Re-imag[en]ing Cervantes’ Don Quixote: a Multi-layered Approach to Editing Visual Materials in a Hypertextual Archive

It is often stated by critics that the Quixote is a theatrical,
graphic, and visual book. Thus, visual representations, like
theatrical performances, popular iconography and book
illustrations, have been recognized as significant contributions
to the understanding of Cervantes’ masterpiece.1 2 Nevertheless,
the thousands of woodcuts, engravings, etchings, drawings,
and lithographs that have accompanied the text are, for the most
part, a little known interpretative tradition, and a much neglected
critical and artistic treasure.3
Obstacles, such as the difficulty of accessing rare books, have
prevented the illustrative tradition from being appreciated by
scholars, students, and users in general. In 2001 the Cervantes
Project (CP) started the creation of a hypertextual archive to
include digital images of the illustrations taken from over 500
of the most significant editions to form the textual iconography
of the Quixote (as permitted by copyright limitations). Our
main objectives are to make the illustrations more accessible
and to establish their contribution to the reception and
interpretation of the text. At the present time, the archive has
digitized, and made available online more than 7,000 images,
supported by a fully searchable database and complemented
by rich metadata and innovative visualization tools.4 (Figure 1
shows the reader’s Web-based access to the collection.)
The multidisciplinary approach of our project enables scholars
to go beyond the literary aspect of Cervantes’ works. As an
invaluable pictorial depository, we emphasize supplying
information regarding the historic value and artistic significance
of the images. The hermeneutic and aesthetic values of each
individual image have been carefully examined by art historians
and the results incorporated in the archive as scholarly
commentary. Additionally, we include biographical commentary
about artists and engravers. These rich scholarly commentaries
will help to boost the study of book illustration art, which has
been to date secondary in Art History, in aspects such as the
evolution of techniques, from the first woodcuts (early 17th
century) to modern mechanical offset (20th century), and the
influence or achievement of an individual engraver, illustrator,
or lithographer. We associate the text and the images through
a taxonomy of episodes, adventures and narrative motifs for
both parts of the Quixote. The 308 taxonomic elements in which
we have divided the 126 chapters of the text encompass as
descriptive categories the totality of the tradition of illustrations
present in the editions in our collection. (See Figure 2 for an
example selection from the taxonomy.)
The reader’s interface offers access to all the levels of editorial
annotation about the artists, textual location, image information
and technique, aesthetic and textual commentary. In addition,
users can obtain, through links, specific information related to
the narrative, biographical information about the creators, and
the meaning of engraving techniques. We are in the process of
incorporating these and other categories into our current search
tool, as well as developing a controlled vocabulary about themes
and characters in the Quixote and about generic area content (flora, fauna, architecture, music, etc.) to facilitate the use and
exploration of the images in the archive by researchers other
than literary scholars. (Figure 3 shows the advanced search
interface.)
As a multidisciplinary project that involves library staff
digitizing the images, computer scientists coding the
applications, Hispanic scholars providing textual editing and
edition metadata, art historians examining the image metadata,
and finally the general public traversing the archive, it is critical
to have an efficient production line and a strategy for
consistently sharing information among the different parties.
After several attempts, we reached a work flow that we found
to work efficiently. The procedure starts with Hispanic scholars
providing bibliographic information and specifying the editions
to be digitized, the pages where the images reside, and the
naming scheme for the electronic files. This information is
recorded on a work sheet and sent to library personnel. Upon
receiving such a work sheet, the library staff responsible for
the digitization reviews the information, clarifies any possible
discrepancies, digitizes the illustrations, and saves them on a
library server as archival TIFF files. A copy of the images is
then transferred to the development site where a computer
scientist processes them into JPEGs, whose sizes are more
appropriate for Web traffic. Afterwards, the references to the
JPEGs are added to the database. At that point, the interfaces
automatically update and these images become available online.
The art historian in charge of image metadata has access to the
original illustrations and then remotely enters the information
using the online Web form. As soon as these commentaries are
finalized, they are published online alongside the images.
A Web-based interface is provided for maintenance of the
collection. The collection maintenance application supports
management functions such as inserting, deleting, or modifying,
project-wide data. The editor's Web form also offers a typology
of illustrations divided into 17 categories, i.e., frontispiece,
vignette, portrait, map, etc., and a 17 item index of engraving
techniques. Much of the design effort in preparing this interface
has focused on specification of the most appropriate handling
of default values and modes. Using an iterative development
methodology, we have focused on the art historian’s tasks,
refining the means for inheriting and modifying values between
entries with the goal of focusing his data-entry activities on
differences rather than similarities. The taxonomic and
typological categories are part of the rich image metadata
approach we have developed to capture and make accessible
the variety of image content present in the iconography of the
Quixote. They are of value both to collection editor and reader
as they are part of the editor's metadata template and of the
reader’s advanced search tool. (Figure 4 shows the tools used
by the editor in specifying metadata.)
The readers’ Web-based interface, introduced earlier, which is
separate from the collection maintainer’s interface, also has
undergone refinement during the project’s development. At
present, it provides a four-layered browsing design: a) an edition
index; b) a thumbnail overview of one entire edition; c) a
low-resolution image together with meta-data; and d) a
high-resolution image. One interesting characteristic in the
public interface is the provision of uncommon relationships
among the artifacts. One particular example is the
ancestor-descent relationship. For instance, an engraver is often
related to another as his master or disciple, or even father or
son; one illustration or a set of illustrations is inherited from
an earlier edition as simply reprints or sometimes after
re-engraving. In cases where ancestral and/or familial ties exist,
we provide a description of the ties, as well as physical links
pointing to the referred artifacts for the purpose of more
in-depth investigation. Navigation along the ties provides an
intuitive way to look closely at the propagation of engraving
skills from one person to another, the effect of the technology
evolution upon art works as reflected in re-engravings, and the
significance, elegance, and popularity of certain artists or
illustrations serving as the archetypes for imitation. At present,
the links are manually encoded on a one-by-one basis. We are
in the process of setting up a keyword-anchor list for automatic
linking.5 (Figure 5 shows the reader’s views of the collection
and metadata.)
The project’s metadata is imbedded into a Dublin Core
framework. We use a MySQL server as the database core to
store the metadata and the references to image files. The
readers’ and maintainers’ Web-based applications occupy a
display tier on top of the database tier.
The availability of the archive will contribute to a more
complete understanding and appreciation of Cervantes’ novel
by initiating new explorations from many perspectives: textual,
artistic, critical, bibliographical, and historical. In particular,
we provide resources and assistance to examine the reception
and evolution of the Quixote’s readings across time, culture,
audience, and milieu. Furthermore, the images can be grouped
according to several layers of content to respond to the users’
need for information selection of a specific critical focus, i.e,
art, geography, history, etc. This is achieved by 1) cataloging
each image using a comprehensive taxonomy of the episodes
and adventures, 2) including multiple levels of textual
annotation about each individual illustration, including
technical, historical, and artistic information, and 3) providing
descriptive and critical commentary related to the content and
textual context of the illustrations. The result is a multi-layered
and multi-directional collection of digital objects recombining
bibliographical, textual and visual materials
(edition-text-image-metadata), and the development of a rich
hypertextual archive encompassing a new form of critical apparatus in which the illustrations are newly re-contextualized
and re-imag[in]ed.
Figure 1: The front page of the collection (left) and the interface to the
iconography collection (right).
Figure 2: An example set of menus showing the selection of an entry in the
thematic taxonomy.
Figure 3: The reader’s advanced search interface.
Figure 4: Screen shots from the editor’s metadata specification toolkit. At top
left is the editor’s view of the collection, which resembles the reader’s. In
addition to allowing selection of editions, this screen allows editing of edition
metadata, lower left. Selecting an edition brings up the middle display, which
flags images still requiring metadata entry. The metadata is then provided
using the screen on the right.
Figure 5: The reader’s view of image magnifications and metadata. At top left
is the collection view, seen before. Selecting an edition from the collection
brings up its images, lower left. Next, selecting an image brings up more detail
about an image, including its metadata, center top. Here, the reader can request
detail about persons mentioned in the metadata (center bottom) and an
additional image magnification (right).
1. Funding and support provided by the National Endowment for the
Humanities, the Cátedra Cervantes at the University of Castilla-La
Mancha, Grupo Santander, Texas A&M University, and the
Cushing Memorial Library, Texas A&M University Libraries.
2. See for instance, Jean Canavaggio, Don Quichotte du livre au
mythe (Paris: Librairie Arthème Fayard, 2005) and José Manuel
Lucía Megías, “Don Quijote de la Mancha: del libro al mito,” En
torno al Quijote: Adaptaciones, imitaciones, imágenes y música
(Madrid: Biblioteca Histórica Municipal de Madrid, 2005) 27-55.
3. Much has been done in conjunction with the 400th centenary
celebration to remedy this situation; in addition to our own project
and digital archive at <http://cervantes.tamu.edu
/> (accessed November 12, 2006) see the Banco de imágenes del
Quijote directed by J.M. Lucía Megías at <http://www.qb
i2005.com/> (accessed November 12, 2006).
4. Fernando González Moreno et al, “La colección de Quijotes
ilustrados del Proyecto Cervantes: Catálogo de ediciones y archivo
digital de imágenes,” Cervantes: Bulletin of the Cervantes Society
of America 25.1 (2005) [2006]: 79-104 and E. Urbina et al, “Visual
Knowledge: Textual Iconography of the Quixote, a Hypertextual
Archive.” " Literary & Linguistic Computing" 21.2 (2006): 247-58. 5. Neal Audenaert et al, “A General Framework for Feature
Identification,” Digital Humanities 2006 Conference Abstracts,
Association for Digital Humanities Organization International
Conference, Université Paris-Sorbonne, (2006) 5-9, and E. Urbina
et al, “Textual Iconography of the Quixote: a Data Model for
Extending the Single-Faceted Pictorial Space into a Poly-Faceted
Semantic Web,” Digital Humanities 2006 Conference Abstracts
(2006) 215-20.