Dilemmas of Corpus Construction beyond Folklore Collections: Threat as a Speech Act in Early Modern Witchcraft Trials

: During the corpus-building operation of the Digital Database of Hungarian Verbal Charms we tried to augment the available material by the inclusion of witness statements of witch trials conducted in early modern Bihar County and the town of Debrecen. My paper explores the kinds of dilemmas and issues we were faced with concentrating especially on generic questions of verbal charms. As regards the exploration of early modern written sources of vernacular language use the most relevant recent approaches came from historical speech act research. Therefore, in the context of the corpus building project I shall also discuss to which extends the results of historical pragmatics, historical speech act research can offer any help (and if so, what kind of help) in solving the generic problems and questions of verbal magic. 1


INTRODUCTION
In constructing the Digital Database of Hungarian Verbal Charms the primary goal of the "Vernacular religion on the boundary of Eastern and Western Christianity: continuity, changes and interactions" ERC research group 2 was to render the verbal charm texts of earlier folklore collections and archives accessible to researchers in the form of a multi-1 Financed from NRDI (National Research Development and Innovation Office) Fund, project number 132535, project title: "Folk belief, vernacular religion, mentality, 16 th -21 st centuries: digital databases, encyclopedic summaries". 2 ERC project No. 324214 "Vernacular religion on the boundary of Eastern and Western Christianity: continuity, change and interaction" operated at the Department of Ethnography and Cultural Anthropology at Pécs University from 1 st September 2013, and from 1 st March 2017, onwards continued and was completed on August 31 st 2018 at the Hungarian Academy of Sciences Institute of Ethnology.http://eastwest.btk.mta.hu/(accessed 31 st August 2019).
dimensional online database, enabling searches from a number of possible perspectives. 3t was also under the auspices of the same ERC project that digitized the documents of the witchcraft trials of the early modern period in Hungary. 4 Thus, when the idea to make up for the lack of historical sources in the Hungarian corpus of verbal charms with early modern material first arose, our choice fell quite naturally on the testimonies of witchcraft trials as they seemed ideal for the purpose. 5The charms, threats and curses that were uttered and recorded as parts of these witness accounts offer us a great opportunity firstly to examine the phenomenon in a context separate from folklore collection and source editions; secondly to map out the early modern lay, mostly oral, tradition; and thirdly to add bewitchment texts to the corpus that previously contained mostly only healing texts. 6n the following, I examine the witness statements of trials conducted in early modern Bihar County and the town of Debrecen,7 and I describe as well the kind of mostly generic questions and issues that we had to face during the corpus-building operation in connection with the above named goals. 8In other words, I mainly seek to answer the question of what kind of texts were we able to select for the Digital Database of Hungarian Verbal Charms from the range of verbal actions noted down in the witness accounts of these witchcraft trials.As regards the exploration of early modern written sources of vernacular language use, and, within that, particularly the verbal interactions recorded in the documents of witch persecution, the most relevant approaches came from historical speech act research.9Therefore, in the context of the corpus building project I shall also discuss to which extends the results of historical pragmatics can offer any help (and if so, what kind of help) in solving the generic problems and questions of verbal magic.

GENRE AND DATABASE
There are two points of view from which it is worth examining questions of folklore genres when creating a genre digital database.On the one hand, the boundaries of the planned database are marked out by notions of the specific genre in question (what kinds of texts are we to admit and what to omit from the database).In this sense we need to determine which are the boundaries whose traits need to be understood in order to lead us to accurate conclusions in our subsequent analyses.
Starting from the 1960's, folklore genres came under a great deal of criticism, particularly as regards their rigid, hierarchic and overly etic categories.Recently, however, a number of papers revising the concepts of folklore genres have pointed out that as long as we treat genres as a flexible, dynamic research tool and remain fully aware of the pre-suppositions and limitations that flow from their nature as heuristic devices, in other words if we bear in mind that they are not objective, constant, unchanging or unchangeable categories, then we are able to take advantage of the ways in which they can assist scholarly analysis and interpretation (Frog 2016;Frog et al. 2016).Finnish folklorist Frog claims that genres remain good interpretative tools if we recognise that a genre "is a construction of individual researchers and disciplinary discourses in dialogue with research materials, research questions, and inherited patterns of use of the term".(Frog 2016:51).
Nevertheless, the question we should ask how flexible can we be when building the corpus of a digital database.If we take a look at the international practice of building databases over recent years, we see that most of these projects aim to digitize the 'inherited frames' and, thus the 'inherited corpora' based on them, in other words they publish online a corpus that had been collected and compiled based on some past notion of folklore. 10The digital database of Hungarian verbal charms is one of these databases, leaning primarily on the two-volume collection Ráolvasások [Incantations] published in 2014(Pócs 2014;;Ilyefalvi 2014).Although Hungarian folklore studies have not been entirely consistent in the past 150 years as regards defining the genre of the verbal charm or the related scholarly terminology, the widely known charm collection compiled by Éva Pócs rests on a unified set of guiding principles.Thus, the corpus of Hungarian verbal charms mostly contains texts that were uttered by people individually, in order to achieve some concrete goal or influence a course of events in a negative or positive direction, independently of the formal-syntactic structure of the text or the emic genre being used.As a result, the corpus includes texts, which the informant would qualify as 'prayers' as long as they were spoken with the purpose of healing, acquiring treasure or money or as an act of exorcism.
10 Although a few databases of textual folklore do try to expand generic boundaries (e.g. the Dutch database of folk tales, which is today moving in the direction of offering a full database of Dutch folk narrative and pools together many prose genres from tales through anecdotes all the way to belief texts), the most widespread approach is not to question the previously established frames and boundaries but simply is aimed to digitize an already existing corpus.One such example is the Finnish rune database, which is in fact the digitized version of the 34 volume book series Suomen Kansan Vanhat Runot [The Ancient Poems of the Finnish People] 1998), which has rendered 89,000 texts accessible to digital research.For a recent summary on digital folklore databases see Ilyefalvi 2018.On the other hand, beyond marking out the boundaries of the corpus, the genre can also set the research questions and, as a result, the possible later directions of research.In folklore studies, genre-based readers, typologies and digital genre databases mostly provide a chance to compare the supposedly analogous textual material of eras and cultures far removed from each other in space and time.Thus, digital genre databases promote comparative folklore investigations and, within that, the methodology of the Historical-Geographic school.Thus, although they usually digitize a pre-existing, 'inherited' corpus, in most cases this body of texts is also expanded e.g. by adding on historical material from different time periods.In the case of the Digital Database of Hungarian Verbal Charms, the textual stock is characterised by the versatility of sources: ever since the 15 th century, texts have survived in a wide variety of historical sources (marginalia, codices, recipe books, anti-superstition documents, treasure hunting books of magic, witness statements of witchcraft trials etc.).If we compare the Hungarian corpus of verbal charms with corpuses of folklore texts where each text originates from roughly the same period or type of source (e.g.ethnographic collection), this versatility makes the Hungarian corpus of verbal charms more suitable for historical-comparative examinations than other corpora of folklore texts.
But how can we recognise verbal charms in historical sources if the genres themselves can and indeed do change across time and space?Frog argues that investigations encompassing different cultures and eras require a notion of genre (as a research tool) based on function and use.11Folklorists have agreed ever since the commencement of the study of verbal charms that the majority of these texts cannot be interpreted according to the "classic" criteria of textual folklore, i.e. purely on the basis of form and content, since the main criterion for defining this genre is its function (Pócs 2014:14;Kljaus 2009:71).It is exactly this functional definition that can bring, and indeed has brought, success in the comparative analysis of charm texts far removed from each other in time and in cultural terms.Essentially, it is function that holds verbal charms together as a genre.Thus, the relevant question to ask may be what are the domains of life that people tried to influence through the use of charms, or what kind of texts and practices they used in order to heal or to harm in various eras and cultures.Based on this, if we can define the function of the verbal charm, it becomes easy to recognise utterances that we would like to include in our database.Speech act theory may help us pinpoint the function of the verbal charm, therefore in the next part of this paper I discuss the opportunities it offers (Searle 2009:25).

SPEECH ACT THEORY AND VERBAL CHARMS
One reason why the linguistic and, within that, the pragmatic point of view deserves closer scrutiny is that speech act theory, proposed by linguist philosophers John L. Austin and later John R. Searle became, from the 1960's onwards, an important theoretical Dilemmas of Corpus Construction beyond Folklore Collections… framework or point of departure for studying ritual forms of speech and language, including verbal charms and prayers (Austin 1999;Searle 1979;2000).All of this was partially tied in with a change of attitude in folklore studies that placed the performative nature and context of the texts in the focus.We may observe, however, that the theory was in fact used on a far larger scale (Frog et al. 2016:25-26).Even during the very first wave of studies in speech act theory, folklorist-anthropologist Ruth Finnegan and later anthropologists of religion Stanley J. Tambiah and Wade T. Wheelock applied speech act theory to approach the genres of ritual language (Tambiah 1968;and Wheelock's useful summary (1982) on the topic).Éva Pócs made use of the various speech acts and used the theory as the organising principle when creating the typology of Hungarian verbal charms (even if she was not directly inspired to do so by speech act theory itself).Her categories, defining the main types of charm, such as the "optative and imperative forms", "wishes and tasks addressed to belief figures"; "declarative forms (denying the patient or the disease, sending away)"; "counting", "declarative statements", "supplications or commands containing action"; "statements of fact in the form of enumeration, command" or "wish" are indeed based on distinctions according to the purpose of the speech act, its illocutionary force (Cf.Pócs 1968; Keszeg 2013:288).12According to Vilmos Keszeg, charms, curses and prayers are one way of talking about belief that he terms "ritual speech act" (Keszeg 2013:310).
In speech act theory, all utterances are performative, i.e. every utterance is an act.At the same time, there are unique speech acts which, when successfully performed, produce factual changes in the world.These are declarative acts which in normal cases can only come about on the basis of some extra-linguistic institutional background, as the listener and the speaker are not, in themselves, enough to perform such an act successfully (Searle 2000:150).Researchers of ritual language were inspired by this insight to define and interpret utterances based on faith in the power of the spoken word as declarative acts in terms of speech act theory.The extra-linguistic frame in this case is provided by the idea or belief that it is possible to change the facts of the world through the power of words.From the perspective of the speech act theory, if we explore and establish the structure and regularities of the various speech acts, more accurately of their illocutionary force, i.e. if we describe the felicity conditions required for their fulfilment, we can successfully distinguish and categorise the various speech acts (Searle 1979:16-18).Thus, the speech act model holds the promise that it might enable us to tell whether a particular utterance is a charm or not, i.e. whether or not it should be included in the database. 13But how can we establish the structure and regularities of illocutionary acts in the case of historical data?

HISTORICAL SPEECH ACT RESEARCH AND EARLY MODERN WITCHCRAFT TRIALS
Irma Taavitsainen and Andreas H. Jucker consider the greatest general problem of historical pragmatic research that standard techniques cannot be used and that in most cases researchers are left to fall back on their intuition.Thus, one central dilemma is to outline and define speech acts, since they are specific to culture and time; all communities have their own kinds and répertoires of speech acts.Therefore, Taavitsainen and Jucker believe that researchers should use the same classification as the speakers themselves.In this way, an 'ethnographic perspective' can ensure that we can genuinely see into the linguistic and cultural practices and norms of the period under examination (Jucker -Taavitsainen 2007:107-108).Therefore, in order to gain a clear understanding of the language use and communication of an era, it is important to focus on the socalled speech act labels that may occur in the various sources, since these represent the users' own, internal categories with regard to the function of one or the other of these speech acts (Leitner 2017:155).We know from historical folklore analyses that in the bewitchment narratives of early modern Hungarian witchcraft trials harming utterances usually receive the speech act label of "threat" from speakers.14"I know for sure that Mrs. Harcsás was going to borrow some peas off me, I refused, and so later she came to me and she threatened me harshly, saying, just you wait, you will gladly give me peas, be off with you; so that night my arm was crippled; in the morning I went straight there and took her some peas, and she taught me herself straight away to pick mullein and bandage it with that, so I did just that and right away I had nothing the matter with me anymore" (1715, Berekböszörmény, Schram 1970:I.72).
A distinct group of bewitchment narratives consists of accounts associated with a threatening or cursing witch, similarly to the above text (Pócs 1995:20;Kristóf 1998:107-108).According to the logic of bewitchment accounts, the threat made by the witch (and in many cases after the subsequent counter-threat of the assaulted party) is followed by a process of atonement or healing by the witch, often accompanied by a new speech act such as a healing charm or by a gesture (Cf.Kristóf 1998:9). 15At the same time, historical pragmatic research has also pointed out that although the recognition of illocutionary force may be made explicit by performative acts (most often verbs, such as speech act verbs), these are not in themselves sufficient to define one or the other of the prevailing speech acts, since the meaning of any of these performative verbs can change, transform, or it can vanish from use altogether (Leitner 2017:156).Dilemmas of Corpus Construction beyond Folklore Collections… Documents of witchcraft persecution in the early modern era cover some 300 years and are not homogeneous from the geographic, legal or cultural perspective.Thus, we cannot speak of this collection of documents as a unified stock of texts.At a later stage of research it would be important to carry out investigations that would take into account the differences in the nature of witchcraft persecution over different geographic areas and different periods of time, exploring, for instance, which speech act verbs are used in witness accounts from region to region in order to refer to acts of bewitchment in different ages and areas (Tóth G. 2020).From this perspective, however, it is more important to deal with the fact that within the same time period informants use different words to refer to the same speech act.Based on the cases in Debrecen and Bihar County, speech act verbs used before a threat may include mondotta, felelte, szidta, összeszólalkozván, kiáltotta ('she said', 'she replied', 'she reprimanded', 'they had words', 'she shouted') etc.
"Another time the witness had words with Mrs. János Zuh, her mistress, and Mrs. János Zuh said, you be quiet, for you will soon cry for regret about this" (1717, Nagyszalonta, Schram 1970:I.85): "Once the witness was coming home from the field and in the streets she saw the witness and attacked her and abused her with her tongue, just you wait, you, born in a dog's bed, for you will regret this.That same evening the witness fell ill and was lying in the manger and Mrs. István Kovács went to her in a tarp and lay on top of the witness and pressed her" (1737, Hencida, Komáromy 1910:493).
"Upon her faith the witness claimed that obviously it [the bewitchment] must have come from Mrs. Sallai, indeed she shouted, you wait, you dog, for as long as you live you will remember me, and soon after that the witness's leg was broken all over" (1715, Kismarja, Schram 1970:I.98).
Historical pragmatists claim that the real function of speech acts, beyond the speech act labels can be best reconstructed based on metacommunicative expressions that the speakers use in connection with the various speech acts (Jucker -Taavitsainen 2013:95). 16It is precisely in this sense that witchcraft trials can contribute valuable extra information when compared to collections of folklore texts or to verbal charm data surviving in written, mostly manuscript form.In the witness accounts the statement that was once uttered is reconstructed along with the context of the original situation through recollection, and the statement is often accompanied by some kind of metapragmatic comment.This is why witness statements of court trials of the early modern period, and within that of witchcraft trials are excellent material for studying the spoken language preserved in these documents.It is not surprising that historical pragmatics, this new sub-discipline that has only become institutionalised in the early 2000's, already offers a dozen different studies based specifically on studying the sources of witchcraft persecution in the early modern period. 17 One of these first studies was co-authored by Elena Semino and Jonathan Culpeper and examines some witchcraft trials from early modern England, focussing on witness accounts made between 1593 and 1664.Analysing the meta-pragmatic commentaries occurring in the texts, the authors offered a comparative analysis of the verbs associated with cursing and threatening (to curse, to wish) found in the trial material and in current usage (Culpeper -Semino 2000).Applying Austin's classification they termed threats by witches as "exercitives" while they referred to contemporary threats as "behabitives" (Austin 1999:145-153; Culpeper -Semino 2000:105-106).The difference between the two speech acts, however, was better highlighted when speech act theory was used in the form further developed by Searle, since Searle based the separation of speech acts on their illocutionary point (Searle 1979: 12-29;Culpeper -Semino 2000:106-107).Accordingly, the bewitching threat of the witch belongs to the group of "declaratives", while threat as used today belongs in the group of "expressives" (Culpeper -Semino 2000:108-110). 18Whenever a witch made a threat, she was not merely hoping that something bad would happen to the person in question, but thought that it would happen as a matter of course, as a direct consequence of this utterance.According to Culpeper and Semino, in early modern England the 'institution' which made it possible for threats to operate and be interpreted as declaratives was the belief that witches were capable of doing harm through words, but even more powerfully the context of witchcraft trials and, within that, the witch's activity type in the Levinsonian sense, as constructed in the context in question (Culpeper -Semino 2000:111-114).On the basis of Culpeper and Semino's analysis, in the following section I shall use examples from early modern Debrecen and Bihar County to examine threatening speech acts.

Threats by witches and counter-threats by their victims
According to Searle, in the contemporary sense we consider any utterance a threat the propositional content of which refers to a future event related to the listener.Its preparatory condition is that the event should not be in the in the listener's interest; according to its sincerity condition the speaker should wish the event to take place, and finally, its essential condition is that the wish or desire should happen to the listener (this is the meaning/purpose of the speech act) (Cf.Searle 2009:25;Culpeper -Semino 2000:110).As regards the structure of the witch's threat, Culpeper and Semino point out that the propositional content is the same as in contemporary utterances, but 17 Regarding the Hungarian literature cf.Mónika Varga's PhD thesis that carries out pragmatic analyses of early modern witchcraft trial documents.(Varga 2018).On the digitization of the witness statements of witchcraft trials see Dömötör -Novák 2016. 18Referring to Blanco Salgueiro's study ("Promises, Threats, and  the first preparatory condition -namely, that the event should not be in the listener's interest -is supplemented by another, that the speaker should be in alliance with some kind of supernatural power -usually the devil -and should indeed be able to exert that power.And while in contemporary threats the sincerity condition is indispensable, in these cases it may be entirely absent.Finally, the essential condition, based on which we distinguish illocutionary acts in Searle's taxonomy, is utterly different: the threat is considered an utterance where the event is going to happen to the listener (Culpeper -Semino 2000:110).
Within the 169 lawsuits from Bihar County and Debrecen, reviewed, there were 42 cases in which I found threats that became 'fulfilled' according to the logic of the narrative, i.e. the witness connected the speech act with the perpetration of some kind of harm or bewitchment.These are the instances that may be considered declarations based on Culpeper and Semino's categories.Of the 42 trials examined there were 23 (more than half of the cases) in which the courts asked specifically about the witch's threat that had mentioned bewitchment or harm.
"Does the witness know or has the witness heard whether Mrs. János Nagy from Sámson ever threatened people either in Sámson or elsewhere, and whether those people ever suffered any kind of disease, or whether she harmed or healed or charmed; whether at night she flew about anywhere astride a poker or broom and damaged cows or healed or chased or harmed them, and what kind of sorcery does witness know about her and from whom?" (1714, Sámson, Schram 1970:I.62).
"If anyone fell ill of some kind of disease after she had threatened someone what kind of disease that was?" (1723, Diószeg, Schram 1970:I.104).
"Does the witness know or has the witness heard or seen clearly and evidently whether she had threatened somebody, "would you give me this or that", and then that person came to harm due to the threat?"(1724, Ottomány, Komáromy 1910:326-327).
Apart from the above examples, there were three court rulings in which the (tribunal) court declared that if anyone was proven to have "eaten up" or "driven to harm" the sufferer with their threat, that person was a witch: "persons that bring harm to the innocent by threats (...) after many threats, harming and bewitching many innocent, god-fearing people in their person and their belongings and animals, or in some cases the healing of such" (1717, Debrecen, Komáromy 1910:278).
"And since the inquisition so far has revealed that she did use threats and her threats were instantly fulfilled and these do reveal beyond doubt her allegiance with the devils and these do actually become fulfilled on the persons she had threatened -and therefore to make sure that the many Christian soul against whom she had made her threats may not be harmed, it is appropriate that the heads of these persons be chopped off by executioner, as is ruled by Titulus If we examine witches' threats in the trial material under discussion in terms of their content, we find that common stereotypical threats include "just you wait" and "you'll regret that": 19 "she had threatened the lad saying, ah, you'll regret that" (1712, Bályog, Schram 1970:I.52).
"amid great threats she threatened the witness, wait, for you will remember this as long as you live!" (1715, Hegyközpályi, Schram 1970:I.69).
"she threatened -wait, for I will pay you back for that" (1723, Hajdúbagos, Komáromy 1910:314).The examples from Bihar County also contain several instances when the phrase 'even the milk you sucked from your mother will turn bitter for you' was used for emphasis or to increase the weight of the harm that was to be done."tell your Wife not to carry on with that spinning because she'll regret it, even the Mother's milk she had sucked will turn bitter" (1723, Váradolaszi, Schram 1970:I.109). 19Mónika Varga found that textual characteristics of threats uttered by persons accused of witchcraft include elements such as "just you wait", "you'll see" or "leave it", which, according to her analysis, only occur in utterances related to the witch (Varga 2018:185).There are a few cases in which the reference to God also appears in the witches' threats.In a more strict classification of folklore genres we might have to consider those utterances as curses, since the perpetrator of the damage is not human and such damage is caused not purely by the force of the words spoken, but requires divine intervention: "just as my hen was crushed that day, may God crush her soul in hell just so" (1717, Szalonta, Schram 1970:I.83).
"until God punishes the person I get angry with, I shall not stand by God, but quarrel with God just as good preachers quarrel" (1717, Szalonta, Schram 1970:I.84).
"I see that God has blessed Thamas Vari with child, but God is also punishing him, I myself strove for the same so that even others are wondering at it" (1717, Szalonta, Schram 1970:I.84).
"they came to words (…) and Mrs. Farmosi said to the witness, you may never die until you summon me to you, but no matter whom you send for me I will not come with them; may God make you so sick that even the milk you sucked from your mother should turn bitter for you" (1726, Érsemjén, Schram 1970:I.117).
The threats attributed to the witch may also make mention of the conflict at the bottom of the accusation: In cases like this, the accused is a person placed by Éva Pócs into the category of a witch who had been denied a favour -in other words, she had asked for something and had not been granted it.The declining of the favour becomes the cause for the conflict (Pócs 1995:20).These threats often make a brief reference to the refused request without any lengthy explanation.According to the bewitchment narratives of the witchcraft trials, the witch is not the only one to apply threats.Of the 42 witness accounts examined, there were 28 in which the victim also used some kind of a threat in the hope of persuading the witch to undo the damage that had been caused.One important element in the counter-threat used by the victim is the naming or identification of the perpetrator.In other words, this is where the accusation is phrased and at the same time a threatening imperative is applied to remedy the damage -since it was believed at the time that bewitchment could only be removed by the person who had caused it in the first place.
"she went at her and said (...) I suspect you, it's you who had eaten me up, so you come and cure me, otherwise you'll see you'll come to grief" (1714, Keserű, Schram 1970:I.58).
As the above examples show, these counter-threats by the victims are often stereotypical and cryptic, similar to those of the witches.One recurring phrase is that the accused will 'come to grief' or 'come to' something, unless they help remedy the damage that was done, but the possible consequences are not explained explicitly.
"she was threatened (…) that unless she cures the leg of the witness she would come to grief" (1708( , Érsemjén, Komáromy 1910:219):219).The majority of threats by the victim express some kind of specific negative future event.Some of them mention the prospect of physical harm (or go as far as threatening to take the life of the accused) which the threatening party would carry out personally, unless the accused undoes the damage caused:  1910:366).
"once István Kapros threatened Kata Józsa to make sure he did not catch her because if he did he would either burn her in the house or crush her" (1744, Sarkad, Komáromy 1910:514).
As regards the content of the victims' counter-threats, it is also common for them to threaten their assumed malefactor with a lawsuit (and one of its possible outcomes, being burnt at the stake).
" "she said, Mrs.Mészáros, born of a bitch, you had eaten up my leg, but the master has said clearly, even if he has to spend all his fortune, he will have Mrs.Mészáros caught and taken to the gaol" (1763, Kismarja, Bessenyei 1997:147).
Certain of the counter-threats made by the victim show us explicitly the possible interpretative frame of witches' threats and their modus operandi.Besides all of this the victims' threats also include some instances that make reference to God when wishing some kind of harm to the accused, although these are fewer in number than in the case of the witches' threats.
"and then she shouted at Mrs. Péter Tóth, this evident witch and whore, as many arrows as God has in his heaven, let all of them strike her soul, to hurt every tendon of this most evident witch" (1727, Zsáka, Komáromy 1910:393).
Within the textual corpus examined there were 14 instances in which the witnesses mentioned some kind of positive change, which led to the end of the bewitchment, 'releasing' or 'dissolving the bond' thus signalling that threatening the witch had succeeded. 20f we make the position of the speaker within the trial our point of departure (the accused witch vs. the victim), it is easy to distinguish a bewitchment threat from a threat in the contemporary sense.Regarding the propositional content of these speech acts, we can again usually distinguish the two speech acts from each other.As regards expanding the digital database by adding the witchcraft trial material, it could follow from the above that all utterances attributed to a witch that can be connected to an instance of damage, regardless of the descriptive speech act, could be incorporated into the database.By this new procedure, however, we would be drawing the boundaries exactly at the same line where witchcraft persecution had done in its own time.What is more, if we examine the individual speech acts in their situational micro-context, we notice a far more complex range of possible interpretations.The dividing line between the two types of threat was often thin, i.e. the same utterance could be interpreted as one or the other speech act by the speaker and the hearer, in line with their own self-interest.This is illustrated by the case of Mrs. Illés Aradi, née Judit Helpári, from Bihar County in the early 18 th century.

Threatening in 'devilish' and 'legal' ways
Judit Helpári was summoned to the feudal magistrate's court on February 16 th 1701, charged with witchcraft.The court, members of which included the prefect of the Diószeg estate, asked the following questions during the hearing: "Whether the witness knew for certain and had heard from Mrs. Illés Aradi's own mouth that she threatened people in the street where she lived, saying, you dogs you will not live in this street, and was this threat directed against the people who had chased her son threatening his life or for other such reason as may have been provoked by her witch nature.Secondly, who was the woman in childbirth whom she had threatened, saying to her, »I will teach you at the hour of your birth«?Did anything happen after this threat on the predicted date of childbirth of that woman and if so, what was it that happened?And did anything else in any other way happen to those whom she threatened?In brief, what kind of witchery can the witness quote against her?" (Schram 1970:I.49).
It emerges from Mrs. Illés Aradi's story that this was probably not the first time she was taken to court upon charges of witchcraft, but documents of her earlier case or cases have not survived for posterity.Witchcraft trials held by feudal magistrate's courts were not common in the region -according to Ildikó Kristóf, we know of only five such instances (Kristóf 1998:38).In the first decade of the 18 th century there were altogether 19 people accused of witchcraft in Debrecen and Bihar County.Six of these cases ended with a death sentence, three with a milder verdict, while in nine cases the conclusion is unknown.The only person who is known to have been acquitted was the protagonist of the present example, Mrs. Illés Aradi (Kristóf 1998:45-46).The accused does not appear again later in the history of Hungarian witchcraft persecution, and at Diószeg itself no witchcraft trial was held for another 20 years.Nevertheless this brief trial documentation is excellently suited for a historical pragmatic analysis, primarily because Judit Helpári's own statement and the court ruling have both survived.In this instance, we note that the fact of the threat was not sufficient to get the accused sentenced, and by employing a suitable linguistic strategy she made a successful appearance in front of the courts.Her acquittal was probably to a large part due to the fact that the witnesses could not enlist any other activity or trait characteristic of witchcraft, the only such case in the county.Thus we may fairly assume that possibly the accused may have displayed other traits of the role of the witch, but we lack relevant sources.The charge consisted of two distinct points and was represented and announced by István Szűcs: "1. this person, scorning God and the laws of the world, declared and professed with her own tongue in the street where she lived, you will not live in this street because of me. 2. which servants included a pregnant woman and she had also said unto her, when the time of childbirth comes, you dog, just tell me so you'll even regret the milk you sucked from your mother, for I fear no judge or council, for even before there was nothing they could do after the confessions of 70 witnesses, of which indeed six should have died, so I fear nothing again this time" (1701, Diószeg, Schram 1970:I. 48).
According to István Szűcs, if Judit Helpári was to deny all of this, that would "be ample proof of it".He was certainly recommending burning her at the stake, in line with the laws of the land, but first to apply torture, in order to extort from her the names of her accomplices.At her first trial, to begin with, Judit Helpári showed little inclination to co-operate and denied the charge.Next, defence counsel for the accused Miklós Faragó came to speak and argued that the accused was not denying that she used threats, but that from this it did not follow directly that she was a witch.
"Although she does not deny that she used threats as mentioned, since her son was chased by certain armed men out to kill him, so driven by maternal love, she said, in an attempt to rescue her son, that if they catch her son, she would make sure they cannot go on living in that street.And from such words it is hard to conclude that, speaking for her son, she would be a witch" (Schram 1970:I. 48).Dilemmas of Corpus Construction beyond Folklore Collections… In spite of the above, István Szűcs called for further investigation of the matter and requested that through the witnesses they should find out and prove what the crime of the accused was and what kind of punishment she deserved.
"In line with my above statement and according to the accepted laws of the town, if the witnesses would provide us with more, I merely wish that in line with my previous speech the laws of the land should establish what sort of punishment she deserves, whether she should be burnt or have a different punishment.After which I expect the law to be applied in a godly manner" (Schram 1970:I. 48).
After the above, Judit Helpári herself said in her defence: "I do understand the wicked hearsay against me, in which his Honour says that I am supposed to have threatened somebody in a devilish manner, which I deny.But, as I said earlier, my dear child having come to such bitter treatment I said, in a legal manner of speaking, that the persons who had so hounded my child and sought him out to bring his death upon him were not to live in the street.And I call upon your Honour, the judge, as straight as you declared my witchery against me, to consider it, in other words we wish to be absolved from the charges levelled by the Officer" (Schram 1970:I. 48-49).
This utterance by the accused points at the essential difference between the two speech acts.It admits the act of threatening, but by way of excuse refers precisely to the fact that this was not "that kind of" threat -not a declarative act.She denies having threatened "in a devilish manner".Her own interpretation of the story, and it is of this that she successfully convinces the other participants, is that through motherly love she had "turned bitter and meaning the thing in a legal manner I said", as a kind of expressive or commissive, and that for this she may, at worst, be reprimanded, but it is no ground for proving that she was a witch.
This was followed by the questions to the witnesses mentioned above.The court was clearly intent on finding out whether the threat had led to any kind of actual harm or bewitchment.Altogether twelve witnesses were heard during the lawsuit, the first two of whom -husband and wife -recalled the conflict almost word by word, including the act of the threat and its propositional content, as we have already read on behalf of the prosecution and the accused, and in the question of the court.
"Did you hear from the mouth of Mrs. Illyés Aradi, I am not afraid of God or the Bishop or the Bishop's Judge or his Council, for if that faithless man Mihalj Balogh could not kill me I am not a bit scared of anyone, but both you and your son will cry even for what you had sucked from your mother and digest it bitterly, and that we will not live in that street because of her.But he does not know of her harming anyone afterwards.""I have heard from Mrs. Illyés Aradi's own mouth that she said, I am not afraid of God, nor the Bishop, nor the Bishop's Judge, nor the Council, for if that faithless man, Mihalj Balog could not kill me I am not afraid of anybody one bit, but both you and your son will cry even for what you sucked out of your mother and digest it bitterly, and that we would not live in that street because of her.But they do not know of her doing harm to anyone after that" (Witness No. 1: István Hati Schram 1970:I.49).
"I have heard from the mouth of Mrs. Illyés Aradi that she said to the people in our street, you dogs, you will not live in this street, because of me, for I am not afraid either of you, nor the Bishop's Judge, nor his council, for I had gone to the Scribe Paul on my knees and I shall be saved if seventy witnesses could not press me down" (Witness No. 2: István Hati's wife, née Zsófia Menyhért, Schram 1970:I.49).
It may have been significant for the outcome of the trial that although the above threat was quoted both by István Hanti and his wife, at the end of his statement István Hanti declared, in effect, that he was not certain as to the impact of the speech act, i.e. whether any kind of harm was done or not after the threat.It is only the twelfth witness, the widow of János Szunyog who managed to produce a bewitchment narrative, which included the fact of bewitchment in the context of the threat.
"Add on to your former statement that after she threatened the son, his leg became covered in ulcers from which he could not recover until he married her daughter, just as Mrs. Illyés Aradi had said, that he would not recover until he married her daughter.But then one night she brought a black liquid to her son and said for him to drink it and having done that I don't know why, for that reason or not, but he recovered" (Witness No. 12: János Szunyog's widow Schram 1970:I.49).
The above described bewitchment story is a new narrative which had not come to the court's attention before, and so it neither confirmed nor disproved the earlier charges.At worst, it added to Judit Helpári's qualities as "cursing", "threatening" or "devilish".Being clearly different from the others, it was not enough.In its judgement the court acquitted the accused and pointed out that the witness statements 'do not sufficiently reveal' Mrs. Aradi's being a witch, but if they heard further threats and 'filthy abuse' from her, they would instantly expel her from the village (Verdict, Schram 1970:I.50).
Furthermore being lucky, Mrs. Illyés Aradi, née Judit Helpár was probably helped by her lawyer and her own linguistic strategy in escaping being burnt at the stake.She had understood very well the essence of the witchcraft trial discourse of Bihar County in the early modern period to save herself.
I hope to have demonstrated with the case of Mrs. Illyés Aradi that without a more thorough understanding of the context and a sufficiently close reading of the contemporary documentation it is impossible to understand the communication characteristic of witchcraft cases.The lessons learnt from the case from Bihardiószeg which I examined supported the finding that at the time and in the context there were at least two ways of threatening a person, and that these were considered to be two so widely different speech acts (expressive/ commissive and declarative) that they could make the difference between life and death. 2121 In her analysis of Scottish witchcraft cases from the early modern period, Magdalena Leitner pointed out that the crucial speech act in the witness statements she had examined was, similarly, that of the threat, and that these threatening speech acts displayed differing degrees of performativity from a pure venting of emotional tension all the way to the curse uttered with the intention of doing harm.Therefore, instead of a rigid demarcation of different threatening speech acts it is more useful to position these various neighbouring speech acts with their different performativity in a common pragmatic space, depending on their situation and interpretation.She points out that the prototypical witches' threat can in many cases not be classified as a declaration but appears more as a simple announcement of an imminent act of supernatural damage (Leitner 2017:153).On the concept of the pragmatic space see: Jucker -Taavitsainen 2000.
As it emerges from the wide variety of individual interpretations, not all witch threats are declarative among the Hungarian early modern cases.Often it is not clear, for instance, whether the bewitchment is supposed to have taken place directly when the witch's threat was uttered or whether it was merely a threat for the future that was, however, certain to take place as it was uttered by a witch.It was also debated by the witnesses and the court and individual interpretations differed as to whether some kind of bewitching practice was also required to accompany the utterance or the threat itself was sufficient for the bewitchment to take place.In the court case of Mrs. Illés Aradi, an utterance by one of the witnesses ("But they do not know of her doing harm to anyone after that") allows us to conclude that the harm was thought to have taken place after the threat had been made (Witness No. 1: István Hati, Schram 1970:I. 49).
What we cannot conclude from the above is that there was no difference between the two (or three) speech acts.The case of Mrs. Illés Aradi proves precisely that the speaker of early modern times made a distinction between one type of threat and the other, to the extent that convincing the judge that the speaker's threat could be interpreted as one and not the other type of threat could save the life of the accused. 22hat does all this mean for the expansion of the database of Hungarian verbal charms?To put the question pragmatically: are we to include in the database as charms with a bewitching intent the threats made by Mrs. Illés Aradi?("both you and your son will cry even for what you had sucked from mother and digest it bitterly, and that we will not live in that street because of her", "he would not recover until he married her daughter").Are we to decide the successful execution of a speech act based on its illocutionary intention or its perlocutionary effect?Even within the same situation, different witnesses can offer different interpretations of what they had seen or heard.While in one statement we read that the verbal interaction that took place is in direct causal connection with the damage suffered, this connection may easily be refuted by the next witness. 23

CONCLUSION
Based on the analysis presented in this paper we can first of all draw the conclusion that speech act theory is not in itself any better suited to decide whether an utterance is a verbal charm or not than was the classic genre theory of folklore studies.At the same time, it has also become clear that while emic interpretations are important and cannot be left out of consideration, they can rarely be used as analytic categories, since they are overly individual, changeable and situation-bound.Approaching these utterances from the angle of speech act theory was, however, suitable to show that in early modern Debrecen and Bihar County a threat could be interpreted as any of three possible speech acts (declarative, expressive/commissive).
What folklore studies can gain from genre databases is primarily the comparison of periods and cultures far removed from each other in time and space.This is precisely why in the first phase of efforts to enhance the corpus of the Digital Database of Hungarian Verbal Charms we did not include all of the bewitchment threats of the early modern witchcraft trials.For the time being, we have annotated and marked data related to bewitchment or verbal magic in the Digital Database of Hungarian Witch Trials.In the first wave, the data we included in the Hungarian verbal charms database from the overall body of witchcraft trial material were those that met the criteria of a narrow folkloristic definition of a verbal charm in terms of both form and content.

"
she threatened him, saying, you wait, you mongrel, you'll even regret what you sucked from your mother" (1715, Hencida, Schram 1970:I.77)."she threatened, you wait, you this and that, you bastard, even the milk you sucked from your mother will turn bitter for you" (1715, Kismarja, Schram 1970:I.90)."but you'll cry for regret, 'cause even what you sucked from your mother will turn bitter" (1717, Szalonta, Schram 1970:I.86).

"
Dosa took an axe in her hand and went to Mrs. Kulcsár's house and said to her, now, you so and so's child, unless you cure my baby while I am at the mill I shall destroy you with this axe in your own house"(1715, Hegyközpályi, Schram 1970:I.68).
"she went for her: you faithless witch, you go cure my son otherwise I will drown you"(1715,  Hencida, Schram 1970:I.78)."going to her house and not finding her at home she said to Ilona Balogh's daughter, tell your mother to cure my little daughter before I get home, because if she does not I shall burn her out of her house" (1715, Kismarja, Schram 1970:I.91)."said the witness, I shall hit the person who does my pressing at night with my axe so hard she won't need any harder" (1722, Szalárd, Schram 1970:I.98)."I will beat up Ferencz Paczai, and fuck that motherfucker and beat him up again" (1722, Szentjobb, Schram 1970:I.101)."yousoandso, you cure me for if you don't, I will find you no matter where you hide, even if I have to go to the end of the world" (1722, Szentjobb, Schram 1970:I.102)."youateme up, auntie, I shall bar your door and burn you in the house" (1723, Váradolaszi, Schram 1970:I.109)."Mywife grew sick very suddenly, all my suspicion goes to you, you did this and I shall cut you up into small pieces you witch, with such and such a soul" (1726, Hegyközpályi, Komáromy she says to Mrs. Göbei, Panda, you go cure my son, for I will have you burnt even if I have to go become other people's servant" (1723, Hajdúbagos, Komáromy 1910:304)."Göbei said, if you want to bind [by bewitchment], for my comadre said they would have you burnt, for if I had cured Samu Pap's servant I would have been burnt by now" (1723, Hajdúbagos, Komáromy 1910:305).
"she said to Anna Berek: by the devil, you all should go and cure my wife, for if you don't, I will get you burnt"(1723, Hajdúbagos, Komáromy 1910:305)."he said to his wife, you deserve to have been burnt a long time ago" (1724, Nagykereki, Schram 1970:I.115)."hesaid,youobviouswitch,youate up my leg, you come and cure it, for if I catch you I will have you destroyed and burnt(1756, Telegd, Komáromy 1910:633)."hesentwordthatsheshouldcure him, otherwise he will have her caught with the permission of the vice magistrate"(1756, Telegd, Komáromy 1910:635)."shechastisedhim:you such and such, you damaged my servant, now go and cure her, otherwise I will have you burnt"(1756, Telegd, Komáromy 1910:635).
"she said, don't you start threatening, Mrs. Csanadi, for if something happens to the child, you will see what happens to you"(1714, Keserű, Schram 1970:I.58).
"Mihály Kovács replied, old woman, don't you start threatening, because if any harm comes to us, we will be sure to find you"(1723, Váradolaszi, Schram 1970:I.108).