Information Technology Reference
In-Depth Information
Formation of the NI Reading Group
In 1990, two graduate students at the MIT Media Lab were trying to talk with
each other about topics that seemed of common interest. One, a humanist new
to computing (Marc), wanted to build programs that could automatically as-
semble short movies from archives of video data. The other, a computer sci-
entist with an interest in literary theory (Mike), wanted to program software
agents that could understand a simulated world, each other, and themselves.
What they found was that while their areas of interest seemed to have com-
mon issues (narrative theory and comprehension, knowledge representation,
story understanding and generation, and user interface design), the discourse
each used to talk about these areas was largely unintelligible to the other. Even
such core ideas as “representation,” “language,” and “communication” meant
different things to each of them. Rather than throw up their hands, they wa-
gered that if they could get each other to read the core texts in their respective
disciplines, they might be able to construct a common language and a useful
discourse. Together with a group of Media Lab students they formed the Nar-
rative Intelligence Reading Group (NI) that met weekly in the basement of the
Media Lab building for six years. From 1990-1993, Marc Davis, Mike Travers,
and Amy Bruckman actively led and facilitated the group, from 1994-1996,
Amy Bruckman and Warren Sack continued the group for another three years.
From 1997 to the present, the Narrative Intelligence Reading Group has func-
tioned as a mailing list and resource for its members and others interested in
its topics (ni@media.mit.edu).
Motivations for a new interdisciplinary discourse
As early graduate students in the Media Lab, we were faced with trying to syn-
thesize an intellectual framework in which we could situate our work. The
desire of the founders and early members of NI to create a common dis-
course and practice connecting artificial intelligence and literary theory also
stemmed from a growing frustration with the limits of our respective dis-
ciplines in their ability to inform the analysis, design, and construction of
computational media.
In artificial intelligence (AI) we encountered a discipline founded on logi-
cist, formalist, and objectivist conceptions of language, cognition, and com-
putation. Alternative approaches to AI included connectionism and situated
action, but both of these schools lacked a coherent theory of representation.
Furthermore, existing theories of representation in AI reflected a bias toward
