Facilitating Text Analysis in Russian Culture: TEI or Topic Maps?

Facilitating Text Analysis in Russian Culture: TEI or
Topic Maps?

Miranda
Remnek

University of Minnesota
m-remn@tc.umn.edu

2002

University of Tübingen

Tübingen

ALLC/ACH 2002

editor

Harald
Fuchs

encoder

Sara
A.
Schmidt

Still under development, the University of Minnesota's Early 19th Century Russian
Readership & Culture Project () is one of relatively few
comprehensive digital archives for the study of Imperial Russian history. Based
on materials gathered over two decades of research that culminated in a
dissertation on The Expansion of Russian Reading Audiences,
1828-1848, the project is distinctive in a number of ways. First, it
pulls together a variety of primary research materials into a single archive.
(To some extent this is not uncommon, but the ENCRRC archive includes not only
texts, images, and scholarly reference materials, but also a large statistical
database of 12,000 subscription records which provide much additional research
data). Second, the search mechanism has been customized to provide simultaneous
access to both English and Cyrillic TEI-encoded texts without the need for a
Cyrillic keyboard--and this in itself is a technical feat.
But the project's most distinctive feature is its use of SGML
<interp> tags to enrich the texts by encapsulating a number of
pre-selected analytical categories. This means that users of the archive are
given not one but two modes of access to the content of the four main text
groups (fiction, journals, memoirs and travel accounts). How is this achieved?
First, researchers are presented with a comprehensive full-text search option,
since the project uses Enigma Corporation's powerful SGML-based DynaWeb software
to deliver the texts to the web (). Second, the project's adoption of SGML-based analytical tagging means that
users have an additional avenue of access to thematic categories. To enable this
approach the interface presents a roster of 10 main categories of analysis, each
divided into small groups of differing sizes for a total of 60 subcategories.
Entries are scripted so that the software can retrieve and display passages that
contain superimposed SGML ID references even though the texts are converted to
HTML for delivery over the web. Space limitations prevent enumeration of all the
subcategories, but a listing of the main themes will give some idea of the
research potential involved: Publishing, Print Categories, Novels, Journals,
Newspapers, Booktrade, Text Access (Bookstores, Coffeehouses, etc), Reading
Publics, Social Groups, and Job Titles.
The provision of these categories works well for researchers in differing fields
of Russian culture. A literary scholar may use the archive to trace references
in the various groups of texts to the distribution of original Russian novels
versus translations of foreign compositions; he or she may then enhance these
findings by searching the database of subscription records from the period
1825-1846, and reviewing biographic and geographic data about the subscribers
connected with both native and translated novels. A history scholar, on the
other hand, may use the materials to trace levels of access to print materials
among less privileged groups (such as lower-level bueaucrats and merchants) not
normally considered part of the contemporary cultural milieu. A women's studies
scholar may use the archive to piece together hard-to-find references to women's
reading, and the mechanisms women used to gain access to texts in a clearly
defined patriarchal society. It should therefore be clear that employment of a
relatively simple SGML-based analysis option has enabled substantial enrichment
of a carefully-selected core of texts such that they are immediately serviceable
for multiple purposes.
In addition to the project's value for different fields of study, another
important benefit resides, as noted, in its presentation of a rich variety of
sources that include encoded images and historical records as well as primary
texts. All these resources are related, moreover, by topic. But the linkage
between them is not always straightforward, and so it has seemed important to
take note of new concepts like SGML Topic Maps (ISO standard 13250, Geneva,
December 2000)--which promises to facilitate the linkage of similar elements in
different types of research data. As Christian Wittern has suggested ("TEI and
Topic Maps," ACH/ALLC 2001), topic maps provide an
"architecture for the semantic structuring of information networks…[that] has
the potential to provide a bridge between… texts encoded with schemes like the
TEI [and] other information resources." More explicitly, Bill Trippe noted
recently in an article published in EContent (August
2001, v. 24, issue 6, p. 45ff), "For proponents, topic maps are the ideal
solution for helping users find information about a topic across a variety of
documents."
Wittern also suggests that topic maps enable an archive to present abstract as
well as concrete representations of knowledge. Until recently the categories
used in the ENCRRC archive were more exclusively objective than the mix of
categories used in our better-known sister project, Women's
Travel Writing, 1830-1930 (), and as such, have been
easier to apply. But WTW's two subjective categories--gender and ethnicity--are
strongly championed by faculty advisers to the project as essential material for
the pursuit of current scholarly trends in women's studies (though certainly
more challenging to the encoder). It thus seemed advisable to problematize more
fully the analytical tools supplied for the ENCRRC archive.
With this in mind, we are redesigning and testing out certain revamped portions
of the ENCRRC archive. Drawing on the work of Hans Holger Rath and Steve Pepper
(including their "Navigating Haystacks, Discovering Needles," Markup Languages,
1999), we hope to achieve a partial implementation of the topic map standard. We
are partially motivated by the hope that this will facilitate an overlay of more
provocative, abstract linkages superimposed on our current, largely objective
interpretive network. A second goal is to determine whether the presentation of
disparate early 19th century Russian history materials in the form of topic maps
will make their analysis more convenient and productive for the researcher than
their current presentation in separate, though proximate, data groups. But as
noted by Trippe, "the proposed standard itself is relatively new… and the
commercial technology supporting the standard is still in its early stages." Our
second goal is therefore our major concern: to explore how well this concept
works in a multi-type archive.