Bibliography and other metaphysical questions
So the next Digital Edition meeting finally took place today. A nice thing happened in the first 15 minutes: Alex was asked by Sabine and Johanna to explain XSLT and XQuery. I am always amazed not only at how much technique Alex knows, but also at his ability of making us understand how these things work (not in detail of course, but the general idea). Just as I am amazed at the complexity of the debates we can have during these sessions. Thankfully, we are through arguing about Geminationsstriche. We found a nice replacement for this formerly everlasting discussion though: the bibliography.
As it is, primary sources (first editions and other prints we use for comparison purposes) are to be listed in the header and referred to in our editorial notes if need be (depending on the specifics of the corpus). But the question we had today concerned primary and secondary literature. The first suggestion was to have a general bibliography for each corpus, on the corpus’ presentation page. I was against it for several reasons (maybe I am wrong on that, though), the main one being that if you provide a bibliography, you have to do it in an exhaustive way and we would not realistically manage to do so for each corpus (or to do it in a harmonious way for all corpora). The second suggestion was that we have for each document (in the very convenient < additional > element in the header) a bibliography that would contain first the primary, then the secondary sources. Again, this was not very satisfactory, on the one hand because of the obvious repetitions that would occur between the different documents of one corpus, on the other hand, because there are books or papers that are of major importance to conceive the commentary of a document, while some are only locally of some help (for situating an event for instance). Those two types of references should not be given the same importance, but it would take us too far to conceive a hierarchical systematic in the secondary literature (all of this, again, in the improbable < additional > element). I was surprised that this turned out to be an important issue arising with the edition of the letters from Chamisso to de La Foye (the corpus we are editing in our DH class of this summer). For the Beausobre letters as well as for the Tieck letters, I had given in my editor’s notes all the bibliographical information which I considered relevant. In fact, there is some information I did not quote systematically my references on, because I did not see the point in quoting a common encyclopedia or dictionary, the source of a simple Latin translation or, concerning the many commentaries that have to do with Solger, my own publications of the last 10 years. The more I think about it after this afternoon’s work session, the more I am starting to consider that we should, after all, provide for a corpus bibliography of secondary works, but more in the form of “The main works we have been referring to for this corpus are… ” followed by no more than 5 to 10 titles.
This has to do with how serious we want to look. I have been looking up the Alfred Escher letters and the Alexander von Humboldt letters and I cannot help but wonder at the differences with our own edition. While we aim at not overloading the page and I even hesitate to add a 10 title-bibliography for each corpus, these editions are not afraid of delivering long text passages and thorough bibliographies. I have been annoying my whole team for weeks with my obsession for a NOT white background for the text, because I find it too tiresome for the reading eye, but this is what all those sites go for (also Carl Maria von Weber, even the all-time best Van Gogh letters). Are we doing something so different?
We necessarily are, because we are a small group with limited time and money: we will never reach or even intend the thoroughness of a complete edition. We select representative corpora and the job of our edition is to empower them with this representativeness and show it. I am not saying that we are deliberately doing bad scholarly work. I only mean that we take decisions (in the rendering of the transcription, in the tagging, in the layout) that have to do, for instance, with the tools we want to develop (for the search, for the collation) and that these are still only in germination phase. Many decisions were taken on a small corpus with the aim of being implemented on a somewhat larger one and will need correction at some point. This certainly is an uneasy position, but it is what it is: a work in progress that will stay so for a while…