If, until the middle of the nineties, the main preoccupation
of the scholarly editor had been that of conserving as
faithfully as possible the information contained in the original
sources, in the last ten years attention has shifted to the
user’s participation in the production of web content. This
development has occurred under pressure from the ‘native’
forms of digital communication as characterised by the term
‘Web 2.0’ (Tapscott and Williams, 2006). We are interested in
harnessing the collaborative power of these new environments
to create ‘fl uid, co-operative, distributed’ (Robinson 2007: 10)
digital versions of our textual cultural heritage. The approach
taken by Hypernietzsche (Barbera, 2007) similarly tries to
foster collaboration around cultural heritage texts. However,
while Hypernietzsche focuses on annotation, what we have
developed is a wiki-inspired environment for documents that
were previously considered too complex for this kind of
Before explaining how this can work in practice, it is worthwhile
to refl ect on why it is necessary to change our means of
editing cultural heritage texts. G.T. Tanselle has recently argued
(2006) that in the digital medium ‘we still have to confront the
same issues that editors have struggled with for twenty-fi ve
hundred years’. That may be true for analysis of the source
texts themselves, but digitisation does change fundamentally
both the objectives of editing and the function of the edition.
For the representation of a text ‘is conditioned by the modes
of its production and reproduction’ (Segre, 1981). The wellknown
editorial method of Lachmann, developed for classical
texts, and that of the New Bibliography for early printed books,
both assumed that the text was inherently corrupt and needed
correction into the single version of a print edition. With
the advent of the digital medium textual critics ‘discovered
that texts were more than simply correct or erroneous’
(Shillingsburg, 2006: 81). The possibility of representing multiple
versions or multiple markup perspectives has long been seen
as an enticing prospect of the digital medium, but attempts to
achieve this so far have led either to complexity that taxes the
limitations of markup (Renear, 1997: 121) or to an overload of
information and a ‘drowning by versions’ (Dalhstrom, 2000).
It is now generally recognised that written texts can contain
complexities and subtleties of structure that defeat the power
of markup alone to represent them (Buzzetti, 2002).
In our talk we would like to present three examples of how
overlapping structures can be effi ciently edited in a wiki: of
a modern genetic text in Italian, of a short classical text, the
‘Sybilline Gospel’, and of a short literary text marked up in
various ways. Our purpose is to demonstrate the fl exibility of
the tool and the generality of the underlying algorithm, which
is not designed for any one type of text or any one type of
editing. However, because of the limited space, we will only
describe here the Sibylline Gospel text. This is a particularly
interesting example, because it not only has a complex
manuscript tradition but it has also been deliberately altered
throughout its history like agenetic text.
The Sibylla Tiburtina is an apocalyptic prophecy that describes
the nine ages of world history up to the Last Judgement. The
fi rst version of this prophecy, which enjoyed a widespread and
lasting popularity throughout the whole medieval era, was
written in Greek in the second half of the 4th century. Of this
lost text we have an edited Byzantine version, dating from the
beginning of the 6th century, and several Latin versions, besides
translations in Old French and German, in oriental languages
(Syric, Arabic and Ethiopian) and in Slavic and Romanian.
The known Latin manuscripts number approximately 100,
ranging in date from the mid 11th century to the beginning
of the 16th. In the course of these centuries, with a particular
concentration between the 11th and 12th centuries, the text
was subjected to continuous revisions, in order to adapt it.
This is demonstrated both by the changing names of eastern
rulers mentioned in the ninth age, which involves the coming
of a Last Roman Emperor and of the Antichrist, and by the
introduction of more strictly theological aspects, especially in
the so-called ‘Sibylline Gospel’, that is an account of Christ’s
life presented by the Sibyl under the fourth age. No critical
edition of the Sibylla Tiburtina based on all its Latin versions
has yet been produced, although this is the only way to unravel
the genesis of the text and the history of its successive
reworkings. A classical type of critical edition, however, would
not be appropriate, nor would it make sense, to establish a
critical text in the traditional sense as one ‘cleaned of the
errors accumulated in the course of its history and presented
in a form most closely imagined to have been the original’.
With this kind of representation one would have on the one
hand the inconvenience of an unwieldy apparatus criticus and
on the other, the serious inconsistency of a ‘critical’ text that
never had any real existence.
To handle cases like this, we are well advanced in the
development of a multi-version document wiki application.
The multi-version document (or MVD) concept is a further
development of the variant graph model described at Digital
Humanities 2006 and elsewhere (Schmidt and Fiormonte,
2007). In a nutshell the MVD format stores the text as a
graph that accurately represents a single work in digital form,
however many versions or markup perspectives it may be
composed of. An MVD fi le is no mere aggregation of separate
fi les; it is a single digital entity, within which versions may be
effi ciently searched and compared. It consists of three parts: 1) The variant-graph consisting of a list of the differences
between the versions
2) A description of each of the versions, including a short
name or siglum, e.g. ‘L10’, a longer name e.g. ‘London, Brit.
Lib. Cotton Titus D.III, saec. XIII’ and a group name.
3) A list of groups. A group is a name for a group of versions
or other groups. For example, in the Sybilline Gospel
text there are three recensions, to which each of the
The wiki application consists of two JAVA packages (outlined
At the lower level the NMerge package implements all the
functionality of the MVD format: the searching, comparison,
retrieval and saving of versions. It can also export an MVD into
a readable XML form. The text of each version is recorded in
a simplifi ed form of TEI-XML, but the MVD format does not
rely on any form of markup, and is equally capable of handling
binary fi le formats.
Building on this package, the Phaidros web application provides
various forms for viewing and searching a multi-version
document: for comparing two versions, for viewing the text
alongside its manuscript facsimile, for reading a single version
or for examining the list of versions. The user may also edit and
save the XML content of each of these views. Since NMerge
handles all of the overlapping structures, the markup required
for each version can be very simple, as in a real wiki. In the
drawing below the differences between the two versions are
indicated by underlined or bold text.
This application is suitable for collaborative refi nement of
texts within a small group of researchers or by a wider public,
and attempts to extend the idea of the wiki to new classes of
text and to new classes of user.
Barbera, M. (2007) The HyperLearning Project: Towards a
Distributed and Semantically Structured e-research and
e-learning Platform. Literary and Linguistic Computing, 21(1),
Buzzetti, D. (2002) Digital Representation and the Text Model.
New Literary History 33(1), 61–88.
Dahlström, M. (2000) Drowning by Versions. Human IT 4.
Retrieved 16/11/07 from http://www.hb.se/bhs/ith/4-00/
Renear, A. (1997) Out of Praxis: Three (Meta) Theories of
Textuality. In Electronic Text, Sutherland, K. (Ed.) Clarendon
Press, Oxford, pp. 107–126.
Robinson, P. (2007) Electronic editions which we have made
and which we want to make. In A. Ciula and F. Stella (Eds),
Digital Philology and Medieval Texts, Pisa, Pacini, pp. 1-12.
Schmidt D. and Fiormonte, D. (2007) Multi-Version
Documents: a Digitisation Solution for Textual Cultural
Heritage Artefacts. In G. Bordoni (Ed.), Atti di AI*IA Workshop
for Cultural Heritage. 10th Congress of Italian Association for
Artifi cial Intelligence, Università di Roma Tor Vergata, Villa
Mondragone, 10 Sept., pp. 9-16.
Segre, C. (1981) Testo. In Enciclopedia Einaudi Vol. 14. Einaudi,
Torino, pp. 269-291.
Shillingsburg, P. (2006) From Gutenberg to Google. Cambridge
University Press, Cambridge.
Tanselle, G. (2006) Foreword in Burnard, L., O’Brien O’Keeffe,
K. and Unsworth, J. (Eds.) Electronic Textual Editing. Text
Encoding Initiative, New York and London.
Tapscott, D., and Williams, A. (2006) Wikinomics: How Mass
Collaboration Changes Everything. Portfolio, New York.
If this content appears in violation of your intellectual property rights, or you see errors or omissions, please reach out to Scott B. Weingart to discuss removing or amending the materials.