Kirsten C. Uszkalo - scholar/guest curator

Troubling Transformations: The Witches in Early Modern England Project

Problem/Question:

Women in early modern England did not suddenly appear as witches, scratching at their neighbors' doors. Nor did the interpersonal tensions between women explode into trials and executions until an accuser came to understand the accused as a witch. The OED defines a witch as a "female magician or sorceress" who is "supposed to have dealings with the devil or evil spirits and to be able by their co-operation to perform supernatural acts." In practice, however, the parameters that defined a witch were less distinct. In fact, a woman first had to be seen as the proper kind of candidate, and publicly labeled a witch before the law could be brought in. The witch's physical body needed to display, or adopt, the key signifier - the witch's mark, her identity needed to extend to include a familiar, and her new job title needed to receive the forthright endorsement of her community, before she could even begin to be publicly persecuted as a witch.

I am looking at published early English witchcraft tracts. I have a number of research question in regards to this work; however, I'd like to present three interrelated questions which focus on the problems and possibilities associated with issues of transformation.

1) How can we trace the familiar which will not even stay in the same shape?

The familiar is, for the most part, an animal spirit which does the witch's dirty work for her. It is through the familiar, rather than spell casting, that witch's hurt their victims. How do we investigate the slippery and morphing familiar whose presence in the text is what publicly signifies a woman as a witch, but whose name, form, and use may change based on who owns it and what use they put it to? The first published witchcraft case is that of Mother Waterhouse and her familiar Sathan (1566). It creates a blueprint for the economics of the familiar which will continue to frame published cases to come. The story began when Elizabeth Francis testified that her grandmother, Eve of Hatfield Peverell, taught her to renounce God and give her blood to Sathan. Sathan appeared as a cat, which she kept in a basket and fed with bread and milk. Francis gave the cat to Mother Waterhouse for whom it first appeared as a cat, then morphed into a frog, next became a frightening dog when her daughter Mary needed it to frighten Agnes Brown, and finally became a devil-dog when it actually appeared to Brown.

2) How can we reconcile the meaning of a witch's mark which keeps moving, has its meaning morph, and mutates its owner?

Women were identified as witches in a number of critically inescapable ways - they could not be hurt unless their protective and unnatural hardness was scratched or pierced; they were insensitive to pain in the place the devil marked them; they could not cry and would be rejected by water (and thereby float); they were old, ugly, and stupid; but most critically, they had a witch's mark - a symbol of their demonic pact. Locating and exposing the hidden witch's mark was considered irrefutable proof that a woman was a witch. The trouble with witch's marks is that they changed. Like the familiar which sucked from them, their form and function morphed. The witch's mark began as a kind of invisible, insensible mark, morphed into a flea-bit of a sore, became an excrescence where a familiar would bit or suck, then became a kind of nipple which lactated unpurified blood, and that nipple moved into the witch's genitals. The mark also changed meaning - from a sign of power, to a way to renew a contract, to a means of placating a feisty familiar, to a kind of sexual organ used to pleasure the beast and hurt its bearer.

3) How can we trace the series of accusations, cross-accusations, relationships, and gossips which turned a restless baby in 1657 into a double execution in March 1662?

How can we trace the movement of accusations? In the case of the Lowestoft witches the suspicions against Amy Denny began in March of 1657 with the illness of Dorothy Durrant's son. In 1659, Elizabeth Durrant dies of maleficium. In November 1661 Samuel Pacy's daughters identify Denny and a mystery witch as those responsible for their bewitching. By the end of November, Cullender is recognized by the community as a witch. By February 1662, Cullender is accused of working alone to hurt Susan Chandler. The dizzying series of cross accusations in the community of Lowestoft provided a network where accusation could not only spread, but multiply. In tracing accusations across a community, and mapping how they spread, we can not only see how ideas about power proliferated, but the real way that the same concepts changed lives, one pin-vomiting fit at a time.

Status of the research:
This work began as an off-shoot of a number of interesting discoveries I made when researching and writing my dissertation "Dark Sisters: Witches and Prophets in Early Modern England" (2006). I began to notice the significant slipperiness shared by witches, familiars, and accusations. Because all three elements can morph within a single tract and across a corpus, they not only resist accessibility, but also interpretation. It became clear that only a scholar with an intimate lexicon of witchcraft lingo and a keen understanding of the meaning of representational slippage could hope to understand the meaning and minutia encoded in the texts. A researcher new to the material may not see those connections.

I have already begun to compile a database of the ways in which these texts connect, looking at the kinds of relationships which appear as significant indicators of meaning within the corpus. Among these are relationship types (mother, sister, sibling, accuser, accused), familiars (shape, type, owners), and descriptors (name, age, significant marks). I have started to use TEI to tag the first texts.

I was part of a poster presentation on the WEME project at the 2006 TEI conference in Victoria, British Columbia. The poster had a positive reception and resulted in some help from Julia Flanders, director of the Brown Women's Writers Project who has generously offered to let us use the encoding for the texts in Women Writers Online as a base.

Measure of success:
The ability to visualize the relationships within the data in a clear, compelling, and beautiful way would be a huge asset to this work. It would be stellar if it were possible to simply trace the transformation of familiars within a set of texts, for example from feline, through several steps ending with the demonic fiend. It would be mind-blowing to be able to connect the three research questions, looking at the ways in which the sexualizing of the witch's mark related to the form of the familiar and how that was manifest in a single accusation trend in a city like Lowestoft.

Texts needed in the collection:
I have already secured permission from the Cornell University Library to use the digitized texts of the Cornell University Library's Witchcraft Collection as part of my research. This freely available, online collection has 102 texts, and includes tracts which range from 6 pages to encyclopedic works of over 600 pages. I have also secured permission from Early English Books Online to use their texts on witchcraft. EEBO's texts are usually limited to subscribers, and although I have not yet received permission to their images, there are open to aiding my research.

This work can be successful conducted with as few as fifty short tracts (of 6 pages and up) or 300 some odd pages; a better scenario would look at all of the secondary textual material available in EEBO and Cornell; in a perfect universe, it would be wonderful to eventually expand to look at records not yet available in digital form, in order to expand the search to the rest of Europe and North America.

Generality:
It seems to me that being able to see patterns changing over time would be useful in a variety of contexts. For other users looking at early modern witchcraft trial documents, it might be interesting to be able to see what kinds of tracts different kinds of testimonies come from (i.e. is it a court document, a narrative, a true story) or where the texts were published (some printers produced more witchcraft tracts than others). They might also want to see a geographic break-down of where and when the various accusations happened across England.

Granularity:
Texts, words, sentences; a view across time and geography would be particularly useful.

Characteristics:
I would like to look at issues such as naming and perform advanced searches on classes of familiars. I would like to set a tool to find information within a certain set of parameters: i.e. a tool that would look for familiars and their location. I believe that as the witch hunt and trial became more influenced by European texts, that the shape of the familiar moved from the more domestic animal (cat, dog, toad, rat), to the more fantastic (tiny cow spirits), to the demon anticipated in the European philosophies on maleficium (bad magic).

The same can be said about witch's marks - they became more sexualized as time passed. A tool that could go out and look for these kinds of patterns would be most useful.

Patterns:
The pattern of transformation across my three different research questions is the single most important aspect of this study. Because this work is all about transformation, it is critical to be able to trace and see how patterns of transformation happen. The transformation of the familiar, within single tracts, and across time and geography; the way the witch's mark transforms from spot to nipple/phallus across time, and the way the accusations spread across villages (and if the relationship of publication and accusation can be traced).

My research is predicated on my theory that the publication of witchcraft tracts encouraged communities to look for witches in their midst. I would like a tool that could display a map of sixteenth- and seventeenth-century England which would display the location of cases of published cases of witch-craft accusations temporally and geographically. The early work we have done on this illustrates that the publication of tracts suggests a different pattern of accusations than do court records. This suggests that witches were not operating where people, because of publications, assumed they were, and that community infamy had more to do with press than with reality.

On a smaller scale, I would also like to be able to map how accusations moved through a community. Although I can trace how that moved from person to person, I am not able to show how word spread without visualization tools. This kind of work will enable me to illustrate not only the kind of influence child and adolescent accusations had on witch-persecutions, but also, in a very real way, how word spread.

Morphology:
As it would be useful to track how accusations moved through communities, and in a greater sense, how they developed through England, from the passing of the 1563 Witchcraft Act (our first case happens in 1566) to the last typical case in 1712 (the Acts against witchcraft were not taken off the record until the mid twentieth-century, it might also be useful to visualize how many tracts cluster around the two peaks in witchcraft accusations in the mid sixteenth- and mid seventeenth-century).

Tags:
I am already using the TEI Guidelines for Electronic Text Encoding and Interchange including date, author, and publisher. However, the TEI guidelines do not provide tags to describe transformative elements such as the shifting form of the familiar Sathan whose name stays consistent, but appears alternatively as a cat, frog, dog, and devil dog. I would get around this by creating a <familiar> tag and assigning it ID and form attributes that map to a supplementary database. The following are examples of how this markup might look in-text:

... <familiar fID="1" form="2">Sathan</familiar> ...
... <familiar fID="1" form="1">toade</familiar> ...
... <familiar fID="1" form="3">black dogge</familiar> ...

In the supplementary database, the fID attribute would match up to a given familiar's unique identifier, in this case "1" for Sathan. Likewise, the form attribute matches a given form's unique ID (e.g., "3" would identify the form as "dog") in the database.

I also need tags to mark people within a text and their relationships to one another, both familial (grandmother, mother and so on) and legal (accusers and victims of witches), and their locations in both time and space.

This work relies heavily on the ability to find patterns between tagged and untagged texts. A tool which could look at what I have tagged and go find similar attributes would be most useful.

Classification:
For the purposes of my research questions, sections of the tracts fall into different categories. Initially, it would be useful to be able to have the system identify places that talk about familiars, accusation, and witch's marks.

Comparisons:
The study is predicated on being able to compare texts in terms of element repetition (e.g. presence, shape location, use of the witch's mark), and compare elements within texts (e.g. the testimony given by Margaret and Philip Flowers seem to repeat elements such as kind of familiar, and align more and more through time and number of examinations?by the last examination they both separately confessed to having a white familiar which sucked from a spot on their left breasts)

Topic extraction:
I'm not sure how this differs from classification.

Lexicon, counts of words, most common occurrences, concordance:
The way familiars and witch's marks are named are as messy and apt to transform as the elements themselves. In terms of having a concordance, I think it would be useful to point to the different names these elements might have (e.g. a familiar might be identified as an imp, a familiar, a spirit, a familiar spirit, a tiny spirit, a demon, a devil, the Devil, a boy and on it goes, whereas a witch's spot might be defined as a the devil's mark, a spot, a blewish spot, a spot that that sucked, a redish spot, a teat, a nipple "lately sucked," a protuberance etc.). Without knowing the kind of language and the way it transforms, scholars can at times only guess what is being described.

Annotation:
Note tools, highlighting, ability to save relationship clusters, mark geographical locations, note time frames.

Collaboration:
I would love to collaborate with Early English Books Online, The Cornell Witchcraft Collection, The Brown Women Writers Online, and the Salem Witch Trials Project. I would like to share the tag-set, and relationship manager (we call it brimstone). We might not be able to share some of the geographic/temporal tools as they would not be supported on their sites.