2.1Objectives

In addition to filling a knowledge gap by investigating a phenomenon which has previously received little attention, this paper aims to make two interrelated contributions to the field. First, by examining the ways in which Roma communities engage in socialising their younger members in cursing practices, we undertake a pragmatic case study into how the socialisation of ritual practices is performed. Since the seminal work of anthropologists such as Durkheim (1912 [1954/2001])Durkheim, Émile 1912 [1954/2001] The Elementary Forms of Religious Life. Trans. Carol Cosman. Oxford: Oxford University Press. and Turner (1967Turner, Victor 1967 The Forest of Symbols: Aspects of Ndembu Rituals. Ithaca, NY: Cornell University Press., 1969Turner, Victor 1969 The Ritual Process: Structure and Anti-Structure. London: Transactions.), it has generally been accepted that it is vital for communities to socialise their youth in the use of ritual practices, as the basic goal of a ritual is to reproduce social structures. Anthropological research has explored the various aspects of this socialisation process, e.g. by studying the rites of passage (e.g. Weil 1986Weil, Shalva 1986 “The Language and Ritual of Socialisation: Birthday Parties in a Kindergarten Context.” Man 21 (2): 329–341. ). However, to the best of our knowledge, relatively little research has been undertaken into the socialisation of ritual as an interactional process (but see, e.g. Réger 1999Réger, Zita 1999 “Teasing in the Linguistic Socialization of Gypsy Children in Hungary.” Acta Linguistica Hungarica 46: 289–315. ), despite the fact that ritual is essentially an interactional phenomenon (Kádár 2013Kádár, Daniel Z. 2013 Relational Rituals and Communication: Ritual Interaction in Groups. Basingstoke: Palgrave Macmillan. ).44.Note, however, that previous research such as Blum-Kulka (1997)Blum-Kulka, Shoshana 1997 Dinner Talk: Cultural Patterns of Sociability and Socialization in Family Discourse. New York: Lawrence Erlbaum. and Schieffelin (1986)Schieffelin, Bambi B. 1986 “Teasing and Shaming in Kaluli Children’s Interactions.” In Language Socialization across Cultures, ed. by Bambi B. Schieffelin and Elinor Ochs, 165–181. Cambridge: Cambridge University Press. has investigated, to a certain degree, the socialisation of certain ritual phenomena such as dinner-table talk, although such research has not pursued an interest in ritual per se. In this paper, we demonstrate that, due to the harm which is associated with ritual cursing (see above), the socialisation of this ritual practice requires adults to actively uphold what experts in the language socialisation of children, such as Eisenberg (1986Eisenberg, Ann R. 1986 “Teasing: Verbal Play in Two Mexicano Homes.” In Language Socialization across Cultures, ed. by Bambi B. Schieffelin and Elinor Ochs, 182–198. Cambridge: Cambridge University Press., 190), define as a ‘safe context’ (Section 3). While this paper is only a case study, it is perhaps not too far-fetched to argue that involvement in interactionally creating a safe context – which is at the centre of our analysis – may be valid for socialising young language users in a variety of rites across different languages and cultures. This may involve even mundane rituals in urban settings, such as socialising adolescents in dealing with unwanted ritual sexual advances, which has been an important aspect in sociological inquiries (e.g. Letendre 2007Letendre, Joan 2007 “ ‘Sugar and Spice but Not Always Nice’: Gender Socialization and Its Impact on Development and Maintenance of Aggression in Adolescent Girls.” Child and Adolescent Social Work Journal 24 (4): 353–368. ).

Second, by examining cursing as a form of ritual teasing, and by paying particular attention to how language users uphold a safe context, we intend to contribute to the continuing research on teasing (see an overview in Haugh 2017Haugh, Michael 2017 “Teasing.” In The Routledge Handbook of Language and Humor, ed. by Salvatore Attardo, 204–218. New York: Routledge. ). We interpret teasing according to Eder’s (1993Eder, Donna 1993 “ ‘Go get ya a French!’: Romantic and Sexual Teasing among Adolescent Girls.” In Gender and Conversational Interaction, ed. by Deborah Tannen, 17–31. Oxford: Oxford University Press., 17) definition, as a “playful remark aimed at another person, which can include mock challenges, commands and threats as well as imitating and exaggerating someone’s behaviour in a playful way”. As the extracts studied in this paper reveal, mock challenges (e.g. Extract 3), threats (particularly all forms of ritual cursing) and mimetic imitations and exaggerations (e.g. Extract 9) are integral parts of cursing. Arguably, the teases that are studied here are noteworthy because of the potential danger and harm that is attributed to ritual cursing. As Haugh (2017)Haugh, Michael 2017 “Teasing.” In The Routledge Handbook of Language and Humor, ed. by Salvatore Attardo, 204–218. New York: Routledge. highlights, relatively little research has been undertaken into the ritual forms of teasing, despite the fact that ritual teasing is worthy of further exploration due to its relative inaccessibility to cultural outsiders. Haugh (2017Haugh, Michael 2017 “Teasing.” In The Routledge Handbook of Language and Humor, ed. by Salvatore Attardo, 204–218. New York: Routledge. , 209) points out that

Research on teasing, particularly in anthropology, suggests that there may be particular ritualised forms of teasing that may not be readily interpretable or accessible to cultural outsiders. For instance, ‘razzing’ amongst Native American Indians (Pratt 1996Pratt, Steven 1996 “Razzing: Ritualized Uses of Humour as a Form of Identification among American Indians.” In Interaction and Identity, ed. by Hartmut B. Mokros, 237–255. London: Transaction.), ‘name-calling’ amongst the Kaluli of Papua New Guinea (Schieffelin 1986Schieffelin, Bambi B. 1986 “Teasing and Shaming in Kaluli Children’s Interactions.” In Language Socialization across Cultures, ed. by Bambi B. Schieffelin and Elinor Ochs, 165–181. Cambridge: Cambridge University Press.), or ‘kin(ship)-based’ teasing amongst Australian Aborigines (Garde 2008Garde, Murray 2008 “The Pragmatics of Rude Jokes with Grandad: Joking Relationships in Aboriginal Australia.” Anthropological Forum 18 (3): 235–253. ).

This inaccessibility could be manifested in the previously discussed stereotypes that surround Romani cursing. We argue that the ritual curses in teasing can serve as playful models of conflict talk, particularly when they are produced by women (Eder 1993Eder, Donna 1993 “ ‘Go get ya a French!’: Romantic and Sexual Teasing among Adolescent Girls.” In Gender and Conversational Interaction, ed. by Deborah Tannen, 17–31. Oxford: Oxford University Press.) as a form of gendered ritual language use.55.While it is beyond the scope of this paper to explore the gendered aspects of cursing and their implications for research into the ideologies of gendered language (Cameron 2005Cameron, Deborah 2005 “Gender and Language Ideologies.” In The Handbook of Language and Gender, ed. by Janet Holmes and Miriam Meyerhoff, 447–467. Oxford: Blackwell.), it is evident that gender inequalities in androcentric Gabor Roma society have a key role in the contextually situated disarming/teasing applicability of curses (see also Section 4). As the central focus of our analysis, we explore how speakers of Romani are socialised to differentiate between serious and non-serious ritual cursing, with the aid of a cluster of contextualisation cues (see an overview in Gumperz 1992Gumperz, John J. 1992 “Contextualization and Understanding.” In Rethinking Context: Language as an Interactive Phenomenon, ed. by Alessandro Duranti and Charles Goodwin, 229–252. Cambridge: Cambridge University Press., 2003Gumperz, John J. 2003 “Contextualization Conventions.” In Sociolinguistics: The Essential Readings, ed. by Christina Bratt Paulston and Richard G. Tuker, 139–155. Oxford: Blackwell.; Levinson 2003Levinson, Stephen C. 2003 “Contextualizing ‘Contextualization Cues’.” In Language and Interaction, ed. by Susan L. Eerdmans, Carlo L. Prevignano, and Paul Thibault, 32–39. Amsterdam, John Benjamins. ) – such as paralinguistic, non-verbal and interactional features (e.g. Haugh 2016Haugh, Michael 2016 “ ‘Just kidding’: Teasing and Claims to Non-Serious Intent.” Journal of Pragmatics 95: 120–136. ) – which are meant to uphold the safe context. As we highlight in Section 3, there is a narrow and somewhat blurred border between the ‘real’ and teasing functions of cursing due to the ritualistic nature of this phenomenon, and therefore contextualisation cues are essential for language users to (acquire the ability to) discern the type of cursing that is being used in a particular interaction. Ritual cursing is important for the pragmatic research of teasing because, due to the power of ritual cursing to cause harm and death, in scenarios where curses are used as teases it is not merely the offence, but rather the more serious feeling of physical threat, that the person being socialised needs to be able to distinguish from humour. In addition to contextualisation cues, we also examine the way in which ritual teasing is kept separate from other parts of the interaction, which contributes to the disarming of the ritual curse.

2.2Data and methodology

This paper is based on fieldwork that the second author undertook over a period of nearly three years during the early 2000s amongst Gabor Roma communities (Szalai 2014Szalai, Andrea 2014 “Ideologies of Social Differentiation among Transylvanian Gabor Roma.” Acta Ethnographica Hungarica 59 (1): 85–112. ). These communities live in Transylvania, a multi-ethnic and multilingual part of Romania, and these communities are neighbours of the ethnic Hungarian minority population. Therefore, they are usually Romani-Hungarian-Romanian trilingual speakers. Their first language (the Gabor Romani dialect) is a lesser known Vlah Romani dialect (Matras 2002Matras, Yaron 2002 Romani: A Linguistic Introduction. Cambridge: Cambridge University Press. ). The second author examined the language ideologies and practices in three closely interconnected Gabor communities in Mureș (in Hungarian: Maros) county. She spent the majority of her time in the settlement which the Gabor Roma refer to as Baro Gav ‘the Big Village’, but she also established contact with many other families in different villages due to the kinship networks of the local Romani families. She also participated in family events which were organised in other Transylvanian settlements (e.g. in Cluj, Oradea, Huedin) and visited Gabor families in other locations. It should be noted that, henceforth, when the generic term ‘Roma’ is used, it refers to the Gabor Roma.

During her fieldwork, the second author encountered cursing being used as a ritual practice for language socialisation on an almost daily basis. However, due to ethical reasons, she had limited opportunity to record cursing in various naturally occurring contexts such as business negotiations and conflict situations; in total, the dataset of teasing encounters which was used in this paper consists of 3 long (i.e. >100 lines) and 32 shorter (<100 lines) interactions in which adults socialise young language users in the practice of ritual cursing. For ethical reasons, the second author was only able to audio record ethnographic interviews and certain types of community events (e.g. death-related rituals and ritual oaths) where the participants gave consent to being recorded, and during such interviews mainly other, non-offensive forms of cursing occurred. Fortunately, for the current research, some interviews transformed into naturally occurring interactions as neighbours called in and became involved. As is attested by anthropological research (e.g. Berta in press; Stewart 1997Stewart, Michael S. 1997 The Time of the Gypsies. Oxford: Westview Press.), in various Roma communities ‘privacy’, in the Western European sense, has less significance: members of extended families, neighbours and friends continuously visit each others’ houses to request help or assistance, to engage in transactions or simply for a friendly chat. Thus, on a number of occasions, unexpected visitors appeared after audio recording had commenced. All newcomers were made aware that conversations were being audio recorded, but soon after the interactions generally became informal as the local people were closely acquainted with the fieldworker.

Since our research focuses on the interactional operation of contextualisation cues, we have transcribed our data by using the following transcription symbols (Atkinson and Heritage 1999Atkinson, J. Maxwell, and John Heritage 1999 “Jeffersons’s Transcript Notation.” In The Discourse Reader, ed. by Adam Jaworsky and Nicolas Coupland, 158–166. London: Routledge.):

3.1Forms of address

In the context of Romani ritual teasing, forms of address are often used as contextualisation cues. For instance, the interaction under investigation begins with such a form of address, as follows:

1 Teri     ((to Kati, shouting from the street))
           
                                                    SO KĂRĂN TUMENGĂ, ŔOMNJE?
           What are you doing, you, Gypsy woman? [lit. ‘What are you doing to yourselves,
           Gypsy women?]
2 Kati     ((to Teri))
           Dikh, katika horbijma.
           Look, I am chatting here ((with the researchers)).
3 Teri     ((Teri is outside in the courtyard, but already initiates the teasing))
           ((in a whiny voice))
           ŔOMEHTE GĂLĀH E PITJŌKA:!
           Pitjoka got married! ((Pitjoka is Zsuzska’s grandmother, a married woman in her late
           fifties at the time of the interview))
4 Kati     In- gălahtar!
           She was taken.
5          Ingărdah la o phuro Pišta, dikhăh?
           She was taken by the old Pista,66.Pista is a recently deceased old man and one of Teri’s relatives. you see?
6 Teri     Ige::n!=
           Yeess!
7 Zsuzska  =Na!
           No!

This was not the first time that day that the interactants had met each other: Teri and her son repeatedly called in, and during each of these visits they stayed for a friendly chat. In the above interaction, Teri begins the conversation with a ritual tease, by using the address term ʻyou Gypsy woman!’ (ŕomnje).77.The noun ŕomnji (in Vocative case: ŕomnje!) includes differentiation on the basis of ethnicity, gender and marital status: it refers to ‘a married Romani woman’. This form of address would be regarded as being derogatory in inter-ethnic settings, but its ritual intra-ethnic use displays solidarity and humour.88.Such expressions tend to be frequently used by socially disadvantaged groups; see Croom (2013)Croom, Adam 2013 “How to Do Things with Slurs: Studies in the Way of Derogatory Words.” Language & Communication 33 (3): 177–204. . It is perhaps not a coincidence that ‘you Gypsy woman’ occurs at the beginning of the interaction: it teases the hostess Kati rather than Zsuzska, and as such its primary aim is to set the tone of the interaction. Importantly, ŕomnje has a contextually embedded meaning because it operates as a playful challenge: due to the importance of gender status in the Romani community (cf. Section 1), it is normally used when addressing married females. Kati was an unmarried girl at the time of the interview, but she fulfilled duties – including babysitting and overseeing the household, including her elder brothers’ wives and children – which were regarded as being typical ‘matronly’ roles.

Following this initial contextualisation cue, Teri and Kati begin to tease the child in lines 3–6, by switching to a much simpler form of interaction: they deliver the clearly false claim that the child’s grandmother (a married woman in her late fifties) has remarried. In line 7, Zsuzska appears to realise that the teasing has started, as she shouts ‘no’ to Teri.

Forms of address also operate as contextualisation cues to benchmark changes in the ritual ‘line’ (Goffman 1967Goffman, Erving 1967 Interaction Ritual: Essays on Face-to-Face Behavior. New York: Anchor Books.) of an interaction: when a socialising rite of aggression becomes too intensive, the speakers may deploy forms of endearment to indicate to the person being teased that the interaction is not serious. For instance, after the initial tease, Teri begins to curse the girl’s grandmother (using her Romani name Pitjoka), and Kati immediately joins in the ritual game; as their curses become increasingly menacing, the child – who initially understands that these curses are part of a game – begins to show signs that she is taking the curses seriously. Once Kati is aware of this, she immediately uses an endearing form of address together with a code switch to the teaser (lines 99, 101), to decrease the intensity of the ongoing ritual by giving the child a cue that the ongoing interaction is not meant to cause harm:

93 Kati     Xal o beng adjeh, hi:::!
            May the Devil eat today, huh!
94 Teri     Mula::h, mula:h, [ja:::j! ((pretends to be crying))
            She has died, she has died, oy!
95 Zsuzska  ((laughs))       [@@@=
96 Kati     Mulah?
            Has she died?
97 Zsuzska  ((partly crying, partly laughing))
            =Na!
            No.
98 Teri     Mulāh tji mami e Pitjōka:!
            Your grandmother Pitjóka died!
99 Kati     ((to Teri, laughing, in Hungarian)
            Ne még mondjad úgy, mindjárt sír!
            Do not tell it to her anymore, she is going to cry!
100 Zsuzska Na!
            No!
101 Kati    ((to Zsuzska, consoling her))
            
                                                    Na:, śej, či mulah!
            No, girl, she has not died!
102 Teri    Mulāh tjo Pōko!
            Your Spider has died!

In lines 93 and 94, Kati and Teri playfully tease Zsuzska by ritually cursing her grandmother. In line 95, Zsuzska responds to the curses with laughter but also begins to cry, indicating that her laughter is ‘more than a simple perception of humour’ (Glenn 2003Glenn, Phillip 2003 Laughter in Interaction. Cambridge: Cambridge University Press. , 23). As the chain of the interactional ritual exchange intensifies (Collins 2004Collins, Randall 2004 Interaction Ritual Chains. Princeton, NJ: Princeton University Press. ), Zsuzska appears to be confused as to whether the cursing is meant to be harmful or not. In line 96, Kati notices this confusion: she asks Zsuzska whether Teri’s claim (performed in an intentionally overexaggerated tone; see Section 3.3) that Zsuzska’s grandmother has died is true, to which Zsuzska immediately answers ‘no’, in line 97. However, Teri’s next curse in line 98 forces her to the brink of tears again, and in line 99 Kati intervenes to decrease the ‘pressure’ (Collins 2004Collins, Randall 2004 Interaction Ritual Chains. Princeton, NJ: Princeton University Press. , 233) of the interactional ritual chain on the child, by advising Teri to reduce the intensity of the curses. Note that she does not say to stop the ritual, but rather suggests making the curses less menacing, and Teri follows this advice in line 102. When making this suggestion, Kati switches from using Romani to Hungarian, a typical indexical of social negotiation (Scotton-Myers 1988Scotton-Myers, Carol 1988 “Code-Switching as Indexical of Social Negotiation.” In Codeswitching: Anthropological and Sociolinguistic Perspectives, ed. by Monica Heller, 151–186. Berlin: Mouton de Gruyter. ): switching to a cultural outsider’s code in line 99 is an indicator that this utterance is not part of the ongoing ritual. In line 100, Zsuzska angrily denies that her grandmother has passed away and, in response, Kati steps out of the ritual line (Goffman 1967Goffman, Erving 1967 Interaction Ritual: Essays on Face-to-Face Behavior. New York: Anchor Books.). In line 101, she supports the teased child by uttering an explicit withdrawal (‘she has not died’), which indicates that the previous claims were meant to be a joke (Haugh 2016Haugh, Michael 2016 “ ‘Just kidding’: Teasing and Claims to Non-Serious Intent.” Journal of Pragmatics 95: 120–136. ), and she uses the form of endearment śej (‘girl’), which would be far too ‘mild’ to be used in aggressive teasing, i.e. it operates as a contextualisation cue for the child that Kati is not now teasing her.

3.2False information delivered via interactional ritual chains

Delivering false information is a fundamental contextualisation cue by which adults can uphold the safe context. As the following Extract (3) illustrates, repeatedly providing false information via a chain of increasingly stronger ritual utterances, could pave the way for the disarmed cursing to be kick-started:

8  Kati     Ingărdah la!(.)
            He took her away!
9           The či na phendāh pehkă Žužkakă!
            And she did not say anything to her Zsuzska!
10 Teri     ((enters the house))
            Na.
            No. ((She did not mention it.)) 
[...]
17 Kati     ((to Teri, offering her a seat))
              Haj [Teri, beš tele!
              Come, Teri, sit down!
18 Teri           [Bešah.
                  Ah, that’s life. 
19 Kati       Beš tele!
              Sit down!
((Teri stops in the middle of the kitchen, facing little Zsuzska, Kati is between them.))
20 Teri     ŔOMĒHTE GĂLĀH E PITJŌKA::!
            Pitjoka has married!
21 Zsuzska  Na.
            No.
22 Teri     O Del [te feril la!
            (May God keep her!) ((the ‘newly-wed’ grandmother))
23 Kati           [Ingărdāh la o:: Pišta, the o Mati, the [o Gabi.
                  Pista, and Mate and Gabi took her away.
24 Teri                                                   [Ingărde la o bară njamcî!
                                                     She was taken away by the great Germans!
25 Zsuzska  ((with laughing intonation)
            Ige::n.=
            Yes. ((Ironically, pretending agreement: Oh, of course!))
26 Kati     ((laughing))
            =O turči!
            The Turks!
27 Teri     ((laughing))
            Merel tji dej! @@[@
            May your mother die!

As was noted in the analysis of Extract (1), even before Teri enters the house, she delivers a saliently false utterance upon seeing Zsuzska, by arguing that a recently deceased relative (the old Pista) has married Zsuzska’s grandmother (line 5). In lines 8–10, Kati and Teri continue the tease. After Kati invites Teri to sit down (lines 10–19), the women again engage in the tease, which ultimately transforms into cursing in line 27. Following Teri’s mock congratulations for the ‘newly-wed’ grandmother (line 22), Kati playfully identifies various men from the child’s family (‘Pista, and Mate and Gabi’) who assisted with the grandmother’s imaginary marriage by helping with her elopement (line 23). One of these men, Gabi, is Zsuzska’s father, with the others being her uncles; all these men are Kati’s own brothers, which further increases the falsehood of the information, thus enabling the child to recognise the interaction as teasing. As a mimetic response in the interactional ritual chain, in lines 24 and 26 Teri and Kati name the ‘great Germans’ and the ‘Turks’ as having eloped with Zsuzska’s grandmother; these are again references to Zsuzska’s above-mentioned relatives, who at that time had travelled to Germany and Turkey on business. This subversive play with ethnic identities further increases the humorous stance adopted in the interaction. Delivering clearly false information proves to be an efficient strategy to engage the child in the ritual: while initially in line 21 Zsuzska is startled by the announcement that her grandmother has married, in line 25 she joins in with the ritual teasing game by uttering a mock agreement, which in turn allows the ritual of cursing to be kick-started in turn 27. Note that the initial tease is not completely harmless. When the interaction took place, both the participants (and the ethnographers) knew that Zsuzska adored her grandmother more than anyone else in her family, i.e. the ‘marriage’ of her beloved grandmother could cause Zsuzska relational and emotional damage.

Ritual falsehood and repetitive chains (Collins 2004Collins, Randall 2004 Interaction Ritual Chains. Princeton, NJ: Princeton University Press. ) are closely interrelated in the rites of socialisation. This is in accordance with Eisenberg’s (1986)Eisenberg, Ann R. 1986 “Teasing: Verbal Play in Two Mexicano Homes.” In Language Socialization across Cultures, ed. by Bambi B. Schieffelin and Elinor Ochs, 182–198. Cambridge: Cambridge University Press. findings regarding teasing in socialisation, i.e. the repetition of clearly false information helps the recipient child to identify the tease and, as such, it operates as a contextualisation cue. This repetitiveness results in a limited choice of topics and a sense of redundancy in our dataset (compared to, for example, scenarios in which teasing is more hidden within the interaction): the playful threat and provocation is usually organised around prototypical themes. For instance, in the interaction analysed in this paper, 61 lines feature curses; that is, 53.1% of the 114 utterances include curses and the majority of the other lines are metapragmatically centred on cursing.

3.3Nonverbal contextualisation cues

The contextualisation of cursing also manifests itself vis-à-vis interrelated nonverbal prosodic features. These include, perhaps most importantly, laughter (see an overview in Glenn 2003Glenn, Phillip 2003 Laughter in Interaction. Cambridge: Cambridge University Press. ): the analysis of our dataset reveals that both adults burst into laughter as they deliver the teasing curses. This not only indicates that these are playful in nature, but also that the adults find it humorous to use otherwise dangerous curses in a safe context. For instance, in the following extract Teri bursts into laughter as she engages in the playful rite of cursing:

49 Teri   [Merel e Pitjōka! @@@@
          May Pitjoka ((grandmother)) die!
50 Kati   ((to Zsuzska))
          
                                                    Del armaje?
          Is she cursing?

Teri’s disarming laughter is noteworthy from a ritual point of view: as previous research, such as Emmons (2000Emmons, Sally L. 2000 “A Disarming Laughter: The Role of Humor in Tribal Cultures. An Examination of Humor in Contemporary Native American Literature and Art.” Doctoral thesis, University of Oklahoma., 107–146) has illustrated, with regards to the socialisation of the ritual acts of aggression, people with various cultural backgrounds tend to engage in this type of laughter. In line 50, Kati metapragmatically confirms (cf. Silverstein 1993Silverstein, Michael 1993 “Metapragmatic Discourse and Metapragmatic Function.” In Reflexive Language: Reported Speech and Metapragmatics, ed. by John Lucy, 33–58. Cambridge: Cambridge University Press. ) the disarmed nature of the curse, by asking for clarification from the child as to whether she has understood what is going on.

As a contextualisation cue, laughter also operates as a form of ‘crossing’ (Rampton 1995Rampton, Ben 1995 “Language Crossing and the Problematisation of Ethnicity and Socialisation.” Pragmatics 5 (4): 485–513. , 495) in our data: for example, there are instances where the adult (typically Teri) delivers curses with a chuckling and rising intonation, and by doing so she imitates the voice of a child who is still unable to properly curse:

81 Kati     Ašun, dah armaje tje mama!
            Have you heard it, she has cursed your mum!
82 Teri     @@@
            ((laughs))
83 Zsuzska  (Ande dej!)
            ( Into mother!)
84 Teri     ((with laughing intonation))
            Merel tji dej@@@!
            May your mother die!

In this part of the conversation, Zsuzska again becomes uncertain as to whether the cursing is meant to cause harm because she is clearly concerned for her mother (line 83), and Teri then softens the situation by repeating her curse in a childish fashion.

Together with laughter, stress and an exaggerated exclamative intonation are also typical prosodic contextualisation cues in our dataset. For instance, in the extract below, which occurs after Extract (5), Teri performs a childish crossing to ensure that the ritual is harmless to the child:

87 Teri   [Merel tji mami:!
          May your grandmother die!

The prosody of mami:! (grandma:a!) imitates how a child would call for her grandmother in an emergency situation; i.e., in the present context it playfully indicates to the child that, while the curse may appear to be harmful, it is actually harmless. The pragmatic power of such contextualisation cues to create a safe context becomes clearly evident in instances when they aid the child to counter-attack with curses – this is arguably a challenging task for a young child when one considers the harm that is attributable to ritual curses. For instance, Zsuzska appears to be particularly clever in one part of the interaction – she not only understands that the ritual cursing is playful, but also launches a mocking counter-attack:

27 Teri     ((laughing))
            Merel tji dej! @@[@
            May your mother die!
28 Zsuzska                   [
                                                    Tji dej, ja?=
                             ((May)) Your mother ((die))), alright?
29 Kati     ((Repeats/echoes little Zsuzska’s words, with laughing intonation))
            =Tji dej, [ja? @ @
            Your mother, alright?

Teri’s use of prosodic contextualisation cues in line 27 are standard for the interaction (see, e.g. Extract 6). However, in line 28 Zsuzska transforms the flow of the interaction, as she playfully requests clarification from Teri regarding whether it is the latter’s mother whom they are cursing. Kati acknowledges this smart counter-attack by repeating Zsuzska’s utterance, whilst laughing, in line 29.

Finally, a related prosodic contextualisation cue that is frequently deployed in our dataset is the lengthening of vowels (cf. Gumperz 1992Gumperz, John J. 1992 “Contextualization and Understanding.” In Rethinking Context: Language as an Interactive Phenomenon, ed. by Alessandro Duranti and Charles Goodwin, 229–252. Cambridge: Cambridge University Press., 235):

94 Teri   Mula::h, mula:h, [ja:::j! ((imitates crying))
          She has died, she has died, oy!

In this instance, Teri playfully imitates crying as she mockingly cries for Zsuzska’s grandmother. It should be noted that, while this represents the default use of lengthened vowels in our dataset, they can occasionally fulfil another function, as the following extract illustrates:

49 Teri      [Merel e Pitjōka! @@@@
              May Pitjoka ((grandmother)) die!
50 Kati      ((to Zsuzska))
             
                                                    Del armaje?
             Is she cursing?
51 Zsuzska   ((in a crying voice))
             Ige::n.
             Yes.
52 Kati      ((to Zsuzska))
             Hi::, dah armaje amară mame!=
             Huh, she cursed our grandmother!
53 Zsuzska   Ige::n!=
             Yes!

When Zsuzska becomes confused and is threatened by the curse, Kati switches to an endearing style of address: in line 52 she aligns herself with Zsuzska by switching to the plural form (she refers to the cursed person as ‘our grandmother’) and also by using the vowel lengthened form hi:: (huh) to elicit Zsuzska’s reflection on the curse. As experts of child language acquisition, such as Blount and Padgug (1977)Blount, Ben G., and Elise Padgug 1977 “Prosodic, Paralinguistic, and Interactional Features in Parent–Child Speech: English and Spanish.” Journal of Child Language 4 (1): 67–86. , have highlighted, vowel lengthening is particularly prevalent in the language behaviour of parents and other caretakers in contexts of uncertainty, and it is likely that in the above interaction hi:: (huh) indexes a protective stance towards Zsuzska.

3.4The ritual contextual frame

So far, our focus has been on the contextualisation cues which are used to ensure that cursing is kept within a safe context. In addition, we will now briefly discuss the way in which cursing as a ritual practice is demarcated from other parts of the interaction. As the extracts that we have studied have illustrated, ritual cursing is emotionally loaded (Collins 2004Collins, Randall 2004 Interaction Ritual Chains. Princeton, NJ: Princeton University Press. ), and the adults in the interaction continuously need to support and reassure Zsuzska – by using both contextualisation cues and metapragmatic comments – that the curses are not harmful. Consequently, it is also important that such rites of aggression are demarcated from ‘ordinary’ (non-harmful) conversation, that is, they take place in a ritual contextual frame (Turner 1979Turner, Victor 1979 “Frame, Flow and Reflection: Ritual and Drama as Public Liminality.” Japanese Journal of Religious Studies 6 (4): 465–499. ).

Initially, the teasing is simply concerned with the alleged marriage of the girl’s grandmother, and ritual cursing does not commence until the teased girl first begins to doubt the truth of the tease. In other words, in terms of the ritual frame, the child is expected to accept the ritual challenge in the form of a counter-challenge (Turner 1979Turner, Victor 1979 “Frame, Flow and Reflection: Ritual and Drama as Public Liminality.” Japanese Journal of Religious Studies 6 (4): 465–499. ), so that the rite of cursing can begin in a playful form with many embedded contextualisation cues. It is relevant here to revisit the following part of the interaction (featured in Extracts (3) and (7)):

24 Teri      [Ingărde la o bară njamcî!
             She was taken away by the great Germans!
25 Zsuzska   ((with laughing intonation)
             Ige::n.=
             Yes. ((Ironically, pretending to agree: ‘Oh, of course!’))
26 Kati      ((laughing))
             =O turči!
             The Turks!
27 Teri      ((laughing))
             Merel tji dej! @@[@
             May your mother die!
28 Zsuzska   [
                                                    Tji dej, ja?=
             ((May)) Your mother ((die))), alright?
29 Kati      ((Repeats/echoes little Zsuzska’s words, with laughing intonation))
             =Tji dej, [ja? @@
             Your mother ((not mine)), alright?

Curses begin when, in line 25, Zsuzska finally ‘accepts’ Teri’s challenges by responding ironically to her tease. This interactional ‘green light’ paves the way for the first ritual curse of the interaction, which is heavily loaded with paralinguistic contextualisation cues (see the previous section). When Zsuzska responds to the curse with a humorous counter-challenge in line 28, and Kati credits this response (line 29), Teri begins to deliver curses which are heavily loaded with contextualisation cues:

30 Teri      ((laughing)) 
             [Merel tji dej! @@@
             May your mother die!

During the interaction, Zsuzska becomes deeply involved in the ritual practice, as a number of the extracts have demonstrated. As the interaction intensifies, Zsuzska takes Teri’s curses to be increasingly menacing and, from line 102, the adults gradually slow down the ritual, ultimately leaving the ritual frame in line 111, as the following extract illustrates:

102 Teri     Mulāh tjo Pōko!
             Your Spider has died!
103 Zsuzska  Na!
             No!
104 Teri     [Mulāh! =
              Died!
105 Kati     [Gălāh ando fōro!
              Went to the town!
106 Teri     =
                                                    Na-j kon t’anel tukă xabe:n!
              There is nobody to bring food to you!
107 Kati     Vi o Kalapošo, vi o Kalapošo mulah?
             The Hatted, has the Hatted also died?
108 Teri     Mulah v’o Kalapošo:!
             The Hatted has also died!
109 Zsuzska  Na.
             No.
110 Kati     @@@ (1.0)
             ((laughs))
111 Teri     ((to her son, leaving, moving towards the door))
             No hajdi!
             Well, come, let’s go!
112          ((to the hosts))
             Ret laśi!
             Good night!
113 Kati     Źe la pačesa!
             Go with peace!

As Zsuzska’s response becomes increasingly defensive – the emotive Na! (No!) would indicate that she is beginning to take the curses seriously – Teri uses a play on language, in line 102, as a contextualisation cue delivering false information (Section 3.2). In this case, she deploys humorous ambiguity when she claims that it was the girl’s grandfather and not her grandmother who had died. The adult teaser assumes the child’s perspective, and in the mock curse she uses the child’s own, simplified idiolectal term (Pōko instead of Pitjōko) for naming her grandfather.99.In several Romani communities, including those of the Gabor Roma, people tend to have a name for in-group (Romani name) and another for out-group (Gaźo/non-Romani name) situations. A person may have more than one Romani name. Some Romani nicknames are reserved mainly for family settings, while others are widely used in the community. The informal, familial Romani nickname of Zsuzska’s grandfather is Pitjōko. She had difficulty pronouncing this name, and therefore she often used a shorter, simplified form of the name Pitjōko>Pōko. This child language form of the name was a source of amusement for the other interactants, as this shortened form is phonologically identical to the noun pōko ‘spider’. The grandfather had another Romani name, Kalapošo ‘Hatted’, and this name was widely used in the community. In this context, when uttered by an adult person, the curse becomes a play on language and a further source of humour, as it could be understood literally to mean: ‘May your spider die!’ Besides, Kati engages in the ritual chain in line 105, by toning down the concept of death (‘Went to town’, rather than ‘died’), and the two adult participants co-construct a reformulated version of the message, by replace another more widely known Romani name with a nickname (Hatted) which is referring to Zsuzska’s grandfather. In this instance, it is Kati who states that Hatted (her own father, Zsuzska’s grandfather) has died, to further downplay the seriousness of the message. The emotive and relational tie between Kati and Hatted is meant to signal that the cursing is playful, as no member of the Roma community would curse their own relatives. Zsuzska continues to feel threatened and acts defensively, and so Kati begins to laugh in line 110. Following a conclusive pause, Teri terminates the ritual in line 111 by telling her son to leave, after which she engages in a leave-taking ritual with Kati. This takes the form of a blessing, and as such it is clearly demarcated from the previous curses.