Paper presented at the session on "VIRTUAL COMMUNITIES",
sponsored by the Society for the Anthropology of Work and the Society of
Psychological Anthropology in the 91th American Anthropological Association
Congress, hold in San Francisco, Dec.2-6, 1992.

Arturo Serra, Ph.D. School of Computer Science 4615 Wean Hall Carnegie
Mellon University Pittsburgh. PA 15213 Tel.: (412) 268 6128 Fax: (412) 268
5016 E-mail:

By content analysis of interviews and written projects gathered in the
community, this study seeks to understand the kinds of cultural knowledge that
support a computer science culture and their differences with other kinds of
cultural knowledge. It also attempts to analyze the meanings of this culture in an
American high technology university. This study is based on two year fieldwork
at Carnegie Mellon University in 1990-1991 as part of a research project
between technologists at CMU and anthropologists from University of
Barcelona. Funding is by Centre Divulgador de la Informatica de la Generalitat,
a Catalonian public computer company.
The topic of this study is the analysis of a North American research university
called Carnegie Mellon as a " computer intensive campus ".
In February 1990 a team of three anthropologists from Barcelona University,
coordinated by professor Maria J. Buxo, arrived at this community in
Pittsburgh, at the invitation of Professor Angel Jordan, a university professor
of Electrical and Computer Engineering and, at that time, Provost of the
institution. We were interested in information technology, especially in
academic organizations. The building of CMU as a &quotcomputer intensive
campus" seemed very innovative to us.
CMU is actually a networked academic community through the ANDREW
system. In the 80s it was the first academic experiment of its kind in the
country. The Andrew system is a distributed computer network connecting each
college, department, and research team in the university. In 1990 there is about
one computer for each member of the university, faculty, students and staff. 90
per cent of faculty use computers to prepare documents, 68 per cent use
electronic mail and 76 per cent use online library information services.
We have tried to understand the so-called "CMU knowledge
revolution". This change has been developed under the influential work of
several CMU professors: among them, Allen Newell, Alan Perlis, Herbert
Simon, Dick Cyert, Raj Reddy, Nico Habermann, Mary Shaw, Angel Jordan.
The Simon's idea of a "sciences of artificial" or "sciences of
design", defined by along the last 20 years, is a good expression of this culture.
For this professor, one of the founding fathers of the Artificial Intelligence, the
design activity is a scientific activity and, the scientist, a designer. From a
European point of view, this statement seems extremely interesting. Usually
science and technology inhabit two different kinds of institutions in Europe, the
humanistic-scientific university and the polytechnic one. "Informatique"
lives mostly in the last one as a technological field.
I have just expent two and one half years doing fieldwork in this university in
three different places: the Engineering Design Research Center (EDRC), the
School of Computer Science and the Andrew network. This "virtual
community" is an INTERNET node, now with more than 5,000 bulletin boards,
many of them dedicated to electronic newsletters, courses, organizations, and
electronic debates, both national and international.
During this time, I made 105 interviews of professors, research scientists,
graduates students and staff from the CMU university community as a whole.
Progressively , I focused my investigation first in the research area of the
university, and then in the School of Computer Science, its main research
center. Finally, I have arrived at two apparently banal but useful conclusions:
First, that the keystone of a research university is their research projects. And
second, that each research project begins with a simple proposal written by a
research team. Then I have designed a methodology to deal with this problem. I
have called it "project analysis".
The basis of this methodology is to try to understand what the goals of a
research activity are, and to consider these goals as the value system of a
research community. The ethnographic model of this kind of community will be
based in its common research projects. We can consider "project analysis"
as a application of the "content analysis"to the technological
Thanks to the friendly collaboration of Dr. Jordan and other professors, and
thanks to the end of the Cold War too, I had access to the documentation of 30
years of defense sponsored research in this school, particularly its original
proposals. The result has been the study of 21 large research projects in the
four basic research areas in this School: "Artificial Intelligence",
"Programming Systems", "Computer Systems" and
"Theory", and the selection of 150 papers, technical reports,books, and
dissertations of professors, graduate students and researchers of the institution
referred to this topic.
After this search, we learned several interesting things:
First of all, Carnegie Mellon has built a kind of high technology university, or
computer university, based on a core research knowledge in computer science
and technology. This knowledge is extended to the rest of the campus through
education in computer technology skills and a daily practice of networked
research and education.
At a first look, this university seems similar to the traditional American
research university . This dominant model of university was analyzed by Talcott
Parsons and Gerald Platt in a book called "The American University"
published by Harvard University Press in 1973.
According to these authors, the "American university", or "full
university", is an institution centered on a faculty of Arts and Sciences, that
conceives of the research activity as a primary academic function. Research and
education is organized in departments comprised of professors and graduate
students. The Arts and Science faculty is organized in three classical categories:
humanities, natural sciences and social sciences, each of one divided in
well-recognized intellectual disciplines. Usually this kind of university in
America has absorbed the professional schools of law, medicine or engineering,
conceiving them as a kind of applied professional complement to the basic core
knowledge on arts and sciences.
But Carnegie Mellon has different characteristics.
In the first place, during the most part of its existence ,90 years, Carnegie
Mellon has been an institute of technology, not a "full university".
Only from 1967, was the institution born with the union of Carnegie Tech and
Mellon Institute renamed "university". That means that during the most
part of its existence, the arts and sciences have been a complement of technology
to improve the knowledge about the design of new technological systems. In
other words, the relation of science and technology is just the inverse situation
than in the dominant universities of the Ivy League.
Second, the leading faculty at Carnegie Mellon has been engineering, not the
faculty of arts and sciences. That engineering culture has introduced the
&quotproblem solving" mentality as a characteristic feature of this institution.
Nevertheless, after the World War II, an important change happened. The
computer field was organized at Carnegie Tech by mathematicians and social
scientists interested in the new machine, not by engineers. Because of that this
new field was called "Computer Science" in Carnegie Tech.
But, at the same time, this new "science" was very pragmatically oriented
from the beginning. This was one of the reasons why it has been funded for 30
years by a federal entrepreneurial agency, the Advanced Research Project
Agency, now DARPA. For decades this agency has supported a kind of
fundamental technological research in Artificial Intelligence, Programming,
Computer System and Theory. This field was called Computer Science, but in
fact this community, known as a "Artificial Intelligence ARPA
laboratory" has centered it research in the knowledge about the design of
technological systems, more than in its discovery as in the traditional sciences.
The contradictory term "scientist of design" expresses this paradoxical
As a result of that context the term "science" has a different meaning in
this community from that in the natural and social sciences.
"Computer Science" at CMU primarily means the creation of
knowledge about what kind of computer system the researcher can design and
how he can build it. As an example of it, we will quote the goals of the research
proposal in Artificial Intelligence at CMU called "Basic Research in
Computer Science: Integrated Architectures for Intelligent
Systems"(1990-1993): " The basic scientific results of this research
will be a technical understanding of what types of total system
organizations are capable of integrated intelligent behavior, as well
as an understanding of which aspects of the total system belong in
the architecture".(CMU-SCS-Basic Research in CS, 1989:6-1). In other
words, this scientific activity is similar to a technical understanding about new
capabilities of the new systems in construction.
As Allen Newell, one of the founding father of CMU Computer Science
community, said last year in a university conference at the SCS: " Science is
in the techniques... .If a domain cannot get beyond having just
discovering... that science is in fact in a preparadigmatic state. It is
in a very early stage. My idea is that discoveries in physics, in
chemistry, in biology all convert routinely into things you can do
later" ("Desires and Diversions", April 12 1991.) Consequently, Newell
spent the last years of his career designing SOAR, a new intelligent architecture.
Discoveries are considered, in the traditional science communities, the
highlights of the discipline. But in Computer Science, at least in CMU,
discoveries are only means to do something different: to increase the knowledge
about what kind of new computer systems are possible and how to design them.
That history began with the invention of the Logic Theory Machine, the first
Artificial Intelligence program in the 50s, and continues now with the design of
the Mach Operating System in the 80s. In fact, ANDREW was also a CMU
Computer Science project .
The general consensus in CMU defined Computer Science as " the study of the
phenomena surrounding computers". Some professors, such as Herbert Simon,
call it a "science of the artificial". For others it is an "experimental
science". But the problem is that in this so-called "science" the computer
scientist must figure out the new system before discovering its empirical
characteristics. He must be a designer before a scientist. In other words, in
computer science the empirical science comes after, not before, the design
In this sense, this cultural knowledge is a technological one in its nature, not a
scientific one. Knowledge is design more than discovery, in the
computer intensive campus.
In this kind of university, the computer design activity precedes the science in a
new kind of innovation cycle, driven by the technological activity.
Usually, the traditional innovation cycle defined by the R&D policy experts is
based in the so called Science &Technology system. This cycle begins with the
Science, as the "basic research", and the Technology is conceived as an
"application" of it.
This model was established by Vanevar Bush and adopted by the National
Science Foundation after the World World II. It has been useful for the period
where the physicists had the leadership in academic research. But the Cold War
is over now and in Computer Science this model does not fit very well.
In this computer culture, fundamental or basic design research has been
growing for decades independent of basic science. The innovation process
begins with this design activity and empirical science follows it. In other words,
the computer technology at CMU is not a mere application of the natural or
social sciences, but increasingly its own foundation.
At the beginning the computer was a simple machine built by mathematicians,
like Pascal or Babbage, as a tool to do calculations. But now mathematicians,
physicists, cognitive psychologists, linguistics, indeed the natural and social
science community are increasingly becoming designers, helping the computer
scientists in building the Universal Machine.
This new research model is redefining what knowledge means in an advanced
information society. In 1988 the Computer Science and Technology Board, a
section of the National Research Council, in a rapport called " The National
Challenge in Computer Science and Technology" said: " Since
computer science is an artificial science (Simon 1981) theoretical
computer science plays a very different role within computer
science than, say, theoretical physics plays within physics.
Theoretical physics seeks to understand the physical universe, which
exists independently. Theoretical computer scientists seek to
understand all possible architectures or algorithms, which computer
scientists create themselves."
This change in the cultural meaning of a key cultural knowledge of Western
civilization, scientific knowledge, could have enormous consequences in the next
future. We are changing from a natural scientific vision of the world, the world
as a &quotnatural order", to a technological one in which the world is
conceived as a man-made machine, as an artifact
This presents a great danger and a great challenge to anthropologists. The final
goal of the Computer Science and Technology community is to design the
Universal Machine. In this sense, the new "artificial world" can have the
appearance of a world of sophisticated machines served by human beings.
Usually, anthropology in computer fields is used to help the engineer in
designing a better system.
But we have seen in CMU that, before a machine exists, a human being plans
for it, designs it, projects for it. The computer as a technology comes certainly
before an empirical science of it, but the computer scientist as a human designer
comes before his computer. The anthropologists can show that machines are
human designs done by human designers. In this sense, the artificial world can
have a different meaning: it is not the machine, but the community that builds
that machine, the world of designers.
As for the applied mathematicians as Turing developed the Universal Machine
model, the applied anthropologists working in computer cultures could develop
a Turing new one we could call: Multicultural Virtual Community . Our goal
would be helping to build new kind of computer-based communities .
Until now the anthropology has adopted the empirical approach of the natural
sciences. Working in the computer cultures, we need now changing to a
computational approach trying to understand, as the theoretical computer
scientists do, not only the cultures that live independently but the possible
virtual communities which the computer anthropologist can promote by itself.
In this sense, the anthropologist working in the computer communities can
become a different kind of designer, a designer of cultural communities.
Arturo Serra.
Pittsburgh, November 30th, 1992.
Some references: