Introduction

1Many significant works of the Middle Ages are still available only in manuscript form. To read these texts, one must be a skilled Latinist and palaeographer, negotiate with a library to gain access to manuscripts, and have the leisure to work through various scribal errors. They almost entirely inaccessible.

2This problem is evident in the case of the commentaries on Peter Lombard’s Sentences, which organizes biblical and patristic sententiae (opinions) into a systematic overview of key debates in theology, becoming the prevailing textbook on the subject in Western Europe after its creation in the middle of the 12th century. From the 13th century until the end of the Middle Ages, it was a standard requirement in formal theological study to comment on this book. Sentences commentaries survive from many of the most influential philosophers and theologians of the late medieval period, such as Alexander of Hales, Thomas Aquinas, and Martin Luther; and there are many more from figures who are much less known. By studying these texts in the context of the others written, one can essentially map the history of Western thought on any of the subjects they address (as shown, for example, by Rosemann 2007). Every commentary forms one piece of an enormous network, and if we could see and link every part, we would have a much clearer idea of the origin and development of philosophical, theological, and social ideas in this period.

3This is practical on a large scale only if there is already an edition of the text, and the slow process of publishing texts under the traditional print model inhibits the study of works beyond a limited canon. This has not favoured authors such as Peter Plaoul, a major figure of the University of Paris in the late 14th century who left little written output, drawing limited attention from modern scholars (see Millet 2009). The edition of Peter’s lectures on the Sentences by Jeffrey C. Witt combats the author’s obscurity and takes on a work that would be intolerably expensive to print, since the complexity of its manuscript tradition prevents straightforward resolution into a single collated text. What we need instead is a series of documentary editions, and the site is suitably labelled editiones, in the plural. Witt makes the case that we can only meet the future needs for learning about the tradition of Sentences commentaries by taking advantage of the tools for electronic publishing available to us (Witt 2015), and this is the first serious project to consider out how to do this in practice.

4The site is also a brilliant application of the principles of ‘progressive publication’ to medieval texts to make Witt’s work public while offering solutions to the scholarly concerns that made this methodology untenable in a printed format, providing a readable draft of the text that he has emended over time. The site has been online since 2011, a time when there were few models for undertaking and presenting such work. Witt’s edition is an inspiration for anyone unconvinced of the benefits of presenting a medieval text online.

‘Progressive Publication’

5Lists of medieval authors and their works (such as Sharpe 1997) are littered with announcements of editions that are nowhere to be found, the instigator having retired, died, or simply lost interest. A later scholar who takes up the project must repeat years of labour. Electronic publishing supplies the means to avoid this situation. Witt calls his model for publicly editing Peter’s unpublished work ‘progressive publication’ (Witt 2013). This idea has already gained traction in the sciences, where it is also called ‘continuous publishing’, but is only slowly making its way to the humanities. It is hardly a radical notion as applied to scholarly editions, as it brings to fruition long-held aspirations of textual critics. The Oxford edition of the Vulgate New Testament, though it began in 1877 and was not finished until 1954, produced almost immediate results by publishing versions of some of its most important manuscripts in a separate series, ‘Old-Latin Biblical Texts’, producing seven volumes between 1883 and 1923 (see Wordsworth 1883). The creators of the Corpus Christianorum originally proposed to publish editions of their texts ‘in rapid succession’ and small print runs to allow the series to stay abreast of the latest research ([Dekkers] 1948, 413; cf. Leemans 2003, 17). With the Internet, this vision can finally be realized, allowing editors to make their initial work public without their inevitable oversights haunting them forever.

6The many benefits of this approach likely make its use seem obvious to outsiders, but there is little precedent in the humanities for making preliminary work public, while many scientific disciplines have practised this for decades. We need to be both more forgiving of editors and more vigilant in how we use manuscript sources if we wish the model to gain acceptance in the field of scholarly editing. Especially when one is dealing with unpublished texts, it takes little effort to produce a useful transcription, but a great deal more to produce a printable edition. Publishing in-progress work entails a certain amount of bravery: one must be prepared to make embarrassing mistakes, of the sort that might prompt a reviewer to question the editor’s linguistic capabilities. Because of the potential for misunderstanding the process, Witt has positioned the site as it stands as a ‘finding aid’ meant to precede a full critical edition (Witt 2013). He makes public only those sections that he feels have reached a reasonable level of accuracy; readers can access the rest of the text only by writing the Witt, who issues individual accounts. This approach allows the text to reach those truly interested in the work, though it is rather labour-intensive and adds complexity to the software. It is an interesting criticism of the general trend favouring processes that allow scholarship without human contact, and a helpful middle ground between the idealist’s approach of placing everything in the open and the typical strategy of making one’s work entirely secret. Nonetheless, one hopes that such precautions will soon be unnecessary; it is more user-friendly to warn readers of the text’s status clearly, as used by projects such as the Shelley-Godwin Archive.

7If anyone is reluctant to cite such an edition, it is worth remembering that our printed editions are far less reliable than we like to think. For example, the influential commentary on the Pauline epistles by Peter Lombard is typically cited from the Patrologia Latina; it is rarely noted that this is an error-prone reprint of a thinly documented edition from the 16th century (Peter Lombard 1535). It is available everywhere online, and is far more accessible to the typical reader than even the most accurate manuscript; but like most early-modern editions, its relationship to its sources is entirely opaque. Even an imperfect working edition is a vast improvement over such texts if it has a documented manuscript basis and follows modern techniques – and, crucially, is widely available. Scholars continue to make assumptions based on the Patrologia Latina and students suffer through its inaccurate texts every day – sometimes even where a more recent critical edition exists, because that edition is only available in a handful of libraries due to its ruinous cost.

The Sentences Commentary Text Archive

8Making texts available more widely available, sooner, is only part of the reason to make a digital edition: another is that print editions simply do not allow us to answer the questions we ask of them. The edition of Peter’s work is only the beginning of an ambitious plan for a Sentences Commentary Text Archive, which will eventually allow precisely the type of comparative study that Sentences commentaries need. As a basis for this, Witt aims to have Sentences commentaries digitized to a common standard. In this initial foray, he took on two projects at once to make Peter’s work available: the editing of the text itself, and the creation of an open-source system used to publish it online, LombardPress, which he has also used for a collaborative edition of Adam Wodeham’s commentary on the Sentences (Hallamma et al. 2011–13).

9Given the circumstances of a nascent field, the mere fact that the edition has appeared at all in a usable form is something of an achievement. There are few people who unite the skills of a philosopher, historian, Latinist, palaeographer, and computer scientist: Witt is evidently one of them. His ideas address the problems faced by all who wish to create digital editions of texts written before the modern era. We have no standards for the publication of medieval works, or even expectations for what online editions should record from a manuscript. Those who set about to begin a digital edition of a work typically find that while it is relatively straightforward to mark up a text using the Text Encoding Initiative (TEI) recommendations, displaying the resulting files in a readable format is difficult. Because TEI-based projects often develop their own guidelines (in choosing, for example, among several competing choices for marking abbreviations), they tend not to be compatible with one another, making it difficult to reuse work from other scholars. The fact that software to support editions must at present be written almost from scratch for each project is one of the reasons why digital editions have not become more common. Indeed, it is unlikely that digital editions will begin to gain traction until a publisher of extended editions exists that supports the use of TEI and manages a stringent peer-review process, as the forthcoming Digital Latin Library intends to do.

10Witt has made a step towards solving these problems in creating a modular system, LombardPress, that he intends to be reusable for other Sentences commentaries, and has made it available under the GNU General Public Licence. Its documentation is still under development, meaning that it is not yet feasible to reapply it to new projects. Further, it relies on software that is complex to maintain over the long term (PHP and MySQL). Nonetheless, few of the problems he seeks to solve are specific to Sentences commentaries, and there is much that his work could contribute to communities aiming to create a common standard for editing premodern documents (such as EpiDoc, which has expanded beyond its original scope of epigraphy). What we now need is a decentralized editing platform developing Witt’s model that can be applied to any premodern text and is based on the most simple technology available, in combination with centralized repositories where sources can be deposited and combined with other works for research and reading.

Editorial Method

11Witt presented the original version of the site as a ‘social edition’ (Witt 2011; inspired by Siemens et al. 2012). The term has not aged well, and Witt now omits any mention of it, but it is still a visible influence. One might argue that the principles of social editing are merely those of making the best use of digital media to improve scholarship and make it more widely known. Witt now frames it in this way:

This site aims to be a stepping stone towards an eventual critical edition of the Sentences commentary of Peter Plaoul. As I have worked on this text, learning more and more about it, both about its length and complexity, it is clear that a supremely polished critical edition of the text is not a one-person job. If it is ever to be completed it will require a team of interested people. But there is a dilemma here. How can people ever become interested in the text if they do not know about its existence or the topics contained within it?(Witt 2013)

12It is not possible to tell how many contributions Witt has received. The comments function is streamlined, helpfully noting the version of the text to which they were referring, but it is rather dull to comb through lists of minor spelling and grammar corrections looking for potentially insightful remarks, and it would be better if one could make desired changes to the text directly. The Wikipedia model of allowing immediate modification is probably not the ideal of most editors; more appropriate to a critical edition directed by a particular editor is the idea of the ‘pull request’ popularized by the Git source control management software, which allows anyone to change their own copy of a file to read precisely as they think it should, then create a proposal explaining their rationale (which the maintainer of the original merges into the main version). Both of these models make it clear who has contributed to a particular file, which might motivate more readers to participate.

13 Also implied in Witt’s model is the ideal of providing the reader with the same resources as the editor. Registered users are able to view images corresponding to the text of the manuscript, including a feature that allows one to compare a transcription (helpfully displayed with the original line breaks) against a photograph. This is what allows the responsible publication of a preliminary text: if one can view the sources, it is easy to confirm the editor’s reading, eliminating the fear that one might be risking one’s reputation by citing a working text.

14Peter’s Sentences commentary is a series of lectures, and some of the surviving witnesses are student notes, varying considerably in places (Glorieux 1939). The edition has so far taken four of seven manuscripts into account, with each source transcribed in full. It is not clear if all witnesses will be included in the final version; an explanation of the editorial method, manuscript selection criteria, and codicological descriptions have yet to be added.

15While Witt does not explain his method in these terms, he is essentially creating a documentary edition of each of these texts, of a sort advocated by others (e.g. Pierazzo 2014; van Dalen-Oskam 2015), with an edited text provided (at least notionally) in parallel. It enforces, further, the idea that every historical version of a text is the edition of its scribe(s). The manuscripts’ original capitalization and punctuation is retained (not always consistently), which can be of great service in editing a text for the first time; editors have made many blunders because they did not take the time to understand scribal punctuation on its own terms. The editions could be made more useful by marking where spelling has been modernized (for example, e is silently expanded on many occasions to ae) and abbreviation are expanded, allowing their use for linguistic research, though it must be admitted that the standard TEI guidelines for these features are too verbose to be usable by editors of heavily abbreviated premodern texts. The approach of creating a critical edition founded on a collection of documentary editions is likely to be the most useful to future scholars.

16 This idea of editiones should be taken further. The focus of the site is still an eclectic reader’s version of the commentary with a traditional apparatus, and we need to reconsider whether these truly help us to answer our research questions. Witt is here doubtless thinking of the demands of readers who simply want to look up a text and move on, but this could be done with much less work. Such readers are rarely interested in textual variants, and simply want something they can understand quickly; it would be less laborious to correct the best witness and display this in the default view (using TEI, this could be displayed in either a diplomatic or normalized form) than to create an entirely separate critical text. There are already many competing methods of punctuating modern editions of medieval Latin, and anyone with enough experience to read Peter Plaoul directly could also work through a corrected version of the medieval punctuation. If one is truly concerned about making the text broadly accessible, this can be achieved even more effectively with a vernacular translation. A critical apparatus, in the meantime, can never meet the needs of every reader interested in textual variants; its function is properly to explain the editor’s rationale for reconstructing the text, and this is better achieved with a commentary. Its format has never promoted usability, since it is defined entirely by the space constraints imposed by print, and it has not improved with a transition to a digital format.

17Witt has the germ of something much better in a tool that graphically shows the variants between any two sources, offering the edited text in parallel with the sources. Where we simply need to show the differences between witnesses, this can be done more usably and accurately with visual tools. The version of the comparison tool available at this writing only allows for comparison of two witnesses, and is not clever enough to ignore variants traditionally deemed insignificant (as of spelling, punctuation, and capitalization). Once these problems are solved (towards which Juxta and CollateX have already made significant progress), this will eventually represent a far more accessible alternative to the critical apparatus normally printed at the bottom of the page.

Presentation

18 It is difficult to go from an encoded text to an edition that a non-technical user can sit down and read. Many editions become wrapped up in what is technically possible rather than focusing on what usable, ending up as a mere technology demonstration. Witt does not, however, try to impress readers with his technical wizardry; one would never know without investigation, for example, that behind the timeline accompanying his engaging biography of Peter is a rigorously structured TEI file.

19 The only part of the website that reveals the underlying markup to the reader is a feature that displays it under individual paragraph menus; the full files are gradually being made available for download as they become more complete. The edition is not, however, at the point of being easier to read than a typical printed edition. Few digital editions produced to date have placed an emphasis on readability, with correspondingly low typographical standards. One quickly becomes lost in the text of the main reading screen, and the ‘Paragraph Menu’ links create unnecessary clutter.

20This situation could be easily improved by increasing the type size and line spacing, reducing the line length, using typographer’s quotation marks, and distinguishing the menu link more clearly from the text (by moving it into to the right margin, for example). More still could be done by taking advantage of responsive design techniques that now make it a relatively straightforward task to make a single website highly usable on desktops, tablets, and phones. This is key for allowing a text to be usable in a classroom setting, while commuting, or as a reference tool in heated conference debates. Yet reading from paper still carries many benefits that screens cannot achieve. As part of LombardPress, Witt has also written a program transforming his editions into a printable format (using the advanced typesetting capabilities of LaTeX). The focus of creating a digital edition of a text should be to publish it in a technology-independent format, meaning that it is not tied to either paper or a particular encoding scheme, and Witt shows that this can achieved with minimal resources.

21 The edition truly excels, even over a printed book, in its finding aids. An outline accompanies each section of the text, allowing one to quickly see the flow of an argument, as well as allowing comparison between disparate witnesses (Witt 2014). There is also a simple search interface, amplified by lists of names, titles of works, subjects, quotations, and unresolved textual issues (a particularly useful innovation). Even many completed editions do not have indices as detailed as these.

Archiving and Licensing

22Computer interfaces are passing: it seems unlikely that much of the software supporting the website will still be functional in another decade, and even if it is, it will need to be updated to meet the expectations of future users. For any digital edition, it is critical to archive one’s findings in an open and well-documented format, as well as deposit them in a repository where other scholars might find them. Editors hope that we have a solution to the first problem in the TEI guidelines, which is highly structured, and helps editors to achieve independence from technology through encouraging to record the meaning of their decisions, while most other software focuses on visual appearance. While the latter approach represents how we have been trained to encode textual meaning over centuries, it is much more subjective and less useful, especially when applying texts to new research questions or making writing accessible to users with print disabilities.

23To fill the need for a reasonably stable referencing system, the paragraphs of each section are numbered. This comes with the problem that paragraph structure often changes as editions progress, and to mitigate this Witt suggests citing the version of the text along with this, which is easy to understand and cite, and the site gives a sample citation for each section of the text. The lack of persistent identifiers is a more serious concern. Even the URI scheme specifically reflects on the site’s PHP coding, with addresses such as petrusplaoul.org/text/textdisplay.php?fs=lectio1. Witt could simplify links and make them more stable by rewriting this as petrusplaoul.org/text/lectio1. It might be simpler to encourage the use of the folio and line numbers of a significant manuscript.

24The text is provided under a Creative Commons Attribution-NonCommercial-NoDerivatives (CC BY-NC-ND) licence. This implies that others are prohibited from producing their own version of the text, against the traditional practice among editors of basing their work on that of others as much as possible; West notes that many editors in his day, to prevent error, sent their editions to printers by modifying a copy of an older edition using proofing marks rather than produce an entirely new typescript (West 1973, 101–2). In any event, the restriction has limited force in many jurisdictions as applied to scholarly editions. The text of a critical edition is not a copyrightable work in some parts of the world, meaning that a critically edited version of a public-domain work is itself in the public domain (Margoni and Perry 2011). In the European Union, scholarly editions have copyright protection for no more than thirty years (even less as implemented in most countries). The Creative Commons Attribution-ShareAlike (CC BY-SA) licence is more reflective of the historical practice of editing; its use formalizes scholarly expectations and ensures that editors have the same rights around the world. Under this licence, other users must credit past editors and make their improved version of the text available under the same terms. This is just as effective in preventing potential exploitation as restricting it to non-commercial use, and can only serve to promote the original editor’s work.

Conclusion

25If Witt had decided to follow the safe and entrenched path of producing an essentially finished text before publishing it in a printed book, few of us would have heard of Peter Plaoul, and we would not be able to read his work on the Sentences with ease for years to come. Even then, it would have been tied to the particular technology of paper, inhibiting its accessibility and limiting its usefulness for research and teaching. Instead, we have an in-progress edition whose relationship to its manuscripts is far more clearly documented than a typical printed edition, and as it develops, it can benefit from the eyes of many readers, rather than having us furtively pencilling corrections into library copies or noting them in reviews. Through his work, Witt is gradually removing the many barriers to the text, bringing it only a translation away from being entirely accessible to a moderately educated audience. This is how we should publish medieval books.

26Witt’s edition felt like a vision of the future when it first appeared in 2011. He has overcome the many layers of technical obscurantism necessary to publish a scholarly edition online, and has helped to establish a model that begins to practise the theories editors have discussed for decades but have never had the means to fulfil. Having succeeded in these things, one desires to be able to read more about the thinking and process behind his work, and especially to know more about what he has discovered in Peter Plaoul. For now, Witt has already brought to light the intellectual side of a remarkable historical figure, giving us another glimpse into university teaching in the 14th century and contributing another piece of the puzzle that is the history of Sentences commentaries.