Relationship Mapping for Art Education and Research

Introduction
The catalogue raisonné, or reasoned catalogue, has long been
a standard tool for representing large art collections. A typical
catalogue raisonné includes images of artworks along with
descriptive metadata, commentary, and background information
(often a biography of the artist) about the collection. More
recently, technological and infrastructural advances (in
particular, cheaper secondary memory, increased network
bandwidth, computational power, and digitization technology)
have enabled development of the digital catalogue raisonné.
Some of these catalogues have been developed primarily to
support the construction of print-based catalogues (Lanzelotte,
et al., 1993; Gladney, et al. 1998), while others are intended
for online use (for example, The Vincent van Gogh Gallery
and Gemini G.E.L.). The Picasso Project (Mallen, 2006) has
developed a digital catalogue raisonné containing 11,000 of
Picasso's artworks, along with 7,000 biographical entries. It is
the most complete and up-to-date collection of Picasso's
prodigious body of work. Building on the premise that the logical structures of the book
do not support scholarly inquiry adequately (McGann, 1997),
we are using the Picasso digital catalogue to facilitate scholarly
work with art collections. Researchers, students, and teachers
in disciplines such as art history, painting, drawing, history of
art, and art appreciation deal with art collections. They analyze
and critique individual works and compare and contrast these
with other works. They identify similarities between pieces of
art and trace threads of influence between artworks, artists,
styles, materials, themes, and social, geo-political, or personal
events. These scholars interpret artworks, identify missing links,
and communicate their findings. In the context of the Picasso
collection, we support scholars in expressing and visualizing
the complex, multifarious relationships between artworks via
a Web-accessible software interface.
Approach
In a series of informal interviews with faculty members from
art education, history, Hispanic studies, art history, and with
local K-12 art teachers we found a diverse set of needs, interests,
and approaches to working with artworks in both education
and research settings. One key theme running through each of
these areas is the need to discover and present relationships
between artworks, although the specific relationships of interest
varied by discipline. The art history scholar wishes to
investigate relationships between artworks displayed together
in an exhibition or to study works composed when an artist was
with a particular lover. The historian wishes to view art in the
context of significant historical events, for example, artworks
created while Europe was anticipating World War I. K-12
teachers are interested in identifying artworks that provide good
examples of specific drawing or painting techniques, such as
the two-point perspective or the use of complementary color
schemes.
In addition to the interviews, we also attended sessions of two
college-level art history survey courses. We observed that
instructors typically showed one or two examples of artworks
from different artists or art movements, discussing each for a
few minutes. In subsequent interviews, the instructors explained
that lack of time constrains their ability to include additional
works. Creating thematic sub-collections based on the
relationships discussed in class could alleviate this problem,
enabling students to study additional examples of materials
covered in the classroom. These observations of classroom
interaction and feedback from educators and researchers have
informed our enhancements for supporting the representation
and visualization of diverse relationships over the Picasso
project’s artwork catalogue.
Picasso’s works cover a broad range of themes, topics, and
materials, thus presenting a rich substrate of artworks for
building a network of semantically diverse, meaningful
relationships. In addition to the image collection, the Picasso
project includes extensive metadata related to these works, such
as its place and date of creation, medium, dimensions, current
location, as well as exhibitions and books in which it has
appeared. We leverage much of this metadata to express
relationships based on ownership, materials, patronage, or
chronology.
Interactive Relationship Visualizer
The Interactive Relationship Visualizer (IRV), an interactive,
Web-based application, enables visualization of relationships
that exist between artworks in the archive. The IRV interface
displays image sub-collections connected by the relationship
of the viewer’s focus. In addition, it presents connections that
exist within artworks in the sub-collection as well as those with
others in the archive, enabling users to navigate the intricately
interconnected hypertextual web defined by these relationships.
While browsing, the the display changes to reflect the dominant
relationship being displayed. In order to express a rich set of
relationships, we are augmenting existing metadata to include
type (such as still life or portrait), art movement (cubism,
fauvism, surrealism), and content (woman, nude, vase, mirror).
The IRV distinguishes two broad categories of relationships,
"inferred" and "specified." Inferred relationships are those which
can be expressed in terms of the metadata elements provided
for each artwork. Some inferred relationships can be expressed
in terms of a single metadata value, such as "artworks created
with oil on canvas." Others require mapping a range of metadata
values onto a higher-level concept and require definitions
involving multiple metadata fields. For example, identifying
"paintings created in Paris around the time of World War II"
is a two-step process. The system must map the timeframe of
World War II to a portion of the traditional calendar and locate
paintings created during this time. It then selects from this set,
those that were created in Paris Finally, relationships such as
"expensive paintings" involve subjective, theory-driven, and
potentially variable definitions. A price that would be expensive
in one art market might be comparatively inexpensive in
another. Inferred relationships provide a powerful mechanism
for exploring, discovering, and expressing relationships between
artworks that leverages existing metadata. In contrast to relationships inferred directly from existing
metadata, other types of relationships must explicitly be stated.
We refer to these as "specified" relationships. For example,
Picasso sketched several rough drafts of large works,
interspersed with smaller works. Thus, a chronological view
of artwork around the time the Guernica was painted results in
a sub-collection that includes these preparatory works as well
as other, unrelated works. Hence, this "preparatory work"
relationship must be expressed between early sketches of the
Guernica and the final masterpiece. Another specified
relationship is images based on a shared subject, for example,
Picasso’s interpretive series of works of Diego Velásquez’s
Las Meninas. Specified relationships afford us the ability to
define and represent relationships between artworks that are
difficult to derive from the descriptive metadata associated with
each work. This category of relationships is critical for the
expression of concepts based in established and novel analytical
approaches to Picasso's work—allowing relationships based
on information beyond that which is encoded in the collection.
The drawback is that participation in specified relationships
must be manually encoded.
Figure 1 displays the IRV system design. The Web interface
employs specific visualizations for displaying different kinds
of relationships. For example, the display of Guernica and its
preparatory images uses a visualization that illustrates the
centrality of Guernica relative to the other images displayed.
In contrast, the display of all artworks in the Las Meninas series
uses a table-like view, since no image is clearly central to this
sub-collection. We employ artwork images from the Picasso
project, reinterpret and extend existing metadata to express
myriad connections between these artworks, and facilitate
visualization of these relationships to support art scholars from
various disciplines.
Future Work
We continue to add new metadata to enrich the
relationships expressed in our archive. While new
attributes enable us to express additional relationships, the
growing number of relationships gets increasingly difficult to
represent visually. We are investigating mechanisms to display
secondary visualizations without overwhelming the presentation
of the primary relationship views. As scholars analyze Picasso’s
works and life, the relationships of their interest are likely to
increase in complexity as well as variety. It is not possible to
express all imaginable relationships among these artworks a
priori, nor is it possible to have all the necessary metadata.
Enabling scholars to define useful metadata as well as
supporting them in forming new relationships will engage them
as partners in this project rather than as mere users.
The IRV has potential for exploration of artwork relationships
in the classroom as well as for evaluating student performance
via homework and papers. For example, students could explore
a relationship and write a short paper about the artworks it
encompasses. An instructor could create a relationship and ask
students to identify the relationship embodied by the included
artworks. Educators need assistance in the form of targeted
features for successful use of the IRV in the classroom setting.
We continue our dialog with instructors to channel the IRV’s
expressive power for enriching education.
Bibliography
Gladney, H. M., et al. "Digital Access to Antiquities."
Communications of the ACM 41.4 (1998): 49-57.
Lanzelotte, R. S. G. "The PROTINARI Project: Science and
Art Team up Together to Help Cultural Projects." Proceedings
of 2nd International Conference on Hypermedia and
Interactivity in Museums, Cambridge, UK (1993). .
Mallen, E., ed. The Picasso Project . Texas A&M University
. Accessed 2006-10-06. <http://picasso.tamu.edu/
>
McGann, J. "The Rational of Hypertext." Electronic Text:
Investigations in Method and Theory. Ed. K. Sutherland. New
York: Oxford UP, 1997. 19-46.
The Vincent van Gogh Gallery . . <http://vggallery.c
om/>
Washington DC: National Gallery of Art . Gemini G.E.L.:
Online Catalogue Raisonné. Texas A&M University. Accessed
2006-09-07. <http://www.nga.gov/gemini/>