysis does not seem to extend to discriminating between different musical structures. Whether singers voluntarily choose, or are instructed by other Vlach Rom to choose, románe djili as opposed to Hungarian songs or, alternatively, whether they insist on expressing the texts of both genres in Romani wherever possible, will depend greatly on the social context of the performance. The age of the singers and the extent to which Romani is used in everyday life will also influence the language of song texts. The elderly, with their knowledge of an older repertoire, are more likely to sing entirely in Romani or at least use very little Hungarian. Members of Community B who are relatively recent migrants to Hungary (Erdös 1989:66) have a much larger Romani Js^vocabulary and speak Romani more than the Rom of Community A and this
ft" OO

carries through into the song texts of each community.
The mulatoso songs known to the two communities I worked with included some Vlach Gypsy songs of non-Hungarian origin (e.g. from Romania) but these were not regarded as a separate category as far as I could determine. The one possible exception was a Serbian song which was in the repertoire of a few older Vlach Rom, one of whom lived in Community B. Members of the family with whom I worked, being distant relatives, had heard the song from her but their knowledge of it was mainly passive. On one occasion when one of the younger Community B women tried to perform a variant of this song with a visiting Rom she lost the melody after a time, no doubt because of its relatively strange musical structure.
The dance songs are labelled khelimaski djili, to use the Romani term or pattogoso (crackling) and pergetöso (rolling) - again words that stem from Hungarian. The Romani term given to dance songs derives from the verb khelel pe (to play). As with the slow mulatoso songs there are many Hungarian songs among the khelimaski djil but the Vlach Rom themselves do not generally make such a clear distinction between them. An important criterion fbT khelimaski djili is that they should be danceable in the Vlach Rom way. Many adopted Hungarian folk songs are now perceived as Vlach Rom and danced to in that manner, whereas those from the popular genre are likely to be designated Hungarian. Thus a Hungarian csárdás [H] is recognised as a Hungarian tune and will be danced to in the Hungarian style but with Romanised steps and movements which results in the Gypsy csárdás. Nevertheless, here too one must take into account the musical occasion, the community and specific individual perceptions. For example, one dance song which, despite its Hungarian song text, was performed in Community A as a Vlach khelimaski djili was rejected as being Hungarian by a Vlach Gypsy woman in another community that I briefly visited. I later discovered that she had earlier had contact with a group of Hungarian researchers making field recordings with Japanese musicologists. Since the emphasis during such short-term visits tends to be on recording only what the researchers themselves consider to be 'authentic' (Kovalcsik 1985:28), I suspect that this may have influenced her judgement.
The Vlach Rom I worked with also keep alive an assortment of Hungarian and western popular songs as part of their occasional dance repertoire. They include a group of Hungarian popular songs from the 1950s that deal with familiar Vlach Rom experiences such as falling in love or being jailed - these are called tangos and are danced to in a non-Vlach style. Another group comprises more modern songs in the magyar nóta or csárdás style, whilst the western imports include a small number of 1950s 'hits', such as Let's Twist Again and Rock Around the Clock, performed as 'party pieces' by middle-aged Rom in a deliberately parodistic style and, for younger Vlach Rom, the current English/American and Hungarian pop-songs, labelled as disco songs. Typically, such foreign imports undergo considerable transformation in performance by the overlaying of a characteristic Vlach rhythmic accompaniment played on spoons, water cans, table tops, etc.
The distinction between Romani songs and adopted non-Gypsy songs is one of the most important and explicitly noted musical boundaries. Nevertheless, there is a subtle generational difference in what is regarded
105
Romani culture and Gypsy icleniiiy
as own and adopted material. Contrary to Stewart's suggestion (1989:95) that the Vlach Rom consider all their own songs to be 'old' (dolmutani), the Rom I have worked with differentiate between 'old' and 'new' songs. The former are often in the repertoire of the older generation whereas new songs usually enter into the repertoire of the younger generation, although both generations share a large number of songs which form a bridge between past and present. We also find that a few Vlach Gypsy songs are sung only by specific families and often kept alive by a single person. The adopted repertoire of the Hungarian Vlach Rom is also an immediate, palpable marker of generational, gender and family differences. Thus older people - mostly, but not exclusively, women - sing mainly Hungarian popular songs with some folk songs among them, whilst middle aged men may sing 1950s hit songs and the younger generation adapt modern pop songs that they hear in discotheques, on the radio, etc.
A Vlach Rom analysis of songs distinguishes between the melody, or 'voice' {glas [R]), the 'words' (vorba [R]), and the particular Romani way of performance including the timbre of the singer's voice. The latter is often regarded as very special, almost as an inherited ethnic characteristic: "You can learn the song but not the way we sing, not our voice," a young Vlach Romani woman told me. However, she also admitted that certain features of performance must be acquired in childhood, suggesting that nurture may matter more than nature. In the Vlach Rom perception there is a stable relationship between melody and text. Every melody has its own words and a good singer should not mix these with other song texts. As one Rom put it, "It is like driving a horse and cart; you must follow a single track instead of moving all over the place," although another person accepted such textual additions, saying, "This was all in this song" or, "In the case of a particular song you can take out verses or add to it." This discourse between the static and flexible verses can lead to debate among singers as to what does and does not constitute a song. Therefore a song, as far as the text is concerned, is perceived both as an externalised and impersonal communal entity and as an individual statement. This supports the analytical observation that a certain degree of flexibility is allowed in text creativity, although not as much freedom as is suggested by some researchers.
The folk analysis accepts individual compositions which 'stem from the life experience of the singers'. When the Vlach Rom talk about a "new song" they are referring both to a new melody with new texts and to an old melody to which a new combination of texts has been fitted. This process, often referred to by the Vlach Rom themselves as "constructing" a new song, is a common enough practice in different parts of the world (e.g. Merriam 1964:181-4; Feld 1982:166). Therefore the western insistence on a 'new song' having a new melody can be misleading when applied cross-culturally.
Whilst one could say that in some instances the Vlach Rom conceptually separate the song from the performance, in practice the two are very much intertwined. Regardless of whether a song is Hungarian or Romani or even a Vlach Romani melody with a Hungarian text, the
„ .performance itself must be perceived of as 'Vlach Gypsy' both in style and spirit. A poor performance can lead to a song that is 'Vlach Gypsy' by all
^.^other measures being rejected by its Rom audience as 'non-Vlach'.

|:tion
Performance style as a model for social interaction All Vlach Rom songs, whether Romani or adopted, have a monophonic structure which, in slow song performance, transforms into heterophony although in north-east Hungary, near the border with Ukraine, Vlach Gypsy songs have shown a recent musical development in which singers spontaneously harmonise the cadences. Contrary to Michael Stewart's observation, the songs are not delivered in unison and they carry a quite different social connotation from that suggested by him. In this, the song performances of Vlach Gypsies resemble the playing of Romungro Gypsy bands which follow the line of the primáš [H], the lead violin player. One of the Rom I know consciously invokes this analogy by crying out, "Banda (H) - wait" when he wants the group to stop singing. Indeed, singing in unison is perceived by the Vlach Rom as being "like Hungarian folk singers on the television" or "priests singing together" which is a gajo, non-Gypsy way of delivery.
In slow songs, a lead singer chooses the song and starts the performance; as the Vlach Rom say, "Only one can drive the horse-cart." It is obligatory first to ask permission to say some 'true' words (Stewart 1987, 1989): "When permission is granted the performance can commence and the lead_singer is then supported by a group of'helpers' who join in slightly behind in each line. Whilst the helping group must never jump ahead of the lead singer, the lead singer reciprocally has to wait for the group to finish .their part before moving on to the next line or verse.
The most frequently verbalised criteria of good performance are tempo and delivery. For a slow song these must be realised 'in a refined way' (finoman [H]), 'slowly' (lokes [R]) and 'peacefully' (csendesen [H]). This applies mainly to the start of a performance after which the dynamics between musical lines are more pronounced. In the area where I worked a basic speed of around 54-60 beats per minute seems to be the desired norm. However, a singer must also pace each line, taking the tempo somewhat faster at the beginning (opre ingrel [R], 'take up') and slowing it at the end (tele ingrel [R], 'take down'). Marked pauses at the end of the second and fourth lines of a stanza and before the very last tone also play a great part in conveying a feeling of tranquillity (Kertész-Wilkinson 1990) although these pauses may also reflect an internal tension or excitement which is expressed by short passages of finger-clicking or clapping that first accelerate in tempo before dying out. The break between the last two tones of a stanza is referred to as a jump (xuttjavel [R]). This term was again explained to me with the simile of driving a horse and cart: "When you want to stop the horse, you must pull back the rein which makes the horse jump a bit." This 'jump' can also be introduced in the middle of a melody line, albeit sparingly, in order to give special emotional intensity and individual character to a performance. Incidentally, certain interjections that are shouted out during singing are the same as commands used for controlling horses; Ho, ho, ho, for instance, can be a warning to singers who press ahead of the lead performer or a sign of appreciation.
Great importance is attached to 'stretching* (cirdel [R]) and to singing with vibrato or rezgetôsen as the Vlach Rom often refer to it in Hungarian. This subtle form of vibrato, which develops into a shake on the longer durational tones, is one of the first things children imitate when they start singing. Singing in an affected way, generally by women, is labelled
107
Romani culture and Gypsy identity
kényes [H] which is a word applied to horses of a frisky, hard-to-handlp tempera-ment. Used in the context of ritual talk between men the same word, however, has a positive connotation as an expression of respect, this time with reference to the noble aspect of a horse's clean 'fussiness'. At the end of a song the lead singer thanks the others for their attention and wishes them good luck and health.
Vlach Romani dance songs also have monophonic melodies but the heterophony of slow song performances is here transformed into a rhythmic polyphony provided by various household utensils; by vocal imitation of instruments, called a mouth-bass; by a technique of rhythmic variation of the tune, called rolling, in stanzas that are without lexical words and by the hand-clapping and foot-stamping of dancers. These different accompaniments variously emphasise the on- and off-beats which go parallel with or against the melody. The dancers imitate these rhythms and step out their own combinations as well; indeed, in a sense they constitute the real focus of the performance.
In the context of large group performances, alongside the requirement to be 'refined', the leader must sing with strong voice (zurales [R]) -another term used when driving horse-drawn carts_- especially as the performance progresses towards its climax when the accolade of "Now s/he really caught it" may be heard. (This seems to be the same whether one is considering the performance of an individual singer's song or the performance event as a whole during which a number of singers may lead.) As regards voice quality, for women a 'thick' quality of timbre is preferred to a 'thin'one since the latter is associated with children's voices. This 'thickness' accords with the Vlach Rom ideal of the female figure and also has connotations of generosity (as opposed to being 'thin-handed' or mean). The members of Community A favour 'high' male voices whereas in Community B a deeper, coarser timbre is more admired.
In the most public contexts the Vlach Rom generally perform only songs that they regard as their own, and mainly in the Romani language, but there are many instances when Hungarian is employed. Even when the Hungarian language is used, the sentiments of the song texts must still accord with Rom public values. Furthermore, at public events the songs performed are mostly those known to all whereas songs belonging to an individual repertoire will be neglected. Nevertheless, there is scope for variation during the whole event as I witnessed both at a wake and at a baptism when visitors were given the chance to perform quasi solo with only close family members helping whilst others listened.
Wakes, along with funerals, are the most significant public occasions for the Vlach Rom, being "bigger than weddings".
Wakes and funerals are the only rites in which musical performances form an obligatory part and are also among the few events where any Rom is welcome, even without an invitation, and feuding must be set aside. Furthermore, the strictest rules are attached to wakes and funerals when only slow songs may be sung whereas during weddings and baptisms, dance songs predominate.
Within the general form provided by the main social context, however, 7 ethere is scope for variations according to the degree of kinship between <^>people and the space allocated for the ritual. During a wake, for example, ^„.^those who are not closely related to the bereaved family may sing slow
108
Song performance: a mode! for social interaction
Singing at a funeral.
The grandson of the deceased leads
the singing. He underlines his feelings
with slrnng physical gestures.
His aunt, standing next to him, tries
to moderate his mouements.
songs with happier, even 'jocular', overtones, especially if a separate room is available. Furthermore, the general rule of limiting the permitted repertoire to slow songs is complemented by a rule demanding the performance-of the deceased person's favourite songs which seem invariably to be dance songs. (Since I had opportunity to study only two wakes I am'reluctant to state that this is always the case without further evidence. Nevertheless, the possibility of incorporating dance songs during the wake as a polarity to slow songs is already evident. Since the dance songs performed during a wake tend to be Hungarian adoptions, this introduces a further paradox in that the deceased is incorporated into the group as a non-Vlach 'outsider'.)
During weddings too there are alternations between dance and slow song performances. The slow songs, in complementary opposition to the shared happiness conveyed by dance songs, mark the different 'nations' and their position within the marriage ritual. At some point the bride's family perform slow songs as a kind of lament for losing a daughter, whilst within the groom's family the younger men sing slower songs together. Those who are not closely related to either groom or bride also have their own 'sessions' of slow song performances, thereby asserting their distinctness among the others.
Thus the musical repertoire is a public demonstration of how the everyday symbolisms of life and death can be reversed yet simultaneously re-established. During a wake the mourners symbolically take part in death by singing slow songs, wearing black clothes, not washing for three days and other displays of'pollution', whilst the deceased is recalled through dance songs which are a metaphor of life. (The complementary polarity between life and death is further emphasised by the fact that non-kin often look for marriage partners for their sons during wakes whilst girls are carefully watched and accompanied everywhere to guard against their being abducted. The combination of death, life and love gives the wake a very intense and unpredictable aura until the morning when the gathering is suspended for a while.) The public rule of mourning extends into the private domain of close relatives of the deceased, who are not allowed to sing dance songs or dance, even in private, for a prescribed period of time. At the end of this period a ritual dance is performed to trample red wine into a garment that has been worn during mourning. Even in this, however, there is some scope for individual choice as to

H  M I!  JÜ  S W
lomani culture and Gypsy identity
whether to abstain for six weeks or a year. This already shows that whilst overall prescriptions exist, they can be stretched in some contexts to incorporate the contradictions or polarities. These contradictions are rationalised as respect for personal needs - all the more so as the Vlach Gypsy kinship system, which reinforces many aspects of musical performance, constantly affirms the right to express and negotiate individual deviations according to rules agreed within each specific community, extended family and nuclear family.
Conflicts, however, remain even in what seem to be clearly private contexts. For example, I once saw a middle-aged Rom man walk angrily out of the room when a young boy began whistling in accompaniment to his performance. Regarding this as an affront both to himself and to the very nature of musical performance, the older man was upholding the public, ritualistic attitude to a musical performance thai the boy interpreted as a private event, although the fact that the two were males of different generations and social status undoubtedly also played a part. In this respect, the incident reflected the use of a private performance to assert authority in the sense that Bloch and Stewart applied solely to public performances. Since that episode, however, the same young 'offender' has grown up to become a much appreciated performer; _during my last visit to his community, in the summer of 1994, middle-aged singers would not sing
without him.
Since the Vlach Rom always consider individuals in relationship to one another, and never as bounded, autonomous entities, the same general rules which govern their everyday actions and feelings in public and private also operate in musical performance. The election of a lead singer is itself governed by a mixture of social rank and musical excellence. At more formal public occasions the leaders of slow songs are normally older men and a few older women, sometimes regardless of musical ability. In this respect they are acting as public representatives of their community and extended family. It is not unusual, however, to have a group of women singing together with a couple of related men joining in softly in the supporting part. The lead is usually passed on according to a hierarchy of social prestige and then community identity. During a wake in Community A no members from Community B took the lead, except one who had lived in Community A for a long time; their low social status put them behind even younger men from other communities. Nevertheless, in most instances those with a reputation for being excellent performers are allowed to escape the social constraints. This may give even a young, unmarried woman the chance to lead if her musical talents are such as to show her family to advantage. In addition, the use of song for consolation can also allow young women to take the lead during singing, especially if they are relatives of the deceased or mourning women. Women may also lead during other sad occasions such as when someone is taken to hospital
or prison.
A male leader's peers may voluntarily provide support but close family members, especially wives, are obliged to show respect by joining the accompaniment. This also gives a woman a chance to draw attention to the unique qualities of her voice and ornamentation and thereby augment her husband's pride in her. A wife will usually do this whenever her husband feels a particular need to sing, even at home. This suggests that musical
II   S   II   II   II   II   M   M   M   M
Singing for an imprisoned relative. This picture has been taken at the climax of a lauf; evening dedicated to the support of both the imprisoned person and his relatives. The lead singer, a young woman, cries bitterly whilst her father (the second man from the front} disguises his emotion by looking at a book. The second support singer (first man in the front) follows her singing with complete emotional involvement. A non-relative observes the event from the bach.
performance both in a wider public and in a very personal context is always regarded, to some degree, as a ritualistic communication and thus subject to seemingly similar rules. Yet different meaning can be ascribed to performances of man and wife in public and private. Whilst in both instances their relationship is based on mutual respect, which the Rom I worked with intertwine with the feeling of love, its function of showing solidarity in front of others in public is in private transformed into one of giving personal psychological support or consolation. It can be perceived as a concealed way of admitting a man's need for his wife's love.
The wife is permitted a wider latitude of choice when the setting is informal where she may even compete with her husband's performance. Furthermore, at some formal as well as smaller public events we find that women in their forties who, by that age have 'grown up' in social standing along with their children, frequently take the lead - especially since they are no longer regarded as posing a sexual threat to the purity of the group.
In the most private contexts we find both men and women performing very personal songs which they admit they would sing only in front of a few trusted Rom.
The most obvious and immediate influence of context is on the song texts. Since all Vlach Romani song texts are built up from sets of more or less fixed couplets, told in the first person and present tense, the differences between expressions of public and private experience are hard to discern for an outsider whether Gypsy or non-Gypsy (see Appendix 1). In public performances the verses sung by men and women alike concentrate on the general Gypsy experience from a male point of view - feeling lonely in the world or having a good time with the 'brothers' at a horse fair. Here a wife is typically portrayed as a disturbing or unreliable figure, as opposed to the 'good' mother or father. The texts of dance songs, in contrast, concentrate on the physical aspect of love. Dance song texts still regard women as unfaithful; however, success in this, for women as for men, is treated with approval in the dancing itself where the aim is to cheat on the partner by getting behind the barrier presented by the partner's arm movements.
Returning to the song text, we find that in a less formal context the same textual couplets can be used to describe individual situations. Men may then transform the feeling of loneliness in the world to one of being
111

ŕ^.fi^ií
■'ü...---i-,£iti!l.
:->*fe*l>i*^t!ft fcfrj.', :-?-.íť.ř. i*. £■&■:£- ^
Romani culture and Gypsy identity
alone among the brothers. Sometimes men even express very private griefs and secrets, such as their powerlessness against or their strong love for and emotional dependence on their wives - the very opposite of the strong and independent character that they strive to project in public - although the ambiguities of the symbols are such that these connotations would be clear only to those who know the individual well. Women in private transfer the general public image of all woman as being bad to another woman and the condition of general loneliness into one of being alone among strangers, that is non-kin, often implying their husbands. In the absence of their husbands, women may also turn the texts around to make fun of the male-dominated public rules, often by means of adopted Hungarian dance songs which express feelings of jealousy, revenge and the like which the Vlach Rom public consciousness ascribes exclusively to non-Gypsies. During a Christmas celebration, which is a formal event amongst related families, one older woman started singing a Hungarian song with the following text:
"I am going to cut off your pumpkin
And get my corn field ploughed.
I am angry with your pumpkin
Because it has climbed into the garden of the woman next door."
^            Later I learned that when she was first married her husband started
■     going after another woman and this song referred to her feelings about this. £v5However, the song was still repeated long after the event, accompanied by j^jmuch joking about its meanings. This kind of remembrance of once painful
.Song performance: ;i model Tor social interaction
or shameful events seems to be a common enough practice among elderly couples. It seems that age permits one to show the imperfect side of human relationships by incorporating and transforming remnants of past pains into jokes.
Quarrels between couples also find their way into songs at less formal or private occasions which could not happen in a public context without incurring disapproval. This appears also to be valid for men singing in public when the aim of the song texts, as Stewart noted, is to express their unity. Nevertheless, even here a competitive element is evident in displays of better delivery of song texts or virtuoso dance steps. Such competitions are often seen at weddings and other happy occasions, as well as at less formal private events where women may be allowed to join in or else compete amongst themselves.
Whilst dance song texts can be more explicit in private than in public, the rules regarding dancing between couples are, more often than not, individually applied due to the sexual connotations of dance. The type and intensity of permitted movements depend largely on the partner's feelings. A jealous man would restrict his wife to dancing with him alone or not at all, whatever the circumstances, whereas a jealous wife can only react in a less formal context. Nevertheless, most men with jealous wives tacitly comply with their wishes and would not risk upsetting them by dancing with others or would do so only in a modest manner.
When we examine melodic variations, including temporal ones, these are less affected by context or gender than by family and community traditions (Appendix 2, Example A). From these influences individuals may develop an idiosyncratic style which they may choose to produce in all contexts (Appendix 2, Examples B and C). Public events are nevertheless important occasions for the Vlach Rom at which to encounter a large number of singers from whom they may learn if they wish.
What emerges from the foregoing is that some parameters of Vlach Rom musical performances, such as the choice of text and who sings with whom, may to some extent be regulated by the context, whereas other features, such as musical and temporal variations, are connected more with the people from whom one learns and with whom one regularly performs.
Thus the structure of various social contexts is both fixed and flexible due to its strong link with the kinship system that guides individual actions, including performance. Hence a less formal context may be transformed into a more formal one by the appearance of an outsider, not
113
rtomam culture ana Gypsy identity
just a visitor from another community but also by the husband entering whilst his wife is singing or even by turning on a tape-recorder. In contrast, the private may involve a large number of people who, being cognate, still represent a homogeneous 'we'. Furthermore, the relationships between individuals, except those between mothers and children, include elements of co-operation as well as competition with others. The paradoxes of human interrelationships can only be enacted simultaneously within musical performance because it operates and proceeds on various layers, such as text, melody, temporal aspect, timbre and kinetics, and thus is able to accommodate similarities and contrasts which would be impossible in speech.
In addition, the performance practices of the slow and dance songs of the Vlach Rom offer further fields for realising different but complementary relationships between people. The two main parts of the slow songs, the 'lead' and the 'support', can also be interpreted as 'masculine' and 'feminine'. Since each part may be performed by either gender, these two polarities become unified in the individual 'human'. This is complemented by the dance songs and the dance which stress the biologically given differences between humans that struggle with each other in order to grow and be united in love. A public group of one gender will tend to glorify their unity at the expense of the other gender, whereas in mixed groups the unity/division of extended families and communities is emphasised. However, it must be borne in mind that such boundaries are fluid and constantly shifting even within the scope of a single event.
Hence, in contrast to the authors cited at the beginning of this article, I regard public performances as representing the diverse statements of a gTOup of individual performers, the less competent as well as the best, who alternate between private and communal expressions through a set of prescribed yet flexible rules. Hence solo performances may be inserted between communal deliveries and either gender may be co-opted to perform at public events out of respect for their social or musical authority. The song texts are left open to interpretation through their use of non-specific symbolism and their grammatical blurring of the difference between the personal and the communal 'ľ. Even when texts and performance practice takes their most general and communal form there is still scope for individual expression through melody, temporal delivery, timbre or a combination of dance steps.
A different picture emerges when one takes account of a full range of contexts. Musical performance encompasses all the socially important rules that divide non-Gypsies and non-Vlach gr-oups from the Rom as well as the internal divisions between genders and generations. In my view, therefore, the social significance of public performances among the Vlach Rom lies partially in the renewal and reinforcement of the ethnic, kinship and gender boundaries of the social structure but also, at the same time, in their transcendence through musical aesthetics and thus in the creation of a model for a more democratic, fuller way of life. The aim in public is to express the same idea, or 'truth' in varied ways, whereas in a private context it is to express different individual experiences, or 'truths', through trie same medium.
C^Vlach Gypsy musical performance is largely a socially created pljaJ)omena. In performance, individual and social identities, personal
Sonfi performance  a model for social interaction
rivalries and the dialectical oppositions between the various factions are enacted and reconciled through the interaction between the communal and individual creativity within the various parameters. A musical performance, however, also has the attribute of creating different relationships between individuals from those emphasised on the social plane. The striving for equality between men - as expressed in everyday existence in the extra-musical speech forms as well as in song texts - is counterbalanced by the subtle hierarchies that emerge in performance structure. Musical performance produces an implicit leadership that would otherwise be challenged in everyday contexts - a leadership that is open mainly, but not exclusively, to men in public, but to all in private. Musical performance also creates not only a subtle hierarchy among men but embodies the possibility of a complementary equality between genders. The possibility of making a different order from the existing ones lies in the central value that Vlach Rom culture attaches to music and individual's musical creativity. Musical authority, which in theory is available to all, can supersede everyday roles and values whilst still acknowledging those roles and values in some of the aspects of musical performance.