Print-based Humanities Research: Is it Time for a Fresh Look in the Digital Age?
Column Editor: Bob Holley (Professor, Library & Information Science Program, Wayne State University, Detroit, MI 48202; Phone: 248-547-0306; Fax: 313-577-7563) <firstname.lastname@example.org>
The other day, I managed to invalidate a major conclusion of my dissertation within five minutes of searching a French database with its new search engine. I completed my dissertation in French literature at Yale University in 1971. I studied a very minor genre, dialogues des morts (dialogs of the dead) that had a brief period of popularity from around 1680-1720. Two or more characters meet in the underworld after their deaths. The characters could have never met in life or would not have been able easily to have a conversation, such as Socrates and Montaigne in the first case and Erasmus and Charles V in second. Bernard le Bovier de Fontenelle used this convention to show off his wit, while François Fénelon emphasized the historical aspects of the genre to educate the heir to the French throne. I based my study upon what I considered to be the definitive list of these dialogs in a book published by Johan S. Egilsrud in Paris in 1934. In part because of this list that contained only nine titles after 1800, one of my conclusions was that the genre died out when readers no longer understood the classical conventions upon which the genre was based.
On May 23, 2011, Sarah G. Wenzel posted a message to the Western European Studies Section List about Gallica, the new search engine for the digital collections of the Bibliothèque nationale de France (BNF). I clicked on the links until I arrived at Gallica (http://gallica.bnf.fr/). To test the search engine, I typed a title keyword search for the topic of my dissertation, “dialogue des morts” (“dialogue of the dead”). Much to my surprise, I discovered seven entries from the 19th century, many more than I would have expected. In fact, only one of these titles appeared in Egilsrud’s “definitive” list. My curiosity piqued, I then searched the full BNF, where a title keyword search discovered 62 entries. I quickly found in the first few pages several more items that Egilsrud had overlooked. A final search with the term “dialogues des morts” in the plural as keywords in the title provided 14 entries in Gallica and 62 in the general BNF online catalog. Thus, the genre appears to have remained more popular than I thought in 1971, a contradiction to one of my major conclusions.
My discovery is disturbing because Egilsrud is one of the key sources for bibliographic information about dialogs of the dead in English, French, and German. The one other dissertation on the subject that I could view in full text cites him. He has multiple entries in both Google Books and Google Scholar. These citations hint at the fact that several scholars have made use of his bibliographic findings without questioning their accuracy.
A key question is why his bibliographies are incomplete. The most obvious answer is that Egilsrud was a bad scholar and missed things that he should have found. A second possibility is that the libraries he consulted, including the BNF for the French entries, had acquired or cataloged the additional items since he published his book in 1934. From looking briefly at the entries in both BNF sources, I’m inclined, however, to advance a third hypothesis that keyword searching has made it possible to find bibliographic information that was lost in the traditional card catalog. Of the nine title entries in Gallica with the term “dialogue des morts,” six were embedded deeply enough in the title that a traditional card catalog search would have been extremely unlikely to have found them. The more powerful searching capabilities available for online resources are able to ferret out keyword occurrences lost in card catalogs or print indexes.
The example above has little significance by itself. I realize that dialogs of the dead are around the 99th percentile in literary importance. What concerns me is that my experience challenges the general belief that Humanities research has longer validity than research in the sciences and even the social sciences. While interpretations can change, I believed that the basic “facts” about texts remained relatively constant. To give another example from my dissertation, the best information that I found in 1971 on Lucian’s dialogues of the dead was contained in a French critical work published in 1882. If the research in three literatures remains to be redone because Egilsrud didn’t uncover many of the examples of the genre because of the limitations of the card catalog, how many additional bibliographies and source documents might be questioned? Would more effective keyword searching of texts also change the conclusions found in pre-digital age research?
Literary criticism often builds upon the work of others. Researchers may not go back to the original sources to reexamine the evidence but rather accept prior studies as being factually accurate. To return again to my graduate school experience, one of my professors told us about the misconceptions concerning a famous work that had entered unquestioned into mainstream criticism because a prominent professor had made unjustified statements that showed that this critic had probably not even read the text. While sloppy scholarship can cause such errors, the best possible pre-digital research might have missed key data because of the limitations to scholarship based upon the card catalog and print resources.
Some of you may be thinking right now: Who cares? I concede some justice in this point. I worry that my doctor may not have the latest information about my complaint and whether the engineers used the right information in building the highway overpass that I use each day. Social science research had better be right if government officials use it to manage the economy or establish social policies. (As an aside, government officials often pay no attention to social science research if it contradicts political goals.) What are the real-world consequences if Francis Bacon wrote Shakespeare’s plays or whether or not Quebec has a post-colonial literature? A first-level response is that any research in any area should be accurate because the goal of the research process is discovering “truth.” On a second level, literary criticism can influence values by alerting readers to biases in literary works. Gone with the Wind has a political agenda, as do many works, and finding the right sources to prove this point may change the reader’s attitude about this novel. Finally, understanding and analyzing texts is a valuable skill for students to learn. A recent column on the Internet, for example, commented that business schools give admission preference for MBA programs to liberal arts degrees over undergraduate business majors because the liberal arts courses teach problem solving and analytical skills. Knowing which sources are needed to resolve a problem or come to a valid conclusion is an important skill that can be developed through examining well-done literary criticism, including the process of fact-checking.
My two final quick thoughts are that many Humanities scholars might find this sort of fact-checking opens up new research possibilities in the digital age as they re-examine earlier scholarship and that these errors may further lessen the importance of print resources.