Cognitive Artifacts in Complex Work - Ambient Intelligence for Scientific Discovery - page 153

well as conventional computing) must successfully support the cognitive func-
tions physical artifacts already assist.
Cognitive artifacts embody what is relevant in distributed work settings be-
cause those who perform work create and use them as an essential part of their
work environment. The life scientists and acute health care clinicians we studied
routinely create and use physical cognitive artifacts as a way to manage uncer-
tainty and complexity. In these technical environments, artifacts reveal processes
that are supported more effectively by paper than computer. In acute health
care, physical artifacts including a hard copy printout and white marker board
serve team members more effectively than electronic displays. In life sciences
research, computer-based and traditional information artifacts do not support
the distributed cognitive processes described in scientific discovery. These two
studies of cohesive work groups and their manual cognitive artifacts therefore
reveal gaps between current practices and technology use.
We present the two cases as models for researching distributed artifact use,
and to show how cognitive artifacts embody implicit requirements for effective
computer-supported information systems. Current computer-supported informa-
tion systems can be inadequate to support distributed cognitive work because
the system developers do not grasp the work that is being performed. We argue
such artifacts must be analyzed in a distributed context, to avoid designing in-
formation technology for discrete tasks at the expense of cooperative work. The
acute care study shows how artifacts can be used as a means to eciently get
at the underlying cognitive work that clinicians perform. As scientific discovery
is also a joint cognitive process, we use this second case to argue for redesign of
information artifacts for distributed cognition in the life sciences laboratory.
Scientific discovery is a cognitive work process conducted in a highly uncer-
tain environment. In much of life sciences, while research projects are generally
planned, experiments follow the lead of findings, not plans. New experiments
are formulated on a daily basis to tease out and learn from phenomena. In ex-
perimental research, we find discovery described as a process; accounts of the
“eureka” event are in reality few, if not nonexistent in modern science. Cognitive
studies of discovery in practice are few, and have only recently emerged in the
literature [7, 8, 32, 19], suggesting much remains unexplored. Cognitive studies
of social reasoning and information use in scientific discovery have focused on
behaviors that facilitate the emergence and recognition of “meaningful novelty”
in research findings and patterns of scientific reasoning. Simon [32], shows the
general importance of discovery as a cognitive process due to its engagement
with ill-structured problems that explore the entire range of human cognitive
resources, with the capability to access deep insights into complex and creative
human thinking.
In terms of distributed cognition, Dunbar's research is noteworthy for its in-
sights into the social process of discovery in natural sciences, and the unexpected
use of distributed information in experimental reasoning. Dunbar [8] describes
the importance ascribed to unexpected findings in lab meetings. Measuring the
contribution of reasoning statements, Dunbar finds that reasoning is widely dis-