Experiences of the Previous Generation: the Thomas Middleton Edition

  1. 1. John Lavagnino

    King's College London, National University of Ireland, Galway (NUI Galway)

Work text
This plain text was ingested for the purpose of full-text search, not to preserve original formatting or readability. For the most complete copy, refer to the original conference program.

I am one of the general editors of a collected edition
of Thomas Middleton’s works, published in 2007—
more than ten years behind schedule. Of course, one of
the most common questions we heard during that period
was: why was it taking so long? (The other common
question was: why don’t you add more features and
more material?) This talk attempts to explain.
Of course, it may just have been too ambitious a schedule
to begin with; most comparable editions have taken
at least as long as ours did. Other people have found the
same task difficult: at least two substantial projects to
produce a collected Middleton were started during the
twentieth century and never finished; the last published
collection appeared in 1885–6. And, as with most largescale
projects, ours would have gone better with more
But we thought we had reasons to believe it would go
faster than it did. From the start, we did not intend to
pursue historical and bibliographical research as far as
possible: collecting the research of the last few decades
and making it available would have been enough, and
while there is much new work in the edition we did not
try to push as far as possible with every single play. We
collected a large group of editors so that few contributors
were working on more than one or two works. And, with
some exceptions that I’ll mention, this edition did not
intend to be particularly innovative in its digital methodology.
It’s true that it was always going to produce both
print and digital editions, not especially common in the
early 1990s; but when we began our idea of a digital edition
was of something you could feed to a concordance
program, not a full-scale online publication.
I think we did largely succeed in addressing delays deriving
from the traditional research. Any group of more
than seventy collaborators will have some whose lives
obstruct their work in some way, but most were able to
complete their portion on schedule and our other delays
made it possible to get everything else completed. We
were doing less than is common for comparable editions
today, as a comparison with Fotis Jannidis’s talk in this
panel will show: we only prepared and published an edited
text, and didn’t include diplomatic transcriptions of
the witnesses as well, something commonly felt to be
important for digital scholarly editions even in the early
We also did offer slightly more digital support for the
work of preparing the edition than is normal for printoriented
editions: I devised a system to allow editors
to prepare their notes using the line numbers of a draft
printing of the text; it then automatically adjusted them
to the right line numbers for the final printing, in which
textual corrections usually made all the line numbers
somewhat different. The normal approach has been for
a staff of student employees to make these adjustments
manually, but we didn’t have such a staff, and in a large
edition it is a large task (our final text has over 60,000
references by line number). The system was a good deal
of work to build, but it had unexpected side effects, since
a system to adjust line numbers in this way is also a system
to check line numbers and lemmata for correctness.
It was perhaps the only really innovative element of our
digital approach.
Because our edition offered only the edited text, we did
not run into the problem of relating it to the sources: so
we largely avoided one of the usual encoding problems.
This is the difficulty arising from the fact that a scholarly
edition has to do more than represent the works included
in an adequate way, possibly covering numerous genres.
It also needs to represent variation in texts in those
genres, something that is challenging with XML, and indeed
is often managed using non-XML systems, such as
version control software. Ideally, the system should be
able to relate the variation to the document structure, but
this is difficult with non-XML approaches.
But if the Middleton edition avoided some difficult problems,
it still faced the software problem. With pre-digital
scholarly editions, one of the biggest costs came from
typesetting; that problem is transferred to software and
its application in the digital era, whether the output is
print or digital. But this isn’t just a shifting of costs. In
a pattern visible in many areas of the digital humanities,
there is a shift: from a process that ends with a static
product, supported by established traditions for distribution
and preservation, to the creation of a machine for
generating what readers see. The digital edition is more
like a customized car than a book, and keeping it running
is a serious problem, as it requires continued attention and vigilance in obtaining replacement parts. Though the Middleton edition is entirely in TEI XML, it
has actually been published only on paper (so far), and
so we might think has avoided the software problem at
the expense of missing out on online publication. But to
preserve the possibility of updated editions and selective
reprints, the machinery that made the printed pages
must be preserved, and it is quite as complicated as a
system for online publication. Perhaps more, because
print makes some demands that online publication does
not raise: fitting as much as possible onto each page, for
example, in our case by printing the text in two columns
and the notes in three. Scholarly editions are recognized
as posing typesetting problems that are often quite complicated;
that doesn’t get easier because software is involved.
Once we’ve built this machinery for a single edition,
could it not then be reused for many? We would get more
of a return from our effort, and there would be more interested
parties to support the infrastructure. In part this
has been done: much of the underlying typesetting work,
to print a text with line numbers and notes keyed to those
line numbers, is done by a layer added on top of Donald
Knuth’s TeX typesetting system that Dominik Wujastyk
and I wrote, EDMAC, and this component has been used
not only by the two of us but by numerous other editions.
We don’t get support from those users, but of course just
being able to list them is a credential we can draw on in
seeking funding, and the possibility of a shared infrastructure
is there.
There are two major problems to be solved in this area,
though. One is the update problem: all software systems
either change or die, and the infrastructure we created
has not been changing. TeX is one of the most stable
systems there is, but even in the TeX world there have
been changes, mostly in font handling, that we haven’t
been keeping up with.
The bigger problem is that of the variety of texts. Even
within the Middleton edition there are many works that
required specialized handling. I don’t just mean ones
where we chose an editorial method different from that
used for the rest of the works: the occasional cases where
we print multiple versions, or the one case of a parallel
text. I mean cases where the work itself required unusual
typesetting because that’s what the work was: one short
theological work that used a six-column parallel layout,
for example, or one place where very long marginal
notes are suddenly necessary (in a passage not even by
Middleton: it’s by Ben Jonson, one of five contributors
to an elaborate pageant). A general framework that can
handle the whole variety observed in texts needs to be
very general indeed; and asking the author to rewrite the
work to make it more tractable is not an option.
We went part of the way towards this goal in EDMAC:
my field is English and Dominik’s is Sanskrit, and we
discovered that we shared almost no basic assumptions
about what an edition is, with the useful result that we
built something very generic. It is only dealing with a
few aspects of scholarly editions, though: the numbering
of lines and the creation of multiple series of notes keyed
to line numbers. Even in this restricted domain there are
practices we did not reproduce: such as one found in the
Monumenta Germaniae Historica series, of numbering
not only the lines of edited text on each page, but also
the lines in the footnotes. (Page images available at the
Monumenta Germaniae Historica digital web site illustrate
the phenomenon.)
In the world of digital curation, the software problem is
familiar, even if more attention has been given to more
fundamental problems about static data formats. (Hunter,
2006, is one work that does address software issues, and
the MLA’s guidelines for scholarly editions also suggest
attention to the problem.) All the more reason, then, why
a shared infrastructure is desirable, rather than one that
is unique to each edition; but the difficulties of creating a
generally-useful system are large, and much more work
is needed to overcome them.
Hunter, Jane (2006). Scientific Publication Packages—
A Selective Approach to the Communication and Archival
of Scientific Output, 1. First published 18 November
2006: http://www.ijdc.net/ijdc/article/view/8.
Knuth, Donald (1984). The TeXbook. Reading, Massachusetts:
Lavagnino, John, and Wujastyk, Dominik (1996).
Critical Edition Typesetting: The EDMAC format for
Plain TeX. San Francisco: TeX Users Group.
Modern Language Association, Guidelines for Editors
of Scholarly Editions, version of 25 September 2007.
cse_guidelines (accessed 14 November 2008).
Monumenta Germaniae Historica digital. http://www.
mgh.de/dmgh/ (accessed 14 November 2008).
Taylor, Gary, and Lavagnino, John, general editors
(2007). Thomas Middleton: The Collected Works and
Thomas Middleton and Early Modern Textual Culture: A
Companion to The Collected Works. Oxford: Clarendon

Conference Info


ADHO - 2009

Hosted at University of Maryland, College Park

College Park, Maryland, United States

June 20, 2009 - June 25, 2009

176 works by 303 authors indexed

Series: ADHO (4)

Organizers: ADHO

  • Keywords: None
  • Language: English
  • Topics: None