The TEI as Luminol: Forensic Philology in a Digital Age

A signifi cant number of digital editions have been published
in recent years, and many of these serve as exemplars for
those of us working within the digital editing community. A
glance at Electronic Text Editing, for example, published in 2006,
indicates that such projects address a wide berth of editorial
problems, ranging from transcriptional practices to document
management to authenticity. And they offer a wealth of
information to editors working on various types of digital
editions.
In this paper, I consider an editorial problem that has yet to be
resolved. My discussion centers on the diffi culties that arise
in editing a single, albeit somewhat unusual, Icelandic saga:
Hafgeirs saga Flateyings. This saga is preserved in an unsigned,
eighteenth-century manuscript, Additamenta 6, folio (Add. 6,
fol.). Today housed in the collection of the Arni Magnusson
Institute for Icelandic Studies in Reykjavík, Iceland, the
manuscript was originally held as part of the Arnamagnæan
Collection in Copenhagen, Denmark. According to the fl yleaf:
“Saga af Hafgeyre fl ateying udskreven af en Membran der
kommen er fra Island 1774 in 4to exarata Seculo xij” (Hafgeirs
saga Flayetings was copied from a twelfth-century manuscript
written in quarto, which came [to Copenhagen] from Iceland in
1774).
While such a manuscript might appear unremarkable, as a
number of paper manuscripts were copied during the late
eighteenth century in Copenhagen, then the capital of Iceland
and the seat of Icelandic manuscript transmission during this
period, only twelve Old Norse/Old Icelandic manuscripts of
those catalogued in the Copenhagen collections are dated to
the twelfth century, while a mere eighteen are dated to 1200
(Kalund 512). The dating on the fl yleaf is therefore unusual, and
as it turns out, suspect as well, since no catalog entry exists
to record the existence of the alleged source manuscript.
Moreover, according to Jorgensen, the motif sequences found in
Hafgeirs saga bear a striking resemblance to those found in the
well-known mythical-heroic saga Hálfdanars saga Brönufóstra
(157). And in a fascinating argument based primarily on this
fact, Jorgensen argues that Add. 6, fol. is a forgery, claiming that
Þorlákur Magnússon Ísfi ord, an Icelandic student studying and
working in Copenhagen during the 1780s, composed and sold
Hafgeirs saga as a copy of an authentic medieval Icelandic saga
(163).
In spite of its questionable origin, Hafgeirs saga stands as a
remnant of Iceland’s literary, linguistic, and textual history, and
Add. 6, fol. can therefore be viewed as an important cultural
artefact. As the editor of the Hafgeirs saga manuscript, my
aim is to provide a ‘reliable’ (see “Guidelines for Editors of
Scholarly Editions” Section 1.1) electronic edition of the text
and the manuscript. But the question, at least until recently,
was how? What is the best way to represent such a text?
Encoding the work according to a markup standard such
as the TEI Guidelines is surely a starting point, but doing so
doesn’t solve one of the primary concerns: How to represent
the manuscript reliably (which presents a complex editorial
problem of its own), while at the same time illuminating the
textual and linguistic ‘artefacts’ that may offer readers insight
into the saga’s origin?
At the 2007 Digital Humanities Summer Institute, Matthew
Driscoll gave a talk entitled “Everything But the Smell: Toward
a More Artefactual Digital Philology.” The talk provided a brief
history of the shift toward ‘new’ philology, and, importantly,
underscored the signifi cance of the material or ‘artefactual’
aspect of new philology, which views manuscripts as physical
objects and thus as cultural artefacts which offer insight into
the “process to which they are witness” (“Everything But the
Smell: Toward a More Artefactual Digital Philology”). The TEI,
Driscoll pointed out, offers support for artefactual philology, and
the descriptive framework of the P5 Guidelines, which defi nes
Extensible Markup Language (XML) as the underlying encoding
language, is ideal for this kind of work. Moreover, as Driscoll
suggests, there is “no limit to the information one can add to
a text - apart, that is, from the limits of our own imagination”
(“Electronic Textual Editing: Levels of Transcription”). To be
sure, this paper does not lapse into what McCarty refers to
as the ‘complete encoding fallacy’ or the ‘mimetic fallacy’ (see
Dahlström 24), but it does agree with Driscoll in arguing that
P5-conformant editions, which can offer a signifi cant layering
of data and metadata, have the potential to move the reader
beyond the aesthetics of sensory experience.
By defi nition, artefactual philology portends a kind of
‘evidentiary’ approach, one that can frame textual features,
including linguistic and non-linguistic features (such as lacunae)
for example, as kinds of evidence. Evidence of what? That is
in the hands of the editors and the readers, but conceivably:
linguistic development, the transmission process, literary merit,
and so on. And when an evidentiary approach to philology is
defi ned within a ‘generative’ approach to a scholarly edition
(see Vanhoutte’s “Generating” 164), a new direction in
electronic editing becomes possible.
This paper explores this new direction. It shows how the
Hafgeirs saga edition employs such a framework to address
the problem of describing linguistic and non-linguistic artefacts,
which are precisely the kinds of evidence that can bear witness
to the composition date and transmission process. And it
demonstrates how the display decisions result in an interactive
and dynamic experience for the edition’s readers.
For example, in addition to elements defi ned within existing
modules of the TEI, the edition’s schema defi nes four new
elements1. Drawn from the perspectives of historical and
socio- linguistics, these elements are intended to aid readers
in evaluating the saga’s composition date. Given the ‘logic of abundance’ (see Flanders 135), encoding the metadata
described in the new elements beside the descriptive data
described by pre-defi ned elements (such as <del> for example)
can be accomplished without sacrifi cing the role of the text as
a literary and cultural artefact. Because the transformation of
the source XML has been designed to display interactively the
various encodings of the text2, readers can view or suppress the
various descriptions and generate their own novel versions of
the text. Readers can display archaisms, for instance, and assess
whether they are “affectation[s] of spurious age” (Einar 39) or
features consistent with the textual transmission process, and
they can view borrowings, for instance, and assess whether
they preclude a medieval origin or are to be expected in a
text ostensibly copied by a scribe living in Copenhagen during
the eighteenth century. Or they can suppress these features
and view the normalized transcription, the semi-diplomatic
transcription, emendations, editorial comments, or any
combination of these.
Ultimately, this paper synthesizes aspects of text editing,
philology, and linguistics to explore a new direction in digital
editing. In doing so, it frames P5 XML as a kind of luminol that,
when transformed, can be used to illuminate new types of
evidence: linguistic and philological data. And the goal is identical
to that of a crime-scene investigator’s: Not necessarily to solve
the case, but to preserve and to present the evidence.
Notes
[1] The <borrowing> element describes a non-native word which
has been adopted into the language. Distinct from the <foreign>
element, a borrowing may have the following attributes: @sourcelang
(source language), @borrowdate (date of borrowing), @borrowtype
(type of borrowing; e.g. calque, loanword). The <modernism>
element describes a word, phrase, usage, or peculiarity of style
which represents an innovative or distinctively modern feature. The
<neologism> element describes a word or phrase which is new to
the language or one which has been recently coined. The <archaism>
element describes an archaic morphological, phonological, or syntactic
feature or an archaic word, phrase, expression, etc.
[2]My discussion of the display will continue my in-progress work
with Garrick Bodine, most recently presented at TEI@20 on
November 1, 2007: “From XML to XSL, jQuery, and the Display of
TEI Documents.”
References
Burnard, Lou, and Katherine O’Brien O’Keefe and John
Unsworth eds. Electronic Text Editing. New York: The Modern
Language Association of America, 2006.
Dalhström, Mats. “How Reproductive is a Scholarly Edition?”
Literary and Linguistic Computing. 19:1 (2004): 17-33.
Driscoll, M.J. “Everything But the Smell: Toward a More
Artefactual Digital Philology.” Digital Humanities Summer
Institute. 2007.
Driscoll, M.J. Text Encoding Initiative Consortium. 15 August
2007. “Electronic Textual Editing: Levels of Transcription.”
<http://www.tei-c.org/Activities/ETE/Preview/driscoll.xml>.
Einar Ól. Sveinsson. Dating the Icelandic Sagas. London: Viking
Society for Northern Research, 1958.
Flanders, Julia. “Gender and the Eectronic Text.” Electronic
Text. Ed. Kathryn Sutherland. Clarendon Press: Oxford, 1997.
127-143.
“Guidelines for Editors of Scholarly Editions.” Modern
Language Association. 25 Sept. 2007. 10 Nov. 2007. <http://
www.mla.org/cse_guidelines>.
Jorgensen, Peter. “Hafgeirs saga Flateyings: An Eighteenth-
Century Forgery.” Journal of English and Germanic Philology.
LXXVI (1977): 155-164.
Kalund, Kristian. Katalog over de oldnorsk-islandske handskrifter
i det store kongelige bibliotek og i universitetsbiblioteket.
Kobenhavn: 1900.
Vanhoutte, Edward. “Traditional Editorial Standards and the
Digital Edition.” Learned Love: Proceedings from the Emblem
Project Utrecht Conference on Dutch Love Emblems and the
Internet (November 2006). Eds. Els Stronks and Peter Boot.
DANS Symposium Publications 1: The Hague, 2007. 157-174.