Recently I asked Dr. Michał Choiński and Dr. Jan Rybicki, professors at the Jagiellonian University, Kraków, Poland, if they would have time to give an interview about their recent Digital Humanities research in Edwards studies. They graciously agreed and granted JESociety a technically-rich peek into their research methods and findings.
Rob Boss: Michał, tell us how Jonathan Edwards first became a research interest to you and why you were interested in his life and works.
Michał Choiński: I spent almost ten years doing research on the rhetoric of the Great Awakening preachers. I was determined to understand the source of the powerful use of language which characterized 18th century colonial revivalism. What was happening there on a purely communicative and sociolinguistic level that allowed a group of speakers to exert such a powerful impact on the minds and hearts of their hearers? So, to answer that question I read a lot of sermons by George Whitefield, Gilbert Tennant, Jonathan Parsons and—obviously—Jonathan Edwards. I also studied responses to these sermons and the contemporary accounts of revival “harvests.” I must say that it was a rather intensive and overwhelming experience to submerge oneself so strongly in these texts for a longer time. But that early colonial “rhetoric of the revival” impressed me a lot as a blend of figurative stratagems, theatrical ploys, and topical diversity that seemed quite unique from a historical point of view. In the book that came out of my research, The Rhetoric of the Revival published in the New Directions of Jonathan Edwards Studies with Yale JEC Centre, I seek to stress precisely the richness of that pulpit oratory, as well as its inventiveness.
Edwards’s sermons exhibit to me impressive literary qualities and I have been puzzled by their figurative complexity ever since I read Sinners as a sophomore in my American literature classes, as a student at the Jagiellonian University. However, having been reading these texts for almost a decade, I came to realize that—to put it simply—I was short of the lifetime needed to acquaint myself with all the relevant materials. Edwards himself wrote more than fifteen hundred sermons. How do you manage such a collection as a reader? Not to mention noticing connections between various groups of texts—say, if you read two thousand sermons by ten authors, can you remember them well enough to see the stylistic patterns they share? That’s why I was so delighted when Jan Rybicki proposed that we team up to use stylometry to study colonial writings and to apply for an OPUS research grant from the Polish Ministry of Higher Education.
Rob Boss: Tell us more about Digital Humanities. What do you see on the horizon of DH?
Jan Rybicki: Digital Humanities is a relatively new concept; perhaps ten years ago, it replaced “humanities computing” and its even narrower predecessor, “literary and/or linguistic computing” when it became clear that many humanists who used digital tools and/or who reflect upon them could not really fit into such a description of the discipline.
The interesting thing is, it is not. DH is not a discipline; it is not even a transdiscipline (like, say, translation studies). Instead, I think it is a product of the relative novelty and rarity of applying information technology to humanities data. This, too, is interesting, since it is difficult to find a more humanities concept that information. Strange, too, is the seemingly triumphant march of DH in the global academia when, at the same time, most of the (traditional) humanities community treats it with mistrust or, worse, as an unworthy yet dangerous rival to the ever-dwindling financial resources of the humanities disciplines. And yet DH centres, even if they are inaugurated with much pomp by university authorities, are far from comfortable financially, frequently unaware of where the necessary money will come from and reliant on a grant-to-grant existence; as a result, their lifespan is often very, very short.
This sorry state of affairs is, in my opinion, rooted in the fundamentally erroneous approach to DH. I do not agree that my own research is special or noteworthy or simply a fun way of spending my time just because I use a computer. The questions I try to answer in digital literary studies are fundamentally literary ones: they are all about language, literary tradition, the creative process, classification of literary phenomena. I just happen to use methods that are computational. There was a time when physics did not use computers; now all physicists do, and yet nobody calls them DIGITAL physicists.
If there is anything truly revolutionary in the way digital humanists work, it is how even the most literary among them tend to work in groups rather than alone. Talks at DH conferences and papers in DH journals are now more often than not authored by more than one person. An ideal DH research project should be a joint venture of several people across the spectrum between computational and literary studies, preferably at varying and different degrees of the two. And this is very, very exciting.
Michał Choiński: Yes, I fully share that enthusiasm about the interdisciplinary nature of DH. Any research in this field often requires teamwork—one person has to do the calculations and graphs, and then another person to interpret the results in the context of a very specific background. So, DH as a research method is inevitably cooperative. I am also quite excited about how DH is gradually entering the methodological mainstream of American Studies. You could find DH panels at a number of recent conferences on American studies: in 2016, Dr. Katie McGettigan organized a DH panel at the IBAAS Conference in Belfast, another panel was put together by Dr. Lauren Tilton at the joint conference of BAAS and EAAS in London in April 2018, and in May Dr. Alex Dunst set up yet another large panel for the GAAS Conference. It is really great to see how the papers delivered at all these panels remain in a synergetic dialogue with one another. Then, you also have special issues of important journals on American Studies dedicated to DH—just this autumn, a huge special issue of American Quarterly came out and it was immediately followed by the publication of a special issue of the German journal Amerikastudien dedicated to DH. And… in 2020 we will have a special issue of Polish Journal of American Studies dedicated to DH in American Studies.
Rob Boss: Would you please give us a brief explanation of R and Stylometry and their use in your teaching and research? How did these tools help in your Edwards research?
Jan Rybicki: It is a statistical programming environment. It is open-access and this means: one, it is free, and two, anybody can write packages, or libraries, or little “programs” aimed at a particular collection of “things” they want to look at using stylistics. There are packages to draw phylogenetic trees and there are packages to count words; but both can be combined to conduct even more advanced experiments on language. In comparison with the (non-free and very often very expensive) statistics packages from the likes of Excel to Statistica, R wastes much less computing power on the graphical user interface and ease of use, so it is much faster. The price for this is the steep learning curve, but this also has the added value of greater flexibility.
Stylo is a package for R oriented to the very specific needs of stylometrists: people who count whatever can be counted in texts and who draw conclusions on the literary features of these texts (i.e. style, or content, or both). It is certainly not the only R package for linguistic analysis; but its success is owed, I think, to the fact that it offers a shortcut around R’s steep learning curve for humanists who would like to conduct their analyses without, perhaps, taking a dreaded computational course. It can be used in very expert ways; but it also offers a GUI that allows users to conduct quite advanced stylometric analyses in no time.
Michał Choiński: I must say that for all my enthusiasm to DH, I approach the application of stylometry to literary studies with some healthy reservations. To me, stylometry is a tool that is immensely helpful in asking the right questions or connecting the kind of dots which are way too large—or too small—for readers to see with the naked eye. Yet, it can never be a substitute for the direct experience of reading. In other words, one of the benefits of stylometry consists in having an idea about what you may look for in the close reading process. Quantitative research can also help us notice connections between groups of authors we’d have hard time arriving at on our own. Or, it can help us to challenge certain stereotypes about writers or groups of texts. For instance, as part our research we investigated the way the image of God changes in Edwards’s sermonic corpus. Again, this is a big collection—and the word “God” appears there more than thirty thousand times. Just imagine going through every single occurrence and making notes of its semantic context by hand. We used a somewhat simplistic piece of software to do that and what emerged was that the image of God in Edwards’s writings is not only very removed from the stereotypical “angry” Almighty from “Sinners” or “Future Punishment”—which most avid readers of Edwards know but which some students of American Literature who only read “Sinners” may find surprising. What struck us, however, was the extent to which this figurative representation of God interacts with the world on the metaphorical level—in other words, it is an image of God who is not casting sinners into hell, but who hears, sees and touches the world—a result surprisingly consistent with Lockean epistemology. And this is an empirical, quantitative result which we can prove with statistics. We also studied the relationship between Jonathan Edwards and Thomas Foxcroft, his literary agent, with a method developed by our colleague, Professor Maciej Eder, Director of the Institute of Polish Language at the Polish Academy of Sciences, called “rolling classify.” It allows to see changes in the stylometric authorial signal in different segments of a text. In this way, we were able to identify those short fragments of Freedom of Will that were probably heavily modified by Foxcroft in the editing process.
Rob Boss: There is a dizzying array of tools and programming languages used in Digital Humanities. What are your recommendations on where a beginner should start?
Jan Rybicki: Doing Digital Humanities is not something you do for its own sake. You need a research problem you want to solve, and the realization that it can help to use a computer and one or more piece of software. A manuscript of an anonymous novel is found is someone’s attic; once it is digitized and the possible suspect authors are identified, statistical analysis of most-frequent word frequencies might not only show the real author, but also provide a good assessment of whether the list of suspects is complete. There is a dispute on the chronology of a certain author? Lexical analysis can help that too. There is a collection of several thousand books and you want to establish some kind of pattern between them—you cannot read several thousand books; even if you could, you would not remember what they were at the end of such an enterprise. Yet once they are digitized, a number of interesting patterns can be observed using quantitative methods which, in turn, might lead to qualitative insight. I could go on…
Rob Boss: Can you tell us about your current work or projects in the near future?
Jan Rybicki: A stylometrist’s work – especially that of someone who is interested in the large-scale analysis of literary data – is never finished because new electronic editions of old or new literary texts continue to appear. I am particularly interested in the strange phenomena associated with literary translation.
Michał Choiński: My current interest gravitates towards the study of regionalisms in American literature with the help of stylometry. Together with Jan Rybicki and Maciej Eder, we have been trying to see if one can investigate the “Southernness” of the writings which originate from the American South. In quantitative terms, is there any stylometric resemblance between the 300 novels written by Southern authors, such as William Faulkner or Flannery O’Connor, in comparison with writers who come from the West or New England? The initial outcomes are quite promising, but I think we will need to apply for a new grant to be able to continue this research.
Rob Boss: Thank you for giving us a fascinating look into your research!
Michał Choiński: Thank you for having us!
Jan Rybicki: Thanks.