A narratological approach to witchcraft trial records: creating experience

ABSTRACT The article uses narratology as a tool to examine the seventeenth-century witchcraft trials in Finland and the area of Finnmark, Northern Norway, to analyse how experience as a category of knowledge and expression surfaces in original court records. This article focuses on courtroom discourse in witchcraft trials: the interrogation, what the accused confessed to in terms of ideas about witchcraft, and how the personal and collective experience of witchcraft may be traced in court record narratives. A narratological approach to courtroom discourse provides an understanding of the legal practice, contextual conditions, and complexity of the entire trial. By focusing on the confession as a narrative, the analysis unearths the ideas about witchcraft expressed by the individual accused person as well as the echo of the questions posed during interrogation, but also a personalized approach traced by features of orality. Thirdly, the narrative voices that we found were used to investigate how they all contribute to the narrative by exchanging information, knowledge, and interpretation so that, eventually, they create a shared understanding of the experience of witchcraft – or lack thereof.

Court records; witchcraft confessions; witness testimonies; Finland; Finnmark; seventeenth century Introduction: narrative, experience, and witchcraft trial records Basing itself on court records of witchcraft trials, this article will focus on courtroom discourse: the interrogation, what the accused confessed to in terms of ideas about witchcraft, and how the personal and collective experience of witchcraft may be traced in narratives in the records. The aim of this article is to shed new light on seventeenthcentury witchcraft trials in Finland and the area of Finnmark, Northern Norway, 1 by applying a narratological approach to analyse how experience as a category of knowledge and expression surfaces in original court records. This is first and foremost a methodological experiment, but it will also shed new light on the trials themselves.
This article works on two important concepts: narrative and experience. Narratology is defined as the study of structures in narrative texts; an exploration of the narrator's function. 2 A narratological analysis will emphasize linguistic conditions: not only what is told, but also how it is told. Our methodology is based on Gérard Genette's influential study Discours du récit [Narrative Discourse] (Paris, 1972). 3 Genette's two subsequent works, Nouveaux discours du récit (Paris, 1983) and Fiction et diction (Paris, 1991), expand his original narratological framework and discuss the boundaries between fictional and factual narratives. 4 Genette's methodology has three levels: history, narrative, and narration. The fundamental relationships between these levels are time, mode, and voice. Based on this, five categories are established that are used to uncover various narrative strategies: voice, focalization, order, speed, and frequency. Genette is particularly known for his handling of voices. 5 The analyses of this article will mainly focus on voices, as explained by Genette, particularly as voices come to the fore in the accused persons' confessions. In witchcraft court records, the confession of the accused person is a narrative embedded within the larger narrative of the entire trial. The confession is a narrative of its own, written down as a coherent story about the learning and performing of witchcraft.
Genette uses the terms "diction" and "factual narratives" for non-fictional texts, stating that because of research requirements, "[narratology] is unlikely to exempt us from having to undertake a specific study of factual narrative (. . .) Such a study would require a largescale inquiry into discursive practices such as those of history, biography, personal diaries, newspaper accounts, police reports, judicial narratives" [our Italics]. 6 Genette underlines the necessity to interpret factual narratives in their historical contexts. This is related to the fundamental difference between factual narratives and fictional narratives. 7 As Lubomír Doležel argues, "historical worlds are subject to restrictions that are not imposed on fictional worlds." 8 In factual narratives, there is a textual layer of reference to historical events that does not occur in fiction.
A narratological analysis pays careful attention to a record's language and the way stories are told, thus providing access to shades of meaning that would otherwise be overlooked. However, as David Herman states: "What makes a story a story cannot be ascribed to narrative form alone, but rather arises from the interplay between the semantic content of the narrative; the formal features of the discourse through which such narrated content manifests itself; and the kinds of inferences promoted via this interplay of form and content in particular discourse contexts." 9 The analysis provides insight into the development of a trial, from when the accused person is brought before the court until a sentence is passed. The authoritative centre of narration renders the voices of various trial participants with their specific qualities. Singling out and getting close to the various voices in this way makes it possible to broaden the understanding of the discourse at stake and of the verbal interaction in the courtroom.
Experience, on the other hand, is understood here as a category of using memories of past events and circumstances to make sense of present ones and form expectations for future ones. 10 It should not be understood as a singular occurrence or event, even when it is discussed on the everyday level, as famously criticized by historians of feminism and colonialism in the 1980s. 11 Rather, experience is a process of observing the world and its events, connecting these observations to previous ones, and then interpreting them in a way that connects one's own collection of previous observations to the new ones but also to the expectations of the surrounding community or culture. The observations are given meaning in a simultaneous, four-way process between the past and the future, and the self and the community, in which experience is continuously being formed in the middle. In this, a historical approach to experience may be different from narratology's emphasis on voice and subjectivity of experience -as we will show. The processes of sharing and personalizing experience eventually consolidate into structures of experience. Understood in this way, experience will always include elements of the worldviews of all participants, even those transcultural influences that influence worldviews only remotely, such as Christianity, Lutheranism or state power. Therefore, it is not fruitful to try to separate clerical or dogmatic from lay influences here, nor elite from popular influences. Rather, experience is about the entanglement of these influences. 12 Looking at experience as a form of knowledge also emphasizes that experience is a continuous interpretative process in which people give meaning to life's events. This is a process of connecting various singular events and circumstances. When this process of making connections -the process of experience -is repeated often enough by a significant number of people and communities, it forms social structures that people come to expect, count on, and despair about. Experience as knowledge is, thus, a cognitive process that takes shape in the form of a narrative. This is why a narrative, and the voices of a narrative, can be a fruitful way of studying experience and vice versa, but also a concept that may challenge both classical and cognitive narratology.
We use court records of witchcraft trials from Finnmark in Norway and from Ulvila, Finland, at the time a part of Sweden, in order to show how narrative and experience analysis works in two different situations. The article will first introduce the sample records as narratives, pointing to their structure but especially to the narrative voices in them. Thereafter, it will focus on the accused persons' confessions, in which the person herself is the narrator. This part of the study focuses on the Norwegian trials, because the Finnish trials, as well as most other court records from Finland, lack a similar confessional part in the narrative structure. Some comparisons can nevertheless be made. Next, we use the narrative voices that we found to see how they all contribute to the narrative by exchanging information, knowledge, and interpretation so that, eventually, they create a shared understanding of the experience of witchcraft -or lack thereof.

The role of the scribe
It is evident, that the treatment of the two cases was very different. Next, we will return to an investigation of the multiplicity of voices, and with them, the construction of narrative, experience and knowledge in the trial records. On the one hand, they have pointed out influence on the part of the judiciary. Norman Cohn, Stuart Clark, and Lyndal Roper, for example, have emphasized the forced discourses based on formulaic questionnaires used by the Inquisition and other Continental courts. 13 Even more explicitly, Rita Voltmer maintains: 'In search of individual voices in the courtroom, be it in confessions or depositions, therefore, we are in constant danger of labelling as local folklore what is really the lore of judges and scribes'. 14 Jürgen Macha and Elvira Topalovic claim that the scribe constructed the court records. 15 Peter Rushton agrees with this. 16 On the other hand, legal records are interpreted as providing glimpses of the imagination of the accused person, and also of a broader sphere of social mentalities. In Elizabeth Cohen's studies, which are based on sources from Italy, the voices of the accused are interpreted as personalized and differentiated according to the scribe's professionalism. She introduces the term 'double modes of reading' in order to emphasize that court records can be analysed on various linguistic levels. 17 Natalie Zemon Davis, working on French pardon tales, claims that her records provide insight into people's mental realm, although their factual content may be more obscure. 18 Malcolm Gaskill argues, based on studies of testimonies in England and New England, that they provide insight into the mental world of the witnesses. 19 Alison Rowlands has studied German witchcraft narratives and maintains that personal testimonies reveal hidden and imaginative worlds. 20 In the recent 'emotional turn' historians have explored these imaginative worlds as emotional, affective or even 'inner'. 21 Carlo Ginzburg goes even further and operates largely on the assumption that suspects consciously told and phrased their own stories, describing how they saw matters. 22 Of existing scholarship, works dealing with analysis of the narratological category of Voice in witchcraft court records, discourse analysis of courtroom discourse, studies of the impact of the personal factor in witchcraft trials, and analysis of comparative witchcraft studies are part of the context of this article. However, we want to emphasize that our focus in this article is a narratological analysis of witchcraft court records, and it is on this point that we have something new to offer. Starting with comparison, studies by Johannes Dillinger and Gunnar W. Knutsen demonstrate regional comparative studies in Germany and Spain, while the Louise Kallestrup's 'Agents of witchcraft' is a transnational study including Denmark and Italy investigating similarities and differences. 23 The personal factor influencing witchcraft trials are discussed by Gustav Henningsen for the Jutland area, however he treats the question as one of source criticism instead of narratological methodology. 24 Discourse analysis of legal documents are dealt with by Natalie Z. Davis, Marion Gibson, Diane Purkiss and Miles Ogborn, based on source material from France, England, Scotland and the Anglo-Caribbean world. All these studies, along with studies of Brian P. Levack and William Monter, have pointed out that the confession narratives were 'negotiated' and 'co-authored'. 25 However, historians in general have not often been interested in Voice as an analytical category, studies by Gunvor Simonsen and Liv Helene Willumsen, based on the West Indies, Scotland and the district of Finnmark in Norway being the notable exceptions. 26 These studies, by close reading of court records, discuss the question not only of what is told, but also how it is told, and try by closer interpretation to distinguish various accents of the voices rendered in the records. The novelty offered in our article is based on what language can convey, and thus founded on methodological grounds: we demonstrate not only how narratology's strong emphasis on linguistics manage to bring forth individualized expression of the speaking person's encounters with the world, but also how this personalized expression, due to the shared character of language enables the emerging of a voice of experience. Because the circumstance in which the court records used in various studies were produced were different, both legally and culturally, comparing our findings to them is complex, but not impossible. 27 The co-authoring process, which used the experience of all those involved, is especially visible in the Finnish trials, but not prominent in the Finnmark trials.
The Nordic countries did not use questionnaires during interrogation in witchcraft trials. Therefore, leading questions on the part of the interrogator came to the fore more often in the form of open questions steered towards demonological content. These questions opened for lengthy answers, not only the one-syllable answers 'Yes' or 'No', and gave an opportunity for the accused person to deliver narratives.
Our view is that confessions in witchcraft trials give access to the accused persons' knowledge when it comes to ideas about witchcraft, mastery of storytelling, and effort to use a personalized language in the courtroom. 28 The scribe is a professional person who strove to record the development of the trial and the words uttered in the courtroom. Not least, this is displayed in the rendering of orality features in the depiction of the accused persons' confessions, even vernacular words, which is a clear individualizing element. 29 Thus, the voices of accused persons come to the fore in the confessions by the rendering of words pronounced in the courtroom and taken down by the scribe. Malcom Gaskill emphasizes the recording of words that were factually uttered, and states that it was not possible for the scribe to distort something which was not said in the first place. 30 The analyses performed below are based on this interpretation of the role of the scribe when it comes to rendered discourse in court records. It should be underlined that the criminological contexts impact the records, and that witchcraft trials are seen as governmental operations on a communal sphere. 31

Witchcraft trials: where the court records were created
In both Finland and Norway, witchcraft was dealt with in secular courts. At the height of the witchcraft trials -from around 1660 to 1700 -the formally accusatorial Swedish system was gradually transforming into a more inquisitorial one, as officials and vicars adopted the role of public prosecutors. Characteristics of the older system started to lose their importance, and the assessment of formal evidence instead of reputation by an educated judge from outside the community was given increasing weight. 32 Rural courts at the end of the seventeenth century in Finland worked with a mixture of medieval lawscompiled by Christopher the Bavarian in the mid-fifteenth century -and a range of more recent ordinances, verdicts, and statutes subsequently issued by the crown. These stipulated punishments for witchcraft: usually a fine of 40 or 60 marks, but if the alleged witchcraft had killed a person, a guilty person could be sentenced to death. In 1608, the king had also appended the Mosaic law of the Bible to the printed editions of the law, widening the possibility of a capital sentence to all persons convicted of witchcraft. Nevertheless, Mosaic law never gained an equal footing with the law of Christopher the Bavarian in Finnish rural courts. 33 As a result, rural district courts did not always follow the law as such, but they followed the lead of the courts of appeal. 34 In the case of sentencing witchcraft, this led to comparatively lenient practices, which the court of appeal approved: according to Marko Nenonen, around half the trials in Finland ended in acquittals, and the rest mostly in fines of varying magnitude. Only 10% ended in capital sentences, of which very few were eventually executed. 35 All proceedings of the lower courts were eventually reviewed by the court of appeal in Turku, which was the only court of appeal in Finland during the seventeenth century. The court records that historians use today were, in fact, mostly created for the court of appeal of Turku to read and review.
In Norway, there was an important difference in the court proceedings from Finland in seventeenth-century witchcraft trials, as lower instance courts were allowed to pass death sentences. Historically, from 1397 to 1523, Sweden, Denmark, and Norway were joined in a personal union under a single monarch, the Kalmar Union. 36 Otherwise, their political and legal administrations were kept apart. In 1536/1537, Lutheranism was established in Denmark and Norway as part of the Protestant Reformation. 37 In 1536, Christian III was placed on Denmark's throne. The union between Denmark and Norway and Sweden lasted until 1523, and the union between Denmark and Norway lasted until 1814. The Nordic countries had codes of law written in the vernacular dating from the Middle Ages. 38 The thirteenth-century national law of Norway, Magnus Lagabøtes Landslov, 39 stated that maleficium was a capital offence. According to Hans Eyvind Naess, this statute seemed not to have been used much until after the Reformation. 40 Throughout the period of the Kalmar Union, Norwegian laws continued to be enforced. There were three levels in the court system: the local courts, the court of appeal, and central legal authorities in Copenhagen. Local courts were allowed to pass a death sentence, and there was no requirement to have the sentence reviewed by the court of appeal. 41 Local courts in Norway appointed magistrates from 1591 onwards. The majority of district governors and bailiffs in Norway were Danish. 42 Danish law stated that denunciations by persons accused of witchcraft should not be accepted as proof in any trial. 43 In addition, new statutes from 1547, 1558, 1576, and 1588 stated that the use of torture was not allowed until a sentence was passed. 44 However, sterner practices on the European Continent regarding procedures in witchcraft cases seem to have influenced mid-sixteenth century Denmark and Norway. In 1584, the king in Copenhagen responded to a letter sent by the bishop in Stavanger, describing the consultation of wise men and women as a deadly sin against God. Healing, signeri, was made a capital offence in the districts of Stavanger and Bergen. 45 Through decrees in 1593 and 1594, this local criminal law was extended all over Norway. The two decrees were standing law until the Witchcraft Decree of 1617. 46 In 1604, a new code of laws was published: Christian IV's Norwegian Code of Laws. 47 This code stated that those practicing witchcraft should not be forgiven. The important Witchcraft Decree of 1617 came next. In this decree, 'true' witches 48 were defined as those who had attached themselves to the devil or consorted with him. A clear demonological definition had thus found its way into the letters of the law. This was repeated in the Recess of 1643 and the Danish Law of 1683. 49 The 1617 decree changed the punishment for seeking acts of healing from a death sentence to the loss of property, and the punishment for healing from a death sentence to exile. 50

Vardø 1662-63, case from Norway
We would first like to discuss the records of two witchcraft trials, one from Finland and one from Finnmark, to see how the elements of narrative and experience impact the interpretation.
For Finnmark, we selected the case of Margrette Jonsdatter, who was accused of witchcraft in a panic that lasted from October 1662 until April 1663, the last and most severe witchcraft panic in Finnmark. Twenty women were sentenced to death during this panic. Margrette lived in Vardø and had previously been a maidservant in Tromsø, so she had travelled to Vardø in search of work. She was denounced for witchcraft by two other women. Margrette was brought before the court the first time at Vardøhus Castle on 8 October 1662, at the beginning of the witchcraft panic. At that time, she was vouched for by people present in court. She was brought before the court again at Vardøhus on 25 October 1662 and 29 January 1663. She refused to confess to witchcraft. Then, on 9 March 1663, she was subjected to the water ordeal and floated. She was brought before the court again on 10 March 1663, together with three other women: Sigri Jonsdatter, Gundelle Olsdatter, and Dorette Poulsdatter. After the water ordeal, she gave a demonological witchcraft confession. She told interrogators about where and from whom she had learned witchcraft, about her entering the devil's pact, and about her participation in witches'gatherings and collective witchcraft operations. Severe torture was used during this panic, and during the night between 10 and 11 March 1663, Dorette Poulsdatter, who refused to confess, was tortured to death on the rack at Vardøhus. Margrette Jonsdatter was brought before the court again on 11 March, found guilty of practicing witchcraft, and sentenced to be executed in fire at the stake. Thus, on the day Margrette got her death sentence, a woman had been killed at Vardøhus the night before.
The verdict and sentence follow the pattern of those of many other trials during this panic: 'Thus, after indictment and responses, and in accordance with the circumstances of the case', 51 it was found to be true that Margrette Jonsdatter, Styrk Olsen's wife, was one of the said witches, 'as she herself had confessed in detail on a most solemn oath before the court, also stating where she learned and subsequently practised her evil craft. Nor did she deviate from this, but said herself that she was a witch. 52 Three other women were convicted the same day as Margrette. The death sentence of these women is recorded in this way: In view of her committed misdeeds, she shall suffer and be punished corporally in this world and since she has also denounced, as she has, in addition to herself, the above Ragnilde Endresdatter, 53 item Siigrj Jonsdatter and Gundele Oelsdatter, also from this same place, maintaining that their craft is just as potent as hers, and they themselves have confessed truthfully in detail before the court that they have learnt witchcraft and consorted with her in various places, according to the lengthy confessions recorded here, we find, in view of the circumstances, that we cannot judge or decide otherwise than that they have forfeited their lives to fire at the stake. 54 The entire court records of Margrette Jonsdatter's trials can be seen as a narrative created by the scribe. A clear timeline is established from the first time she was brought before the court until her conviction. Factual information is taken care of, such as the dates she was brought before the court and the water ordeal was carried out. Thus, it becomes clear that she had been rumoured to practise witchcraft for a long time and imprisoned for more than a month at Vardøhus before the final stage of her trial took place. Also, light is shed on court practice and the use of circumstantial evidence for the whole of Norway in the form of the water ordeal. 55 Witnesses are not brought before the court. As witchcraft as a crime was impossible to prove, it was important to get a confession from her own mouth. Margrette's own voice is heard first in her denial of witchcraft and later in her confession. The water ordeal seems to have had an impact on her decision to confess. The voice of the law is heard in the verdict and sentence.
The scribe structured and compiled the text. The courtroom discourse that is rendered displays several intertwined voices. As a narrative, the text takes care of linearity -the order in which events are told. Voices of participants in the trial are rendered in a reliable way. The voice of the interrogator surfaces either in direct questions or in shadow questions, which are deduced by scrutinizing the answer given. The interrogator clearly steered the interrogation in a demonological direction, starting with when and from whom Margrette learned witchcraft, and later going into details about entering the devil's pact and the performance of collective witchcraft. The scribe used a neutral style, in which his own voice does not emerge. When it comes to the mood of these court records, the scribe used no distancing devices when Margrette's confession was recorded, not even when unrealistic elements were confessed to. This signals belief in witchcraft on the part of the scribe, and probably on the part of the court's officials. The court records, seen as a narrative, show the complexity of the trial and were written in a style that heeded legal conventions, keeping the records together as a coherent text by using narrative structures.
When viewed as a narrative, the court records of Margrette Jonsdatter's trial display a linguistic format suitable for documenting court practice and courtroom discourse. By aiming at giving an account of the development of the trial as a judicial event and of the words uttered in the courtroom, the records as an entire text appear as consistent and comprehensible. By paying attention to both the phases of the trial and the voices heard in the courtroom, the scribe managed to create a document that appropriately states the reasons for the sentence passed.

Ulvila 1677, case from Finland
The Finnish sample records come from the district court summer session in Ulvila on the west coast of Finland, which took place in July 1677. The local bailiff (Länsman) informed the court that lodger Thomas Eriksson had publicly accused his household mistress, Brita Michelsdotter, of riding a goat to Blåkulla, the place where witches' sabbath was held, on Easter mornings in three previous years. Brita herself was not present, but her husband testified for her, saying that she was completely innocent. Likewise, a burgher and a schoolteacher from the nearby town of Pori testified that despite the rumours they knew of nothing witchcrafty or superstitious to be practiced by either Brita herself, her parents, or any of her siblings. A locally living gentry, Captain Reinholdt Böning, repeated having heard Thomas's story that he had seen Brita riding a goat, but that he had not seen anything himself. Brita's husband confessed that three years ago, another rumour had been spread by two other local women, called Margreta and Brita, which stated that they had seen his wife riding a goat. The matter was postponed to the next session, when Brita, the other women, and other possible witnesses were to appear. 56 This trial took place well over ten years after the trial in Finnmark, and it was not the first of its kind in Finland or Sweden either. The first major occurrences in Finland had taken place in Åland in 1666-70. 57 During the later 1660s, the phenomenon also spread to what we now call Sweden proper: Dalarna and Stockholm. This was, however, the first case in the parish of Ulvila. 58 The procedure was nevertheless the same as it was in cases of traditional maleficium: the case was postponed until all parties could be summonsed. In the meantime, everyone went home, which also stands in strong contrast to the imprisonment and torture that took place in Finnmark.
In September, the court reconvened for a new session, and the investigation continued. Brita was present, and she was exhorted to give God his honour and confess the truth. She said she was innocent, and was 'asking God now as before to gently save her from such deeds'. 59 The other Brita, called as a witness in the previous session, took an oath and said she knew nothing about these accusations except that Thomas Ericksson, now a soldier instead of a lodger, had talked about having seen someone riding Brita Michelsdotter's goat on an Easter morning. 'After a while', the court record states, 'she added that when Thomas Ericksson was summonsed to court, he was very sad and miserable, but as he had travelled away [as a soldier] he had been glad to be spared appearing in court'. 60 Brita Michelsdotter's husband Arvid Sachariasson called for a Henrich Thomasson, who had told him of the rumours against his wife. Having taken his oath, Henrich said that he had heard from his son-in-law, Erick Monsson, when Brita's husband had told the pastor in church about the goat-riding affair, and that 'he [Arvid] should not make a big deal out of it, for they say it was his own wife who had been riding'. 61 Erich took his oath in and said he had heard this from his aunt, Gertrud Markusdotter, the previous winter, but since the aunt was over sixty years old and lived more than four miles 62 away, she would not be able to come and testify in court. Margareta, who was named as a witness in the July session was also not present but confined after childbirth. She had, however, sworn to a local juror that she knew nothing about the matter. Finally, the local pastor testified that Thomas Ericksson had confessed to him that he had fabricated the story only to scare the children, not knowing anything real about a goat or anyone riding it.
The twelve local jurors testified that both Brita and her husband had always had a good reputation, had taken the Holy Communion as often as possible, and had never done anything forbidden. On the other hand, the jury noted, Thomas Ericksson had a previous conviction for theft, had purposefully got himself written up as a soldier to avoid the investigation, and had admitted being responsible for the talk. The court acquitted Brita of all suspicion and noted that she could try suing Thomas for slander 'as best she could', when and if he returned from the battlefields. 63 Like the Norwegian court record, the Finnish one is also a narrative that was put together, constructed, and edited by the court scribe. It, too, is a linear narrative -not of whatever was interpreted as witchcraft or the detection thereof, but of the events in the two court sessions. The narrative begins with the bailiff presenting the matter in court. For the rest of the narrative, the role of the protagonist shifts, and witnesses seem to drive their agendas. Among them, Brita's husband emerges as the protagonist: he had made the rumour public in the first place. He demanded that all those who had spread the rumour present their evidence, so it could be proven to be nothing. The narrative describes him calling for the witnesses and asking them to testify to certain things, one by one, in chronological order. The narrative ends with a sentencing section that was clearly separated from the previous text and presented as an item of its own, and that closed the case.
The purpose of the record was not to prove that witchcraft had happened -indeed, it ended up proving the opposite -but that the court proceedings had been properly conducted, that all reasonably attainable evidence had been gathered, and that a sentence had been passed according to the law. That is the master plot 64 of the court records, told in the narrative voice of the scribe, in as impersonal a tone as one would expect from a court scribe.
For the purpose of showing that all reasonable evidence had been taken into account, the scribe included traces of the original testimonies in his narrative. These form subnarratives and sub-plots, as the individual witnesses told what they knew or did not know of Brita's flight to Blåkulla, and how had they come to know it. The main narrative moves from one sub-narrative to the next by inserting introductions and explanations such as 'Arvid Sachariasson called for Erich Monsson as witness' and 'Erich took his oath and said'. These sub-narratives were translated from oral, rural, everyday Finnish into at least semilegal Swedish, and presented through the scribe's third-person narrative. However, they change and multiply the points of view of the narrative. The court record thus comprises multiple layers of narrative, with different voices telling the story; a multivocal chorus. 65 Brita's own voice is present in the story only briefly, as she denied all. The strongest voice is that of the scribe, although it is never personal. The voices of the other witnesses are merely traces, implicated in the scribe's third-person narrative. However, they are worth a closer inspection, and we will return to them in the sub-section 'Narrative as a way to construct and evaluate experience'.
The differences between a torturous trial in Finnmark with imprisonment and forced individual confessions and the more tolerant and multivocal one in Ulvila, as well as the outcomes of the trials, are apparent from the first sight. Neither type of trials was unique in Europe or North America. Historians have explained such differences referring to use of torture and circumstantial evidence, local circumstances, legal and judicial cultures, political conditions, and personal influences of judges and witch-finders, among other things. 66 Both the cases we consider in this article took place late in the general scheme of witch hunts in Europe -but not so late in Scandinavia, Norway or Finland, even less so globally. As far as narratology is concerned, this means that stories of witchcraft had been widely circulating for decades if not centuries, they were familiar to everyone, as were the 'scripts' of confessions and testimonies. The witchcraft narrative script may have lost some of its momentum as it became more known and more mundane, but it is also clear that it had not lost all of it, since executions still took place at this time in both Finnmark and Finland. The ways of creating an understanding of what had happened and whether it could be understood as witchcraft -through narrative and experience -were pertinent to all these factors, and we will be looking at that next.

The accused person's confession as a narrative
In this section, narratological structures in the accused person's confession will be discussed, as well as the confession being given as a result of pressure. Margrette Jonsdatter showed resistance for a long time. She was brought before the court six times before she finally confessed and denounced others. 67 When she finally confessed, she had been submitted to the water ordeal and floated, and she knew that severe torture took place at Vardøhus if an accused woman did not confess. Thus, taking in consideration that she first denied knowing witchcraft, was then exposed to pressure over several months, and finally had circumstantial evidence pointing against her, her confession must be interpreted as an enforced narrative. For her, continued denial would have meant gruesome torture. She well knew that a demonological confession would seal her death sentence, but she still delivered it, clinging to her ability as a storyteller.
Margrette's confession contains the main elements of a demonological narrative, namely the devil's pact, a witches' gathering, and the collective operation of witchcraft. In addition, her confession also provides a bit of information about her life, as she confessed that she had learned witchcraft from an old woman who worked at the same place she worked at in Tromsø, a town in Troms district south of Vardø. She had been given the evil power in cheese and milk, after which she became sick. She tried her art on a sheep, which died. This was common in Norwegian confessions. The ritual of learning witchcraft seems to be familiar to Margrette, and it is told without any distancing devices, like a recounting of a factual event. The same is true for the part where she enters into the devil's pact, in which the devil appeared in the likeness of a man, asked for her service, and promised her good fortune with cattle. Her first refusal is common in a demonological narrative, and the same goes for her turning from God in order to serve her new god, Zacharias. Margrette's mentioning of a personal demon, an Apostle, is found in several other European countries. 68 In Finnmark, it was a common idea that a woman who had entered into a devil's pact received a personal demon.
The witches' gathering that Margrette confessed to was located at the mountain Domen, just outside Vardø: She also confesses that two years ago on the eve of St Hans's [i.e., St John's], and also the year before that, she went to Dommen together with the following witches (. . .). On the said Dommen, they danced and drank beer and wine, and they also played board games, 69 each one with her own god, and the aforementioned wicked Satan played his fiddle for them, and the said Giertrud served them beer and wine in pewter cups, and they were there for an hour or an hour and a half. 70 The devil played various instruments. Often, he played a fiddle, 71 sometimes described as a red fiddle. In the confessions, different types of fiddles are mentioned, all of them known as traditional musical instruments in Norway. There was also mention in the records of a langeleik. 72 An additional detail about Margrette's participation is that she lost a shoe when she was dancing with Gammel-Erik 73 on Domen, and that he gave her a new shoe. This seems to be a motif from the widespread Cinderella story, which comes from folk belief. The gathering on Domen is a pleasant meeting with music, dancing, and drinking. When a detail like a 'pewter jug' is mentioned, it may be a sign that these meetings had an aura of sophistication. 74 Several details -like the pewter cup, Satan playing a particular type of fiddle, and the loss of a shoe while dancing -point to an elaborated narrative that was orally transmitted. Worship of the devil is subdued; the devil is mainly an entertainer. There is no sex and no cannibalism. Shapeshifting at the gathering itself is not mentioned, but it occurs in other parts of Margrette's confession. We see a merging of folkloric and demonological notions. The narrative has found content and a form that are adjusted to the local culture, while preserving the core elements of demonology.
Margrette also confessed to the performance of collective witchcraft, in the form of chasing the fish from the shore, casting a spell on the district governor's sledge, and trying to cast a spell on the ship of Captain Jens Ottesen. The latter operation was unsuccessful, because the crew 'evoked the Lord'. 75 One of the events is narrated in a beautiful way: She took part in beating the fish away from the shore with stalks of seaweed last Easter. And before the court, she also firmly denounces the following person, Søren Christensen's wife Giertrud from Krogen, saying that she was also there at the time, true enough in her own likeness, and she wore a black jacket, a red Bøffelbay 76 skirt, also a red cap with golden lace, and white linen around her neck. Besides, she sat on the water, holding and surrounded by seaweed. Margrette says that she herself was in the likeness of a gull, while Gundle was in the likeness of a seal, Dorette, Waarøe Hans's wife, was in that of a porpoise, and Sigrj Jonsdatter was in that of a bluefin, and they were all holding stalks of seaweed, applying their craft to drive the fish from the shore, and this they did from land and from the Islands. 77 Such an image displays Margrette's abilities as a storyteller. Narrative structures surface in her confession. Linearity -order -is created through the establishing of chronological events. The speed of the narrative, the alternating between quick and slow-moving parts, is seen among others in the description above of chasing the fish from the shore, in which the text comes to a standstill by the many words used in order to paint an image. Frequency -elements of repetition -is seen in the repeated words that are used when recounting the witches' gathering: beer, dancing, and drinking made the Domen party a good one. The mood of the narrative, the narrator's attitude towards what is told, is clear, as Margrette certainly believed that witchcraft works. Orality features are many, seen in the details she has spiced her confession with, the embellished expressions, the additive sentence structures, and the closeness to the human lifeworld that comes to the fore. 78 She obtained her mastering of telling stories from the oral field.
As for individual experience, Herman emphasizes that, in general, 'narratives can serve particular communicative purposes only on the basis of a progress of negotiation between storytellers and their interlocutors'. The confessions in the Finnmark witchcraft records support his point of view. 79 The confessions are enforced narratives carrying a stamp of a personalized experience. This individualization may be traced linguistically for instance in the occurrence of orality features in the voice of the accused person and signals an interpretation of ideas about witchcraft unique for each person.
Margrette Jonsdatter's confession has a personal touch. The pieces of information about where she learned witchcraft and from whom, her reaction to drinking the beer, and the name of her personal god come from her own knowledge. The same goes for the description of chasing the fish, destroying the sledge of the district governor, and partying at Domen. She makes the story her own. The demonological narrative, known all over Europe, has different variants in different countries and areas. The Finnmark variant is mild compared to those of other European countries. Still, the major elements of a demonological narrative are present, and by 1663, it was certainly known among common people in Finnmark.
When Margrette delivered her confession as a rich narrative, she showed that she had the strength to use her language even in a hopeless situation. She clung to a narrative she knew from oral transfer and in this way managed to stand upright to the bitter end, and even give the story a colourful flavour. Thus, on the one hand, she gave the interrogators what they wanted to hear: a confession from her own mouth that included the denunciation of other women, which made the panic continue. On the other hand, the context in which this story was told made it an enforced narrative. First, she showed that she could narrate; afterwards, she was severely punished because of it. This is the paradox in demonological witchcraft trials. The confession that showed pride and storytelling abilities was also what qualified the woman for a death sentence. We think it is wrong to say that the confessing woman in a demonological trial was cooperative; rather, she was forced, in a very subtle way, to show her knowledge and qualities in storytelling. Thus, in our view, a woman living in a local community and accused of witchcraft in the 1600s was a victim of strong forces in society, whether these were located in the state, the courts, or the church. To relate a demonological narrative in the courtroom in a seventeenth-century witchcraft trial, like Margrette Jonsdatter did, had only one outcome.

Narrative as a way to construct and evaluate experience
Confessions and testimonials in witchcraft trials enable the investigation not only of what the parties knew, but also of how they came to know things and what they considered to be knowledge about witchcraft. If experience is a process in which singular empirical events are given meaning, it must be an essential process for the creation of meaningful knowledge. 80 This can be investigated by looking at the connection between the competing narratives and the process of experiencing itself.
In literature studies, narrative is constructed by 'putting together a sequence of events' and in recent 'cognitive' narratology by adding experientiality to the sequence. However, narrative is also supposed to transmit experience to the reader. 81 Likewise, in the Early Modern courtroom, the narrative had the same dual function. It was constructed by smaller narratives that told a sequence of stories, but the experiential part that made it credible consisted of small, personalized details. At the same time, the narrative, or narratives, constructed an interpretation of what exactly was going on; that is, an experience. A properly constructed narrative could create or destroy an experience of witchcraft, as will be shown below.
Returning to Brita's case, we note that two other witchcraft cases were being investigated in the same district court session in Ulvila. Therefore, it is evident that there was a general belief in the existence of witchcraft. For the Early Modern inhabitants of Ulvila, as well as for the elite judge driving the investigation, witchcraft existed, was potentially dangerous to the community and its individual members and was therefore worthy of the court's time. The other cases concerned traditional maleficium and did not include Blåkulla travel, but the concepts had already been mixed in the popular culture. 82 It was therefore not necessary for the narrative to establish a credible explanation for the existence of witchcraft any more than it was necessary to establish that Brita had a goat. Both were considered realistic in seventeenth-century Finland and self-evident in the neighbourhood.
What had to be established, with credibility, was that witchcraft and Blåkulla travel had taken place here, with Brita committing the act. In witchcraft trials, as in many other types of narratives, this was done by adding local and personal details to the narrative; details that were recognizable and familiar in everyone's previous experience. With them, the narrative gained legitimacy in the eyes of not only the participants in court but also of who later read the record in the court of appeal. Part of this personalization and localization process was giving the witnesses a credible voice, albeit in a second-hand description. 83 The scribe also used the narrative devices that mark the switching between the main narrative and sub-narratives as a device to move between his own impersonal tone and the personal tones that portray the individuality and personality of the witnesses.
The narrative voices correlated with not only the subject matter of the narrative but also the social status and role of the witnesses. An example is the brief authoritativeness in the testimonies of the schoolmaster and the burgher from Pori town, as well as of Captain Böning, as they stated they had heard the rumour but knew nothing personal. This is what one would expect from a person in their position. Their testimonies also served to establish that a rumour widely circulated which held that Brita had indeed flown to Blåkulla. In seventeenth-century culture, a rumour or a reputation was evidence of at least something that had to be investigated. It suggested Brita could have flown to Blåkulla, and if no one denied it, it was at least a potential truth. The social status of these witnesses also gave them a relatively impersonal voice -they merely reported that a rumour existed, and no one asked how they knew that the rumour existed. This made it possible for them to present the case and make the accusatory rumours public without a personal risk of getting involved in the investigation.
The change in the narrative occurred in the beginning of the second part of the investigation, in September. After Brita Michelsdotter's personal declaration of innocence, the investigation narrative continues with the other Brita as the witness. As the scribe described how she hesitated and 'after a while' said Thomas had been glad to go soldiering in the battlefields and escape the investigation, a suspicion was cast on the source of the rumours. Thereafter, the rumours were traced to their sources, with the narrative repeating again and again that they were hearsay by having Henrich, Erich, and Margreta all stating who told them about these rumours. This soon began to make clear that Thomas was the only one who had ever claimed having seen anything -everything else was just hearsaybut it also served to further establish the extent of the rumour. Thereafter, it was no longer quite as much of a surprise when the local pastor -again in the blunt authoritative voice of an upper-class official figure -gave his testimony that Thomas had already admitted that he had invented the story to scare and entertain children. This was as much of a 'storyworld disruption' 84 as the formula of legal documents allowed, delivered in the pastor's authoritative bluntness. It resolved the narrative, but it was made more credible by the preceding hesitation of Brita the witness and the evasion of personal knowledge by Henrich and Erich.
The construction of the narrative, with different voices being given to different witnesses, also reflects the construction of experience in the court room. The different voices make it possible to see how different approaches to the same events were brought in for discussion. The next voices then comment and add their own experience, all made as credible as possible by a personal touch and personal detail added by the witnesses in court and preserved by the scribe when drawing up the main narrative of the court record. The emergent narrative is actually quite 'modern', in the sense that, instead of a general moral stance, it showcases the individual experience, based on which the common and shared experience is constructed. 85 However, whereas cognitive narratology emphasizes the individual and subjective experiencer behind a textual mediation of experience 86 in a historical exploration of experience in Finnish court records, the individual point of view does not emerge. Moreover, in a historical investigation of experience the point of interest shifts: instead of 'what it was like' of cognitive narratology, a historian of experience asks 'how did people make it into what it was like'. In this context, experience was a communally constructed and shared evaluation of various viewpoints. While it was possible for any participant to individualize and own it, the 'experiencer' in the court records was not an individual but a collective -'the court', including the officials, the parties involved and the wider audience who all had to be convinced for the matter to be efficiently solved.
Nevertheless, the individual experience was not enough for the construction of an acceptable truth. Hesitation, taking time, and adding afterthoughts, as well as the emphasis on the witnesses' part on the rumour being second-hand, were crucial in pointing that out. These characteristics of the witnesses' narrative voices were used to emphasize the instability and insecurity of individual and personal, unvetted experience.
As the narrative was formed, first orally in the court room and then in written form by the scribe, individual experiences were evaluated and turned into communal knowledge. The return to an authoritative voice by the pastor and, even more so, by the court as a community in the final sentencing then pronounced the eventual, and trustworthy 'Truth', which had been put together from piecemeal individual testimonies, publicly vetted and communally approved in the courtroom discussion.

Conclusion
In this article, we have, first, examined how a witchcraft trial document was crafted. In the case of court records of a witchcraft trial, the entire document can be seen as a narrative, with the function of the scribe being similar to the function of the narrator in structuring and compiling the text. Our focus has been on narrative discourse, particularly the category of Voice. The Narrator delegates Voice to the various participants of the trial, be that the Interrogator, the Accused Person, the Scribe, or the Law. A narratological approach to courtroom discourse used in this way will provide an understanding of the legal practice, contextual conditions, and complexity of the entire trial.
Second, we examined how meaning in a recorded witchcraft trial is expressed through the narratological category of voice by examining the confession of an accused person. In witchcraft court records, the confession of the accused person is a narrative embedded within the larger narrative of the entire trial. Rendered in indirect discourse, the confession is a narrative of its own, written down as a coherent story about the learning and performing of witchcraft. In the confession of Margrette Jonsdatter, we hear about how she first learned witchcraft from an old woman, about the ritual of taking witchcraft in through food, about trying out her newly won art to see if it was working, about the devil appearing and the devil's pact, about the witches' gathering at Domen and the events taking place there, and about the performance of collective witchcraft against boats, sledges, and even fish in the sea. By focusing on the confession as a narrative, the analysis unearths the ideas about witchcraft expressed by the individual accused person as well as the echo of the questions posed during interrogation. Such a personalized approach enables an understanding of the mentalities of the time, of how words uttered during the confession contain glimpses of experience and emotion. Features of orality surface in the confessions, thus underlining the scribe's professional recording of a document wherein individual voices may be heard.
Third, we have examined accents of court records of two cases from Norway and Finland that support and shed light on the creation and mediation of experience in a court record narrative, and on the influence of narratological structures in the confessions and thus on the evaluation of experience and the outcome of the trials. By dividing the narrative into sub-narratives, narrative voices expressing various accents emerged, all recounting their own version of the experience of witchcraft that was under investigation. Individually, the narratives served as a way to evaluate the individual experiences of the witnesses. Voice, order, and repetition or frequency were used to eventually turn the individual testimonies into credible and, ultimately, shared experiences of witchcraft, other forms of disturbance, or normality. Following the interplay between individual and accumulated narrative voices, the process of creating shared concepts of 'truth' and trust was made visible. In the Norwegian case, the long confession given by Margrette displays a very personal but coherent experience of witchcraft. She appears with her own linguistic accents preserved. In the Finnish case, however, the multiple short narratives by various witnesses provided a much more fragmented picture. However, as the witnesses' narratives were combined into a new, common narrative, they also communicated a shared and shareable experience of witchcraft. While our two cases are representative of the witchcraft trials in these countries -in that Norwegian records usually include a long confession by the witch, but the Finnish do not -the difference between these cases also serves to highlight the potential of narrative as a method for examining the creation and mediation of 'experience' in various types of court records. In both countries, the scribe has, through his diligent and professional work, created sources of a fascinating kind that can be used for the analysis of nearly 400-year-old courtroom discourse as well as for the history of mentalities.
Narratological analysis has been used before in historical and literary studies, 87 and later anthropologists or ethnographers. 88 Among historians, narratological analysis has mainly been used to show how witchcraft narratives appeared and how the narrative scripts guided interrogations and confessions in the trials, 89 but at times also to analyse narrative Voice. 90 This methodological experiment shows also that narratological methods can be used in showing how the scripts guided the actors in court to not only to create a story, but to interpret what they had seen or heard as 'truth' through shared experience. In the future, this may also help to investigate why some narratives, stories and rumours did not lead to trials and executions.