Cognitive Artifacts in Complex Work - Ambient Intelligence for Scientific Discovery - page 165

- Generating inquiries ( searches , questions) prompted by experimental find-
ings
- Seeking validation of specific experimental findings: Seeking clarifications,
alternative findings , prior literature , other valid approaches
- Finding relevant published articles , and scanning and reading for specific
questions (very little broad reading)
- Preparing article manuscripts -writing sections of articles , sharing findings
and making sense of data, reviewing current literature, internal review of
drafts
- Research conversation - Phone calls and emails with other scientists, current
and potential collaborators, maintaining social networks
The goal of the research project, and the end goal of these practices, can be
considered original scientific discovery, with an outcome of publication of find-
ings. But cognitive research shows discovery a highly distributed, social cognitive
process with no clear milestones of activity. Scientific discovery is a process that
depends on both individual and shared knowledge practices. It builds on tacit
knowledge, experience with research practice and literature, and the ability to
frame findings from experiments within a research context. Dunbar [8] describes
discovery as a communicative process, with no specific associated cognitive arti-
facts. Information artifacts (manuscripts, lab notebooks, research articles) con-
tribute to this process, but are not central. We examine cognitive artifacts to
better understand this relationship.
The triggering event for research discovery is the initial discovery of unex-
pected experimental findings. Dunbar [8] shows that over 70% of scientists' in-
ductions, 50% of their deductions, and 70% of their causal reasoning statements
were devoted to unexpected findings. Life scientists revealed the importance of
“new findings in the lab” as a triggering event for specific information tasks.
“I have weekly meetings with my supervisor to brainstorm or to discuss
my progress and what my research findings could mean.”
“Weekly meetings ...to discuss findings from articles that may back up
my research.”
Dunbar [8] also shows the importance of distributed cognition. In molecular
biology lab meetings, more than 50% of the reasoning process was distributed,
shared and generated among a number of scientists. This suggests that dis-
tributed reasoning (in collaborations of small research teams) significantly con-
tributes to the process of discovery. However, no studies are found that examine
information use and information artifacts in this process.
Analogies were frequently generated in lab discussions, to formulate hypothe-
ses, design experiments, or to explain results to other scientists. However, an
interesting cognitive operation was found that scientists tended to forget their
analogies and other “cognitive scaffolding” used in reasoning, and they recalled
and subsequently relied on the results of their reasoning. While analogies are
important in leading scientists to discovery, they were undervalued and even in-
visible. Further, the identification and appropriation of analogies in the process