Court officials? Or perhaps the nature of Puritan religion itself?3
Inspired by the “new” social history, later scholarship moved away from identifying the parties most responsible for the crisis. These authors sought to understand how this could have happened, or, in particular, what set the trials in Salem apart from other contemporary witch hunts.4 Today, academics are less concerned about finding fault or assessing blame and instead recognize that there was not a single cause but rather a variety of reasons why the Salem witch trials occurred. Moving beyond why it happened or who or what was to blame, historians now use the example of Salem as a reflection of larger themes in colonial life.5 Digital history—the latest trend in historiography—provides the opportunity to challenge or confirm previous scholarship and also to ask new questions.6
SOCIAL NETWORK ANALYSIS OF THE SALEM WITCHCRAFT PAPERS
The Salem witchcraft papers, which include testimonies, court transcripts, depositions, and arrest warrants, are an invaluable primary source that provides a key window into this historical moment. In 1938, President Roosevelt’s Works Progress Administration provided funding for an extensive new compilation of Salem witchcraft materials. Under the supervision of historian Archie N. Frost, multiple archives were searched and their handwritten records transcribed. Forty years later, in 1977, historians Paul Boyer and Stephen Nissenbaum refined and published these transcriptions in The Salem Witchcraft Papers: Verbatim Transcripts of the Salem Witchcraft Outbreak of 1692.7 In the early 2000s, Boyer and Nissenbaum’s transcripts were digitized by the University of Virginia, making them even more accessible to historians today.8
For the students to complete a sociogram, they first had to complete a social network analysis of the Salem Witchcraft Papers. A social network analysis is the study of structure and relationships within communities. The mode of analysis has a long history in sociology and mathematics and is a recent addition to the humanities and history. Social network analysis concentrates on relations among members of a community rather than on individual attributes of those members. The objects of the study (persons, institutions, airports, etc.) are commonly referred to as actors or nodes. In this assignment, the nodes were individual persons involved in the Salem witch trials—that is, anyone who appeared in the court transcripts, depositions, warrants, and other primary documents that make up the Salem Witchcraft Papers. The lines that connect the nodes to each other represent relationships. These ties, also referred to as edges, can represent specific kinds of relationships. For this assignment, students labeled each relationship between various actors in one of five ways—accuser, coaccused, witness against, witness for, or court official. (See fig. 1.1.)
INTERPRETING CHALLENGING DOCUMENTS
Before students could close read the Salem Witchcraft Papers to construct a data set needed to create the visual map of nodes and edges, they first required instruction on how to better comprehend these seventeenth-century documents. Although the original handwritten documents had previously been transcribed, these were verbatim transcriptions. Early American printers often capitalized the first letters of words in ways that can appear quite arbitrary to a modern reader. Spelling and grammar could also get creative as no standardized spelling existed in 1692. Misspellings, contractions, and sentences that run on for many lines with little or no punctuation can become unwieldy. Moreover, documents that were not originally meant for public consumption tend to be more erratic in their spelling, grammar, and use of capitalization. This is particularly true of legal documents, like those from a court proceeding, many of which were hastily recorded and under considerable stress.
To assist with this task, students were responsible for annotating court transcripts and depositions regarding Bridget Bishop, the first accused witch to be executed during the Salem witch trials. A number of free annotation software programs are available online. Annotation software allows students to highlight and paraphrase discrete lines in a challenging text for stronger and more active reading comprehension. Students in my class used MIT’s Annotation Studio (http://www.annotationstudio.org/). It should be noted, however, that one does not have to use specific annotation software for this part of the assignment; even copying and pasting the documents to be interpreted into Google Docs works for this task. In fact, a tool like Google Docs that many students are already familiar with may be even more useful. The goal of this part of the assignment is not to teach how to use a new computer program but to practice a skill—in this case, the annotation of arcane court documents—that can become internalized after considerable practice.
PRELIMINARY PAPER
After obtaining firsthand experience reading and successfully interpreting seventeenth-century documents, each student chose one of the accused witches and wrote a short paper (two to three pages) on this individual. Students used the digitized court documents as the basis of their evidence to respond to two questions. First, why did people think this person was a witch? And, second, how did he or she respond to those accusations? Students were not allowed to write about Bridget Bishop because they had already annotated documents from her court docket. This necessitated additional annotation of different primary sources in order to complete the short-response essay.
Figure 1.1. Zoomed in, using Palladio, to the center of the sociogram of individuals involved with the Salem witch trials. The larger the node (circle), the more often that individual appears in the court documents. The darker nodes belong to individuals accused of witchcraft.
MINING THE DATA
The third step in this scaffolded project was the creation of a large data set from which we as a class could begin to analyze the connections and ties among individuals involved in the Salem witch trials. Each student was assigned another accused witch—ideally, one they had not written about in the previous assignment; they then read through all of the digitized documents related to that individual and recorded data on an Excel spreadsheet. Their task was to record each person named in their specific documents and identify how that person was connected to their accused witch. Were they the individuals who made the accusation? Were they accused of witchcraft in the same document? Were they mentioned as someone who was injured by the accused witch? Were they a petitioner who defied the court in declaring that this individual was not a witch? Other information such as gender, age, marital status, and place of residence was also recorded as network attributes. Students were provided with a template in Excel to help them keep track of this information and remind them of what information they needed to record. In later class discussions, students identified even more categories beyond those provided on the Excel template that could help historians ask questions beyond the Salem trials in a study of general colonial life in New England. Overall, 30 students were able to record the information for nearly 600 different actors (nodes) and over 1,900 relationships (edges) between those actors.9
After merging students’ data into one master Excel file, we were ready to create a sociogram of the recorded connections and attributes of individual actors. We used Palladio—a free online visualization tool in active development by Stanford University.10 Like Annotation Studio, Palladio runs on all internet browsers and is, therefore, platform-independent. The website offers helpful tutorials for those wishing to learn the program, which is fairly intuitive compared to other social network analysis programs. For example, other programs require a separate file for nodes and one for edges; in Palladio, the master data set need only be saved as a single .csv file and then dragged and dropped into the program’s data window. My goal was not to make students experts at Palladio in a few weeks’ time but to familiarize them enough with the program’s filter functions so they could manipulate the visual map we had created and make some preliminary observations. (See fig. 1.2.)
Figure 1.2. A complete sociogram of individuals involved with the Salem witch trials, using Palladio.
FINAL