Opening the creaky binding of a 400-year-old book or straining to make out 18th-century handwriting might never lose its allure to some. But other scholars are putting aside the magnifying glasses and book cradles and turning to computers to help them better understand old texts and objects. And in this process, they're revealing new insights into history.
For researchers ready to dive into data with the help of a computer, Bowdoin's Special Collections & Archives is a treasure trove. Among the many collections in the archives is the personal library of James Bowdoin III, consisting of more than 2,000 volumes and thousands more pamphlets, which the philanthropist and statesman bequeathed to the College in 1811. He collected these publications during his lifetime to give to a liberal arts college.
There is also the collection from botanist and artist Kate Furbish, who lived most of her life — 1834 to 1934 — in Brunswick, Maine. Furbish spent decades traveling through Maine to collect plant specimens which she used as inspiration for delicately beautiful scientific illustrations. She donated approximately 1,300 of her wildflower drawings to Bowdoin in 1908.
Bowdoin's library also holds the papers of Civil War hero and Reconstruction figure Gen. Oliver Otis Howard, and those of his two brothers, including thousands of their letters. All three men were Bowdoin graduates. The Howard papers offer a broad and yet intimate portrait of a family in the second half of the 19th century.
This past year, students, staff, and faculty made digital forays into these three collections. In the interactive data visualization class taught by Visiting Associate Professor of Computer Science Clare Bates Congdon, students built computer tools to expand the ways in which Furbish researchers can interact with her plant drawings. Associate Professor of Digital Humanities Crystal Hall began an initial textual analysis of James Bowdoin III's library, and librarian Karl Fattig has been using powerful digital tools to explore the Howard brothers' letters. "The materials are here," Hall said. "The opportunities are exciting."
The increasingly close relationship between the library's collections and computer science reflects the growing prominence of a field called digital humanities. At Bowdoin, students are exposed to digital humanities concepts both within and outside of traditional compuer science courses. The Digital and Computational Studies Department offers courses that teach students in any major how to use digital technology for research projects in history, biology, sociology, English, political science, and more. Meanwhile, computer science faculty are also incorporating the humanities into their technical courses.
Using digital techniques helps students and faculty uncover new findings — within texts, between two texts, or among a collection of papers — that would otherwise have remained hidden.
The importance of metadata (data about data)
No digital humanities project would be possible without metadata. Kat Stefko, director of the George J. Mitchell Department of Special Collections & Archives at Bowdoin, said that her team balances being good caretakers of old and fragile things with making those materials as accessible as possible to the public. They do this in part by creating metadata to catalogue, label, and describe every object they hold. This metadata, recorded ands stored in a systematic order, can be easily searched and sorted.
"Library catalogers are the unsung heroes of digital humanities," Stefko said. "Because our librarians have done such a good job of describing the library collections over the years, digital humanists now have the ability to work with these robust metadata sets."
In addition, each time students work with a collection, they can add to the objects' metadata. "Metadata is never done," Stefko pointed out. "The more researchers work with collections, they more they understand it and can contribute to its description."
Indeed, as more digital humanists work with the library's collections, Stefko has observed a new kind of archive opening up — that of metadata itself. "It's no longer just metadata, but also a collection in its own right," she said.
In Clare Congdon's class, From Data to Visualization: Designing Interactive Approaches to Understanding Information, students design interactive, easy-to-use tools for large data sets, with the goal of aiding future researchers, or "helping people answer questions about their data," as Congdon puts it. "Vision is one of our most powerful senses," she continued. "Perception and cognition are related, so if we can present the data well, we're accessing our powerful vision systems to do a lot of work for us to understand it."
Data visualization tools can create pictures from data — like interactive bar charts, pie charts, and scatter plots that users can manipulate. "In making data visualization tools interactive, the viewer is enabled to look at the data from different perspectives, for example, turning on or off different features of the data, or looking at particular date ranges," Congdon said.
Last fall, the data visualization class broke into small groups, with one set focused on metadata from the Furbish collection. When Congdon and her students set their gazes on an archive like Furbish's, they don't just see the botanist's gorgeous flower illustrations. They also focus on the the thousands of "data points," associated with the drawings — like the plant's common and scientific names, and where and when the specimens were collected.
Early on in the semester, Stefko met with students in the class to introduce them to the biography and botanical work of Kate Furbish, and to tell them about some of the questions she would like answered about the collection. For instance, were there particular places in Maine Furbish traveled to frequently, or where she found many intriguing plants? In what years did she have the most luck?
"We have an enormous archive of her work, but very little personal information about her," Stefko said. Only one diary remains of Furbish's productive years during which she gathered plants in Maine. "She traveled all around the state. Is there a richer story we can tell? What can the data fill in for us?"
To help answer these kinds of questions, Congdon's data visualization students last fall put together an interactive map allowing people to see what was found, how much of it was found, where, and when.
For future student projects, Stefko wonders whether students might design tools that can answer even more questions: Did Furbish use multiple specimens to create a single drawing? And if she used, say, three samples, did she only have three samples? Or were some plants — or some parts of plants —more desirable as models than others? Because Furbish's herbarium of dried specimens still exists, students could use image-recognition software to match real specimens to drawings.
Stefko also wondered whether the techniques of digital humanities could establish Furbish as more than a mere "amateur botanist." Furbish has not always been taken seriously as a scientist, perhaps because she did not have a formal botany education, or because she worked in a field dominated by men. Yet, evidence suggests she followed rigorous scientific methods. "She corrected her plant labels, making sure her specimens stayed up-to-date with the current taxonomy," Stefko said. "And she found and exchanged rare plant samples with the day’s leading academic botanists. Perhaps through digital humanities techniques we could tease out the complexity of the question of whether Furbish was an amateur or a scientist."
James Bowdoin III built his library not just to satisfy a strong thirst for knowledge but also to fulfill his mission to launch an excellent academic institution. "Bowdoin was collecting books to donate to the college to be a foundational piece of a liberal arts college at the turn of the 19th century," Professor Crystal Hall said.
After Bowdoin died in 1811, Bowdoin College received not only Bowdoin's library of 783 titles bound in 2,048 volumes, but also his mineral collection, scientific devices, and art collection. With this gift, the College library likely became the third largest academic library in New England at the time. "His gift to the College really changed the character of the school, making it a much more consequential and serious place of study," Stefko said.
Based on early research Hall has done with the library's metadata she's been struck by the diversity of concepts the library represents. "There are many ideas here that aren't connected," she said. "When the pamphlets are included we get a sense of just how broad that conceptualization of a liberal arts college would have been."
With a couple of students, Hall has begun exploring the collection, and along the way adding metadata tags to the material they encounter. "We're conveying information to a future scholar," she said.
While we know a lot about James Bowdoin, his personal library presents some unanswered questions. For instance, what was his methodology? How did he know what to look for, and where to find it? What did he consider to be essential knowledge for a liberal arts education?
A future researcher might begin answering this last question by comparing James Bowdoin's library with the personal libraries collected by some of his peers to glean insights into how Bowdoin conceived of a liberal arts college, and whether his approach reflected current intellectual thinking or perhaps was more progressive.
Textual Analysis Tools
Researchers can use various tools to search, sort, or otherwise analyze texts. One can start by simply looking up the most frequently used words, for instance. Creating word clouds lets one see how prominent certain words were in a specific amount of time.
Papers, books, and other texts can also be quickly sorted into categories, such as by date or whether they were written by women or men, or say were written by people residing in Washington DC, Philadelphia or somewhere else. This data is all part of a book's metadata.
In addition, topic modeling programs can scan lots of text to find groupings of words — or words commonly used in proximity to one another. This helps researchers gain a sense of recurring themes in texts, and how those themes change from year to year, or over decades.
In this way, students can ask an almost infinite number of questions of the collections — and start to pile up answers. This sort of work is now being done with the letters of Civil War general Oliver Otis Howard and his brothers. "Interested students will have their own encounters with these letters and discover what aspects they want to look at," Karl Fattig said.
Meagan Doyle, the library's digital archivist, oversaw the National Historical Publications and Records Commission’s digitization of the 80,000 or so letters, photographs and diaries in the Oliver Otis Howard collection. Most of the letters were written by O.O. Howard or sent to him, but the project also includes the collected letters of his two brothers: Charles Henry Howard and Rowland Bailey Howard.
For each of the 80,000 items there is associated metadata: including the date, recipient, writer, and place from which it was sent. Because there are so many letters written and received by the Howard family across a century, there are thousands and thousands of data points. The scale of the metadata set alone supports projects to discover patterns, trends, and changes over time within the collection.
Doyle offered a few questions that she'd like to see students investigate: For instance, she's curious about when and where O.O. Howard traveled, and who was writing to him when, and from where. "He was receiving correspondence from diverse people, and these people change over time," she said. "It'd be interesting to see who he was corresponding with a lot, when, and why. A lot of that is hidden right now."
In addition, about 1,300 of the handwritten letters from the three Howard brothers have been transcribed into digital text by a family descendent, opening them to even more expansive research. Karl Fattig, who is Bowdoin's system and digital initiatives librarian, has been delving into these transcriptions with textual analysis tools, like word clouds and topic modeling. His aim is to master these techniques so he can teach Bowdoin professors and students how to use them.
As he becomes more knowledgeable about the digital tools, Fattig has also become more familiar with the Howard family. "One thing that jumped out at me," he said, "is the bond between the mother and sons. Their father died when they were young."
This intimacy became evident to him when he searched for the most frequently occurring words in the collection. These handful of words refer to writing, specifically letter writing, and staying in touch, which Fattig believes reflects the boys' desire to stay in close communication with their mother, Eliza. These "phatic expressions," he said, are evoked most frequently when the boys are students at Bowdoin, when they would have been away from their mother for a long time.
The sentimental statements decrease in frequency over the years, Fattig noticed. He hypothesized that as the boys grew older, their confidence grew and their anxiety lessened about their positions in the world, and they leaned on their mother less for reassurance and sympathy.
In addition, Fattig has observed that like many women of her era, Eliza had a less formal education than her sons. He's interested in using digital tools to reveal whether the language and vocabulary the mother and sons use in their letters are different, or whether the sons adjusted their language in communication with their mother.
Anther possible angle a future Howard researcher could take is to examine the role of religion in the men's lives, and how their religious views changed over time. Olivier Otis Howard, in the late 1850s when he was in his 20s, became a born-again Christian. In the letters, "they shared their religion with one another," Fattig said. He added that he thinks it would be fascinating to analyze how Howard's religious language changes before and after his evangelical conversion.
Bowdoin College's collections of James Bowdoin III's library, Furbish drawings, and Howard family papers present a unique invitation for students to engage with history.
“Data can be dry and boring, but when it relates to a person important to the history of the college, it can make the data more accessible and interesting to a student," Stefko said. “Data, in the hands of digital humanists, has transformative possibilities—it can open up a new window on our history, it can allow us to imagine and understand our past in new and more nuanced ways, and it can tell us new things about distant historical figures. In my mind, one of the most powerful aspects of digital humanities is the possibilities it brings for humanizing data.”
"The students are glimpsing lives of people caught in a moment," she added.