Temporal Modelling

Temporal Modelling

Johanna
Drucker
Media Studies, University of Virginia
jrd8e@virginia.edu

Bethany
Nowviskie
SpecLab, University of Virginia
bethany@virginia.edu

2003

University of Georgia

Athens, Georgia

ACH/ALLC 2003

editor

Eric
Rochester

William
A.
Kretzschmar, Jr.

encoder

Sara
A.
Schmidt

Temporal Modelling is an interface designed specifically to meet the needs of
humanities scholars wanting to interpret or analyse the subjective experience of
temporality in historical documents or imaginative artifacts. Since its
inception, Temporal Modelling has had two goals: 1) to provide a responsive
interface for visualizing complex temporal relations in humanities data and 2)
to experiment with an alternative approach to content modelling in humanities
computing. Two years into the project, we think we can demonstrate progress on
both fronts. We’ve designed a working prototype suited to humanities scholars
that generates XML output concurrent with a user's graphical modelling. Our
session describes the interwoven conceptual and technical development of this
project (research, visual design development, conceptualization process, and
prototype production), and also aims to make it known to a community who we hope
will be among its primary users and first adopters.
Temporal Modelling demonstrates an alternative approach to content modelling for
humanities computing. In the usual sequence of humanities computing events, a
content model is developed and then used to create a richly marked data set.
Visualizations or graphical displays follow as a result, not as a point of
input. By contrast, we’ve created a tool that allows visualization to act as a
procedure, not an outcome, of interpretation. Graphical sketching integrates
interpretation into digitization concretely, creating a content model.
Temporal Modelling consists of two parts. The first we call a composition space
or PlaySpace: an interface with a set of visual, graphical elements for making
models of temporal relations. The mechanisms and processes of the composition
space focus on: the positioning of temporal objects (such as events, intervals,
and points in time) on the axis of a timeline; the labelling of those objects
using text, color, size, and quality; the relation of objects to specific
temporal granularities (the standards by which we mark hours, seasons, aeons);
and, in complex interaction, the relation of objects to each other. These can be
further modified by a process we call “inflection”—the assignment of attributes
with either semantic or syntactic values. For example, a humanities scholar
attempting to chart a sequence of events traced in family letters, in which many
temporal events are referenced relationally, rather than by date, could use such
a system in a preliminary stage of analysis, creating a visual scheme to
represent events or references. In addition, however, Temporal Modelling
contains tools of interpretation and analysis to characterize events through
such inflections as anticipation, regret, foreshadowing or causality.
A user’s interpretation is captured and formalized into a structured data scheme
that develops concurrently, as the visualization proceeds. User gestures and
image renderings are translated into an XML schema which can be exported, used
to design a DTD, or transformed through use of XSLT or other manipulations. In
the second part of our project, a complementary DisplaySpace will use the same
graphical elements to display “published” models from the PlaySpace, as well as
supporting display from imported structured data. DisplaySpace models can be
manipulated for contrast and comparison in their “published” form or reloaded
into the PlaySpace for further refinement. The composition space enables
understanding through iterative visual construction in an editing environment
that implies infinite visual breadth and depth. In contrast, the display space
channels energy into iterative visual reflection by providing a responsive,
richly-layered surface in which subjectivity and inflection in temporal
relations are not fashioned but may be reconfigured.
The objects, actions, and relations defined by our schemata and programming are
not married inextricably with specific graphics and on-screen animations or
display modes. Just as we have provided tools for captioning and coloring (and
the ability to regularize custom-made systems with legends and labels), we have
also made possible the upload and substitution of user-made standard vector
(SVG) graphics for the generic notation systems we've devised. This is more than
mere window-dressing. Our intense methodological emphasis on the importance of
visual understanding allows the substitution of a single set of graphics
(representing inflections for, say, mood or foreshadowing) to alter radically
the statements made possible by Temporal Modelling’s loose grammar. Users are
invited to intervene in the interpretive processes enabled by our tool almost at
its root level.
Temporal Modelling can be used to model date-stamped, or empirical data, but its
strength is in its ability to structure the representation of subjective
experience. Several of the design features of our project embody this
conviction, most notably the “now slider”. This element embeds an individual
point of view (or several, depending on the project) into the graphical model
and allows the interpretation to be played forward or backward (“progressed” or
“regressed”). The model of events changes depending on the position of these
individual now-sliders. An event once anticipated may give rise to an unforeseen
set of outcomes and be transformed into an interval charged with regret or
melancholy. Temporal Modelling attempts to embody the subjectivity of an
interpretation, not merely depict a subjective approach.
Our initial conceptualization was informed by reading across a range of
disciplines and fields. We were interested in a historical, trans-cultural
inventory of ways time and temporality have been conceptualized. We grounded our
fundamental distinction between time as an a priori category and temporality as
a relational conception. Our readings were drawn from logic, religion,
anthropology, and philosophy, as well other humanities and social sciences (see
References). In addition, we made an inventory of visualizations for showing and
analyzing data that have a temporal dimension (). Because we are intent on
creating both a composition space and a display space that can utilize the same
set of visual elements, we wanted to consider conventions for presenting
date-stamped information while concentrating our design on a system suited to
humanities documents whose temporal references and/or relations do not conform
to empirical models.
After this research survey, we created a basic conceptual schema for designing
the Temporal Modelling project. This involved several steps 1) defining temporal
primitives (entities, actions, and relations); 2) developing a graphical
vocabulary for their presentation; 3) developing a labelling and annotation
system that allows for customization by individual users. Our assertion is that
representations based on empirical approaches assume an objective, homogeneous,
continuous, and uni-directional notion of time. We wanted to design a system
grounded in subjective, heterogeneous, dis-continuous, and multi-directional
temporalities. To accomplish this, we conceived of a number of unique design
features in addition to the now slider. These include “stretchy” timelines with
variable scales and granularities, branching and alternative narratives of
temporal events according to catastrophic (event-driven) and continuous
(unfolding) models. Designs for ruptured and discontinuous time are also in the
plans. The technical implementation of this design involved creating a tightly
integrated relation between visualization and digitization of information. In
addition, at each instance, the design of specific elements raises interesting
intellectual problems at the intersection of philosophical and computational
concerns—for instance, is “regret” a semantic attribute of an event, or a
syntactic one, engaged in relational effects? Such issues are central to the
intellectual and programming structure of this project.
Temporal Modelling is an experiment in using speculative methods— approaches to
conceptualization and design of computational tools or projects with uncertain
outcomes. The Speculative Computing Lab at the University of Virginia defines
its mission as undertaking projects that have a risk of failure. These projects
are not necessarily grounded in discipline-specific research and have a
tool-based and interpretive aim rather than a collections or archive development
goal. Speculative projects are motivated by the desire to foreground these
interpretive practices, particularly the subjective practices that are central
to traditional humanities. Temporal Modelling is an Intel-sponsored project of
the Speculative Computing Lab in the Media Studies Program at the University of
Virginia.

REFERENCES

J.
F.
Allen

Time and Time Again: The Many Ways to Represent
Time

International Journal of Intelligent Systems

6
4
341-355
July 1991

J.
Burg
, et al
Using Constraint Logic Programming to Analyze the
Chronology in A Rose for Emily

Computers and the Humanities

34
4
377-392
December 2000

J.
T.
Fraser

Time, The Familiar Stranger

Massachusetts UP
1987

C.
S.
Jensen
et al
A Glossary of Temporal Database Concepts

Proceedings of ACM SIGMOD International Conference on
Management of Data 23, 1, March, 1994

1994

P.
W.
Jordan

Determining the Temporal Ordering of Events in
Discourse.

Unpublished masters thesis for

Carnegie Mellon Computational Linguistics
Program
1994

H.
Price

The View from Nowhen

Time’s Arrow and Archimedes’ Point.

New York
Oxford UP
1996

Teri
Reynolds

Spacetime and Imagetext

Germanic Review

73
2
161–74
Spring 1998

F.
A.
Schreiber

Is Time a Real Time? An Overview of Time Ontology in
Informatics

Real Time Computing

1992

M.
Steedman

The Productions of Time

draft tutorial notes 2.0: University of
Edinburgh